The Stock–Flow Consistent Approach

The Stock–Flow Consistent

Approach
Selected Writings of Wynne Godley

Edited by

Marc Lavoie
Professor, Department of Economics, University of Ottawa, Canada

and

Gennaro Zezza
Professor, Department of Economics, University of Cassino, Italy
© Selection and editorial matter Marc Lavoie and Gennaro Zezza 2012
Individual chapters © Wynne Godley 2012
All rights reserved. No reproduction, copy or transmission of this
publication may be made without written permission.
No portion of this publication may be reproduced, copied or transmitted
save with written permission or in accordance with the provisions of the
Copyright, Designs and Patents Act 1988, or under the terms of any licence
permitting limited copying issued by the Copyright Licensing Agency,
Saffron House, 6–10 Kirby Street, London EC1N 8TS.
Any person who does any unauthorized act in relation to this publication
may be liable to criminal prosecution and civil claims for damages.
The authors have asserted their rights to be identified as the authors of this
work in accordance with the Copyright, Designs and Patents Act 1988.
First published 2012 by
PALGRAVE MACMILLAN
Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited,
registered in England, company number 785998, of Houndmills, Basingstoke,
Hampshire RG21 6XS.
Palgrave Macmillan in the US is a division of St Martin’s Press LLC,
175 Fifth Avenue, New York, NY 10010.
Palgrave Macmillan is the global academic imprint of the above companies
and has companies and representatives throughout the world.
Palgrave® and Macmillan® are registered trademarks in the United States,
the United Kingdom, Europe and other countries.
ISBN 978-1-349-33275-5 ISBN 978-0-230-35384-8 (eBook)
DOI 10.1057/9780230353848
This book is printed on paper suitable for recycling and made from fully
managed and sustained forest sources. Logging, pulping and manufacturing
processes are expected to conform to the environmental regulations of the
country of origin.
A catalogue record for this book is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Godley, Wynne.
The stock-flow consistent approach : selected writings of Wynne
Godley / edited by Marc Lavoie and Gennaro Zezza.
p. cm.
Includes bibliographical references and index.

1. Macroeconomics. 2. Stock-flow analysis. 3. National income–

Accounting. I. Lavoie, M. (Marc) II. Zezza, Gennaro. III. Title.
HB172.5.G623 2011
339–dc23 2011030614
10 9 8 7 6 5 4 3 2 1
21 20 19 18 17 16 15 14 13 12
Contents

List of Tables vii

List of Figures viii

Acknowledgments xii
Notation xiv
Notes on Contributors xxi

Introduction 1
Marc Lavoie and Gennaro Zezza

Part I Early Views on the Stock–Flow Coherent

Approach
1 Inflation Accounting of Whole Economic Systems 21
K.J. Coutts, W.A.H. Godley and G.D. Gugdin
2 Time, Increasing Returns and Institutions in
Macroeconomics 39
Wynne Godley

Part II Stock–Flow Coherence and Economic

Theory
3 An Important Inconsistency at the Heart of the Standard
Macroeconomic Model 65
Wynne Godley and Anwar Shaikh

4 Weaving Cloth from Graziani’s Thread: Endogenous

Money in a Simple (but Complete) Keynesian Model 81
Wynne Godley
5 Macroeconomics without Equilibrium or Disequilibrium 90
Wynne Godley
6 Kaleckian Models of Growth in a Coherent Stock–Flow
Monetary Framework: A Kaldorian View 123
Marc Lavoie and Wynne Godley

v
vi Contents

Part III Stock–Flow Coherence and Economic Policy

7 A Simple Model of Three Economies with Two Currencies:
The Eurozone and the USA 159
Wynne Godley and Marc Lavoie
8 Maastricht and All That 189
Wynne Godley
9 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 194
Wynne Godley and Marc Lavoie
10 Seven Unsustainable Processes: Medium-Term Prospects
and Policies for the United States and the World 216
Wynne Godley
Wynne Godley – A Bibliography 255
Index 271
Tables

1.1 The balance sheet 24

3.1 The ex ante flow of real funds 70
3.2 Simulated price and real variable changes in the
face of an increase in money supply 73
4.1 Model transaction matrix and glossary 83
5.1 Flow of funds at current prices 93
5.2 Balance sheets 94
6.1 Balance sheets 125
6.2 Transactions matrix 126
7.1 Balance sheet matrix 162
7.2 Transactions-flow matrix 163
9.1 Transactions-flow matrix of a simple closed economy model 195
9.2 Steady-state values of variables for some parameter values 199
10.1 Implications of the six projections for the United States 238
10.2 Percentage shortfall of GDP in 2003 compared with base
projection 239
A.1 Estimated long-run coefficients using the ARDL approach 245
A.2 Error correction representation for the selected ARDL model 246
A.3 ARDL estimates selected based on the Schwarz Bayesian
Criterion 247
A.4 Dynamic forecasts for the level of px 248
A.5 Flow matrix describing flow variables of ‘stripped down’
model of U.S. Economy 250

vii
Figures

2.1 The growth-maximizing mark-up 49

2.2 How inflation resolves competing claims in national
income 54
2.3 A full steady-state of a real stock–flow model 57
5.1 Simulation 1: Effect on income and expenditure flows 103
5.2 Simulation 1: Effect on wealth and its components 104
5.3 Simulation 1: Effect on banks’ balance sheets 105
5.4 Simulation 1: Effect on interest rates 106
5.5 Simulation 1: Bonds and money as shares of wealth 107
5.6 Simulation 1: Effect on government liabilities 108
5.7 Simulation 2: A step in government expenditure 109
5.8 Simulation 2: Changes to components of the banks’
balance sheet 109
5.9 Simulation 2: Banks’ bonds and the rate of interest 110
5.10 Simulation 2: Components of wealth 110
5.11 Simulation 2: Allocation of interest bearing assets of
households 111
5.12 Simulation 3: Response of money to income shocks 112
5.13 Simulation 3: Response of inventories to sales shocks 112
5.14 Simulation 3: Response of money rates to banks’ bond
holdings 113
5.15 Simulation 3: Allocation of household wealth 113
5.16 Simulation 3: Response of government liabilities 114
5.17 Simulation 4: Interest rates 115
5.18 Simulation 4: Households’ portfolio responses to interest
rate changes 115
5.19 Simulation 4: Banks’ holdings of bonds and the money
rate of interest 116
5.20 Simulation 4: Additions to government liabilities and
their make-up 116
6.1 Higher propensity to consume, normal regime 140
6.2 Higher propensity to consume, effect on equities 141
6.3 Higher propensity to consume, effect on money to
wealth ratio 142
6.4 Higher propensity to consume, puzzling regime 142
6.5 Higher interest rate, normal regime 143
6.6 Higher interest rate, puzzling regime 144

viii
 Figures ix

6.7 Stronger preference for equities, effect on equity prices

and rate of return 145
6.8 Stronger preference for equities, effect on other variables 146
6.9 Higher wage share 148
6.10 Larger issues of equities 149
6.11 Higher retention ratio 149
7.1 Effect on the domestic product of each country of an
increase in the propensity of the ‘&’ country (Italy) to
import products from the ‘$’ country (USA) 174
7.2 Depreciation of the euro, measured in dollars, following
an increase in the propensity of one euro, country to
import products from the ‘$’ country 174
7.3 Effect on various balances of an increase in the
propensity of the ‘&’ country (Italy) to import products
from the ‘$’ country 175
7.4 Evolution of the assets and liabilities of the ECB
following an increase in the propensity of the ‘&’
country (Italy) to import products from the ‘$’ country 176
7.5 Relative evolution of the debt-to-GDP ratio of each
government, following an increase in the propensity of
the ‘&’ country (Italy) to import products from the ‘$’
country, in a world where pure government
expenditures grow at an exogenous rate 177
7.6 Effect of an increase in the propensity of the ‘&’ country
(Italy) to import products from the ‘$’ country, when the
‘&’ interest rate is left to be endogenous 180
7.7 Effect on the current account balances arising from an
increase in the propensity of the ‘&’ country (Italy) to
import products from the ‘$’ country, when the ‘&’
interest rate is endogenous 181
7.8 Effect on the domestic product of each country of an
increase in the propensity of the ‘&’ country (Italy) to
import products from the ‘$’ country, when government
expenditures of the ‘&’ country are assumed to be
endogenous 182
7.9 Effect on current account balances of an increase in the
propensity of the ‘&’ country (Italy) to import products
from the ‘$’ country, when government expenditures of
the ‘&’ country are assumed to be endogenous 182
7.10 Relative evolution of the debt-to-GDP ratio of each
government, following an increase in the propensity of
the ‘&’ country (Italy) to import products from the ‘$’
country, in a world with growth, where expenditures of
the ‘&’ government are endogenous 183
x Figures

9.1 Impact of an increase in the nominal interest rate, from

3% to 10%, on the ratio of real deficit to real GDP and
on the ratio of public debt to GDP, when the real growth
rate is still 2.5% 202
9.2 Evolution of the inflation rate, following a reduction in
the target rate of inflation, from 2% to 1.5% 207
9.3 Evolution of the ratio of actual output to potential
output, following a reduction in the target rate of
inflation, from 2% to 1.5% 207
9.4 Evolution of the growth rate of real pure government
expenditures, following a reduction in the target rate of
inflation, from 2% to 1.5% 208
9.5 Evolution of the growth rate of real output and of the
growth rate of pure real government expenditures,
following an increase in the propensity to consume out
of disposable income 209
9.6 Evolution of the inflation rate, following an increase in
the real rate of interest, from 1% to 7% 209
9.7 Evolution of the growth rate of output and of the growth
rate of real pure government expenditures, following an
increase in the real rate of interest, from 1% to 7% 210
9.8 Evolution of the ratio of public debt to GDP, following
an increase in the real rate of interest, from 1% to 7% 211
9.9 Evolution of the main balances, following the
appearance of a trade account deficit that stands forever
at 1% of GDP 213
10.1 General government structural balance 219
10.2 Adjusted fiscal ratio and GDP 220
10.3 Current balance of payments and balance of trade
manufactures 221
10.4 Adjusted trade ratio and GDP 222
10.5 Combined fiscal and trade ratio 223
10.6 Combined fiscal and trade ratio and GDP 224
10.7 The three major financial balances 225
10.8 Real private expenditure and disposable income 227
10.9 Analysis of private financial deficit 227
10.10 Private financial balance and net lending to private sector 228
10.11 Growth in real (deflated) stock of money (M3),
compared with a year earlier 229
10.12 The three major financial balances, actual 1970–1999Q1
and projections implied by CBO 230
 Figures xi

10.13 Private financial balance and growth of nonfinancial

debt, actual 1970–1999Q1 and projections implied
by CBO 232
10.14 Private debt, actual 1960–1999Q1 and projection
implied by CBO 233
10.15 U.S. net foreign assets, actual and projected 233
10.16 Breakdown of net foreign assets 235
10.17 ‘Interest’ rates on foreign assets and liabilities 235
10.18 The private financial balance on six different
assumptions 236
10.19 Private debt on six different assumptions 237
10.20 Goldilocks resuscitated 241
A.1 Selected assets and liabilities as a percentage of private
disposable income 243
Acknowledgments

The following copyrights and permissions are acknowledged gratefully

by the editors of the volume.
Chapter 1: Ken J. Coutts, Wynne Godley and Graham D. Gudgin
(1985) ‘Inflation Accounting of Whole Economic Systems.’ Studies in
Banking and Finance [Supplement to Journal of Banking and Finance,
Amsterdam: North Holland] 9(2) (June): 93–114. Copyrights belonging
to the authors. Permission granted by K.J. Coutts and G.D. Gudgin.
Chapter 2: Wynne Godley (1993) ‘Time, Increasing Returns and Insti-
tutions in Macroeconomics.’ In S. Biasco, A. Roncaglia and M. Salvati
(eds), Market and Institutions in Economic Development: Essays in Hon-
our of Paolo Sylos Labini (Basingstoke: Palgrave Macmillan), pp. 59–82.
Copyrights belonging to the author.
Chapter 3: Wynne Godley and Anwar Shaikh (2002) ‘An Important
Inconsistency at the Heart of the Standard Macroeconomic Mode.’ Jour-
nal of Post Keynesian Economics 24(3) (Spring): 423–443. Copyright ©2002
by M.E. Sharpe, Inc. Used by permission. All Rights Reserved. Not for
reproduction.
Chapter 4: Wynne Godley (2004) ‘Weaving Cloth from Graziani’s
Thread: Endogenous Money in a Simple (but Complete) Keynesian
Model.’ In R. Arena and N. Salvadori (eds), Money, Credit and the Role
of the State: Essays in Honour of Augusto Graziani (Aldershot: Ashgate), pp.
127–135. Reprinted by permission of the Publishers. Copyright © 2004.
Chapter 5: Wynne Godley (1997) ‘Macroeconomics without Equi-
librium or Disequilibrium.’ Working Paper No. 205, Annandale-on-
Hudson, NY: The Levy Economics Institute, August. Permission granted
by Dimitri Papadimitriou, President of the Levy Economics Institute.
Chapter 6: Marc Lavoie and Wynne Godley (2001–2002) ‘Kaleckian
Models of Growth in a Coherent Stock–Flow Monetary Framework: A
Kaldorian View.’ Journal of Post Keynesian Economics 24(2) (Fall): 277–312.
Copyright ©2002 by M.E. Sharpe, Inc. Used by permission. All Rights
Reserved. Not for reproduction.
Chapter 7: Wynne Godley and Marc Lavoie (2001) ‘A Simple Model
of Three Economies with Two Currencies: The Eurozone and the USA.’
Cambridge Journal of Economics 31(1) (January): 1–23. Used by permission
of Oxford University Press.

xii
 Acknowledgments xiii

Chapter 8: Wynne Godley (1992) ‘Maastricht and All That.’ London

Review of Books 14(19) (8 October): 3–4. Permission granted by the London
Review of Books.
Chapter 9: Wynne Godley and Marc Lavoie (2007) ‘Fiscal Policy in a
Stock–Flow Consistent (SFC) Model.’ Journal of Post Keynesian Economics
30(1) (Fall): 79–100. Copyright ©2007 by M.E. Sharpe, Inc. Used by
permission. All Rights Reserved. Not for reproduction.
Chapter 10: Wynne Godley (1999) Seven Unsustainable Processes:
Medium-Term Prospects and Policies for the United States and the World,
Strategic Analysis (Annandale-on-Hudson, NY: The Levy Economics
Institute) January. Permission granted by Dimitri Papadimitriou, Pres-
ident of the Levy Economics Institute.
Notation

We chose to use a common notation for all chapters which is as close

as possible to the one used in Godley and Lavoie (2007), Monetary
Economics.
As a general rule, lower-case variables correspond to variables mea-
sured as quantities, prices, interest rates etc., while uppercase variables
correspond to aggregates measured at current prices.
A star (∗ ) denotes a target value for a variable, while an e superscript
denotes an expected value.

Symbol Description

Bs Treasury bills supplied by government

Bd , Bhd Bills demanded by households
Bh , Bhh Bills held by households
Bb , Bbd Bills actually demanded by banks
B$, B&, B# Bills issued by the $, &, # countries, respectively
Bij Bills issued by the j-th country and held by the i-th
country’s households, with i, j = $, &, #
BECBi Bills held by the ECB, and issued by the i-th country,
with i = $, &, #
BFED$ Bills held by the central bank of the $ country, issued
by the $ country
C, Cd Consumption goods demand by households, in
nominal terms
c, cd Consumption goods demand by households, in real
terms
Cs Consumption goods supply by firms, in nominal terms
CAB Current account balance
CG Capital gains
CGe Expected capital gains of the current period
DEBT Sum of the debts of the private sector, government
sector and foreign sector
DEF Government deficit
DF Debt of the financial sector (Chapter 10)
DP Debt of the private sector (Chapter 10)
E Value of equities

xiv
 Notation xv

Symbol Description

es Number of stock equities supplied by firms

ed Number of firms equities demanded by households
EC Contributions for social insurance
FT Total profits of firms, inclusive of interest payments on
inventories
ft Total profits of firms in real terms, inclusive of interest
payments on inventories
FTnipa Profits, as measured by national accountants
F, Ff Realized entrepreneurial profits of production firms
F Sum of bank and firm profits
Fe Expected entrepreneurial profits of firms
FN Profits of firms, net of interest payments to banks
fn Profits of firms in real terms, net of interest payments to
banks
FfT Target entrepreneurial profits of production firms
FD, FDf Realized dividends of production firms
FU, FU f Realized retained earnings of production firms
FU T
f
Target retained earnings of production firms
Fb Realized profits of banks
FbT Target profits of banks
FDb Dividends of banks
FECB Profits of the ECB
Fi Profits of the central bank distributed to the government of
the i-th country, with i = $, #, &
G Pure government expenditures in nominal terms
g Pure government expenditures in real terms
gT Real total government expenditures (inflation accounted)
GG Government expenditures on goods
GS Government expenditures on services
GT Total government expenditures, inclusive of interest
payments on debt
GD Government debt (public debt)
gd Government debt in real terms
GL Gross lending to the private sector (Chapter 10)
gr Steady-state growth rate of the economy
grg growth rate of pure government expenditures in real terms
gr k Growth rate of net capital accumulation
H, Hs High-powered money, or cash money, supplied by the central
bank
xvi Notation

Symbol Description

H$, H&, H# Cash money in the $, &, # countries, respectively

Hid Cash money demanded by households in the i-th
country, with I = $; &; #
Hbd Reserves demanded by banks
Hhs Cash money supplied to households by the central
bank
Hbs Reserves supplied to banks by the central bank
Hd Cash money demanded by households
Hh , Hhh Cash money held by households
HC Historic costs
HCe Expected historic costs
HUC Historic unit cost
HUCe Expected historic unit cost
HWC Historic wage cost
i, id New fixed capital goods demanded by firms
(investment flow), in real terms
Id New fixed capital goods demanded by firms
(investment flow), in nominal terms
Is , I New fixed capital goods supplied by firms, in nominal
terms
ik Incremental investment/sales ratio (= i/s)
in Realized stock of inventories, in real terms
ine Short-run target level (expected level) of inventories, in
real terms
inT Long-run target level of inventories, in real terms
IN Realized stock of inventories, at current unit costs
IM Imports, in nominal terms
IM C Imports of computers
IM N Imports of other goods and services
IM O Imports of oil
im Real imports
INT h Interest payments received by households
INT f Interest payments paid by firms
INT g Interest payments paid by the government
int e Interest payments expected by households (Chapter 3)
int T Interest payments planned by firms (Chapter 3)
K Value of fixed capital stock, in nominal terms
k Fixed capital stock, in real terms (number of machines)
 Notation xvii

Symbol Description

kf Fixed capital stock, in real terms (number of machines)

of firms
kb Fixed capital stock, in real terms (number of machines)
of banks
L, Ld Loans demanded by firms to private banks
Ls Loans supplied by banks to firms
l Leverage ratio = L/K
Ms Money supplied by the government (Chapter 3) or the
banks
Md , Mhd Money deposits demanded by households
M, Mh , Mhh Money deposits actually held by households
M1 Checking account money deposits of households
mh Real money balances held by households
ML Mean lag
mc Marginal cost
mpk Marginal product of capital
mpl Marginal product of labour
N, Nd Demand for labour
Ns Supply of labour
NAFA Net acquisition of financial assets from the private
sector
NIT Net indirect taxes
NHUC Normal historic unit cost
NUC Normal unit costs
OTG Other net transfers from the government to the private
sector
p Price level
pb Price of bonds (Chapter 3)
ph Real price of a (second-hand) house
ps Price index of sales
pk Price of fixed capital goods
pe , pef Price of firms equities
pr Labour productivity, or trend labour productivity
PX Private expenditures (consumption, investment and
change in inventories)
px Private expenditures in real terms
q The valuation ratio of firms (Tobin’s q ratio)
Ra Random number modifying expectations
REP Repayments of private debt
xviii Notation

Symbol Description

r, rb Nominal rate of interest on bills

ri Rate of interest on bills in the i-th country, with
i = $, #, &
rr b Real rate of interest on bills
re Rate of return on equities
rf Rate of profit (= FT/K−1 )
rl Rate of interest on bank loans
rr l Real rate of interest on bank loans
rm Rate of interest on deposits
rr m Real rate of interest on term deposits
rr c Real rate of interest on bank loans, deflated by the
cost of inventories index
rk Dividend yield (or rate of cash flow = FU /K-1)
S Sales in nominal terms
SC Sales valued at cost
Se Expected sales in nominal terms
s Realized real sales (in widgets)
se Expected real sales
sf Ratio of retained profits to profits net of interest
payments (Chapter 6)
SAV h Household saving
T Taxes
TRpf Net transfers from the private to the foreign sector
TRgf Net transfers from the government to the foreign
sector
u Rate of capacity utilization
UB Unemployment benefits
UC Unit cost of production
V , Vh Wealth of households, in nominal terms
Vi Wealth of households in the i-th country, where
i = $, &, #
V iG Net worth of the government sector in the i-th
country, where i = $, &, #
VECB Net worth of the ECB
Ve Expected wealth of households, in nominal terms
VF Net financial wealth held abroad
Vnc Wealth of households, net of cash
Vnc e Expected wealth of households, net of cash
v Wealth of households in real terms
W Nominal wage rate
WB, WBd , WBs The wage bill
 Notation xix

Symbol Description

X Exports in nominal terms

XA Exports of agricultural goods
XC Exports of computers
XN Exports of other goods and services
x Real exports
xr$ Dollar exchange rate: value of one dollar expressed in
euro
xr¤ Euro exchange rate: value of one euro expressed in
dollars
Y National income, in nominal terms
Yd Aggregate demand
Ys Aggregate supply
Y fc Full-capacity output
Yf Household income
Y sa National income including stock appreciation
Y hr Regular income of households
Y hs Haig–Simons nominal income (including all capital
gains)
y Real output
yfc real full capacity output
Ypf Factor income paid abroad
Yfp Factor income received from abroad
YD Disposable income of households
YDe Expected disposable income
YDg Disposable income of the government
YDhs Haig–Simons nominal disposable income (including all
capital gains)
YDr Regular disposable income
YDer Expected regular disposable income
YDsa Private disposable income, including stock
appreciation
yd e Expected real disposable income
yd r Realized real regular disposable income
yd er Expected real regular disposable income
yd, yd hs Haig–Simons realized real disposable income
(including inflation losses)
yd hse Haig–Simons expected real disposable income
YP Nominal personal income
z Dichotomic variable or some numerical parameter
xx Notes on Contributors

Greek letters

(greek
Symbol letter) Description

α (alpha) Consumption parameters

α0 Autonomous consumption
α1 Propensity to consume out of regular income
α2 Propensity to consume out of past wealth
α3 Implicit target wealth to disposable income
ratio of households
β (beta) Reaction parameter related to expectations
γ (gamma) Partial adjustment function that applies to
inventories and fixed capital
ε (epsilon) Another reaction parameter related to
expectations
θ (theta) Personal income tax rate
λ (lambda) Reaction parameters in the portfolio choice of
households
μ (mu) Import propensity or parameter
ξ (xi) Reaction parameter tied to changes in interest
rates/share of investment that can be
financed externally (Chapter 2) or by
issuing equities (Chapter 6)
π (pi) Price inflation rate
π π  =p/p
πT Target inflation rate
πc Inflation rate of unit costs
σ (sigma) Various measures of inventories to output (or
sales) ratio
σs Realized (past period) inventories to sales
ratio
σse Expected (past period) inventories to sales
ratio
σN Normal (past period) inventories to sales ratio
σT Target (current) inventories to sales ratio
τ (tau) Sales tax rate
ϕ (phi) Costing margin in pricing
φ Share of entrepreneurial profits in sales
ψ (psi) Target retained earnings to lagged investment
ratio/ratio of retained profits (Chapter 2)
Contributors

Ken Coutts is Assistant Director of Research at the Faculty of Economics

in the University of Cambridge. His main research interests include open-
economy models of debt dynamics, visible and invisible earnings in
the balance of payments, macroeconomic policy, and industrial pric-
ing behaviour, particularly, within the context of open economies. Ken
Coutts was a long-time member of the Cambridge Economic Policy
Group and of Department of Applied Economics, where he produced
several works in collaboration with Wynne Godley, including Indus-
trial Pricing in the United Kingdom (1978). With Bob Rowthorn and Bill
Martin, he has recently written about the prospects for the UK balance
of payments.

Wynne Godley was Director of the Department of Applied Economics

at the University of Cambridge, UK, from 1970 to 1989, and a Professor
of Applied Economics from 1980 to 1993. Subsequently, he was Distin-
guished Scholar at the Levy Economics Institute of Bard College, New
York, USA, and then a Visiting Research Associate at the Cambridge
Endowment for Research in Finance. Besides his theoretical work, Godley
is mainly known for his numerous public interventions on public pol-
icy as long-time leader of the Cambridge Economic Policy Group, and
for having identified, in the late 1990s, the unsustainable processes that
eventually led to the Global Financial Crisis.
Graham Gudgin is Honorary Research Associate at the Centre for
Business Research, in the Judge Business School at the University of
Cambridge and Senior Economic Advisor with Oxford Economics. He
worked with Wynne Godley as a member of the Cambridge Economic
Policy Group from 1978 to 1985. He moved to Northern Ireland in 1985
from Cambridge as Director of the Northern Ireland Economic Research
Centre, a post which he held until 1998 when he was seconded as Spe-
cial Adviser to the Northern Ireland First Minister on economic policy.
He is the author of a large number of books, reports and journal arti-
cles on regional economic growth in the UK and on the growth of small
firms, and has also written widely on electoral systems and gerryman-
dering. He is currently working with Ken Coutts on an assessment of the
macro-economic impact of market liberalization policies in the UK over
the past 30 years.

xxi
xxii Notes on Contributors

Marc Lavoie is Professor in the Department of Economics at the Uni-

versity of Ottawa, where he started teaching in 1979. Besides having
published over 175 papers in refereed journals and book chapters, he
has written a number of books, among which are Foundations of Post-
Keynesian Economic Analysis (1992) and Introduction to Post-Keynesian
Economics (2006) (with French, Spanish, Japanese and Chinese ver-
sions). With Wynne Godley he co-authored three articles and one book
chapter, as well as Monetary Economics: An Integrated Approach to Money,
Income, Production and Wealth (2007). His work spans across many fields,
including monetary economics, macroeconomics, growth theory, pric-
ing theory, consumer theory, the economics of language and sports
economics.
Anwar Shaikh is Professor of Economics at the Graduate Faculty of Polit-
ical and Social Science of the New School University, Associate Editor of
the Cambridge Journal of Economics, and was a Senior Scholar and member
of the Macro Modeling Team at the Levy Economics Institute of Bard Col-
lege from 2000 to 2005. He is the author of three books, the most recent
being Globalization and the Myths of Free Trade (2007). Recent articles
include ‘The First Great Depression of the 21st Century’, Socialist Register
2011 (Fall 2010) and ‘Reflexivity, Path-Dependence and Disequilibrium
Dynamics’, Journal of Post Keynesian Economics (Fall 2010).
Gennaro Zezza is Associate Professor at the University of Cassino, Italy,
and Research Scholar at the Levy Economics Institute of Bard College. He
is a member of the Levy Institute’s Macro-Modeling Team and co-author
of its Strategic Analysis reports, which deals with medium-term projec-
tions of the U.S. and the world economy. From 1987 onwards, Zezza
worked with Godley in the United Kingdom, Denmark, Italy and the
United States on theoretical and empirical stock-flow-consistent mod-
els. His other research interests include economic growth, monetary
economics, econometrics, poverty and vulnerability. He has several pub-
lications in edited books and journals, including Metroeconomica and the
Journal of Post Keynesian Economics. He holds a degree in Economics from
the University of Naples.
Introduction
Marc Lavoie and Gennaro Zezza

Origins

In December 2009, when Gennaro Zezza was visiting him in the North-
ern Ireland home of his daughter Eve, Wynne Godley asked Zezza to help
compile a list of papers which at the time he felt were his most impor-
tant contributions. The volume, titled Collected Writings, contained eight
papers. The list was then sent to Marc Lavoie, who suggested to add
a couple of papers, the paper which is now Chapter 2 of this volume
and a paper written with Francis Cripps, ‘A Formal Analysis of the Cam-
bridge Policy Group Model’, published in Economica in 1976, as it was
thought that these two papers would help understand the evolution and
the continuity in Godley’s thought. These suggestions were accepted,
and Wynne Godley was to write a one-page introduction for each of the
chosen papers. The project was left aside for a while, until Wynne Godley
passed away on May 13, 2010.
In July 2010, after having met at the Levy Economics Institute during
the Minsky summer seminar, we decided to resurrect the project, giving it
a slightly different twist, by focusing on the stock–flow coherent method,
which had been at the heart of Wynne’s contribution over the years
and which seemed to attract the attention of several students present
at the summer seminar. The Collected Writings project thus became a
Selected Writings volume, where the selection of the papers was based on
their linkages with the stock–flow coherent method. We also decided to
homogenize the notation of the variables of the various articles, so as to
make it easier for readers to follow the arguments of the papers. As a con-
sequence, because we felt that two papers either did not fit well with the
rest of the selected papers or were being overly technically demanding,
we decided to drop out the 1976 Economica paper mentioned above as
well as one of the papers that Wynne had himself selected, the one that
he had written with William D. Nordhaus, ‘Pricing in the Trade Cycle’,
published in the Economic Journal in 1972, despite its original and prac-
tical analysis of cost-plus pricing. In the meantime, with all the news

1
2 Introduction

covering the eurozone crisis, in particular the Greek crisis, we stumbled

upon Wynne’s 1992 literary paper on the Maastricht treatise, which is
Chapter 8 of this volume. We thought the paper showed how prescient
Godley was of the troubles that the current European setup would lead
to, and thus we decided to add it to the other papers. We also opted to
add a 1997 working paper that demonstrated the appeal of the stock–
flow coherent framework but that had been left out in the cold. In total
we thus have ten papers in this book of selected writings, seven of which
were chosen by Wynne himself and three of which were added by us.
The selected papers have been regrouped into three parts. Part I con-
tains two papers that portray Godley’s early views on the stock–flow
consistent approach; Part II contains four papers that are at the theo-
retical core of the stock–flow consistent approach; and Part III deals with
papers that are at the juncture of economic policy and the stock–flow
coherent approach.

Contents

Chapter 1, ‘Inflation Accounting of Whole Economic Systems’, pub-

lished in 1985, is a paper which is nearly impossible to find, even on
the Internet, and most likely readers will be very grateful to find it here.
Indeed Godley kept referring to it when we were working with him,
but Marc Lavoie started to doubt its existence when the inter-library
loan system could not retrieve it and when Internet searches delivered
no reference to it. The paper was written for a conference on national
accounting in inflationary conditions, held in Dorga, near Bergamo,
Italy, in January 1984. The conference was sponsored by the Journal of
Banking and Finance, and the proceedings were published in the journal’s
supplement, Studies in Banking and Finance. The paper had been written
with Ken Coutts and Graham Gudgin, the latter having been recruited
by Godley to the Cambridge Economic Policy Group between 1978 and
1985, while the former had become a close collaborator in the mid-1970s,
being in particular a co-author of the book Industrial Pricing in the United
Kingdom, published in 1978, which was an update and extension of the
Godley and Nordhaus (1972) Economic Journal article.
Whereas the principle of stock–flow consistency and the idea that there
should never be any black hole in the accounting were already present
in the Macroeconomics book written with Francis Cripps in 1983, Wynne
Godley always considered that this 1985 paper was his first comprehen-
sive attempt at writing down consistent accounting, in a whole economic
system as the title of the paper says, and thus constituted the first step
 Introduction 3

towards proper stock–flow consistent accounting. The purpose of the

Bergamo paper was to set out the main principles of measuring stocks
and flows consistently both in nominal terms and at constant prices
(real flows and stocks), based on an ex post version of the Hicksian mea-
surement of real income, or what Godley and Lavoie (2007) called the
Haig–Simons measure of income. The paper was an extension of the
chapters of the Godley and Cripps book that had dealt with inflation
accounting, and it defined the conditions for inflation neutral effects on
income distribution and aggregate demand – which was of great concern
following several years of double digit inflation in the UK. The 1985 paper
offered an additional innovation, relative to the 1983 book: it includes a
stock matrix where all the financial assets must have a liability as an exact
counterparty and where only tangible assets appear in the net wealth of
a (closed) economy. As recalled to one of us by Ken Coutts, ‘Wynne felt
that the paper was a useful step, in deriving the main measurement con-
cepts in algebraic form, towards building the stock-flow macroeconomic
models that properly integrate the financial system with real demand
and output.’1 As such, and not only because it is nearly impossible to
get one’s hands on it, this paper clearly belongs to the selected writings.
Wynne Godley kept working on this topic in the 1980s, but with
the intention of writing, in collaboration with Ken Coutts, a complete
monograph that would integrate both the real and the financial sec-
tors. During his sabbatical as a visiting professor at the University of
Aalborg in Denmark, a stay about which Godley always talked very
fondly, a conference in the honour of John Hicks, ‘Fifty Years after the
IS-LM’, was held there in September 1987. Godley presented a paper
that was both a critique of IS-LM and an announcement of things to
come, ‘IS-LM and Real Stock-Flow Monetary Models: a Prelude to Applied
Macroeconomic Modelling’, written with Michael Anyadike-Danes and
Ken Coutts. This was in fact a prelude to Chapter 2, ‘Time, Increasing
Returns and Institutions in Macroeconomics’, a paper published in 1993
which was presented at a conference held in Rome in 1990, when foreign
scholars were invited to celebrate Paolo Sylos Labini – the famous Italian
economist.2 In this paper Godley announces that he is preparing a sub-
stantial monograph (in collaboration with Ken Coutts) and that there
exists ‘a simulation model in which banks’ operations are fully articu-
lated with income, expenditure and transfer flows together with asset
demand functions’. But only the main equations of the model are being
shown in the 1993 paper, with a steady-state solution, along with justi-
fications of the behavioural equations imposed on the various sectors –
households, corporations, banks and the government.
4 Introduction

That paper is key in understanding the transition towards the fully

developed models that will be constructed during Wynne Godley’s stay at
the Levy Institute, in particular the 1996 working paper that will be God-
ley’s first analysis of his simulation experiments with a fully integrated
model. This 1993 paper is the sketch of an alternative macroeconomic
theory, which at the time Godley was calling the real stock–flow monetary
model. Besides the usual concerns with basic accounting issues and infla-
tion accounting, the 1993 paper also illustrates the fascination exercised
by the work of both Alfred Eichner and Adrian Wood, in particular the
1975 book of the latter, A Theory of Profits. While both of these authors
accepted cost-plus pricing, which was at the heart of Godley’s work on
industrial pricing, they provided an explanation of the mark-up which
Godley was evidently looking for and quite ready to embrace: the mark-
up depended on the financing needs of growth. Finally, the paper also
illustrates Godley’s frustration with neoclassical theory, which he could
never associate with the real world that he had experimented when work-
ing at the Treasury. In particular, already at that time, Godley could not
accept the mainstream claim that fiscal policy is impotent, a topic which
is the subject of Part III.
Godley was also concerned and frustrated with the timeliness aspects
of neoclassical theory. We thus move to Part II of the book, with the four
chapters on economic theory and the stock–flow approach, introduc-
ing Chapter 3, ‘An Important Inconsistency at the Heart of the Standard
Macroeconomic Model’, written with Anwar Shaikh. This paper was pub-
lished in 2002, but was written as early as 1998 when it came out as a
working paper at the Levy Economics Institute. Anwar Shaikh, a profes-
sor at the New School University, in New York City, was a frequent visitor
and guest scholar at the Levy Institute, which is situated two hours north
of New York City, and so had several occasions to engage in discussions
with Godley, in particular about the prospects of the US economy but
also about the merits of the stock–flow coherent approach that Godley
was advocating. Shaikh – a Marxist – and Godley were both critical of
neoclassical theory, presumably for different reasons (!), but they shared
a common interest since both had underlined the weaknesses of empir-
ical works ‘demonstrating’ the validity of neoclassical theory. Shaikh
(1974) had shown that the neoclassical production function appeared to
‘work’ simply because it was reproducing national accounting relations,
while Godley (with Anyadike-Danes in 1989) had shown in a very sim-
ilar fashion that econometric regressions pertaining to demonstrate the
neoclassical view that higher wages led to reduced employment also ulti-
mately relied on national accounting identities for their good fit and
 Introduction 5

could still demonstrate this result even when it was assumed away by
construction.
In their 2002 paper, Godley and Shaikh push the standard mainstream
macroeconomic model, that of Patinkin, with an exogenous money sup-
ply and firms issuing private bonds, to its limits. The usual assumption is
to drop off the bond market from the analysis, by invoking Walras’s law.
But Godley and Shaikh show that things are not so simple, because it can-
not be taken for granted that the flow of interest payments arising from
past bond issues will equal the profits made by firms on their capital,
and hence all profits might not be distributed to households, in con-
trast to what is usually assumed. This makes for an inconsistent model,
which can easily be corrected, but as Godley and Shaikh demonstrate, if
the inconsistency is removed, then the famous neoclassical dichotomy
between real and nominal variables no longer holds, and an increase in
the exogenous stock of money could lead to a fall in prices.
As the authors point out themselves, they do not advocate a cor-
rected stock–flow coherent version of the neoclassical model. And so we
move to the positive theoretical contributions of Wynne Godley, based
on alternative models. In Chapter 4 we present ‘Weaving Cloth from
Graziani’s Thread: Endogenous Money in a Simple (But Complete) Key-
nesian Model’, which is the simplest stock–flow coherent model that can
be built while incorporating private banking money. The paper was pre-
sented at a conference held in Naples in 2003, on the occasion of the
70th birthday of Augusto Graziani, another major figure among Italian
economists, and was published in 2004. Graziani is considered the leader
of the Italian circuit school, that is, the monetary theory of production,
and he is the author who first pointed out that Joan Robinson had an
extensive discussion of monetary issues in her 1956 Accumulation of Cap-
ital, that anticipated what then became the post-Keynesian monetary
theory, usually attributed to Nicholas Kaldor. Godley and Graziani had
found that they held very similar views about how money entered the
economy, mainly through the credit granted to firms that needed to pay
for their labour costs and intermediate goods. For production to occur,
entrepreneurs need to borrow, and in the simplest models, outstanding
loans will be equal to the value of inventories, which themselves will
be equal to the money balances that households decide to hold. Both
Graziani and Godley very firmly believed that money was endogenous
at a time when this hypothesis was not fashionable at all. One of us had
earlier underlined the links between the monetary theories presented in
the Godley and Cripps (1983) book and the work of monetary circuitists
(Lavoie 1987). Indeed, as pointed out in the first note of Chapter 2,
6 Introduction

Godley had been a guest lecturer at the University of Naples, at the invi-
tation of Graziani at the end of the 1980s, and had given three lectures
on time, credit money and the neoclassical synthesis.
The 2003 conference in Naples brought some dissatisfaction to Godley
because he felt that he had been misunderstood and unfairly treated by
the discussant at the conference. However, he was quite proud to having
been able to produce, at long last, a short and simple treatment – in less
than ten pages – of what he considered to be his essential ideas about
macroeconomic theory at the time. The paper presents the transactions-
flow matrix, with its ex post budget constraints and its counterparties
to each transaction. The model contains less than 25 equations, taking
care of both nominal and constant-price variables, while entertaining the
presence of inventories and a consumption function based on some real
wealth to disposable income target ratio. A main feature of the model,
as Godley points out himself in the conclusion of Chapter 4, is that ‘it is
impossible for the supply of money to differ from the amount of money
which people want to hold, or find themselves holding, without either
the need or the place for any mechanism to bring this about’, meaning
that this equality held without the need for modifications in the interest
rate – a point also made, but only in a heuristic way, by members of the
monetary circuit theory.
Godley was happy to have produced a shortened version of his views
because over the 1996–1998 period he had written three papers that
incorporated similar ideas but in a much more complex form, with a rich
balance sheet of financial assets and by adding the public sector. As we
pointed out earlier, the 1996 Levy Institute working paper called ‘Money,
Finance and National Income Determination: An Integrated Approach’,
was Godley’s first attempt to present a complete fully integrated model,
with all its equations and with charts representing the impact of sim-
ulation experiments. For this reason it is fairly well-known by those
interested in the stock–flow coherent framework. In particular, it is this
paper that was brought to the attention of Marc Lavoie by Anwar Shaikh
and that led to Lavoie’s discovery of the stock–flow coherent approach.
The 1998 working paper is also well-known because it is the version that
eventually got published in 1999 as ‘Money and Credit in a Keynesian
Model of Income Determination’ in the Cambridge Journal of Economics,
and so it had a wide diffusion. But Godley also wrote another working
paper, nearly forgotten, which can be found here in Chapter 5, called
‘Macroeconomics without Equilibrium Or Disequilibrium’, which came
out in 1997. The models of these three papers have a similar but not an
identical structure. The 1996 paper has the most complicated balance
 Introduction 7

sheet: in addition to cash, demand deposits and time deposits, it has

equities and bonds. The 1999 published paper removes equities, while
the 1997 paper removes both equities and bonds, thus getting rid of the
difficulties associated with capital gains.
In our opinion, the 1997 paper of Chapter 5 has been unfairly
neglected. It is the first paper where one finds the simultaneous pres-
ence of both the balance sheet matrix and the transactions-flow matrix,
which will become the hallmark of the stock–flow coherent approach.
The paper has several graphs that explain the logic of the stock–flow
coherent approach as it applies to the initial periods of the transition,
when the model is moving from its initial stationary state towards its new
equilibrium. In this paper the focus of the analysis is clearly on the short
run, and on the interrelations that arise in an economy where loans,
inventories, portfolio holdings, the government debt and many other
variables are all taken into account. One of us had already taken note of
this paper in the past, and tried to induce Wynne to edit it and submit it
for publication, but for some reason he did not seem to see much interest
in this endeavour. The present Selected Writings is an excellent opportu-
nity to get the paper out of oblivion. One reason perhaps for which the
paper did not get the attention that it merits is the poor quality of the
charts. Thus we were forced to reconstruct the model to get high-quality
graphs, but luckily for us the parameter values could be found in the
appendix of the original paper. It needs to be pointed out however that
in a few cases we were unable to reproduce the original charts, and so
the current chapter is slightly different from the original paper.
The section on stock–flow coherence and economic theory ends with
the paper that marks the beginning of the collaboration between Marc
Lavoie and Wynne Godley. This is Chapter 6, ‘Kaleckian Models of
Growth in a Coherent Stock–Flow Monetary Framework: A Kaldorian
View’, which was published in 2001–2002. An extended version of
this article appeared as a Levy Institute working paper in June 2000.
As is recalled in the preface of Godley and Lavoie (2007), Godley and
Lavoie met for the first time in December 1999, when Godley made
a presentation at the University of Ottawa based on his 1999 Cam-
bridge Journal of Economics paper. At the time Lavoie was working on
improving Kaldor’s (1966) growth model where corporate firms issue
shares and keep retained earnings, by adding money balances and
introducing endogenous rates of capacity utilization, as in neo-Kaleckian
growth models, while keeping track of capital gains on the stock market.
Accounting for capital gains caused some difficulties and hence Lavoie
asked Godley if he could solve these accounting puzzles. These were
8 Introduction

child’s play for Godley, who then offered to build a discrete time ver-
sion of the model and provide simulations, thus leading to the paper of
Chapter 6.
We think it is fair to say that this Kaleckian–Kaldorian growth model
is now considered the archetypal post-Keynesian stock–flow coher-
ent model, incorporating the key features of the stock–flow coherent
approach with portfolio choice, as identified by Tobin (1982) in his
Nobel lecture on what macroeconomics ought to be, while entertaining
post-Keynesian behavioural equations and closure, such as an explicit
investment equation, taking into account both real and monetary fac-
tors, the presence of retained earnings, the principle of effective demand,
endogenous money, imperfect information, procedural rationality, cost-
plus pricing and liquidity preference. The paper has inspired many
contributions that have improved upon the model, by adding a pub-
lic sector (Zezza and Dos Santos 2004), or by extending the results of
the model to several different regimes, as in Taylor (2004: 272–278, 303–
305) or in van Treeck (2009). It should be pointed out that Wynne Godley
saw two drawbacks in the model of Chapter 6. First, he was somewhat
frustrated by the fact that its consumption function was only based on
flows of income and capital gains, leaving aside any component based
on the stock of wealth. In all of his other models, the consumption
function depends on disposable income and wealth, and hence implies
a stock–flow norm, which is not the case here. The second drawback,
as perceived by Godley, was that the model assumed away inventories,
and hence assumed a period long enough for producers to respond ade-
quately to demand. Again, this was in contrast to his other models, for
instance those found in Chapters 4 and 5. Both of these drawbacks were
remedied in the growth model found in Godley and Lavoie (2007).
It is a well-known fact that Wynne Godley was mostly interested in
public policy, and in particular in imbalances related to trade and open
economies. When working at the UK Treasury, he was in operational
charge of short-term economic forecasting, realizing quite early that
trade performance was an important constraint on growth. He had little
patience for sophisticated economic theory, besides some form of crude
Keynesianism. When he moved to the Department of Applied Economics
at the University of Cambridge in 1970, Godley and Francis Cripps, who
had previously worked with Kaldor, formed the Cambridge Economic
Policy Group (CEPG), along with Robert Neild, with the view of carrying
this policy-oriented work into academia. The work of the CEPG had at
least three distinct characteristics. The first was their use of the funda-
mental accounting identity, which says that the private accumulation of
 Introduction 9

financial assets by the domestic private sector has to equal the sum of
the public deficit plus the external current account surplus. This feature
was no other than the aggregate flow version of what became later the
stock–flow coherent framework. The second feature was their determi-
nation to contribute to public discussions of economic policy. The third
feature was their concern with open-economy considerations and their
(heretic) belief in managed trade in contrast to free trade. The third and
last section of the book is precisely devoted to those three characteristics:
stock–flow coherence, open economies and economic policy.
Chapter 7, ‘A Simple Model of Three Economies with Two Curren-
cies: the Eurozone and the USA’, written again with Marc Lavoie and
published in the Cambridge Journal of Economics in January 2007, is
a theoretical stock–flow coherent model but it has practical relevance
for eurozone economic policy and the eurozone institutional setup.
Although Wynne himself had chosen it as his only paper that dealt
with open-economy considerations, we wondered for a while whether
it would not be more appropriate to replace it with the two-country
working paper that had been written in 2003 for the Cambridge Endow-
ment for Research in Finance, as the latter paper was technically more
fancy, distinguishing between nominal and real variables, and intro-
ducing sophisticated import and export quantity and price equations.
As we thought more about it however, we came to realize the wisdom
of Wynne’s decision: despite dealing only with nominal variables, the
model of Chapter 7 is really useful to understand the weaknesses of the
current eurozone setup and how they could be remedied, and hence it
shows the usefulness of a complete stock–flow coherent framework.
The Chapter 7 model describes three economies: the USA is presumed
to be on a flexible exchange rate with the eurozone, and that eurozone
is made up of two countries, each with its own government, but with
the two countries sharing a central bank – the ECB. The paper then
examines what happens if one country of the eurozone – say Germany –
now benefits from a favourable shock on its exports to the USA. As one
would expect, the trade and current accounts of Germany will move
into a surplus position, as will its fiscal balance because of higher eco-
nomic activity. What is perhaps less obvious is that the trade and current
accounts of the other eurozone country, as well as its fiscal balance, will
all move into a deficit position, even though there has been no change
whatsoever in the parameter affecting the import and export equations
of this second eurozone country – say Spain. This is because the stronger
German exports will lead to a stronger euro currency and hence under-
mine the exporting capability of the rest of the eurozone. In a flexible
10 Introduction

exchange rate regime, one would expect the current account balances
to be brought to zero, as they do in a two-country model subjected to
the standard or near-standard conditions. The same result occurs here:
the eurozone as a whole sees its current account balance being brought
back to zero; however disequilibria within the eurozone remain: because
of the intra-eurozone fixed exchange rate, there is no mechanism bring-
ing back the various balances of the two individual countries to zero.
In the paper, Godley and Lavoie argue that, besides the impact on eco-
nomic activity and employment and hence on government balances, this
does not really matter and should have no impact on domestic interest
rates as no one in the financial markets should care about the current
account deficits of individual euro countries, as long as the ECB is ready
to take in more securities issued by Spain, the deficit country. But this is
precisely what a long series of European treaties, starting with the Maas-
tricht treaty, have forbidden. It took the Greek crisis, and impending
Irish, Portuguese, Spanish and Italian crises, to force the ECB to modify
its conventions and accept to purchase Greek and other sovereign debt
in May 2010 and thereafter.
Chapter 8, ‘Maastricht and All That’, published in the London Review
of Books in 1992, when Godley was approaching the end of his term
as head and member of the Department of Applied Economics at Cam-
bridge, continues on the same theme. Although the Maastricht treaty
has been criticized by many other economists, and while this criticism
may explain in part why the UK decided not to join the eurozone, this
paper offers a set of questions and answers about the Maastricht treaty
that are just as relevant today as they were 20 years ago. The weaknesses
of the proposed European setup that he outlines in this paper are pre-
cisely the weaknesses that we have been able to witness first hand since
the beginning of 2010. Godley identifies two fatal flaws in the eurozone
which was then proposed and which actually ended up being set up.
Both flaws are related to the implicit belief that modern economies are
self-adjusting. Europe thus only needs an independent unique central
bank, with a series of governments that balance their budgets.
Godley first points out that such a European Central Bank implies
that governments lose their national sovereignty: national governments
become mere regional or local authorities, like the State governments
in the USA or the provincial governments in Canada. They don’t have
the power to issue their own currency, get advances from or sell their
securities to the central bank, nor do they have the ability to devalue
their currency or set domestic interest rates. Indeed, various articles for-
bade explicitly the ECB from directly purchasing central government
 Introduction 11

securities or from indirectly doing so by pursuing outright open mar-

ket operations and purchasing central government securities. This flaw
in the European construction was later to be repeatedly underlined by
members of the so-called neo-chartalist branch of post-Keynesianism,
most notably by Stephanie Bell (2003), now Kelton, and by Kelton and
Wray (2009), before the Greek crisis erupted.3 The second flaw, related
to the first one, is the absence of a true federal European government,
with a fiscal capacity that would way exceed that of the current European
community, that would be able both to engage in strong counter-cyclical
reflation policies both at the overall level in case of a global crisis and
at the regional level when some countries or regions would suffer from
some structural setback.
Presumably, such a strong central European government was never
in the blueprints in part because in the 1990s and up to the global
financial crisis it was believed by a vast majority of economists that
monetary policy was enough to promote self-adjustments and that fis-
cal policy was destabilizing or at best useless. Wynne Godley strongly
opposed this view in all his contributions, and he did once more in his
paper (Chapter 9) ‘Fiscal Policy in a Stock–Flow Consistent (SFC) Model’,
which he published with Marc Lavoie in 2007. The paper was part of
a Journal of Post Keynesian Economics symposium organized by Louis-
Philippe Rochon, on the failings of monetary policy, just before the start
of the global financial crisis. All symposium papers argued that mone-
tary policy on its own, as conceived at the time, was insufficient, and
many argued that fiscal policy had an important role to play to achieve
full employment and price stability. The paper (Chapter 9) made the
additional point that if the fiscal stance was not set at an appropriate
level, but usually not the balanced-budget condition vowed by most
mainstream economists, full employment and low inflation could not
be achieved in a sustainable way. By contrast, appropriate fiscal policy
could achieve full employment and price stability even in a model based
on an assumed NAIRU. Another interesting conclusion of the paper is
that in an economy described within an appropriate stock–flow con-
sistent framework, debt dynamics remain sustainable even when real
rates of interest (net of taxes) exceed the real rate of growth of the
economy. Debt dynamics remain stable despite the assumption that
the government is setting its fiscal stance to achieve full employment
at all time; there is no need to generate unemployment by imposing
cutbacks and taking austerity measures. The dynamics of the model
will be such that, eventually, a primary budget surplus will be gener-
ated, thus insuring, as in the standard analysis, convergence towards a
12 Introduction

constant debt to GDP ratio, but without imposing limits on deficit to

GDP ratios.
Only a couple of years later did Lavoie discover that Godley had already
arrived at similar results in a paper devoted to ‘The dynamics of public
sector deficits and debt’ published with his Cambridge colleague Bob
Rowthorn in 1994. Godley had apparently forgotten all about it! Both
papers dismissed the mainstream method of ‘taking the economy as com-
posed from a single household that owns the government and the single
firm, and finds itself fully employed and sees no point in going into debt
against itself’.4 Bill Martin (2008), a former collaborator of Godley at the
CEPG and a former UK chief economist for UBS, later wrote a neat revised
analytical version of the model of Chapter 9, showing why the govern-
ment did not need to give itself an austerity rule to achieve a sustainable
debt path. There are essentially two reasons for this. First, in the Godley
and Lavoie model, as in all other Godley SFC models, there is an assumed
stock–flow norm that rules the private or household sector wealth tar-
get. Second, it is assumed that the government reduces its discretionary
expenditures in line with the increase in interest payments when full
employment is achieved. As long as the private sector saving behaviour
is stable, public budget solvency is automatically insured, provided the
wealth to disposable income target ratio is smaller than the inverse of
the real rate of interest on debt net of taxation. Another way to put it
is to say that the propensity to consume out of wealth must be greater
than the product of the propensity to save out of current income times
the real rate of interest net of taxes.
This book concludes with what is perhaps Wynne Godley’s best-know
piece among non-specialists, reproduced in Chapter 10, ‘Seven Unsus-
tainable Processes: Medium-Term Prospects and Policies for the United
States and the World’, which was published in January 1999 as a Strate-
gic analysis of the Levy Institute. This piece is considered by many to
be the announcement, justified by economic analysis, that the Clinton
boom years of the late 1990s were about to come to an end. Furthermore,
many of the unsustainable processes described in the 1999 paper were
still present in the 2000s, and so some observers also consider this 1999
paper to be a warning that the growth associated with the US housing
boom of 2001–2006 was to peter out as it did with the subprime financial
crisis. The seven unsustainable processes occurring in the United States
at the end of the 1990s were the following: (1) the fall in private saving
into ever deeper negative territory, (2) the rise in the flow of net lending
to the private sector, (3) the rise in the growth rate of the real money
stock, (4) the rise in asset prices at a rate that far exceeds the growth of
 Introduction 13

profits (or of GDP), (5) the rise in the budget surplus, (6) the rise in the
current account deficit, (7) the increase in the United States’s net foreign
indebtedness relative to GDP.
Godley had focused on processes (6) and (7) in a 1996 public pol-
icy brief, but here his outlook was even darker and obviously much
more general, underlining in particular the rising indebtedness of the pri-
vate sector, tying it with the huge surpluses that the government sector
was accumulating, making use once again of the fundamental identity
linking the private decumulation of financial assets to the government
surplus and the external current account deficit. The model he used,
which is still at the core of the Levy Institute macroeconomic model, was
based on the analysis of private expenditure rather than its individual
components (consumption, investment and the change in the stock of
inventories), and was therefore strongly connected to the New Cambridge
approach that Godley proposed, along with Francis Cripps and others,
in the 1970s. A key assumption of New Cambridge was that the change
in net financial assets of the private sector as a whole would be stable in
relation to GDP. This implied, among other things, that an expansion-
ary fiscal policy aimed at full employment would generate a deficit in the
current account, unless other policies – such as exchange rate manage-
ment – were put in place. Since estimated equations in macroeconomics
are usually grounded in some assumptions on behaviour, it was difficult
to justify an equation for private expenditure as a whole as that estimated
in Chapter 10, which was the result of independent decisions on con-
sumption by households and investment by firms. Godley had not made
this point very clear, but the analysis of his ‘debt identity’ in Chapter 10
gives some insights, since, in a stylized world, the sum of the debt of
the private, government and foreign sector is also equal to the assets of
the financial sector, and since financial assets are ultimately exchanged
for financial liabilities, total debt of non-financial sectors is also equal to
the overall liquidity (money) in the economy – which is the total debt of
the financial sector. The analysis of the change in debt positions is thus
mirrored in the evolution of money creation (or destruction), which is
crucial for sustainable growth. Deviations from stable stock–flow norms
for the private sector as a whole are therefore a signal of imbalances call-
ing for policy intervention, and if this not recognized, a crash will sooner
or later occur.
As the 2000 stock market crash became ever nearer, Godley focused
more and more on the unsustainable rise in household borrowing and
hence in households debt, a feature which would also be associated with
the housing boom that followed the stock market crash. While Godley
14 Introduction

was relatively silent on current affairs for many years after he arrived
at the Levy Institute in 1994, the ‘Seven Unsustainable Processes’ Paper
launched an uninterrupted series of frequent medium-term forecasts and
analyses, underlining a variety of explosive ratios, either with Godley on
his own or with various co-authors such as Randall Wray and Bill Martin,
but mainly with Alex Izurieta, Gennaro Zezza and Claudio Dos Santos.
Chapter 10 gives a good illustration of the method used and advocated
by Godley in his practical work.

Motivations

As we noted in the beginning, this book is a selection of Godley’s contribu-

tions, focusing on the stock–flow–consistent (SFC) methodology, and the
selection was guided by the interest expressed by many young researchers
in this line of research. The majority of chapters were chosen to show
how Godley contrasted his approach to the mainstream, while develop-
ing a coherent theoretical alternative, fully integrated with the work of
Tobin and the Post-Keynesian tradition. This may give the impression
that Godley’s main interest was to contribute to the theoretical debate
on macroeconomics.
On the contrary, Godley’s work always originated from some problems
or puzzles in the ‘real world’ economy, rather than from curiosities on
the intricacies of economic models. With his background at the Trea-
sury, working with primitive econometric models and in a world where
national accounts were still being developed, he had a very strong sense
of how economic variables should be evolving for growth to be sustain-
able, and the analysis of stock–flow ratios – such as foreign debt to GDP –
or flow–flow ratios – such as the financial balances to GDP ratios for the
main economic sectors – was of guidance to spot alarming trends which
called for policy intervention. The development of a model was always
targeted to address the problem under scrutiny, although Godley always
gave special care to starting from a complete and coherent set of accounts.
Godley was into economic theory, but with a practical purpose.
Since the publication of Monetary Economics (Godley and Lavoie 2007),
a growing number of young researchers is becoming interested in the
stock–flow–consistent (SFC) methodology, but – given the current crite-
ria for publications in scientific journals – there is a risk, in our view,
of developing ever more sophisticated and complex models, where the
researcher sees her goal as solving some technical problem, rather than
addressing some relevant issue for a ‘real’ economy, as it is the case of
several papers in the mainstream tradition.5 On the contrary, the SFC
 Introduction 15

methodology was always used by Godley, in his own words, as a ‘tool

for thinking’ on what was happening to the economy, and he never
made a claim that what he had developed was the ‘true’ model. As it
happens, this set of tools has proven more effective in predicting the last
two recessions than other models (Bezemer 2009 and 2011).
A second point we want to stress for future research is related to
the importance of accounting consistency. Although Godley’s method-
ology is now popularized as ‘stock–flow consistent’, the tightness of
the accounting was not proposed originally by Godley, but is required
by national accounting practices (and by logic!) and should therefore
be common – at least in an implicit form – to any macroeconomic
model.6 A consistent SFC model, grounded in national accounting prac-
tices, should provide a better description of the real world, as compared
to models based on (shaky) microeconomic foundations of rational
behaviour. Godley used to say that accounting tightness played the same
role for his models as micro foundations did for mainstream models: they
both provided scientific rigour. To achieve it, model accounting must
indeed be tight.7
If model accounting is set up correctly, the role of econometric estima-
tion of behavioural equations becomes less crucial than in mainstream
models. Godley had little interest in following the latest fashionable
tricks in econometrics. He had strong priors on what the values of model
parameters should be, and his intuition was partly due to a deep knowl-
edge of how real economies reacted to shocks, and partly due to the fact
that model parameters usually imply, or influence, values of stock–flow
ratios at steady state, and his choice of a ‘good’ econometric estimate
for a model was mainly guided by the plausibility of the implied overall
model simulation properties, rather than by fancy econometric testing.
Godley devoted a lot of efforts in producing a simple model which
could be used in undergraduate teaching. His first attempt (Godley
and Cripps 1983) was perhaps too different from standard textbooks,
and still too complex, and he devoted several papers and seminars to
attacking the IS-LM textbook model (see Chapter 3 in this volume, as
well as Godley, Anyadike-Danes and Coutts 1987) and providing simple
alternatives also based on diagrams, as in Chapter 2. He was satisfied
with the result achieved with Marc Lavoie in Monetary Economics, but
thought that more work should be done to popularize the SFC approach.8
Recently, a web site has been set up to encourage the diffusion of the
SFC approach.9 It is now becoming a community for sharing simulation
models over different software platforms, for solving models over the
Internet and so on.
16 Introduction

Another promising avenue of research puts together the SFC approach

at the macro level with agent-based modelling at the micro level (Kinsella
et al. 2011; Seppecher 2011).
Summing up, we hope that this collection of selected writings will
be useful not only for reconstructing the evolution of Godley’s analysis
over the years, but also as a guidance for developing further robust the-
oretical, and empirical, work: new tools for thinking on how to achieve
sustainable, full-employment growth.10

Notes
1. Letter to Marc Lavoie, dated 27 of August 2010.
2. There is now an active Paolo Sylos Labini (PSL) Association, which pub-
lishes among other things the PSL Quarterly Review, which replaces the Banca
Nazionale del Lavoro (BNL) Quarterly Review. The PSL Quarterly Review devoted
its 2009 volume to the reproduction of ten articles previously published in the
BNL Quarterly Review and that were deemed by its editor Alessandro Roncaglia
to enlighten the debate over the roots of the global financial crisis. Two arti-
cles, by Godley and also Godley and Izurieta, were thus chosen, as can be
found in the bibliography.
3. These theoretical links between Godley, Bell-Kelton and Wray are not surpris-
ing since their paths crossed each other while they were researchers at the
Levy Institute.
4. As pointed out in an email to Godley by Ekkehart Schlicht, dated 19 March
2007. Schlicht (2006) had independently argued along similar lines in an
earlier paper.
5. See the analysis in Klein and Romero (2007).
6. The System of National Accounts (SNA) (see European Commission and oth-
ers, 2009) sets the standards at the international level. For instance, Tables
2.13 and 2.14 (European Commission and others 2009, pp. 31–32) present
detailed integrated flows and stocks accounts for a whole economy, which
can be compared to Godley’s simplified versions in Chapter 5, Chapter 7 or
Chapter 10.
7. Anyone who has worked on an SFC model will be familiar with the exhausting
task of finding the wrong sign, or the missing variable, among the many
accounting identities of a simulation model.
8. Claudio Dos Santos has been active in this line of research. See Dos Santos
and Zezza (2008) and Macedo e Silva and Dos Santos (2011).
9. At http://sfc-models.net
10. Marc Lavoie has already explained in Godley and Lavoie (2007: xxxix–xliii)
how he met Wynne Godley and how they started working together. Here I
(Gennaro Zezza) recall my own experience. I met Wynne Godley in 1986. I
had just obtained my degree in Economics in Naples, with Augusto Graziani,
and I had obtained a small grant to complete a research on the relevance of
invoice currencies in trade. I decided to visit Cambridge and its library, and I
was admitted as a research student at the Department of Applied Economics
(D.A.E.). I thought it would be polite to introduce myself to the director of
 Introduction 17

the Department of Applied Economics, and so I went to meet him, without

suspecting that this meeting would change my life. I became a junior research
officer at the Department of Applied Economics in 1988, and both my kids
were born in Cambridge. Later, I worked with Godley in Denmark and the
United States.
On our very first meeting, Wynne asked me to write a paper on the Ital-
ian economy using a simple stock–flow consistent model (Godley and Zezza
1986), and this marked the beginning of an innumerable sequence of whole
days spent working together at his computer. This was quite shocking to a
young Italian scholar like me, who was used to meet his professors only for
very short meetings. I therefore had the privilege to watch Wynne’s method of
analysis from the inside: he started with an insightful observation, developed
a model, and then wrote a narrative to illustrate the finished product.

References

Bell, S.A. (2003) ‘Common Currency Lessons from Europe: Have Member States
Forsaken Their Economic Steering Wheels.’ In L.P. Rochon and M. Seccareccia
(eds), Dollarization: Lessons from Europe and the Americas (London: Routledge),
pp. 70–91.
Bezemer, D.J. (2009) ‘ “No One Saw This Coming”: Understanding Financial Crisis
Through Accounting Models.’ University Library of Munich, MPRA Paper n.
15767.
Bezemer, D.J. (2011) ‘The Credit Crisis and Recession as a Paradigm Test.’ Journal
of Economic Issues 45 (1) (March): 1–18.
Coutts, K.J., W. Godley and W.D. Nordhaus (1978) Industrial Pricing in the United
Kingdom (Cambridge: Cambridge University Press).
Dos Santos, C.H. and G. Zezza (2008) ‘A Simplified, “Benchmark”, Stock-Flow
Consistent Post-Keynesian Growth Model’, Metroeconomica 59 (3): 441–478.
European Commission, I.M.F., O.E.C.D., United Nations and World Bank (2009)
System of National Accounts 2008, New York, available at http://unstats.un.org/
unsd/nationalaccount/docs/SNA2008.pdf
Godley, W. (1996) ‘Money, Finance and National Income Determination: An Inte-
grated Approach.’ Working Paper No. 167, The Levy Economics Institute of Bard
College.
Godley, W. (1998) ‘Money and Credit in a Keynesian Model of Income Deter-
mination.’ Working Paper No. 242, The Levy Economics Institute of Bard
College.
Godley, W. (1999) ‘Money and Credit in a Keynesian Model of Income Determi-
nation.’ Cambridge Journal of Economics, 23 (4) (July): 393–411.
Godley, W., M.l Anyadike-Danes and K.J. Coutts (1987) ‘IS-LM and Real Stock
Flow Models: A Prelude to Applied Macroeconomic Modelling.’ Department of
Applied Economics working paper.
Godley, W. and M. Anyadike-Danes (1989) ‘Real Wages and Employment: a Scep-
tical View of Some Recent Empirical Work.’ Manchester School of Economic and
Social Studies 57 (2) (June): 172–187.
Godley, W. and F. Cripps (1976) ‘A Formal Analysis of the Cambridge Economic
Policy Group Model.’ Economica 43 (September): 335–348.
18 Introduction

Godley, W. and F. Cripps (1983) Macroeconomics (London: Fontana).

Godley, W. and M. Lavoie (2003) ‘Two-Country Stock-Flow-Consistent Macroeco-
nomics Using a Closed Model within a Dollar Exchange Regime.’ Working Paper
No. 10, Centre for Financial Analysis and Policy, University of Cambridge.
Godley, W. and M. Lavoie (2007) Monetary Economics: An Integrated Approach to
Credit, Money, Income, Production and Wealth (Basingstoke: Palgrave Macmillan).
Godley, W. and W.D. Nordhaus (1972) ‘Pricing in the Trade Cycle’, Economic
Journal 82 (September): 853–882.
Godley, W. and A. Shaikh (1998) ‘An Important Inconsistency at the Heart of the
Standard Macroeconomic Model.’ Working Paper No. 236, The Levy Economics
Institute of Bard College.
Godley, W. and G. Zezza (1986) ‘A Simple Real Stock Flow Monetary Model of
the Italian Economy.’ Working paper, University of Cambridge, Department of
Applied Economics.
Kaldor, N. (1966) ‘Marginal Productivity and the Macro-Economic Theories of
Distribution.’ Review of Economic Studies 33 (October): 309–319.
Kelton, S.A. and L.R. Wray (2009) ‘Can Euroland Survive?’ Public Policy Brief
No. 106, The Levy Economics Institute of Bard College.
Kinsella, S., M. Greiff and E.J. Nell (2011) ‘Income Distribution in a Stock-Flow-
Consistent Model with Education and Technological Change.’ Eastern Economic
Journal 37 (1): 134–149.
Daniel B. Klein and Pedro R. Romero (2007) ‘Model Building versus Theorizing:
the Paucity of Theory in the Journal of Economic Theory.’ Econ Journal Watch
4 (2) (May): 241–271.
Lavoie, M. (1987) ‘Monnaie et production: une synthèse de la théorie du circuit’,
Économies et Sociétés 21 (9) (September): 65–101.
Lavoie, M. and W. Godley (2000) ‘Kaleckian Models of Growth in a Coherent
Stock-Flow Monetary Framework: a Kaldorian View.’ Working Paper No. 242,
The Levy Economics Institute of Bard College.
Macedo e Silva, A.C. and C.H. Dos Santos (2011) ‘Peering Over the Edge of the
Short Period? the Keynesian Roots of Stock-Flow Consistent Macroeconomic
Models.’ Cambridge Journal of Economics 35 (1): 105–124.
Martin, B. (2008) ‘Fiscal Policy in a Stock-Flow Consistent Model: a Comment.’
Journal of Post Keynesian Economics 30 (3) (Summer): 649–668.
Schlicht, E. (2006) ‘Public Debt a Private Wealth: Some Equilibrium Considera-
tions.’ Metroeconomica 57 (4) (November): 494–520.
Seppecher, P. (2011) ‘Flexibilité des salaries et instabilité macroéconomique (un
modèle multi-agents avec monnaie endogène).’ Working paper.
Shaikh, A. (1974) ‘Laws of Production and Laws of Algebra: the Humbug
Production Function.’ Review of Economics and Statistics 56 (1) (February):
115–120.
Taylor, L. (2004) Reconstructing Macroeconomics; Structuralist Proposals and Critiques
of the Mainstream (Cambridge, MA: Harvard University Press).
Van Treeck, T. (2009) ‘A Synthetic, Stock-Flow Consistent Macroeconomic Model
of Financialisation.’ Cambridge Journal of Economics 33 (3) (May): 467–493.
Wood, A. (1975) A Theory of Profits (Cambridge: Cambridge University Press).
Zezza, G. and C.H. Dos Santos (2004) ‘The Role of Monetary Policy in
Post-Keynesian Stock-Flow Consistent Growth Models.’ In M. Lavoie and
M. Seccareccia (eds), Central Banking in the Modern World: Alternative Perspectives
(Cheltenham: Edward Elgar), pp. 183–208.
Part I
Early Views on the Stock–Flow
Coherent Approach
1
Inflation Accounting of Whole
Economic Systems
K.J. Coutts, W.A.H. Godley and G.D. Gugdin

List of symbols

Δ = means a first difference, and a subscript (–1) indicates a value

lagged one period N.P. Uppercase letters indicate nominal values
while lowercase letters indicate real values, except where
indicated below.
Bb = Bonds held by banks at issue price.
Bh = Bonds held by persons at issue price.
C = Consumers’ expenditure at market prices.
CG = Capital gains on equities.
E = Equities at current market prices.
FT = Gross profits of companies including stock appreciation.
FN = Profits net of interest payments on bank loans.
G = Government expenditure at market prices.
GD = Stock of government debt.
H = Stock of high-powered money.
IN = Stock of inventories at historic cost.
INT g = Interest payments on government debt.
K = Stock of fixed capital at current replacement values.
L = Stock of bank loans.
M = Stock of bank deposits.
p = Price index.
q = Ratio of the equity market valuation of capital goods to their
current replacement cost.
rb = Interest rate on government bonds
rm = Interest rate on bank deposits.
rg = Interest rate on government debt.

21
22 Early Views on the Stock–Flow Coherent Approach

rl = Interest rate on bank loans.

rr = Real interest rate
S = Total final expenditure at market prices.
V = Stock of household wealth at market values.
WB = Income from employment.
Ysa = National income including stock appreciation.
YD = Conventional definition of private disposable income
YDg = Government income net of transfer payments.
YDsa = Private disposable income including stock appreciation.
Y = Total output at market prices.

Introduction

Until recently macroeconomic theory, and also the compilation of

national income accounts, has used a framework where the logical con-
straints on the description of whole economic systems apply to flow
variables alone. Estimates of some stock variables (e.g. the stock of capi-
tal and the stock of money) have been made for a long time but national
balance sheets which interrelate all stock and flow variables consistently
are only now becoming available. And despite the increased impor-
tance attached in theoretical work to the way stock variables behave
and to sectoral budget constraints, discussion of these habitually pro-
ceeds without reference even to hypothetical sets of accounts where all
stocks and flows are consistently represented.1 These comments would
apply to the relatively simple business of accounting in terms of current
prices. A fortiori little attention has been paid to the logic and method-
ology of consistent accounting of whole economic systems at constant
prices.
In the inflation accounting of whole economic systems the accoun-
tants appear to be somewhat ahead of the theorists. The first estimates
(so far as we know) of consistent, inflation accounted sectoral income
flows at current prices for the UK were produced in 1979 by Taylor and
Threadgold, while in 1983 Hibbert produced his important stock and
flow accounts by sector at constant and current prices for a large number
of countries.
The purpose of this paper is to give an abstract description of a
measurement system where all stocks and flows are coherently related
to one another both at current and at constant prices. The system
described makes many simplifying assumptions, our aim simply being to
make crystal clear the main logical relationships between the principal
concepts.
 Inflation Accounting of Whole Economic Systems 23

The definition of income

Discussions relating to the measurement of income are haunted by Hicks’

famous2 definition of income as ‘the maximum amount of money which
the individual can spend this week and still expect to be able to spend
the same amount in real terms in each ensuing week’. Income so defined
can obviously never be precisely measured because it can never be ascer-
tained, particularly when measuring whole economies, what peoples’
expectations about their future real spending actually were or would
have been.
We shall therefore stick strictly to ex post concepts in accordance with
all measurement practice hitherto, recognizing that these will need fur-
ther interpretation when they come to be used. Specifically, we shall
define sectoral nominal disposable income such that it equals nominal
expenditure plus the nominal change in financial wealth; sectoral real
income is defined as real expenditure plus the change in real financial
wealth.

Consistent accounting of nominal stocks and flows

The following simplifying assumptions will be made.

(i) The economy is closed.

(ii) Companies do not save, do not hold financial assets and therefore
pay out post-tax profits as interest and dividends, implying that they
borrow to pay for all capital expenditure.
(iii) Fixed capital is assumed to be indestructible, enabling us to ignore
depreciation.
(iv) In order that interest receipts on banks’ assets exactly match interest
payments on their liabilities we assume that commercial banks have
no net worth, incur no expenses and make no profits.
(v) Financial liabilities other than equities include only bank loans and
government debt. There are no notes or coins and government lia-
bilities consist only of banks’ reserve assets and ‘bills’ which do not
change their nominal value when interest rates change. This lat-
ter assumption contrasts sharply with theorists’ usual assumption
that government liabilities other than high-powered money take the
form of perpetuities, the capital value of which changes inversely
and fully in proportion to changes in nominal interest rates. Our
assumption that debt is ‘capital certain’ is much nearer reality than
the assumption that it consists entirely of consols, at least for the
UK, and departures from it are discussed at the end of the paper.
24 Early Views on the Stock–Flow Coherent Approach

Table 1.1 The balance sheet

Persons Banks Companies Government Total

Equities E(= p·k·q) −E 0

Deposit money M −M 0
Bonds Bh Bb −Bh − Bb 0
High powered money H −H 0
Bank loans L −L 0
Inventories IN IN
Fixed capital K K

Total V 0 IN + K−E−L −GD IN + K

(vi) All commodity prices move together and only on the first day of
each accounting period.

Initially we make two further assumptions:

(vii) All taxes are indirect and ad valorem.

(viii) Households neither invest nor borrow.

Balance sheet consistency

As our definition of income hinges crucially on the maintenance
of wealth, it seems appropriate to start by considering the struc-
ture of national balance sheets. A logically complete set of balance
sheets, though subject to our simplifying assumptions, is represented
in Table 1.1.
Table 1.1 shows the correspondence between financial assets held by
the personal sector and the liabilities of government and of companies
(other than banks). The intermediary role of banks is also shown, as is
the correspondence between the tangible assets of companies and both
the financial liabilities of companies and the assets of banks. Equities are
assumed to be valued at market prices, and fixed capital at replacement
cost. Since government liabilities are assumed to be capital certain, gov-
ernment debt exactly equals the sum of bonds and high-powered money
(the reserve assets of banks).
Table 1.1 gives two definitions of total financial wealth (V ) at current
market prices. Summing vertically,

V ≡ M + Bh + p · k · q (1.1)

where M is the total of bank deposits, Bh is bonds held by households, k

is the stock of capital at constant prices, p is the commodity price index
 Inflation Accounting of Whole Economic Systems 25

and q is the ratio of the asset market valuation to the replacement cost
of capital.
Summing horizontally,

V ≡ GD + L + p · k · q (1.2)

where the net liabilities of the government (GD) are equal to high-
powered money (H) and bonds held by the banks (Bb ) and by house-
holds (Bh ):

GD ≡ H + Bh + Bb (1.3)

Flow identities
Coming next to flow identities, still in current prices, the point we wish
to stress is that income and expenditure should be defined in ways which
are consistent with balance sheets at market prices. This means that
capital gains on equities must be included within income, if the flow
change in the value of stocks is to be consistent with the balance sheet
changes in stocks at current market prices. We start with the division
of the post-tax national income between profits (F T ) and income from
employment (WB):

Y − T ≡ FT + WB (1.4)

Balance sheet consistency requires that profits be defined as final sales

(S) less purchases (in this case entirely income from employment) plus
the change in the value of inventories valued at cost.

FT ≡ S − T − WB + Δin · p + in–1 · Δp (1.5)

So long as the production period is shorter than the accounting period,

the whole stock of inventories must be turned over in each period. Firms’
inventory investment – the amount they will actually have paid out to
acquire their end period stock – will therefore be equal to the change in
the volume of inventories valued at period t prices (p · Δin), plus the cost
of replacing the whole opening stock at period t prices, (in–1 · Δp).
It is very common in theoretical work to leave inventories out of the
story and to define profits simply as S − WB. This omission makes the
flow system inconsistent with the stock system. It also leads to results
which are logically inconsistent with the present assumptions that all
investment is financed by borrowing and that all profits are distributed.
If fixed investment and the change in the value of inventories is entirely
financed by the issue of equity and by bank borrowing, firms’ excess of
26 Early Views on the Stock–Flow Coherent Approach

receipts of funds over outlays includes both S − WB and receipts from

new borrowing plus new equity issues. For consistency, profits must be as
defined in Equation (1.5) i.e., including stock appreciation. By assumption
these must necessarily all have been distributed as dividends and interest.
Next writing out final sales at market prices in full:

S ≡ G + C + p · Δk (1.6)

we reach a definition of the national income at current market prices,

measured from expenditures:

Ysa ≡ G + C + p · Δk + p · Δin + in–1 · Δp (1.7)

While Equation (1.7) is a consistent way of representing total factor

income, there is another crucial aggregate flow concept.
Equation (1.7) clearly does not represent the flow of production (Y) at
current market prices. To obtain this, stock appreciation i–1 · Δp must
indeed be deducted:

Y ≡ G + C + p · Δk + p · Δin (1.8)

Sectoral income at current prices

The definition of the government’s disposable income (YG) at cur-
rent prices is simply tax receipts less interest payments, since this is
equal to government expenditure plus the change in the government’s
indebtedness.

YDg ≡ T − rg · GD–1 ≡ G − ΔGD (1.9)

The conventional definition3 of private disposable income (YD) is total

factor income plus interest received from the government:

YD ≡ Ysa − T + rg · GD–1 ≡ C + ΔM + ΔBh + p · Δk (1.10)

However, this definition of private disposable income does not equal

consumption plus the change in household wealth, because the change
in the value of equities in any period does not equal fixed investment at
current prices. To obtain a definition of disposable income which makes
savings equal to the change in wealth it is necessary to add in the change
in the market value of the opening stock of equity k–1 · (p · q − p–1 · q–1 )
as well as any difference between the cost of investment in the period
and the equity market valuation of it p · Δk · (q − 1).
Defining nominal capital gains as the sum of these two terms:

CG ≡ k–1 · (Δp · q–1 + Δq · p) + p · Δk + p · Δk · (q − 1) (1.11)

 Inflation Accounting of Whole Economic Systems 27

we have a definition of nominal personal disposable income which

is fully consistent with the flow of consumption and the change in the
stock of wealth at current market prices:

YDsa ≡ Ysa + CG + rg · GD–1 − T ≡ C + ΔV (1.12)

Note that we have dropped the prime symbol from YDsa to denote
that this is the concept of disposable income which is consistent with
changes in balance sheets.
As a check we can add Equations (1.9) to (1.12):

YDsa + Yg ≡ Ysa + CG ≡ C + G + [ΔV − ΔGD] (1.13)

The term ΔV – ΔGD is indeed equal to capital investment (p·Δk) plus

the change in the value of inventories plus capital gains as defined in
Equation (1.11).

An alternative (though equivalent) definition of personal disposable income

We already know from Equation (1.4) that total factor income equals
income from employment plus profits.

Ysa − T ≡ WB + FT (1.4)

Gross profits (FT ) may be subdivided into interest paid on bank loans
(rl · L–1 ) and net profits (FN ):

FT ≡ FN + rl · L–1 (1.14)

The interest paid by the government may be also subdivided:

rg · GD ≡ rh · H + rb · Bb + rb · Bh (1.15)

Banks are assumed to pay out all the interest they earn:

rm · M ≡ r h · H + r b · B b + r l · L (1.16)

Combining Equations (1.4), (1.14), (1.15) and (1.16):

Ysa − T ≡ WB + rm · M–1 + rb · Bh−1 + FN [−rg · GD–1 ] (1.17)

Using Equation (1.17) we can now define personal disposable income,

making explicit its various sources:

YDsa ≡ WB + rm · M–1 + rb · Bh−1 + FN + CG ≡ C + ΔV (1.18)

In words, personal disposable income equals income from employ-

ment plus property income stemming from the three kinds of financial
asset held directly by the personal sector – money, bonds and equity –
plus capital gains on equities.
28 Early Views on the Stock–Flow Coherent Approach

Consistent accounting in real terms

So far we have outlined an accounting system which is consistent in
the sense that the difference between income and current expenditure is
equal to the change in stocks as defined in the balance sheets. In order
to investigate the effect of inflation on real stocks and flows it is first
necessary to translate the nominal system into real terms, preserving
the same budget constraint that changes in real stocks should equal real
income less real current expenditure. To achieve this, real income must
be defined to include not only capital gains but also the gains or losses
on financial assets and liabilities due to inflation.

Real income
How should real income be defined? The usual way, even now, is simply
to divide money disposable income by a price index. One habitually sees
the real disposable income of the personal sector written:
Y + INTg − T ≡ p · (y + intg − t) (1.19)
(where INT g is nominal interest payment by the government), or that
of the government sector:
T − INTg ≡ p · (t − intg ) (1.20)
However, this is inadequate because it takes no account of the losses due
to inflation on financial assets or the gains due to inflation on financial
liabilities. The result is that the difference between real income and real
expenditure is not equal to the change in real wealth.
The easiest route to seeing this is, perhaps, by demonstrating that the
change in real wealth Δv is not given by dividing p into the change in
nominal wealth.
Δν = ΔV /p (1.21)
The change in real wealth is given by:
Δν = ΔV /p − V–1 / − p–1 = V /p − V–1 /p − V–1 (1/p–1 − 1/p) (1.22)
= V /p − V–1 (p–1 /p)/p–1 (1.23)
or
Δν = ΔV /p − ν–1 · Δp/p (1.24)
We now can use Equation (1.25) with Equation (1.12) to produce a
consistent definition of real personal disposable income:
ydsa ≡ ysa + rg · GD–1 /p − t + cg − π  · v–1 ≡ c + Δv (1.25)
where π’ = Δp/p
 Inflation Accounting of Whole Economic Systems 29

Equation (1.25) says that, for real income to equal real consumption
plus the change in real financial wealth, the erosion in the opening stock
of wealth as a result of inflation must be deducted from the sum of the
real flows.
Income is inclusive of stock appreciation, as must be the case if the
flows are to be consistent with changes in the stocks.

Nominal and real interest rates

The logical relationship between nominal and real interest rates is given
by the famous equation of Fisher (1896, 1930):

1 + rr = (1 + r) · p–1 /p (1.26)

or

r · p–1 /p = rr + π  (1.27)

Using this expression we can divide the nominal interest on govern-

ment debt into two parts: one representing the real interest flow and the
other the erosion of the real stock of debt by inflation.

rg · GD–1 rg · (p–1 · gd–1 )

= = (rrg + π  ) · gd–1 (1.28)
p p

Hence we can re-express the definition of real personal disposable

income as:

ydsa ≡ ysa − t + cg + rrg · gd–1 − π  · (v–1 − gd–1 ) ≡ c + Δv (1.29)

In this equation the interest flow on government debt has been split
into two components. One component is equal to the erosion of the
value of existing debt by inflation. The other element, rrg · gd–1 , is the
real interest on the real stock of debt. This means that real personal dis-
posable income can be defined in the following way. It includes real
post-tax national income (including stock appreciation) plus real capital
gains and real interest on government debt. From this the losses on total
financial wealth due to inflation must be deducted, offset by the compen-
satory component of nominal interest payments on government debt.
It might be noticed that since the difference between total wealth and
government debt must be private debt (v–1 – gd –1 ) in Equation (1.30) is
equivalent to (l–1 + k–1 · q–1 ). Thus:

ydsa ≡ ysa − t + cg + rrg · gd–1 − ρ · (l–1 + k–1 · q–1 ) (1.30)

30 Early Views on the Stock–Flow Coherent Approach

Real public disposable income

Turning next to the definition of real public disposable income, we need
again to take account of the erosion of existing debt by inflation. The
change in real government debt can be defined as the deflated flow of
new debt less the change in real value of the existing real debt i.e.:
ΔGD
Δgd = − π  · gd–1 (1.31)
p
Now the deflated change in new debt is given by:
ΔGD G − T + rg · GD–1
≡ ≡ g − t + (rrg + π  ) · gd–1 (1.32)
p p
Hence combining the last two expressions we get:

gd = g − t + rrg · gd–1 (1.33)

Thus, only if real government disposable income is defined as follows

is the budget constraint preserved in real terms:

ydg ≡ t − rrg · gd–1 (1.34)

Real national income

Real national income is obtained as the sum of personal and government
disposable income:

ydsa ≡ ysa − t + cg + rrg · gd–1 − π  · (l–1 + k–1 · q–1 ) (1.35)

and

ydg ≡ t − rrg · gd–1 (1.36)

Combining the above expressions and substituting the real stock of

inventories for the real stock of company debt gives:

ydsa + ydg ≡ ysa − π  · in–1 + cg − π  · k–1 · q–1 (1.37)

Real personal plus government disposable income are equivalent to

real national income excluding stock appreciation, plus any increase in
the value of the stock of fixed assets over and above that required to
increase the replacement cost of the existing capital stock by the rate of
inflation.
Real national expenditure can be obtained by deflating nominal
expenditure by the price index. Nominal expenditure is defined in the
conventional way as in Equation (1.7):

Ysa ≡ C + G + p · Δk + p · Δin + in–1 · Δp (1.38)

 Inflation Accounting of Whole Economic Systems 31

Dividing through by p we have:

ysa ≡ c + g + Δk + Δin + π  · in–1 (1.39)

or on the conventional definition excluding stock appreciation:

y ≡ ysa − π  · in–1 ≡ c + g + Δk + Δin (1.40)

The conditions for inflation neutrality

Introduction and assumptions
In this section we examine in an accounting sense, the ways in which
inflation neutrality might be maintained in an economy with rising
prices. It is not our intention to introduce a model of macroeconomic
behaviour, but merely to track the consequences of inflation within the
context of a minimum number of additional behavioural assumptions.
We will examine a hypothetical period in which no real changes occur
but in which prices rise. We will also assume that:

(a) Real rates of interest are invariant to the rate of inflation. This
assumption prevents any redistribution through inflation of real
income between debtors and creditors, because debtors are exactly
compensated for inflation losses on existing assets.
(b) The net profit mark-up on historic costs of production plus interest
is invariant to the rate of inflation. This assumption taken together
with (a) implies no redistribution of wealth or income between wages
and profits as a result of inflation.

The compensating processes

In general, if an economic system is to be inflation neutral there must
be mechanisms which ensure that all nominal stocks and flows are fully
adjusted when inflation occurs, so that their real values are unaffected.
The accounting system deployed in sections 1 and 2 requires three
separate compensating mechanisms relating to private debt, government
debt and equities respectively.

(i) Nominal interest rates must rise enough so that the nominal income
stream provides full compensation for the erosion of the real value
of the stock of debt and also full compensation for the loss, through
inflation, of the real income flow. Since we assume that inven-
tories are financed by loans, companies must increase their bank
borrowing fully in line with inflation. As we shall demonstrate,
maintenance of real stocks and flows implies that companies’ nom-
inal revenue receipts from sales will be less than their current cash
32 Early Views on the Stock–Flow Coherent Approach

outgoings, by an amount which exactly equals the increase in the

value of their stocks. Also the amount of extra borrowing exactly
matches the compensation element in higher interest payments.
(ii) Given the real and nominal flows of government expenditure and
tax receipts, the compensation of existing holders of government
debt through higher nominal interest rates requires additional nom-
inal borrowing by the government. Again the scale of the additional
nominal borrowing will be exactly equal to the addition to the nom-
inal interest flow and will be exactly such as will raise nominal
government debt in proportion to the price level. The additional
interest payments are paid either to banks or direct to the personal
sector. Our assumption that all interest receipts by banks are passed
on to depositors means that the personal sector ultimately receives
the entire additional flow.
(iii) Since under assumption (b) the flow of real profits is invariant to
inflation, it is reasonable to infer that the market value of equity in
real terms, if given by the stream of expected profits discounted by
the shareholders’ required rate of return, would also be invariant to
inflation.

Let us now examine these propositions formally. As a starting point

we have the definition of national income from the expenditure side:

Y ≡ S + ΔIN (1.41)

where S is the sum of consumer and government spending on goods and

services, plus investment in fixed capital. Subtracting taxes from both
sides, and substituting Equations (1.4 and 1.14) in Equation (1.41) gives:

S − T ≡ Y − T − ΔIN

or

S − T + ΔIN ≡ FN + (rl · L–1 + WB) (1.42)

Equation (1.42) demonstrates that companies outgoings fall short of

their revenue by an amount which equals ΔIN, requiring additional
borrowing.
To simplify our analysis we assume that the length of the accounting
period and the production period coincide so that we may equate ΔWB =
ΔIN = ΔL.4 Equation (1.42) may therefore be expressed as:

S − T ≡ FN + (1 + rl ) · WB–1 (1.43)
 Inflation Accounting of Whole Economic Systems 33

If we now define the real rate of interest with respect to cost inflation,5
1 + rrl = (1 + rl ) · WB–1 /WB, Equation (1.43) may be re-expressed as:

S − T ≡ FN + (1 + rrl ) · WB (1.44)

A final rearrangement of Equation (1.44) is made by assuming that net

profits are set as a mark-up on wage and interest costs:

FN = φ · (1 + rl ) · WB–1 = φ · (1 + rrl ) · WB (1.45)

which gives:

S − T = (1 + φ) · (1 + rrl ) · WB (1.46)

Equation (1.46) says that the value of final sales, given the net profit
mark-up φ and the real interest rate, increases directly (and immediately)
in proportion to current wage costs. The net profit flow in real terms may
now be written as:

fN = φ · (1 + rrl ) · wb (1.47)

If the mark-up φ is not affected by wage inflation then net real profits
are unaffected as well. We may therefore infer that the real value of the
equity given by the stream of real profits discounted by shareholders’
required real rate of return is also invariant to the rate of inflation.
Now recall Equation (1.37), which defines real personal disposable
income:

ydsa = ysa − t + rrg · gd–1 + cg − π  · (l–1 + k–1 · q–1 ) = c + Δv (1.48)

Since personal disposable income ysa − t is the sum of real wages and
profits we can write:

ysa − t = (FN + rl · L–1 + WB)/p = fN + rrl · l–1 + wb + π  · l–1 (1.49)

Hence:

ydsa = fN + wb + rrg · gd–1 + rrl · l–1 + (cg − π  · k–1 · q–1 ) (1.50)

Now capital gains on equities completely offset the erosion of existing

equities by inflation (i.e., cg − π  · k–1 · q–1 ). This can be shown recalling
Equation (1.11), i.e.,
CG k–1 · (p · q–1 + q · p)
cg = = + k · (q − 1)
p p
= q–1 · k–1 + q · k–1 + k · (q − 1) (1.51)
34 Early Views on the Stock–Flow Coherent Approach

Equation (1.52) shows three sources of real capital gains: the first two
relating to the existing capital stock and the third to new investment.
The first term is the revaluation of existing equity in line with price
inflation. The second arises from changes in equity prices over and above
any changes in commodity prices (and hence on top of the replacement
cost of the real capital stock). Third is any discrepancy in the valuation
of new investment between the value at replacement cost in real terms
and its valuation in the equity market.
Under our conditions for inflation neutrality, that real interest rates
and the profit mark-up remain constant, the real value of equity will
be constant and hence q will be constant. If the market value of equity
were originally equal to the replacement cost of the capital stock, infla-
tion would leave q = 1 and the last two terms of Equation (1.51) would
be zero.
Substituting Equations (1.51) and (1.49) in Equation (1.48) gives:
ydsa = wb + fN + rrl · l–1 + rrg · gd–1 = c + Δv (1.52)
In view of the banks’ balance sheet identity and the assumption that
banks earn no net income Equation (1.52) may also be expressed as:
ydsa = wb + fN + rrm · m–1 + rrb · bh−1 = c + Δv (1.53)
When real profits and the real interest rate are invariant to the rate of
inflation, the erosion of the real value of the personal sector’s net asset
holdings is entirely compensated for through adjustments of nominal
interest rates on capital certain holdings and through capital gains on
equity. The personal sector’s real disposable income (allowing for the
erosion by inflation of existing assets) may therefore be written entirely
in terms of real flows and real rates of interest on inherited real stocks.
We no longer need any terms representing inflation gains or losses to obtain a
definition of real sectoral income fully consistent with changes in real wealth.
We may summarize our complete set of macroeconomic accounts
using Equations (1.33), (1.34) and (1.53) to express the flow of funds
and change in balance sheet for a fully inflation accounted, inflation
neutral economy in real terms:
(wb + fN + rrm · m–1 + rrb · bh−1 − c) − (Δin + Δk) = g − ydg (1.54)
or,
Δv − (Δl + Δ(k · q)) = Δgd (1.55)
Some qualifications
It remains to list briefly some of the main circumstances under which, in
practice, inflation will alter real stocks and flows even if nominal interest
 Inflation Accounting of Whole Economic Systems 35

rates rise in line with the rate of inflation and the profit mark-up on
historic costs plus interest payments is constant.

(a) The holders of assets will not be fully compensated for the erosion
in their real value, even if real interest rates are unaltered, unless
they are capital certain. This is because a rise in nominal rates causes
the market value of bonds to fall. If inflation and nominal interest
rates rise to new constant rates, holders of bonds will suffer a once
and for all real and nominal capital loss. Thereafter bond holders
are compensated for the continuing rise in commodity prices by a
combination of an increasing bond price towards redemption, and a
higher yield.6
(b) Holders of capital certain assets will not be fully compensated by
a rise in nominal interest rates in line with inflation if direct taxes
are imposed in differently on all nominal income flows. (This is a
problem we sidestepped in the main exposition, by assuming all taxes
to be indirect.) If holders of assets are to be compensated for the
erosion of the real value of their assets, it is necessary to exempt from
taxation that part of the interest flow or any capital gains which is
topping up the real value of the stock of assets.
(c) The real flow of dividends will not be fully maintained unless the
mark-up on historic costs plus current interest payments is invariant
to inflation implying that prices change simultaneously and fully
in line with costs. There is plenty of empirical evidence that, in
reality, price changes lag behind cost changes. Therefore if nomi-
nal rates of interest rise when inflation increases, there is likely to be
some redistribution of real income and wealth away from owners of
equity and in favour of income from employment. Whether or not
the distribution is away from rentiers as well depends on the scale
of the increase in nominal interest rates and also on the extent to
which such an increase causes a reduction in the nominal value of
bonds.
(d) Even though the price of housing (which in the UK is the largest sin-
gle asset held by the personal sector) may be assumed to rise roughly
in line with inflation, the value of mortgage debt is unlikely to rise
simultaneously fully in proportion. This is because mortgages (in the
UK at least) are held for an average of ten years or so. The turnover of
mortgages will however, tend eventually to maintain the real value
of the entire stock of mortgage borrowing.

A final point. We know that if real asset stocks are to be fully main-
tained through a period of accelerating inflation, it is necessary for the
36 Early Views on the Stock–Flow Coherent Approach

nominal public sector deficit to increase to an extent which exactly com-

pensates for the erosion through inflation of the real public debt. We
have described the conditions under which increases in interest rates by
themselves would bring about this compensation.
But it is always open to governments to maintain the real value of
their debt directly via the fiscal system, by raising expenditure or cutting
taxation. This kind of compensation would, however, redistribute real
income and wealth between sectors even if it succeeds in maintaining
the aggregate real income flow.
We end by emphasizing that this paper has been entirely concerned
with measurement and accounting. Nothing it contains should be inter-
preted as normative. For instance, the fact that the public sector’s
nominal deficit must rise if its real debt is to remain constant through an
inflationary period does not, of itself, imply that the government ought
to act in any particular way.

Appendix: Inflation and real capital losses on fixed coupon

bonds and perpetuities
As noted in the text, owners of capital-certain bills with variable interest rates can
be directly compensated for the fall in the real value of bills which occurs when
prices rise, by a sufficient increase in the interest rate. What effect will inflation
have on the real value of bonds which carry a fixed coupon rate of interest and on
the real income derived from their ownership? The following three consequences
might be supposed if an increase in inflation is accompanied by a rise in the
market yields on bonds. Firstly the nominal and real price of bonds would fall.
Secondly the real income derived from bonds would decline each period so long
as inflation continued since the nominal income derived is fixed. Finally, the real
value of bonds would also be eroded in line with inflation independently of fur-
ther changes in yields. These consequences might be supposed to be particularly
acute in the case of perpetuities where the nominal price of a consol would move
inversely with the rise in the market yield, and subsequently remain unchanged
in nominal value implying further erosion in real terms.
Though plausible, this intuition is incorrect. If the market yield fully reflected
the rate of inflation (i.e. in preserving the real market yield) bond holders would
suffer a once-for-all fall in nominal and real capital but no continuing erosion of
real wealth. There would also be a step fall in the real income earned by applying
an unchanged real yield to a lower real value of capital, since some real capital
is ‘destroyed’ by inflation. Real income would not however continue to fall even
though inflation persisted. These propositions apply whether the bond has a finite
redemption date or is a perpetuity.
Consider first the valuation of a bond in the absence of inflation. Let:
r = the coupon rate of interest on bonds,
B = the par of redemption value of bonds,
d = the redemption date of the bonds,
 Inflation Accounting of Whole Economic Systems 37

R = the current market yield on bonds,

Vn d = the opening market price of bonds with redemption date d.
Investors are willing to pay, at time t, the present value of the stream of returns
on a bond, when discounted at the market yield. Hence:


n
rB B
Vt−1 , d = + (A.1)
(1 + R)i (1 + R)n
i=1

where n = d − (t–1 )
Evaluating Equation (A.1) gives a formula relating yield and market price.
(r(1 + R)n + R − r)B
Vt−1 , d = (A.2)
R(1 + R)n
In the case of perpetuities, as n tends to infinity Equation (A.2) becomes:
rB
Vt−1 , d = (A.2a)
R
where n = d − (t–1 )
Assuming that R > r the capital appreciation per period is given as:
(R − r)B
ΔV = Vti − Vt−1 = (A.3)
(I + R)n
The total return per period, expressed as a proportion of the opening market
price of the bond, consists of interest income, rB, and capital gain, ΔV .
Hence from Equations (A.2) and (A.3),
 
(R − r)B r(1 + R)n−1 B + (R − r)B
(rB + ΔV )/V–1 = (rB + ÷ (A.4)
(1 + R)n R(1 + R)n
The total return per period equals the market (redemption) yield of the bond.
Now suppose commodity prices are increasing at σ per cent per period, that r
is the real market yield (also equal to the fixed coupon rate) and R is the nominal
yield such that:

R = r(1 + σ ) + σ (A.5)

Hence from Equation (A.4) in every period during which a constant rate of
inflation prevails the total return consists of:

RV–1 = r(1 + σ )V–1 + σ V–1 = rB + V − V–1 (A.6)

The first term after the first equality sign is a nominal income stream that grows
at a rate equal to the inflation rate, preserving a constant real income. The second
term is the return which exactly compensates the bond holder for the erosion of
the opening real value of the bond. The total return accrues partly as a nominal
income stream, rB and partly as nominal capital appreciation. Rearrangement of
Equation (A.6) shows that the maximum ‘spendable’ real income, while main-
taining real capital intact, requires that part of the nominal income stream be
saved and re-invested.

σ V–1 = r(B–1 − (1 + σ )V–1 ) + ΔV (A.7)

38 Early Views on the Stock–Flow Coherent Approach

This ensures that after the initial fall in the nominal value of the bond which
accompanies the increase in R, both the nominal income and capital subsequently
grow at the rate of inflation until redemption. If the bond is a perpetuity the
nominal value grows at the inflation rate entirely through reinvestment of the
balance of income received over ‘spendable’ income, i.e. in Equation (A.7), ΔV = 0
and:
σ V–1 = r(B − (1 + σ )V–1 (A.7a)
But using Equations (A.2a) and (A.5) in Equation (A.7a) we obtain:
[r(1 + σ ) + σ ]V–1 − r(1 + σ )V–1 = σ v–1 (A.8)
We have therefore shown that, with capital uncertain fixed coupon bonds, if
the real yield is invariant to the inflation rate, there is a once-for-all fall in the real
value of bonds. Since by assumption the real yield is unchanged the real income
flow rate falls entirely because of the ‘destruction’ of part of the real stock of bonds.
The real capital loss is greatest for perpetuities.

Notes
1. But see Buiter (1983), Godley and Cripps (1983) and Jump (1980) for excep-
tions.
2. Hicks (1939, p. 179).
3. Note however that Equation (1.10) would not be regarded as conventional by
everyone because Ysa and YD both include stock appreciation.
4. The argument which follows does not depend critically on this assumption
and may easily be relaxed as shown in Godley and Cripps (1983).
5. The logic of the argument strictly speaking requires that real interest be defined
with respect to price inflation. If the real interest rate rr remains constant when
cost inflation occurs then so also will the real interest rate defined with respect
to price inflation.
6. The appendix sets out the reasoning which supports this proposition

References

Buiter, W.H. (1983) ‘Measurement of the Public Sector Deficit and Its Implications
for Policy Evaluation and Design.’ IMF Staff Papers XXX, no. 2.
Fisher, I. (1930) The Theory of Interest (New York).
Fisher, I. (1896) ‘Appreciation and Interest.’ Publications of the American Eco-
nomic Association, August.
Godley, W. and F. Cripps (1983) Macroeconomics (Oxford University Press).
Hibbert, J. (1983) ‘Measuring the Effects of Inflation on Income, Saving and
Wealth.’ Report prepared by Mr Hibbert acting as consultant to the secretariat
of the OECD and Eurostat (OECD, Paris).
Hicks, John R. (1939) Value and Capital (Oxford).
Jump, Gregory V. (1980) ‘Interest Rates, Inflation Expectations and Spurious Ele-
ments in Measured Real Income and Saving.’ American Economic Review 70 (5)
(December): 990–1004.
Taylor, C.T. and A.R. Threadgold (1979) “‘Real” National Saving and Its Sectoral
Composition, Bank of England.’ Discussion paper no. 6, October.
2
Time, Increasing Returns and
Institutions in Macroeconomics1
Wynne Godley

Throughout my working life I have looked to Sylos Labini for inspiration,

presuming to find an affinity with him in major lines of enquiry we have
both pursued. I am thinking, most particularly, of his work on a series of
related questions concerning the behaviour of firms under conditions of
oligopoly, the well-attested practice of mark-up pricing and the meaning
(if any) of the aggregate neoclassical production function.
It seems to me that Sylos Labini’s findings are extremely radical in
their implications. They are indigestible if not, in the end, lethal to
the neoclassical paradigm although this is not exactly (so far as I know)
the conclusion he himself draws explicitly. Yet when, having produced
a destructive critique of the neoclassical production function, he asks,
‘When will economists finally accept their own logic?’ I do believe he is
not just sniping from the sidelines at the neoclassical paradigm (NCP),
he is shaking at one of its foundation stones. For this reason my short
answer to his question is ‘Never’ or at least ‘Not until we have a new
paradigm’.
In this chapter written in his honour I shall make my own character-
ization and critique of the NCP. Then, knowing that no paradigm can
be successfully contradicted but only ousted, I shall sketch an alterna-
tive synthesis based on my eclectic understanding of Sylos Labini and a
number of other economists mainly from Italy or from Cambridge, Eng-
land. It attributes distinctive motivations and functions to corporations
and banks as well as to individual agents and therefore belongs to a spe-
cific historical and institutional setting. Production and investment will
be conceived of as processes taking time (therefore requiring finance)
with production subject to increasing returns. Except in financial mar-
kets, equilibria are defined in terms of real stock–flow ratios, while the
prices of goods are generally set as a mark-up on costs, the mark-up being

39
40 Early Views on the Stock–Flow Coherent Approach

determined as part of firms’ strategy to maximize market shares. The

model provides a framework within which the determinants of income
distribution can be characterized as well as the process of polarization
between relatively successful and relatively unsuccessful economies and
regions, which is possibly the most important feature of the capitalist
world economy today and which the NCP cannot handle.
What is the NCP anyway? I think the best way to start is by consider-
ing the market-clearing, full-employment macroeconomic equilibrium
in the form in which it now appears towards the beginning of virtually
every undergraduate textbook – namely as the intersection of an aggre-
gate demand curve with a vertical aggregate supply curve. Underpinning
this simple diagram there has to be a system of at least ten equations for
which the locus classicus is Modigliani (1944) as amended by Modigliani
(1963). As Modigliani makes clear, but the textbooks conceal, the equi-
librium described by this intersection is, in the technical sense, a general
equilibrium (GE) in which ‘prices’ have been found which simultane-
ously clear all the four ‘markets’ which comprise it. Written formally the
following equilibrium conditions must all be satisfied, y s = y d , M s = M d ,
N s = N d , and, by implication, Bs = Bd where y is real product, M the
stock of money, N labour, B the stock of bonds and the superscripts s
and d denote respectively supply and demand.
I am convinced that this concept of general equilibrium in a monetary
economy constitutes the primal scene2 – the primitive imaginary vision
of the world – out of which the whole of mainstream macroeconomics
now flows. At one extreme are ‘monetarists’ of various hue who believe
that the classical version of this simple model does, or should, or can
somehow be made to describe the real world. Almost all other modern
macroeconomists, while forming a huge spectrum, have as their essen-
tial activity the study of what happens if parts of the machine do not
function properly, e.g. are subject to rigidities or time lags. For instance,
much work has been concerned with the effects on the solution of this
model if the various prices do not clear markets or clear them imper-
fectly. If wages are not flexible the labour market may not clear; this
is what most students now understand as Keynesian economics. If the
price of goods is not flexible, the market for goods may not clear, perhaps
generating ‘classical’ unemployment.
Now Sylos Labini (like Kaldor and Pasinetti in different ways) makes
a devastating case against the empirical relevance or even meaningful-
ness of the aggregate neoclassical production function. What I want to
emphasize here is the system role which the production function fulfils
and therefore just why the Sylos Labini critique is so important. What the
 Time, Increasing Returns and Institutions in Macroeconomics 41

production function does for all equilibrium systems – whether markets

clear or not – is to bring labour into instantaneous equivalence with real
product in such a way that alternative quantities of each can potentially
be traded against one another. The production function is necessary for
this equivalence so that labour can instantaneously be translated into the
profit-maximizing quantity of product which firms are therefore moti-
vated to supply. Without the production function no neoclassical model
will start up; the blood supply to its head is cut off.
The fact that production cannot, in reality, be instantaneous leads
to a range of problems which have often been raised (for instance by
Hicks (1980–81) and Tobin (1980)) but not resolved within the frame-
work under discussion. These concern the historical time period over
which all these things are supposed to happen. Since, in the absence of
futures markets, only one price is allowed in each market per trading
episode, severe problems arise if the entire process, by which I mean the
finding of prices and also acting on them, does not take place instan-
taneously. Yet the shorter time allowed, the less plausible becomes the
implied behaviour of some variables, particularly investment which in
reality takes months to implement.
The critique of neoclassical macroeconomics made by Hicks (1989) is as
fundamental as that of Sylos Labini. According to Hicks, drawing heavily
on Kaldor (1939), the realistic way to characterize ‘flex-price’ markets is to
postulate the existence of intermediate traders who are always prepared
either to buy or sell and are enabled to do so by virtue of the fact that they
hold stocks. In the case of ‘fix-price’ goods the holding of stocks both by
manufacturers and retailers is also an essential part of the market process.
But then, as Hicks points out with characteristic moderation, the theory
needed to explain the functioning of markets ‘could not be developed
without a considerable change of view... [since]... the traditional view
that market price is at least in some way determined by an equation
of demand and supply had to be given up’. The question is no longer
whether or not, or how, the equilibrium condition is fulfilled, because
the equilibrium condition itself has been destroyed. The demand–supply
relationship, now being a difference equation (since it contains a term
in stockbuilding) ‘can only be used in a recursive manner to determine a
sequence’. By contrast, the neoclassical equilibrium, even in its ‘Keyne-
sian’ version, is an instantaneously interdependent system from which
all notion of unidirectional causality, working sequentially through time,
is removed. Thus the multiplier must be seen as holding good instan-
taneously, not as a process in which investment generates income in
one period and consumption, employment and further additions to
42 Early Views on the Stock–Flow Coherent Approach

income in subsequent periods as it was certainly originally conceived

by Richard Kahn.
The inclusion of stockbuilding as an indispensable concept will lead to
the resolution of another paradox of the NCP. The neoclassical equilib-
rium deals exclusively in spot transactions which are all simultaneously
determined and executed; hence there is no need for money or finance.
Yet the solution to the macroeconomic model cannot be found unless a
stock of money exists and is willingly held. It is only when trading, as
well as production and investment, takes time and a system requirement
for money and finance is called into existence.
I have reached a point when I am prepared to make a declaration. I
want to say of neoclassical macroeconomics what I have sometimes said
of certain kinds of fiction; I know that the world is not like that and
I have no need to imagine that it is. In particular, I do not believe that
there exists a market in which goods in aggregate and labour in aggregate
can be exchanged for one another provided only that the price of each
is right in relation to some given stock of ‘money’.
But my objection goes beyond scepticism that the world we live in is
being described realistically. My additional concern is that the NCP is
prejudicial with regard to the understanding of some of the most impor-
tant processes going on in the world today. Thus in the ‘classical’ version
of the NCP real output is determined by supply side factors alone; fiscal
policy is entirely impotent and the government can only affect anything
by changing the money supply; even then all it can do is affect the price
level. The idea that fiscal policy is impotent, which seems to be based
entirely on this model, has been extremely influential in contemporary
political discussion; it is not just a provisional result suitable for a week
or two in an elementary class.
Then the abolition of time prejudices the perception of inflation as an
evolutionary process; the equilibria generate ‘explanations’ of price lev-
els not changes, and theories of inflation cannot be convincingly coaxed
forth. As if this were not enough, the whole construction leads by virtue
of its axioms to the conclusion that wage and price flexibility, in combi-
nation with free trade, will generate full employment and convergence,
if not equalization, of living standards between countries and between
regions within countries. In sum, while the absence of processes occur-
ring in historical time means that the NCP does not encourage students
to go and look up figures in books, if and when they are forced to do so
their vision is likely to have been for ever distorted.
I shall have further complaints to make and some of these will appear
a contrario in the following section.
 Time, Increasing Returns and Institutions in Macroeconomics 43

Sketch of an alternative macroeconomic theory

I think it is true to say that anti-neoclassical Keynesian economists,

among whom I number Sylos Labini and Kaldor, notwithstanding their
penetrating and suggestive insights, have not succeeded in creating an
alternative paradigm. There is, for instance, no Kaldorian textbook;
Kaldorian ideas in their positive mode have not been put together in
a way which covers the syllabus.3
In the rest of this paper I am going to put forward a sketch4 of an
alternative synthesis of which almost all the component parts were first
thought of by other people. It is based on my eclectic understanding (in
particular) not only of Sylos Labini but also of Graziani, Hicks, Keynes,
Kaldor, Pasinetti, Tobin and Adrian Wood. I shall call it a real stock–flow
monetary model – RSFM for short.

A methodology for modelling capitalist institutions

within a complete macroeconomic system

It will be a central feature of the proposed model that distinct func-

tions and motivations will be attributed to several different institutional
sectors. To keep the deployment within reasonable bounds, the follow-
ing analysis will be confined to the closed economy case although it is
emphasized that by doing this many of the most important and difficult
problems, both of economic policy and of macroeconomic theory, are
assumed away. Our economy will therefore consist here of four sectors
only: corporations, banks, the government and households.
If all these sectors are to have different motivations (in particular if
banks and corporations are not just passive agents for households), there
is no option but to start by mastering the purely accounting relationships
between them. The accounting may be tedious when taken by itself.
But if all the logical relationships are mastered at the beginning it will
become clear that the number of behavioural relations which drive the
model remains quite small; indeed the model is very parsimonious – at
its simplest there are only ten to fifteen equations but these (given the
exogenous variables) generate a well-defined dynamic path for at least a
hundred variables.

Basic accountancy of the RSFM model

The accounting framework bears some resemblance to Tobin (1982) an

important difference being that Tobin’s 1982 model has no commercial
banks nor, a fortiori, any need for them.
44 Early Views on the Stock–Flow Coherent Approach

In contrast to the basic NCP model (and also to the existing conven-
tions of national income accounting) the accounting of the RSFM model
starts off with a consistent set of sectoral and national balance sheets
where every financial asset, valued at current market prices, has a coun-
terpart in a financial liability somewhere else; the net worth of each
sector is obtained by summing its assets and liabilities. Budget constraints
describe for each sector how net flows of expenditure, factor income and
transfers have counterparts in changes in asset stocks.
The entire system of stocks and flows valued at current prices is now
deflated using appropriate price indices. It is a distinctive feature of the
RSFM model that, although it is denominated and motivated in real
terms, it is not a ‘fix-price’ model. The rate of inflation will turn up as an
argument of many functions.
While the steps (which have been omitted) are particularly tedious,
the process of inflation accounting yields neat and intuitive results. For
instance the government’s budget constraint at current prices is (at its
simplest):

H + B = G − T + rb · B–1 (2.1)

where H is high-powered money, B bonds, G government expenditure, T

transfers, rb the nominal rate of interest on bonds, and  a first difference
operator.
In real terms the government budget constraint is simply

h + b = g − t + rrb · b–1 + rrh · h–1 where (2.2)

h = H/p − H–1 /p–1 etc. (2.3)

and where ex post real interest rates are defined by variants of the Fisher
equation, e.g.
r −π
rrb = b (2.4)
1+π
−π
rrh = (2.5)
1+π
where π is the inflation rate.
A consistent system of accounts which includes corporations and
banks requires, at its simplest, that household wealth must have at least
four component parts:

v = h h + m + bh + e (2.6)

where v is real wealth, hh (real) cash, m real interest bearing money, bh

real bonds and e the real value of equity.
 Time, Increasing Returns and Institutions in Macroeconomics 45

The definition of real household disposable incomes is such that if all

income is spent the change in the real stock of wealth is unchanged.
Thus defining real disposable income

yd = v + c (2.7)

where c is consumption, consistent accounting yields

yd = y − t + rrh · h–1 + rrm · m–1 + rrb · b–1 + rre · e–1 (2.8)

where all real rates are defined to comprise real capital gains.
I go into a little more detail with regard to the accounting of corpo-
rations because I want to bring emphasis to the strategic identity which
describes the outcome of the pricing decision for the distribution of the
real national income.
The consistent treatment of corporations within the accounting matrix
leads to a slightly unconventional definition of profits. The conventional
definition of gross (accounting) profit at current prices, assuming the
economy to be closed, can be derived precisely from the appropriation
account of the aggregated company sector.

FT = S − T − WB + IN (2.9)

where FT is accounting profit, S (sales) the sum of public and private

consumption and fixed investment, WB the wage bill, and IN end-period
inventories valued at cost.
I propose the following definition of entity profit (FN ) since interest
on inventories is an inevitable cost arising from the fact that production
takes time. This means that firms have to borrow from themselves to
finance inventories if they do not borrow from banks, thereby forgoing
interest.

FN = S − T − WB + IN − rl · IN–1 (2.10)

where rl is the nominal rate of interest on loans. This differs from

the conventional definition of economic profit which deducts stock
appreciation5 from accounting profit. The definition in Equation (2.10)
will only be equal to the conventional definition in the special case where
the real rate of interest is exactly zero, i.e. where the nominal rate of
interest exactly equals the rate of cost inflation.
Note that Equation (2.10) may alternatively be written:

S = FN + T + (1 − σ ) · WB + (1 + σ · rl ) · WB–1 (2.11)
46 Early Views on the Stock–Flow Coherent Approach

where σ is the ratio of inventories to sales. Equation (2.11) says that

sales are equal to profits plus the historic cost (including the unavoidable
interest cost) of producing the goods sold.
We now define τ as the indirect tax rate,  the mark up of price on
historic cost, UC the wage bill per unit of output and rr w the real rate of
interest defined with respect to cost inflation rrw = rl − πc /1 + πc , where
πc is the rate of change in UC. These terms may be substituted6 into
Equation (2.11) to obtain an adding-up identity describing the compo-
nents of prices with no lagged terms although the mark-up is on historic
costs; the trick has been to collapse time with the real interest rate.

p = (1 + τ ) · (1 + ) · (1 + σ · rrw ) · UC (2.12)

Alternatively, dividing through by p, we have:

1 = (1 + τ ) · (1 + ) · (1 + σ · rrw ) · uc (2.13)

where uc is real unit wage costs.

And multiplying through by output:

y = (1 + τ ) · (1 + ) · (1 + σ · rrw ) · wb (2.14)

The accounting identities described in Equations (2.12)–(2.14)7 have

a crucial importance in the RSFM model. They provide the essential link
between, on the one hand, fiscal policy, real interest rates and firms’
pricing decisions and, on the other, the distribution of the real national
income as a four-way split between the government, firms, banks and
households. They will also provide a framework for the discussion of
inflation as resolving rival claims on the real national income.

Behaviour of the major sectors

Corporations
I start with four stylized facts:

(a) Prices of goods are extremely sticky relative to normal costs defined as
the unit prime costs which would have obtained at normal capacity
utilization.
(b) Labour productivity is positively associated with production. This
is a stylized fact confirmed by virtually every observation across
industries and economies and during long as well as short periods.
(c) As follows from (a) and (b), profits fluctuate pro-cyclically in the short
term.
 Time, Increasing Returns and Institutions in Macroeconomics 47

(d) The predominant source of finance for fixed investment is retained

profits.

With these stylized facts in mind, consider the main functions which
describe the behaviour of corporations.
There is, first of all, the short-period production decision which is well
described by a model similar to that known as the Keynesian cross.8 That
is, firms aim to meet sales plus normal inventory accumulation plus the
replacement of inventories arising from earlier mistakes. Thus,

y = se + ine − in–1 (2.15)

where the superscript e means expected values.

Desired inventories are in some ratio to expected sales:

ine = σ T · se (2.16)

There is no need for a complex model to explain short-term expec-

tations about sales, which will usually be indicated by the size of order
books. Mistakes about sales are quickly signalled by the ex post movement
of inventories.
I do not believe that neoclassical supply constraints ever bind in the
sense that an extra order is not met because, whereas the previous order
was profitable, this one is not.9 On the other hand, there do arise straight-
forward capacity constraints when no more can physically be produced;
this is signalled by lengthening order books and delivery dates.
There is a system requirement10 that inventories be financed from
outside the corporate sector. This can easily be seen if the corporate
appropriation identity, Equation (2.10), is rearranged

S − FN − T − rl · IN–1 − WB = IN (2.17)

Equation (2.17) is saying that receipts from sales in each period fall
short of profits plus all costs of production in the same period by exactly
the change in the value of stocks (IN), a concept which includes stock
appreciation. In other words, the circular flow of income, as conven-
tionally thought of and taught, has a hole in it. It will, accordingly, be
one of the main system functions of commercial banks to provide the
finance for inventory accumulation in any economy which is growing,
whether the growth is real or purely inflationary. It will also (a fortiori) be
part of their function to accommodate firms when inventories change
involuntarily because of mistaken expectations.
We next come to the whole nexus of decisions concerning pricing,
investment, marketing, choice of technique and so on. Our task, not a
48 Early Views on the Stock–Flow Coherent Approach

small one, is to find a replacement for the mechanisms which motivate

supply strategies and determine the distribution of income in the NCP
without resorting to the contingent and theoretically empty ‘stickiness’
or ‘rigidity’ which marks American Keynesianism or, more particularly,
the ‘quantity rationing’ theories such as that proposed by Malinvaud. It
has been a great gap in modern macroeconomic discussion that, while
the existence of mark-up pricing is widely accepted, few people seem to
have any idea as to why the mark-up is what it is.
I believe that Adrian Wood in his wrongly neglected A Theory of Prof-
its has given the key to understanding how these decisions are made.
Wood’s theory derives from the basic hypothesis that the aim of corpo-
rations is not to maximize profits11 but to maximize their own market
shares. He first points out that there is a financial constraint defined, as
a matter of arithmetic, by the profit margin, the growth rate of sales, the
level of investment and the availability of external funds.
Specifically, it must be the case that the minimum level of (real) profits
is given by
(1 − ξ ) · i
f ≥ (2.18)
ψ
where f is real profits, i fixed investment, ξ the share of investment
that can be financed externally and ψ the maximum proportion of
profits which can be retained in order to keep shareholders happy and
predators at bay. Note that the definition of profits here is harmonious
with that described in the accounting section (Equation (2.13) above)
because interest payments which have to be made as a result of holding
inventories are not available for financing fixed investment.
Defining (ϕ) the share of profits in sales12
f
ϕ= (2.19)
s
we divide Equations (18) and (19) through by sales to obtain
(1 − ξ ) · i/s
ϕ≥ (2.20)
ψ
and multiplying and dividing the rhs by s we infer that the profit
margin must be at least large enough to satisfy
(1 − ξ ) · ik/gr
ϕ≥ (2.21)
ψ
where ik is the incremental investment/sales ratio (i/s), (henceforth to
be called the ‘investment coefficient’) and g is the growth rate of sales.
 Time, Increasing Returns and Institutions in Macroeconomics 49

ik1
ik1 ik2
ik2
ik3
ik3

ik4 ik4

ϕ
ik5
ik5

gr

Figure 2.1 The growth-maximizing mark-up

Next it is postulated that when forming a rolling plan for consecutive

(say) three- or four-year periods, firms must consider the price they will
charge, together with selling costs, marketing strategy, the technique of
production and the scale of investment. The hypothesis here is that the
attainable profit margin will be negatively related to the growth rate of
planned sales and positively related to the investment coefficient.

ϕ ≤ f (gr, ik); f1 < 0; f2 > 0 (2.22)

As illustrated in Figure 2.1, the inequalities, Equations (2.21) and (2.22),

may be drawn as a series of loci in a space defined by the mark-up (ϕ)
and the sales growth rate (gr).
The upward sloping lines radiating from the origin (described by
Equation (2.21)) are frontiers to the right of which (given the invest-
ment coefficient) insufficient funds are available to pay for investment.
The downward sloping lines described by Equation (2.22) are ‘oppor-
tunity frontiers’ which (given the investment coefficient) describe the
maximum obtainable profit margin. There is a presumption that the
loci describing these slope downwards at an increasing rate, that is, for
any given level of investment an addition to the growth rate can only
be achieved at ever-increasing cost in terms of the profit margin since
selling costs must be assumed to be higher. The (ex ante) maximum
feasible growth rate may now be read off13 together with its counter-
part mark-up (ϕ) as the highest point of intersection on the horizontal
50 Early Views on the Stock–Flow Coherent Approach

scale between any compatible pair of loci. Thus the ex ante profits which
are the counterpart of growth maximization simultaneously determines
both the level of investment and the mark-up on historic costs.
Wood’s theory is explicitly a ‘long period’ theory with no relationship
to any real macroeconomic system evolving in historical time. But I don’t
think there is any difficulty in adapting it to be realistic. As I have tried
to show in a number of empirical studies relating to the UK, industrial
prices are generally related, not to actual unit costs, but to normal unit
cost, defined as the costs which would have obtained at normal capacity
utilization. (I must note that Sylos Labini has expressed some dissatis-
faction with this ‘finding’.) Accordingly the Wood hypothesis can be
realistically embodied by use of a simple adaptation of the identity as
shown in Equation (2.12).14

p = (1 + τ ) · (1 + ) · (1 + σ · rr) · NUC (2.23)

where NUC is normal unit wage costs and  is no longer an ex post number
but the mark-up parameter.
What Wood has given is a rationale for the size of the mark-up on
normal costs. This can harmoniously be brought into a realistic macroe-
conomic model so long as actual unit costs are clearly distinguished from
normal unit costs. All short-run departures of actual from normal unit
costs will turn up as short-run procyclical fluctuations in actual relative
to normal profit margins in accordance with the stylized facts set forth in
the introduction to this section. And in the long run the profits of indi-
vidual companies ex post may be swollen or even obliterated if market
shares turn out to be different from those on which plans were based; the
same proposition can be made more emphatically when firms are com-
peting in world markets since the profitability of the industry of whole
countries may be at risk.
Similar considerations apply to investment; just as the theory can only
explain the normal profit margin, it can only explain ‘normal’ invest-
ment. If short-term expectations about the growth of sales are disturbed,
so will investment be disturbed, not least because realized profits (and
thereby internally generated funds) will be below normal in times of
cyclical recession.

The household sector

This section is extremely brief because it contains little which is not quite
conventional.
The household sector is assumed to consume; also to acquire wealth
and allocate it between various assets. Recalling the definition of real
 Time, Increasing Returns and Institutions in Macroeconomics 51

personal disposable income as that which can be consumed while leaving

the real stock of wealth intact (yd = c + v), the consumption function
in a real stock flow model may, at its most stylized, be written

c = c(yd e , v–1 ) (2.24)

or

v = v(v ∗ , v–1 ) (2.24a)

In a hypothetical stationary steady state, a consumption function of

this form implies that enough wealth would have been acquired relative
to income (v = v*) and therefore that consumption would be equal to
disposable income and saving zero. While the result is a commonplace
in modern theory it is very different from the consumption function
required for the neoclassical equilibrium, where saving and investment
must always be positive.
Nor do I have anything special to say about households’ portfolio
choice. There will be a transactions demand for cash, and it is reasonable
to suppose that mistaken expectations with regard to income normally
have a counterpart in unplanned changes in cash balances. Otherwise,
(ignoring the speculative demand for money), it is to be expected (as
Tobin has taught us) that the demand for each type of asset which makes
up total wealth is an increasing function of its own real rate of return and
a decreasing function of all other rates.

Commercial banks
Commercial banks play several crucial roles in the RSFM model. Above
all they provide loans15 which finance various types of transaction, in
particular revolving finance for inventories, for the installation of capital
equipment, for mortgage finance and consumer credit for households.
Perhaps their most fundamental function (as noted off and on down the
centuries from Adam Smith to Graziani) is to monetize inventories; it
has already been pointed out that the circular flow of income is incom-
plete because an economy which is growing in nominal terms requires
a continuing injection of cash from outside the production system to
pay wages in advance of sales. Thus, in the RSFM model, bank loans are
not part of the portfolio decision of the private sector; there is a system
requirement for banks to provide them for the holding of inventories
and the installation of capital equipment. Inventories exist, on the one
hand, as an inevitable consequence of the fact that production and dis-
tribution take time. But on top of this they fluctuate because mistakes
52 Early Views on the Stock–Flow Coherent Approach

are made in the short term by firms as to how much is going to be sold,
and banks have to accommodate this in addition to the perennial need
for revolving finance.
While banks fulfil an indispensable role by providing finance for
inventories, they also have to accommodate the portfolio choice of the
personal sector including the mistakes it makes, which result in short-
term fluctuations on holdings of cash. So (to sum up) in addition to
providing finance for loan expenditure and accommodating portfolio
choice, banks must also be able to accommodate two different kinds of
mistake – those by firms with regard to their production decision and
those by households with regard to their income.
How are banks able to perform these functions, and how are they moti-
vated to do so? The answer to these questions is too intricate to admit of a
statement which would have a focus commensurate with the rest of this
chapter. My conjectures as to how the banking system works are based
on a simulation model in which banks’ operations are fully articulated
with income, expenditure and transfer flows together with asset-demand
functions. The model exploits the constraints imposed by the systematic
accounting of the whole macroeconomy which must be absent from any
model of the banking system in isolation. To summarize broadly. Banks
are motivated to maximize their balance sheets, since at the margin they
can ensure that the return on all their assets exceeds the return on all
their liabilities. My finding is that, given all the stock and flow account-
ing constraints and some condition concerning reserve ratios, there will
always be a configuration of interest rates such that banks will be able,
and also motivated, to supply loans and mediate the non-bank sectors’
portfolio allocation requirements. The whole system, in which there are
enough equations to generate all sectoral assets and liabilities and all
rates of interest, requires assumptions about banks’ portfolio choice as
well as that of the non-bank sector. For instance, bank lending rates must
always be higher than bond yields (otherwise banks would not want to
lend to the private sector) and rates on interest bearing bank deposits
must be lower than bond yields (otherwise neither the public nor the
banks would want to hold bonds). This implies, in contrast to the con-
ventional way of describing the demand for money in a macroeconomic
system (in which there is no need for more than one rate of interest)
that households must have the opportunity to own at least two different
kinds of interest bearing asset and that their rates of interest must be able
to change relative to one another.
I do not think it makes sense to postulate that markets for the assets
and liabilities of the banking system are cleared individually, in the sense
 Time, Increasing Returns and Institutions in Macroeconomics 53

that there is a demand for and supply of interest bearing bank deposits
which is brought into equivalence by the relevant rate of interest. It is,
rather, the case that the whole spectrum of banks’ assets and liabilities,
together with their associated rates of interest, is determined in a single
interdependent process.
An important implication of giving the banks these roles and motiva-
tions is that the stock of credit money can under no circumstances be
treated as an exogenous variable. Given all the stock and flow accounting
constraints, and given also that both households and banks have their
own schedules describing portfolio allocation, the stock of credit money
emerges, determined endogenously, as part of a complicated interactive
process. I believe the habitual way of describing the determination of ‘the
money supply’ in textbooks (slipped in towards the end) as the product
of the total stock of high-powered money and a money multiplier to be
incorrect. If the banking system is modelled as the provider of loans on
a scale determined by industrial needs while simultaneously mediating
non-banks’ portfolio and cash requirements, the ratio of the total stock
of high-powered money to credit money must be variable.

Inflation

A good starting point for the conflictual theory of inflation is the set of
accounting Equations (2.12) through (2.14).

p = (1 + τ ) · (1 + φ) · (1 + σ · rrw ) · UC (2.12)
1 = (1 + τ ) · (1 + ) · (1 + σ · rrw ) · uc (2.13)
y = (1 + τ ) · (1 + ) · (1 + σ · rrw ) · wb (2.14)

These equations only measure ex post numbers. The variable UC in

Equation (2.12) contains the nominal wage per unit of output averaged
over some period of time and is therefore not a good measure of any kind
of behaviour, being a messy weighted average of new settlements and old
settlements made in previous periods which happen to survive for vari-
ous proportions of the present period. To understand wage behaviour it
is necessary to start with the concept of the wage settlement – the money
wage award at the moment it is made. Still less can the concept of the
real wage comprised in Equation (2.13) be thought of as a behavioural
variable; it is doubly endogenous – nominal wage settlements averaged
over the period divided by the price level averaged over the same period.
These various shares can generally be improved or defended to a greater
or less degree: by the government imposing taxes, by firms through the
54 Early Views on the Stock–Flow Coherent Approach

mark-up on costs, by workers bargaining for money wage settlements

and by creditors charging higher nominal rates of interest. The degree
of power which the various groups can exercise is very roughly in accor-
dance with the order I have described them above. The government is
in a particularly strong position to appropriate whatever proportion it
wishes by imposing indirect taxes at the point of sale or direct taxes on
income. Similarly firms are in a good position to determine their own
shares by the mark-up they charge.
Workers, on the other hand, are relatively vulnerable so long as wage
settlements take place at discrete intervals. The hypothesis is that the
key behavioural feature is the size of the real value of the money wage
settlement at the time it is made. Inflation is the process which erodes
the value of the money wage on the scale which makes the ex post real
wage conform with the quantity of real resources available from total
production. This analysis, if correct, emphasizes the irrelevance of the
ex post real wage as an agent of inflation. It must be emphasized that
if the time between wage settlements shortens – if, in particular, wages
are indexed – the workers’ share becomes less vulnerable but only at the
expense of exploding inflation.
Finally, and potentially important both because of its magnitude and
its causal impact, comes the share of creditors. Everything depends on
whether nominal interest rates keep pace with inflation or whether real
interest rates fall as inflation rises. Figure 2.2 shows how inflation would
resolve the various claims on the assumption that nominal interest rates
are invariant to the inflation rate and that wage rates are renegotiated at
Percentage of total cost

τ–τ

φ–φ
100%

rr – rr
Negative real interest
w–w

Price inflation per bargaining period

Figure 2.2 How inflation resolves competing claims in national income

 Time, Increasing Returns and Institutions in Macroeconomics 55

fixed intervals of time. Inflation rates on the horizontal axis are defined
in units of the time over which wages are not renegotiated.
In this table the vertical axis shows the share of income which would
go (ex ante) to each of the four sectors on the assumption that there is
no inflation at all. Thus w – w is the share implied by the real value of
the money wage settlement, rr – rr the share of rentiers (the nominal
rate equals the real rate), φ − φ and τ − τ show the shares appropriated
by firms and the government.
As can be seen from the horizontal axis the sum of ex ante claims
exceeds total resource availability, but at higher rates of inflation, both
the real wage and also the real return to creditors are progressively
reduced. The inflation rate is determined at the point of intersection
of the upper horizontal line where the sum of claims exactly equals the
total availability of resources.
The figure also shows what would happen to the inflation rate, given
all the other assumptions, if the rate of indirect tax were increased. The
share of profits remains the same, and that of the government increases,
while the share both of real wages and of creditors (whose real rate of
return has gone to zero and ultimately would become negative) falls.
I shall not adumbrate any strong theory about what determines the
value of the money wage settlement although this clearly plays a key
role in the process. The motive to go for a high settlement will, however,
be higher the greater is the possibility that the consequence will be that
real wages are actually raised as a result. Thus the incentive to go for
higher wages will be stronger if (for instance) money rates of interest do
not keep pace with inflation or if the government is holding down the
wage rates of its own employees relative to those employed in the private
sector.
It is more likely than not that a high pressure of demand for labour will
be associated with high or rising real wage settlements but the evidence
on this is not clinching.

Fiscal policy and the solution of the whole model

So long as the assumption is retained that the economy is closed, the

government’s fiscal policy occupies the key role in determining the real
output flow. The constraint is not (normally) inflation but the physical
capacity of the economy to produce.
The role of fiscal policy is most readily conveyed by considering the
hypothetical full steady state of the model. We virtually bypass the prob-
lem of inflation (without assuming it away) by considering the entire
56 Early Views on the Stock–Flow Coherent Approach

system in real (inflation accounted) terms. The full steady state, which
has a clear provenance from Christ (1967), Blinder and Solow (1973) and
Tobin and Buiter (1976) lends itself readily to analytic representation.
The important difference from Tobin and Buiter’s ‘Model 2’ is that no
assumption is made that prices are constant. The full stationary equilib-
rium of the RSFM model, if it ever came to pass, would be described not
by market-clearing conditions but by real stock conditions, in particular

in = in∗ , k = k∗ , v = v ∗

where i is real inventories, k the real stock of capital equipment, v the

real stock of household wealth, and where asterisks denote equilibrium
values which are functions of appropriate flow and other variables.
The intuition is that in the full steady state no changes in either real
stocks or flows are occurring, and hence, a fortiori, the real flow of
government payments must equal real receipts.

g + rrb · b–1 = t + rrh · h–1 (2.25)

Assuming for simplicity that real interest payments are taxed

t = θ · (y + rrb · b–1 ) (2.26)

y ∗ = [g − rrh · h–1 + rrb · b–1 · (1 − θ )]/θ (2.27)

the full steady state becomes

y ∗ = [g − rrh · h–1 + rrb · b–1 · (1 − θ )]/θ (2.28)

A graphical representation of this equilibrium has more expository

power than a very long algebraic expression. In Figure 2.3 I have (for
once) ‘netted out’ the banking sector and assumed the inflation rate to
be constant (not necessarily zero).
The figure represents a stationary steady state where all real stocks and
flows are constant. It is assumed that government expenditure, and any
one out of rrb∗ , h∗ or b∗ are exogenous. The SE quadrant shows how steady
state output y ∗ is determined at the intersection of government outflows
and government receipts. The NW quadrant shows at v ∗ the total stock
of wealth as determined by the consumption function in a stationary
state. It also shows how the stock of real wealth is allocated between
equities and cash (both decreasing functions of the real rate of interest)
and bonds (an increasing function of the real rate). The radiating line
in the SW quadrant is the product of the equilibrium rate of interest
and the equilibrium stock of bonds. The real interest flow is determined
 Time, Increasing Returns and Institutions in Macroeconomics 57

m k r
k* b*
r*

m*

v* y*
Assets y = Real output
r *b*
g
rr b*.b*

t = θ (y* + r *b *) + rr h .h
Govt flows

Figure 2.3 A full steady-state of a real stock–flow model

in the SW quadrant by the vertical line running south from the inter-
section of the b∗ function with rr h * in the NW quadrant. The horizontal
line describing the real interest flow extends eastwards from the SW into
the SE quadrant.
As will be apparent from Equation (2.27) the rate of inflation (rr h ) is an
independent argument of the expression determining the output flow. In
normal times, since rr h and h are both such small numbers their product
will be negligible. In periods of hyperinflation, however, the ‘inflation
tax’ will dominate the expression and make it useless.
While this deployment concentrates on the hypothetical notion of a
stationary steady state, the economy should be thought of as being in a
state of multiple adjustment towards (constantly changing) equilibrium
stocks and flows. Each short period opens with stocks of real human,
tangible and financial wealth distributed among agents and institutions
for reasons inherited from the past; these encapsulate history. Decisions
by governments, corporations and banks to produce, employ, set prices,
lend and so on interact sequentially with households’ decisions to work,
consume and accumulate stocks of wealth. Flows generated by these
interactions in each period generate, in turn, end period stocks which
constitute the starting-point for the next period. The state of multiple
adjustment is not a disequilibrium in any useful sense. Rather the move-
ment in the direction of steady-state values should be thought of as itself
taking place at desired rates. Time lags are not contingent; they are part
of the model.
It may be objected that the steady state will be long and potentially
difficult of achievement because so much is going on and the stabil-
ity conditions are so complex. I have demonstrated elsewhere that the
mean lag in the response of the entire system of changes in its exoge-
nous variables is determined by equilibrium stock–flow ratios alone and
is completely independent of all the individual adjustment processes.
58 Early Views on the Stock–Flow Coherent Approach

Specifically, it can be demonstrated that the mean lag approximates

(gd ∗ /y ∗ )
closely to θ where gd ∗ is the steady-state stock of government debt
∗ ∗
(= v – k ). This expression may be evaluated (at least for the UK) to be in
the region of 1–2 years. Extensive simulation experiments have led me
to conjecture that over a wide range of parameters the dynamics of the
system will not be such as to undermine the usefulness of the mean lag
theorem; in other words the movement of the whole economy towards
its steady-state value (which can be relatively simply characterized) is so
rapid that we can retain some intuitive understanding of what is going
on despite the apparent complexity of the model as a whole.

Conclusion

Some of the main features of the model outlined above may be listed.

(1) The RSFM model is conceived as evolving sequentially through real

time.
(2) The function of price determination in macroeconomic analysis is
not to clear markets in goods and labour but to distribute income,
with inflation resolving the rival claims of various sectors. Inflation
is, however, a contingent process influenced by, among other things,
the opportunities open for shares to be changed.
(3) Industrial corporations, operating under conditions of imperfect
competition and dynamically increasing returns to scale, aim to max-
imize their own growth rates and market shares. Their decisions with
regard to investment, pricing, and dividend distribution are taken
with this objective in view.
(4) As production and investment take time, firms have a requirement
for finance beyond what can be acquired from retained profits and
from issues of securities particularly when the economy is growing.
This finance is created by commercial banks.
(5) A key factor determining the sequences is the way in which real flows
are generated and interact with one another to create real stocks.
Hence in the closed economy case fiscal policy has an all-important
role.
(6) If the economy is opened, exports, imports and international trans-
fers must be included with the other flows, and (it has to be
admitted) crucially important problems then arise which I have not
discussed here. In the long period it will be the success or failure
of corporations, with or without active help from governments, to
compete in world markets which will govern the rise and fall of
nations.
 Time, Increasing Returns and Institutions in Macroeconomics 59

Notes
1. This paper is a revised and condensed version of three lectures called ‘Time,
Credit Money and the Neoclassical Synthesis’ Which I gave in Naples in April
1988 at the invitation of Professor Augusto Graziani. It is a summary of a
substantial monograph in the course of preparation in collaboration with Ken
Coutts. I am grateful for penetrating comments from K. Coutts, A. Graziani,
G. Harcourt, S. Mundle, N. Norman and G. Zezza.
2. ‘The primal scene’ is a technical term in psychoanalysis; it is the imagi-
nary perception, postulated by Freud, by the infant of its own parents at
intercourse.
3. Albert Eichner was bravely trying to do this when he died two years ago.
4. As mentioned earlier, a monograph is in the course of preparation. The model
to be deployed has already been tested by simulation experiments.
5. Defined as the change in the value of inventories (IN) less the value of the
change (in · UC) when UC is unit wage cost.
6. For a full derivation of this expression, together with the way in which it can
be used to derive the distribution of the real national income, see Godley and
Cripps (1983), Appendix to Chapter 9. It may be useful to note that output is
equal to sales plus the change in inventories (y = s + in).
7. Equation (2.12) has a strong affinity in both form and meaning with that
proposed by Graziani in, for example, (1989). Graziani has no σ but only
because he defines the period in question as the production period. In the
Graziani system there is a three-way split between profits, wages and the real
return on money.
8. It is one of many crosses (the pun belongs to Robert Clower) elementary
students must bear that the first equilibrium they learn is the stock–flow equi-
librium of the Keynesian cross. This, in every textbook I know of, passes by
sheer prestidigitation into the IS-LM equilibrium, where markets are cleared
by prices and of which the nature is therefore totally different.
9. Even where firms are working at such high rates of utilization that marginal
costs rise sharply, they prefer to lengthen order books rather than turn down
sales.
10. This is an ancient problem which becomes invisible in the neoclassical world
precisely because in that world production takes place instantaneously. It used
to be a real cause for concern that wages couldn’t be paid until the harvest
was gathered and sold.
11. Actually in a world of increasing returns and imperfect competition, growth
maximization may not really be different from profit maximization. Accord-
ing to Kaldor (1980), ‘It is on account of the economies of large scale
production that a rising market share means success and a falling market
share spells trouble. And it is on that account that in a growing market a
business cannot stand still: it must grow if it wishes to survive.’
12. The relationship between φ and ϕ in Equation (2.15) is, of course ϕ = /(1+φ).
13. The formal conditions required to make such a maximum possible have been
derived by Ken Coutts and will be given in our forthcoming monograph. The
necessary conditions are not very strong ones.
14. Equations similar to Equation (2.23) are commonly to be found in economet-
ric models although there is usually held to be a lag between changes in wage
60 Early Views on the Stock–Flow Coherent Approach

costs and changes in prices. I think it implausible that changes in prices sys-
tematically lag behind costs. However it can be inferred from Equation (2.11)
that Equation (2.23) is identically equal to:

p = (1 + τ ) · (1 + ) · [(1 − σ ) · UC + σ · (1 + r) · UC–1 ]

I believe it is because the rate of interest is not normally included in the econo-
metric specification that a lag (for which there is no theoretical justification)
usually turns up in equations of this kind.
15. It may seem strange that it should be necessary to emphasize this. How-
ever, there is virtually no reference to (private) loan expenditure in the
NCP macroeconomic literature. For instance, Modigliani in his celebrated
1963 model provides a mock-up of a banking system but its only assets are
high-powered money and government bonds.

References

Blinder, A.S. and R. Solow (1973) ‘Analytical Foundations of Fiscal Policy.’ In

A.S. Blinder, R. Solow et al.(eds), The Economics of Public Finance (Washington:
Brookings Institution).
Christ, C.F. (1967) ‘A Short-Run Aggregate Demand Model of the Interdepen-
dence of Monetary and Fiscal Policies with Keynesian and Classical Interest
Elasticities.’ American Economic Review, May.
Clower, R.W. (1985) ‘The Keynesian Cross Revisited.’ Working Paper, 383,
Department of Economics, University of California.
Dornbusch, R. and S. Fischer (1984) Macroeconomics, 3rd edition (Tokyo:
McGraw-Hill).
Fischer, S. (1987) ‘1944, 1963 and 1985.’ In R. Dornbusch, S. Fischer, J. Bossons
(eds), Macroeconomics and Finance (Cambridge, Mass.: MIT Press).
Friedman, M. (1968) ‘The Role of Monetary Policy.’ American Economic Review 78
(March).
Godley, W. and F. Cripps (1983) Macroeconomics (London: Fontana).
Graziani, A. (1989) ‘The Theory of the Monetary Circuit.’ Thames Paper, 12.
Hicks, J. (1937) ‘Mr Keynes and the “Classics”: A Suggested Interpretation.’
Econometrica 5.
—— (1980–1981) ‘IS-LM: An Explanation.’ Journal of Post Keynesian Economics
3 (January).
—— (1989) A Market Theory of Money (Oxford: Basil Blackwell).
Kahn, R. (1931) ‘The Relation of Income Investment to Unemployment.’ Economic
Journal 41 (June): 173–198, reprinted in Selected Essays on Unemployment and
Growth (Cambridge: Cambridge University Press, 1972), pp. 1–27.
Kaldor, N. (1939) ‘Speculation and Income Stability.’ Review of Economic Studies
7: 1–27, reprinted in Essays on Economic Stability and Growth, Vol. II (London:
Duckworth, 1960).
—— (1980) ‘The Foundations of Free Trade Theory and Their Implications
for the Current World Recession.’ In E. Malinvaud and J.-P. Fitoussi (eds),
Unemployment in Western Countries (London: Macmillan), pp. 85–100.
Leijonhufvud, A. (1983) ‘What Was the Matter with IS–LM?’ In Fitoussi J.-P. (ed.),
Modern Macroeconomic Theory (Oxford: Basil Blackwell).
 Time, Increasing Returns and Institutions in Macroeconomics 61

Lucas, R.E. (1973) ‘Some International Evidence on Output-Inflation Tradeoffs.’

American Economic Review 63 (June).
Malinvaud, E. (1977) The Theory of Unemployment Reconsidered (Oxford: Basil
Blackwell).
Modigliani, F. (1944) ‘Liquidity Preference and the Theory of Interest and Money.’
Econometrica 12.
—— (1963) ‘The Monetary Mechanism and Its Interaction with Real Phenomena.’
Review of Economics and Statistics 45.
Pasinetti, L.L. (1974) Growth and Income Distribution (Cambridge: Cambridge
University Press).
Patterson, D.M. (1988) ‘Stock-Flow Consistent Accounting: A Macroeconomic
Perspective.’ Economic Journal 98 (September).
Sargent, Th. (1987) Macroeconomic Theory (San Diego: Academic Press).
Sylos Labini, P. (1988) ‘The Great Debates on the Laws of Return and the Value of
Capital: when Will Economists Finally Accept Their Own Logic?’ BNL Quarterly
Review 166 (September).
Tobin, J. (1980) ‘Money and Finance in the Macroeconomic Process.’ Journal of
Money, Credit and Banking 14.
Tobin and W. Buiter (1976) ‘The Long Run Effects of Fiscal and Monetary Policy on
Aggregate Demand.’ In J.L. Stein (ed.) Monetarism (Amsterdam: North-Holland
Press).
Wood (1979) A Theory of Profits (Cambridge: Cambridge University Press).
Part II
Stock–Flow Coherence and
Economic Theory
3
An Important Inconsistency at the
Heart of the Standard
Macroeconomic Model
Wynne Godley and Anwar Shaikh

The problem stated

The standard neoclassical model is the foundation of most mainstream

macroeconomics. Its basic structure dominates the analysis of macroeco-
nomic phenomena, the teaching of the subject and even the formation of
economic policy. And, of course, the modern quantity theory of money
and its attendant monetarist prescriptions are grounded in the model’s
strict separation between real and nominal variables.
It is quite curious, therefore, to discover that this model contains an
inconsistency in its treatment of the distribution of income. And when
this seemingly small discrepancy is corrected, without any change in all
of the other assumptions, many of the model’s characteristic results dis-
appear. Two instances are of particular interest. First, the strict dichotomy
between real variables and nominal variables breaks down, so that, for
example, an increase in the exogenously given money supply changes
real variables such as household income, consumption, investment, the
interest rate, and hence real money demand. Second, since the price
level depends on the interaction of real money demand and the nom-
inal money supply, and since the former is now affected by the latter,
price changes are no longer proportional to changes in the money sup-
ply. Indeed, we will demonstrate that prices can even fall when the
money supply rises. The link to the quantity theory of money, and to
monetarism, is severed.
In its most basic form, the model encompasses four ‘markets’: com-
modities, labour, private bonds and money.1 These arenas are bound
together by the (implicit) household and business sectors’ budget con-
straints, which link what agents plan to spend with what they expect
to receive. When cast in Walrasian terms, these budget constraints

65
66 Stock–Flow Coherence and Economic Theory

aggregate into the familiar expression known as Walras’s Law, which

states that the sum of the planned demands for the four items must
equal the sum of their expected supplies – that is, that excess demands
in the four arenas must sum to zero (Buiter 1980; Clower 1979). This lat-
ter result is then used to justify the dropping of any one market from the
formal description of the model, on the grounds that equilibria (or even
particular disequilibria) in any three determine the state of the fourth.
In the standard form depicted in Equations (3.1) through (3.11) of the
next section, it is the bond market that drops out of view (McCafferty
1990, p. 46).
As is well-known, the standard model exhibits a block recursive struc-
ture beginning from equilibrium in the labour market and moving to
real output demand and its components, including the real demand for
money, and ending finally in nominal wages and prices. The price level
in particular is determined by the conjunction of the real demand for
money and a given nominal money supply. Since the former is a func-
tion of real variables such as output and the interest rate, and since the
block recursive structure implies that real variables are unaffected by the
money supply (because they are analytically upstream of nominal rela-
tions), it follows that doubling the money supply must double prices
so as to keep the real money supply equal to an unchanged real money
demand. This is acknowledged to be an absolutely central result of the
model (McCafferty 1990, p. 53). Yet it turns out to be very generally false.
The source of the problem lies in the apparently innocuous assumption
that all of the real net income of the business sector (the real value of the
net product) is somehow distributed to households. In the case of wage
income, this is straightforward, since firms pay workers for their labour
services. But when we ask how profits are to be distributed, we find that
within the logic of the model they can only be distributed in the form
of interest payments on the bonds issued by firms, for there is no other
instrument available in the model. Firms borrow money from households
by issuing bonds, and are then obliged to pay interest on them at the
rate determined by the model. The difficulty is that these aggregate real
interest payments will generally differ from aggregate real profits. This
in turn implies that household income (wage and interest income) must
generally differ from business income (wages and profits).
It is a simple matter to correct the model by explicitly writing real
household income as the sum of real wage and interest income (the
latter being the interest rate times the real value of bonds). On the side
of businesses, this implies that the value of new bonds issued by firms
(their new borrowing) in a given period can differ from the value of
 Standard Macroeconomic Model 67

the investment expenditures they plan to make, precisely because their

total out payments to households can differ from their own net income.
Budget constraints, after all, only require that the overall sum of inflows
equal overall outflows. With these minor changes, the model becomes
consistent.
But, although the correction appears minor, its consequences are not.
The full employment core of the original model is preserved, so that real
wages, employment and output continue to be the same. This means
that real profits are also unaffected. But now a change in the price level
(due, say, to a change in the money supply) changes the real value of
bonds outstanding, and hence changes the level of real interest flows.2
Since real interest flows enter into household income, this affects real
consumption demand, real investment demand (which is the difference
between the unchanged real output and changed consumption demand),
and the interest rate (which must adjust to make real investment demand
come out right). Because real money demand is affected in opposite ways
by real household income3 and the interest rate, both of which change
in the same direction, its overall direction of change is ambiguous. It can
rise or fall in the face of an increase in the money supply so that prices
can change less or more than the money supply. This property alone is
sufficient to sever any simple linkage between the two. As noted earlier,
we can show that even under perfectly plausible parameter values, prices
can actually fall when the money supply increases.
The problem that we have identified is noted in passing in Patinkin’s
(1965) seminal text, but is then buried in footnotes. In an effort to
maintain a forced equality between aggregate household income and
aggregate value added, he is driven to make a series of ad hoc behavioural
assumptions. He does not remark on the contradictions to which
these give rise. We comment on his proposed solutions in the section
‘Patinkin’s attempts to grapple with the issue’.
One implication of our results is that the bond market can no longer be
‘dropped’ out of the story. This is because real interest payments depend
on the number of bonds, which requires us to deal explicitly with the
determinants of this quantity. It is true, of course, that Walras’ Law still
allows us to infer the state of excess demand in the bond market from
that in the other three arenas. But this implicit relation between the sup-
ply and demand for bonds does not in itself allow us to determine their
respective levels. For that, and hence for the determination of real inter-
est flows, the bond market becomes a structurally necessary part of the
model. This is possible because a description of the bond market actually
requires two conditions: Walras’s Law, which in this model reduces to the
68 Stock–Flow Coherence and Economic Theory

requirement that the bond market be in equilibrium; and an investment

finance constraint for the firm, which provides us necessary addi-
tional equation. We will see that these two conditions derive from the
implicit budget constraints of the household and business sectors (Buiter
1980).

A formal exposition

The standard neoclassical macroeconomic model

We start with the standard exposition of the model, elaborated to as to
make explicit its underlying assumption that household income is identi-
cal to value added – that is, that profits are always completely distributed.
Thus, we explicitly express consumption and money demand functions
in terms of household income (Equations (3.4) and (3.6)), and then
add the condition that household income equals value added (Equation
(3.11)). This has no effect on the results at this stage in the argument,
but it does prepare us for what follows. In general, lowercase refers to
real and uppercase to nominal variables.

Theory of the firm

ys = f (k, Nd ) [aggregate production function, (3.1)
with given real capital stock k]
Nd = Nd (W/p) [p = mc, where mc = W/mpl, (3.2)
mpl = f (Nd ) from short-run
profit-maximizing]
id = id (r) [id (r) = investment demand] (3.3)

Theory of the household

cd = cd (yh ) [consumption function, from (3.4)
utility-maximizing behaviour]
Ns = Ns (W/p) [labour supply of households, from (3.5)
utility-maximizing behaviour]
Md /p = md (yh , r) [money demand function of households, (3.6)
from optimal portfolio formation]
Definitions and equilibrium conditions
yd = c d + i d [definition of aggregate demand] (3.7)

yd = y s [commodity market equilibrium] (3.8)

Nd = N s [labour market equilibrium] (3.9)

 Standard Macroeconomic Model 69

Md = M [money market equilibrium, the money stock M (3.10)

being taken as given]
Distribution condition
yh = y s [household income assumed to equal value added, (3.11)
that is, all profits are distributed]

where, respectively, yd and ys are real commodity demand and supply,

Nd and Ns are labour demand and supply, yh is real household income,
cd and id are real consumption and investment demand, Md is nomi-
nal money demand, r is the real (and nominal) interest rate, W and p
are nominal wages and profits, and M is the exogenously given money
supply.
Note that we have 11 endogenous variables defined above (M being
exogenous), and 11 independent equations.
A fundamental characteristic of the model is that it is block recursive.
Thus, Equations (3.2), (3.5) and (3.9) determine the equilibrium real
wage (W/p)* and real employment N*, and through Equations (3.1) and
(3.8) the latter determines real output and real demand y*. The preceding
variables then determine equilibrium household income yh *, consump-
tion c*, investment i*, the interest rate r*, and real money demand
(Md /p)∗ = m∗d = md (y ∗ , r ∗ ), by means of Equations (3.3), (3.4), (3.6), (3.7)
and (3.11). This last variable, in conjunction with the given money sup-
ply M and Equations (3.6) and (3.10) allows us to determine nominal
money demand Md = M, the nominal price level p = Md /md (yh∗ , r ∗ ), and
the nominal wage W = p · (W/p)*. The significance of block recursion
is that equilibrium values of downstream variables have no effect on
those of upstream ones. Therefore, a change in the supply of money M
must change the equilibrium price level p in the same proportion and
direction, because p = M/m∗d , and the equilibrium real output y ∗ and
interest rate r ∗ , which determine equilibrium real money demand m∗d ,
are upstream of p (and independent of M). It is this particular property that
is the foundation for the monetarist aspect of the model. And it is precisely this
property that does not survive.

Finding the bond market

Although interest rates play an important role in the operations of the
model, there is no representation of interest payments. Where the subject
is mentioned at all, it is generally dismissed on the grounds that Wal-
ras’s Law allows us to drop the bond market out of explicit consideration
(Barro 1990, p. 108; McCafferty 1990, p. 46; Modigliani 1963, p. 81;
Patinkin 1954, p. 125; 1965, p. 230). But Walras’s Law only permits us
70 Stock–Flow Coherence and Economic Theory

Table 3.1 The ex ante flow of real funds

Households Firms Total

Consumption and −cd −id −yd = −(cd + id )

investment
Sales Ys ys
Wages (W/p) · Ns −(W/p) · Nd −(W/p) · (Nd − Ns )
Financial payments int e int T (int e − int T )
Changes in bonds −(pb /p) · (bd − b0 ) (pb /p) · (bs − b0 ) −(pd /p) · (bd − bs )
Changes in money −(Md − M)/p −(Md − M)/p
Total 0 0 0

to deduce that there will be equilibrium in the bond market if the other
three markets are in equilibrium. It does not tell us what the equilibrium
quantity of bonds, and hence what the equilibrium level of interest pay-
ments, will be. Most important, it does not permit us to drop the flow of
interest payments out of sight.
The issues involved can be brought into focus by considering the ex
ante budget constraints that underlie the whole model, because then we
are forced to explicitly account for the planned uses and expected sources
of funds (including borrowing) for each sector. In Table 3.1, each column
represents a particular sector’s uses (negative signs) and sources (positive
signs). If sectors are consistent in making their plans,4 each column, and
hence the overall sum of columns, must sum to zero.
The row sums of the matrix are another matter, since they represent
the discrepancy between ex ante expenditures planned on a particular
activity by a given sector and the ex ante receipts expected from the
same activity by another sector. There is no reason here for individual
rows to sum to zero, since plans by one sector need not match antici-
pated receipts by another. All that is required is that the overall sum of
the rows be zero,5 since this is merely the overall column sum. The lat-
ter requirement implies that ex ante discrepancies must add up to zero,
which in this context is simply Walras’s Law.
In Table 3.1, flows are presented in real terms, and the initial number
of bonds is denoted by b0 (so that bd − b0 represents the change in bond
holdings desired by households, and bs − b0 represents the change in
bond issue expected by firms). Of crucial significance are the yet undefined
flows of real financial payments int e expected by households and int T
planned by firms. The flow of funds matrix implies that in addition to
the equations of the model there are two further equations implicit in the
 Standard Macroeconomic Model 71

model. We can derive these equations from any two of the three column
sums in the model (since the third is just the sum of the first two). Taking
the firms’ and total columns give us the most familiar results.
Thus, if we take the column sum for firms, recognizing that ys −(W/p)·
Nd = real profits = f , and that f − int T = undistributed profits, we find
that the sectoral budget constraint of firms is equivalent to an investment
finance constraint, which says that the real value of new bonds issued
must equal the excess of investment needs over undistributed profits.

(pb /p) · (bs − b0 ) = id − [ys − (W/p) · Nd ] − int T = id − (f − int T )

[investment finance constraint] (3.12)

For the other equation we take the total column sum (and reverse
signs), which gives us an expression recognizable as Walras’ Law
(Equation (3.13)), except for the presence of the yet undefined financial
payments flows. Indeed Equation (3.13) is exactly the form of Walras’
Law that Buiter (1980) derives.6 We will return to that point shortly.

(yd − ys ) + (W/p) · (Nd − Ns ) + (Md − M)/p + (pb /p) · (bd − bs )

− (int e − int T ) = 0 [Walras’ Law] (3.13)

Real financial payments appear in both of the preceding relations.

But what determines them? The answer lies in the fact that the model
assumes that firms issue new bonds, in which case they must also pay
interest on these same bonds. Since bonds are the only instruments for the
disbursement of profits, these interest flows are the only financial payments
dictated by the logic of the model. If, in a Walrasian spirit, we assume that
borrowing is planned at the beginning of the period and that the corre-
sponding interest rate flows are expected during that same period, and
if we note that the price of bonds pb = 1/r, then7 :

int e = interest payments expected by households

= r · (pb /p) · bd = bd /p = real value of bonds demanded.

int T = interest payments planned by firms

= r · (pb /p) · bs = bs /p = real value of bonds supplied (3.14)

Substituting the expressions for real financial payments (Equation

(3.14)) into Walras’s Law (Equation (3.13)) allows us to combine the
resulting bond market terms into one expression concerning excess
demand in the bond market: (pb /p) · (bd − bs ), where pb = pb · (1 − r) =
the net price of bonds. Note that the three equilibrium conditions in
72 Stock–Flow Coherence and Economic Theory

Equations (3.8) through (3.10), along with Walras’s Law in Equation

(3.13), imply the bond market equilibrium condition bd = bs . With this
elaboration, the model is completely specified.
The trouble is that now the overall model, built around the familiar
core in Equations (3.1) through (3.11) from which all the standard results
derive, is inconsistent. This is because the standard form assumes that
household income yh = the value of net output y = wages + profits. But
in actuality, yh = wages + interest payments = (W/p) · Ns + r · (pb /p) · bd =
(W/p) · Ns + bs /p, so the two expressions for yh are not equivalent because
real interest payments will not generally equal real profits. The former is deter-
mined in the bond and money markets, and the latter is determined by a
given capital stock and the full employment marginal product of capital.
They would be equal only by accident.
Removing the inconsistency is straightforward. One only has to sub-
stitute the second, proper, expression for yh into what was formerly
Equation (3.11) of the original model. The consistent model then
consists of Equations (3.1) through (3.10), the corrected definition of
household income (Equation (3.11’)), Equations (3.12) and (3.13) mod-
ified to reflect the definitions of financial payments in Equation (3.14)
into account, and an explicit definition of bond price pb :

yh = wages + interest payments = (W/p) · Ns + bs /p

[household income] (3.11’)
(pb /p) · (bs − b0 ) = id − ([ys − (W/p) · Nd ] − r · (pb /p) · bs )
[investment finance constraint] (3.12’)

(yd − ys ) + (W/p) · (Nd − Ns ) + (Md − M)/p + (pb /p) · (bd − bs ) = 0

[Walras’ Law] (3.13’)

where pb = pb · (1 − r) = net price of bonds.

pb = 1/r (3.14’)

Now the model is consistent. But its behaviour is substantially dif-

ferent. This is because household income depends on the real value of
interest payments, which means that a rise in the money supply affects
both the price level and the level of real household income (through
the real value of interest flows, in Equation (3.11’)). Complex interac-
tions then become possible (see the Appendix). For instance, it becomes
possible for a rise in the money supply to raise real household income.
This would in turn raise real consumption and ceteris paribus, also raise
 Standard Macroeconomic Model 73

Table 3.2 Simulated price and real variable changes in the face of an increase in
money supply

M y yh b c i r W p

3.8 0.981 0.981 1.087 0.589 0.393 0.172 3.934 2.768

4.2 0.981 0.965 0.943 0.579 0.402 0.044 3.558 2.504
(+10.5%) (0%) (−1.6%) (−13.2%) (−1.7%) (+2.3%) (−25.56%) (−9.6%) (−9.5%)

real money demand (Equations (3.4) and (3.6)), both of which depend
positively on real household income. Because real output, and hence
aggregate demand (Equation (3.8)) is unaffected, the fact that consump-
tion demand has risen implies that real investment demand must fall
and hence the interest rate must rise. Therefore a rise in the money supply
can raise the interest rate and ‘crowd out’ investment.
Real household income and the interest rate move together but have
opposite effects on real money demand (Equation (3.6)), so the overall
effect is ambiguous. But the important point is that real money demand
md (yh , r) generally changes when the money supply changes. Since the
price level p = M/md (yh , r), this means that neither the magnitude, nor
even the direction, of price changes is a simple reflection of changes
in the money supply. The Appendix shows that some real effects can
be substantial, and that prices can even fall when the money supply
increases. This latter case is illustrated in Table 3.2.

Patinkin’s attempts to grapple with the issue

The crux of the problem arises from the fact that within the logic of
the neoclassical model, profits and real interest payments are differently
determined and hence will not generally be equal. The standard form
of the model, in which these two flows are simply assumed to be equal,
produces a system that is over-determined and hence generally incon-
sistent. This difficulty can be resolved by making the two flows distinct,
which renders the model consistent. But then its standard results, par-
ticularly those pertaining to the so-called dichotomy between real and
nominal variables, and to the putative effects of a change in the money
supply, no longer hold.
Conversely, the standard results require that real business financial out
payments int T = real profits mpk ·k at all times, where financial payments
at least encompass real interest flows r · pb · bs /p. Only then will house-
hold income yh = net value added y, and the value of newly issued bonds
74 Stock–Flow Coherence and Economic Theory

equals the value of investment (from Equation (3.12)). Since all the rel-
evant variables are either given exogenously or determined within the
model, one must propose an additional mechanism to bring about the desired
result. We will see that this is precisely what Patinkin attempts to do.
Throughout his text, Patinkin (1965) assumes that all profits will be
automatically distributed. But the problems we have raised also seem to
have troubled him, because he does make an attempt, albeit very cur-
sory, to justify this crucial assumption. He notes that the assumption of
the full distribution of profits requires the further assumption that any
excess of profits over interest payments is ‘appropriated by entrepreneurs’
(Patinkin 1965, p. 201), which would then ensure that total financial out
payments by firms int T = real profits mpk · k. Nowhere does he even men-
tion the fact that the difference between profits and interest payments
can be positive or negative, which would require entrepreneurs to always
pay themselves bonuses in the first case, and always assess themselves
penalties in the second. Moreover, he does not note that if entrepreneurs
did happen to behave in such a manner, the excess profits they paid
themselves would be taken from funds that would otherwise be used
for investment, and that they have to be made up by extra borrowing by
their firms. They would simply be robbing Peter to pay Paul. The implicit
behavioural assumptions become even more strained when one consid-
ers the case in which interest flows exceed profits, for then entrepreneurs
must be supposed to reduce their own incomes (via a penalty) so as to
make up the difference. But most important of all, there is absolutely
no motivation within the model’s own microfoundations for any such
behaviour. Given Patinkin’s emphasis (and that of neoclassical macroe-
conomics in general) on the importance of microfoundations, this is very
telling indeed.
One implication of the assumed automatic full disbursement of profits
is that firms must finance investment entirely through borrowing in the
bond market (Equation (3.12) in the case where undistributed profits
f − int T = 0). This in turn implies that in both real and nominal terms
the total value of bonds equals the value of the stock of capital. Just a
few pages later, Patinkin runs headlong into the further problems caused
by this assumption. And once again, he is forced to make another set of
ad hoc assumptions in order to keep these new difficulties at bay. In the
course of a discussion of the effects of a doubling of the money supply,
Patinkin derives the familiar result in which nominal variables (W, p)
are doubled, but real variables such as output y, the interest r (and hence
bond price pb = 1/r), and the real money supply M/P are unchanged. The
real value of the planned bond supply pb · bs /p has been assumed to be a
 Standard Macroeconomic Model 75

function of these real variables, so it too must be unchanged. But with

pb unchanged and p doubled, it must then be the case that the number
of bonds issued by firms bs must somehow double as nominal variables
double (Patinkin 1965, pp. 216–217). So, in a footnote, he says: ‘There is
an implicit assumption here that all the firms’ capital equipment must
be replaced during the period in question’ (ibid., p. 217, footnote 13).
But what can it mean that the firm’s capital equipment must be
‘replaced’, and how could this resolve the present difficulty? The answer
lies in recognizing that with y and r unchanged, real net investment is
unchanged. But with p doubled and real investment unchanged, nomi-
nal investment is doubled. Thus, firms will have to issue a new quantity
of bonds equal to the changed nominal value of new investment. How-
ever, with the price level doubled, the nominal value of new capital will
also have doubled, so if firms are to maintain a stock of bonds equal to
the value of the capital stock, as required by the distributional assumption,
they must sell a quantity of new bonds equal to the changed nominal
value of the capital stock. These two distinct requirements are generally
inconsistent.
One step towards rendering the two distinct financial relations con-
sistent is to assume that all capital turns over in one period,8 so that
real investment and the real capital stock are always equal. Then, with
i = k = k, if firms issue new bonds to finance new investment (pb · bs =
p · i), then this will also ensure that the change in the nominal value
of bonds will match the change in the nominal value of the given cap-
ital stock (pb · bs = p · k). Then, if the initial value of bonds equalled
the initial value of the capital stock, this equality would be maintained
throughout as long as the capital stock turned over completely in each
period. It should be noted that in this case bonds would also have to be
one-period bonds with a price pb = 1/(1 + r), not the consols with a price
pb = 1/r, which Patinkin assumes throughout.
Understandably uneasy about the previous solution, Patinkin proposes
an alternative one:
Alternatively, we can assume that firms immediately write up their
capital equipment in accordance with its increased market value, sell
additional bonds to the extent of this increased value, and pass on the
explicit capital gains to their respective entrepreneurs. Conversely, in
the event of a decrease in prices, entrepreneurs must make good the
implicit capital loss, and firms then use these funds to retire bonds.
In this way the nominal amount of bonds outstanding can always be kept
equal to the current value of the firms’ assets. (Patinkin 1965, p. 217,
n. 13, emphasis added)
76 Stock–Flow Coherence and Economic Theory

Recall that the crux of the problem is that the assumed automatic dis-
tribution of profits requires that the nominal value of bonds remain equal
to the nominal value of the capital stock. So now Patinkin abandons the
bedrock assumption that firms issue bonds to finance new investment
in favour of the assumption that they instead issue or retire bonds
to match changes in the nominal value of the existing capital stock:
pb · bs = p · k > p · i, since in general k > i.
A simple numerical example illustrates the difficulty facing Patinkin.
Suppose that initially pb = 5, p = 1, i = 10, k = 100, and that a change
in the money supply produces p = 1. Then if new bonds are issued to
finance the changed value of new investment, pb · bs = p · i = 10, so
bs = 2. Alternatively, if new bonds are issued to realize capital gains on
the stock of capital, pb · bs = p · k = 100, so bs = 20. The two solutions
are inconsistent unless one assumes that all capital turns over in one
period (k = i at all times), or one abandons the notion that firms issue
bonds to finance nominal new investment in favour of the assumption
that bonds are issued to ‘pass on the explicit capital gains [from the
increased value of the capital stock] to entrepreneurs.’
In all of these instances, Patinkin’s strained and behaviourally unmoti-
vated assumptions are driven entirely by the need to avoid the contradic-
tions generated by the a priori assumption that household income always
be the same thing as the aggregate net income of firms. This assump-
tion is essential to the derivation of the famed dichotomy between real
and nominal variables. But we have seen that any such forced equality
between household income and aggregate value added is not sustain-
able within the logic of the model. Patinkin’s discussion only confirms
this fact.

Summary and conclusions

Our central finding has been that the famous dichotomy between real
and nominal variables, which emerges from the standard neoclassical
macroeconomic model, rests on extraordinarily shaky foundations. Writ-
ing out the ex ante flow of funds corresponding to the model reveals that
its standard form embodies inconsistent assumptions about the treat-
ment of the distribution of non-wage income. Firms are assumed to
disburse all of the profits, but the only instrument available is the inter-
est on the bonds they have issued. Contrary to the implicit assumption
within the model, the resulting interest flows will not generally equal
profits.
 Standard Macroeconomic Model 77

The revealed inconsistency is easily rectified by distinguishing between

household income (wages and interest payments) and net value added
(wages and profits). But then, leaving all other assumptions unchanged,
the model’s behaviour changes dramatically. In particular, real variables
such as consumption, investment, the interest rate and real money
demand become intrinsically linked to nominal variables such as the
price level and the money supply. One striking consequence is that a
rise in the money supply can actually lead to a fall in prices – even
under the standard assumptions about money demand functions. It fol-
lows that monetarism cannot be grounded in a consistent neoclassical
model.
It should be noted that our main concern here has been to exam-
ine internal consistency of the standard neoclassical macroeconomic
model. Although we do not advocate this model, it is our hope that
our colleagues in the neoclassical tradition will recognize it as a consis-
tent exposition of their own framework and modify their own claims
correspondingly.

Appendix: Numerical simulation of the consistent

neoclassical model
The corrected model:

ys = a · kβ · nd
1−β
(A.1)

mpl = (1 − β) · Nd = W/p (A.2)

id = γ0 − γ1 · r (A.3)
cd = α · yh (A.4)

Ns = σ0 · (W/p)σ 1 (A.5)
Md /p = λ0 + λ1 · yh − λ2 · r (A.6)
yd = cd + id (A.7)
yd = ys (A.8)
Nd = Ns (A.9)
Md = M (A.10)
yh = (W/p) · Ns + (r · pb · bd )/p [household income] (A.11’)
(pb /p) · (bs − bd ) = id − (ys − (W/p) · nd − r · pb · bs /p)
[investment finance constraint] (A.12’)
78 Stock–Flow Coherence and Economic Theory


(yd − ys ) + (W/p) · (Nd − Ns ) + (Md − M)/p + (pb /p) · (bd − bs ) = 0

[Walras’ Law] (A.13’)


where pb = pb · (1 − r) = net price of bonds.

pb = 1/r (A.14’)

We have 14 endogenous variables (ys , Nd , id , cd , Ns , yd , yh , Md , r, W, p,

pb , bs , and bd ) and 14 independent equations. The three equilibrium conditions
and Walras’ Law (Equations (A.8) through (A.10), and A.13) together imply bond
market equilibrium bd = bs .
Parameter values:

a = 0.97 β = 0.4 k = 3.86 γ0 = 0.4054 γ1 = 0.75 a = 0.6

σ0 = 0.4 σ1 = 0.1 λ0 = 0.20 λ1 = 1.65 λ2 = 2.6 b0 = 0.9

Initial values (note that initially values have been chosen so that household
income is initially equal to net value added – that is, all profits are initially
distributed):
M = 3.8

ys = yd = 0.981 Ns = Nd = 0.414 Md = M = 3.8

bd = bs = 1.087 cd = 0.589 id = 0.393
yh = 0.981 [note that yh = ys initially] r = 0.172
pb = 5.81 W = 3.934 p = 2.768

Now, when the money supply rises by 10.5% to M = 4.2, real output and
employment are unchanged, household income changes only slightly (from 0.981
to 0.965), and yet there are substantial changes in the interest rate (it drops from
17.2% to 4.4%), and the price level actually falls by 9.5%.
M = 4.2(+10.5%)

ys = yd = 0.981 Ns = Nd = 0.414 Md = M = 4.2

bd = bs = 0.943 cd = 0.579 id = 0.402
yh = 0.965 r = 0.044 pb = 22.721
W = 3.558 p = 2.504(−9.5%)

Analysis of the consistent model helps us understand how this sort of result can
occur. Equilibrium in the labour market together with the aggregate production
function (Equations (A.1), (A.2), (A.5) and (A.9)) yield equilibrium real output
y ∗ , the real wage bill (W/p)∗ · N ∗ , and real profits f ∗ = y ∗ − (W/p)∗ · N ∗ = mpk∗ ·
k∗ , none of which are affected by nominal changes. Then equilibrium in the
commodity market and its associated relations (Equations (A.3), (A.4), (A.7) and
(A.8)) gives us:

y ∗ = cd∗ + id∗ = α · yh + γ0 − γ1 · r (A.15)

 Standard Macroeconomic Model 79

A comparable result can be derived from money market equilibrium and its
associated conditions (Equations (A.6) and (A.10)).

M/p = λ0 + λ1 · yh − λ2 · r (A.16)

Note that the two derived relations do not reduce to the familiar I-S, L-M pair
because real household income yh is not generally equal to real (full employment)
output y ∗ . The former depends on the real demand for bonds, and it is precisely
this dependence that prevents us from ‘dropping’ the bond market out of sight.
From Equations (A.8) through (A.10) and (A.13’) we get bd = bs = b, so from
Equations (A.12’), (A.14’), (A.7), (A.8), (A.4) and (A.11’),

(1/r) · (b/p − b0 /p) = id − ys + (W/p) · N + b/p

= −cd + yh = (1 − α) · yh

so

(b/p − b0 /p) = r · (1 − α) · yh (A.17)

Since pb = 1/r, r · pb = 1, so from Equation (A.11’), b/p = yh − (W/p)∗ · N ∗ .

Substituting this into Equation (A.17) yields:

[yh − (W/p) · N − b0 /p] = r · (1 − α) · yh

[1 − r · (1 − α)] · yh = [b0 /p + (W/p)∗ · N ∗ ] (A.18)

where, since the propensity to consume α < 1, yh > 0 if r ≤ 1.

Combining Equation (A.18) with each of Equations (A.15) and (A.16) then gives
us two nonlinear equations in 1/p and r,9 whose intersection determines the
equilibrium values of p∗ , r ∗ . Note that the value of the money stock M enters
directly into the equilibrium values via Equation B, as does the initial number of
bonds b0 via both equations.

1/p = [{(1 − r + r · α)] · (y ∗ − γ0 + γ1 · r)/α} − (W/p)∗ · N ∗ ]/b0 (A)

1/p = [(1 − r + r · α) · (λ0 − λ2 · r) + λ1 · (W/p)∗ · N ∗ ]/
[(1 − r + r · α) · M − λ1 · b0 ],
for (1 − r + r · α) · M  = λ1 · b0 (B)

Given the particular linear functional forms used in this appendix, one can
impose restrictions on r (such as, y ∗ > γ0 − γ1 · r > 0 since the right-hand side is
investment demand id , and 1 − r + r · α > 0 since that is necessary for yh > 0, and
so on). There are multiple intersections possible for such nonlinear curves, hence,
multiple possible equilibria. Plotting these curves and their shifts in the face of
changes in the money supply M or in the initial bond stock b0 demonstrates that
the possible effects are quite complex.

Notes
1. The desired holdings of money are counterposed to an exogenous supply of
money, which is not really a market.
80 Stock–Flow Coherence and Economic Theory

2. Real interest payments r · pb · b/p = b/p = the real value of bonds outstanding,
where r = the rate of interest, pb = the price of bonds = 1/r, b = the number of
bonds, and p = the price level.
3. In the standard model, only households hold money. But this is not essential
to our results.
4. Clower (1979, p. 297) calls this assumption ‘a fundamental convention of
economic science.’
5. Sectoral budget constraints imply that individual columns, and hence both
the sum of column sums and the sum of row sums, equal zero.
6. Buiter (1980, equation 14, p. 6) actually lists the financial payments as ‘divi-
dend’ payments expected and planned. This is odd because the model contains
bonds but no equity (were it the other way around, there would be no rate of
interest in the model). In leaving these ‘dividend’ payments unexplained, he
sidesteps the inconsistency that we have identified.
7. An alternate assumption is that interest flows in a given period are on the
stock of bonds inherited from the previous period (b0 ). In this case, int e =
int T = r · (pb /p) · b0 = b0 /p = current real value of the opening stock of bonds.
Then Equation (3.13) takes the familiar form of Walras’s Law, since the term
(int e − int T ) drops out. But the dependence of investment finance on interest
payments (Equation (3.12)), and hence on undistributed profits, still remains.
And so the basic contradiction in the standard model continues to exist.
8. Formally, the number of new bonds issued is given by the investment finance
relation pb · (bs − b0 ) = p · k. In the standard model, with r = 1/pb and i = k
unchanged, a change in the money supply implies pb · bs = p · k in this
particular period alone. Hence, only if capital turns over in one period – that
is, if there is no fixed capital – does this also imply that the outstanding stock
of bonds will have doubled.
9. The first of these is straightforward, and results in Equation (A). For the second,
we get M/p = λ0 + λ1 · yh − λ2 · r = M/p = λ0 + λ1 · [{b0 /p + (W/p)∗ · N ∗ }/{1 − r ·
(1 − α)}] − λ2 · r, which, after rearrangement, yields Equation (B).

References

Barro, R.J. (1990) Macroeconomics, 3rd edition (New York: John Wiley).
Buiter, W.H. (1980) ‘Walras’ Law and All That.’ International Economic Review 21
(1) February: 1–16.
Clower, R.W. (1979) ‘The Keynesian Counterrevolution: A Theoretical Appraisal.’
In P.G. Korliras and R.S. Thorn (eds), Modern Macroeconomics: Major Contributions
to Contemporary Thought. (New York: Harper and Row), pp. 289–304.
McCafferty, S. (1990) Macroeconomic Theory. (New York: Harper and Row).
Modigliani, F. (1963) ‘The Monetary Mechanism and Its Interaction with Real
Phenomena.’ Review of Economics and Statistics 45 (supplement): 79–107.
Patinkin, D. (1954) ‘Keynesian Economics and the Quantity Theory of Money.’
In K. Kurihara (ed.), Post-Keynesian Economics. (New Brunswick, NJ, Rutgers
University Press), pp. 123–152.
——. (1965) Money, Interest, and Prices: An Integration of Monetary and Value Theory,
2nd edition (New York: Harper and Row).
4
Weaving Cloth from Graziani’s
Thread: Endogenous Money
in a Simple (but Complete)
Keynesian Model1
Wynne Godley

Graziani’s thread

One of Graziani’s main themes runs as follows. In order to finance pro-

duction, the entrepreneur must obtain the funds necessary to pay his
workforce in advance of sales taking place. Starting from scratch, he
must borrow from banks, at the beginning of each production cycle,
the sum which is needed in order to pay wages, creating a debt for
the entrepreneur and, thereby, an equivalent amount of credit money,
which sits initially in the hands of the labour force. Production now takes
place and the produced good is sold at a price which enables the debt
to be repaid inclusive of interest, while hopefully generating a surplus –
that is, a profit – for the entrepreneur. When the debt is repaid, the
money originally created is extinguished. An entire monetary circuit is
now complete.
This account of the monetary circuit has a number of extremely impor-
tant and distinctive features. It emphasizes, in particular, that a) there is
a gap in (historical) time between production and sales which generates
a systemic need for finance; b) bank money is endogenously determined
by the flow of credit and c) total real income must be considered to be
divided into three parts – that received by entrepreneurs, that received
by labour and that received by banks. We have already travelled an infi-
nite distance from the (yes, silly) neoclassical world where production
is (must be) instantaneous, where money must be exogenous and fixed
and has no counterpart liability, and where the distribution of income is
determined by the marginal products of labour and capital – a construc-
tion which depends entirely on the assumption that all firms sit peren-
nially on a single aggregate neoclassical production function frontier.

81
82 Stock–Flow Coherence and Economic Theory

Weaving cloth from this thread

In what follows there is not one breath of criticism of the Graziani con-
struct, which is at once simple, elegant and fruitful. What I propose
to do here is adapt the model so that the main insights (as I under-
stand them) are carried across into a world where aggregate production
is a continuous set of overlapping individual processes and in which the
production period can vary. I shall fill out a whole macroeconomic frame-
work where, in a number of sequences, various stock variables (money,
debt and inventories) generate and are generated by flows (incomes and
expenditures). All of Graziani’s insights are retained.
I start with a transactions matrix which defines all the current price
flows (occurring in some given period of time) to be used in the model
and which describes the accounting relationships between them. I must
make it clear that this model is far too simple to be realistic. There is, for
instance, no international trade, no government, no financial asset other
than credit money and no fixed investment. I am making the smallest
possible model capable of embodying the key features I wish to illustrate.
It is always possible to add more and more realistic features, but at the
cost of dramatically increasing variables and equations; but this would
not advance my present purpose.
The matrix (see Table 4.1), following the methodology advocated by
Backus, Brainard, Smith and Tobin (1980), reveals the accounting struc-
ture of the model. Its key feature is that all columns and all rows sum
to zero thereby enforcing the fundamental principle that all balances
between income and expenditure generate equivalent changes in stocks
of financial assets and liabilities and, more generally, that ‘everything
comes from somewhere and everything goes somewhere’. Without a
comprehensive accounting framework of this kind, the system proper-
ties of macroeconomic models can never be securely tied down. This
framework makes it mandatory, for instance, to make it explicit how
investment is financed – a key process which is systematically ignored
in most conventional macroeconomics. For a more elaborate case where
firms undertake fixed investment as well as inventory investment, and
this is financed, not just by bank loans but also by undistributed profits
and issues of equity, see Godley (1996).
In the following I first give the whole model written out formally and
then rapidly run through it (the glossary is in Table 4.1):

y = s + Δin (4.1)
s=c (4.2)
 Weaving Cloth from Graziani’s Thread 83

Table 4.1 Model transaction matrix and glossary

Firms Banks

Households Current Capital Current Capital 

0
Consumption (= Sales) −C +S 0
Inventory investment +IN −ΔIN 0
Wages +WB – WB 0
Firms’ interest −r · L–1 +r · L–1 0
payments
Firms’ profits +F −F 0
Bank profits +Fb −Fb 0
Change in stocks of: 0
Money −ΔM rM 0
Loans rgL −ΔL 0
 0 0 0 0 0 0

Notes: C, c = consumption M, m = credit money S, s = sales

F = firms’ profits n = employment UC = unit wage cost
F b = banks’ profits p = price of goods W = wage rate
IN, in = inventories pr = productivity WB = wage bill
σ = opening volume of r = nominal rate of YD, yd = personal
inventories as proportion interest income
of sales volume
L = bank loans rr = real rate of y = output
interest

Uppercase denotes values, lowercase volumes. The star (*) denotes a desired quantity.
Subscripts s and d denote supply and demand; h means (money) ‘held’.

Δin = γ1 · (in∗ − in–1 ) (4.3)

in∗ = σ ∗ · s (4.4)
N = y/pr (4.5)
WB = W · N (4.6)
UC = WB/y (4.7)
σ = in–1 /s (4.8)
p = (1 + φ) · [UC · (1 − σ ) + σ · (1 + r) · UC–1 ] (4.9)
S = s·p (4.10)
F = S − WB + ΔIN − r · IN–1 (4.11)
IN = in · UC (4.12)
84 Stock–Flow Coherence and Economic Theory

Ld = IN (4.13)
Ls = L d (4.14)
Ms = L s (4.15)
Fb = r · L–1 (4.16)
YD = F + Fb + WB (4.17)
YD Δp
yd = − · m–1 (4.18)
p p
Mh = Mh−1 + YD − C (4.19)
C = c·p (4.20)
c = α0 + α1 · yd + α2 · m–1 (4.21)
m = Mh /p (4.22)
 
UC
r = (1 + rr ) · −1 (4.23)
UC–1

The production decision, provisionally assuming perfect foresight, is

based on sales (entirely taking the form of consumption) plus the change
in inventories, Equations (4.1) and (4.2). Inventories are assumed to
move towards some desired ratio to sales, Equations (4.3) and (4.4).
Employment is determined by output and productivity, Equation (4.5),
the wage bill by employment times the wage rate, Equation (4.6), and
unit labour costs by the wage bill divided by output, Equation (4.7).The
only tricky bit of the story concerns Equation (4.9) the way in which
prices distribute the value of sales proceeds between profits and costs.
The process needs a short section to itself, which follows.
Of the physical objects sold this period (s), a certain proportion, σ ,
was made last period (Equation (4.8)); the rest were made this period.
The objects made, but not sold, last period (σ · s) constitute the stocks
with which firms start the period and the unit wage cost of production
of these stocks was UC–1 , while the unit wage cost of objects made this
period ((1 − σ ) · s) is UC. It is a fundamental set of assumptions, Equations
(4.13)–(4.15), that inventories, since they involve outlays by firms in
advance of sales, are always financed by loans which are extinguished
when sales are made; also that there is a counterpart to every loan in
the form of credit (or bank) money. Firms had loans outstanding at the
beginning of the period equal to inventories valued at cost, so they have
to pay interest on these loans to the banks. The total historic cost (HC),
including interest, of producing what was sold this period may therefore
 Weaving Cloth from Graziani’s Thread 85

be written:

HC = (1 − σ ) · s · UC + σ · s · UC–1 + r · σ · s · UC–1 (4.24)

Firms’ profits are equal to the value of what they sell less the historic
cost of production:

F = S − HC

which may be written using a mark-up:

S = (1 + φ) · HC (4.25)

Then from Equations (4.24), (4.25) and (4.10) we have a key identity
which describes how the price implies (or is implied by) the profit mark-
up on historic unit costs, i.e. Equation (4.9). For the rest, the value of
sales, inventories and consumption are given by Equations (4.10), (4.12)
and (4.20) while, given the price decision, profits are given by Equation
(4.11) – the residual item in column 2 of the transactions matrix. Banks
are assumed to charge interest on loans outstanding, Equation (4.16), but
not to pay interest on money and this is how banks’ profits are generated.
Banks’ profits, like firms’ profits, are all distributed to households and
these receipts together with the wage bill make up nominal personal
disposable income, Equation (4.17). Real disposable income, as defined
in Equation (4.18) above, is always equal to real consumption plus the
change in the real stock of money Equation (4.22) – the only form of
wealth in this model. To spell this out, note first the identity Equation
(4.19) which says that nominal money held at the end of each period is
equal to the opening stock of money plus nominal disposable income
less nominal consumption.
Next, noting that changes in the nominal stock of money, like changes
in the value of any stock variable, can be decomposed into prices and
quantities, we can write:

ΔM = M − M–1 = m · p − m–1 · p–1 = Δm · p + Δp · m–1

Hence, using Equations (4.19) and (4.20), we obtain:

YD Δp · m–1
Δm = − −c
p p

The consumption function Equation (4.21) makes consumption depend

on real income and the opening stock of money (the only form of
wealth in this model) plus an exogenous component. Note that since
86 Stock–Flow Coherence and Economic Theory

real income is defined as above the consumption function can be

alternatively written as a wealth adjustment function:
 
α
Δm = α2 · − 0 + α3 · yd − m–1
α2

where α3 = (1 − α1 )/α2 , implying a long-run wealth target (achieved

when ΔM = 0),
α
m∗ = − 0 + α3 · yd
α2
Finally banks are assumed to adjust nominal loan rates so as to
maintain real interest rates at some given level, Equation (4.23).
The model is now complete. Treating as exogenous the variables α0 ,
r r , W, p and φ we have an equation in every variable – all stocks and
all flows, both real and nominal. The model may be solved (given ini-
tial conditions) as a fully interdependent dynamic system evolving in a
determinate way though real time. And conditional on any given con-
figuration of exogenous variables it will reach a full steady state when
the real wealth target is met (m = m∗ ).
The full steady state for output is given by
α0
y∗ =
1 − α 1 − α4
where
α0 · σ ∗
α4 =
(1 + φ) · (1 + σ ∗ · r)

Some major implications

One conclusion of central importance may be indicated via re-perusal of

the equations of the formal model, where we have one equation in the
money supply generated by loans to firms, Equation (4.15), and another
in the money which households find themselves holding, Equation
(4.19) – yet there is no equation which brings the two into equivalence
with one another. This equivalence is invariably and exactly guaranteed,
however, by the system properties of the model taken as a whole. The
use of a comprehensive double-entry system, and the combination of
national income concepts with flow of funds concepts, guarantees that
every row and every column sum to zero (see Table 4.1). From this it fol-
lows ineluctably that as soon as every variable except one is determined,
that last variable must be determined as well. And that is the position
 Weaving Cloth from Graziani’s Thread 87

we now find ourselves in. Every row and every column is indeed deter-
mined in the model as summing to zero except row 7, which shows the
supply of money and money holdings, each determined by a different
process. Yet because all other rows and columns sum to zero, it follows
that there is neither need nor place for an equation to make these two
numbers equal to one another; the system ensures that this is invariably
and exactly true. This conclusion confirms the view reiterated endlessly
by (for instance) Kaldor, Wray and Moore.
The necessary equivalence of money created with money held gains
a new dimension, augmenting the theoretical foundations of monetary
theory in a very fundamental way, when expectations are introduced
into the story. Suppose that (as in reality) firms do not know exactly what
their sales are going to be, and that therefore they base their production
decision on expected sales and intended inventory changes. To the extent
that sales expectations are not fulfilled, inventories will take the rap –
they will differ from their intended values to the extent that realized
sales differ from expected sales, and the amount of loan finance will be
comparably different as well. Then next period, starting from a position
in which inventories are out of kilter, too high or too low, the production
decision will be modified to take account of this. The firm will thus be
responding to quantity signals when making its key decisions, not price
signals. No elaborate theory of expectations is needed to underpin this
account, as mistakes are quickly remedied as a result of the palpable fact
that inventories have turned out to be excessive or inadequate.
A very similar story may be told about the consumption and the
implied intended end-of-period money holdings by households. The
consumption decision has to be taken in partial ignorance of what real
income is going to be. If income turns out to be different from what
was expected, then the accumulation of money (wealth) will be differ-
ent from what was intended to an equal extent. It is the unexpected
accumulation or depletion of the stock of money (perhaps a letter from
the bank manager) which gives a quantity signal to the household that
it must modify its consumption behaviour.
Note that in each case (that of producers and that of consumers)
we have, by introducing the notion of unintended stocks, abolished
the need for the equilibrium conditions (or disequilibrium conditions)
which are so fundamental to the traditional neoclassical theory. Produc-
ers themselves set prices; they do not need to know a hypothetical price
which will bring aggregate demand into equivalence with aggregate sup-
ply. And households will invariably be found to be holding that amount
of money which is created by the need for business finance. As already
88 Stock–Flow Coherence and Economic Theory

mentioned, there is neither need nor place for an equilibrium condition

which makes the ‘demand’ for money (whatever that may mean) equal
to the supply, and which determines the rate of interest in the process.
And while, in this model, expectations take on a centrally important
theoretical function, their practical importance is not very great because
mistakes are easy to rectify. The destruction of the key equilibrium con-
dition used by neoclassical authors by including inventory investment
in the demand/supply equation was emphasized by Hicks (1989).
A second major implication of the equations in this model is that with
only a small number of further steps we may derive an expression which
precisely describes the distribution of the real national income between
the three major sectors.
First define the rate of cost inflation,
UC
πc = −1
UC–1
and the real rate of interest defined with respect to cost inflation,
1+r
rr = −1
1 + πc
These two equations may be substituted in the price equation (Equation
(4.9)) to obtain,

p = (1 + φ) · (1 + σ · rr ) · UC

We may now divide by p and multiply by y, real output, to obtain

an expression which precisely describes the division of real output (or
income) between real profits, the real wage bill (wb = WB/p) and the real
income of banks, the creditors of the system – all in one single period of
time

y = (1 + φ) · (1 + σ · rr ) · wb

This equation, although in itself nothing more than an accounting

identity, is extremely useful when it comes to analysing the distribution
of income, both empirically and theoretically. No one of these shares can
change without the sum of the other two changing by an equal amount;
and no pair of shares can change without there being a precise impli-
cation for the third. If the profit mark-up could be fixed, rather as an
indirect tax rate can be fixed, and if banks could adjust the nominal rate
of interest on loans so that the real rate (as defined here) remained fixed,
it would follow that the nominal wage bargain is completely impotent
as a means of changing real wages; the real wage bill would simply be a
 Weaving Cloth from Graziani’s Thread 89

residual. Alternatively if the profit mark-up had to be adjusted in such a

way that prices remain constant, as a result, say, of foreign competition,
then we have a way of gauging the effect of nominal wage changes both
on real wages and on real profits. This description of income distribution
may also be useful for the analysis to which Graziani has given consider-
able amount of thought (see Graziani 1985) of the interaction between
the aspirations of the three sectors to collar various shares of real income
and the way in which inflation resolves conflicts between them.

Conclusion

In this chapter, starting from the ‘monetary circuit’ theory of how and
why credit money is generated, I have taken a single step towards the
incorporation of its insights into the simplest imaginable macroeco-
nomic model which is yet complete in the important sense that all rows
and all columns of the transactions matrix sum to zero. One important
conclusion is that it is impossible for the supply of money to differ from
the amount of money which people want to hold, or find themselves
holding, without either the need or the place for any mechanism to
bring this about.

Note
1. I am grateful to Ken Coutts, Carluccio Bianchi, Marc Lavoie and Gennaro Zezza
for comments on an earlier draft.

References

Backus, B. and T. Smith (1980) ‘A Model of US Financial and Non-Financial

Economic Behaviour.’ Journal of Money Credit and Banking 12.
Godley, W. (1996) ‘Money, Finance and National Income Determination: An
Integrated Approach.’ Working Paper No. 167, Levy Institute
Graziani, A. (1985) ‘Interet monétaire et interet réel’ in Production, circulation et
monnaie (Paris: Presses Universitaires de France).
Hicks, J.R. (1989) A Market Theory of Money (Oxford: Clarendon Press).
5
Macroeconomics without
Equilibrium or Disequilibrium
Wynne Godley1

Introduction

This paper uses a simulation model2 to describe the role which banks
have to play when decisions by households and firms are taken under
conditions of uncertainty, and when production, distribution and invest-
ment all take time. The first objective of the study is to supplement the
narrative method used perforce by Keynes and his followers before the
computer age. But it also adumbrates an alternative way of looking at
the world – alternative, that is, to the neoclassical paradigm which is
used by ‘IS/LM’ Keynesians, new Keynesians, monetarists of both kinds,
quantity rationers and almost all writers of modern textbooks. Its title
emulates Kaldor (1985) and its contents derive largely from Hicks (1989)
and from Tobin’s work read seriatim.
The neoclassical synthesis (NCS) is characterized in all its versions by
three uncomfortable features. First, the concept of an exogenous money
stock which can be ‘controlled by the Fed’ is required if this class of
models is to be capable of solution. The entire apparatus of IS-LM dia-
grams, which is still the workhorse of macro teaching, requires that the
‘money supply’ is not merely exogenous but fixed 3 . Bank loans have no
essential role, if any, to play4 . Second, the NCS takes it as axiomatic
that prices send all the signals which govern action, even when the
signalling system doesn’t work well because of rigidities, imperfections,
asymmetries in information flows and so on. And expectations, which
have become such an important part of economics in recent years, are
invariably expectations about prices. Third, mainstream thinking, as
Hicks pointed out with increasing emphasis in his later works, has no
satisfactory way of handling real time. The theory of exchange, even
when inter-temporal ‘trade’ is assumed to occur, cannot characterize

90
 Macroeconomics without Equilibrium or Disequilibrium 91

the Hicksian ‘traverse’ – the whole sequence of events which carries the
community, often chaotically, through history. These three uncomfort-
able features of the NCS constitute a syndrome which has its roots in
a vision of the universe as consisting, in its essence, of a single market
where individual maximizing agents exchange goods, labour, money
and ‘bonds’ with one another. The NCS sponsors the belief that strong
conclusions can be drawn about how the real world actually works
(e.g. what determines the level of real output and employment) from
assumptions about supposedly rational behaviour in advance of any
empirical study.
In the model proposed here, there is literally no such thing as a ‘sup-
ply’ of bank money, at least in the sense required for the solution of
the IS-LM model – that is, a supply distinct from demand, with an equi-
librium condition equating the two and thereby determining ‘the’ rate
of interest. Banks accept money and undertake to exchange it in vari-
ous ways. They respond to the fluctuating needs of firms for revolving
finance and of households in the management of their transactions and
the disposition of their wealth, while remaining profitable and solvent.
Governments can no more ‘control’ stocks of either bank money or cash
than a gardener can control the direction of a hosepipe by grabbing at
the water jet. Decisions by households, firms and banks are mainly based
not on price but on quantity signals which often take the form of real-
ized stocks of wealth or inventories. Expectations concern such diverse
things as sales, income and wealth. Historical time is intrinsic because
the past, in the form of state variables, is inherited by each period; then
a transition to a new state takes place which becomes the inheritance of
the subsequent period. Simulation is used because unruly sequences can-
not be penetrated by static or equilibrium methods; the method makes
it possible to pin down exactly why the sequences occur as they do.
Nothing, it is maintained, can be known about the real world unless it
is actually studied empirically, hence no greater claim is made for the
model presented here than that it is an elementary schema laying out
a rigorous space within which empirical macroeconomics can proceed.
The starting point we lay down is a realistic, if simplified, characteriza-
tion of the institutional framework within which all modern capitalist
economies operate.
Our model, looked at one way, is the extreme antithesis of the Wal-
rasian model. Yet agents’ disparate plans, expectations and outcomes
are all reconciled with one another in it – though obviously not by a
heavenly auctioneer calling prices; the reconciliation occurs through
the agency of banks when they allow loans and all kinds of money to
92 Stock–Flow Coherence and Economic Theory

expand and contract, without anyone even noticing, in response to the

uncoordinated needs of firms and households. The model does have an
important Walrasian feature however. It is based on a complete, if simpli-
fied, system of stock and flow accounts set in a double-entry framework
where every variable performs a different role according to the context
in which it occurs e.g. according to whether it describes an asset or a
liability. It then follows that there is always one variable which is deter-
mined by two different equations which must both give the same answer
when the model is solved. Such completeness is perhaps a hallmark of
any properly constructed model of a whole system.
The first section of the paper displays the accounting framework of
the model, the second describes the behaviour of the four sectors which
make it up, the third shows how the whole thing works using four numer-
ical simulations and a concluding section draws some strands together.
An appendix contains a glossary and lists the equations used to gen-
erate the simulations. The reader is invited to skip, first time round,
to the simulation results which give a quick general idea of what is
at issue.

The accounting framework

The following matrices set out the stock and flow accounts on which the
model is based. The major simplifications are that the economy is closed,
there is no fixed investment, no fixed capital and no equity: households
do not borrow and firms do not hold money; all bonds are ‘bills’ of
which the capital value does not change when interest rates change;
money wages and productivity are constant. While these assumptions
make the model unrealistic as a representation of the real world, there
remains enough to characterize precisely the main, very basic, features
of a monetary economy. Although simplified, the model is not arbitrary,
for it is complete in its own terms; everything visibly goes somewhere
and comes from somewhere.
The flow matrix shows how the model comprises four sectors: house-
holds, firms, government and banks; it also defines most of the symbols
to be used. Households receive all factor income plus interest payments
on their assets. What they do not spend on consumption has an identi-
cal counterpart in changes in wealth, somehow allocated between four
assets – cash, non-interest bearing money, interest bearing money and
government bills. Firms produce and sell goods and services, accumulate
inventories, borrow from banks, pay wages and distribute profits. Banks
 Macroeconomics without Equilibrium or Disequilibrium 93

Table 5.1 Flow of funds at current prices

Firms: Banks Row

Sum
Households Current Capital Current Capital Govt. 0

Consumption −C +C
Government +G −G 0
Expenditure
[Sales] [S]
Stockbuilding +ΔIN −ΔIN 0
Tax −T +T 0
Wages +WB −WB 0
Profits +F −Ff −Fb
Interest on loans −rl · L–1 +rl · L–1
Interest on +rm · M–1 −rm · M–1 0
money
Interest on bills +rb · Bh–1 +rb · Bb–1 +rb · B–1 0
[Disposable [Yd]
income]
ΔStock of cash −ΔHh −ΔHb +ΔH 0
Δstock of current −ΔM1 +ΔM1 0
deposits
Δstock of −ΔM +ΔM 0
demand
deposits
ΔStock of bills −ΔBp −ΔBb +ΔB 0
Δstock of loans +ΔL −ΔL 0
Column Sum: 0 0 0 0

have credit money (both kinds) as liabilities and loans, bills and cash for
assets. Their transactions in assets may all be looked on, reading hori-
zontally, as residuals which make the row in question sum to zero; they
can be seen this way because, since every other column sums to zero, the
banks’ transactions must do so as well. Banks’ profits are the excess of
interest receipts over interest payments. The government spends, taxes
and pays interest on its debt. Any deficit has, as its counterpart, a change
in cash plus bills in some combination.
Table 5.2 shows the stock (balance sheet) counterpart of the flow
matrix. Every financial asset is matched by a financial liability. Total
household wealth is equal to the sum of money plus bills (reading verti-
cally) or equivalently (reading horizontally) to the stock of government
debt plus the stock of inventories valued at cost – the only tangible asset
in this model. It is transactions in assets in Table 5.1 which heave the
stock variables in Table 5.2 from one period to another.
94 Stock–Flow Coherence and Economic Theory

Table 5.2 Balance sheets

Households Firms Banks Government Total

Inventories +IN +IN

Cash +Hh +Hb −H 0
Current deposits +M1 −M1 0
Demand deposits +M −M 0
Bills +Bh +Bb −B 0
Loans −L +L 0
Column sum V 0 0 GD 0

Notes: V = Household wealth

GD = Total government liabilities

Behavioural assumptions

In this section, the behaviour of the four sectors of the model will be
described for the most part verbally, but equations will be used when
precision calls for them. Moving from the world of accountancy to that
of behaviour requires that each concept be given a different function
according to the context in which it occurs and suffixes will be appor-
tioned accordingly; for instance the suffix e denotes an expected value,
a star indicates a desired value and so on. Only those symbols which
describe ex post realized values will have no suffix. The simulation model
is given, as a complete system of about 40 equations, in the appendix.5

The behaviour of firms

The following schema describes the main decisions firms take6 and
shows why bank finance is required if normal business is to proceed.
‘Firms’ here comprise the distributive chain as well as producers nar-
rowly defined. The manufacturing firm makes goods over a period of
time which intermediary traders stock, advertise, guarantee and mar-
ket, normally holding prices fixed – certainly in response to short-run
fluctuations in demand – and the whole chain of agents is in a state of
uncertainty about what the value of sales and profits will actually be. It
will be assumed that firms are operating within the normal range of out-
puts at which running costs per unit of output are constant and that they
base their decisions about production and prices on the quantity they
expect to sell at the price they choose plus any adjustment to inventory
levels they wish to see7 .
 Macroeconomics without Equilibrium or Disequilibrium 95

Realized sales are determined by actual consumption plus government

expenditure and realized inventories by planned inventories modified
by the difference between expected and actual sales. Realized profits are
then given by residual as shown in column 2 of the transactions matrix,
Table 5.1, namely,

Ff = S − T − WB − rl · IN–1 + ΔIN (5.1)

where Ff is profits of firms, S is final sales (consumption plus government

expenditure), T is indirect taxes, WB is the wage bill, rl is the rate of
interest on loans and IN inventories valued at cost.8 It will be assumed
that profit margins are set like tax rates, as some proportion of the pre-
tax value of sales – an assumption which is broadly consistent with the
stylized facts, since the share of profits in total final sales, though cyclical
and subject to trends, is a quite well behaved number. It will also be
assumed that realized profits are all distributed to households. These
assumptions have two very important logical implications. First, if profits
are a constant share of sales, then it must also be the case that prices are
a constant mark-up on the historic cost of production. Second, if profits
are all distributed, it must also be the case that bank loans expand and
contract, $ for $, with inventories.
To show this, note first that as, taking all firms together, wages are the
only cost of production, the end period value of inventories is the pro-
portion (σ ) of the wage bill incurred each period which is not embodied
in sales that period.

IN = σ · WB (5.2)

Putting Equations (5.2) into (5.1) we get an alternative, more intuitive,

way of describing profits which makes the time factor more explicit and
intelligible.

Ff = S − T − (1 − σ ) · WB − σ–1 · (1 + rl ) · WB–1 (5.1a)

In words, profits are equal, by definition, to the excess of receipts from

ex-tax sales over what it cost, historically, to produce what was sold.
The third term on the right-hand side of Equation (5.1a) describes the
proportion of costs incurred this period which is embodied in sales this
period; the fourth term describes the costs incurred last period which
will be embodied in sales this period, including the interest cost which
arises from the fact that production takes time.
96 Stock–Flow Coherence and Economic Theory

Defining the last two terms in Equation (5.1a) as historic cost (HC) we
can write:

S = T + Ff + HC (5.1b)

or,

S = (1 + τ ) · (1 + φ) · HC (5.1c)

where τ is the tax rate and φ the rate of profit mark-up. It can now be seen
why, with historic cost pricing and full distribution of profits, changes
in inventories valued at cost must always be matched exactly, $ for $, by
changes in loans from outside the production sector. This now follows
directly from the definition of profits in Equation (5.1), for if all profits
are distributed, the cash flow derived from sales falls short of what is
needed for taxes, wages and interest payments by exactly the amount of
the increase in inventories. Injections of revolving finance from outside
are thus essential if firms are to undertake production in advance of sales
and also extract (and distribute) profits from the business as sales are
made and profits realized.
How are firms’ expectations about sales formed? The question proba-
bly doesn’t have a good or general answer. The assumption underlying
this paper is that we live in a contingent world about which economic
theory cannot tell us very much and which can only be understood bet-
ter as a result of laborious empirical study, with pattern recognition a
key element in the type of cognition required. The important thing here
is that we have a way of dealing with the fact that sales never turn out
as expected. The signal to which firms respond is not a price signal but,
typically, a quantity signal; it is in response to realized sales and therefore
inventory levels that firms decide whether or not to increase or reduce
production, change prices or, in a more complete model, invest. Mean-
while bank loans expand and contract buffer-wise to the extent that
expectations are falsified.

Behaviour of households
Consumption is determined by the stock of wealth inherited from the
previous period together with the expected flow of disposable income,
ignoring, for the time being, consumer credit and asset price changes.

C = C(YDe , V–1 ) 0 < C1 , C2 < 1 (5.3)

This, given the accounting relationship describing wealth accumulation,

ΔV = YD − C (5.4)
 Macroeconomics without Equilibrium or Disequilibrium 97

necessarily implies a precise value for the desired long-run wealth-

income ratio.
As shown in Table 5.1, any addition to wealth must be allocated
between four assets – cash (Hh ), non-interest bearing money (M1), inter-
est bearing money (M) and bills (Bh ) – and the way this happens in the
model owes everything to James Tobin and his associates.
Households aim to apportion their wealth between the assets avail-
able to them, in proportions which depend on the rates of interest on
offer subject to their having enough spendable money (current deposits
and cash) for transactions and to take the strain when unexpected
things happen. In order to understand (or model) the process it is abso-
lutely essential to distinguish interest bearing from non-interest bearing
money, the two being held for very different reasons.9
In the model, since cash holdings are nowadays so unimportant, they
are assumed to be some straightforward proportion of consumption
which is unaffected by interest rates. Intended holdings of other assets
are described by the following functions where the suffixes e and h denote
that the variable in question, lifted out of the accounting matrix into the
world of behaviour, denote what households ‘expect’ or ‘hold’. The word
‘hold’ contrasts with the usual, perhaps prejudicial, expression ‘demand
for’ money or other assets.
 
M1∗h YDe
e = M1 r m , r b , e ; M11 < 0; M12 < 0; M13 > 0 (5.5)
Vnc Vnc
 
Mh YDe
e = M r m , r b , e ; M1 > 0; M2 < 0; M3 > 0 (5.6)
Vnc Vnc
 
Bh YDe
e = B r m b, r , e ; B1 < 0; B2 > 0; B3 < 0 (5.7)
Vnc Vnc

where rm , rb are the rates of interest on respectively money and bills and
Vnc is wealth net of cash holdings. M1h , holdings of non-interest bearing
money, has a star which means that the function describes an aspiration.
It is essential that the income terms in these equations be scaled by
wealth, otherwise the share of M1 in wealth (at given interest rates) will
rise through time with income.10 The constraints and adding-up prop-
erties hardly need emphasizing; the sum of constants must be one since
total wealth must equal the sum of its parts, the sum of coefficients on
each interest rate (reading vertically) must be zero, and the sum of coef-
ficients on the income term must be zero as well. The sum of coefficients
on the interest rates in Equations (5.6) and (5.7) reading horizontally
must be approximately zero too because there can be no great difference
98 Stock–Flow Coherence and Economic Theory

between raising the own rate of interest and reducing the sum of all other
interest rates.
It is assumed that the planned holdings described in Equations (5.6)
and (5.7) go through but that holdings of non-interest bearing money
perform a ‘buffer’ role. The aspiration is given in Equation (5.5) but the
actual outcome modifies this to the extent that income expectations are
falsified.

M1h = M1∗h + YDe − YD (5.8)

As any two of the three Equations (5.5) to (5.7) imply the third, we can
represent holdings of interest bearing money as the residual between net
wealth and total holdings of the other two assets.

Mh = Vnc − M1h − Bhh (5.9)

In the simulation model, holdings of M1 are constrained to be non-

negative. If actual income falls short of expectations by enough to
eliminate holdings of M1, Equation (5.9) ensures that households delve
into their demand deposits.
As with firms, we don’t have a very strong view about how expectations
are formed. Under certain circumstances expectations can be important,
for instance if whole generations alter their savings patterns. But nor-
mally, as is the case with firms, households are kept on the rails by the
regular information they receive about their stocks of wealth. Nothing
guides people more remorselessly than the monthly bank statement.

The banks
Banks may be said, without putting an excessive strain on language, to
‘supply’ loans although it seems more natural to say that they ‘make’
them. But they do not, in any sense recognizable to common parlance,
‘supply’ money unless what is being referred to is a loan.11 What they
do is exchange assets for one another or for loans. Presented with a valid
cheque banks will make (it is part of what they undertake to do by tak-
ing you on as a client) the appropriate entries in whatever account is
designated or hand cash over the counter without question; presented
with cash, they will make a counterpart addition to a current or deposit
account or reduce a loan. The making of these exchanges has nothing
in common with the exchange of money for goods and services (say
haircuts) where the business makes a profit by appropriating some pro-
portion of what is sold. Banks make their profits in a completely different
way – by receiving a higher rate of interest on their assets than what they
 Macroeconomics without Equilibrium or Disequilibrium 99

pay on their liabilities. In what follows, the assumption that banks take
a passive role with regard to this switching will be emphasized by using
the suffix x, denoting exchange, rather than the usual s for supply. It
has already been pointed out that if firms distribute all their profits, they
must be getting finance from banks on a scale which matches the value
(reckoned at cost) of inventories one for one, and it is an assumption of
the model that this is what in fact happens.
To formalize, the banks’ balance sheet constraint is:

Bbh = M1x + Mx − Lm − Hbh (5.10)

which says that their holdings of bills and cash plus the loans they
have made must exactly equal the money they have exchanged into,
or accepted as, deposits of bank money.
Banks’ profits (Fb ) are given by the excess of receipts of interest on their
assets (loans and bills) over payments of interest on money.

Fb = rl · Lm−1 + rb · Bbh−1 − rm · Mx−1 (5.11)

In the absence of equity capital, banks’ profits all simply flow to the
household sector.
We next assume that banks have to hold reserves, in the form of cash,
in some fixed proportion to their liabilities.

Hbh = ρ · (M1x + Mx ) (5.12)

To guarantee that banks make profits, two conditions have to be met.

The first is that the rate on loans exceeds the rate on money. In practise
the rate on loans is higher than the rate on bills as well, otherwise banks
would make higher profits by holding bills than by making loans. This is
modelled by making the loan rate exceed the money rate by some mark-
up, but when this is insufficient to get the loan rate above the bill rate a
trigger mechanism is introduced to make it do so.
This has been modelled, Heath Robinson style, as follows:

rl = rl1 · z1 + rl2 · z2 (5.13)

Where,

rl1 = (1 + φb1 ) · rm (5.13a)

rl2 = (1 + φb2 ) · rb (5.13b)

z1 and z2 take on the value 0 or 1 depending on whether rl1 is greater or

smaller than the bill rate and φb1 , φb2 denote rates of mark-up.
100 Stock–Flow Coherence and Economic Theory

The second condition necessary to ensure that banks make profits is

that their bill holdings are normally positive – they do not have to borrow
for long from the government at penal rates. We model this by making
banks raise the money rate of interest in steps whenever their bill hold-
ings fall below a certain level (relative to their liabilities) and reduce the
rate on money whenever bonds are above this critical level.
More precisely,

Δrm = (z3 − z4) times some small number (5.14)

where z3 and z4 take on the value 0 or 1 depending whether the banks’

bill to asset ratio is above or below the critical level.
The remaining equations describing the dealings of the banks with
households and firms are:

Hhx = Hhh (5.15)

M1x = M1h (5.16)
Mx = M h (5.17)
Bhx = Bhh (5.18)
Lm = L r (5.19)

It has already been pointed out, in section 2, that as every row in

the transactions matrix sums to zero and every column excluding banks
sums to zero, it follows that the column describing banks’ transactions
must sum to zero as well. This property of banks’ balance sheets means
that banks can exchange, with impunity, any one kind of asset for any
other and simultaneously make loans, on any scale whatever. None of
the equations above is an equilibrium condition in the ordinary sense. There
are enough equations in the model for banks to be able to respond
immediately and profitably to any configuration whatever of asset hold-
ings desired by households and simultaneously the loan requirements
of firms.

The government
The government’s budget constraint is simple and traditional

ΔH + ΔB = G + rb · B–1 − T (5.20)

The government has three policy instruments at its disposal: the flow
of Government expenditure; the rate of tax, in our model all indirect,
levied on all types of expenditure and the rate of interest on bills. The
 Macroeconomics without Equilibrium or Disequilibrium 101

announcement of a bill rate of interest implies that the government will

exchange any quantity of bills at that rate of interest for cash.

Bx = Bbh + Bhh (5.21)

And this, given that households’ bill holdings are determined in

Equation (5.7), means we can write:

Bbx = Bx − Bhh (5.22)

where the suffix x means that the asset has been passively exchanged
for something else (cash in this case). The governments’ bill liabili-
ties are what is left over as a residual from all the government’s other
transactions.
We have at last completed the Walrasian circle! We have an equation
both in banks’ holdings of bills (Equation (5.10), A.25 in the appendix)
and also in the government’s exchange of them with banks (Equation
(5.22) above, A.37 in the appendix). After much travail we have estab-
lished a logical architecture such that the two are found, indeed, to be
equal to one another when the model is solved. It is not immediately
obvious that this should be so for the two equations come, as it were,
from two quite different directions. From one direction banks’ bill hold-
ings are the residual of the relatively active components of the banks’
balance sheet (all three kinds of money plus loans). From the other direc-
tion, banks’ holdings of bills are the residual of all the government’s
other transactions. At each instant of time, the bills which the govern-
ment finds that it has sold to banks is the same number as the bills which
the banks find, for entirely different reasons, that they have needed to
buy from the government. The two versions of Bb will only be equal if
the accounting in all the rest of the model is complete and watertight.
To achieve this is easier said than done.
In the neoclassical model it is habitual to use the same government
budget constraint as here (Equation 5.20) and then to declare any one
of cash, bills or interest rates exogenous whereat the other two become
endogenous (see, for instance, Modigliani (1963)). In the absence of his-
torical time, there is nothing untoward about this. As pointed out in the
footnote to the second paragraph of this paper, the neoclassical model
in its market-clearing version can be solved using alternative assump-
tions about the stock of money which will, yes, make no difference to
any component of the model’s solution except the ‘price level’. Set in
historical time, however, with banks providing loans, exchanging assets
and keeping guard, with an inevitable time lag in their response, over
102 Stock–Flow Coherence and Economic Theory

any untoward changes in the structure of their balance sheets, the posi-
tion is entirely different. The government’s ex post deficit is a residual
over which the government has no direct control and the banks’ hold-
ings of bonds are a residual over which the banks have no direct control.
The total stock of cash is thus a residual made up of two other residuals,
neither of which can be directly controlled! So much for the dogma
contained in every modern textbook, on which the whole neoclassical
edifice rests, which says that the stock of cash is ‘controlled by the Fed’
with the stock of bank money (both kinds in an ugly lump) determined
thence by the money multiplier. In our model, notwithstanding that
there is a rigid fractional reserve rule in place, the entire chain of causal-
ity is reversed compared with this story! Credit money holdings have two
starkly different component parts: interest and non-interest bearing; one
is determined as part of households asset allocation decision, the other
by households’ fluctuating needs which in any short period are bound
to move in unexpected ways. The stock of cash (excluding that held by
households) is then determined by the stock of bank money (both kinds
together) via the fractional reserve ratio; banks must swap cash for bills
until their reserve requirements are met.

Long-run properties of the model

Before coming to the simulations, it remains to point out that, in
accordance with the famous insight of Carl Christ (1967) subsequently
embellished by Blinder and Solow (1975) and Tobin and Buiter (1976),
the full steady state of any properly specified stock–flow model of a closed
economy will be one in which (as all stock variables are then constant)
the tax receipts exactly equals government outlays. Hence, if taxes are
levied in some proportion to income (or sales), the steady-state flow of
GDP must be equal to government outlays times the reciprocal of the
tax rate. The steady-state stock of wealth is determined in the consump-
tion function and the steady-state stock of government liabilities will be
equal to wealth less private sector loans.

Simulations

In this section the model’s properties are demonstrated using numeri-

cal simulations. Although the results are conditional on rather arbitrary
values which have been attributed to variables and parameters, our con-
jecture is that, once a comprehensive system of stock and flow accounts
has been designated, the behaviour of the model will be very broadly
the same whatever parameters are chosen provided, of course, that they
 Macroeconomics without Equilibrium or Disequilibrium 103

10.0

8.0

6.0 Disposable
GDP income
4.0 Stockbuilding

2.0

Consumption
0.0
0 1 2 3 4 5

Figure 5.1 Simulation 1: Effect on income and expenditure flows

assume stock–flow norms – wealth-income ratios for households and

inventory-sales ratios for firms. For all its shortcomings, the simulation
method has the merit that it is always possible to track down exactly
why the results are what they are. If, for instance, interest rates rise
unexpectedly in response to a particular kind of shock, we can go back
and see whether this is because the model has unacceptable features (in
which case we have to change the model) or perhaps because we hadn’t
realized, when doing thought experiments, that once all the ramifica-
tions are made explicit, we get anomalous results which make us wish to
change the model we were previously carrying in our minds.

Simulation 1: a step up in inventory levels

The first simulation follows through the effects of a once-for-all rise in
the desired ratio of inventories to production, the main purpose being to
show what happens when loans generate income flows as well as money
stocks. In the very short term, as Figure 5.1 shows, the rise in the level
of inventories causes a blip to stockbuilding and hence to production.
There is no simultaneous effect on consumption in period 1 because, it
is assumed, the rise in income was unexpected by consumers.12 How-
ever in period 2, consumption rises in response to the addition to wealth
during period 1. Eventually a new steady state will be reached in which
GDP, disposable income, consumption and the stock of wealth all end up
roughly where they started. Figure 5.2 shows the initial effect on house-
holds’ balance sheets. The top line represents the addition to household
104 Stock–Flow Coherence and Economic Theory

Total wealth
12.5

10.0

Interest bearing
7.5 money

5.0
Bonds held by
Non–interest bearing households
2.5 money

Cash
0.0

1 3 5 7 9

Figure 5.2 Simulation 1: Effect on wealth and its components

wealth which has taken place as a consequence of the shock and is

equal to the cumulative excess of disposable income over consumption
in Figure 5.1. The four lower lines show how wealth is allocated between
the four financial assets. As the addition to income was unexpected,
no active portfolio choice is immediately made and consequently the
entire accretion fetches up, in period 1, as an addition to holdings of
non-interest bearing bank money. In the present instance the notion of
the initial rise in money being a response to an increased ‘demand’ for it
is particularly wide of the mark; holdings of non-interest bearing money
have gone up by default13 because income recipients have been caught
napping.
In period 2 the process of asset allocation begins. There is a tiny addi-
tion to holdings of cash by households which is needed to finance the
higher flow of consumption, but otherwise the initial accretion of non-
interest bearing money starts to be salted away into interest bearing
deposits and bills in proportions which depend on relative interest rates.
Holdings of non-interest bearing money, although reduced in period 2,
remain higher than they were before because of the continuing need to
finance a higher flow of transactions.
Figure 5.3 shows the counterpart changes (always compared with
what would otherwise have happened) in the banks’ consolidated
 Macroeconomics without Equilibrium or Disequilibrium 105

20.0

15.0 Loans

10.0
Money (both kinds)

5.0
Reserves
0.0

–5.0
Bonds

0 1 2 3 4 5

Figure 5.3 Simulation 1: Effect on banks’ balance sheets

balance sheet. The top line shows the (addition to the) stock of loans,
assumed equal to the cumulative total of the addition to stockbuilding
in Figure 5.1; as the stock of inventories is higher for ever, so is the stock
of loans. The second line gives the addition to deposits of both kinds
taken together and the third line shows the addition to banks’ reserves,
assumed to be 10% of total deposits. The lower line then shows how,
as a logical necessity given everything else, banks are initially obliged
to reduce their holdings of bills; they have to do this to the extent that
the rise in loans and reserve requirements exceeds the amount of bank
money that households wish to hold.14
Figure 5.4 shows (using the solid lines and the right-hand scale) the
three interest rates on money, loans and bills, together with banks’ hold-
ings of bonds expressed as a proportion of their assets (the dotted line
using the left-hand scale). Banks will always set the loan rate of interest
above the bill rate, otherwise it would be more profitable for them to
hold bills rather than make loans; and the bond rate is always higher
than the money rate otherwise households would never hold bills. This
hierarchy will be satisfactory to banks because the rate on each category
of their assets (excluding mandatory reserves) is higher than that on each
category of their liabilities.15 A crucial further assumption is that banks
avoid being forced ‘into the bank’ i.e. having to borrow from the central
bank at a penal rate; to do this they will keep their bills in some positive
ratio to their liabilities – to be termed ‘the defensive asset ratio’. They
106 Stock–Flow Coherence and Economic Theory

0.02625
0.050
Loan rate (LH scale)
0.02500
Banks’ bonds as 0.040
share of money
0.02375 (dotted line,
RH scale) 0.030

Bond rate (LH scale)

0.02250
0.020

0.02125
0.010
Money rate (LH scale)
0.02000
0.000
1 5 9 13 17 21 25

Figure 5.4 Simulation 1: Effect on interest rates

will respond to a decline in this ratio – a quantity signal – by getting

households to switch out of bills into money by raising the money rate
of interest.
In the simulation model it was assumed that banks raise or lower the
rate they are prepared to pay on deposits to an extent which depends on
the distance of the defensive asset ratio from the desired norm. To pro-
tect banks’ profits when deposit rates are raised, loan rates must be raised
simultaneously. Figure 5.4 shows how money and loan rates change rela-
tive to the bill rate until the defensive asset ratio is restored to its original
level.
Figure 5.5 shows the counterpart of everything described so far in terms
of changes to the two largest components of household wealth; because
of the change in the pattern of interest rates, with loan and money rates
permanently higher than they were before, there is permanent switch out
of bills and into interest bearing money, each expressed as proportion
of wealth. According to this way of thinking, by the way, it is only for
the brief moment before households react to higher incomes by spending
more or by investing actively, that it is true to say that ‘every loan creates
a deposit’.
In the new steady state, the ratio of wealth to income is restored to
its original level. But since loans and inventories are higher than before,
the total stock of government debt (cash plus bills) has to be lower by
 Macroeconomics without Equilibrium or Disequilibrium 107

0.3030

0.6975
0.3000

0.6900 Interest bearing

money as share of
0.2970
wealth (RH scale)
0.6825
0.2940

0.6750
Bonds 0.2910
as share of wealth
0.6675 (LH scale)
0.2880

0.6600
2 6 10 14 18 22 26

Figure 5.5 Simulation 1: Bonds and money as shares of wealth

the amount of the increase in private debt; this can be read off the balance
sheet matrix, Table 5.2. The dynamic intuition here is that between the
two steady states, total income and output are all the time higher than
they otherwise would have been. As government expenditure on goods
and services and tax rates are unchanged, there has to be a reduction
in government indebtedness throughout the period which is illustrated
in Figure 5.6. Yet the total stock of cash must be higher in the new
steady state because banks’ reserve requirements rise (Figure 5.3) while
households cash holdings fetch up (virtually) unchanged. Therefore
(always assuming fixed bill rates of interest) more than all of the fall
in government liabilities takes the form of lower bill holdings.
The story of simulation 1 is almost complete. It remains to point out
that since government debt is lower in the new steady state, the flow of
government interest payments (given bill rates) will also be lower and therefore
the flow of aggregate income will be slightly lower as well. We shall defer
discussion of whether and in what sense the total stock of base money
could be ‘controlled’ under these or other circumstances. A question!
How should we think about the limit to the loan making process? One
answer is that an increase in the loan rate will, in reality though not
in this model, choke off the demand for loans. The second is that as
money rates nudge the bill rate, the government may be unable to sell
bills at all except at a higher rate of interest. In other words, it may be
108 Stock–Flow Coherence and Economic Theory

0.30
Effect on cash

0.00

–0.30

Effect on total liabilities

–0.60

–0.90
Effect on government bonds
–1.20

2 5 8 11 14 17 20 23 26 29 32

Figure 5.6 Simulation 1: Effect on government liabilities

unrealistic to suppose that the bill rate of interest can validly be treated
as exogenous beyond a certain point; perhaps it has eventually to move
up if the private demand for loans rises beyond a certain point.

Simulation 2: a step in government expenditure

The second simulation explores the consequences of lifting government
expenditure on goods and services in a single step, everything else given.
In this experiment, so as better to isolate the asset allocation decisions,
perfect foresight on the part both of firms and households is assumed;
expected sales and disposable income are assumed to be equal to the
actual values generated by the model.
Figure 5.7 shows the addition to government expenditure and the con-
sequential additions to GDP and wealth. There is a small overshoot in
period 1 because of the relatively rapid adjustment of inventories towards
their new level. Wealth, on the other hand, adjusts relatively slowly.
Government debt (implied but not directly shown in the figure) and
government interest payments rise throughout the transition period.
Figure 5.8 shows what happens to the components of the banks’ bal-
ance sheet. The top dotted line shows the addition to loans – a rapid
response occasioned by the need of industry for finance. The addition to
the total stock of money (taking both kinds together) rises more slowly
(along with wealth) and banks’ reserves rise step by step with money.
It then has to be the case that banks’ holdings of bills, their defensive
 Macroeconomics without Equilibrium or Disequilibrium 109

10.0

Addition to GDP
8.0

6.0
Addition to wealth
4.0
Addition to government expenditure
2.0

0.0

1 4 7 10 13 16 19 22 25

Figure 5.7 Simulation 2: A step in government expenditure

30.0

22.5

15.0 Addition to
loans
7.5
Addition to money (both kinds)
0.0
Addition to cash
Change to bond holdings
–7.5
0 2 4 6 8 10 12

Figure 5.8 Simulation 2: Changes to components of the banks’ balance sheet

assets, initially fall by the difference between loans and reserves on the
one hand and money on the other. The way banks respond to the fall in
their bills holdings is shown in Figure 5.9.
The fall in the defensive asset ratio sparks off a rise in the money (and
hence loan) rate of interest which starts to be reversed as soon as that ratio
is restored. What brings interest rates down again? The answer is that
after period 9 (by when the defensive asset ratio is restored) the flow of
disposable income falls progressively relative to the stock of wealth – that
this is happening is clearly implied in Figure 5.7. The fall in disposable
income relative to wealth means that ex ante holdings of money also fall
progressively (see Equations 5.5–5.7 in section 3) and this, in turn, means
110 Stock–Flow Coherence and Economic Theory

0.0400 0.02080

Banks’ bonds as a proportion of

0.0360 their liabilities 0.02060

0.0320 0.02040

0.0280 0.02020

Rate of interest
0.0240 on money 0.02000

0.0200 0.01980

1 4 7 10 13 16 19 22 25

Figure 5.9 Simulation 2: Banks’ bonds and the rate of interest

15.0 90.0

Non-interest bearing
money as a proportion of wealth (LH scale)
14.0 89.0

13.0 88.0

12.0 87.0
Interest bearing assets
(money plus bonds) as a
11.0 proportion of wealth (RH scale) 86.0

10.0 85.0

1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Figure 5.10 Simulation 2: Components of wealth

that the rate of interest on money falls without any further change in
the banks’ defensive asset ratio.
Figure 5.10 shows, using the left-hand scale, how holdings of non-
interest bearing money immediately rise for transactions purposes; and
 Macroeconomics without Equilibrium or Disequilibrium 111

0.2625
0.6375
Interest bearing money as a
proportion of wealth (RH scale) 0.2600
0.6300

0.2575
0.6225

0.2550
0.6150
Bonds as a proportion of
wealth (LH scale) 0.2525
0.6075

0.3500
0.6000
1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Figure 5.11 Simulation 2: Allocation of interest bearing assets of households

there has to be a corresponding fall in interest bearing forms of wealth

for ‘adding up’ reasons. Holdings of non-interest bearing money (as a
share of wealth) gradually fall back to their original level as wealth rises.
The next Figure 5.11 shows the response of households’ holdings of
interest bearing money bonds to the initial shock and to the subsequent
changes in interest rates. There is an initial downwards blip in both series
to accommodate the immediate need for non-interest bearing money;
thereafter the two series balloon in opposite directions, then subside
again. With given bill rates of interest, the entire system settles down
with income flows higher than before but with all stock–flow ratios and
relative interest rates exactly where they started.

Simulation 3: introducing random expectations

In this third simulation we put the whole system under severe strain by
assuming that expectations of sales by firms and also expectations of dis-
posable income by households are subject to violent random processes.
No pretense is made that expectations are really formed in this way; the
object of the exercise is to find out how banks would deal with such
chaotic behaviour if they had to.
Figure 5.12 shows, for each period, the gap between actual and
expected disposable income together with the change in deposits of
112 Stock–Flow Coherence and Economic Theory

0.90

Expected less
0.60 actual income

0.30

0.00

–0.30

–0.60 Change in money (both kinds)

i.e. liabilities of banks

Figure 5.12 Simulation 3: Response of money to income shocks

0.60
Change in inventories
0.40 (and hence loans)

0.20

0.00

–0.20

–0.40
Expected less actual sales

Figure 5.13 Simulation 3: Response of inventories to sales shocks

non-interest bearing money which act as buffers, moving each period

in the opposite direction to the expectations gap. Figure 5.13 shows a
similar divergence between actual and expected sales by firms, whose
inventories fluctuate in a similarly shock absorbing way.
The banks have no difficulty accommodating the sharp and disparate
series of shocks to these two components of their balance sheet, but they
do have to move money rates (and therefore loan rates) about sharply in
order to achieve this.
Figure 5.14 shows the defensive asset ratio with the money rate of
interest and Figure 5.15 shows the consequential changes to holdings of
 Macroeconomics without Equilibrium or Disequilibrium 113

0.0600

0.02060
0.0525
Banks’ bonds as a proportion of their liabilities
0.02040
0.0450

0.02020
0.0375

0.02000
0.0300 Money rate of interest

0.01980
0.0225

0.01960

Figure 5.14 Simulation 3: Response of money rates to banks’ bond holdings

0.6450
0.300
0.6375 Money (of both kinds)
as a share of wealth (RH scale) 0.290
0.6300
0.280
0.6225
0.270
0.6150
Bonds held by households 0.260
0.6075 as a share of wealth (LH scale)
0.250

Figure 5.15 Simulation 3: Allocation of household wealth

interest bearing money and bills (expressed as a share of a wealth) which

take place as a consequence.
The last Figure 5.16 shows what happens to changes in government
liabilities – that is, total bill issues (the solid line) and total issues of
high-powered money (the dotted line). They have to move about in this
wild way as a unique counterpart to the operations of households, firms
and banks. The story is as follows. The government has a pre-determined
fiscal policy (public expenditure and tax rates are pre-determined) and
has also determined the rate of interest on bills. Banks, in order to
remain profitable, have to keep adjusting loan and money interest
rates so as to keep their bill holdings in the right parish while house-
holds are continuously responding to relative interest rates by shifting
114 Stock–Flow Coherence and Economic Theory

0.150

0.100 Change (first difference) in bonds

0.050

0.000
Change in cash
–0.050

–0.100

Figure 5.16 Simulation 3: Response of government liabilities (bonds and cash)

between interest bearing money and bills. But households, firms and
banks can only continue to function, in this chaotic world, if the gov-
ernment is continuously exchanging high powered money for bills on
demand. There has to be an active frontier at the spot where the residuals
pile up.

Simulation 4: changing the bond rate of interest

It is re-emphasized that the model, as it stands at present, generates
no negative feedback from higher interest rates either to-expenditure or
to asset prices which makes this simulation particularly unrealistic. The
model will only illustrate a limited set of responses and interactions, but
this it does very precisely.
The purpose of the fourth simulation is to show how money and loan
rates respond when the government changes the bill rate of interest.
Figure 5.17, which should be read with Figure 5.18, shows what happens
when bill rates are changed up and down in a rather wild way. When
bill rates go up there is a tendency for households to move out of money
into bills. This reduces the banks’ defensive asset ratio, causing them to
put up money rates of interest. Then, to preserve their profit margins,
banks put up loan rates, normally by the same amount as money rates.
But this is not the end of the story, for there is nothing so far to prevent
loan rates from falling, on occasion, below bill rates. For this reason it is
assumed, in the model, that banks set loan rates slightly above bill rates
when the normal margin over money rates is insufficient to generate
the normal hierarchy of rates. According to simulation 4, there is a brief
period towards the end of the period during which money rates exceed
bond rates. This arises because the model only allows the money rate
to adjust in stages towards any new desired level. But this may not be
 Macroeconomics without Equilibrium or Disequilibrium 115

0.0450

Loan rate
0.0400
Bond
0.0350 rate

0.0300
Money rate
0.0250

0.0200

Figure 5.17 Simulation 4: Interest rates

0.090 1
Additions to bond holdings
as a share of wealth
0.060
ii

0.030

0.000

–0.030

Additions to interest bearing

–0.060 money as a share of wealth

–0.090

Figure 5.18 Simulation 4: Households’ portfolio responses to interest rate changes

unrealistic? There will surely be at least some delay in the response of

households to relative interest rate changes which could produce such
an outcome.
Figure 5.19 shows changes in holdings of interest bearing money and
bills (expressed as shares of wealth) as a result of shifting relative interest
rates.
Figure 5.20 shows, in the top solid line, the total addition to the stock
of government liabilities and, in the two lower lines, the breakdown of
116 Stock–Flow Coherence and Economic Theory

0.20
0.0400

Banks’ bonds as a share

0.10 of money (LH scale) 0.0360

0.00
0.0320

–0.10
0.0280

–0.20 Money rate

of interest (RH scale) 0.0240

–0.30
0.0200

Figure 5.19 Simulation 4: Banks’ holdings of bonds and the money rate of interest

5.0

Total
4.0 liabilities

3.0

Bonds
2.0

1.0

0.0
Cash (“high powered money”)

Figure 5.20 Simulation 4: Additions to government liabilities and their make-up

this into bills and high-powered money. The rise in the total comes about
because, as can be seen in Figure 5.17, there is, by assumption, a pro-
gressive addition to interest rates and therefore to government interest
payments; and no particular significance attaches to this.
 Macroeconomics without Equilibrium or Disequilibrium 117

Concluding section

There are many ways in which the model deployed here could be
expanded, depending on which particular aspect of macroeconomics
one wished to explore. It could, for instance, provide a framework for the
study of fixed investment, the capital stock and equity; it could include
another country, comprise foreign trade multipliers for each of them and
establish a framework for studying the determination of exchange rates;
and it could represent productivity changes and wage and price infla-
tion. A start with some of these points have already been made in two
working papers (Godley 1996a and 1996b).
To come down to it, the present paper claims to have made, so
far as I know for the first time, a rigorous synthesis of the theory
of credit and money creation with that of income determination in
the (Cambridge) Keynesian tradition. My belief is that nothing the
paper contains would have been surprising or new to, say, Kaldor,
Hicks, Joan Robinson or Kahn. I look forward to hearing what Good-
hart says, particularly about the institutional aspects of the operations
described; but what I have written has been informed in part by a care-
ful reading of his work and I have no reason to suppose that we are
in significant disagreement. The paper could not have been written
without Tobin’s monumental contribution to the subject. Yet, in the
perception I at present have, and which may turn out to be quite mis-
guided, Tobin never makes the final step – essential to my story here –
where bank loans are required to enable industry to function at all;
the raison d’etre of Tobin’s banks, so far as I can see, is to enlarge the
asset choice of households and facilitate the agility with which it can
be made.

Appendix
The following table defines the variables and parameters of the model and gives
the numbers which have been attributed to each of them in order to obtain an
initial steady state. The number of equations exceeds the number of endogenous
variables listed below because the variables in the model describe values which
are expected, desired, exchanged etc. The suffixes are:

e Expected
h Held
r Required
m Made
x Exchanged
** Long-run steady state
* Short-run aspiration
118 Stock–Flow Coherence and Economic Theory

Exogenous variables
br = Banks’ normal bill holdings as a proportion
of money 0.0378
ρ = Fractional reserve ratio 0.1
G = Government expenditure 25
Ra1, Ra2 = Random variables with 0 mean
and normal distribution
rb = Rate of interest on bills 0.023
φ = Profit mark-up 0.1
φ1 , φ2 = Mark-ups of the loan rates 0.02, 0.01
σ = Inventory-sales ratio 0.275
τ = Rate of indirect tax 0.25

Endogenous variables
B = Total bill issue 67.5
Bb = Bills held by banks 1.2
Bh = Bills held by households 66.4
C = Consumption 107.8
F = Total profits 9.8
Fb = Banks’ profits 0.1
H = Total cash 13.8
Hb = Banks’ cash 3.1
Hh = Households’ cash 10.7
IN = Inventories 26.4
L = Bank loans 26.4
M = Interest bearing money 27.7
M1 = Non-interest bearing money 2.9
S = Total sales valued at market prices 132.8
SC = Total sales valued at cost 96.0
T = Yield of taxes 26.5
V = Wealth 107.8
WB = Wage bill 95.9
Y = GDP 132.8
p = Ratio of final sales to ditto at cost 1.38
rm = Rate of interest on money 0.02
rl = Rate of interest on loans 0.024

Parameters

The equations of the model

Firms

Ff = S − T − WB + ΔIN − rl · IN–1 (A.1)

S = C+G (A.2)
 Macroeconomics without Equilibrium or Disequilibrium 119

WB = SCe + IN ∗ − IN–1 (A.3)

SCe = SC[+Ra1] (A.4)

Se = (1 + τ ) · (1 + φ) · (WB − ΔIN ∗ + rl · IN–1 ) (A.5)

Se
p= (A.6)
SCe
SC = S/p (A.7)

ΔIN = γ · (IN ∗∗ − IN–1 ) (A.8)

IN ∗∗ = σ · WB (A.9)

ΔIN = ΔIN ∗ − (SC − SCe ) (A.10)

ΔLr = ΔIN (A.11)

Households

YD = F + WB + rm · M–1 + rb · Bhh−1 (A.12)

F = Ff + Fb (A.13)

ΔV = YD − C (A.14)

C = α1 · YDe + α2 · V–1 (A.15)

YDe = YD∗∗ [+Ra2] (A.16)

G + rb · B–1
YD∗∗ = −G (A.16a)
τ/(1 + τ )

ΔV e = YDe − C (A.17)

Vnc = V − Hhh (A.18)

e = Ve − He
Vnc (A.19)
hh

Hhh = αc · C (A.20)

M1∗h YDe
e = λ01 − λ11 · rm − λ21 · rb + λ31 · e (A.21)
Vnc V
 
M YDe
A22a) eh = λ02 + λ12 · rm − λ22 · rb − λ32 · e (A.22a)
Vnc V

Bhh YDe
e = λ03 − λ13 · rm + λ23 · rb − λ33 · e (A.23)
Vnc V

Mh = Vnc − M1h − Bhh (A.22)

M1h = M1∗h + YD − YDe (A.24)

120 Stock–Flow Coherence and Economic Theory

Banks

Bbh = M1x + Mx − Lm − Hbh (A.25)

Fb = rl · Lm−1 + rb · Bbh−1 − rm · Mx−1 (A.26)
rl1 = (1 + φb1 ) · rm (A.27a)
rl2 = (1 + φb2 ) · rb (A.27b)
rl = rl1 · z1 + rl2 · z2
z1 = 1, rl1 > rb ; 0, rl < rb
z2 = 1, rl1 > rb ; 0, rl > rb (A.27)
Δrm = z3 · A1 − z4 · A1
z3 = 0, br < top; 1, br > top
z4 = 0, br > top; 1, br < top
Bbh
br = (A.28)
M1x + Mx
Hbh = ρ · (M1x + Mx ) (A.29)
M1x = M1h (A.30)
Mx = Mh (A.31)
Hhx = Hhh (A.32)
Lm = Lr (A.33)

Government

ΔBx = G + rb · Bx−1 − T − ΔHx (A.34)

T = S · τ/(1 + τ ) (A.35)
Hbx = Hbh (A.36)
Bbx = Bx − Bhx (A.37)
Hx = Hhx + Hbx (A.38)
Bbx = Bbh (A.39)

Notes
1. I owe a special debt to George McCarthy who has helped and guided me
throughout in all manner of ways. I am also indebted to Stephanie Clark,
Anwar Shaikh and Malcolm Sawyer for extensive discussions; and to Robert
Solow and Lance Taylor who both wrote careful critiques of an earlier draft.
2. All the simulations were carried out using MODLER software.
 Macroeconomics without Equilibrium or Disequilibrium 121

3. Of course the IS-LM model can be re-solved using alternative assumptions

about the money stock. But this is not the same thing as ‘increasing the
money supply’ as a process in time. The IS-LM diagrams make me think of
childrens” ‘pop up’ books which generate three dimensional images out of
two dimensional space; you can get a series of images but only by closing the
book and opening it at different pages!
4. Surely the absence of the asset side of banks’ balance sheets – the loans they
make – is a lacuna of great significance in Patinkin (1956)
5. The equations listed in the text only have an expository function and do not
constitute a complete system. The appendix model has equations numbered
A1, A2, etc.
6. The schema is very spare, concentrating largely on accountancy. But even this
is quite a big job! For instance, to bring enough precision to the model for
numerical solutions to be obtained, ten equations were needed to describe
firms, fifteen for households and eleven for banks. It was, in particular,
an intricate business getting inventories in materia pari with wages and expected
sales.
7. This obviously contrasts with the neoclassical assumption that firms are all on
their production frontiers producing at the spot where price equals marginal
cost. As Hicks (1989, p. 22) put it ‘There is no need to assume that there is
a single optimum output for which the plant is designed; it is better, being
more realistic, to think of it as having a regular range of outputs . . . which it is
. . . fitted to produce [and] …over that range marginal cost is simply running
cost per unit of output…which could be considered constant…’ The limit to
production is a matter beyond the scope of this paper but we protest that this
is not realistically described by a putative limit beyond which it is unprofitable
to fulfil an additional order.
8. The interest cost of holding inventories must be included among costs partic-
ularly if the definition of profits in Table 5.1 is to survive meaningfully when
inflation is introduced into the model. The term rl · IN–1 is identical to stock
appreciation (IVA) when the rate of interest equals the inflation rate. The uni-
versal convention used by national income accountants is simply to deduct
stock appreciation from gross profits but that is a crude and often inappropri-
ate adjustment e.g. when real interest rates are negative or fluctuate a great
deal.
9. It often happens that the two are added together in neoclassical texts,
notwithstanding that they are chalk and cheese, because together they
constitute the liabilities of the banking system and are therefore the end
product of the ‘money multiplier’ on which so much is supposed to
hang.
10. Was it a slip in Brainard and Tobin (1968) to make this argument in income
alone? This incomplete formulation has found its way into a number of texts.
11. For instance, one might perfectly well respond to the question ‘How could
you afford it?’ by saying ‘I got the money from a bank’. But this response
states that a bank loan has been granted which stands as a liability (i.e not
money) in the books of the respondent.
12. For the present simulation we assume that expectations are ‘model consistent’
that is, expected disposable income is the disposable income which the model
would generate, given the exogenous variables, in a steady state.
122 Stock–Flow Coherence and Economic Theory

13. This is surely what D.H. Robertson (1931) meant by ‘automatic lacking’!
14. N.B. The bottom line says that bill holdings are lower than they otherwise
would have been, not that they are negative!
15. See Godley and Cripps (1982) pp. 161–162.

References

Arena, R. and Graziani, A. (eds) (1985) Production, circulation et monnaie (Paris:

Presses Universitaires de France).
Backus, Brainard, Smith and Tobin (1980) ‘A Model of U.S. Financial and Non-
Financial Economic Behaviour.’ Journal of Money, Credit and Banking 12.
Blinder, A.S. and R.M. Solow (1973) ‘Does Fiscal Policy Matter.’ Journal of Political
Economy: 319–337.
Brainard, W.C. and J. Tobin (1968) ‘Pitfalls in Financial Modelling.’ American
Economic Review 38: 98–154.
Christ, C. (1967) ‘A Short Run Aggregate Model of the Interdependence and
Effects of Monetary and Fiscal Policy.’ American Economic Review 57, Papers
and proceedings.
Godley, W. (1996a) ‘Money, Finance and National Income Determination: An
Integrated Approach.’ Jerome Levy Institute Working Paper No. 167.
Godley, W. (1996b) ‘A Simple Model of the Whole World with Free Trade,
Free Capital Movements and Floating Exchange Rates.’ Jerome Levy Institute
mimeo.
Godley, W. and T.F. Cripps (1983) Macroeconomics (Fontana and OUP).
Hicks J.R. (1974) The Crisis in Keynesian Economics (Jahnsson Lectures).
Hicks J.R. (1989) A Market Theory of Money (Oxford: Clarendon Press).
Kaldor N. (1985) Economics without equilibrium (M.E. Sharpe).
Modigliani F. (1963) ‘The Monetary Mechanism and Its Interaction with Real
Phenomena.’ Review of Economics and Statistics: 79–107.
Patinkin D. (1956) Money, Income and Prices (Harper & Rowe).
Robertson D.H. (1940) Essays in Money and Interest (Fontana).
Tobin J. (1969) ‘A General Equilibrium Approach to Monetary Theory.’ Journal of
Money, Credit and Banking (February).
Tobin J. (1982) ‘Money and Finance in the Macroeconomic Process.’ Journal of
Money, Credit and Banking 14.
6
Kaleckian Models of Growth in a
Coherent Stock–Flow Monetary
Framework: A Kaldorian View
Marc Lavoie and Wynne Godley

This paper integrates a stock–flow monetary accounting framework, as

proposed by Godley and Cripps (1983) and Godley (1993, 1996, 1999),
with Kaleckian models of growth, as proposed by Rowthorn (1981), Dutt
(1990), and Lavoie (1995). Our stock–flow accounting is related to the
social accounting matrices (SAM) originally developed by Richard Stone
in Cambridge, with double-entry bookkeeping used to organize national
income and flow of funds concepts. We present a consistent set of sectoral
and national balance sheets where every financial asset has a counterpart
liability, and budget constraints for each sector describe how the bal-
ance between flows of expenditure, factor income and transfers generate
counterpart changes in stocks of assets and liabilities. These accounts are
comprehensive in the sense that everything comes from somewhere and
everything goes somewhere, or to put it more formally, all stocks and
flows can be fitted into matrices in which columns and rows all sum to
zero.1 Without this armature, accounting errors may pass unnoticed and
unacceptable implications may be ignored.
The paper demonstrates the usefulness of this framework when deploy-
ing a macroeconomic model, however simple. The approach was used
by Godley (1996, 1999) to describe an economy that tended towards a
stationary steady state, with no secular growth. In this paper, the same
methodology is used to analyse a growing economy.
A useful starting point for our study is the so-called neo-Pasinetti model
proposed by Kaldor (1966). In Kaldor’s model, the budget constraint
of the firm plays an important role in determining the macroeconomic
rate of profit, for a given rate of accumulation. In addition, through his
‘valuation ratio’, which is very similar to what later became known as
Tobin’s q ratio, Kaldor provides a link between the wealth of households

123
124 Stock–Flow Coherence and Economic Theory

and the financial value of the firms on one hand, and the replacement
value of tangible capital assets on the other.
One drawback to Kaldor’s 1966 ‘neo-Pasinetti’ model, as Davidson
(1968) was quick to point out, is that it does not describe a mone-
tary economy, for Kaldor assumed that households hold their entire
wealth in the form of equities and hold no money deposits. This
assumption gave rise to the bizarre conclusion that households’ propen-
sity to save has no effect on the steady-state macroeconomic profit
rate, a conclusion that gave the model its name.2 To take money
into account, Davidson proposed the concept of a ‘marginal propen-
sity to buy placements out of household savings’ (1972, p. 272; cf.
1968, p. 263), whereas Skott (1981) set out explicit stock–flow norms
linking the two components of wealth (money and equities) to the
consumption decision. The Skott model, in its various incarnations
(1988, 1989), is closest to the model used here, since Skott uses
explicit budget constraints with money/credit stocks for both firms and
households.
Our model extends Kaldor’s 1966 model by assuming that firms obtain
finance by borrowing from banks as well as by issuing equities. It includes
an account of households’ portfolio behaviour à la Tobin (1969), where
the proportion of wealth held in the form of money balances and equities
depends on their relative rates of return. It also includes an investment
function, which makes the rate of growth of the economy largely endoge-
nous. The model is Kaleckian because, in contrast with both Cambridge
models of growth à la Robinson and Kaldor, and also with classical mod-
els of growth (Duménil and Lévy 1999; Moudud 1999; Shaikh 1989),
rates of utilization in the long period are not constrained to their nor-
mal or standard levels.3 Our model develops a Kaldorian view because
it includes many features, such as mark-up pricing, endogenous growth
and flexible rates of utilization, as well as endogenous credit money and
exogenous interest rates, which Kaldor (1982, 1985) emphasized towards
the end of his career.4
The first section of this paper presents our social accounting matrices
and the second section gives the behavioural equations of the model.
The third section describes experiments in which we explore the effect
of changes in the propensity to consume, liquidity preference, the rate
of interest, the rate at which securities are issued, the retention ratio and
the real wage on variables such as the rate of accumulation, the rate of
profit, the rate of capacity utilization, Tobin’s q ratio and the debt ratio
of firms.
 A Kaldorian View 125

Table 6.1 Balance sheets

Households Firms Banks 

Money +Md −Ms 0

Equities +ed · pe −es · pe 0
Capital +K +K
Loans −Ld +Ls 0
 (net worth) +V K − (Ld + es · pe ) 0 +K

The social accounting framework

We have made many drastic simplifications in the service of trans-

parency. Our postulated economy has neither a foreign sector nor a
government, whereas banks have zero net worth. Firms issue no bonds,
only equities, and hold no money balances, implying that whenever
firms sell goods, they use any proceeds in excess of outlays to reduce
their loans. No loans are made to households, and there is no inflation.5
The balance sheet matrix of this economy is presented in Table 6.1,
whereas Table 6.2 gives the flow matrix that describes transactions
between the three sectors of the economy and which distinguishes, in
the case of firms and banks, between current and capital transactions.
Note that capital gains, which eventually have an effect on the stocks
of the balance sheet matrix, do not appear in the transactions matrix of
Table 6.2 since capital gains are not transactions. Symbols with plus signs
describe sources of funds, and negative signs indicate uses of funds. The
financial balance of each sector – the gap between its income and expen-
diture reading each column vertically – is always equal to the total of
its transactions in financial assets, so every column represents a budget
constraint.
The subscripts s and d have been added to relevant variables (denoting,
very roughly speaking, ‘supply’ and ‘demand’), the purpose of which is
to emphasize that each variable must make behavioural sense wherever it
appears. The inclusion of these subscripts in no way qualifies the obvious
fact that each row of the flow matrix must sum to zero; but we shall be
at pains to make explicit the means by which this equivalence comes
about. The watertight accounting of the model implies that the value of any
one variable is logically implied by all the other variables taken together. It also
implies that any one of the columns in Table 6.2 is logically implied by the
sum of the other four.
126 Stock–Flow Coherence and Economic Theory

Table 6.2 Transactions matrix

Firms Banks

Households Current Capital Current Capital 

Consumption −Cd +Cs 0

Investment +ls −ld 0
Wages +WBs −WBd 0
Net profits +FD −(FU + FD) +FU 0
Interest on −rl· · Ld –1 +rl · Ls–1 0
loans
Interest on +rm · Md –1 −rm · Ms–1 0
deposits
 in loans +Ld −Ls 0
 in money −Md +Ms 0
Issue of equities −ed · pe +es · pe 0
 0 0 0 0 0 0

In writing out our system of equations, each endogenous variable will

only appear once on the left-hand side (LHS), facilitating the counting
of equations and unknowns and making it easier for the reader to recon-
struct the whole model in his or her mind. When a variable does appear
on the LHS for a second time – therefore in an equation that is logically
implied by other equations – that equation will be numbered with the
suffixes a, b and so on.
Take the first column of Table 6.2. The regular income of households,
Yhr , is defined as the sum of all the positive terms of that column, wages
WBs , distributed dividends FD, and interest received on money deposits
rm · Md−1 , where rm is the rate of interest on money deposits, and Md−1
is the stock of money deposits held at the end of the previous period.
Yhr ≡ WBs + FD + rm · Md –1 (6.1)
Cs + Is ≡ WBd + F (6.1a)
From the first column of Table 6.1, we know that the wealth, V , of
households is equal to the sum of money holdings plus the value of
equity holdings:
V ≡ M d + ed · pe (6.2a)
where ed is the number of equities and pe is the price of equities. We can
rewrite Equation (6.2a) as:
Md ≡ V − [ed · pe ] (6.2)
 A Kaldorian View 127

where  is a first difference operator.

The second term on the right-hand side (RHS) of Equation (6.2) can
be written as:

[ed · pe ] ≡ (ed · pe ) − (ed –1 · pe–1 ) ≡ ed · pe + pe · ed –1 (6.2b)

which says that the change in the value of the stock of equities is equal
to the value of transactions in equities (ed · pe ) plus capital gains on
equities held at the beginning of the period (pe · ed−1 ).
We define the capital gains that accrue to households in the period
as CG:

CG ≡ pe · ed –1 (6.3)

The change in wealth, using column 1 of Table 6.2 again, as well as

Equations (6.1), (6.2), (6.3) and (6.2b), can be written as:

V ≡ Yhr − Cd + CG (6.4)

where Cd is consumption.
Rearranging Equation (6.4) allows us to retrieve the Haig–Simons
definition of income, Yhs , according to which income is the sum of
consumption and the increase in wealth.

Yhs ≡ Cd + V ≡ Yhr + CG (6.4a)

The current account of the firm sector, shown in column 2 of Table 6.2,
yields the well-known identity between national product and national
income.

Cs + Is ≡ WBd + F (6.1a)

where Is is investment and F is total profits. This equation, since it is

logically implied by the other four columns of Table 6.2, was dropped
when we came to solve the model.
Total profits F are made up of distributed dividends FD, retained earn-
ings FU, and interest payments on bank loans rl · Ld –1 , where rl is the rate
of interest on loans Ld –1 outstanding at the end of the previous period:

FU ≡ F − FD − rl · Ld –1 (6.5)

The capital account of the firm sector is given in column 3 of Table 6.2,
which shows the financial constraint of firms:

Ld ≡ Id − FU − es · pe (6.6)

128 Stock–Flow Coherence and Economic Theory

Equation (6.6) says that investment Id must be financed by some

combination of retained earnings, sale of new equities and additional
borrowing from banks.6 This is the budget constraint of firms that was
introduced by Kaldor (1966).
Our banking system is the simplest possible one. There is no govern-
ment sector, so a fortiori there is no government debt, no high-powered
money and no currency. This is a pure Wicksellian credit economy, where
all money takes the form of bank deposits. As an added simplification,
banks do not make profits, so the rate of interest on money deposits and
the rate of interest on loans are identical. With these assumptions, the
banks’ balance sheet is given by:

M s = Ls (6.7)

whereas its appropriation account implies:

rm = r l (6.8)

Behavioural relationships

Firms
Firms have four categories of decision to take. They must decide what
the mark-up on costs is going to be (see Coutts et al. [1978] and Lavoie
[1992, chapter 3]). In the present model, it is assumed that prices are set
as a mark-up on unit direct costs that consist entirely of wages. We have
a simple mark-up rule:

P = (1 + ϕ) · W/pr (6.9)

with p the price level, w the nominal wage rate, ϕ the mark-up, and where
pr is output per unit of labour such that:

Nd ≡ Y/pr (6.10)

where Nd is the demand for labour and output, Y, is:

Y ≡ Cs + Is (6.11)

We shall assume that the parameters in the above equations are all con-
stant, implying constant unit costs and constant returns to scale. The
wage rate is also assumed to be exogenous (and constant), and the mark-
up stays the same regardless of the degree of capacity utilization. These
are very strong assumptions made in order to bring a limited range of
problems into sharp focus. It will be not be difficult to amend them in
 A Kaldorian View 129

a later model. We also define units in such a way that the price level is
equal to unity, so that there is no difference between nominal and real
values.
Under these assumptions the main purpose of the pricing decision is to
determine the share of income between profits and wages. For instance,
since the total wage bill is WBd = (W/pr) · Y = W · Nd , and the total wage
income of households is WBs ≡ W · Ns , and since there is assumed to be
an infinitely elastic supply of labour,

Ns = N d (6.12)

total profits are given by:

F = {ϕ/(1 + ϕ)} · Y (6.13)

Entrepreneurs must next decide how much to produce. It is assumed

that firms fully adapt supply to demand within each period. This implies
that sales are always equal to output, and hence aggregate supply S is
exactly equal to aggregate demand, given by the sum of consumption
Cd and investment Id . We thus have the first of our two equilibrium
conditions, where equilibrium is achieved by a quantity adjustment
(an instantaneous one), as is always the case in standard Keynesian or
Kaleckian models:

Cs + Is = Cd + Id (6.14)

The third kind of decision made by firms concerns the quantity of

capital goods that should be ordered and added to the existing stock
of capital K – their investment. Because we have a growth model, the
investment function is defined in growth rates. We shall identify the
determinants of the rate of accumulation of capital gr k , such that:

Id = K = grk · K–1 (6.15)

Investment functions are controversial. In Kaldor (1966) there was

no investment function, the growth rate being exogenous. In Robinson
(1956) there was an investment function, where the rate of capital accu-
mulation depends on the expected profit rate. Some authors believe that
it is more appropriate to take the rate of capacity utilization and the nor-
mal rate of profit (rather than the realized one) as the determinants of
the investment function (Bhaduri and Marglin 1990; Kurz 1990). These
models usually assume away debt and money. Obviously, in a monetary
model, the interest rate and the leverage ratio should play a role. The
possibilities are endless.7
130 Stock–Flow Coherence and Economic Theory

We have decided to use the investment function recently tested empir-

ically by Ndikumana (1999). His model is inspired by the empirical
work of Fazzari and Mott (1986–1987), which they present as a Kalecki–
Steindl–Keynes–Minsky investment function. In the Ndikumana model,
there are four variables that explain the rate of accumulation: the ratio
of cash flow to capital, the ratio of interest payments to capital, Tobin’s q
ratio and the rate of growth of sales. We shall use the first three of these
and replace the fourth by the rate of capacity utilization, which was one
of the variables implicitly used by Fazzari and Mott.8 Before setting out
the investment function, we make the following five definitions.
The rate of capacity utilization u, which is the ratio of output to full-
capacity output Yfc :

u ≡ Ys/Yfc (6.16)

where the capital to full-capacity ratio σ is defined as a constant:

Yfc ≡ K/σ (6.17)

Tobin’s q ratio, which is the financial value of the firm divided by the
replacement value of its capital9 :

q ≡ V /K = (Ls + es · pe )/K (6.18)

The leverage ratio l, which is the debt-to-capital ratio of the firms:

l ≡ Ld /K (6.19)

The rate of cash flow rk , which is the ratio of retained earnings to capital:

rk = FU /K–1 (6.20)

The investment function, or, more precisely, the rate of capital accu-
mulation gr k , is given by Equation (6.21), with γ0 comprising exogenous
investment (‘animal spirits’) and all other γ ’s being (positive) parame-
ters. The parameters are all assumed to take effect after one period, on
the assumption that investment goods must be ordered and that they
take time to be produced and installed, and that entrepreneurs make
their orders at the beginning of the period, when they have imperfect
knowledge concerning the current period.

grk = γ0 + γ1 · rk–1 − γ2 · rl–1 + γ3 · q–1 + γ4 · u–1 (6.21)

In this model, as in the model tested by Ndikumana (1999), interest

payments have two negative effects; they enter the investment function
 A Kaldorian View 131

twice, once directly, but also indirectly, by reducing cash flow and there-
fore the ability to finance investment internally. The direct effect of high
interest payment commitments is to reduce the creditworthiness of firms
and increase the probability of insolvency, which may cause firms to slow
down their expansion projects; this is because entrepreneurs will be more
prudent, to ensure that they stay in business (Crotty 1996, p. 350); and
banks will be more reluctant to provide loans to firms with high debt
commitments.
Tobin’s q ratio is not usually incorporated into heterodox growth mod-
els with financial variables. For instance, it is not present in the models of
Taylor and O’Connell (1985) and Franke and Semmler (1989), although
these models do have some mainstream features, such as a fixed money
supply. The valuation ratio, however, is to be found in the investment
functions of Rimmer (1993) and Delli Gatti et al. (1990). The latter
refer to their investment function as a Keynes–Davidson–Minsky the-
ory of investment determination, citing Davidson (1972) and Minsky
(1975).10 Thus, it is clear that various Post Keynesians have considered
the introduction of the valuation ratio (the q ratio) as a determinant of
investment, although Kaldor himself did not believe that such a ratio
would have much effect on investment.11
Introducing the valuation ratio may reduce the rate of accumulation
decided by entrepreneurs whenever households show little desire to save
or to hold their wealth in the form of equities. As pointed out by Moore
(1973, p. 543), such an effect ‘leads back to the neoclassical conclusions
of the control of the rate of accumulation by saver preferences, albeit
through a quite different mechanism. A reward to property must be
paid … to induce wealth owners to hold voluntarily, and not to spend on
current consumption, the wealth accumulation that results from busi-
ness investment.’ We shall see that some of the usual conclusions of
Keynesian or Kaleckian models can indeed be overturned, depending on
the values taken by the reaction parameters, when the valuation ratio is
included as a determinant of the investment function.
There is nothing in the model to force the q ratio towards unity.
We could have written the investment function by saying that capital
accumulation is a function γ3 of the difference (q − 1). But this is like
subtracting γ3 from the constant in the investment function; it does not
imply q converges to unity in steady-state growth. For this to happen,
we would need to claim that the change in the rate of accumulation is
a function of the difference (q − 1). Formally, we would need to write
the difference equation: dgr k = γ (q − 1), so that gr k becomes a constant
when q = 1. In stationary neoclassical models, this result is achieved by
assuming that I = I(q − 1), as in Sargent (1979, p. 10).
132 Stock–Flow Coherence and Economic Theory

One may wonder where expectations enter the investment func-

tion, since (nearly) all the determinants of investment are one-period
lagged variables. For instance, in the investment functions of Taylor and
O’Connell (1985) and Franke and Semmler (1989), the rate of accumu-
lation depends on the current rate of profit augmented by a premium
that represents expectations of future rates of profit relative to the cur-
rent one. As a first step, these authors assume the premium to be an
unexplained constant. In elaborations of the model, the premium is an
inverse function of the debt ratio. In other words, it is assumed that
expected future rates of profit decline when debt ratios rise. We have a
similar mechanism by virtue of the term γ2 · rl · l–1 , on the grounds that
an increase in debt commitments will slow down accumulation. In addi-
tion, a change in the exogenous term in the investment function, γ0 ,
can represent a change in expectations regarding future profitability or
future sales relative to current conditions.
Finally, we consider the fourth category of decisions that firms must
take. Once the investment decision has been taken, firms must decide
how it will be financed. Which variable ought to be considered the resid-
ual one? Franke and Semmler (1991, p. 336), for instance, take equity
financing as a residual. However, they note that the recent literature on
credit and financial constraints may suggest, rather, that ‘debt financing
should become the residual term to close the gap between investment
and equity finance’, and this is exactly what will be done here.12 Firms
borrow from the banks whatever amount is needed once they have used
up their retained earnings and the proceeds from new equity issues. As
Godley (1996, p. 4) suggests, bank loans ‘provide residual buffer finance’.
This has already been given a formal representation in Equation (6.6),
which gave the budget constraint of firms: Ld = Id − FU − es · pe .
We propose two behavioural equations: one determining the split
between distributed dividends and retained earnings, and the other
determining the amount of new equities to be issued. Distributed div-
idends are a fraction (1 − sf ) of profits realized in the previous period
(net of interest payments). Again, a lag is introduced on the ground that
firms distribute dividends each period on the basis of the profits earned
the previous period, having imperfect knowledge of soon-to-be-realized
profits.
It is assumed, however, that these distributed dividends are upscaled
by a factor that depends on the past rate of accumulation, to take into
account of the fact that the economy is continuously growing.

FD = (1 − sf ) · (FT–1 − rl–1 · Ls–2 ) · (1 + grk–1 ) (6.22)

 A Kaldorian View 133

This formulation of the dividend decision, though without the lags,

can be found in Kaldor’s 1966 model (FD = (1 − sf ) · FT ). Similarly, Kaldor
assumes that firms finance a percentage ξ of the investment expendi-
tures, regardless of the price of equities, or of the value taken by the
valuation ratio.13 This is clearly an oversimplification, but we shall adopt
it as an approximation, with a lag, so that:

es · pe = ξ · I–1 (6.23)

With the above two equations, and remembering that Kaldor assumes
away bank debt, Kaldor (and Wood [1975]) arrives at the following deter-
mination of the overall rate of profit: rf = grk · (1 − ξ )/sf , where rf = FT /K
is the overall rate of profit, and where gr k is the exogenous rate of
accumulation.
This equation is the source of Kaldor’s (1966) surprising belief that the
rate of household saving has no effect on the rate of profit, for a given
rate of growth. By contrast, when there is bank debt and money, the
budget constraint (omitting time lags) is telling us that:

(Id /K) = grk = sf · (FT − rl · Ld )/K + ξ · Id /K + (Ld /Ld ) · (Ld /K)

In the steady-state case, where bank debts are growing at the same rate
as the capital stock, that is, when Ld /Ld = grk , the equilibrium value of
the rate of profit is given by a variant of Kaldor’s equation:

rf = grk · (1 − ξ − l)/sf + rl · l

Thus, in steady-state growth, the rate of profit is positively related to

the rate of accumulation gr k and to the rate of interest on bank loans
rl .14 The problem here, however, is that the debt ratio of firms, l, can be
considered as a parameter, given by history, only in the short period. In
the long period, the debt ratio is among the endogenous variables, to be
determined by the model and dependent, among other things, on the
rate of household saving and the growth rate of the economy, so that
the above expression is hardly informative.15 Simulations will allow us
to observe the actual relationship between the rate of profit, the rate of
growth and the debt ratio.16

Banks
Banks make loans on demand and, obviously, they accept and exchange
deposits as well as pay and receive interest.

Ls = L d (6.24)
134 Stock–Flow Coherence and Economic Theory

The equality between loan demand and loan supply should be inter-
preted as representing the equality between the effective demand for
loans and the supply of loans.17 All creditworthy demands for loans
are granted in this system. In the present model, when debt commit-
ments increase, the symptoms of the crumbling creditworthiness of
firms, accompanied by a shift in the effective demand for loans (and
possibly in the notional demand for loans), appear as a downward shift
of the investment function Equation (6.21), under the negative effect of
the rl · l term representing debt commitments.
It would have been possible to make the rate of interest on loans a pos-
itive function of the debt ratio of firms, introducing a kind of Kaleckian
effect of increasing risk, but this would have simply compounded the
negative effect of high leverage ratios on investment.

Households
Households must decide how much they wish to consume and save,
thereby determining how much wealth they will accumulate. They must
also decide the proportions of their wealth they wish to hold in the form
of money and equities. We have already discussed, in the first section,
the budget constraint that households face when making these decisions.
Here we focus on behaviour.
Using a modified version of the Haig–Simons definition of income,
consumption is held to depend on expected regular household income
and on capital gains, which occurred in the previous period. When they
make their spending decisions, households still do not know exactly
what their income is going to be.18 The consumption equation is then:
∗ + (α /α) · CG
Cd = α1 · Yhr (6.25)
1 –1
with 0 < α1 < 1, α > 1, and
∗ = (1 + gr ) · Y
Yhr (6.26)
–1 hr –1

gr = Yhr /Yhr –1 (6.27)

where the asterisk (*) symbol represents expected values.

Expected regular household income is assumed to depend on the real-
ized regular household income of the previous period, and on the rate
of growth, gr, of regular household income the previous period. The
implication of such a consumption function is that unexpected income
increases are not spent in the current period, rather, they are saved, much
in line with the disequilibrium hypothesis put forth by Marglin (1984,
chapter 17) and other nonorthodox authors. This unexpected saving
 A Kaldorian View 135

is held entirely in the form of additional money deposits since the allo-
cation of wealth to equities has already been decided on the basis of
expected income. Thus actual money balances are a residual – they con-
stitute an essential flexible element of the system (Godley 2000, p. 18;
Lavoie 1984, p. 789).
Our consumption function is nearly the same as that suggested by
Kaldor (1966, p. 318) in a footnote to his neo-Pasinetti article, where
there is a single saving propensity for the household sector applying
equally to wages, dividends and capital gains. Here the propensity to
consume applies uniformly to wages, dividends and interest income. It
is doubtful, in a world of uncertainty, whether households would treat
accrued capital gains – that is, nonrealized capital gains – on the same
footing as regular income. Indeed, some empirical studies have found
no relationship between consumption and contemporaneous capital
gains. However, ‘studies that have included lagged measures of capital
gains have often found a significant impact’ (Baker 1997, p. 67). As a
result, we have assumed that only lagged capital gains enter the con-
sumption function, and that a smaller propensity to consume applies to
these gains.
It would have been possible to introduce a third element in the con-
sumption function, namely the stock of wealth accumulated previously,
V–1 , with a certain propensity to consume out of it, say α2 , an addi-
tion akin to the mainstream models of consumption (the life cycle and
the permanent income hypotheses). In models dealing with stationary
steady states without growth, such an addition is a necessary require-
ment, because, if the α1 coefficient is less than one, wealth must be
rising relative to income, without limit (Godley 1999, p. 396). How-
ever, in a growth model, wealth is continuously growing, and hence, the
standard Keynesian consumption function, with α1 < 1 and α2 = 0, is
adequate. In a growing economy, Equation (6.25), where consumption
only depends on flows of regular or accrued income, still makes it possi-
ble to incorporate the theory of credit, money and asset allocation into
that of income determination in a coherent way. We shall therefore stick
with the Kaldorian consumption function for the time being.19
Coming to households’ portfolio choice, we follow the methodol-
ogy developed by Godley (1999), and inspired by Tobin (1969).20 It is
assumed that households wish to hold a certain proportion λ0 of their
expected wealth V * in the form of equities (and hence a proportion [1−λ0 ]
in the form of money deposits), but that this proportion is modulated
by the relative rates of return on bank deposits and equities, and by the
transactions demand for money (related to expected household income).
136 Stock–Flow Coherence and Economic Theory

The two asset-demand functions are thus:

∗ /V ∗ )
(pe · ed )∗ /V ∗ = λ0 − λ1 · rm + λ2 · re−1 − λ3 (Yhr (6.28)

Md∗ /V ∗ = (1 − λ0 ) + λ1 · rm − λ2 · re−1 + λ3 (Yhr

∗ /V ∗ ) (6.28a)

where the λs are parameters, the * symbol again represents expected

values, and re(−1) is the rate of return obtained on equities in the previous
period. The rate of return on equities of the current period is defined as
the ratio of dividends received plus capital gains over the value of the
stock of held equities in the previous period.

re ≡ (FD + CG)/(pe–1 · ed –1 ) (6.29)

The two asset-demand functions are homogeneous in wealth, that is,

the proportions of the two assets being held does not vary in the long
run with the absolute size of wealth although, by virtue of the final
term in each function, there is a transactions demand for money that
can make a temporary difference. The two asset functions sum to one
because households are assumed to make consistent plans, symmetric to
the adding-up condition of Equation (6.2aA). Portfolio plans, under the
adding-up assumption, are thus:

Md∗ ≡ V ∗ − (ed · pe )∗ (6.30)

Equation (6.30) implies that one of the two asset–demand functions must
be dropped for the model to solve. And this is indeed what is done in the
simulations, Equation (6.28a), describing the money demand function
has been dropped and replaced by Equation (6.30).
Expected regular household income was defined by Equation (6.26).
Expected capital gains are assumed to depend on past capital gains and
the rate of accumulation of capital in the previous period, so that:

CG∗ ≡ (1 + grk–1 ) · CG–1 (6.31)

On the other hand, for households to have consistent plans, the expected
level of wealth must be in line with its expected budget constraint.
The realized budget constraint of households was already defined by
Equation (6.4). The following equation is its equivalent, within the realm
of expectations:

V ∗ = V–1 + Yhr
∗ + CG∗ − C
d (6.32)

When expectations and plans are fulfilled, the ratios targeted in

Equations (6.28) and (6.28a) will be exactly realized. In this case, the only
 A Kaldorian View 137

element of flexibility resides in the price of equities pe , since all the other
elements, including e – the number of equities – are pre-determined.
The price of equities will rise until the targeted ratio is attained since
there cannot be any discrepancy between the number of shares that
have been issued and the number of shares that households hold. In
other words, there has to be a price-clearing mechanism in the equity
market, such that:

ed = es (6.33)

What happens when expectations about regular income are mistaken? As

pointed out above, an extra element of flexibility resides in the amount of
money balances held by households. On the basis of their expectations,
regardless of whether they are realized or not, households invest in the
stock market in such a way that:

pe · ed = (pe · ed )∗ (6.34)

System-wide implications
We now have the same number of equation as unknowns, including
equations in both the ‘demand’ (Equation (6.2)) and the ‘supply’ of
money (Equation (6.7)). So the whole model is now closed and there is
therefore neither a need nor a place for an equilibrium condition such as:

Ms = M d (6.7a)

However, from the balance sheets of Table 6.1 we know that the equality
between the money deposits households find themselves holding and
the money deposits supplied by banks – which are equal to the loans
they have made – must invariably hold. Indeed, this property of the
model provides a way in which its accounting logic can, in practice, be
tested. Having solved the model, we can check the accounting, using the
simulations, to verify that the numbers do indeed generate Ms = Md . It
is only when an accounting error has been committed, that the equality
given by Equation (6.7A) will not be realized. With the accounting right,
the equality must hold. And in the present model, the equality holds
with no need for any asset price or interest rate adjustment.
If household income, and hence household wealth, turns out to be
different from expected levels, the adjustment factor is the amount of
money left with households, Md , compared with Md *.21 For instance,
suppose that actual household income is higher than its expected level:
Yhs > Yhs *. As a result, because consumption does not depend on actual
current income, there will be a corresponding gap between the actual and
138 Stock–Flow Coherence and Economic Theory

expected change in wealth: V > V *. As a consequence, the amount

of money held by households will be higher than what they expected
to hold by exactly the amount that income has been underestimated.
Formally, we have:22

V = V–1 + Yhr + CG − Cd (6.4)

V ∗ = V–1 + Yhr
∗ + CG∗ − C
d (6.32)

V − V ∗ = Yhs − Yhs
∗ (6.I)

V = M d + ed · pe (6.2a)
V ∗ = Md∗ + (ed · pe )∗ (6.30)

V − V ∗ = Md − Md∗ (6.II)

Md = Md∗ + (Yhs − Yhs

∗ ) (6.2c)

Equation (6.2c) shows that the planned demand for money can be differ-
ent from the realized one. In other words, we know that it is possible to
have: Ms > Md *. But this has no bearing on whether or not an excess sup-
ply of money can arise. This inequality is due to mistaken expectations;
it has no causal significance of its own. In particular, it cannot be said
that the excess money supply, defined here as Ms − Md *, can be a cause
of an excess demand on the goods market, or of an excess demand on
the equities market (which would push down financial rates of return).
It is for a moment, surprising that the stock of money people fetch up
with, whether or not they have made wrong predictions, is identically
the same amount as the loans that firms find that they have incurred –
although this follows from a distinct set of decisions. Our model is so
simple that it reveals with unusual clarity why this must be so. Kaldor’s
(1982) intuition – that there can never be an excess supply of money –
is vindicated.
Kaldor’s assertion has often been called into question. Some authors
have noted that, because money deposits are created as a result of loans
being granted to firms, money supply could exceed money demand.
Coghlan (1978, p. 17), for instance, says that: ‘If we accept that advances
can be largely exogenous . . . then the possibility must exist that bank
deposits can grow beyond the desires of money holders.’ That claim is
wrong, however. As shown here, and as explained informally by Lavoie
(1999), such a misunderstanding arises as a result of ignoring the overall
constraints imposed by double-entry financial bookkeeping.23
Finally, it should be pointed out that the seeds of our generalization
of Kaldor’s 1966 model to a monetary economy can already be found in
 A Kaldorian View 139

Joan Robinson’s works (1956, 1971).24 Robinson endorsed Kaldor’s neo-

Pasinetti theorem, with the proviso that ‘the banking system is assumed
to be generating a sufficient increase in the quantity of money to offset
liquidity preference’ (1971, p. 123). She had argued earlier that banks
must provide residual finance by writing that ‘banks must allow the total
of bank deposits to increase with the total of wealth,’ and that banks must
‘lend to entrepreneurs (directly or by taking up second-hand bonds), the
difference between rentier saving and rentier lending’ (Robinson 1956,
p. 277).25

Experiments

The model presented above was solved numerically and subjected to a

series of simulation experiments. First we assigned values to the var-
ious parameters using reasonable stylized facts. Then we solved the
model, and found a steady-state solution through a process of suc-
cessive approximations. Having found a steady state, we conducted
experiments by modifying one of the exogenous variables or one of
the economically significant parameters of the model at a time. The
advantage of this approach is that it is always possible to find out
exactly why the model generates the results it does. The disadvantage
is that we can only analyse local stability: we do not know if there
are other equilibria, or if these other equilibria are stable. What we
do show is that over a reasonable range of parameter values, includ-
ing, obviously, the values that we chose, the model does yield a stable
solution.
We quickly discovered that the model could be run on the basis of two
stable regimes.26 In the first regime, the investment function reacts less
to a change in the valuation ratio – Tobin’s q ratio – than it does to a
change in the rate of utilization. In the second regime, the coefficient of
the q ratio in the investment function is larger than that of the rate of
utilization (γ3 > γ4 ). The two regimes yield a large number of identical
results, but when these results differ, the results of the first regime seem
more intuitively acceptable than those of the second regime. For this
reason, we shall call the first regime a normal regime, whereas the second
regime will be known as the puzzling regime. The first regime also seems
to be more in line with the empirical results of Ndikumana (1999) and
Semmler and Franke (1996), who find very small values for the coefficient
of the q ratio in their investment functions, that is, their empirical results
are more in line with the investment coefficients underlying the normal
regime.
140 Stock–Flow Coherence and Economic Theory

Changes in the propensity to consume

Let us first consider changes in the propensity to consume. We shall
spend more space on this issue, because it is a particularly touchy one, as
indicated in the previous section. The paradox of thrift – a higher propen-
sity to consume or a lower propensity to save leads to faster growth – is
a crucial component of the Keynesian/Kaleckian school, in contrast to
the classical/Marxian models of growth or to the neoclassical models of
endogenous growth, where the opposite occurs. Here, whether the para-
dox of thrift occurs or not depends on the value taken by the coefficient
of the q ratio in the investment function.
In the normal regime the paradox of thrift holds. An increase in the
propensity to consume leads to an increase in the rate of accumulation,
both in the short period and in the long period, despite the fall in the
q ratio.
The logic of this result is the following. The increase in the propensity
to consume leads to higher rates of utilization and higher rates of profit,
both of which encourage entrepreneurs to increase the rate of accumu-
lation. The higher profits of entrepreneurs allow them to reduce their
dependence on debt and reduce the leverage ratio l. All of these effects
are shown in Figure 6.1, where, as in all following figures, the various
series are expressed as a ratio of the steady-state base case.

1.150
Profit rate
Utilization rate
1.100

Rate of accumulation
1.050

1.000

0.950
q-ratio
Debt ratio
0.900

1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 6.1 Higher propensity to consume, normal regime

 A Kaldorian View 141

1.100
Rate of return on equities

1.050

1.000
Growth rate of equity prices

0.950

0.900

0.850
1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008

Figure 6.2 Higher propensity to consume, effect on equities

On the other hand, the initial fall in savings is accompanied by a falling

demand for equities, which initially slows down the rate of increase in
the price of equities, and hence reduces the q ratio and the rate of return
on equities re (see Figure 6.2).27 The initial fall in re increases the demand
for money as a share of wealth. However, as profits and capital keep on
growing, the rate of return on equities recovers, and hence, in the new
steady state, the money-to-wealth ratio is lower than in the previous
steady state (Figure 6.3). Because entrepreneurs hardly react to the fall
in the q ratio, accumulation keeps going strong: its steady-state rate is
higher than that of the initial steady state, but it is lower than the pre-
viously achieved peak (Figure 6.1). The paradox of thrift holds in this
regime.
In the puzzling regime, the paradox of thrift does not hold. The faster
rate of accumulation initially encountered is followed by a floundering
rate, due to the strong negative effect of the falling q ratio on the invest-
ment function. The turnaround in the investment sector also leads to
a turnaround in the rate of utilization of capacity. All of this leads to a
new steady-state rate of accumulation, which is lower than the rate exist-
ing just before the propensity to consume was increased (see Figure 6.4).
Thus, in the puzzling regime, although the economy follows Keynesian
or Kaleckian behaviour in the short-period, long-period results are in
line with those obtained in classical models or in neoclassical models of
142 Stock–Flow Coherence and Economic Theory

1.030

1.020

1.010
Money to wealth ratio
1.000

0.990

0.980

1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 6.3 Higher propensity to consume, effect on money to wealth ratio

1.160

1.120

Utilization rate
1.080

1.040

Rate of accumulation
1.000

0.960

1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 6.4 Higher propensity to consume, puzzling regime

endogenous growth: the higher propensity to consume is associated with

a slower rate of accumulation in the steady state. In the puzzling regime,
by refusing to save, households have the ability over the long period
to undo the short-period investment decisions of entrepreneurs (Moore
 A Kaldorian View 143

1973). On the basis of the puzzling regime, it would thus be right to say,
as Duménil and Lévy (1999) claim, that one can be a Keynesian in the
short period, but that one must hold classical views in the long period.

Changes in the interest rate on loans and deposits

The key difference between the behaviour of the normal and the puzzling
regimes is the effect of a change in the (real) interest rate on loans (and
deposits). Recall that an increase in the interest rate has two effects on
effective demand. On one hand, as is shown in mainstream IS/LM mod-
els, an increase in the rate of interest has a negative effect on investment.
But on the other hand, an increase in interest rates has a favourable effect
on consumption demand and hence on the rate of capacity utilization,
since more income is now being distributed to households. This effect is
underlined in the models of stationary steady states presented by God-
ley (1999), where a higher interest rate leads to a higher stationary level
of output. The positive effect on effective demand, for a given level of
investment, is also present in Skott (1988), in a model that is closely
related to the present one.
In our model, with the chosen parameters, the negative investment
effect is initially strongest in both regimes. In the normal regime the
negative effect of the higher debt commitments carries over to the long
period (Figure 6.5). However, in the puzzling regime, despite the heavier

1.025

1.000

Utilization rate
0.975

0.950

0.925

Rate of accumulation
0.900

1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 6.5 Higher interest rate, normal regime

144 Stock–Flow Coherence and Economic Theory

1.120

1.080
q-ratio
Debt ratio
1.040

1.000
Rate of accumulation
0.960

0.920

1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 6.6 Higher interest rate, puzzling regime

debt commitments due to both the higher rate of interest and the higher
leverage ratio l, an increase in interest rates eventually drives up the
steady-state rate of accumulation to a level that exceeds the growth rate
associated with the lower rate of interest (Figure 6.6) – a rather surpris-
ing and counterintuitive result. This counterintuitive result justifies the
name puzzling, which we have attributed to this second regime.
In both regimes, despite an initial downward move, the steady-state
rate of utilization ends up higher than its starting value (see Figure 6.5).
In addition, the q ratio is quickly pushed upwards (see Figure 6.6), as
more disposable income allows households to spend more on equities.
This effect has particularly strong repercussions on capital accumulation
in the second regime, which explains why the increase in the rate of
interest drives up the steady-state rate of growth.
It may also be noted that in the normal regime, the higher lending rates
of interest are associated in the long period with lower rates of return on
equities, whereas in the puzzling regime there is a positive long-period
link between lending rates of interest and rates of return on equities.

Changes in the propensity to hold equities

The other experiments show little difference of behaviour between the
first and second regimes. For instance, in both regimes, a shift in liquid-
ity preference, out of money deposits and into equities, symbolized by
 A Kaldorian View 145

an increase in the λ0 parameter of the portfolio equations, leads to an

increase in the short- and long-period rate of accumulation. The view of
liquidity preference in the present model is consistent with that offered
by Mott (1985–1986, p. 230), according to whom ‘liquidity preference is
a theory of the desire to hold short- versus long-term assets’. Here, money
deposits are the short-term asset, whereas equities are the long-term one.
Our experiments give considerable support to the Post Keynesian belief
that liquidity preference, defined in a broad way, does matter in a mon-
etary economy. The favourable effect of lower liquidity preference can
be observed independently of any change in the confidence or animal
spirits of entrepreneurs or their bankers (as proxied by the γ0 coefficient
in the investment equation, or by the level of the real rate of interest).
Our model allows us to identify the mechanisms by which pure liquidity
effects can affect the real economy.
The favourable effect of the increasing desire of households to hold
equities instead of money can be attributed to two standard effects. On
one hand, the increase in the stock demand for equities pulls up the
price of equities and creates capital gains (Figure 6.7). These gains are
then partly consumed, thus raising the rate of capacity utilization, and
hence, in the next period, it shifts up the investment function. On the
other hand, the increase in the demand for equities pushes up the q ratio,

Growth rate of equity prices

2.00

1.60

1.20

Rate of return on equities

0.80

0.40

1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008

Figure 6.7 Stronger preference for equities, effect on equity prices and rate of
return
146 Stock–Flow Coherence and Economic Theory

1.050
Rate of accumulation

1.025
q-ratio

1.000

Debt ratio
0.975

Money to wealth ratio

0.950

1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 6.8 Stronger preference for equities, effect on other variables

an increase that also contributes to shift up the investment function. All

of these effects are accompanied by a lower money-to-wealth ratio and a
lower debt ratio, which also contributes to the faster accumulation rate
of the economy (all of these effects are shown in Figure 6.8).
There is a feedback loop that operates as a result of the initial increase
in the desire of households to hold securities; there is an acceleration
in the rate of growth of the economy and the rate of utilization rises.
All of this drives up the rate of return on securities re , thus reinforcing
the desire of households to reduce their money deposits relative to their
overall wealth.
Mott (1985–1986, p. 231) asserts that ‘liquidity preference is governed
primarily by the profitability of business’. In all of our experiments, the
steady-state values of the rate of accumulation and the rate of return
on equities moved in the same direction. Since the demand for equities
depends on the rate of return on equity, we may say that there is indeed
a link between the good performance of the economy and the preference
of households for long-term assets.28

Changes in real wages

A typical Kaleckian effect is also to be found in the present model. Assume
that there is a decrease in the mark-up ϕ, which, ceteris paribus, implies
 A Kaldorian View 147

that there is an increase in the real wage of workers, relative to their pro-
ductivity, (w/p)/pr.29 This means that the share of wages is now higher,
whereas that of profits is lower. In standard Kaleckian growth models,
an increase in the real wage leads to an increase in the long-period rate
of accumulation and in the long-period rate of capacity utilization (Dutt
1990; Lavoie 1995; Rowthorn 1981). The same result is obtained here.
The increase in real wages leads to an increase in consumption
demand, because firms will now be distributing more income to house-
holds while retaining less. As a consequence, the rate of capacity
utilization is pushed upwards. Note that the increase in capacity uti-
lization will only be felt one period later since consumption depends
on expected regular household income, rather than on realized regular
income.
Initially, in the short period, despite the increase in the rate of utiliza-
tion, the rate of profit of businesses falls, because of the lower mark-up.
This short-period result is in contrast with the result achieved in time-
continuous Kaleckian models, because in these models everything is
simultaneous, so that firms react immediately to the higher rate of uti-
lization by speeding up their rate of accumulation, generating higher
rates of profit in the process.
In the present model, by contrast, the rate of capital accumulation set
by firms depends on the variables of the previous period, and as a result
the increase in the rate of utilization induced by rising real wages has no
immediate effect on accumulation. In later periods, however, the rate
of accumulation starts recovering from the lower rate of profit initially
induced by the lower mark-up. Over time, the faster accumulation helps
to improve profitability. In the long period, the rate of accumulation is
much higher with higher real wages, whatever the regime of the model.
In the normal regime, the more likely one, the rate of profit does not
totally recover.30 This last result, as pointed out above, is in contrast
with the time-continuous Kaleckian models of growth. In addition, the
lower mark-up set by firms leads to a higher debt ratio, a not-so-obvious
result. All of these effects are shown in Figure 6.9.

Changes in parameters controlled by the firms

When discussing the behaviour of firms, it was assumed that firms had
the ability to set the number of equities they wished to issue each period –
a rule was given according to which firms financed ξ per cent of their
investment by issuing new shares – and that firms chose a retention ratio
on profits (net of interest payments). What happens when firms decide
to change these percentages?31
148 Stock–Flow Coherence and Economic Theory

1.080
Utilization rate Debt ratio

1.050

1.020

Rate of accumulation
0.990
Profit rate
0.960

0.930

1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007

Figure 6.9 Higher wage share

First, consider the case when the ξ ratio is increased. Firms issue more
securities. This leads to an initial fall in the rate of growth of equity
prices, and hence to a fall in the q ratio. This fall induces a capital loss,
and hence, a slowdown in consumption demand growth. This slowdown
leads to a fall in the rate of utilization, and hence, in the cash flow of
firms. The fall in these two determinants of the rate of accumulation, as
well as the fall in its third determinant – the q ratio, leads to a permanent
slowdown in the rate of accumulation, as shown in Figure 6.10. The only
positive effect of issuing more securities is that the debt ratio is reduced,
but this appears to be a second-order effect (not shown here).
If the model correctly describes the behaviour of a true economy, the
reluctance of companies to issue equities may appear to be well-founded.
Larger issues of equities have detrimental effects on a monetary economy,
leading to a fall in the growth rate, the rate of profit and the rate of return
on equities. Reciprocally, when companies buy back their shares from
households, as done in the late 1990s, it should have a positive effect on
the overall economy.
Let us now consider the case of an increase in the retention ratio of
firms. This increase has two contradictory effects on effective demand.
On one hand, it automatically increases the cash flow that is available
to firms to finance their investments, thus pushing up the investment
function. In addition, firms have to borrow less, and hence can reduce
 A Kaldorian View 149

1.040

1.000
Rate of accumulation
q-ratio
0.960

0.920

Growth rate of equity prices

0.880

0.840

1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008

Figure 6.10 Larger issues of equities

1.080

1.040
Cash-flow ratio

1.000
Rate of accumulation

0.960 Debt ratio

Consumption growth rate

0.920

1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008

Figure 6.11 Higher retention ratio

their debt ratios. On the other hand, households are left with less regular
income, and hence, the rate of growth of consumption demand slows
down. With the chosen parameters, the positive effects on the rate of
accumulation initially overwhelm the negative ones, but over the long
period, an increase in the retention ratio does have a negative effect
150 Stock–Flow Coherence and Economic Theory

on the rate of growth of the economy. All of these effects are shown
in Figure 6.11. In the steady state, there is also a negative effect on the
overall rate of profit and the rate of return on equities.

Conclusion

Post Keynesian economics, as reported by Chick (1995), is sometimes

accused of lacking coherence, formalism and logic. The method pro-
posed here is designed to show that it is possible to pursue heterodox
economics, with alternative foundations, which are more solid than
those of the mainstream. The stock–flow monetary accounting frame-
work provides such an alternative foundation that is based essentially
on two principles. First, the accounting must be right. All stocks and all
flows must have counterparts somewhere in the rest of the economy. The
watertight stock–flow accounting imposes system constraints that have
qualitative implications. This is not just a matter of logical coherence; it
also feeds into the intrinsic dynamics of the model.
Second, we need only assume, in contrast to neoclassical theory, a very
limited amount of rationality on the part of economic agents. Agents act
on the basis of their budget constraints.32 Otherwise, the essential ratio-
nality principle is that of adjustment. Agents react to what they perceive
as disequilibria, or to the disequilibria that they take note of, by mak-
ing successive corrections.33 There is no need to assume optimization,
perfect information, rational expectations or generalized price-clearing
mechanisms.
Another feature of the present analysis is the simulation method. With
simulations, a full model can be articulated and its properties ascer-
tained and understood, without the need to resort to reduced forms.
The simulation method enables one to penetrate, with one’s understand-
ing, dynamic models of far greater complexity than can be handled by
analytic means. Indeed, even practitioners of multidimensional stabil-
ity analysis resort to simulations to figure out how their models behave
(see, for instance, Flaschel et al. 1997). Nonlinearities can be easily intro-
duced. For instance, we can program behaviour to change whenever a
variable exceeds or drops below some threshold level, as in the model
of Godley (1999). In that model, the steady state was stationary. It is
quite possible, however, to superpose the present model to that previ-
ous model, to obtain a growth model with highly complex but coherent
features. These would include a government sector, a detailed banking
sector, and consumption and production that occur in real time, with
 A Kaldorian View 151

inventories, and with output supply not being generally equal to output
demand.
Although narration and verbal explanation are in order – indeed
essential – we are suggesting a method that has much rigour and demon-
strability. In our methodology, we can justify every point by reference to a
precise system of relationships. If others disagree, they can be challenged
to say precisely what simplification or parameter is inappropriate. Every
relationship can be changed, and one can find out whether the change
makes any difference to the results. This method ought to be helpful to
resolve some controversial issues. For instance, we have shown how and
why an excess supply of money can never occur.

Notes
1. This method was first put to use by Backus et al. (1980), as far as we know.
2. ‘The rate of profit in a Golden Age equilibrium . . . will then be independent
of the “personal” savings propensities . . .. In this way, it is similar to the
Pasinetti theorem . . .. It will hold in any steady growth state, and not only in
a “long-run” Golden Age’ (Kaldor 1966, p. 318).
3. As in other Kaleckian models, it will be assumed that parameters are such that
the rate of capacity utilization does not exceed unity.
4. See Lavoie (1998) for an analysis of Kaldor’s 1966 model with endogenous
rates of capacity utilization. There is evidence that Kaldor (1982, pp. 49–50)
was aware of stock–flow accounting constraints.
5. See Palley (1996) for an analysis of household debt. Of course it would be
possible within the present model to suppose that households borrow to
speculate on the stock market.
6. To avoid any confusion with the simplifying accounting assumptions used
in other works (such as Backus et al. 1980, p. 268; Dalziel 1999–2000, pp.
234–235), it should be pointed out that retained earnings are not imputed
to shareowners as if they were dividends or as if they were an issue of new
equities to existing shareowners, and capital gains are not imputed to existing
shareowners in the form of an implicit equities issue.
7. For instance, the investment function proposed by Dutt (1995) includes the
cash flow ratio, the debt ratio and the rate of utilization.
8. The suggested investment function is also supported by the empirical work
of Semmler and Franke (1996).
9. Some authors prefer to define the q ratio as: q’ = (es · pe )/(K − L). We then have
q’ = (q − l)/(1 − l).
10. ‘[The article] “Money, Portfolio Balance, Capital Accumulation and Economic
Growth,” written in 1965 . . . presented an alternative approach to money and
capital accumulation more in tune with Keynes’s General Theory and Treatise
on Money. This alternative to Tobin’s 1965 accumulation analysis involved
utilizing the forward market price for capital (that is, the market price of
existing real capital relative to the cost of producing real capital) as the rele-
vant “invisible hand” ratio directing the entrepreneurial determination of the
152 Stock–Flow Coherence and Economic Theory

rate of investment or disinvestment in real capital. This ratio, is of course, the

equivalent of the famous q-ratio that Tobin was to discover in 1968’ (Davidson
1992, p. 111).
11. ‘The stock exchange value of a company can fall to say one half of the value of
the assets employed in the business. But this does not change the decision as
to whether it is worthwhile to undertake some investment or not; the implicit
rate of return would only become relevant to the firm’s decisions if the normal
method of financing investment were to be the issue of ordinary shares for
cash – which in fact plays a very small role. Most of the profits come from
ploughed back profits, in which case the expected internal rate of return is
relevant and not the implicit rate of return’ (Kaldor, November 9, 1983, in a
letter to one of the authors).
12. It is also what Flaschel et al. (1997, p. 357) end up doing themselves.
13. Alternative formulations would have been possible. For instance, Marris
(1972) and Skott (1988) assume that the stock of issued securities grows at
a constant rate gs . That rate could also be assumed to be higher when the
valuation ratio exceeds unity.
14. Here, because there is no price inflation, all growth rates are in real terms: the
rate of interest is the real rate of interest.
15. Since the propensity to save of households has an effect on the debt ratio l,
it also has an effect on the rate of profit, even if there is no change in the rate
of growth. Thus, as guessed by Davidson (1968, 1972), introducing money
into Kaldor’s neo- Pasinetti model does change the main feature that gave it
its name!
16. Computing the steady-state value of the debt ratio l yields an extraordinarily
complicated equation, even in such a simple model.
17. The expression ‘effective’ demand for loans, to denote the demand from
creditworthy customers, is utilized by both Lavoie (1992, p. 177) and Wolfson
(1996, p. 466).
18. This is one of the crucial aspects that distinguish the present model from that
of Skott (1988).
19. It should be pointed out, however, that Kaldor was fully aware that wealthy
households could consume without ever having to declare any taxable
income. Even if a portion of realized capital gains were to become part of
taxable income, these wealthy families could dodge taxation altogether by
borrowing their way into consumption, getting loans for consumption pur-
poses, secured on the basis of their large assets, thus slowly depleting their
net assets. This is why Kaldor wished to have an expenditure tax replacing
the income tax.
20. See Panico (1993, 1997) and Franke and Semmler (1989, 1991) for models that
purport to integrate Tobin’s portfolio adding-up constraint approach with
Kaldor’s growth models.
21. This assumption can be found in Godley (1996, p. 18): ‘It is assumed that mis-
taken expectations about disposable income turn up as differences in holdings
of [money deposits] compared with what was targeted.’
22. Equation (6.2c) is the result of combining Equations (6.I) and (6.II), which,
given Equations (6.34) and (6.4a), arise from the subtractions shown below.
23. By contrast, Godley (1999) shows how, in a world with a more sophisticated
banking system, the path of loans and deposits can diverge. But the question
 A Kaldorian View 153

of the equality between the demand for, and the supply of, money is an
entirely different issue.
24. See Rochon (1999, chapter 4) for an overview of Robinson’s unjustly neglected
analysis of endogenous credit money.
25. The reader will see some similarity with Davidson’s (1972, p. 335) analysis
of growth when the so-called excess flow-demand for securities is negative.
See also Dalziel (1999–2000) for a symmetrical analysis when the excess flow-
demand for securities is positive.
26. Some parameter values yielded unstable behaviour.
27. Figure 6.2 shows substantial cyclical fluctuations in the stock market, which
are due to the mechanical way in which portfolio decisions are taken. Still
the variables do converge to their steady-state values.
28. From the budget constraint of firms, and from the definition of the rate of
return on equities, it can be shown that, in the steady state, re = {rf − rl · l +
grk · (q − 1)}/(q − l).
29. In the simulations of the model, the mark-up ϕ is reduced, whereas the nom-
inal wage rate w is simultaneously increased, to keep output prices constant
at p = 1.
30. In the puzzling regime, however, the steady-state rate of profit with higher
real wages is much higher than that with low real wages.
31. It turns out that the regime of the model does not matter.
32. Other authors, mainly heterodox ones, have made use of balance sheets, to
secure appropriate accounting foundations, and of Tobin’s adding-up con-
straint, to achieve portfolio equilibrium, for instance, Franke and Semmler
(1989, 1991). But although the stock matrix is given a great deal of atten-
tion, the flow matrix is sometimes left out, especially when dealing with the
banking sector.
33. Duménil and Lévy (1995, p. 370) strongly advocate the same adjustment
principle.

References

Backus, D., W.C. Brainard, G. Smith, and J. Tobin (1980) ‘A Model of U.S. Financial
and Nonfinancial Economic Behavior.’ Journal of Money, Credit, and Banking 12
(2): 259–293.
Baker, D. (1997) ‘Conceptual and Accounting Issues in the Analysis of Saving,
Investment, and Macroeconomic Activity.’ In Robert Pollin (ed.), The Macro-
Economics of Saving, Finance and Investment (Ann Arbor: University of Michigan
Press), pp. 35–82.
Bhaduri, A. and S. Marglin (1990) ‘Unemployment and the Real Wage: The Eco-
nomic Basis for Contesting Political Ideologies.’ Cambridge Journal of Economics
14 (4) (December): 375–390.
Chick, V. (1995) ‘Is There a Case for Post Keynesian Economics?’ Scottish Journal
of Political Economy 42 (1) (February): 20–36.
Coghlan, R. (1978) ‘A New View of Money.’ Lloyds Bank Review 129 (July): 12–27.
Crotty, J.R. (1996) ‘Is New Keynesian Investment Theory Really Keynesian? Reflec-
tions on Fazzari and Variato.’ Journal of Post Keynesian Economics 18 (3) (Spring):
333–358.
154 Stock–Flow Coherence and Economic Theory

Coutts, K., W. Godley, and W. Nordhaus (1978) Industrial Pricing in the United
Kingdom (Cambridge: Cambridge University Press).
Dalziel, P. (1999–2000) ‘A Post Keynesian Theory of Asset Price Inflation with
Endogenous Money.’ Journal of Post Keynesian Economics 22 (2) (Winter):
227–246.
Davidson, P. (1968) ‘The Demand and Supply of Securities and Economic Growth
and Its Implications for the Kaldor-Pasinetti Versus Samuelson-Modigliani
Controversy.’ American Economic Review 58 (3) (May): 252–269.
——. (1972) Money and the Real World (London: Macmillan).
——. (1992) ‘Davidson, Paul (born 1930).’ In P. Arestis and M. Sawyer (eds), A
Biographical Dictionary of Dissenting Economists (Aldershot, UK: Edward Elgar),
pp. 109–115.
Delli Gatti, D., M. Gallegati, and L. Gardini (1990) ‘Real Accumulation and Finan-
cial Instability: A Model of Profit Flows, Debt Commitments and Capital Asset
Prices.’ Studi Economici 41 (2): 101–126.
Duménil, G. and D. Lévy (1995) ‘Vrais et faux débats.’ Recherches économiques de
Louvain 61 (3): 359–394.
——. (1999) ‘Being Keynesian in the Short Term and Classical in the Long Term:
The Traverse to Classical Long-Term Equilibrium.’ Manchester School 67 (6)
(December): 684–716.
Dutt, A.K. (1990) Growth, Distribution and Uneven Development (Cambridge:
Cambridge University Press).
——. (1995) ‘Internal Finance and Monopoly Power in Capitalist Economies: A
Reformulation of Steindl’s Growth Model.’ Metroeconomica 46 (1) (February):
16–34.
Fazzari, S.M. and T.L. Mott (1986–1987) ‘The Investment Theories of Kalecki and
Keynes: An Empirical Study of Firm Data, 1970–1982.’ Journal of Post Keynesian
Economics Winter 9 (2): 171–187.
Flaschel, P., R. Franke and W. Semmler (1997) Dynamic Macroeconomics: Instability,
Fluctuation, and Growth in Monetary Economics (Cambridge: MIT Press).
Franke, R. and W. Semmler (1989) ‘Debt-Financing of Firms, Stability, and Cycles
in a Dynamical Macroeconomic Growth Model.’ In W. Semmler (ed.), Financial
Dynamics and Business Cycles: New Perspectives. (Armonk, NY: M.E. Sharpe), pp.
38–64.
——. (1991) ‘A Dynamic Macroeconomic Growth Model with External Financ-
ing of Firms: A Numerical Stability Analysis.’ In E.J. Nell and W. Semmler
(eds.), Nicholas Kaldor and Mainstream Economics: Confrontation or Convergence?
(London: Macmillan), pp. 335–359.
Godley, W. (1993) ‘Time, Increasing Returns and Institutions in Macroeconomics.’
In S. Biasco, A. Roncaglia, and M. Salvati (eds.), Market and Institutions in Eco-
nomic Development: Essays in Honour of Paulo Sylos Labini. (New York: St. Martin’s
Press), pp. 59–82.
——. (1996) ‘Money, Finance and National Income Determination: An Integrated
Approach.’ Working Paper No. 167, Jerome Levy Economics Institute of Bard
College, Annandale-on-Hudson, New York.
——. (1999) ‘Money and Credit in a Keynesian Model of Income Determination.’
Cambridge Journal of Economics 23 (2) (April): 393–411.
——. (2000) Monetary Macroeconomics. Unpublished manuscript, Jerome Levy
Economics Institute of Bard College, Annandale-on-Hudson, New York.
 A Kaldorian View 155

Godley, W. and F. Cripps (1983) Macroeconomics. (London: Fontana).

Kaldor, N. (1966) ‘Marginal Productivity and the Macro-Economic Theo-
ries of Growth and Distribution.’ Review of Economic Studies 33 (October):
309–319.
——. (1982) The Scourge of Monetarism (Oxford: Oxford University Press).
——. (1983) Personal correspondence, November 9.
——. (1985) Economics Without Equilibrium (Armonk, NY: M.E. Sharpe).
Kurz, H.D. (1990) ‘Technical Change, Growth and Distribution: A Steady-State
Approach to Unsteady Growth.’ In H.D. Kurz (ed.), Capital, Distribution and
Effective Demand: Studies in the Classical Approach. (Cambridge, UK: Polity Press),
pp. 210–239.
Lavoie, M. (1984) ‘The Endogenous Flow of Credit and the Post Keynesian Theory
of Money.’ Journal of Economic Issues 18 (3) (September): 771–797.
——. (1992) Foundations of Post-Keynesian Economic Analysis (Aldershot, UK:
Edward Elgar).
——. (1995) ‘The Kaleckian Model of Growth and Distribution and Its
Neo-Ricardian and Neo-Marxian Critiques.’ Cambridge Journal of Economics 19
(6) (December): 789–818.
——. (1998) ‘The Neo-Pasinetti Theorem in Cambridgian and Kaleckian Models
of Growth and Distribution.’ Eastern Economic Journal Fall 24 (4): 417–434.
——. (1999) ‘The Credit-Led Supply of Deposits and the Demand for Money:
Kaldor’s Reflux Mechanism as Previously Endorsed by Joan Robinson.’
Cambridge Journal of Economics 23 (1) (January): 103–114.
Marglin, S.A. (1984) Growth, Distribution, and Prices (Cambridge: Harvard Univer-
sity Press).
Marris, R. (1972) ‘Why Economics Needs a Theory of the Firm.’ Economic Journal
82 (March): 321–352.
Minsky, H.P. (1975) John Maynard Keynes (New York: Columbia University Press).
Moore, B.J. (1973) ‘Some Macroeconomic Consequences of Corporate Equities.’
Canadian Journal of Economics 6 (4) (November): 529–544.
Mott, T. (1985–1986) ‘Towards a Post-Keynesian Formulation of Liquidity Prefer-
ence.’ Journal of Post Keynesian Economics Winter 8 (2): 222–232.
Moudud, J.K. (1999) ‘Finance in Classical and Harrodian Cyclical Growth Mod-
els.’ Working paper 290, Jerome Levy Economics Institute, Bard College,
Annandale-on-Hudson, New York.
Ndikumana, L. (1999) ‘Debt Service, Financing Constraints, and Fixed Invest-
ment: Evidence from Panel Data.’ Journal of Post Keynesian Economics Spring 21
(3): 455–478.
Palley, T.I. (1996) Post Keynesian Economics: Debt, Distribution and the Macro
Economy (London: Macmillan).
Panico, C. (1993) ‘Two Alternative Approaches to Financial Model Building.’
Metroeconomica 44 (2) (June): 93–133.
——. (1997) ‘Government Deficits in Post-Keynesian Theories of Growth and
Distribution.’ Contributions to Political Economy 16: 61–86.
Rimmer, R.J. (1993) Income Distribution in a Corporate Economy (Aldershot, UK:
Edward Elgar).
Robinson, J. (1956) The Accumulation of Capital (London: Macmillan).
——. (1971) Economic Heresies: Some Old-Fashioned Questions in Economic Theory
(London: Macmillan).
156 Stock–Flow Coherence and Economic Theory

Rochon, L.-P. (1999) Credit, Money and Production: An Alternative Post-Keynesian

Approach (Cheltenham, UK: Edward Elgar).
Rowthorn, R.E. (1981) ‘Demand, Real Wages and Economic Growth.’ Thames
Papers in Political Economy Autumn: 1–39.
Sargent, T.J. (1979) Macroeconomic Theory (New York: Academic Press).
Semmler, W. and R. Franke (1996) ‘The Financial-Real Interaction and Investment
in the Business Cycle: Theories and Empirical Evidence.’ In G. Deleplace and
E.J. Nell (eds.), Money in Motion: The Post Keynesian and Circulation Approaches
(London: Macmillan), pp. 606–634.
Shaikh, A. (1989) ‘Accumulation, Effective Demand and Marx, Keynes and
Kalecki.’ In W. Semmler (ed.), Financial Dynamics and Business Cycles: New
Perspectives (Armonk, NY: M.E. Sharpe), pp. 65–86.
Skott, P. (1981) ‘On the Kaldorian Saving Function.’ Kyklos 34 (4): 563–581.
——. (1988) ‘Finance, Saving and Accumulation.’ Cambridge Journal of Economics
12 (September): 339–354.
——. (1989) Conflict and Effective Demand in Economic Growth (Cambridge:
Cambridge University Press).
Taylor, L. and S.A. O’Connell (1985) ‘A Minsky Crisis.’ Quarterly Journal of
Economics 100 (Supplement): 871–885.
Tobin, J. (1969) ‘A General Equilibrium Approach to Monetary Theory.’ Journal of
Money, Credit, and Banking 1 (1) (February): 15–29.
Part III
Stock–Flow Coherence and
Economic Policy
7
A Simple Model of Three
Economies with Two Currencies:
The Eurozone and the USA1
Wynne Godley and Marc Lavoie

Introduction

This paper presents a model which describes three countries trading

merchandise and financial assets with one another. It is initially assumed
that all three countries have independent fiscal policies but that two of
the countries share a currency – hence the model can be used to make
a preliminary analysis of the conduct of economic policy in ‘the euro-
zone’ vis-à-vis the rest of the world –‘the USA’. It is assumed, as seems
most realistic nowadays, that the exchange rate between the eurozone
and the USA is freely floating. The main conclusion will be that, if all
three countries do indeed operate independent fiscal policies, the sys-
tem will work under a flexible rate regime, but only so long as the
European Central Bank (ECB) is prepared to modify the structure of
its assets by accumulating an ever-rising proportion of bills issued by
any ‘weak’ euro country. For instance, if one of the ‘euro’ countries
starts importing too much and makes no modification to fiscal pol-
icy, the ex ante effect will be to raise the proportion of bills issued by
that country and held by the ECB – in successive stages and without
limit. If this becomes unacceptable, at least within the confines of the
model (and always given the assumption about three independent fis-
cal policies), the interest rate of the deficit country must give way and
become endogenous. But this would bring about an exploding situa-
tion, as the interest rate of the weak country would need to increase
for ever. There is, in the end, only one lasting solution to this prob-
lem within the existing structure – to endogenise the fiscal policy of
the deficit country. Faced with rising imports, unable to devalue and
trapped by an unaccommodating central bank, the ‘weak’ country would
have, in the end, no alternative but to operate a restrictive fiscal policy

159
160 Stock–Flow Coherence and Economic Policy

that would have strong negative effects on output and employment.

The only alternative is to modify the structure of the eurozone arrange-
ments, either by forcing euro countries enjoying surpluses to pursue
expansionary fiscal policies or by expanding the spending power of
the European Union government, so it can engage in induced equal-
ization payments that transfer fiscal resources from surplus to deficit
countries.
Our model is at once complex and elementary. There is no investment;
firms hold neither tangible nor financial assets; there are no commer-
cial banks; there is no inflation, and prices are not explicitly modelled;
exchange rate expectations are omitted. Still, despite all this, the model
contains no less than 80 equations, many of which are accounting iden-
tities, but several of which are needed to describe the intricate financial
relationships between the various sectors of this three-country economy.
A large number of variables will be endogenous in all three countries,
namely, gross domestic product, disposable income, sales and consump-
tion; household wealth and its distribution between the various available
financial assets; taxes, government debt and the money supply; imports,
exports and the capital gains arising from exchange rate fluctuations;
the trade account, current account and capital account of the balance of
payments, and the exchange rate itself. By contrast, the crucial exoge-
nous variables of the main closure of the model will be the following:
government expenditures (net of debt servicing), tax rates, interest rates,
import propensities, propensities to consume and household portfolio
parameters.
Our model is an exemplar of what we believe to be stock–flow con-
sistent monetary macroeconomics. The method that we advocate is
based on a system-wide logic, where there are ‘no black holes’. It relies
on the intrinsic dynamics of stock–flow consistency, with explicit bud-
get constraints and adding-up constraints, and a multiplicity of sectors
and financial assets, as also advocated by Tobin (1982) and recently
reasserted by Dos Santos (2005). The approach integrates real activity
with its financial counterpart – hence the denomination: consistent mon-
etary macroeconomics. The relationship between real flows and financial
flows and balance sheets yields macroeconomic laws, the implications of
which are sometimes being omitted, but that constrain the possible evo-
lution of the described model (Godley and Cripps 1983, p. 18). We pay
considerable attention to the evolution of the various variables through
the short run and onto the long run, through the use of simulations. This
method, as applied to open-economy models, can be found in various
two-country incarnations in the works of Godley (1999a), Godley and
 A Simple Model of Three Economies with Two Currencies 161

Lavoie (2003, 2005–2006), Izurieta (2001, 2003), Lavoie (2003, 2006) and
Mouakil (2005). There is also a similar three-country euro/USA model by
Lequain (2003). Besides various minor discrepancies, the main differ-
ence between Lequain’s model and ours is that Lequain assumes a fixed
exchange rate between the euro and the dollar, whereas we consider a
flexible exchange rate regime.

The model

As pointed out by Taylor (2004, p. 206), ‘the best way to attack a prob-
lem in economics is to make sure the accounting is right’. We thus start
by presenting the balance sheet of our three-countries world economy
(Table 7.1, the stock matrix), as well as its flow matrix, which we call the
transactions-flow matrix (Table 7.2), in a way akin to the method put for-
ward by Godley (1999b). This latter matrix integrates national income
and product accounts (NIPA) with flow of funds accounts. Both of these
matrices will be highly useful when we outline the accounting identities
which must constrain the behaviour of this ‘world’ economy. The main
principle ruling the matrices is that every variable must be accounted for.
As a result, all rows and columns of the transactions-flow matrix must
sum to zero.2 As to the stock matrix, all rows of financial assets also ought
to sum to zero; columns do sum to zero when net worth is taken into
account. Notational symbols will be made clear as we run through the
equations of the model.
The national income identity for each country is3 :

Y$ = C$ + G$ + X$ − IM$ (7.1)
Y# = C# + G# + X# − IM# (7.2)
Y& = C& + G& + X& − IM& (7.3)

where Y is Gross Domestic Product (GDP), C household consumption,

G government expenditure, X exports and IM imports. The suffixes refer
to the three countries ($, # and &) being described. It is assumed that the
last two (# and &) share a single currency (the euro ¤).
Personal income, YP, is defined to include capital gains due to changes
in the exchange rate.4 The rate xr$ is the value of the dollar in euros
(that is the number of euros per dollar). This rate allows us to transform
all dollar-denominated variables into euros and hence allows us to sum
the components of the rows of our two matrices when the rows comprise
entries denominated in different currencies. The rate xr¤is the reciprocal
 Table 7.1 Balance sheet matrix

US$ Euroland

Households Govt. FED Exchange rate Households # Govt. # Households & Govt. & ECB Sum

Cash +H$ −H$ 0

+H# −H# 0

162
+H& −H& 0
Bills +B$$ −B$ +BFED$ xr$ +B#$ +B&$ +BECB$ 0
+B$# xr$ +B## −B# +B&# +BECB# 0
+B$& xr$ +B#& +B && −B& +BECB& 0
Net worth −V $ −V $G 0 xr$ −V # −V #G −V & −V &G −VECB 0
Sum 0 0 0 0 0 0 0 0 0
 Table 7.2 Transactions-flow matrix

USA Euroland

# Euro country & Euro country

Exch.
Hhlds Firms Govt FED Rate Hhlds Firms Govt Hhlds Firms Govt ECB Sum

Cons. −C$ +C$ −C# +C# −C& +C& 0

Govt. +G$ −G$ +G# −G# +G& −G& 0
Exp.
Trade −IM$ xr$ +X$# +X$& 0
+X$ xr$ −IM#$ −IM&$
+X&# −IM&# 0
−IM#& +X#& 0

163
GDP +Y$ −Y$ +Y# −Y# +Y& −Y& 0
Taxes −T $ +T $ −T # +T # −T & +T & 0
Interest +r$·B$$ −r$·B$ +r$·BFED$ xr$ +r$·B#$ +r$·B&$ +r$· BECB$ 0
payments
+r#·B#$ xr$ +r#·B## −r#·B# +r#·B&# +r#·BECB# 0
+r&·B#& xr$ +r&·B# & +r&·B&& −r&·B& +r&· BECB& 0
CB profits +F$ −F$ +F# +F& −FECB 0
Changes
in:
Cash −H$ +H$ −H# −H& +H#+H& 0
$ bills −B$$ +B$ −BFED$ xr$ −B#$ −B&$ −BECB$ 0
# bills −B$# xr$ −B## +B# −B&# −BECB# 0
& bills −B$& xr$ −B#& −B&& +B& −BECB& 0
Sum 0 0 0 0 0 0 0 0 0 0 0 0
164 Stock–Flow Coherence and Economic Policy

of xr$ (xr¤is the number of dollars per euro). We have:

YP$ = Y$ + r$−1 · B$$d−1 + r#−1 · B$#d−1 + r&−1 · B$&d−1

+ xr¤(B$#s−1 + B$&s−1 ) (7.4)

YP# = Y# + r$−1 · B#$d−1 + r#−1 · B##d−1 + r&−1 · B#&d−1

+ xr$(B#$s−1 ) (7.5)

YP& = Y& + r$−1 · B&$d−1 + r#−1 · B&#d−1 + r&−1 · B&&d−1

+ xr$(B&$s−1 ) (7.6)

where xr is the change in the relevant exchange rate, while r−1 is
the rate of interest (the suffix denoting the country to which the rate
refers) set at the end of the previous period, and which applies to
B−1 , the opening stock of government bills owned by each country’s
households.
The notational principle is that, when there are two currency sym-
bols (say $#), the first denotes the country in which a bill is held,
while the second denotes the country from which the bill originates.
Subscript d denotes ‘demand’, in which case the bills are denominated
in the currency of the country where the bills are held. For instance,
B$#d stands for bills held by households in the ‘$’ country, but issued
by the ‘#’ government, the bills being valued in ‘$’ currency. By con-
trast, subscript s will denote ‘supply’, and bills supplied are denominated
in the currency of the country where the bill in question originated.
For instance, B$#s will stand for bills held by households in the ‘$’
country and issued by the ‘#’ government, the bills being valued in ‘#’
currency.
Taxes, T , are defined as a proportion of personal income YP:

T $ = θ $ · YP$ (7.7)

T # = θ# · YP# (7.8)

T & = θ & · YP& (7.9)

where θ is the relevant tax rate. Since personal income was defined above
to include capital gains, taxation is assumed to apply both to regular
income and to capital gains.5
Personal disposable income, YD, is then what is left over from personal
income after taxes have been paid, and, since it takes into account capital
 A Simple Model of Three Economies with Two Currencies 165

gains, it is equivalent to a Haig–Simons measure of income

YD$ = YP$(1 − θ $) (7.10)

YD# = YP#(1 − θ #) (7.11)

YD& = YP&(1 − θ &) (7.12)

Wealth accumulation by the private sector is equal to the discrep-

ancy between disposable income (including net capital gains) and
consumption:

V $ = YD$ − C$ (7.13)

V # = YD# − C# (7.14)

& = YD& − C& (7.15)

where V is wealth.
The consumption functions are:

C$ = α1 · YD$ + α2 · V $−1 (7.16)

C# = α1 · YD# + α2 · V #−1 (7.17)

C& = α1 · YD& + α2 · V &−1 (7.18)

The lagged stock variable supplies the essential dynamic component

which will generate sequences in real time. Note that by virtue of the
identities (7.13)–(7.15), the consumption functions can alternatively be
written as wealth adjustment functions:

V = α2 (α3 · YD − V−1 ) (7.16a)–(7.18a)

where α3 = (1 − α1 )/α2 . Imports are determined by sales, S, and relative

prices (proxied by the exchange rate, when relevant).6 The notational
principle here is that the first currency symbol attached to a variable,
whether it is an export or an import, denotes the country into which
that object flows, the second denotes the country from which it flows.
Exports are valued in the currency of the exporting country; imports
are valued in the currency of the importing country. For instance, IM$#
are the imports into the ‘$’ country from the ‘#’ country, valued in ‘$’
currency. Imports into the USA increase when ‘$’ sales increase, and they
166 Stock–Flow Coherence and Economic Policy

decrease when the euro (xr¤) rises.7

im$# = μ0$# + μ1$# · s$ − μ2#$ · xr¤ (7.19)

im$& = μ0$& + μ1$& · s$ − μ2#$ · xr¤ (7.20)
im#$ = μ0#$ + μ1#$ · s# − μ2#$ · xr$ (7.21)
im#& = μ0#& + μ1#& · s# (7.22)
im&$ = μ0&$ + μ1&$ · s& − μ2&$ · xr$ (7.23)
im&# = μ0&# + μ1&# · s& (7.24)

where bold (lowercase) letters denote natural logs.

Sales are made up of domestic consumption and government expen-
ditures, as well as export sales.

S$ = C$ + G$ + X$ (7.25)
S# = C# + G# + X# (7.26)
S& = C& + G& + X& (7.27)

Remembering that the exports of one country to another need be equal

to the imports of the other country from the first country, provided they
are valued in the same currency, exports are defined in the following
way:

X$ = IM$# · xr$ (7.28)

X$& = IM$& · xr$ (7.29)
X#$ = IM#$ · xr¤ (7.30)
X#& = IM#& (7.31)
X&$ = IM&$ · xr¤ (7.32)
X&# = IM&# (7.33)

Exports and imports by each country are aggregated, generating the

following identities:

X$ = X#$ + X&$ (7.34)

X# = X$# + X&# (7.35)
X& = X$& + X#& (7.36)
 A Simple Model of Three Economies with Two Currencies 167

IM$ = IM$# + IM$& (7.37)

IM# = IM#$ + IM#& (7.38)
IM& = IM&$ + IM&$ (7.39)

We now move to standard if simplified portfolio equations. While we

assume perfect capital mobility, we only assume imperfect asset substi-
tutability. This means that agents will be modifying the structure of their
portfolio as interest rates change, but that no further change will be forth-
coming once the structure of interest rates remains given. This implies
that rates of return on different assets can remain unequal in equilib-
rium. We assume that households only hold domestic currency but that
they may choose to hold bills issued by any of the three governments.
The array of asset demands for ‘$’ residents is:

H$d /V $ = λ00$ − λ01$ · r$ − λ02$ · r# − λ03$ · r& (7.49a)

B$$d /V $ = λ10$ + λ11$ · r$ − λ12$ # · r# − λ13$ · r& (7.40)
B$#d /V $ = λ20$ − λ21$ · r$ + λ22$ · r# − λ23$ · r& (7.41)
B$&d /V $ = λ30$ − λ31$ · r$ − λ32$ · r# + λ33$ · r& (7.42)

where H is cash, more specifically the high-powered money issued by the

central banks.8 The subscript d denotes demand, and hence all assets are
valued in the domestic currency of the country in question, as pointed
out above. For simplicity, and also because formalizing such behaviour is
controversial, we assume away speculative behaviour based on expected
changes in exchange rates.
For ‘#’ residents the array of asset demands is:

H#d /V # = λ00# − λ01# $ · r$ − λ02$ · r# − λ03$ · r& (7.50a)

B##d /V # = λ10# − λ11# · r$ + λ12# · r# − λ13# · r& (7.43)
B#$d /V # = λ20# + λ21# · r$ − λ22# · r# − λ23# · r& (7.44)
B#&d /V # = λ30# − λ31# · r$ − λ32# · r# + λ33# · r& (7.45)

For ‘&’ residents the array is:

H&d /V & = λ00& − λ01& · r$ − λ02& # − λ03& · r& (7.51a)

B&&d /V & = λ10& − λ11& · r$ − λ12& · r# + λ13& · r& (7.46)
B$#d /V & = λ20& − λ21& · r$ + λ22& · r# − λ23& · r& (7.47)
B&$d /V & = λ30& + λ31& · r$ − λ32& · r# − λ33& · r& (7.48)
168 Stock–Flow Coherence and Economic Policy

The parameters in each array are constrained according to Tobinesque

principles, so that the sum of constants (the λi0 s) is equal to one, and
the sum of each of the other columns of parameters is zero. Note that, so
long as all interest rates are held constant, every ratio in the three arrays
must be a given constant.9
As the demand for money in each country is implied logically by the
sum of the three bill demand functions, this must be represented in the
model as a residual (which is why the previous money demand functions
did not carry standard equation numbers).

H$d = V $ − B$$d − B$#d − B$&d (7.49)

H#d = V # − B##d − B#$d − B#&d (7.50)
H&d = V & − B&&d − B&$d − B&#d (7.51)

The main closure: exogenous interest rates and

fiscal policies

Having identified the asset-demand equations, we must now find out

how demands and supplies of assets will get into equivalence; the way
this happens will determine possible closures. In our main closure, we
shall assume that interest rates on bills and fiscal policies are all exoge-
nous. It may be useful at this stage to take a quick glance back at the
balance sheet representing our overall world economy in Table 7.1.
There are three governments issuing bills (B$, B# and B&) but only two
central banks. The ‘Fed’ only holds bills issued by the ‘$’ government,
under the assumption that the ‘$’ currency is the international currency,
while the ECB holds bills issued by the ‘#’ and ‘&’ governments, as well as
bills issued by the ‘$’ government, which constitute its foreign reserves.
Since there is a common currency in the eurozone, we should further
assume, as long as interest rates are administered by the ECB, that the
rates of interest r& and r# are equal to one another.
It is a major, if obvious, feature of the model that total assets supplied
by each country are determined by their governments’ budget restraints.
Following the customary procedure of having only one variable on the
left-hand side of each equation, we have written these supplies as follows:

B$s = G$ + r$–1 · B$s–1 − T $ − F$ (7.52)

B#s = G# + r#–1 · B#s–1 − T # − F# (7.53)
B&s = G& + r&–1 · B&s–1 − T & − F& (7.54)
 A Simple Model of Three Economies with Two Currencies 169

Each government must issue a net amount of new bills whenever

its expenditures (pure expenditures plus debt service) exceed its rev-
enues (collected taxes plus the central bank profits F, which are entirely
returned to government).10
Central banks make profits out of the interest payments they get on
their assets, since their only liabilities are cash money, while operating
costs are assumed to be negligible. It is further assumed that the two euro
countries are of similar size, so that the profits of the ECB (called FEBC)
are split 50–50 to the two euro governments, as indicated below:

F$ = r$–1 · BFED$s–1 (7.55)

F# = (1/2)FEBC (7.56)
F& = (1/2)FEBC (7.57)
FEBC = r#–1 · BECB#s–1 + r&–1 · BECB&s–1 + r#–1 · BECB$s–1 · xr$
(7.58)

We now enter what can only be called the ‘Chinese puzzle’ stage of
our analysis.11 To ensure that we have an equation for every variable
and that no variable is ever determined by more than one equation (in
other words to facilitate the counting of equations and unknowns), we
write our model with a different variable on the left-hand side of each
equation. The way in which we have chosen to arrange the equations
may seem to be arbitrary on first reading, and there may indeed be more
than one way of doing this. But however it is done, it will be found to
be impossible to write down every supply on the left-hand side of an
equation without duplication; it will always be found that there are two
asset supplies too many. We shall solve this problem when we reach it.
As can be read from the last three rows of Table 7.1, the overall amount
of bills supplied by each government is identically made up of a series
of components, one of which will appear on the left-hand side of each
equation:

B$$s = B$s − B#$s − B&$s − BFED$s − BECB$s (7.59)

B$#s = B#s − B##s − B&#s − BECB#s (7.60)
BECB&s = B&s − B$&s − B#&s − B&&s (7.61)

The constraints on central banks, besides Equation (7.61), appear in

the following equations, starting with the Fed, which purchases bills
on the open market according to its needs in supplying cash (Equation
7.62). It is assumed that the ‘$’ Treasury supplies ‘$’ bills to the Fed on
170 Stock–Flow Coherence and Economic Policy

demand (Equation 7.63).

BFED$d = H$s (7.62)

BFED$s = BFED$d (7.63)

And then with the ECB, first recalling Equation (7.61), we need to have:

BECB&d = BECB&s (7.64)

BECB#s = BECB#d (7.65)
BECB#d = H#s + H&s − BECB$d − BECB$s · xr$ (7.66)
BECB$d = BECB$s · xr$ (7.67)

According to Equation (7.64), the ECB picks up any residual ‘&’ bill
left out by the market, as calculated through Equation (7.61). Equation
(7.65) indicates, as was the case for the Fed through Equation (7.63),
that the Treasury of country ‘#’ supplies the ECB with the ‘#’ bills that
it demands. These two equations contribute to keeping the euro rates
of interest, r# and r&, at the constant level set by the ECB. Equation
(7.66) reflects the balance sheet constraint of the ECB, and is expressed
in differences, since capital gains or capital losses on previously held
foreign reserves occur whenever the exchange rate xr$ gets modified, as
reflected in Equation (7.67). These capital gains are part of a revaluation
account and get reflected in an increase in the net worth of the ECB,
this net worth being called VECB in Table 7.1. The capital gains or losses
of the central bank have no impact whatsoever on the other elements
of its balance sheet; in particular, they have no impact on the money
supply. Because we are in a pure flexible exchange rate regime, BECB$s ,
the supply of foreign reserves, can be taken as a constant, since the cen-
tral bank lets market forces determine the exchange rate, without ever
intervening. Indeed, because these foreign reserves play no fundamental
role in the dynamics of a model with flexible exchange rates, we shall
assume that they are equal to nil in the numerical simulations. Under
such conditions, Equation (7.66) can be rewritten in a simpler form as:

BECB#d = H#s + H&s − BECB&d (7.66a)

As we shall see, the values of H#s and H&s are pre-determined by the
respective demands for cash, while BECB&d is itself pre-determined by
the amount of ‘&’ bills that the markets do not take up, but BECB#d can
take up any residual amount. There is no natural restriction whatever
about the share of each type of bill in the balance sheet of the ECB.
 A Simple Model of Three Economies with Two Currencies 171

There were 12 equations representing the demand for assets by house-

holds. We infer from the fact that interest rates are exogenous that
households’ demand for money is always met, making the ‘supply’
of money endogenous. In other words, the central banks are ready
to exchange money for bills on any scale whatever, at the chosen
interest rate.
H$s = H$d (7.68)
H#s = H#d (7.69)
H&s = H&d (7.70)
The nine bill supplies may initially be written as follows. First, we have
the five asset supplies where no exchange rate is involved:
B##s = B##d (7.71)
B&&s = B&&d (7.72)
B$$s = B$$d (7.73)
B#&s = B#&d (7.74)
B&#s = B&#d (7.75)
Next, we have the four supplies which have an exchange rate term:
B$&s = B$&d · xr$ (7.76)
B#$s = B#$d · xr¤ (7.77)
B&$s = B&$d · xr¤ (7.78)
B$#s = B$#d · xr$ (7.79a)
Out of these nine equations, there are two where the supply has already
appeared in a previous equation, that is, B$$s and B$#s , in Equations
(7.73) and (7.79a) above, are also on the left-hand side of Equations
(7.59) and (7.60). But we know that, since every row and every column
in the matrix sum to zero, there is always one equation which must be
‘dropped’ if the model is to be capable of solution. In this particular
case, the coherence of the accounting system as a whole will ensure that
Equation (7.73) holds, so long as every other equation is satisfied. In
other words Equation (7.73) is the ‘redundant’ equation, and it can be
‘dropped’ from the computer model.
The only way to include Equation (7.79a) in the model is to invert it
so that the exchange rate appears on the left-hand side.
xr$ = B$#s /B$#d (7.79)
172 Stock–Flow Coherence and Economic Policy

Equation (7.79) seems, for a moment, to imply that the exchange rate
is determined in a unique market, the market for B$# bills. But this is not
the case. The exchange rate, like every other endogenous variable, can
only appear a single time on the left-hand side of an equation. But the sys-
tem is a fully interdependent one such that the solution of the model as
a whole requires and ensures that every equation in which the exchange
rate appears is satisfied at the same time. Thus Equations (7.76), (7.77)
and (7.78) must all simultaneously be satisfied, and all have a causal
status equivalent to that of Equation (7.79). And the exchange rate deter-
mined in Equation (7.79) will be found to satisfy all the trade equations
in which it appears and also to influence personal consumption through
its effect on capital gains.
Finally, we give the definition of the exchange rate that yields the value
of the euro in dollars:

xr¤ = 1/xr$ (7.80)

We now have a complete 80-equation model (including the redundant

Equation 7.73) of a flexible exchange rate economy, composed of the
US and the Fed on the one hand, and a two-country eurozone with its
single ECB, on the other. The exogenous variables are G, θ and r (for
each country). This represents the main closure of our model. Output in
each country together with consumption, sales, imports, exports, debt
service, government debt, wealth and its allocation between the available
assets, as well as the exchange rate, are all endogenous.
In this model, up to a limit that remains to be determined, countries
are free to pursue fiscal policies of their choice, and central banks are free
to pursue monetary policies of their choice, since fiscal parameters and
interest rates are fully exogenous.
Before we move onto numerical simulations, we should dispose of an
objection which is sometimes brought up against some of the supply and
demand balance conditions which are to be found in the present model.
The objection is that specific rules forbid the ECB from directly financing
Treasuries of euro governments. Article 21.1 of the Statute of the European
System of Central Banks and of the European Central Bank (ECB 2004) points
out that any type of credit facility ‘with the ECB or with the national
central banks in favour of Community institutions or bodies, central
governments … shall be prohibited, as shall the purchase directly from
them by the ECB or national central banks of debt instrument’. Now, this
would seem to be a very peculiar arrangement, which would prevent the
Eurosystem from financing the fiscal deficits of euro governments. But
very similar rules apply in the USA. As noted by a Vice-President of the
 A Simple Model of Three Economies with Two Currencies 173

Federal Reserve Bank of New York, Michael Akhtar (1997, p. 37), ‘the
Federal Reserve is prohibited by law from adding to its net position by
direct purchases of securities from the Treasury that is, the Federal Reserve
has no authority for direct lending to the Treasury. As a consequence,
at most the Desk’s acquisition at Treasury auctions can equal maturing
holdings’. But nobody has ever argued that it was impossible for the
US Treasury to have its central bank finance part of its deficit through
purchases of federal securities on the open market. In other words, the
deficit is indirectly financed by the Fed, just as euro government deficits
can be financed indirectly through the Eurosystem. This is currently most
easily done through the repo market. In other words, save for the absence
of commercial banks, we believe that the structure of our financial system
does not contradict existing European institutions and rules and practise.

The impact of a negative external shock in the case

of the main closure

We start our numerical simulations by dealing with our main closure,

the one where all interest rates are assumed to remain exogenous, and
where pure government expenditures are also exogenous. (To facilitate
reading, from now on the ‘&’ country will be considered to be Italy while
the ‘#’ country will be called Germany). We can simulate a situation
where, starting from a full stationary state, Italy (‘&’) finds that its import
propensity has risen; here, more specifically, the imports from the USA.
This will slow down the Italian economy, as the current account balance
deteriorates and net exports fall. What will happen is that, because of
the resulting budget deficit, by Equation (7.54), the Italian government
unloads bills into the market, which will be absorbed by the ECB. There
will be no change in the Italian interest rate as long as the ECB central
bank is willing to let go German bills while accumulating Italian bills on
its balance sheet, which is what has been assumed in the model.
Figure 7.1 shows the effect on the domestic product of each coun-
try of this increase in the propensity of Italy to import goods from the
USA. All national products reach a new stationary level; Italy, which
now imports more, winds up with a lower GDP; the USA benefit from
a temporary boost in its GDP, owing to its improved exports, but GDP
returns towards its initial level as the exchange rate provides an adjust-
ment mechanism, with a stronger dollar relative to the euro, as shown in
Figure 7.2, bringing back down the overall net exports of the USA. The
other euro country, Germany (‘#’), is the one which has most gained
from the weakening external position of the first euro country, Italy. As
174 Stock–Flow Coherence and Economic Policy

108.0
# gross domestic product

107.0

106.0
$ gross domestic product

105.0

& gross domestic product

104.0

103.0
1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.1 Effect on the domestic product of each country of an increase in the
propensity of the ‘&’ country (Italy) to import products from the ‘$’ country (USA)
(main closure)

1.040

1.030

1.020
Value of dollar in euros (xr$)

1.010

1.000

0.990

1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.2 Depreciation of the euro, measured in dollars, following an increase

in the propensity of one euro, country to import products from the ‘$’ country
(main closure)
 A Simple Model of Three Economies with Two Currencies 175

0.90

0.60 # fiscal balance

0.30 # current account balance

$ current account
balance
0.00 $ fiscal
balance

–0.30 & current account balance

–0.60 & fiscal balance

1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.3 Effect on various balances of an increase in the propensity of the ‘&’
country (Italy) to import products from the ‘$’ country (main closure)

shown in Figure 7.1, Germany gets a permanently higher GDP level, as

a result of the weakened euro, which helps to boost its exports.
Figure 7.3 shows in detail the effects of the initial external shock aris-
ing out of Italy on the various internal and external balances of our three
countries. What is remarkable is that the flexible exchange rate regime,
in contrast to the standard two-country model (Godley and Lavoie 2003),
does not succeed in bringing back internal balance and external balance
into equilibrium. The achieved results are more akin to a fixed exchange
rate model, which in a way is not surprising since the two euro countries
are congealed, between themselves, into a fixed exchange rate situation.
After benefiting from some short-run fiscal surpluses and current account
surpluses, the USA (‘$’) goes back towards zero balances (and hence by
required accounting symmetry, so do the balances of Euroland, taken
overall). But the balances of the individual euro countries do not go back
towards zero values. The current account balance of the ‘weak’ euro coun-
try, Italy, reaches a stationary deficit, while that of the other euro country,
Germany, reaches a stationary surplus (of equal size). What happens to
the internal balances is even more appalling: the fiscal deficit of the
weak country gets ever worse, while the fiscal surpluses of the other euro
country get ever larger.
176 Stock–Flow Coherence and Economic Policy

0.90

0.60 Change in & bills held by ECB

0.30

Change in the monetary base of the ECB

0.00

–0.30 Change in # bills held by ECB

–0.60

1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.4 Evolution of the assets and liabilities of the ECB following an increase
in the propensity of the ‘&’ country (Italy) to import products from the ‘$’ country
(main closure)

The model has reached a quasi-stationary state, where output and

household wealth are constant, but where there is still change in some
flows and stocks.
Under these circumstances, how can interest rates in these two euro
countries remain the same? The answer can be found in Figure 7.4. Since
output in both countries quickly returns to stationary values, the endoge-
nous movements in the monetary base only occur in the short run, as
the monetary base itself (H#+H&) quickly returns to a constant value.
This implies that the asset counterpart to the monetary base must also
remain constant. This will be achieved through two symmetric mecha-
nisms: the ECB will acquire ever-growing amounts of bills issued by the
Italian (‘&’) government, the weak country which faces a fiscal deficit;
and it will sell back ever-growing amounts of bills initially issued by the
other euro government, the German (‘#’) government that runs fiscal
surpluses. As a result, the amount of bills being supplied to households
will correspond to the amounts being demanded, and interest rates can
continue to be pegged.
One may wonder whether such a mechanism can continue for long,
since Italy will be faced with ever-growing ratios of external debt to GDP,
 A Simple Model of Three Economies with Two Currencies 177

as well as ever-growing public debt-to-GDP ratios. First, it should be

noted that nobody is likely to take any notice of the negative current
account balance of Italy, even if information on that aggregate was forth-
coming. So, even though the external debt-to-GDP ratio is rising, Italy
does not and cannot have a balance of payments financing crisis within
the described closure. But what about the rising public debt-to-GDP
ratio?
An additional experiment has been conducted by postulating a grow-
ing economy, driven by exogenously growing pure government expen-
ditures. This economy is then compared with a similar growing one,
where, once again, it is assumed that Italy sees a rise in its propensity to
import US goods. Figure 7.5 compares the two economies. It can be seen
that growth will not solve the problems of the Italian economy. While
the American economy reaches a constant public debt-to-GDP ratio, this
ratio is still ever growing in the case of the Italian economy. This is so
despite assuming an exogenous growth rate that exceeds the interest rate
(3.5% versus 3% in the numerical simulation).

1.200

1.125 & Debt to GDP ratio

1.050

$ Debt to GDP ratio

0.975

# Debt to GDP ratio

0.900

0.825
1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.5 Relative evolution of the debt-to-GDP ratio of each government, fol-
lowing an increase in the propensity of the ‘&’ country (Italy) to import products
from the ‘$’ country, in a world where pure government expenditures grow at an
exogenous rate (main closure)
178 Stock–Flow Coherence and Economic Policy

Once again, notwithstanding the issue of rising public and external

debt-to-GDP ratios, one may question whether it is realistic to assume
an ECB that accumulates a growing proportion of its assets in the form of
bills issued by one country, while the proportion of bills held by the ECB
and issued by the other country would be declining. We would argue that
such a situation is quite possible. The practical aspects of the monetary
policy of the ECB are carried out in a very decentralized manner. While
the ECB decides on the conduct of monetary policy, such as the target
rate of inflation or the target overnight rate, ‘the practical aspects, such
as the execution of open market operations and the administration of
the standing facilities …, are carried out in a decentralised manner by the
national central banks’ (ECB 1999). The components of the balance sheet
of the ECB itself hardly change through time, unless some discretionary
decision is taken. By contrast, the components of the consolidated bal-
ance sheet of the Eurosystem, which is made up from the ECB and the
national central banks, change endogenously. What will happen in the
actual European monetary system is that the Italian national central bank
will reduce its advances to the Italian domestic banking system as its
stock of domestic Treasury bills will rise, while the German national cen-
tral bank will increase its advances to its domestic banking system as its
stock of domestic Treasury bills falls. The re-proportioning of the balance
sheet of what we call the ECB will thus occur in reality through the books
of the Eurosystem.

Alternative closures: equations and experiments

Theory
Still, despite the arguments made above, let us assume that the central
bank refuses to accommodate the modification in the structure of its
assets brought about by the high propensity to import of Italy, refusing
to absorb Italian bills.12 One option is for the rate of interest on Italian
bills to become endogenous and for the rate of interest on those bills,
r&, to rise. But it can be shown that this rise in the Italian rate of interest
makes the Italian budget deficit still worse thereby making necessary an
even larger rise in the Italian rate of interest. The process (which will
be illustrated with a numerical simulation) never comes to an end. The
model needs to be slightly modified in order to represent this. What now
happens is that the amount of Italian bills taken in by the ECB becomes
a constant, thus removing Equation (64) with BECB&d on its left-hand
side. We need to invert that equation so that the supply of Italian bills
 A Simple Model of Three Economies with Two Currencies 179

to the ECB is equal to that constant:

BECB&s = BECB&d = constant (7.64R)

But now Equation (7.61) needs to be rearranged, since otherwise

BECB&s would appear on the left-hand side of two equations. So
Equation (7.61) needs to be rewritten as:

B&&s = B&s − B$&s − B#&s − BECB&s (7.61R)

Because B&&s appears on the left-hand side of Equation (7.72), it must

be inverted as well, leading now to:

B&&d = B&&s (7.72R)

which implies that the Italian rate of interest (r&) is now endogenous and
becomes the left-hand side variable of the portfolio equation defining the
demand for Italian bills by Italian households in Equation (7.46). This is
our second closure.
Within the confines of the model, there is an alternative solution to
the central bank’s refusal to accommodate, which keeps interest rates
exogenous – namely to endogenise the Italian fiscal policy. Rearranging
variables in a manner similar to what was done above, the new equations
of this third model will now be:

BECB&s = BECB&d = constant (7.64G)

B&s = B&&s + B$&s + B#&s + BECB&s (7.61G)
G& = B&s − r&−1 · B&s−1 + T & + F& (7.54G)

Under these conditions, with the ECB refusing to accommodate

(Equation (7.64G)), fiscal policy in Italy must become endogenous, as
shown by Equation (7.54G). This is our third closure. The fiscal stance
of the Italian government depends on how many extra bills can be
unloaded on financial markets (as determined by Equation 7.61G). The
Italian government is financially constrained. It is as if some kind of loan-
able funds constraint existed. The intuition is that fiscal policy always
adjusts so that the total supply of Italian assets is such that the market
absorbs them willingly at given interest rates. Alternatively, this third
closure can be understood as a situation where the rising debt of the
deficit country is judged to be unsustainable by ‘the markets’, and that,
as a result of this, fiscal policy of the deficit country and the amount of
new government debt issued is ‘constrained’ by the financial markets.
180 Stock–Flow Coherence and Economic Policy

Practice
Let us now run simulations based on the two alternative closures
identified above. Given an increase in the propensity to import of
Italy, assume first that the ECB refuses to purchase additional amounts
of Italian bills, so that BECB& is assumed to be a constant, as in
Equation (7.64R). The new endogenous variable is the rate of interest
on Italian bills, which now adjusts to clear the market for Italian bills (as
in Equation 7.72R).
As the Italian government runs into its deficit and the amount of Ital-
ian bills being supplied to the markets rises, while demand does not
change, the rate of interest needs to rise. But this rise, as can be seen
from Figure 7.6, must continue in every period. As a result, as shown in
Figure 7.7, the current account balance of Italy does not reach a station-
ary negative level, as it did in the main closure; the current account
balance worsens ever more quickly, leading to an unsustainable sit-
uation. Indeed, the model explodes and diverges from a long-period
equilibrium. Freeing interest rates, as in all open-economy models of
this kind, leads to instability (see Izurieta 2003).

0.075

Interest rate on & bills

0.060

0.045

0.030
Interest rate on # bills

0.015

0.000
1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.6 Effect of an increase in the propensity of the ‘&’ country (Italy) to
import products from the ‘$’ country, when the ‘&’ interest rate is left to be
endogenous
 A Simple Model of Three Economies with Two Currencies 181

2.0
# current account balance

1.0

$ current account
0.0 balance

–1.0
& current account
balance
–2.0

–3.0

1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975

Figure 7.7 Effect on the current account balances arising from an increase in the
propensity of the ‘&’ country (Italy) to import products from the ‘$’ country, when
the ‘&’ interest rate is endogenous

There is thus a need for another closure. In this third closure, the new
endogenous variable is the amount of pure expenditures by the Italian
government. Its fiscal policy is adjusted in such a way that the Italian
government is running a deficit only if the financial markets are will-
ing to take in more Italian bills at the given interest rates, as shown in
Equation (7.54G). In other words, as long as there is no change in the
demand for Italian bills by the private sector, knowing that the ECB keeps
constant its own stock of Italian bills, the Italian government must run
a balanced budget.
Such a fiscal policy certainly has negative effects on the Italian econ-
omy: owing to the negative shock on its net exports, the Italian
government is forced to reduce its expenditures, leading to a substan-
tial reduction in the ‘&’ (Italy) output, as shown in Figure 7.8. However,
the restrictive fiscal policy is highly efficient in another sense, since it
brings back the economy towards a super-stationary state, where all stocks
remain constant. This can be seen from Figure 7.9, which represents the
current account balances of our three countries. Cuts in Italian govern-
ment expenditures bring back the current account balance of the ‘weak’
euro country towards equilibrium. These fiscal cutbacks in one country
182 Stock–Flow Coherence and Economic Policy

# gross domestic product

106.0
$ gross domestic product

104.0

102.0

100.0 & gross domestic product

98.0

96.0
1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.8 Effect on the domestic product of each country of an increase in the
propensity of the ‘&’ country (Italy) to import products from the ‘$’ country, when
government expenditures of the ‘&’ country are assumed to be endogenous

0.40 $ current account balance

0.20
# current account balance

0.00

& current account balance

–0.20

–0.40

–0.60
1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.9 Effect on current account balances of an increase in the propensity of

the ‘&’ country (Italy) to import products from the ‘$’ country, when government
expenditures of the ‘&’ country are assumed to be endogenous
 A Simple Model of Three Economies with Two Currencies 183

have feedback effects on the current account balance of the other euro
country; the current account surplus of Germany is gradually brought
down to zero, as is the current account balance of the rest of the world,
represented by the ‘$’ country.
We may again experiment with the case of a growing economy by
assuming that the pure government expenditures of Germany and the
USA are exogenously growing at a given rate. Starting from a full steady
state, where all variables of the model are growing at a steady rate while
the current accounts and the government budgets are balanced, we can
compare this situation with one where the propensity of Italy to import
US goods is increased permanently. What will be the impact on the
public debt-to-GDP ratio of each country? Figure 7.10 shows that, in
contrast to the main closure, the third closure with endogenous Ital-
ian fiscal policy brings about stabilized public debt-to-GDP ratios for all
countries. In particular, while the negative external shock to the Italian
economy does lead to higher public debt-to-GDP ratios, the ratio does
converge to a constant value, showing that the new situation is clearly
sustainable.

1.050
& Debt to GDP ratio

1.035

1.020

1.005

0.990 $ Debt to GDP ratio

0.975 # Debt to GDP ratio

1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986

Figure 7.10 Relative evolution of the debt-to-GDP ratio of each government, fol-
lowing an increase in the propensity of the ‘&’ country (Italy) to import products
from the ‘$’ country, in a world with growth, where expenditures of the ‘&’
government are endogenous
184 Stock–Flow Coherence and Economic Policy

Conclusion

We have presented a model with nearly 80 equations, which we believe

can adequately represent the behaviour of a whole world comprising two
regions, linked by a flexible exchange rate. One of the regions operates
under a single central bank, with a common currency for two countries.
The latter can be taken to represent two eurozone countries, with the
ECB (or the Eurosystem). The other region, or the rest of the world, can
be assumed to represent the USA, with the Federal Reserve Bank.
Our model shows that the dynamics of a three-country and two-
currency model are somewhat peculiar compared with a more standard
model with only two countries. In a flexible exchange rate regime with
only two countries, the flexibility of the exchange rate, given the proper
parameter values, would allow any country to recover from a negative
external shock (Godley and Lavoie 2003). After some periods of cur-
rent account deficits and budget deficits, a weakened country would
eventually get back on track with balanced budgets and balanced exter-
nal current accounts. However, in the case of the eurozone and its
common currency, while the eurozone taken as a whole also recov-
ers from an external negative shock affecting one country (the overall
Euroland current account balance and Euroland fiscal balance converge
towards zero as the overall world economy approaches a stationary state),
each country from the currency union taken individually pursues a
diverging path.
The peculiarity of the common currency case is particularly clear when
one considers the path taken by the current account balance and the
budget balance of the ‘weakened’ euro country. Its current account bal-
ance tends towards a constant deficit value, while its budget deficit keeps
growing through time, even though the world taken as a whole con-
verges towards a (quasi-)stationary state. Hence, in this case, the ‘twin
deficit’ or ‘twin surplus’ property of quasi-stationary states does not
exactly hold: the two balances are different in size, although they are
of the same sign. Despite the fact that income and wealth are no longer
changing, which is a defining feature of the quasi-stationary state, the
current account deficit and the budget deficit of the weakened country
are not equal to each other. This is compensated by the fact that a sym-
metric phenomenon is affecting the other euro country, which in the
quasi-stationary state will be experiencing unequal budget surpluses and
current account surpluses, as a result of its boosted net exports due to the
weakened euro caused by the original negative shock on the weakened
euro country.
 A Simple Model of Three Economies with Two Currencies 185

This situation can go on as long as the ECB agrees to modify the

proportion of its assets, accumulating an ever-growing proportions of
bills issued by the weakened country (and as long as financial markets
accept that the weakened country experiences a rising public debt-to-
GDP ratio and a rising external debt-to-GDP ratio, if measures of the
latter are provided and analysed). If the ECB were to decline to do so,
either the interest rate on bills issued by the weakened country or the fis-
cal policy of the weakened country would have to become endogenous.
The first alternative, our second closure, produces rising interest rates
and leads to exploding dynamics. This is clearly not the way to go. The
second alternative, our third closure, whereby the weakened economy
reduces its government expenditures to eliminate its budget deficit, leads
to lower income and output levels (as McCombie and Thirlwall (1994)
would have it, when interpreted in growth terms), but also to a new
super-stationary state, with wiped-out current account deficits. A fourth
closure with exogenous interest rates, not discussed here but developed
in Lavoie (2003, pp. 120–121), shows that any downward adjustment
mechanism in pure government expenditures when faced with a fiscal
deficit would pull the trick. This is perhaps the justification that partisans
of the Maastricht rules putting limits on fiscal deficits would be looking
for. If the ECB is forbidden from accommodating market-driven changes
in the composition of its assets, or if the ECB rules that it will not accu-
mulate additional stocks of securities issued by governments that have
excessively large debts according to rating agencies, then fiscal policy in
the ‘weak’ countries must be endogenous for stability to prevail, for oth-
erwise it would seem that the only alternative is to let interest rates on
euro bills to diverge from country to country in an unsustainable way.
Now this would seem to be a rather dismal state of affairs, from a
progressive standpoint. However, it should be noted that balanced fis-
cal and external positions for all could as well be reached if the euro
country benefiting from a (quasi) twin surplus as a result of the neg-
ative external shock on the other euro country decided to increase its
government expenditures, in an effort to get rid of its budget surplus.
This case, where the surplus countries rather than the deficit countries
adjust, as many authors have underlined, would eliminate the current
downward bias in worldwide economic activity. Now this would require
an entirely new attitude towards government deficits. One would need
an anti-Maastricht approach, that would run against the Stability and
Growth Pact and its neoliberal obsession with fiscal balance and gov-
ernment debt reduction.13 For instance, one would need a new Pact
that would discourage fiscal surpluses.14 National governments that ran
186 Stock–Flow Coherence and Economic Policy

budget surpluses would pay large proportional automatic levies to the

European Union, who would be compelled to spend the sums thus col-
lected in the deficit countries. In this manner, the ‘weak’ and the ‘strong’
members of the eurozone could converge towards a super-stationary
state, with balanced budgets and current accounts, through an increase
rather than a decrease in government expenditures and economic activity.
Alternatively, the present structure of the European Union would
need to be modified, giving far more spending and taxing power to
the European Union Parliament, transforming it into a bona fide federal
government that would be able to engage into substantial equalization
payments which would automatically transfer fiscal resources from the
more successful to the less successful members of the euro zone. In
this manner, the eurozone would be provided with a mechanism that
would reduce the present bias towards downward fiscal adjustments of
the deficit countries. This raises the profound question as to whether in
the long term it is possible to have a community of nations which have
a single currency which does not have a federal budget of substantial
size, and by implication a federal government to run it – a point that
was made very early on in Godley (1992).

Notes
1. Preliminary versions of this paper were presented on two different occasions,
once by each author, at workshops organized in 2004 by Professor Jacques
Mazier, from the University of Paris 13 Villetaneuse. We are very grateful to
Professor Mazier for having given us the opportunity to present our work, and
for all the questions and comments that were made by the various workshop
participants. We are also thankful to Alex Izurieta for comments made on an
earlier draft. All simulations were conducted and all figures were constructed
with the MODLER software.
2. To save space, notation dealing with lags has been omitted from the matrix.
3. It is assumed that the price of goods and services does not change. This will
mean that there is a formal but not lethal inconsistency in our model, because
we implicitly assume that exchange rates influence export and import prices.
4. There are no capital gains arising from changes in the prices of securities,
since we have assumed that households hold no (long-term) bonds, only
(short-term) bills, the price of which is assumed to remain constant within
the period. See Godley (1999a), Lavoie (2003) and Lequain (2003) for open-
economy models with bonds.
5. The assumption is only there to reduce the number of equations. It is possi-
ble to introduce any other assumption about capital gains taxation without
changing the main results of the model.
6. Imports could depend on GDP (as in Godley and Lavoie (2003), dispos-
able income (as in Lavoie (2003)), or any other such aggregate. It makes no
difference to the dynamics.
 A Simple Model of Three Economies with Two Currencies 187

7. It is of course a drastic simplification to write such an equation. A full alterna-

tive would require a relationship between the exchange rate and import prices
and add a relationship describing the price elasticity of demand for import
volumes. This is done in Godley and Lavoie (2003). The present formulation
is a kind of reduced form, the purpose of which is to cut down on the number
of equations required by a three-country model.
8. Since there are no commercial banks, H could represent deposits of house-
holds at the central bank.
9. A term involving a transactions demand for money, Y/V or YP/V, has been
omitted for simplification.
10. Since all central bank profits are returned to government, this explains that
the net worth of the Fed is zero, as shown in Table 7.1. Also, the Fed holds
no foreign assets that could generate capital gains or losses, and we assumed
away, for simplicity, a starting amount of own funds.
11. The reader is asked to sympathize with us here because of the length of time
(and the degree of sustained concentration) it took to find an arrangement
which satisfied the logic of the model and to which the computer would yield
a solution.
12. Assets eligible for collateral at the national central banks are determined by
the ECB. These rules specify that eligible collateral does not include assets
rated below A– by rating agencies. It has been argued by Atkins and Schieritz
(2005) that this implies that the securities of some European governments
(small governments with large debts and deficits) could become ineligible.
This possibility would thus correspond to the described closure whereby the
ECB refuses to accumulate additional Italian bills.
13. Along the lines of some of the suggestions made by the working
group ‘Alternative Economic Policy for Europe (Euromemorandum-Group)’,
as can be found on http://www.memo-europe.uni-bremen.de/euromemo/
indexmem.htm
14. There is some similarity here with Davidson’s (1982, p. 223) proposal that
rules should be designed ‘to provide automatic mechanism for placing a major
burden of trade imbalance adjustments on the surplus nations’.

References

Akhtar, M.A. (1997). Understanding Open Market Operations (New York: Federal
Reserve Bank of New York, Public Information Department).
Atkins, R. and M. Schieritz (2005) ‘ECB Targets Its Problem Nations.’ Financial
Times, 8 November.
Davidson, P. (1982) International Money and the Real World (London: Macmillan).
Dos Santos, C. (2005) ‘A Stock–Flow Consistent General Framework for Formal
Minskyan Analyses of Closed Economies.’ Journal of Post Keynesian Economics
27 (4): 711–736.
ECB (1999) ‘The Operational Framework of the Eurosystem: Description and First
Assessment.’ ECB Monthly Bulletin May.
ECB (2004) Protocol on the Statute of the European System of Central Banks and
of the European Central Bank (unofficial consolidated version, 1 June 2004),
http://www.ecb.int/ecb/legal/pdf/ en_statute.pdf
188 Stock–Flow Coherence and Economic Policy

Godley, W. (1992) ‘Maastricht and All That.’ London Review of Books 14 (19), 8
October.
Godley, W. (1999a) ‘Open Economy Macroeconomics Using Models of Closed
Systems.’ Working Paper 281, The Jerome Levy Economics Institute.
Godley, W. (1999b) ‘Money and Credit in a Keynesian Model of Income Determi-
nation.’ Cambridge Journal of Economics 23 (4): 393–411.
Godley, W. and F. Cripps (1983) Macroeconomics (London: Fontana).
Godley, W. and M. Lavoie (2003) ‘Two-Country Stock–Flow Consistent Macroeco-
nomics Using a Closed Model within a Dollar Exchange Regime.’ CERF Working
Paper 10, University of Cambridge.
Godley, W. and M. Lavoie (2005–2006) ‘Comprehensive Accounting in Sim-
ple Open Economy Macroeconomics: Endogenous Sterilization Or Flexible
Exchange Rates.’ Journal of Post Keynesian Economics 28 (2): 241–276.
Izurieta, A. (2001) ‘Can Countries Under a Common Currency Conduct Their
Own Fiscal Policies?’ Working Paper 337, The Jerome Levy Institute of Bard
College.
Izurieta, A. (2003) ‘Dollarization as a Tight Rein on the Fiscal Stance.’ In Rochon,
L.P. and Seccareccia, M. (eds), Dollarization: Lessons from Europe and the Americas
(London: Routledge), pp. 143–164.
Lavoie, M. (2003) ‘A Fully Coherent Post Keynesian Model of the Euro Zone.’ In
Arestis, P., Baddeley, M. and McCombie, J. (eds), Globalisation, Regionalism and
Economic Activity (Cheltenham: Edward Elgar), pp. 98–126.
Lavoie, M. (2006) ‘A Fully Coherent Post-Keynesian Model of Currency Boards.’ In
Gnos, C. and Rochon, L.P. (eds), Economic Policies: Perspectives from the Keynesian
Heterodoxy (Cheltenham: Edward Elgar).
Lequain, M. (2003). ‘A Three-Country Study of the Euro Zone versus the Rest of
the World: the Implications of a Monetary Union in An Open Environment.’
Paper presented at the Annual Conference of the Eastern Economic Association,
New York, February.
McCombie, J.S.L. and A.P. Thirlwall (1994). Economic Growth and the Balance-of-
Payments Constraint (New York: St Martin’s Press).
Mouakil, T. (2005) ‘Les dysfonctionnements du système moneétaire interna-
tional: une modélisation post-keynésienne.’ Paper presented at the ‘Économie
politique de la gouvernance’ Conference, Dijon, December.
Taylor, L. (2004) ‘Exchange Rate Interderminacy in Portfolio Balance, Mundell–
Fleming and Uncovered Interest Rate Parity Models.’ Cambridge Journal of
Economics 28 (2): 205–227.
Tobin, J. (1982) ‘Money and Finance in the Macroeconomic Process.’ Journal of
Money, Credit, and Banking 14 (2): 171–204.
8
Maastricht and All That
Wynne Godley

A lot of people throughout Europe have suddenly realized that they know
hardly anything about the Maastricht Treaty while rightly sensing that
it could make a huge difference to their lives. Their legitimate anxiety
has provoked Jacques Delors to make a statement to the effect that the
views of ordinary people should in future be more sensitively consulted.
He might have thought of that before.
Although I support the move towards political integration in Europe, I
think that the Maastricht proposals as they stand are seriously defective,
and also that public discussion of them has been curiously impoverished.
With a Danish rejection, a near-miss in France, and the very existence
of the ERM in question after the depredations by currency markets, it is
a good moment to take stock.
The central idea of the Maastricht Treaty is that the EC countries should
move towards an economic and monetary union, with a single currency
managed by an independent central bank. But how is the rest of eco-
nomic policy to be run? As the treaty proposes no new institutions
other than a European bank, its sponsors must suppose that nothing
more is needed. But this could only be correct if modern economies
were self-adjusting systems that didn’t need any management at all.
I am driven to the conclusion that such a view – that economies are
self-righting organisms which never under any circumstances need man-
agement at all – did indeed determine the way in which the Maastricht
Treaty was framed. It is a crude and extreme version of the view which for
some time now has constituted Europe’s conventional wisdom (though
not that of the US or Japan) that governments are unable, and there-
fore should not try, to achieve any of the traditional goals of economic
policy, such as growth and full employment. All that can legitimately
be done, according to this view, is to control the money supply and

189
190 Stock–Flow Coherence and Economic Policy

balance the budget. It took a group largely composed of bankers (the

Delors Committee) to reach the conclusion that an independent cen-
tral bank was the only supra-national institution necessary to run an
integrated, supra-national Europe.
But there is much more to it all. It needs to be emphasized at the start
that the establishment of a single currency in the EC would indeed bring
to an end the sovereignty of its component nations and their power to
take independent action on major issues. As Mr Tim Congdon has argued
very cogently, the power to issue its own money, to make drafts on its
own central bank, is the main thing which defines national indepen-
dence. If a country gives up or loses this power, it acquires the status of a
local authority or colony. Local authorities and regions obviously cannot
devalue. But they also lose the power to finance deficits through money
creation while other methods of raising finance are subject to central
regulation. Nor can they change interest rates. As local authorities pos-
sess none of the instruments of macroeconomic policy, their political
choice is confined to relatively minor matters of emphasis – a bit more
education here, a bit less infrastructure there. I think that when Jacques
Delors lays new emphasis on the principle of ‘subsidiarity’, he is really
only telling us we will be allowed to make decisions about a larger num-
ber of relatively unimportant matters than we might previously have
supposed. Perhaps he will let us have curly cucumbers after all. Big deal!
Let me express a different view. I think that the central government
of any sovereign state ought to be striving all the time to determine the
optimum overall level of public provision, the correct overall burden of
taxation, the correct allocation of total expenditures between compet-
ing requirements and the just distribution of the tax burden. It must
also determine the extent to which any gap between expenditure and
taxation is financed by making a draft on the central bank and how
much it is financed by borrowing and on what terms. The way in which
governments decide all these (and some other) issues, and the qual-
ity of leadership which they can deploy, will, in interaction with the
decisions of individuals, corporations and foreigners, determine such
things as interest rates, the exchange rate, the inflation rate, the growth
rate and the unemployment rate. It will also profoundly influence the
distribution of income and wealth not only between individuals but
between whole regions, assisting, one hopes, those adversely affected
by structural change.
Almost nothing simple can be said about the use of these instruments,
with all their inter-dependencies, to promote the well-being of a nation
and protect it as well as may be from the shocks of various kinds to which
 Maastricht and All That 191

it will inevitably be subjected. It only has limited meaning, for instance,

to say that budgets should always be balanced when a balanced budget
with expenditure and taxation both running at 40% of GDP would have
an entirely different (and much more expansionary) impact than a bal-
anced budget at 10%. To imagine the complexity and importance of a
government’s macroeconomic decisions, one has only to ask what would
be the appropriate response, in terms of fiscal, monetary and exchange
rate policies, for a country about to produce large quantities of oil, of
a fourfold increase in the price of oil. Would it have been right to do
nothing at all? And it should never be forgotten that in periods of very
great crisis, it may even be appropriate for a central government to sin
against the Holy Ghost of all central banks and invoke the ‘inflation tax’ –
deliberately appropriating resources by reducing, through inflation, the
real value of a nation’s paper wealth. It was, after all, by means of the
inflation tax that Keynes proposed that we should pay for the war.
I recite all this to suggest, not that sovereignty should not be given up
in the noble cause of European integration, but that if all these functions
are renounced by individual governments they simply have to be taken
on by some other authority. The incredible lacuna in the Maastricht pro-
gramme is that, while it contains a blueprint for the establishment and
modus operandi of an independent central bank, there is no blueprint
whatever of the analogue, in Community terms, of a central government.
Yet there would simply have to be a system of institutions which fulfils
all those functions at a Community level which are at present exercised
by the central governments of individual member countries.
The counterpart of giving up sovereignty should be that the compo-
nent nations are constituted into a federation to whom their sovereignty
is entrusted. And the federal system, or government, as it had better
be called, would have to exercise all those functions in relation to its
members and to the outside world which I have briefly outlined above.
Consider two important examples of what a federal government, in
charge of a federal budget, should be doing.
European countries are at present locked into a severe recession. As
things stand, particularly as the economies of the USA and Japan are
also faltering, it is very unclear when any significant recovery will take
place. The political implications of this are becoming frightening. Yet the
interdependence of the European economies is already so great that no
individual country, with the theoretical exception of Germany, feels able
to pursue expansionary policies on its own, because any country that did
try to expand on its own would soon encounter a balance of payments
constraint. The present situation is screaming aloud for co-ordinated
192 Stock–Flow Coherence and Economic Policy

reflation, but there exist neither the institutions nor an agreed frame-
work of thought which will bring about this obviously desirable result.
It should be frankly recognized that if the depression really were to take
a serious turn for the worse – for instance, if the unemployment rate
went back permanently to the 20–25% characteristic of the Thirties –
individual countries would sooner or later exercise their sovereign right
to declare the entire movement towards integration a disaster and resort
to exchange controls and protection – a siege economy if you will. This
would amount to a re-run of the inter-war period.
If there were an economic and monetary union, in which the power to
act independently had actually been abolished, ‘co-ordinated’ reflation
of the kind which is so urgently needed now could only be undertaken
by a federal European government. Without such an institution, EMU
would prevent effective action by individual countries and put nothing
in its place.
Another important role which any central government must perform is
to put a safety net under the livelihood of component regions which are
in distress for structural reasons – because of the decline of some indus-
try, say, or because of some economically adverse demographic change.
At present this happens in the natural course of events, without any-
one really noticing, because common standards of public provision (for
instance, health, education, pensions and rates of unemployment ben-
efit) and a common (it is to be hoped, progressive) burden of taxation
are both generally instituted throughout individual realms. As a con-
sequence, if one region suffers an unusual degree of structural decline,
the fiscal system automatically generates net transfers in favour of it. In
extremis, a region which could produce nothing at all would not starve
because it would be in receipt of pensions, unemployment benefit and
the incomes of public servants.
What happens if a whole country – a potential ‘region’ in a fully
integrated community – suffers a structural setback? So long as it is a
sovereign state, it can devalue its currency. It can then trade successfully
at full employment provided its people accept the necessary cut in their
real incomes. With an economic and monetary union, this recourse is
obviously barred, and its prospect is grave indeed unless federal bud-
geting arrangements are made which fulfil a redistributive role. As was
clearly recognized in the MacDougall Report which was published in
1977, there has to be a quid pro quo for giving up the devaluation option
in the form of fiscal redistribution. Some writers (such as Samuel Brittan
and Sir Douglas Hague) have seriously suggested that EMU, by abolish-
ing the balance of payments problem in its present form, would indeed
 Maastricht and All That 193

abolish the problem, where it exists, of persistent failure to compete

successfully in world markets. But as Professor Martin Feldstein pointed
out in a major article in the Economist (13 June), this argument is very
dangerously mistaken. If a country or region has no power to devalue,
and if it is not the beneficiary of a system of fiscal equalization, then
there is nothing to stop it suffering a process of cumulative and termi-
nal decline leading, in the end, to emigration as the only alternative to
poverty or starvation. I sympathize with the position of those (like Mar-
garet Thatcher) who, faced with the loss of sovereignty, wish to get off the
EMU train altogether. I also sympathize with those who seek integration
under the jurisdiction of some kind of federal constitution with a federal
budget very much larger than that of the Community budget. What I
find totally baffling is the position of those who are aiming for economic
and monetary union without the creation of new political institutions
(apart from a new central bank), and who raise their hands in horror at
the words ‘federal’ or ‘federalism’. This is the position currently adopted
by the Government and by most of those who take part in the public
discussion.
9
Fiscal Policy in a Stock–Flow
Consistent (SFC) Model
Wynne Godley and Marc Lavoie

In our book Monetary Economics (Godley and Lavoie 2007, chapter 11),
we claimed that a particular level of government expenditure relative to
tax rates, and also relative to gross domestic product (GDP), is essential if
stable, noninflationary growth and full employment are to be achieved.
We argued, on the basis of simulation models, that monetary policy on
its own was unable to maintain full employment and low inflation for
more than a short period of time, unless fiscal policy was appropriate.
Our conclusions conflict with those of the ‘new consensus,’ which holds
that a correct setting of interest rates is the necessary and sufficient condi-
tion for achieving noninflationary growth at full employment, leaving
fiscal policy rather in the air. This has led different countries to adopt
different targets for the nominal budget deficit and government debt as
proportions of (nominal) GDP measured ex post.1 But the rationale for
such targets has never been clear (at least to us).
In this paper, we deploy a simple stock–flow consistent (SFC) model
that will enable us to outline the way in which the fiscal stance (as
defined below) should be determined as the necessary, though not always
sufficient, condition for the achievement of the major objectives of
macroeconomic policy. We also show that the new emphasis on mon-
etary policy may be misplaced. In theory, although in practice this
may be an entirely different issue, fiscal policy can achieve everything
the central banks claim they are able to do through monetary policy.
In other words, just as the success of monetary policy is judged on
the basis of medium-term achievements, and not on the monthly or
quarterly variations of the inflation rate, there is a similar role to be
played by fiscal policy on the medium-term evolution of output and
employment.

194
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 195

A simple SFC growth economy model

An outline of the SFC model

The matrix shown in Table 9.1 describes the accounting structure of the
basic model we use. All variables in this matrix are measured at current
prices. The counterpart real variables will be defined in the text that fol-
lows. As always in a transactions-flow matrix, each row and each column
must sum to zero.
All variables are defined in the matrix apart from r (the nominal
interest rate), V (private wealth) and GD (government debt). For sim-
plification, the accumulation of capital by firms has been assumed
away.
In what follows, the numbered equations correspond with those
directly entering the model (i.e., those required by the computer to
obtain a solution). Equations introduced using capital letters (A, B etc.)
are auxiliaries that hopefully aid the exposition. Although the model is
very simple, its exposition is slightly intricate because decisions by the
private sector are assumed to be taken entirely in real terms, whereas
those of the government regarding interest rates and tax rates together
with targets for budget balances are measured in nominal terms.
We assume that the economy we describe is closed, including a
government and an aggregated private sector:

y ≡ g + px, (A)

where y is real GDP, px is real private expenditure, and g is real pure

government expenditure, meaning here that g does not include debt
servicing. Lowercase letters are used throughout to describe real variables
and uppercase to describe nominal variables.

Table 9.1 Transactions-flow matrix of a simple closed economy model

Households Firms Government Sum

Private expenditures −PX +PX 0

Government expenditures +G −G 0
Income (GDP) +Y −Y 0
Taxes −T +T 0
Interest +r · GD−1 −r · GD−1 0
Change in wealth/debt −V +GD 0
Sum 0 0 0 0
196 Stock–Flow Coherence and Economic Policy

Real (inflation accounted) disposable income is given by:

yd ≡ y + rr · v−1 − t, (9.1)

where yd is real disposable income, rr is the real rate of interest, v is the

accumulated stock of real financial wealth owned by the private sector,
and t is the deflated flow of tax payments.2
It is assumed that real private expenditure is functionally related to
real disposable income, the inherited stock of financial wealth, and the
real interest rate:

px = α1 · yd + α2 · v−1 , (B)

where spending out of income is negatively related to the real interest

rate:

α1 = α10 − ι · rr−1 . (9.2)

It is recognized that this is an impoverished representation of the way

in which monetary policy works. In the real world, monetary policy
temporarily affects demand, in addition, via its effect on the value of
assets and also on the exchange rate.
As the change in the real stock of wealth is equal by definition
to real disposable income less expenditure – that is, in line with the
Haig–Simons definition of real disposable income:

v ≡ yd − px [≡ real private saving], (C)

Equation (B) can equivalently be written as a wealth adjustment

function:

v = v−1 + α2 · (v ∗ − v−1 ). (9.3)

This implies that the desired real stock of financial wealth, ν*, is a
determinate proportion of disposable income:

v ∗ = α3 · yd, (9.4)

where,

α3 = (1 − α1 )/α2 . (9.5)

As we are going to make suggestions about policy in the real world, it

is important to note here that the coefficient α3 is intended to refer to a
long-run tendency. In the short run, the ratio of desired financial wealth
to disposable income will fluctuate, for instance, because of capital gains
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 197

and losses and also credit cycles. It is precisely from such (normally)
short-term influences that we wish to abstract, because there will only
be rare occasions on which it will be appropriate to use fiscal policy to
offset them.
It follows that private expenditure enters the equation system in the
following form:

px ≡ yd − v, (9.6)

because yd and ν are already determined in Equations (9.1) and (9.3).

Nominal taxes, T , are raised as a proportion, θ, of nominal private
factor income, Y, plus nominal interest receipts:

T = θ · (Y + r · V−1 ) (9.7)

where Y is nominal GDP, V is the nominal stock of financial wealth, and

r is the nominal interest rate:

Y ≡ y·p (9.8)

and

V ≡ v · p, (9.9)

where p is the price level.

Nominal and real interest rates are related according to the Fisher
formula:

rr ≡ (1 + r)/(1 + π) − 1, (9.10)

where π is defined as the rate of price inflation, which is a given in our

little model:

π ≡ p/p−1 . (9.11)

The economy is assumed to grow at a rate, gr, and to be at a level that

corresponds with full employment as well as low and stable inflation.
In the wording of mainstream economics, the output gap is zero at all
times and the economy is at the NAIRU. We do not actually believe that
such conditions usually occur, or that the NAIRU is a useful concept, but
we set out these conditions for the sake of discussion. Another way to
understand Equation (9.12) is to say that, although the economy may not
be performing at full employment at all times, we are trying to ascertain,
as will be clear later, the fiscal stance that needs to be adopted if the
economy is to be at full employment on average.

y = y−1 · (1 + gr). (9.12)

198 Stock–Flow Coherence and Economic Policy

The real tax yield is:

t ≡ T /p. (9.13)

Total real government outlays, gT , are given by:

gT ≡ g + rr · gd−1 , (9.14)

where gT is real government expenditure gross of real interest payments,

and gd is the real stock of government debt. The government’s real, infla-
tion accounted, deficit is therefore equal to the change in the real stock
of debt:

gd ≡ gT − t [≡ the real deficit]. (9.15)

We can now derive the remaining government variables at current

prices. Total government outlays, GT , are given by:

GT = G + r · GD−1 , (9.16)

where G is nominal government expenditure on goods and services, and

GD is nominal government debt.

G ≡ g · p. (9.17)

The nominal budget deficit, DEF, is:

DEF ≡ GT − T , (9.18)

and the nominal stock of government debt is:

GD = GD−1 + DEF, (9.19)

To complete the model, we now only have to invert Equation (A), thereby
making the real flow of government expenditure on goods and services
endogenous.

g ≡ y − px. (9.20)

In other words, we assume that, for a given tax rate, pure government
expenditures take up any slack that could exist between potential (or
full-employment) output and private expenditures. We have recently
become aware that a paper by Schlicht (2006) shows a remarkable degree
of affinity with the present work, both in its modelling strategy and in
its conclusions.
Our model is now complete in the sense that it can be solved for
the level and growth of government expenditure and the budget deficit
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 199

conditional on any configuration of assumptions regarding r and θ – the

policy variables – as well as gr, α10 , α2 , ι and π .
Note finally that nominal private saving, or the net accumulation of
financial assets, is given by:

NAFA ≡ (Y + r · V−1 − T ) − X. (9.21)

This identity will provide a useful check that the accounting of the
model is correct because nominal private saving should be found to
be equal to the (nominal) budget deficit (DEF) although there is no
(individual) equation to make this happen.3

Some arithmetical results

In this section, we confine ourselves to solutions that describe grow-
ing steady states, in which all real stocks and flows are growing at the
same rate while all nominal stocks and flows are growing at a different,
higher rate. We first set forth a base run in which real output and all
other real flows and stocks grow at 2.5% per annum, thus assuming that
this is known to be the rate at which the productive potential of the
economy is growing. In addition, we make arbitrary but uncontrover-
sial assumptions about the tax rate (25%), the inflation rate (2%), the
nominal rate of interest (3%) and all the parameters that control pri-
vate expenditures relative to wealth. We can then infer levels for various
key ratios of the economy, as identified in Table 9.2: real interest rate;

Table 9.2 Steady-state values of variables for some parameter values

gr = 2.5%; θ = 25%; α2 = 0.2; α10 = 0.9; ι = 0.2

π = 2%|r = 3% π = 6%|r = 7% π = 2%|r = 10%

rr 0.98% 0.94% 7.84%

(1 − θ ) · rr 0.73% 0.71% 5.89%
α1 0.88 0.88 0.74
g/y = G/Y 25.9% 26.3% 22.4%
gT /y 26.3% 26.6% 29.5%
t/y = T /Y 25.3% 25.7% 27.2%
gd/y 1.0% 1.0% 2.3%
(GT /Y) 27.1% 28.9% 31.3%
DEF/Y 1.8% 2.2% 4.1%
r · GD−1 /Y 1.2% 2.6% 8.9%
Primary surplus/Y −0.6% 0.4% 4.8%
gd/y = GD/Y 40.9% 40.5% 93.5%
200 Stock–Flow Coherence and Economic Policy

after-tax real interest rate; propensity to spend out of disposable income;

the ratio of pure government expenditures to GDP; the ratio of real total
government expenditure (including debt servicing) to real GDP; the ratio
of tax to GDP; the ratio of real government deficit (or change in real gov-
ernment debt) to real GDP (which is the difference between the former
two ratios); the ratio of nominal total government expenditure to GDP;
the ratio of nominal government deficit to GDP; the ratio of nominal
debt service to GDP; the ratio of primary surplus to GDP both in nomi-
nal terms and, finally, the ratio of government debt to GDP, which, given
our starting hypotheses, is also the ratio of private wealth to GDP.
Solutions to the baseline model, given the assumptions about exoge-
nous variables, are shown in the second column of Table 9.2. Under the
circumstances, to sustain full employment and a zero output gap, pure
government expenditures as a ratio of GDP must reach 25.9%. Govern-
ments must run deficits: the nominal deficit as a ratio of nominal GDP
(DEF/Y) must be 1.8%, the primary deficit as a ratio of nominal GDP must
be 0.6%, and the real deficit as a ratio of real GDP has to be 1.0% (gd/y).
These are all endogenously generated numbers that follow ineluctably
from the assumptions we have made.
These results are not enormously altered if the assumptions about
exogenous variables are changed, unless the changes are very large. For
instance, if we assume an inflation rate of 6%, with a consequential
increase of 4 percentage points in the nominal interest rate, as r moves
up from 3 to 7% – thus keeping the real interest rate approximately con-
stant – the ratio of pure government expenditures to GDP barely moves,
going from 25.9 to 26.3%. The ratio of real deficit to real GDP does not
change, whereas the ratio of nominal deficit to GDP moves up from 1.8
to 3.2%, with the primary surplus remaining what it was. As to the ratio
of debt to GDP, it also barely changes, going from 40.9 to 40.5%.

Some analytical results

Simple but tedious computations can help explain these results. We can
derive the following steady-state values for three of the main real ratios
of our economy:
The ratio of government expenditure to GDP:
 ∗   
g gd gr − rr(1 − θ ) + π/(1 + π)
=θ+ . (9.22)
y y (1 + gr)
The ratio of public debt to GDP:
 ∗
gd (1 − α1 )(1 − θ )(1 + gr)
= . (9.23)
y gr + α2 + [(1 − α1 )θπ/(1 + π)] − (1 − α1 )(1 − θ)rr
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 201

The ratio of real deficit to real GDP:

 
gd ∗ gr(1 − α1 )(1 − θ )
= . (9.24)
y gr + α2 + [1 − α1 )θπ/(1 + π)] − (1 − α1 )(1 − θ)rr
With no inflation (π = 0), and with the real rate of growth equal to the
real rate of interest net of tax (gr = (1 − θ )rr), these steady-state solutions
get highly simplified:
 ∗
g
=θ (9.22 )
y
 ∗
gd (1 − α1 )(1 − θ )(1 + gr)
= (9.23 )
y α1 gr + α2
 
gd ∗ gr(1 − α1 )(1 − θ )
= . (9.24 )
y (α1 gr + α2 )
In this case, taking the derivative of Equation (9.24 ) with respect to
gr, it is rather obvious that an increase in the real rate of growth of the
economy, accompanied by an equal increase in the real rate of interest
net of tax, will lead to a decrease in the ratio of public debt to GDP, as
long as the propensity to spend out of disposable income is higher than
that out of wealth (α1 > α2 ).4 Only when the growth rate of the economy
gets down to nil – the stationary state – should the real deficit become
zero and the real budget be balanced.
Inspection of Equation (9.23) also shows that – keeping all the other
parameters constant, including the real interest rate – an increase in the
propensity to save out of wealth (α2 ), in the tax rate (θ ), and in the
inflation rate (π ) leads to a lower ratio of steady-state public debt to
GDP, whereas an increase in the real rate of interest (rr) leads to a higher
ratio of steady-state debt to GDP, as one would suspect.

A surprising result
Our simple SFC model can, however, provide us with a more surpris-
ing result. It is usually asserted that, for the debt dynamics to remain
sustainable, the real rate of interest must be lower than the real rate of
growth of the economy for a given ratio of primary budget surplus to
GDP. If this condition is not fulfilled, the government needs to pursue
a discretionary policy that aims to achieve a sufficiently large primary
surplus. We can easily demonstrate that there are no such requirements
in a fully consistent stock–flow model such as ours. The last column of
Table 9.2 shows what occurs if the nominal rate of interest is pushed to
10%, thus raising the real rate of interest rr to 7.84%. Even if we rein-
terpret this condition as meaning that the real rate of interest net of tax
202 Stock–Flow Coherence and Economic Policy

has to be smaller than the real rate of growth, as does Feldstein (1976),
the real rate of interest net of tax, 5.89%, is still way above the real rate
of growth of the economy, which stands at 2.5%. An increase in the real
interest rate induces, in our fiscally generated full-employment model,
a substantial increment in the ratios of public debt to GDP and deficit
to GDP, as many of us would suspect. But this process reaches a limit.
The ratio of (real) primary surplus to GDP achieves a positive figure in the
steady state (here, +4.8%), as traditional analysis would have it when the
rate of interest is larger than the rate of growth. But this is not achieved
in the model by the exogenous imposition of a large primary surplus.
Instead, the only behavioural requirement that has been imposed upon
the public sector is a high enough level of pure government expenditure,
such that full-employment output is verified in each period.
The numbers in the last column of Table 9.2 were not obtained by
relying on the steady-state values of Equations (9.23)–(9.25), although
they correspond to these equations. They were obtained by running our
first model with a simulation program, MODLER. Figure 9.1 illustrates
the transition of our economy from the initial steady state, with low
real interest rates, toward the new steady state, with real interest rates

0.120
Debt to GDP ratio 0.90
(reft-hand scale)
0.100
0.80
0.080
0.70
0.060
0.60
0.040
Real deficit to real GDP ratio
(left-hand scale) 0.50
0.020
0.40
1985 1991 1997 2003 2009 2015 2021 2027 2033 2039 2045 2051

Figure 9.1 Impact of an increase in the nominal interest rate, from 3 to 10%, on
the ratio of real deficit to real GDP and on the ratio of public debt to GDP, when
the real growth rate is still 2.5%
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 203

standing at 7.84%. Clearly, despite the overly high real interest rates, the
ratio of real deficit to real GDP converges, and so does the ratio of public
debt to GDP. The model yields stable, nonexplosive, results.
We have run further experiments, with real rates as high as 25%, and
the model still held up. The ratio of debt to GDP would then rise to
absurd numbers, at about 240%, but the ratio of real deficit to real GDP,
after spiking to above 30% for one period, would be brought back to a
steady ratio of about 7.5%.
Defining the government’s fiscal stance as the ratio of real government
outlays relative to the average tax rate (i.e., (g + rr · gd)/θ ), it follows from
the model that not only must the fiscal stance be set at a particular level
at any point of time for full employment to be achieved, but once full
employment has been achieved, the fiscal stance must grow (by 2.5% per
annum) through time, as long as the real rate of growth in productive
potential remains at 2.5%.
It also follows clearly from Figure 9.1 that if central banks, for whatever
reason, have decided to kick real interest rates up, there will be definite
repercussions on the ratio of deficit to GDP and on the ratio of public
debt to GDP, even if full employment is preserved at all times through
an appropriate choice of the fiscal stance. It makes no sense to put limits
on deficit or debt ratios, as in the Maastricht rules and Gordon Brown’s
golden rules, outside the context of how any economy actually works.

A fiscal policy alternative to the new consensus on

monetary policy

It has been pointed out by a variety of authors that the role of fis-
cal policy has been considerably reduced over the past 20 years or so,
prominence being given to monetary policy to achieve both a target
rate of inflation and a level of demand compatible with potential out-
put or full-employment output. Authors in the new consensus tradition
have been particularly silent with regard to the role that fiscal policy
ought to play. As Arestis and Sawyer point out, ‘the “new consen-
sus” model (or equivalent) provides little role for fiscal policy’ (2004,
p. 455). This is particularly puzzling, because, according to their sur-
vey of central bank empirical results, any negative impact on the rate
of inflation works through reductions in aggregate demand, and these
require very large changes in interest rates to be of any significance. As
a consequence, they conclude by saying that ‘fiscal policy remains a
potent tool for offsetting major changes in the level of aggregate demand’
(ibid., p. 461). Here we wish to show that fiscal policy can in principle
204 Stock–Flow Coherence and Economic Policy

achieve what new consensus authors claim that monetary policy can
achieve.
Some authors say that fiscal policy has been discredited as a short-
term regulator of aggregate demand, because of its well-known logistical
problems, such as lags in legislation, implementation, and effects, as
well as because of the politics involved. Although those concerns are
certainly relevant and worth discussing, we do not wish to address them
at this stage, as we mainly attempt to make a series of theoretical points.
Suffice it to say for the moment that central bankers, now and ever since
the empirical works of Milton Friedman, recognize that monetary policy
usually takes from 12 to 24 months to impinge on inflation. There are
bound to be lags as well with fiscal policy, but fiscal policy has proven
incredibly effective where it has been used relentlessly, for instance, in
the case of the Reagan fiscal expansion in the 1980s and the Bush fiscal
expansion following September 11, 2001.
If lags in the implementation of fiscal policy are to be reduced, there
is clearly a need for institutional change, whereby plans for govern-
ment expenditures – in particular, government investment – would be
prepared way in advance, ready to go when required. Others, such as
Mitchell and Juniper (2007) or Wray (1998), have argued in favour of
public service employment programs that would kick off the moment
output demand falls behind full-employment output.

A fiscal policy reaction function

We start with the simple model that was presented in the previous
section, adding two behavioural equations. First, we now make the rate
of price inflation endogenous, by assuming that inflation reacts to the
output gap, as it does in the much-acclaimed vertical Phillips curve anal-
ysis first introduced by Friedman. New consensus authors, as recalled
by various Post-Keynesian economists in their critiques of the new con-
sensus (Lavoie 2006; Setterfield 2005), usually assume some variant of
the vertical Phillips curve, which, in its most simplified form, can be
presented as:

π = π−1 + ε + γ · (y − yfc )/y (9.25)

or,

π = ε + γ · (y − yfc )/y. (9.25 )

We assume here, although we have denied the relevance of this accel-

erationist view of inflation on numerous occasions (e.g., Godley and
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 205

Lavoie 2007, pp. 301–304, 387–388), that the change in the rate of
inflation depends on the output gap, as usually defined by mainstream
economists, and on some cost-side determinant ε, which we will detail
no more. Thus, yfc stands for potential output and y now stands for the
demand-led actual output, with γ measuring the sensitivity of inflation
to the relative output gap. As we said in the introduction, we introduce
such a vertical Phillips curve as a means of exploring the relevance of fis-
cal policy, in a world – with the accelerationist theory of inflation – that
is most favourable to mainstream economics. If we can demonstrate that
fiscal policy is of supreme relevance within that framework, then a for-
tiori it should play a substantial role in a (Post-Keynesian) world devoid
of the accelerationist hypothesis.
Because we now clearly distinguish between potential output and
actual output, as determined by demand, we need to rewrite two
equations of our simple model. Equations (9.12) and (9.20), which, for
convenience, we repeat here,

y = y−1 · (1 + gr) (9.12)

g ≡ y − px (9.20)

get replaced by Equations (9.12-2) and (9.20-2):

yfc = yfc−1 · (1 + gr) (9.12–2)

y = g + px. (9.20–2)

We thus need an additional equation that will explain real pure gov-
ernment expenditures, g. In analogy with the reaction function of the
central bank, which determines the nominal or the real interest rate set
by the central bank, we define a fiscal reaction function, which defines
the growth rate of real pure government expenditures, calling gr g this
growth rate. We thus have the following two equations:

g = g−1 · (1 + grg ) (9.26)

grg = gr − β1 · π−1 − β2 · (π−1 − π T ). (9.27)

The growth rate of real pure government expenditures gr g is thus

anchored by the growth rate of potential output gr. It is lower than gr
when the lagged inflation rate is rising and when the actual inflation rate
is above the target inflation rate π T , a target presumably set together by
the central bank and the government. Because of Equation (9.25) and its
accelerationist hypothesis, to say that the growth rate of real pure gov-
ernment expenditures is lower when the rate of inflation rises implies
206 Stock–Flow Coherence and Economic Policy

that this growth rate will tend to be lower when actual output overtakes
potential output.
Obviously, this kind of fiscal policy mimics the various central bank
reaction functions that have been proposed since the 1990s. In particular,
gr, the rate of growth of potential output, or the natural rate of growth,
plays a role that is similar to that of the natural rate of interest in the new
consensus reaction function equations. It is assumed that governments
react to lagged inflation rates, rather than to actual or expected inflation
rates, on the realistic grounds that fiscal policy may have a reaction time
somewhat longer than monetary policy.

Experiments with the fiscal policy reaction function model

We can conduct various experiments with our slightly more sophis-
ticated SFC model. As usual, we start from a baseline case, where
steady-state positions have been reached – with capacity, real output,
and real government expenditures all growing at 2.5%, along with the
real stocks of the economy. Inflation, as before, is assumed to run at 2%.
The nominal and real interest rates, as before, are set at 3% and nearly 1%,
respectively. Experiments have shown that the behaviour of the model
hardly changes whether nominal or real interest rates are considered to
be the exogenous variable. In the figures that will be shown, it has been
assumed that the central bank has given itself as a policy to keep the real
rate at a constant level, so that Equation (9.10) needs to be reversed into
Equation (9.10-2), which becomes the central bank reaction function:

r = rr + π + π · rr. (9.28)

As a first experiment, let us assume that the central bank is unhappy

with its current inflation target, and has managed to successfully lobby
the government into accepting a lower inflation target, say π T = 1.5%.
What will then happen? Figures 9.2, 9.3, and 9.4 show the impact on
some of the main variables of the model. First, Figure 9.2 shows that fiscal
policy is able to smoothly get the rate of inflation down to its new lower
target. The lower rate of inflation is achieved by getting the economy
to operate at lower than full employment for a number of periods (the
years on the charts may be imagined as being quarters), as can be seen in
Figure 9.3, thus creating downward pressure on demand inflation. All of
this is accomplished through fiscal policy, as the government lowers the
rate of growth of real pure government expenditures in the initial time
periods, as can be seen in Figure 9.4. By doing this, the rate of growth
of output demand follows the path of the growth rate of government
expenditure, but with less amplitude. In the latter periods, pure real
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 207

Old target
0.0200

0.0190

0.0180

0.0170 Inflation rate

0.0160

New target
0.0150
1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Figure 9.2 Evolution of the inflation rate, following a reduction in the target rate
of inflation, from 2 to 1.5%

1.020

1.010

Actual output to potential output ratio

1.000

0.990

0.980

0.970

1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Figure 9.3 Evolution of the ratio of actual output to potential output, following
a reduction in the target rate of inflation, from 2 to 1.5%
208 Stock–Flow Coherence and Economic Policy

0.0375

Growth rate of real government expenditures

0.0300

0.0225

0.0150

0.0075

0.0000
1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Figure 9.4 Evolution of the growth rate of real pure government expenditures,
following a reduction in the target rate of inflation, from 2 to 1.5%

government expenditures and real output must grow at a pace that is

faster than the natural rate of growth, as actual output and employment
must catch up with potential output and full employment. In the end,
the lower inflation target has been achieved by forcing the economy to
operate at less than full employment for a number of periods. The output
thus lost has been lost forever.
As a second experiment, let us assume that households decide to raise
their propensity to consume out of disposable income (the α1 coefficient
is moved up, through a higher α10 ). This should initially lead to an
increase in aggregate demand and, hence, in inflation. Indeed, inflation
rises, only to gradually go back to its target level. Figure 9.5 shows the
evolution of the growth rate of pure government expenditure, and that
of actual output, as fiscal policy attempts to mitigate the inflationary
effects of the increase in private spending.
As a third and final experiment, let us assume that the central bank,
acting on the lobby of rentiers, decides to raise the real rate of interest
from 1 to 7%. What will occur? Figure 9.6 shows the evolution of the
inflation rate. With the initial increase in interest outlays out of govern-
ment debt, there is an increase in private expenditure, which leads to a
brief and small increase in the inflation rate, as can be seen in Figure 9.6.
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 209

0.050
Growth rate of real output

0.040

Growth rate of pure real government expenditures

0.030

0.020

0.010

0.000

1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Figure 9.5 Evolution of the growth rate of real output and of the growth rate of
pure real government expenditures, following an increase in the propensity to
consume out of disposable income

0.0250
Inflation rate

0.0225

0.0200
Inflation target

0.0175

0.0150

0.0125

1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Figure 9.6 Evolution of the inflation rate, following an increase in the real rate
of interest, from 1 to 7%
210 Stock–Flow Coherence and Economic Policy

0.30

0.20

0.10

0.00
Growth rate of real pure government expenditure
–0.10
Growth rate of real output

–0.20

1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Figure 9.7 Evolution of the growth rate of output and of the growth rate of real
pure government expenditures, following an increase in the real rate of interest,
from 1 to 7%

However, immediately afterward, the inflation rate drops briskly, finally

coming back to its initial target level after some overshooting. What hap-
pens is that, as can be seen in Figure 9.7, as the private sector reacts with a
one-period lag to the new higher real interest rate, they decide to reduce
their propensity to spend out of disposable income, thus plunging the
economy into a recession. The fiscal authorities, also with a lag, try to
maintain the economy close to full employment, by hiking up the rate of
growth of real pure government expenditures. Eventually, the economy
comes back to full employment, at the natural rate of growth. However,
as can be seen from Figure 9.8, all of this adjustment can only occur if
the government, and financial markets, accept letting the ratio of public
debt to GDP double, from about 41% to nearly 85%. As to the ratio of
real deficit to real GDP (not shown here), it peaks for a while at 9%, while
its steady-state level rises from 1 to 2%. Once again, despite the fact that
the real rate of interest after tax is much higher than the trend real rate
of growth of the economy, all adjustments are sustainable and the model
remains stable.
The lesson to be drawn from all of this is that fiscal policy is, in theory,
capable of achieving full employment at some target inflation rate. It
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 211

0.90

Government debt to GDP ratio

0.80

0.70

0.60

0.50

0.40
1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Figure 9.8 Evolution of the ratio of public debt to GDP, following an increase in
the real rate of interest, from 1% to 7%

is not clear what advantage monetary policy has, besides the fact that
target interest rates can be easily altered every month or even every week.
Indeed, by bringing back fiscal policy as the main tool to affect in the
real rate of interest, from 1 to 7% aggregate demand, monetary policy
would now have an additional degree of freedom to set the real interest
rate, which is a key determinant of distribution policy. The real interest
rate could be set at its fair level, which, according to Pasinetti (1981),
is equal to the trend rate of growth of labour productivity (see Lavoie
and Seccareccia 1996). With such a fair rate of interest, the earnings of
one hour of labour, when they are saved, allow its owner to obtain a
purchasing power that is equivalent to that obtained with the earnings
of one hour of labour in the future.

The simplified model again, with a foreign sector

In this section, we open the economy, postulating a foreign sector that

exports (X and x) and imports (IM and im) goods and services. Export
and import prices move with domestic prices but imports always exceed
exports by 5%, so that with exports rising at the same rate as GDP there
is always a trade deficit equal to 1% of GDP (given the assumed ratios of
212 Stock–Flow Coherence and Economic Policy

trade to GDP). All the other assumptions about exogenous variables that
were entertained in the first section are retained. This implies that the
following equations are modified or added:
g ≡ y − px − (x − im) (9.20–2)
x ≡ x−1 · (1 + gr) (9.28)
im ≡ x · 105% (9.29)
X ≡ x·p (9.30)
IM ≡ im · p. (9.31)
The balance of payments on current account is equal to the trade bal-
ance plus or minus the flow of interest payments abroad, which are given
by r · VF−1 , where VF is the stock of overseas financial wealth, changes in
which are equal each period to the current account balance (CAB). This
implies the following equalities:
CAB ≡ X − IM + r · VF−1 (9.32)
VF ≡ VF−1 + CAB. (9.33)
The redundant equation, which was NAFA ≡ DEF in the closed econ-
omy, is now equal to
NAFA ≡ DEF + CAB. (D)
This is a well-known flow of funds identity, of which forecasters and
analysts now make use (see Godley 1999).
We start this highly simple open-economy model from a situation
where trade is balanced, assuming the country neither holds foreign
assets nor owes debt to foreigners. Then, in the second period, we
impose upon it the conditions that were described in Equations (9.28)
and (9.29) – that is, we impose a perpetual trade deficit. The solutions
of this model have two important properties. First, the model converges
to stable ratios when the current account balance reaches (nearly) –2.5%
of GDP. This is perhaps a surprising result, for it is commonly assumed
that if a country is indebted to the rest of the world, stability can only
come about if the balance of trade is positive. Second, the solutions show,
rather obviously, that if there is a chronic current account deficit of 2.5%
relative to GDP, then, other things being equal, the budget deficit must
be 2.5 percentage points higher than would otherwise be the case. In the
present case, with the ratio of current account deficit to GDP moving up
through time from 1 to 2.5%, as can be ascertained from Figure 9.9, the
government budget deficit must move from 2.8 to 4.3% of GDP.
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 213

0.0375
Nominal government deficit to GDP ratio

0.0250
Net accumulation of financial asset to GDP ratio

0.0125

0.0000

Trade balance to GDP ratio

–0.0125
Current account balance to GDP ratio

–0.0250

1952 1961 1970 1979 1988 1997 2006 2015 2024 2033 2042 2051

Figure 9.9 Evolution of the main balances, following the appearance of a trade
account deficit that stands forever at 1% of GDP

Conclusion

The purposes of this paper are, first, to insist that there exist rules that
must govern the conduct of fiscal policy as the counterpart of stable
growth without inflation or unemployment and to make suggestions as
to how those rules should be formulated. In addition, external trade
or current account deficits have implications for deficit ratios and debt
ratios. Finally, we are tentatively drawing two unconventional conclu-
sions: that an economy (described within an SFC framework) with a real
rate of interest net of taxes that exceeds the real growth rate will not nec-
essarily generate explosive interest flows, even if the government makes
no discretionary attempt to achieve primary budget surpluses, and, sec-
ond, that it cannot be assumed that a debtor country requires a trade
surplus if interest payments on debt are not to explode.
We have shown that fiscal policy can deliver sustainable full employ-
ment at a target inflation rate within an SFC framework with some
arbitrary interest rate. It follows from our model that if the fiscal stance
is not set in the appropriate fashion – that is, at a well-defined level
and growth rate – then full employment and low inflation will not be
achieved in a sustainable way. As far as we know, new consensus authors
have shown only that monetary policy could provide full employment at
214 Stock–Flow Coherence and Economic Policy

some target inflation rate over a short period, with fiscal policy left hang-
ing in the air. They have yet to demonstrate such a result over the long
run within an SFC framework.

Notes
1. Obvious examples are the Maastricht rules in the European Union, Gordon
Brown’s ‘golden’ rule in the United Kingdom, and various rules forbidding or
attempting to forbid government deficits. For a previous formal critique, see
Godley and Rowthorn (1994).
2. This expression is logically equivalent to the following, which we commonly
used in our book (Godley and Lavoie 2007), yd ≡ (Y +r ·V−1 −T )/p −p ·v−1 /p,
where uppercase letters describe nominal variables (i.e., Y is nominal income,
T is nominal tax payments, V is nominal wealth, r is the nominal interest rate,
and p is the price level).
3. In the wording of our book (Godley and Lavoie 2007), as can be seen from the
one before the last row of Table 9.1, the redundant equation is DEF = NAFA.
4. This effect will be further enforced because an increase in rr leads to an induced
fall in the propensity to consume out of disposable income, the α1 coefficient,
according to Equation (9.3).

References

Arestis, P. and M. Sawyer (2004) ‘On the Effectiveness of Monetary Policy and of
Fiscal Policy.’ Review of Social Economy 62 (4) (December): 441–463.
Feldstein, M. (1976) ‘Perceived Wealth in Bonds and Social Security: A Comment.’
Journal of Political Economy 84 (2) (April): 331–336.
Godley, W. (1999) ‘Seven Unsustainable Processes: Medium-Term Prospects and
Policies for the United States and the World.’ Strategic Analysis, Levy Economics
Institute of Bard College, Annandale-on-Hudson, NY.
Godley, W. and M. Lavoie (2007) Monetary Economics: An Integrated Approach
to Credit, Money, Income, Production and Wealth (Basingstoke, UK: Palgrave
Macmillan).
Godley, W. and B. Rowthorn (1994) ‘Appendix: The Dynamics of Public Sector
Deficits and Debt.’ In J. Michie and J. Grieve Smith (eds.), Unemployment in
Europe (London: Academic Press), pp. 199–206.
Lavoie, M. (2006) ‘A Post-Keynesian Amendment to the New Consensus on
Monetary Policy.’ Metroeconomica 57 (2) (May): 165–192.
Lavoie, M. and M. Seccareccia (1996) ‘Central Bank Austerity Policy, Zero-Inflation
Targets, and Productivity Growth in Canada.’ Journal of Economic Issues 30 (2)
(June): 533–544.
Mitchell, B. and J. Juniper (2007) ‘Towards a Spatial Keynesian Macroeconomics.’
In P. Arestis and G. Zezza (eds.), Advances in Monetary Policy and Macroeconomics
(London: Palgrave Macmillan), pp. 192–211.
Pasinetti, L.L. (1981) Structural Change and Economic Growth (Cambridge:
Cambridge University Press).
 Fiscal Policy in a Stock–Flow Consistent (SFC) Model 215

Schlicht, E. (2006) ‘Public Debt as Private Wealth: Some Equilibrium Considera-

tions.’ Metroeconomica 57 (4) (November): 494–520.
Setterfield, M. (2005) ‘Central Bank Behaviour and the Stability of Macroeconomic
Equilibrium: A Critical Examination of the “New Consensus”.’ In P. Arestis,
M. Baddeley, and J. McCombie (eds.), The New Monetary Policy: Implications and
Relevance (Cheltenham, UK: Edward Elgar), pp. 23–49.
Wray, R. (1998) Understanding Modern Money: The Key to Full Employment and Price
Stability (Cheltenham, UK: Edward Elgar).
10
Seven Unsustainable Processes:
Medium-Term Prospects and
Policies for the United States and
the World
Wynne Godley

The US economy has now been expanding for nearly eight years, the
budget is in surplus, and inflation and unemployment have both fallen
substantially. In February the Council of Economic Advisers (1999) fore-
cast that GDP could grow by 2.0 to 2.4% between now and the year 2005,
and this forecast has since been revised upwards (Office of Management
and Budget 1999). Many people share the CEA’s optimistic views. For
instance, in his New Year message (Financial Times December 29, 1998)
Alan Blinder compared the United States’s economy to one of its mighty
rivers – it would ‘just keep rolling along’; and President Bill Clinton con-
cluded his Economic Report of the President with the words ‘There are no
limits to the world we can create, together, in the century to come.’ This
paper takes issue with these optimistic views, although it recognizes that
the US economy may well enjoy another good year or two.
During the last seven years a persistently restrictive fiscal policy has
coincided with sluggish net export demand, so rapid growth could come
about only as a result of a spectacular rise in private expenditure relative to
income. This rise has driven the private sector into financial deficit on an
unprecedented scale. The Congressional Budget Office (CBO) is project-
ing a rise in the budget surplus through the next 10 years, conditional on
growth’s continuing at a rate fast enough to keep unemployment roughly
constant, and this implies that it is government policy to tighten its
restrictive fiscal stance even further (Congressional Budget Office 1999a,
1999c). At the same time, the prospects for net export demand remain
unfavourable. But these negative forces cannot forever be more than off-
set by increasingly extravagant private spending, creating an ever-rising
excess of expenditure over income.
If spending were to stop rising relative to income without there being
either a fiscal relaxation or a sharp recovery in net exports, the impetus
216
 Seven Unsustainable Processes 217

that has driven the expansion so far would evaporate and output would
not grow fast enough to stop unemployment from rising. If, as seems
likely, private expenditure at some stage reverts to its normal relation-
ship with income, there will be, given present fiscal plans, a severe and
unusually protracted recession with a large rise in unemployment.
It should be added that, because its momentum has become so depen-
dent on rising private borrowing, the real economy of the United States
is at the mercy of the stock market to an unusual extent. A crash would
probably have a much larger effect on output and employment now than
in the past.
A long period of stagnation in the United States, still more recession,
would have grave implications for the rest of the world, which seems to
be depending, rather irresponsibly, on the United States to go on acting
as spender of last resort indefinitely.
This paper makes no short-term forecast. Bubbles and booms often
continue much longer than anyone can believe possible and there could
well be a further year or more of robust expansion. The perspective taken
here is strategic in the sense that it is only concerned with developments
over the next 5 to 15 years as a whole. Any recommendations regarding
policy do not have the character of ‘fine-tuning’ in response to short-
term disturbances. They ask, rather, whether the present stance of either
fiscal or trade policy is structurally appropriate looking to the medium-
and long-term future.
A sustained period of stagnation or recession, through its adverse effect
on the national income, could drive the budget back into deficit with-
out there being any relaxation of policy, yet to counteract an endemic
recession, it will be necessary to relax fiscal policy, making any emerging
deficit even larger. Further relaxation of monetary policy could not sus-
tain the expansion, except temporarily and perversely by giving a new
lease on life to the stock market boom. While a relaxation in the stance
of fiscal policy will ultimately have to be made, this by itself will not
be enough to generate balanced growth in the medium term because, as
matters stand, this would be accompanied by a continuing rise in the
United States’s external deficit and indebtedness. There is probably no
way in which sustained and balanced growth can be achieved in the
medium term except through co-ordinated fiscal expansion worldwide.
The difference between the consensus view and that put forward here
could not exist without a profound difference in the view of how the
economy works. So far as the author can observe, the underlying theo-
retical perspective of the optimists, whether they realize it or not, sees all
agents, including the government, as participants in a gigantic market
process in which commodities, labour, and financial assets are supplied
218 Stock–Flow Coherence and Economic Policy

and demanded. If this market works properly, prices (e.g., for labour
and commodities) get established that clear all markets, including the
labour market, so that there can be no long-term unemployment and
no depression. The only way in which unemployment can be reduced
permanently, according to this view, is by making markets work better,
say, by removing ‘rigidities’ or improving flows of information. The gov-
ernment is a market participant like any other, its main distinguishing
feature being that it can print money. Because the government cannot
alter the market-clearing price of labour, there is no way in which fiscal or
monetary policy can change aggregate employment and output, except
temporarily (by creating false expectations) and perversely (because any
interference will cause inflation).
No parody is intended. No other story would make sense of the
assumption now commonly made that the balance between tax receipts
and public spending has no permanent effect on the evolution of the
aggregate demand. And nothing else would make sense of the debate
now in full swing about how to ‘spend’ the federal surplus as though
this were a nest egg that can be preserved, spent, or squandered without
any need to consider the macroeconomic consequences.
The view taken here, which is built into the Keynesian model later
deployed, is that the government’s fiscal operations, through their
impact on disposable income and expenditure, play a crucial role in
determining the level and growth rate of total demand and output.
The circumstances that have generated a budget surplus combined with
falling unemployment are not only unusual but essentially temporary.
No decision to ‘spend’ a surplus can be taken without regard for the
impact on aggregate demand. In any case, there may soon be no surplus
to spend.
This paper first looks at where the current growth has come from,
examining, in turn, fiscal policy, foreign trade, and private income
expenditure and borrowing. This examination shows that current
growth is associated with seven unsustainable processes in the United
States: (1) the fall in private saving into ever deeper negative territory,
(2) the rise in the flow of net lending to the private sector, (3) the rise
in the growth rate of the real money stock, (4) the rise in asset prices
at a rate that far exceeds the growth of profits (or of GDP), (5) the rise
in the budget surplus, (6) the rise in the current account deficit, (7) the
increase in the United States’s net foreign indebtedness relative to GDP.
The paper then presents a number of medium-term scenarios based on
models of the United States and world economies, considers some of their
implications, and discusses appropriate policy responses. The appendixes
contain notes on the models used and some econometric results.
 Seven Unsustainable Processes 219

Fiscal policy

In the United States the public discussion of fiscal policy concentrates

almost exclusively on the operations of the federal government. Yet state
and local governments account for about a third of all public expenditure
and taxes; moreover, their budgets are generally in surplus so that these
authorities are now in substantial credit – a fact easily verifiable from
the national income and product accounts (NIPA), which show them to
be large net receivers of interest and dividend income. In what follows,
government inflows and outflows – and debts – will always refer to the
operations of the ‘general government’ (the combined federal, state, and
local governments).
The stance of fiscal policy is usually measured by the general govern-
ment structural balance, that is, the size of the budget surplus or deficit,
preferably corrected for the business cycle and for inflation. The gov-
ernment’s fiscal stance is said to be neutral if the deficit is small and
does not increase, as a share of GDP, through time. Figure 10.1 portrays

1.0

0.5

0.0
Percent of potential GDP

–0.5

–1.0

–1.5

–2.0

–2.5

–3.0

–3.5

–4.0
1982 1984 1986 1988 1990 1992 1994 1996 1998

Figure 10.1 General government structural balance

Note: The vertical line is drawn at 1992 to mark the transition from expansionary
to restrictive fiscal policy.
Source: OECD Economic Outlook, December 1998.
220 Stock–Flow Coherence and Economic Policy

the adjusted budget deficit since 1982, showing that fiscal policy was
expansionary until 1992 but has been restrictive since then.
The data illustrated in Figure 10.1 may be supplemented with an alter-
native but closely related measure of fiscal stance, namely, the ‘fiscal
ratio’ or the ratio of government spending to the average rate of taxa-
tion. When the budget is balanced, this fiscal ratio will be exactly equal
to GDP; it will exceed GDP when the budget is in deficit and fall short
of it when the budget is in surplus.1 The advantage of measuring fis-
cal stance this way is that it makes it easy to make simple inferences
about fiscal policy. For instance, we can infer that, with a neutral fiscal
stance, real government expenditure, given the average tax rate, must
rise through time at the same rate as GDP; alternatively, tax rates must
fall if real government expenditure is held constant.
Corrected for inflation and the business cycle, the adjusted fiscal ratio
(AFR) rose, between 1960 and 1992, at an average rate of 3.6% per
annum, while GDP rose at an average rate of 3.3%. As Figure 10.2 shows,

8,000

7,000
Adjusted final rate (IMR)
$ Billion (at 1992 prices)

6,000 GDP

5,000

4,000

3,000

2,000
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1961 1964 1969 1972 1977 1980 1985 1988 1992 1997 2000

Figure 10.2 Adjusted fiscal ratio and GDP

Note: In this and the following figures, the vertical line is drawn at 1992Q3 unless
otherwise indicated.
Source: Citibase and author’s calculations (see text for details).
 Seven Unsustainable Processes 221

during the last seven years the average growth rate of the AFR was 0.9%,
while GDP continued to rise at an average rate of 3.3%. By this measure,
fiscal policy since 1992 has been far more restrictive than during any
seven-year period in the last 40 years.

Foreign trade and payments

Nor has there been much stimulus to the economy from net export
demand. As Figure 10.3 shows, the current balance of payments has
been in continuous and growing deficit throughout the last seven years.
It also shows that the deterioration in the balance of trade in manufac-
tured goods was enough, by itself, to account for the whole deterioration
in the current account. Manufactures continue to dominate changes in
international trade despite the greatly diminished role of manufacturing
as an employer of labour and a generator of value added. In the last few
years a perceptible part of the overall deterioration has been caused by
trade in computers, which was nearly $30 billion in deficit in 1998.
The view taken here, for which prima facie support is provided in
Figure 10.3, is that the growing current account deficit in the United

2.0

1.0 Balance of trade in manufactures

0.0
Percent of GDP

–1.0

–2.0

Current balance of payments

–3.0

1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000

Figure 10.3 Current balance of payments and balance of trade manufactures

Note: 1998 parlty estimated.
Source: Citibase and author’s estimates.
222 Stock–Flow Coherence and Economic Policy

States has little to do with domestic saving and investment patterns,

although there is an accounting identity that links the national saving
with the current balance of payments. The growing deficit is mainly
the consequence of an increasingly successful invasion of US markets by
foreign manufacturers and increased outsourcing of intermediate prod-
ucts. This long-standing adverse trend in trade has been aggravated by
the recent collapse of Asian markets and the appreciation of the dollar
since 1996.
Whatever the cause, there is no question but that over the ‘Goldilocks’
period as a whole net export demand has made only a weak contribution
to the growth of aggregate demand; since the beginning of 1998 its con-
tribution has been negative, even after allowing for the improvement in
the US terms of trade, which, taken by itself, had a beneficial effect on
the real national income.
To get an overall impression of the effect of foreign trade on aggre-
gate demand during the past 40 years, Figure 10.4 shows a measure of
international trade performance that will be called the ‘adjusted trade
ratio’ (ATR). The ATR is constructed according to the same principles
as the AFR, that is, it is the ratio of exports and foreign transfers to

8,000

7,000
$ Billion (at 1993 prices)

6,000

5,000

4,000
GDP
Adjusted trade ratio (ATR)
3,000

2,000
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1960 1964 1968 1972 1976 1980 1984 1986 1992 1998 2000

Figure 10.4 Adjusted trade ratio and GDP.

Source: Citibase and author’s calculations (see text for details).
 Seven Unsustainable Processes 223

8,000

7,000
$ Billion (at 1993 prices)

Combined final and trade ratio (CFTR)

6,000
Adjusted final ratio (AFR)

5,000

4,000

3,000 Adjusted trade ratio (ATR)

2,000
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001

Figure 10.5 Combined fiscal and trade ratio.

Source: Citibase and author’s calculations (see text for details).

the average import propensity, with all variables corrected for inflation,
relative prices, and the business cycle.2
Figure 10.5 combines the fiscal and trade ratios into a ‘combined fis-
cal and trade ratio’ (CFTR).3 The CFTR measures the extent to which
these exogenous factors, taken together, fed the growth of aggregate
demand; it shows, that is, the extent to which government expendi-
ture plus exports pumped funds into the economy relative to the rate at
which taxes and imports siphoned funds out of it.
The view taken here is that since stocks of assets and liabilities are
unlikely to rise or fall indefinitely relative to income flows, the GDP
should normally track the CFTR roughly one for one, albeit erratically.
The theoretical basis for this view, which has a respectable pedigree
starting with Carl Christ (1968) and Blinder and Solow (1973), may be
conveyed using a hydraulic analogy. If water (government expenditure
plus exports) flows into a receptacle at some given rate, and if a certain
proportion of the water (tax payments and imports) flows out of it at
some other rate, the level of the water in the receptacle will change. If
the water reaches a stable level (regardless of what that level is), at the
point at which it stabilizes, outflows must be exactly equal to inflows.
Whenever the inflow of government expenditure plus exports is equal
224 Stock–Flow Coherence and Economic Policy

8,000

7,000
Combined final and trade ratio (CFTR)
$ Billion (at 1993 prices)

6,000 GDP

5,000

4,000

3,000

2,000
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001

Figure 10.6 Combined fiscal and trade ratio and GDP.

Source: Citibase and author’s calculations (see text for details).

to the outflow of taxes plus imports, the level of aggregate income and
output must be equal to the CFTR.4
How does this story square with the facts? Figure 10.6 shows the CFTR
together with GDP since 1961. Between 1961 and 1992 GDP did indeed
track the CFTR one for one, if erratically. Since the beginning of 1992,
while GDP has risen 3.3% per annum, the CFTR has risen only 0.6% per
annum. However sceptical the reader may be concerning our stock/flow
model, there is no gainsaying the facts displayed in Figure 10.6 – net
demand from the government and net exports since 1992 have been
much weaker than in any other period since 1960.

Private saving, spending and borrowing

How could the economy expand so fast after 1992 seeing that the arterial
flows that normally make it grow were so sluggish? An answer is sug-
gested in Figure 10.7, which shows the three major financial balances:
the private financial balance between total income and expenditure, the
general government balance, and the current balance of payments. As
every student of the NIPA knows, these three balances must, by account-
ing identity, sum to zero. In Figure 10.7 public borrowing is given a
 Seven Unsustainable Processes 225

7.5 Private financial balance

General government balance

5.0 (written as a default)
Percent of GDP

2.5

0.0

–2.5

Current balance of payments

(written as a surples)
–5.0

Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000

Figure 10.7 The three major financial balances.

Source: Citibase and author’s estimates.

positive sign so as to make it crystal clear that the private deficit is always
exactly equal to the public surplus plus the balance of payments deficit.
The intuition that underlies this rearrangement of the numbers is that
public deficits and balance of payments surpluses create income and
financial assets for the private sector whereas budget surpluses and bal-
ance of payments deficits withdraw income and destroy financial assets.
This method of presenting the figures makes the way financial assets and
income are created for the private sector quite transparent.
As the budget balance during the last seven years has changed by a
larger amount than ever before (at least since the early 1950s) and has
reached a record surplus (2.2% of GDP in the first quarter of 1999) and as
the current balance of payments has deteriorated rapidly, it comes as no
surprise to find that the private sector balance has moved south as well,
again by a record amount and reaching a record deficit (5.2% of GDP in
the first quarter of 1999).
The scale of the private financial deficit, though subject to revision,
cannot be called into question (significantly) by any redefinition of per-
sonal income, saving, consumption, or investment. The private financial
deficit measures something straightforward and unambiguous; it mea-
sures the extent to which the flow of payments5 into the private sector
226 Stock–Flow Coherence and Economic Policy

arising from the production and sale of goods and services exceeds pri-
vate outlays on goods and services and taxes, which have to be made in
money. While capital gains obviously influence many decisions, they do
not by themselves generate the means of payment necessary for transac-
tions to be completed; a rise in the value of a person’s house may result
in more expenditure by that person, but the house itself cannot be spent.
The fact that there have been capital gains can therefore be only a par-
tial explanation of why the private sector has moved into deficit. There
has to be an additional step; money balances must be run down (surely
a very limited net source of funds) or there must be net realizations of
financial assets by the private sector as a whole or there has to be net bor-
rowing from the financial sector. Furthermore, a capital gain only makes
a one-time addition to the stock of wealth without changing the flow of
income. It can therefore, by its very nature, have only a transitory effect
on expenditure. It may take years for the effect of a large rise in the stock
market to burn itself out, but over a strategic time period, say 5 to 10
years, it is bound to do so.
While Figure 10.7 implies that private expenditure has been exceeding
income by growing amounts, it tells us nothing directly about what has
caused the expansion of the economy. For all that Figure 10.7 contains,
the growth in private expenditure relative to income could have been
accompanied by an absolute fall in both series. Figure 10.8, which shows
private income and expenditure separately, puts it beyond doubt that it
is the rapid relative rise in private expenditure that has been the main
driving force behind the US expansion since 1991–1992.
It has occasionally been said that the rise in private expenditure rela-
tive to income is the expected and healthy consequence of the budget
tightening that, by reducing interest rates, has stimulated investment.
This would be the explanation suggested by many modern textbooks on
macroeconomics.
But, as Figure 10.9 demonstrates, this explanation is clearly incorrect.
There has been a moderate increase in business investment, which rose
from 9% of GDP in 1992 to 11% at the beginning of 1999. But, in 1992
the business sector was so substantially in surplus (that is, undistributed
profits were so substantially in excess of investments) that it has only
just moved south of the zero line; in recent quarters almost all busi-
ness investment was financed from internally generated funds. As the
figure shows, most of the fall in the private balance and the entire deficit
has taken place in the household sector. It is the excess of personal con-
sumption and housing investment over personal disposable income that
is now much larger than ever before.
 Seven Unsustainable Processes 227

7,000

6,000
$ Billion (at 1993 prices)

5,000

Real private disposable income

4,000

Real private expenditure

3,000

2,000

1,000
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000

Figure 10.8 Real private expenditure and disposable income.

Source: Citibase and author’s estimates.

4.0

2.0
Percent of GDP

0.0

–2.0
Business saving less investment

Household saving less investment

–4.0

Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000

Figure 10.9 Analysis of private financial deficit.

Source: Citibase and author’s estimates.
228 Stock–Flow Coherence and Economic Policy

20.0
Percent of private disposable income

15.0
Not lending to
private sector

10.0

5.0

0.0

Private financial balance

–5.0

–10.0
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000

Figure 10.10 Private financial balance and net lending to private sector.
Source: Citibase and author’s estimates.

The descent of the private sector into financial deficit means that the
sector as a whole has become a net borrower (or a net seller of financial
assets) on a record and growing scale. Figure 10.10 illustrates the inverse
relationship between the flow of net lending to the non-financial pri-
vate sector (derived from the Flow of Funds) and the balance of private
income and expenditure (derived from the NIPA). It shows, in particular,
that the annual rate of net lending rose fairly steadily from about 1%
of disposable income ($40 billion) at the end of 1991 to 15% (over $1
trillion) in the first quarter of 1999. However, while the private finan-
cial deficit was far larger than ever before, the flow of net lending in the
fourth quarter was some way from being a record. Ignoring the possi-
bility that the figures will be revised, the reasons for this may be, first,
that leasing of motor vehicles has increased (the underlying purchase
presumably now consisting of fixed investment by the financial sec-
tor). Second, households have also been able, up to a point, to make
net realizations of capital gains without borrowing and without causing
the market to move against them because firms have been repurchas-
ing equity while foreigners have been making large net purchases of
US stocks.
 Seven Unsustainable Processes 229

12.0

10.0

8.0

6.0
Percent change

4.0

2.0

0.0

–2.0

–4.0

–6.0
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1960 1965 1970 1975 1980 1985 1990 1995

Figure 10.11 Growth in real (deflated) stock of money (M3), compared with a
year earlier.
Source: Citibase and author’s estimates.

Figure 10.11 shows the annual rate of growth in the real stock of money
(M3) compared with the year earlier. It rose from minus 3% in 1992 to
nearly 10% at the beginning of 1999. The growth rate of the real money
stock during the past year far exceeds the high rates of the mid 1980s and
has reached the extremely high rates of the early 1970s. The expansion
in money supply growth is the flip side of the credit expansion illustrated
in Figure 10.10 and confirms that the growth of net lending did indeed
continue up to the first quarter of 1999.

The strategic prospects

The central contention of this paper is that, given unchanged fiscal pol-
icy and accepting the consensus forecast for growth in the rest of the
world, continued expansion of the US economy requires that private
expenditure continues to rise relative to income. Yet while anything can
happen over the next year or so, it seems impossible that this source of
growth can be forthcoming on a strategic time horizon. The growth in
230 Stock–Flow Coherence and Economic Policy

net lending to the private sector and the growth in the growth rate of the
real money supply cannot continue for an extended period. Moreover,
if, per impossibile, the growth in net lending and the growth in money
supply growth were to continue for another eight years, the implied
indebtedness of the private sector would then be so extremely large that
a sensational day of reckoning could then be at hand. In sum, if a truly
strategic view is taken, covering the next 10 to 15 years, one is forced to
the conclusion that the present stance of policy is fundamentally out of
kilter and will eventually have to be changed radically.

Projections based on CBO forecasts

To illustrate the scale of the problem, some simulations were done that
show what has to be assumed about private income, expenditure, and
borrowing to validate the CEA’s forecasts. Figure 10.12 gives projections
of the three major financial balances between now and 2008. It was
assumed, with the CBO’s April projection, that GDP grows at 2.0 to 2.4%

8.0
Private financial balance

General government balance

4.0
Percent of GDP

0.0

–4.0 Current balance of payments

–8.0

Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1970 1975 1980 1985 1990 1995 2000 2005 2010

Figure 10.12 The three major financial balances, actual 1970–1999Q1 and pro-
jections implied by CBO
Note: Data after 1999Q1, where the vertical line is now drawn, are author’s
projections.
Source: Citibase, Flow of Funds, and author’s projections.
 Seven Unsustainable Processes 231

and that inflation is stable at 2.1%. The projected surplus of the general
government was derived by taking an average of the two projections that
the CBO makes, adding (about) 1% of GDP to allow for the surpluses of
state and local governments and scaling the result to harmonize with
national income concepts.
For the balance of payments, a projection of output over the next five
years (to 2004) in every foreign country or country ‘bloc’ was made, using
consensus forecasts and adding them together using US shares in each
bloc’s imports.6 For the rest of the period (from 2004 to 2008), it was
assumed that (non-US) world output grows at its long-term average rate.
Estimates of the US balance of trade were then derived using standard
equations describing the behaviour of export and import volumes and
prices, assuming no further change in exchange rates. These projections
are believed to be noncontroversial, given the medium-term outlook
for the United States’s main markets and given its well-attested high
income elasticity of demand for imports. The projected trade balance
improves perceptibly after 2004 because of an assumed recovery in world
production and trade, but the effect of this on the balance of payments
is muted by a rise in factor income payments as net indebtedness soars
toward $6 trillion, or nearly 45% of GDP.7
Accordingly, the growing budget surplus projected by the CBO, taken
in conjunction with the balance of payments projections shown in
Figure 10.12, carries the implication, since the three balances must sum
to zero, that the private sector deficit continues to rise for the next six or
seven years and even then does not fall significantly.
Figure 10.13 shows the history of the private surplus and net lending
(reproducing, for the past, the material in Figure 10.10) together with
projections of both series into the medium-term future. If expenditure
continues to rise relative to income, the flow of net lending must go on
rising as well.
The implications of these forecasts are revealed in Figure 10.14, which
shows the implied level of debt relative to disposable income. If the
flow of net lending continues at 15% or more of disposable income,
the rise in the burden of indebtedness would accelerate away from its
present record level of about 1.6% of disposable income, nudging 2.5%
toward the end of 2008, and still rising rapidly after that. While the
stock exchange boom has generated so much wealth that the exist-
ing level of indebtedness may not, in general, pose a threat to private
balance sheets at the moment, the same thing could hardly be true if
indebtedness rises two or threefold (or more) compared with its existing
level.
232 Stock–Flow Coherence and Economic Policy

22.5

15.0
Percent of private disposable income

Not lending to
private sector

7.5

0.0

Private financial
–7.5 balance

–15.0
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1970 1975 1980 1985 1990 1995 2000 2005 2010

Figure 10.13 Private financial balance and growth of nonfinancial debt, actual
1970–1999Q1 and projections implied by CBO
Note: Data after 1999Q1 are author’s projections.
Source: Citibase, Flow of Funds, and author’s estimates.

Digression on the external debt and deficit

Should expansion, against the odds, continue in the medium term in the
way foreseen by the Council of Economic Advisers, the consequences for
the United States’s balance of payments and net foreign indebtedness
could be serious. It is often assumed that balance of payments deficits
have a powerful tendency to correct themselves, but this runs contrary to
the experience of many countries (for instance, Denmark and Australia
within the last 20 years) where the accumulation of foreign debt led
eventually to a painful period of retrenchment. There is certainly no
tendency at the present time for the dollar to fall in the way needed to
generate an improvement in net export demand – quite the contrary.
Figure 10.15 shows the scale of the United States’s foreign indebted-
ness implied by the balance of payments projections in Figure 10.12. In
the 1999 Economic Report of the President (ERP), the CEA takes the pos-
sibility of a chronic, rising deficit very calmly. The ERP notes that the
deficit, by virtue of an accounting identity, is always exactly matched,
 Seven Unsustainable Processes 233

2.5

Percent of private disposable income

2.0

1.5

1.0

0.5
Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1 Q1
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Figure 10.14 Private debt, actual 1960–1999Q1 and projection implied by CBO
Note: Data after 1999Q1 are author’s projections.
Source: Citibase, Flow of Funds, and author’s estimates.

30.0

20.0

10.0
Percent of GDP

0.0

–10.0

–20.0

–30.0

–40.0

–50.0
1970 1975 1980 1985 1990 1995 2000 2005 2010

Figure 10.15 U.S. Net foreign assets, actual and projected

Note: Data after 1998 are author’s projections.
Source: Author’s calculations using official benchmark.
234 Stock–Flow Coherence and Economic Policy

one for one, by an inflow of capital, aka net borrowing, from abroad.
The ERP also argues that this borrowing from abroad may not be a bad
thing if it gives rise to profitable investment that raises US productiv-
ity. To support this point, the ERP contains a chart (19–11) that shows
the scale of inward and outward direct investment in recent years. It
also argues that holdings of US equities by foreigners should not ‘count’
as debts.
The ERP is not convincing on either of these points. The figures relating
to net foreign direct investment do not support the notion that this did
anything for US productivity. For one thing, US direct investment abroad
has generally exceeded foreign direct investment in the United States, so
net direct investment has made a negative contribution to the financing of
the current account deficit. For another, foreign direct investments in the
United States have performed poorly, if their profit record is anything to
go by. It is because the rate of profit earned on foreign direct investments
in the United States was so much below that on US direct investments
abroad that it was not until last year that net payments of factor income
across the exchanges finally turned negative, although the United States
became a net debtor in 1989.
And while it is true that equity issued by a corporation is not part of its
indebtedness, US equities held by foreigners have not been issued by the
United States as a country. Equities give rise to payments of factor income
by the United States to foreigners in just the same way as government
bonds do and they can as easily be liquidated.
Figure 10.16 analyses net holdings of overseas assets into direct invest-
ments, private holdings of financial assets, and government holdings of
financial assets. It will be seen that, so far as direct investment is con-
cerned, the United States has remained a creditor, with net assets valued
at current market prices averaging around 4% of GDP in recent years.
All the large changes have been in holdings of financial assets; net pri-
vate holdings fell rapidly, to minus 11% of GDP at the end of 1997, and
government holdings fell to minus 7.5% of GDP.
Figure 10.17 shows payments and receipts of factor income derived
from financial assets and liabilities expressed as a proportion of the rele-
vant stock (lagged one year), and these quasi-interest rates are compared
with the rate on one-year US Treasury bonds. The rate of ‘interest’ on
financial liabilities has consistently exceeded that on assets and is also
in excess of the normal growth rate. Accordingly, if the trade forecasts
are correct, net payments of factor income by the United States will rise
steadily from now on, accelerating the growth in the current account
deficit and the rise in the United States’s net indebtedness. The process
 Seven Unsustainable Processes 235

8.0
Net direct investments
(at market prices)
4.0
Percent of GDP

0.0
Net government assets

–4.0
Net private financial assets

–8.0

–12.0
1983 1985 1987 1989 1991 1993 1995 1997

Figure 10.16 Breakdown of net foreign assets (stocks).

Source: Citibase, Flow of Funds, and author’s projections.

12.0

10.0 Factor income payments

(at percentage of financial liabilities)

8.0
Percent

6.0

Factor income receipts (at

percentage of financial modes)
4.0

US treasury bonds
one-year rate
2.0
1984 1986 1988 1990 1992 1994 1996

Figure 10.17 ‘interest’ rates on foreign assets and liabilities.

Source: Survey of Current Business, and author’s estimates.
236 Stock–Flow Coherence and Economic Policy

6.0
Percent of GDP (comparility management)

3.0
6
5
4
0.0 3

–3.0 2

–6.0

1
–9.0

1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010

Figure 10.18 The private financial balance on six different assumptions

Note: Data after 1998 are author’s projections.
Source: Citibase and author’s projections.

described is clearly unsustainable and will eventually have to be checked,

preferably before an exchange crisis forces the issue.

Alternative scenarios based on different assumptions about private

sector behaviour
To illustrate a range of outcomes, Figures 10.18 and 10.19, which
should be read together, show alternative scenarios based on six dif-
ferent assumptions about private sector behaviour. The numbers from
which the figures have been drawn are taken from simulations of two
econometric models, one describing the US economy, the other describ-
ing production in and trade between the eleven country blocs that taken
together constitute the whole world. Although based on computer mod-
els that cannot be made readily accessible to the reader, it is hoped that
the figures, together with the argument in the text, will carry prima facie
evidence. The heart of the argument is that if the seemingly impossi-
ble rise in indebtedness shown in Figure 10.14 is required to keep the
US economy rising at 2.4% per annum (the minimum needed to keep
unemployment from rising), any slower growth in net lending will cause
a slowdown in output large enough to cause unemployment to rise.
 Seven Unsustainable Processes 237

2.50 1
Percent of private development income
2.25
2
2.00

3
1.75 4
5

1.50
6

1.25

1.00
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010

Figure 10.19 Private debt on six different assumptions

Note: Data after 1998 are author’s projections.
Source: Citibase, Flow of Funds, and author’s projections.

Figures 10.18 and 10.19 show six different possibilities regarding the
future course of the private financial balance and their counterparts
in terms of private indebtedness. No pretense of knowledge is made
regarding the likelihood of any of these outcomes. On the contrary, it
is emphasized that yet other outcomes, not illustrated in the figure, are
perfectly likely to occur and the turning point could come earlier or later.
Despite these great uncertainties the important conclusion remains that
the present stance of fiscal and trade policies will have to be radically
changed at some stage during the first decade of the new millennium.

Implications for the United States

Projection 1, which reproduces (what must be held to be) the CBO’s ver-
sion of future events, has already been dismissed as implausible in view
of the apparently absurd increase in private indebtedness it implies. The
other projections are meant to encompass a fair range of plausible out-
comes based on different assumptions about future levels of indebtedness
and the behaviour of the stock market. Projection 2 bears an approxi-
mate resemblance to the projection recently published by the IMF (World
Economic Outlook 1999), which puts the private sector deficit at about 5%
of disposable income in 2003, and it is for this reason that it is included
238 Stock–Flow Coherence and Economic Policy

Table 10.1 Implications of the six projections for the United States

General Balance of
Average government payments
growth of GDP, Unemployment balance in 2003 in 2003
Projection 1998–2003 rate in 2003 (% of GDP) (% of GDP)

1 2.34 4.8 2.9 −5.8

2 1.51 6.0 1.1 −3.9
3 1.18 6.5 0.4 −3.6
4 0.79 7.3 −0.5 −3.3
5 0.49 8.1 −1.4 −2.9
6 −0.10 9.6 −2.9 −2.8

here. In the author’s opinion the growth of indebtedness implied by the

IMF projection is still implausibly large – and unsustainable in the long
term. At the other extreme, projection 6 is based on the assumption that
there is a 40% break in the stock market in the fourth quarter of 1999
and that this is accompanied by a fall in net lending and a decline in
indebtedness to levels last seen in the mid 1990s.
As already mentioned, every one of projections 2 through 6 implies
an unacceptably low growth rate, taking the average over several years
for the United States. Table 10.1 shows the indicators that are of great-
est interest. Column 2 shows, for each projection, the implied average
growth rate between 1998 and 2003, the purpose being to convey the
character of the whole period rather than to forecast what will happen
in any particular year. Columns 3, 4, and 5 give counterpart numbers
for unemployment, the general government balance, and the balance
of payments. As these numbers are generated mechanically out of the
particular paths assumed to construct the long-run projections, they
should not be interpreted literally as describing what might happen in
the particular year 2003.
The figures speak for themselves. They say that the United States now
runs a serious risk of suffering a prolonged period of stagnation (or
worse), with rising unemployment throughout the next five years and
beyond. The budget surplus could wither even if there were no relax-
ation of fiscal policy. And, the balance of payments looks set to remain
in substantial deficit.

Implications for the rest of the world

To generate the results in the table above, projections from our model
of the US economy were used, in a process of successive approximation,
 Seven Unsustainable Processes 239

Table 10.2 Percentage shortfall of GDP in 2003 compared with base

projection

Projection

Country Bloc 2 3 4 5 6
United States −3.8 −5.5 −7.5 −9.2 −11.6
Western Europe −0.7 −1.0 −1.3 −1.6 −2.0
Japan −0.7 −1.0 −1.3 −1.6 −2.1
South America −2.5 −3.6 −4.8 −5.8 −7.3
Other developed countries −2.3 −3.3 −4.5 −5.5 −6.7
Asia −2.1 −3.0 −4.1 −5.0 −6.3
China −1.8 −2.6 −3.5 −4.3 −5.4
Middle East −2.0 −2.9 −3.8 −4.5 −5.0
Russia −1.0 −1.4 −1.9 −2.3 −2.0
Africa −1.3 −1.9 −2.5 −3.0 −3.7
Eastern Europe −0.4 −0.4 −0.6 −0.8 −1.2
World (except United States) −1.2 −1.8 −2.3 −2.9 −3.6

in conjunction with a model of world trade and production. More pre-

cisely, each projection of developments in the United States incorporates
assumptions about world production (required to generate projections
of US exports) that have been modified by the implied fall in exports to
the United States compared with what otherwise would have happened.
Table 10.2 shows how output in each country bloc might be affected
going from one projection to the next.
The world model from which these estimates are derived is extremely
simple, with ripples generated solely by the foreign trade multiplier
effects (that is, by income reductions caused by falling exports) and by
changes in commodity and oil prices. But the model does have the great
advantage of being comprehensive, in the sense that it encompasses the
entire world and its solutions allow (if crudely) for the interdependence
of world production and trade.
The extent to which foreign countries are affected by recession in
the United States depends on the extent of their openness, in partic-
ular to direct and indirect trade with the United States, and on the
extent to which they are net exporters or importers of raw materials
and energy. Thus Japan is comparatively unaffected partly because, as a
big net importer of raw materials and fuel, it gains substantially from an
improvement in the terms of trade, which boosts real income.
It is noteworthy that by far the greatest impact of a severe recession in
the United States would be experienced in South America, Asia, and
‘other developed countries,’ in particular Canada and Australia. The
240 Stock–Flow Coherence and Economic Policy

effects on Europe are relatively small but large enough, in the worst case,
to add 2 million or more to unemployment.
For all the crudity of the models being used, the figures in the table
sound an alert, to put it moderately, as to the potentially grave effects of
a severe recession in the United States on the rest of the world, much of
which will still be reeling from the blows suffered in 1997 and 1998.

Policy considerations

The main conclusion of this paper is that if, as seems likely, the United
States enters an era of stagnation in the first decade of the new mil-
lennium, it will become necessary both to relax the fiscal stance and to
increase exports relative to imports. According to the models deployed,
there is no great technical difficulty about carrying out such a program
except that it will be difficult to get the timing right. For instance, it
would be quite wrong to relax fiscal policy immediately, just as the credit
boom reaches its peak. As stated in the introduction, this paper does not
argue in favour of fiscal fine-tuning; its central contention is rather that
the whole stance of fiscal policy is wrong in that it is much too restrictive
to be consistent with full employment in the long run. A more formidable
obstacle to the implementation of a wholesale relaxation of fiscal policy
at any stage resides in the fact that this would run slap contrary to the
powerfully entrenched, political culture of the present time.
The logic of this analysis is that, over the coming five to ten years, it will
be necessary not only to bring about a substantial relaxation in the fiscal
stance but also to ensure, by one means or another, that there is a struc-
tural improvement in the United States’s balance of payments. It is not
legitimate to assume that the external deficit will at some stage automati-
cally correct itself; too many countries in the past have found themselves
trapped by exploding overseas indebtedness that had eventually to be
corrected by force majeure for this to be tenable.
There are, in principle, four ways in which the net export demand can
be increased: (1) by depreciating the currency, (2) by deflating the econ-
omy to the point at which imports are reduced to the level of exports, (3)
by getting other countries to expand their economies by fiscal or other
means, and (4) by adopting ‘Article 12 control’ of imports, so called after
Article 12 of the GATT (General Agreement on Tariffs and Trade), which
was creatively adjusted when the World Trade Organization came into
existence specifically to allow nondiscriminatory import controls to pro-
tect a country’s foreign exchange reserves. This list of remedies for the
external deficit does not include protection as commonly understood,
 Seven Unsustainable Processes 241

namely, the selective use of tariffs or other discriminatory measures

to assist particular industries and firms that are suffering from relative
decline. This kind of protectionism is not included because, apart from
other fundamental objections, it would not do the trick. Of the four
alternatives, we rule out the second – progressive deflation and resulting
high unemployment – on moral grounds. Serious difficulties attend the
adoption of any of the remaining three remedies, but none of them can
be ruled out categorically.
While a proper discussion of all these issues would be beyond the scope
of this paper, a final simulation is presented in which the problems that
have been raised are assumed to have been solved. The story is put for-
ward with great diffidence for no one knows better than the author how
little is really known about how all the key relationships – import and
export price and volume elasticities, the elasticity of responses of gov-
ernment payments and receipts, and so forth – will behave in the future.
However, the models can be used to give answers, of a kind, concerning
the possible magnitudes of policy changes that may be required.
The data illustrated in Figure 10.20 were derived by superimposing on
projection 5 whatever fiscal expansion plus (effective) dollar devaluation

6.0
Private financial balance
General government balance
4.0

2.0
Percent of GDP

0.0

–2.0

–4.0 Current balance at payments

–6.0
1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010

Figure 10.20 Goldilocks resuscitated

Note: Data after 1998 are author’s projections.
Source: Citibase, Flow of Funds, and author’s projections.
242 Stock–Flow Coherence and Economic Policy

is necessary to generate the growth of output assumed in the CBO pro-

jections (growth just enough to keep unemployment close to its present
low level) and an improving balance of payments. Specifically, it was
necessary to raise total general government outlays (including transfers
but not interest payments) in stages by about 16% – corresponding to
about $400 billion per annum at current prices – compared with what
the CBO is at present projecting.8 Also necessary was an effective 20%
depreciation of the dollar at the end of 1999, which ‘sticks’ throughout
the rest of the period.
As Figure 10.20 illustrates, these changes, which generate a 2.4%
average growth rate between 2000 and 2008, are compatible with the
reversion of balance of payments to zero by the end of the period
(notwithstanding greatly increased factor income payments abroad). But
they also imply (given that the private financial balance recovers to its
normal level) that the budget of the general government goes into deficit
to roughly the extent that was normal in the 1970s and 1980s. Parallel
simulations of the world model suggest that the net effect of these mea-
sures on output in the rest of the world would be positive, but small; the
positive effect of higher US output on other countries’ exports would
just offset the reduction of their net exports as a result of the dollar
depreciation.

Appendix 1: A private expenditure function

This paper has so far been written as though it were more or less self-evident
that total private expenditure has a systematic and predictable relationship with
total private disposable income and the flow of net lending to the private sector.
A relationship of this kind (once known as the New Cambridge equation) was
presented in Fetherston and Godley (1978) and criticized by Alan Blinder (1978)
on the grounds that the aggregation of consumption with investment did not
make sense in terms of any known theories of consumption and investment taken
individually. Blinder did, however, generously conclude that:

To the credit of the New Cambridge group … the one feature of the model
that Fetherston and Godley clearly label as absolutely essential to New Cam-
bridge is also the one feature that should elicit the greatest interest on this side
of the Atlantic: the unusual specification of aggregate private expenditure. I
rather doubt that the sum of consumption and investment spending can be
explained very well by the sum of disposable income plus retained earnings,
and its lagged value, in the U.S. But, if it can be, American Keynesians will
have to reexamine the prevailing models of consumer and investor behavior.
An empirical study of this question in the U.S. would be most welcome, and
would really decide whether there is anything in New Cambridge that we in
America should import.
 Seven Unsustainable Processes 243

2.50

2.00
Total indebtedness

1.50
Private debt
Percent

1.00

0.50 Government debt

Net foreign wealth

0.00

–0.50
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000

Figure A.1 Selected assets and liabilities as a per cent of private disposable income
Source: Citibase, Flow of Funds, and author’s projections.

What follows is a justification of the aggregation in question, at least for the

limited purpose of underpinning the conclusions drawn in this paper.
Figure A.1 shows government debt, foreign wealth, and (non-financial) private
sector net debt (all expressed as a proportion of private disposable income) since
1960. The net indebtedness of the US general government is now just under 30%
of GDP, compared with 43% for the federal government taken by itself.9 Net
foreign assets, in the absence of official figures for more than a short period, has
been calculated by cumulating the current account balance using the published
figure for 1983 as a benchmark. The figure for private indebtedness was taken
from the Flow of Funds accounts.
It will be seen that total indebtedness, DEBT, taking government, foreign, and
private sectors together, was relatively stable as a per cent of private disposable
income, never far from a 1.7% mean, while movements of its component parts
tended to offset one another. The variance in the ratio of DEBT to income (.027)
was smaller than that of either private debt, DP (.029), or foreign wealth, VF
(.053). The variance of government debt, GD, was only 0.009, but that is to be
expected given that the variance of DEBT was relatively small and that DP and
VF happened to move in opposite directions.
Write the debt identity

DEBT = DP + GD + VF (A.1)

where DP is private debt, GD is public debt, and VF is net foreign wealth. Its first
difference is:

DEBT = (GL − REP) + (G − T ) + CAB (A.2)

244 Stock–Flow Coherence and Economic Policy

where GL is gross lending to the private sector, REP is repayments of private debt,
G is government outlays, T is government receipts, CAB is the current balance of
payments, and  is a first difference operator.
Equation A.2, by simple rearrangement of national income identities, becomes:

DEBT = (YD + GL) − (PX + REP) = M (A.3)

where YD is total private disposable income, PX is total private expenditure (that

is, consumption and investment combined), and M is liquid financial assets. The
terms in the first parentheses on the right-hand side of equation 3 describe the
total cash flow into the private sector each period, and the terms in the second
parentheses describe total outflows. So the change in total debt (DEBT ), taken
across the economy, is defining an accumulation of liquid financial assets (M ) by
the private sector.
If it could be established that the desired stock of the category of assets defined
by this cash flow identity is reasonably stable relative to income, it would follow,
on hydraulic principles alone, that outflows (i.e., total expenditures) would track
inflows one for one with a mean lag equal to the stock/flow norm (Godley and
Cripps 1982, 64–65). The numbers in Figure A.1 are consistent, prima facie, with
the hypothesis that there is a stock/flow norm that is fairly small and stable.
To derive an estimating equation, a norm for the stock/flow ratio was postu-
lated:

M ∗ = α1 · YD (A.4)

Some adjustment process was also postulated:

M = F1 (M ∗ − M−1 ) (A.5)

Where F1 describes a function.

Equations A.1, A.4 and A.5 imply a relationship between inflows, outflows, and
stocks:

PX = F2 (L, YD, M−1 ) (A.5)

where L is net lending (GL – REP) and F2 is another function.

For the estimating equation, PX, YD, and M were deflated using a price index
for private expenditure as a whole (to become px, yd, and m). Real net lend-
ing was disaggregated into its three major component parts – consumer credit
(DP1), other personal borrowing, mainly mortgages (DP2), and business borrow-
ing (DP3). Terms have also been included that describe real stock market (pe /p)
and real (second-hand) house prices (ph /p).10 The inclusion of the stock mar-
ket and house price terms, given that the real stock of financial assets is an
argument, implies that a step rise in stock prices, even if the step is an endur-
ing one, has only a transitory effect on the expenditure flow; a permanent
rise in the (level of the) expenditure flow would require a perpetual increase in
stock prices.
 Seven Unsustainable Processes 245

Table A.1 Estimated long-run coefficients using the ARDL approach

Regressor Coefficient Standard error T-Ratio [Prob]

yd .70610 .048770 14.4780[.000]

DP1 1.0434 .17809 5.8584[.000]
DP2 .19396 .077611 2.4992[.014]
DP3 .31557 .055472 5.6889[.000]
Pe /p 230.0381 35.3642 6.5048[.000]
ph /p 9.9135 1.7325 5.7221[.000]
v−1 .080823 .023421 3.4509[.001]
Constant −672.5665 92.7038 −7.2550[.000]

Notes: ARDL(2,1,0,0,0,0,0) selected based on the Schwarz Bayesian Criterion

Dependent variable is px
113 observations used for estimation from 1969Q2 to 1997Q2

Using quarterly data from the first quarter of 1968 to the third quarter of
1998, an autoregressive distributed lag model (ARDL) of equation 6 was specified,
with up to five quarterly lags in each variable. Model selection criteria were used
to choose the preferred order of lags. This procedure, described in the Microfit
econometric software package, provides a single equation approach to cointegra-
tion analysis. The specification selected has two lags on the dependent variable
and one on DP1, the variable that describes the (deflated) net flow of consumer
credit. From the selected specification the estimated long-run coefficients (shown
in Table A.1) were obtained. The results, written as an error correction model, are
given in Table A.2.11 The full model written out in levels, with all its lagged terms,
is shown in Table A.3.
The formal diagnostics around this equation seem satisfactory. The standard
error is low (0.5%). There is neither significant serial correlation of the residuals
nor heteroscedasticity (notwithstanding that the equation uses levels not logs).
And, fitted up to the first quarter of 1997, it gives excellent forecasts of total private
expenditure during the subsequent six quarters – when private expenditure was
behaving in such a strange way (Table A.4).
No claim is being made regarding any broad theoretical significance for this
equation. In no fundamental sense can either lending flows or asset prices be
treated or thought of as exogenous; and there is a high degree of simultaneity
between many of the variables on both sides of the equation. The central point
in the present context is that as the stock of liquid financial assets does not, as an
empirical matter, fluctuate wildly and is not high relative to the flow of income, it
is acceptable to bypass the specification of (several) consumption and investment
functions as well as the labyrinthine interrelationships between the household
and business sectors, for instance, the distribution of the national income between
profits, proprietors’ income and employment income, the retention of profits, and
the provenance of finance for investment.
246 Stock–Flow Coherence and Economic Policy

Table A.2 Error correction representation for the selected ARDL model

Regressor Coefficient Standard error T-Ratio [Prob]

px−1 .22704 .051476 4.4107[.000]

yd .54410 .054261 10.0274[.000]
DP1 .44503 .073148 6.0840[.014]
DP2 .082732 .033340 2.4815[.015]
DP3 .13460 .028539 4.7166[.000]
pe /p 98.1202 15.3473 6.3933[.000]
ph /p 4.2285 .82000 5.1567[.000]
v−1 .034474 .011109 3.1031[.002]
ecm−1 −.42654 .041758 −10.2146[.000]
constant −286.8757 46.0978 −6.2232[.000]

List of additional variables created:

px = px − px−1
yd = yd − yd−1
DP1 = DP1 − DP1−1
DP2 = DP2 − DP2−1
DP3 = DP3 − DP3−1
pe /p = pe /p − pe−1 /p−1
ph /p = ph /p − ph−1 /p−1
v−1 = v−1 − v−2
constant = constant − constant−1
ecm = px − 0.70610yd − 1.0434DP1 − 0.19396DP2 − 0.31557DP3 −
230.0381pe /p − 9.9135ph /p − 0.080823v−1 + 672.5665constant

R-Squared .81401 R-Bar-Squared .79578

S.E. of Regression 19.6165 F-stat. F(9, 103) 49.6032[.000]
Mean of Dependent 30.7577 S.D. of Dependent Variable 43.4082
Variable
Residual Sum of Squares 39250.2 Equation Log-Likelihood −490.8834
Akaike Info. Criterion −501.8834 Schwarz Bayesian Criterion −516.8840
DW-statistic 2.2098

Notes: ARDL(2,1,0,0,0,0,0) selected based on the Schwarz Bayesian Criterion

Dependent variable is px
113 observations used for estimation from 1969Q2 to 1997Q2

The equation is consistent with the view taken in this paper that aggre-
gate private expenditure responds in a coherent way to aggregate income given
various assumptions about the future course of asset prices and of net lend-
ing to the private sector – both of which may now be close to their cyclical
peak.
 Seven Unsustainable Processes 247

Table A.3 Autoregressive Distributed Lag Estimates ARDL(2,1,0,0,0,0,0) selected

based on the Schwarz Bayesian Criterion

Dependent variable is px
113 observations used for estimation from 1969Q2 to 1997Q2

Regressor Coefficient Standard error T-Ratio [Prob]

px−1 .80051 .068115 11.7522[.000]
px−2 −.22704 .051476 −4.4107[.000]
yd .54410 .054261 10.0274[.000]
yd −1 −.24292 .066562 −3.6495[.000]
DP1 .44503 .073148 6.0840[.014]
DP2 .082732 .033340 2.4815[.015]
DP3 .13460 .028539 4.7166[.000]
pe /p 98.1202 15.3473 6.3933[.000]
ph /p 4.2285 .82000 5.1567[.000]
v−1 .034474 .011109 3.1031[.002]
constant −286.8757 46.0978 −6.2232[.000]

R-Squared .99962 R-Bar-Squared .99959

S.E. of Regression 19.6165 F-stat. F(9, 103) 27169.5[.000]
Mean of Dependent 4110.1 S.D. of Dependent 966.3492
Variable Variable
Residual Sum of Squares 39250.2 Equation −490.8834
Log-Likelihood
Akaike Info. Criterion −501.8834 Schwarz Bayesian −516.8840
Criterion
DW-statistic 2.2098

Diagnostic Tests

Test statistics LM Version F Version

A: Serial correlation CHSQ(4) = 4.7786[.311] F(4, 98) = 1.0818[.370]

B: Functional form CHSQ(1) = .86880[.351] F( 1, 101) = .75255[.378]
C: Normality CHSQ(2) = 2.0566[.358] Not applicable
D: Heteroscedasticity CHSQ(1) = 2.3784[.123] F(1, 111) = 2.3865[.125]

A: Lagrange multiplier test of residual serial correlation

B: Ramsey’s RESET test using the square of the fitted values
C: Based on a test of skewness and kurtosis of residuals
D: Based on the regression of squared residuals on squared fitted values
248 Stock–Flow Coherence and Economic Policy

Table A.4 Dynamic forecasts for the level of px

Based on 113 observations from 1969Q2 to 1997Q2.

ARDL(2, 1, 0, 0, 0, 0, 0) selected using Schwarz Bayesian Criterion.
Dependent variable in the ARDL model is px included with a lag of 2.
List of other regressors in the ARDL model:
yd yd −1 DP1 DP2 DP3
pe /p ph /p v−1 constant

Observation Actual Prediction Error

1997Q3 6157.3 6149.7 7.5873
1997Q4 6217.0 6217.5 −.47760
1998Q1 6371.4 6335.0 36.3867
1998Q2 6432.5 6434.8 −2.3231
1998Q3 6507.1 6523.2 −16.0857

Summary Statistics for Residuals and Forecast Errors

Estimation Period Forecast Period

1969Q2 to 1997Q2 1997Q3 to 1998Q3
Mean .7999E-8p 5.0175
Mean Absolute 15.4662 12.5721
Mean Sum Squares 347.3472 329.1871
Root Mean Sum Squares 18.6373 18.1435

Dynamic forecasts for the change in px

Based on 113 observations from 1969Q2 to 1997Q2.

ARDL(2, 1, 0, 0, 0, 0, 0) selected using Schwarz Bayesian Criterion.
Dependent variable in the ARDL model is px included with a lag of 2.
List of other regressors in the ARDL model:
yd yd −1 DP1 DP2 DP3
pe /p ph /p v−1 constant
Observation Actual Prediction Error
1997Q3 78.5513 70.9640 7.5873
1997Q4 59.7202 67.7851 −8.0649
1998Q1 154.4126 117.5483 36.8643
1998Q2 61.1098 99.8188 −38.7099
1998Q3 74.6196 88.3821 −13.7625

Summary Statistics for Residuals and Forecast Errors

Estimation Period Forecast Period

1969Q2 to 1997Q2 1997Q3 to 1998Q3

Mean .7999E-8 −3.2171

Mean Absolute 15.4662 20.9978
Mean Sum Squares 347.3472 633.8902
Root Mean Sum Squares 18.6373 25.1772
 Seven Unsustainable Processes 249

Appendix 2: Note on the models employed

A ‘stripped down’ quarterly stock/flow model of the U.S economy was used to
derive alternative medium-term scenarios. For all its shortcomings, this model
has the merit of consistency, the accounting being watertight in the sense that
everything comes from somewhere and goes somewhere, while all financial bal-
ances have precise counterparts in changes in stock variables. Table A.5 describes
the accounting structure of the model and shows all variables measured at current
prices. All stock and most flow variables were also deflated to derive conventional
measures of real income, expenditure, and output.
The model describes the processes by which the fiscal operations of the gov-
ernment, the net demand for exports, and the flow of private credit generate
(ex ante) stocks and flows of financial assets for the private sector. The spend-
ing response of the private sector to its inherited asset stocks and its current
flow receipts interacts with tax receipts, imports, and other cyclically sensi-
tive variables to resolve the ex ante dispositions of all the three major sectors
through the solution of a system of (dynamic) simultaneous equations. This
is all good old Keynesian stuff, except that careful track is kept of stock vari-
ables, which not only enter the expenditure function but generate flows of
interest payments by the government as well as flows of factor income across the
exchanges.
This model has only a limited application because it takes so much as exoge-
nous, for instance, interest rates, exchange rates, asset prices, world commodity
prices, the flow of net lending, and the rate of wage inflation. The main objective
on the present occasion is to obtain a quantitative sense of the scale and duration
of the slowdown that will follow when the lending cycle turns down or if there
were a downward adjustment of stock market prices. An equally important objec-
tive is to obtain a sense of the interdependence of the whole stock/flow system;
it emphasizes, in particular, that the size of the budget surplus cannot be sensibly
judged outside the context of what happens to the whole configuration of stocks
and flows.
To the private expenditure function described in Appendix 1 were added
conventional import and export price and volume equations and a simplified
representation of the response of tax receipts and ‘entitlement’ programs to the
business cycle. This latter part of the exercise was carried out in a particularly
crude way, using rules of thumb regarding elasticities of the fiscal system with
regard to changes in real income and inflation, but it at least has the merit that it
approximately reproduces the responses set out in the CBO’s Economic and Budget
Outlook: Fiscal Years 2000–2009 (Congressional Budget Office 1999b).
The output of this model of the United States was used, in a process of successive
approximation, in conjunction with a model of world trade and production. This
world model, which was devised by Francis Cripps in (1979), divides the world
exhaustively into eleven blocs (made up of one or more countries). At its heart
there resides a matrix describing exports and imports of manufactures between
each pair of blocs that is used to derive the share of each bloc’s exports in the
imports of every other bloc. Each bloc’s supply of and demand for and also trade
in energy and raw materials are also recorded, but not on a bilateral basis. The GDP
of each bloc is determined by the sum of its domestic expenditure and balance of
trade. Each bloc’s domestic demand is determined by its real income (that is, real
 Table A.5 Flow matrix describing flow variables of ‘stripped down’ model of U.S. Economy

Sector Income/ Production Financial General Interest pool Foreign 

expenditure Government

Private expenditure −PX +PX 0

Government expenditure +GG −GG 0
on goods
Government expenditure +GS −GS 0
on services
Exports: agriculture +XA −XA 0
Exports: computers +XC −XC 0
Exports: other goods and +XN −XN 0
services

250
Imports: oil −IM O +IMO 0
Imports: computers −IMC +IMC 0
Imports: other goods and −IMN +IMN 0
services
Memo: gross domestic [= GDP]
product
Net indirect taxes −NIT +NIT 0
Total factor income +Y −Y 0
Unemployment benefit +UB −UB 0
Other government +OTG −OTG 0
(domestic) transfers
Direct tax −T +T 0

(continued)
 Table A.5 Continued

Sector Income/ Production Financial General Interest pool Foreign 

expenditure Government

Contributions −EC +EC 0

Private interest payments +INTh −INTh 0
Private transfers abroad −TRpf +TRpf 0

Memo: private disposable [YD]

income
Government interest −INTg +INTg 0
payments
Factor income payments −Ypf +Ypf 0
abroad

251
Factor income received +Yfp −Yfp 0
from abroad
Government transfers −TRgf +TRgf 0
abroad
Net lending to +L −L 0
non-financial private
sector
Sectoral financial balances −M +DF +GD +VF 0
= changes in asset or Change in Change in Change in Net change
liability stocks liquid assets liabilities of government in overseas
= cash flow financial debt = general assets =
surplus sector government current
deficit balance of
payments
252 Stock–Flow Coherence and Economic Policy

output adjusted for the terms of trade). Imports into each bloc are determined by
its output, using an imposed (but estimated) income elasticity of demand; total
exports of manufactures from each bloc are determined by its (projected) share
in each of the other ten bloc’s imports. World demand for energy and raw mate-
rials is brought into equivalence with supply through market-clearing processes
which determine world prices both for energy and raw materials relative to that
of manufactures.
The US and the world models are both extremely transparent and easy to use.
For instance, it is possible to enter a new assumption about the course of US
output exogenously into the world model and compare the results with a previous
solution in about ten seconds. The solution of the world model itself, using a
pentium laptop, takes about one second.
To generate the consistent results described in the main text, the two models
were used in tandem. For instance, the more pessimistic projections of devel-
opments in the US incorporate assumptions about world production (required
to generate projections of US exports) that have been modified, using the world
model, by the implied fall in exports to the United States compared with what
otherwise would have happened.

Acknowledgments
I am grateful to Ken Coutts, Jay Levy, Will Milberg, Jamee Moudud, and Randy
Wray for their help and penetrating comments; also to Bill Martin, with whom
I co-authored a survey covering some of the same ground last December. The
US model was solved using Modler software produced by the Alphametrics
Corporation. Alphametrics also supplied the software to solve the world model.

Notes
1. Define G as government spending, T as tax receipts, and θ = T /Y as an average
tax rate where Y is GDP. The fiscal ratio G/θ is exactly equal to Y when the
budget is in balance (G = T ). When the fiscal ratio exceeds GDP, there is a
deficit (G > T ); and when it is lower, there is a surplus (G < T ). The fiscal ratio
shown in the figures has been adjusted for the business cycle by correcting
relevant components of G and T and adjusted for inflation by appropriate
deflation of both stocks and flows.
2. The ATR is X/μ where X is exports of goods and services plus all transfers
corrected for price changes and μ is the average import propensity corrected
for the business cycle.
3. The CFTR is (G + X)/(θ + μ), with everything adjusted for price changes as
well as for the business cycle.
4. Because, to spell it out, if G + X = T + IM, where IM equals imports, and if
T + IM = (θ + μ) · Y, it must also be the case that Y = (X + G)/(θ + μ).
5. There are small exceptions to this of which the most important is probably
imputed rent. The figures in Figure 10.7 have been purged of capital consump-
tion by the government – a notional item that has recently been included in
the NIPA to allow for the benefits derived from the publicly owned capital
stock.
 Seven Unsustainable Processes 253

6. For a brief description of the model of world trade and production on which
these projections are based and a list of the blocs into which the world is
divided for the model, see pages 16 and 25.
7. The projections are almost identical to those made by Blecker (1999).
8. The same result could have been obtained by cutting taxes on a corresponding
scale.
9. No official estimate of general government debt held by the public seems to
exist. The Organization for Economic Cooperation and Development (OECD)
calculates and publishes its own estimate of general government debt, which
is much higher than the one given here. Perhaps the OECD excludes stocks of
debt arising from the Social Security surplus of state and local governments
although it includes the corresponding surplus of the federal government.
The OECD estimate seems inconsistent with the measured net flow of interest
payments, implying impossibly low interest rates.
10. It has so far proven impossible to locate any measure of second-hand house
prices before the first quarter of 1968.
11. The tables have been reproduced as in the original file, with a change in
notations for variables to keep homogeneity with the rest of the book. In
Table A.2 there must be a mistake which was in the version of the software
used by Godley at the time, since the equation includes the change in the
constant constant, which is obviously zero. This variable should instead be
the intercept in the equation [Eds].

References

Blecker, Robert A. (1999) ‘The Ticking Debt Bomb.’ Briefing Paper (Washington,
D.C.: Economic Policy Institute).
Blinder, A.S. and R.M. Solow (1973) ‘Does Fiscal Policy Matter?’ Journal of Political
Economy 81 (2): 319–337.
Blinder, A.S. (1978) ‘What’s “New” and What’s “Keynesian” in the “New
Cambridge” Keynesianism.’ In K. Brunner and A.H. Meltzer (eds), Public Poli-
cies in Open Economies. Carnegie Rochester Conference, vol. 9 (Amsterdam:
North-Holland).
Brunner, K. and A.H. Meltzer (eds) (1978) Public Policies in Open Economies.
Carnegie Rochester Conference, vol. 9 (Amsterdam: North-Holland).
Christ, C. (1968) ‘A Simple Macroeconomic Model with a Government Budget
Restraint.’ Journal of Political Economy 76 (1): 53–67.
Congressional Budget Office (1999a) An Analysis of the President’s Budgetary Pro-
posals for Fiscal year 2000 (Washington, D.C.: Government Printing Office),
April.
——. (1999b) The Economic and Budget Outlook: Fiscal Years 2000–2009
(Washington, D.C.: Government Printing Office).
——. (1999c) The Economic and Budget Outlook: An Update (Washington, D.C.:
Government Printing Office), July.
Council of Economic Advisers (1999) Economic Report of the President. (Washington,
D.C.: Government Printing Office).
Cripps, T.F. (1979) ‘Appendix A to Chapter 1.’ Cambridge Economic Policy Review
no. 5 (April).
254 Stock–Flow Coherence and Economic Policy

Fetherston, Martin J. and W. Godley (1978) “‘New Cambridge” Macroeconomics

and Global Monetarism.’ In K. Brunner and A.H. Meltzer (eds), Public Policies
in Open Economies. Carnegie Rochester Conference, vol. 9 (Amsterdam: North-
Holland).
Godley, W. and T.F. Cripps (1982) Macroeconomics (Oxford: Fontana and OUP).
International Monetary Fund. (1999) World Economic Outlook. Washington, D.C.
Martin, Bill and W. Godley (1998) ‘America and the World Economy.’ Research
Group Occasional Paper no.3. Phillips & Drew (December).
Office of Management and Budget (1999) Mid-Session Review. Budget of the United
States Government (Washington, D.C.: Government Printing Office).
Organization for Economic Cooperation and Development (1998) OECD Economic
Outlook. Paris.
Ott, D. and A. Ott (1965) ‘Budget Balance and Equilibrium Income.’ Journal of
Finance 20 (1).
Tobin, J. and W. Buiter (1976) ‘Long-Run Effects of Fiscal Monetary Policy on
Aggregate Demand.’ In J. Stein (ed.), Monetarism (Amsterdam: North Holland).
Wray, L. Randall (1998) Understanding Modern Money: The Key to Full Employment
and Price Stability (Aldershot, UK: Elgar).
Wynne Godley – A Bibliography

Papers and books

1964
(with C. Gillion) ‘Measuring National Product.’ National Institute Economic Review
27 (February): 61–67.
(with C. Gillion) ‘Pricing Behaviour in the Engineering Industry.’ National Institute
Economic Review 28 (May): 50–52.
(with J.R. Shepherd) ‘Long-Term Growth and Short-Term Policy.’ National Institute
Economic Review 29 (August): 26–38.
(with D.A. Rowe) ‘Retail and Consumer Prices 1955–1963.’ National Institute
Economic Review 30 (November): 44–51.

1965
(with W.A.B. Hopkin) ‘An Analysis of Tax Changes.’ National Institute Economic
Review 32 (May): 33–42.
(with J.R. Shepherd) ‘Forecasting Imports.’ National Institute Economic Review 33
(August): 35–42.
(with C. Gillion) ‘Pricing Behaviour in Manufacturing Industry.’ National Institute
Economic Review 33 (August): 43–47.

1972
(with William D. Nordhaus) ‘Pricing in the Trade Cycle.’ Economic Journal 82 (327)
(September): 853–882.
(with J. Rhodes) ‘The Rate Support Grant System.’ Proceedings of a Conference on
Local Government Finance, Institute of Fiscal Studies, Publication n.10.

1973
(with T. Francis Cripps) ‘Balance of Payments and Demand Management.’ London
and Cambridge Economic Bulletin, n.82 (January).

1974
‘Demand, Inflation and Economic Policy.’ London and Cambridge Economic Bulletin,
n.84 (January).
(with Francis Cripps) ‘Budget Deficit and Demand Management.’ London and
Cambridge Economic Bulletin, n.84 (January).
‘The Concept of a Par Economy in Medium-Term Analysis.’ In G.D.N. Worswick
and F.T. Blackbaby (eds.) The Medium Term: Models of the British Economy
(London: Heinemann).

255
256 Wynne Godley – A Bibliography

1975
(with A. Wood) ‘Profits and Stock Appreciation.’ Economic Policy Review, n.1
(February).

1976
‘Costs, Prices and Demand in the Short Run.’ In M.J.C. Surrey (ed.) Macroeconomic
Themes, Edited Readings in Macroeconomics with Commentaries (Oxford: Oxford
University Press), pp.306–309.
(with T. Francis Cripps) ‘A Formal Analysis of the Cambridge Economic Policy
Group Model.’ Economica 43(172) (November): 335–348.
‘The Strategic Problems of Economic Policy.’ Economic Policy Review, n.2 (March):
1–19.
(with T. Francis Cripps; Martin J. Fetherston) ‘What Is Left of New Cambridge?.’
Chapter 6, Economic Policy Review, n.2 (March): 46–49.
‘The Measurement and Control of Public Expenditure.’ Chapter 8, Economic Policy
Review, n.2 (March): 58–63.

1977
‘Inflation in the United Kingdom.’ In Krause, Lawrence B. and Salant, Walter
S. (eds.) Worldwide Inflation: Theory and Recent Experience (Washington D.C.,
Brookings Institution).
(with R. May) ‘The Macroeconomic Implications Of Devaluation And Import
Restriction”, Economic Policy Review, n.3 (March), chapter 2: 32–42.
(with A. McFarquhar; D. Silvey) ‘The Cost of Food and Britain’s Membership of
The EEC.’ Economic Policy Review, n.3 (March), chapter 3: 43–46.
(with Christopher Taylor) ‘Measuring the Effect of Public Expenditure.’ In
M. Posner (ed.) Public Expenditure (Cambridge: Cambridge University Press).

1978
(with Martin J. Fetherstone) ‘‹New Cambridge› Macroeconomics and Global Mon-
etarism: Some Issues in the Conduct of UK Economic Policy.’ In K. Brunner
and A.H. Meltzer (eds) Public Policies in Open Economies, Carnegie-Rochester
Conference Series on Public Policy, vol. 9 (1): 33–65.
(with T. Francis Cripps) ‘Control of Imports as a Means to Full Employment and
the Expansion of World Trade: the UK’S Case.’ Cambridge Journal of Economics,
2 (3) (September) pp.327–334.
(with Ken J. Coutts; William D. Nordhaus) Industrial Pricing in the United Kingdom
(Cambridge: Cambridge University Press).
(with T. Francis Cripps; Martin J. Fetherstone) ‘Simulations with the CEPG Model.’
In M. Posner (ed.) Demand Management (London: Heinemann).
(with R. Bacon; A. McFarquhar) ‘The Direct Costs to Britain of Belonging To the
EEC.’ Economic Policy Review, n. 4 (March), chapter 5: 44–49.

1979
‘Britain’s Chronic Recession – Can Anything Be Done?.’ In W. Beckerman (ed.)
Slow Growth in Britain (Oxford: Clarendon Press): 226–235.
 Wynne Godley – A Bibliography 257

‘The System of Financial Transfers in the EEC’ in Whitby, M. (1979) The Net Cost
and Benefit of EEC Membership: a Workshop Report, (Ashford Kent: Centre for
European Agricultural Studies, Wye College), Seminar Paper 7: 20–33.

1981
‘Monetarism in Three Countries: United Kingdom.’ In D. Crane (ed.), Beyond the
Monetarists: Post-Keynesian Alternatives to Rampant Inflation, Low Growth and
High Unemployment (Toronto: James Lorimer): 36–41.

1983
‘Keynes and the Management of Real National Income and Expenditure.’ In
D. Worswick and J. Trevithick (eds.) Keynes and the Modern World (Cambridge:
Cambridge University Press): 135–156.
(with T. Francis Cripps) Macroeconomics (London: Fontana).
(with Michael Anyadike-Danes) ‘Nominal Income Determination, Financial
Assets and Liabilities and Fiscal Policy.’ Brazilian Review of Econometrics 3(2):
105–130.

1984
‘Confusion in Economic Theory and Policy—Is There a Way Out?.’ In J. Corn-
wall (ed.) After Stagflation: Alternatives to Economic Decline (Oxford: Basil
Blackwell):63–85.

1985
(with Ken J. Coutts; Graham D. Gudgin) ‘Inflation Accounting of Whole Economic
System.’ Studies in Banking and Finance, vol. 9(2) (June):93–114.
‘Unemployment in Europe: the Strategic Problem of the Mid Eighties.’ In Neuer
Protektionismus in der Weltwirtschaft und EG-Handelspolitik: Jahreskolloquium:
99–108.

1987
(with Michael Anyadike-Danes) ‘A Stock Adjustment Model of Income Determi-
nation with Inside Money and Private Debt.’ In M. De Cecco and J.P. Fitoussi
(eds.) Monetary Theory and Economic Institutions (London: Macmillan):95–120.
(with Nicos M. Christodoulakis) ‘Macroeconomic Consequences of Alternative
Trade Policy Options.’ Journal of Policy Modeling 9(3) (Fall): 405–436.

1988
‘The Sensibility of Contemporary Institutions.’ Theology 91(740) (March):89–94.
‘Manufacturing and the Future of the British Economy’ In T. Baker and P. Dunne
(eds.) The British Economy After Oil: Manufacturing or Services? (London, New
York, and Sydney: Croom Helm, in association with Methuen): 5–14.

1989
(with Gennaro Zezza) ‘Foreign Debt, Foreign Trade and Living Conditions, with
Special Reference to Denmark.’ Nationalokonomisk Tidsskrift 127(2): 229–235.
258 Wynne Godley – A Bibliography

(with Michael Anyadike-Danes) ‘REal Wages and Employment: A Sceptical View

of Some Recent Empirical Work”, The Manchester School of Economic and Social
Studies 57(2) (June): 172–187.
(with Michael Anyadike-Danes) ‘Real Wages and Employment: Response to
Nickell’s Comment.’ The Manchester School of Economic and Social Studies 57(3)
(September): 285.
‘The British Economy during the Thatcher Era.’ Economics 25(108) (Winter):
159–162.
(with Ken J. Coutts) ‘The British Economy Under Mrs Thatcher.’ The Political
Quarterly 60(2), (April-June):137–151.

1990
(with Ken J. Coutts; Robert Rowthorn; Gennaro Zezza) ‘Britain’s Economic Prob-
lems and Policies in the 1990s.’ Economic Study n.6, London, Institute for
Public Policy Research (IPPR).
(with Ken J. Coutts) ‘Prosperity and Foreign Trade in the 1990s: Britain’s Strategic
Problem.’ Oxford Review of Economic Policy 6(3) (Autumn): 82–92.
‘The British Economy under Mrs. Thatcher: A Rejoinder.’ The Political Quarterly
61(1) (January): 101–102.
‘When the party had to stop.’ Professional Observer, (May): 32–37.

1992
(with Gennaro Zezza) ‘A Simple Real Stock Flow Model Illustrated with the Danish
Economy.’ In H. Brink (ed.) Themes in Modern Macroeconomics, (London and
Basingstoke: Macmillan): 140–179.
‘Godley, Wynne (born 1926).’ In P. Arestis and M. Sawyer (eds.), A Biographical
Dictionary of Dissenting Economists (Aldershot: Edward Elgar): 193–201.
‘Britain and the Danger of the EMU.’ Samfundsokonomen, n.7 (November).
(with Ken J. Coutts) Does Britain’s Balance of Payments Matter Any More?.’ In
J. Michie (ed.) The Economic Legacy: 1979–1992, (London: Academic Press):
60–67.
(with Ken J. Coutts; J.G. Palma) ‘The British Economy under Mrs. Thatcher.’
Economic Bulletin for Latin America.
The Godley Papers: Economic Problems and Policies in the 1980s and 90s, (London:
New Statesman and Society).

1993
‘Time, Increasing Returns and Institutions in Macroeconomics.’ In S. Biasco, A.
Roncaglia and M. Salvati (eds.) Market and Institutions in Economic Development:
Essays in Honour of Paolo Sylos Labini, (New York: St. Martins Press): 59–82.

1994
(with William Milberg) ‘US Trade Deficits: The Recovery’s Dark Side?’ Challenge
37(6) (June): 40–47.
(with Bob Rowthorn) ‘Appendix; The Dynamics of Public Sector Deficit and Debt.’
In J. Michie and J. Grieve Smith (eds), Unemployment in Europe (London:
Academic Press): 199–206.
 Wynne Godley – A Bibliography 259

1995
‘The U.S. Balance of Payments, International Indebtedness, and Economic Policy.’
Levy Institute Public Policy Brief, n.23.

1996
‘America’s Unaccountable Admiration for Mrs Thatcher’s economics.’ In J. Eatwell
(ed.), Global Unemployment: Loss of Jobs in the ‘90s (Armonk, N.Y.: M.E. Sharpe):
21–40.

1997
(with George McCarthy) ‘The Boskin Commission’s Trillion-Dollar Fantasy.’
Challenge 40(3) (May-June): 14–21.
(with George McCarthy) ‘The Boskin Commission’s Trillion-Dollar Fantasy: Erra-
tum.’ Challenge 40(4) (July-August, p. 108.
‘The Hole in the Treaty.’ In P. Gowan and P. Anderson (eds.), The Question of Europe
(London and New York: Verso Books): 173–177.
‘The United Kingdom and the Community Budget.’ In R.T. Griffiths (ed.), The
Economic Development of the EEC: Economic Development of Modern Europe Since
1870, vol. 12 (Edward Elgar: Cheltenham): 427–441.

1998
(with George McCarthy) ‘Fiscal Policy Will Matter..’ Challenge 41 (1) (January-
February): 38–54.
‘Using Figures to Guide Macroeconomic Policy.’ In I. Begg and S.G.B. Henry (eds),
Applied Economics and Public Policy (Cambridge: Cambridge University Press):
258–263.

1999
(with Bill Martin) ‘America’s New Era.’ Economic Outlook, 24 (1) (October):14–19.
(with L. Randall Wray) ‘Can Goldilocks Survive?.’ Levy Institute Policy Notes,
1999/4.
‘How negative can US saving get?.’ Levy Institute Policy Notes, 1999/1.
‘Money and Credit in a Keynesian Model of Income Determination.’ Cambridge
Journal of Economics 23(4) (July): 393–411.
‘Seven Unsustainable Processes: Medium Term Prospects and Policies for the US
and the World.’ Levy Institute Strategic Analysis, January.

2000
‘Drowning in Debt.’ Levy Institute Policy Notes, 2000/6.
‘Interim Report: Notes on the US Trade and Balance of Payments Deficits.’ Levy
Institute Strategic Analysis, January.
(with L. Randall Wray) ‘Is Goldilocks Doomed?.’ Journal of Economic Issues 34(1)
March: 201–206.
260 Wynne Godley – A Bibliography

2001
(with Alex Izurieta) “As the Implosion Begins...? Prospects and Policies for the US
Economy: A Strategic View”, Levy Institute Strategic Analysis, July.
(with Alex Izurieta) ‘”As the Implosion Begins...?”: A Rejoinder to Goldman Sach’s
J. Hatzius’ “The Un-Godley Private Sector Deficit” in US Economic Analyst (27
July)..’ Levy Institute Strategic Analysis, August.
‘Fiscal Policy to the Rescue.’ Levy Institute Policy Notes, 2001/1.
‘The Developing Recession in the United States.’ Banca Nazionale del Lavoro
Quarterly Review 54(219) (December) pp. 417–425.
(with Alex Izurieta) ‘The Developing US Recession and Guidelines For Policy.’ Levy
Institute Strategic Analysis, October.
(with Marc Lavoie) ‘Kaleckian Models of Growth in a Coherent Stock-Flow Mon-
etary Framework: A Kaldorian View.’ Journal of Post Keynesian Economics 24(2)
(Winter), 2001–2002: 277–311.
‘Godley, Wynne (born 1926).’ In P. Arestis and M. Sawyer (eds.), A Biographi-
cal Dictionary of Dissenting Economists, Second Edition (Cheltenham: Edward
Elgar).

2002
(with Anwar Shaikh) ‘An Important Inconsistency at the Heart of the Standard
Macroeconomic Model.’ Journal of Post Keynesian Economics 24(3) (Spring):
423–441.
‘Kick-Start Strategy Fails to Fire Sputtering Us Economic Motor.’ Levy Institute Policy
Notes, 2002/1.
(with Alex Izurieta) ‘Strategic Prospects and Policies for the US economy.’ Levy
Institute Strategic Analysis, April.
(with Alex Izurieta) ‘The Case for a Severe Recession.’ Challenge 45(2) (April):
27–51.

2003
(with Alex Izurieta) ‘Coasting on the Lending Bubble. Both in the UK and the US.’
Cambridge Endowment for Research in Finance Strategic Analysis.
“The US Economy: A Changing Strategic Predicament”, Levy Institute Strategic
Analysis, March.
‘Sauver Masud Kahn.’ Revue française de psychanalyse 67(3): 1015–1028.

2004
‘Commentary to: Anne-Marie Sandler, "The Case of Masud Khan: Institutional
Responses to Boundary Violations".’ International Journal of Psycho-Analysis
85(1): 27–44.
(with Dimitri B. Papadimitriou; Anwar M. Shaikh; Claudio H. Dos Santos; Gen-
naro Zezza) ‘Is Deficit Financed Growth Limited? Policies and Prospects in an
Election Year”, Levy Institute Strategic Analysis, April.
(with Alex Izurieta; Gennaro Zezza) ‘Prospects and Policies for the US economy:
Why Net Exports Must Now Be the Motor for US Growth.’ Levy Institute
Strategic Analysis, August.
‘Weaving Cloth from Graziani’s Thread: Endogenous Money in a Simple (but
Complete) Keynesian Model.’ In R. Arena and N. Salvadori (eds.), Money,
 Wynne Godley – A Bibliography 261

Credit and the Role of the State: Essays in Honour of Augusto Graziani (Aldershot:
Ashgate): 127–135.
‘Corrigenda: Money and Credit in a Keynesian Model of Income Determination.’
Cambridge Journal of Economics 28(3) (May): 469.
(with Alex Izurieta) ‘The US Economy: Weaknesses of the "Strong" Recovery.’ Banca
Nazionale del Lavoro Quarterly Review 57(229) (June): 131–139.
(with Alex Izurieta) ‘L’economia statunitense: debolezza della “forte” ripresa.’
Moneta e Credito 57(226) (June): 151–160.
(with Alex Izurieta) ‘Fragilités de la reprise économique.’ Problèmes économiques, #
2861.

2005
‘Commentaire à: Anne-Marie Sandler, Réponses institutionnelles aux transgres-
sions: le cas de Masud Khan.’ L’Année Psychanalytique Internationale:15–31.
‘Imbalances Looking for a Policy.’ Levy Institute Policy Notes, 2005/4.
‘Some Unpleasant American Arithmetic.’ Levy Institute Policy Notes, 2005/5.
(with Dimitri B. Papadimitriou; Claudio H. Dos Santos; Gennaro Zezza) ‘The
United States and Her Creditors: Can the Symbiosis Last?”, Levy Institute
Strategic Analysis, September.
(with Marc Lavoie) ‘Comprehensive Accounting in Simple Open Economy
Macroeconomics with Endogenous Sterilization or Flexible Exchange Rates”,
Journal of Post Keynesian Economics 28(2) (Winter), 2005–2006: 241–276

2006
(with Gennaro Zezza) “Debt and Lending: A Cri de Coeur”, Levy Institute Policy
Notes, 2006/4.
(with Marc Lavoie) “Features of a Realistic Banking System within a Post-Keynesian
Stock-Flow Consistent Model”, in M. Setterfield (ed.), Complexity, Endoge-
nous Money and Macroeconomic Theory: Essays in Honour of Basil J. Moore
(Cheltenham: Edward Elgar): 251–268.

2007
(with Marc Lavoie) Monetary Economics: An Integrated Approach to Credit, Money,
Income, Production and Wealth (Basingstoke: Palgrave Macmillan).
(with Marc Lavoie) ‘A Simple Model of Three Economies with Two Currencies:
The Eurozone and the USA.’ Cambridge Journal of Economics 31(1) (January):
1–23.
(with Dimitri B. Papadimitriou; Greg Hannsgen; Gennaro Zezza) ‘The US Econ-
omy: Is There a Way Out of the Woods?.’ Levy Institute Strategic Analysis,
November.
(with Dimitri B. Papadimitriou; Gennaro Zezza) ‘The US Economy: What’s Next?.’
Levy Institute Strategic Analysis, April.

2008
(with Dimitri B. Papadimitriou; Gennaro Zezza) ‘Prospects for the United States
and the World: A Crisis That Conventional Remedies Cannot Resolve.’ Levy
Institute Strategic Analysis, December.
262 Wynne Godley – A Bibliography

2009
‘The Developing Recession in the United States.’ PLS Quarterly Review 62(248–251):
87–95.
(with Alex Izurieta) ‘The US Economy: Weaknesses of the "Strong" Recovery.’ PSL
Quarterly Review 62(248–251): 97–105.
(with Alex Izurieta) ‘L’economia statunitense: debolezza della “forte” ripresa.’
Moneta e Credito 62(245–248): 91–100.

Working papers
(with Christopher Taylor) ‘Public Spending and Private Demand.’ Economics
Reprint No. 339, 1971.
(with T. Francis Cripps) ‘Local Government Finance and Its Reform: A Critique
of the Layfield Committee’s Report.’ Working paper, Department of Applied
Economics, 1976.
(with T. Francis Cripps) ‘The Planning of Telecommunications in the United
Kingdom.’ Working paper, 1978.
(with Ken J. Coutts) ‘Introduction to a Synthesis of Macroeconomic Theory Based
on Tobin’s Nobel Lecture.’ University of Cambridge Department of Applied
Economics working paper, 1984.
(with Ken J. Coutts) ‘Outline for a Reconstructed Basis for Macroeconomic Theory.’
University of Cambridge Department of Applied Economics working paper,
1984.
(with Nicos M. Christodoulakis) ‘A Dynamic Model for the Analysis of Trade Pol-
icy Options.’ University of Cambridge Department of Applied Economics
working paper, 1986.
(with Ken J. Coutts; Robert Rowthorn; Terry S. Ward) ‘The British Economy: Recent
History and Medium Term Prospects.’ Working paper, 1986.
(with Gennaro Zezza) ‘A Simple Real Stock Flow Monetary Model of the Italian
Economy.’ Working paper, Department of Applied Economics, 1986.
(with Ken J. Coutts; Juan Carlos Moreno-Brid) ‘Industrial Pricing in UK Manufac-
turing Industry under Conditions of "Stagflation".’ University of Cambridge
Department of Applied Economics working paper n. 8781, May 1987.
(with Michael Anyadike-Danes; Ken J. Coutts) ‘IS-LM and Real Stock Flow Models:
A Prelude to Applied Macroeconomic Modelling.’ Working paper, 1987.
(with Gennaro Zezza) ‘A Simple Real Stock Flow Model Illustrated with the Dan-
ish Economy.’ University of Cambridge Department of Applied Economics
working paper n. 8901, February 1989.
‘Time, Increasing Returns and Institutions in Macroeconomics.’ University of
Cambridge Department of Applied Economics working paper n. 9023, May
1990.
‘Britain and the Danger of EMU’ and ‘A Macro View of the Danish Economy.’
University of Cambridge Department of Applied Economics working paper
n. 9111, April 1991.
‘US Foreign Trade, the Budget Deficit and Strategic Policy Problems: A Background
Brief”, Levy Institute Working paper, n. 138, 1995.
‘A Simple Model of the Whole World with Free Trade, Free Capital Movements,
and Floating Exchange Rates.’ Unpublished manuscript, Levy Institute of
Economics, 1996.
 Wynne Godley – A Bibliography 263

‘Money, Finance and National Income Determination: An Integrated Approach.’

Levy Institute Working papers, n. 167, June 1996.
‘Macroeconomics without Equilibrium or Disequilibrium.’ Levy Institute Working
papers, n. 205, August 1997.
(with Bill Martin) ‘America and the World Economy.’ Research Group Occasional
Paper no.3, Phillips & Drew, December, 1998.
(with Anwar Shaikh) ‘An Important Inconsistency at the Heart of the Standard
Macroeconomic Model.’ Levy Institute Working papers n. 326, May 1998.
‘Open Economy Macroeconomics Using Models of Closed Systems.’ Levy Institute
Working papers, n. 281, August 1999. 1
(with Bill Martin)’America’s New Era.’ Phillips & Drew Research Group Occasional
Paper NO. 7, September 1999.
(with Marc Lavoie) ‘Kaleckian Models of Growth in a Stock-Flow Monetary Frame-
work: A Neo-Kaldorian Model.’ Levy Institute Working papers n. 302, June
2000.
(with Alex zurieta) ‘Strategic Prospects for the US economy: A New Dilemma.’
Cambridge Endowment for Research in Finance Working Paper Series, No 4,
University of Cambridge, Cambridge, November 2002.
(with Alex Izurieta) ‘Balances, Imbalances and Fiscal Targets. A New Cambridge
View.’ working paper, Cambridge Endowment for Research in Finance,
University of Cambridge, 2003.
(with Marc Lavoie) ‘Two-Country Stock-Flow-Consistent Macroeconomics Using a
Closed Model within a Dollar Exchange Regime.’ Centre for Financial Analysis
and Policy Working paper n. 10, November 2003.
(with Marc Lavoie) ‘Features of a Realistic Banking System within a Post-Keynesian
Stock-Flow Consistent Model.’ Centre for Financial Analysis and Policy
Working paper n. 12, 2004.
(with Marc Lavoie) ‘Simple Open Economy Macro with Comprehensive Account-
ing a Radical Alternative to the Mundell Fleming Model.’ Centre for Financial
Analysis and Policy Working paper n. 15, April 2004.
‘Towards a Reconstruction of Macroeconomics Using a Stock Flow Consistent
(SFC) Model.’ Centre for Financial Analysis and Policy Working paper n. 16,
May 2004.
(with Marc Lavoie) ‘Simple Open Economy Macro with Comprehensive Account-
ing: A Two Country Model.’ Cambridge Endowment for Research in Finance
Working paper n. 20, February 2005.
‘Prolegomena for a Sensible Macroeconomics.’ Paper prepared for the conference
The Keynesian Legacy in Macroeconomic Modelling, Università degli Studi di
Cassino, Italy, September 2005. http://ius.unicas.it/mc2005/papers/godley.pdf
(with Marc Lavoie) ‘Prolegomena to Realistic Monetary Macroeconomics: A The-
ory of Intelligible Sequences.’ Levy Institute Working paper n. 441, February
2006.

Memoranda
‘Measurement, Forecasting and Control of Public Expenditure.’ Memorandum to
Commons Expenditure Committee, 1970/71 (3rd Report), 1970.
‘Implied Price Relationships, Cmnd. 4234, Cmnd. 4578 and the National Income
and Expenditure Account.’ memorandum to Commons Expenditure Committee,
1970/71, 1971.
264 Wynne Godley – A Bibliography

‘Measurement, Forecasting and Control of Public Expenditure.’ Memorandum to

the Select Committee on Expenditure of the House of Commons (3rd Report), 1971.
‘Note on the Treasury Memorandum "Public Expenditure and Demand in Real
Resources", 1971–1972’ (7th Report), 1971.
(with B. Stafford) ‘Notes on Accounting Conventions with Particular Reference to
the Treatment of Receipts.’ Submitted to Commons Expenditure Committee,
1970/71 (3rd Report), 1971
‘Notes on "Public Expenditure to 1976–77" (Cmnd. 5178).’ Memorandum to
Commons Expenditure Committee, 1972/73, 1972.
‘Public Expenditure and Demand on Real Resources.’ Memorandum to the Select
Committee on Expenditure of the House of Commons (7th Report), 1972.
‘Public Expenditure and Economic Management.’ Memorandum to the Select
Committee on Expenditure of the House of Commons (7th Report), 1972.
‘The Need for Further Information about Public Expenditure and What It Buys.’
Memorandum to Expenditure Committee, 1971/1972 (8th Report), 1972.
‘The Supply and Disposition of Real Resources.’ Note submitted to Commons
Expenditure Committee, 1972/73 (11th Report), 1972.
‘The Measurement and Control of Public Expenditure.’ Memorandum to Expenditure
Committee, November, 1973.
‘Public Expenditure and Inflation.’ Memorandum to Expenditure Committee, (9th
Report), Session 1974, H.M.S.O., 1974.
(with T.Francis Cripps; Martin J. Fetherston) ‘Public Expenditure and the Manage-
ment of the Economy.’ Memorandum to Expenditure Committee, (9th Report),
Session 1974, H.M.S.O., 1974.
‘Reflections on the Control of Local Government Expenditure and Its Finance.’
Evidence submitted to the Layfield Committee, January, 1975.
‘Reflections on the Control of Local Government Expenditure and Its Financing:
Evidence Given to the Committee of Inquiry into Local Government Finance.’
Department of Applied Economics, University of Cambridge, 1975.
(with Maurice FitzGerald Scott) ‘The Arguments for and Against Protectionism.’
Bank of England, Paper presented to the Panel of Academic Consultants, no. 10,
1980.
‘The Sensibility of Contemporary Institutions.’ Sermon before the University,
Kings College Chapel, 31 May 1987.
‘Success in International Trade as the Key to Sustained Growth.’ London, TSB
Group, 1993.
‘The Panel of Independent Forecasters February 1993 Report (Submission by
W.Godley).’ London, HM Treasury, 1993.
‘The Panel of Independent Forecasters July 1993 Report (Submission by W.Godley).’
London, HM Treasury, 1993.
‘The Panel of Independent Forecasters October 1993 Report (Submission by
W.Godley).’ London, HM Treasury, 1993.
‘The Panel of Independent Forecasters February 1994 Report (Submission by
W.Godley).’ London, HM Treasury, 1994.
‘The Panel of Independent Forecasters May 1994 Report (Submission by W.Godley).’
London, HM Treasury, 1994.
‘The Panel of Independent Forecasters November 1994 Report (Submission by
W.Godley).’ London, HM Treasury, 1994.
 Wynne Godley – A Bibliography 265

‘The Panel of Independent Forecasters May 1995 Report (Submission by W.Godley).’

London, HM Treasury, 1995.

Articles in Magazines and Newspapers, Letters to newspapers

‘The Barber “Package” under the Microscope.’ The Financial Times, 3 November
1970.
(with Christopher Taylor) ‘Heavier Tax Burden and Reduced Consumption.’ The
Times, 22 February 1971.
(with Christopher Taylor) ‘Four Year Forecast of Public Spending.’ The Times,
8 December 1971.
(with Francis Cripps) ‘Effects of the Cut in SET.’ The Times, 5 April 1971.
(with Francis Cripps) ‘Can We Achieve a Balance in Our Resources.’ The Times,
9 December 1971.
(with Francis Cripps) ‘Need for Unconventional Methods.’ The Times, 9 January
1973.
(with Francis Cripps) ‘Why the Government’s Economic Strategy Is a Dangerous
Gamble.’ The Times, 5 September 1973
(with Francis Cripps) ‘Drastic Steps to Restore Balance.’ The Times, 6 September
1973.
‘Revamping Rate Support.’ The Guardian, 3 December 1973.
(with Francis Cripps) ‘Why Britain Needs a Fresh Set of Principles to Manage the
Economy.’ The Times, 22 January 1974.
(with Francis Cripps) ‘Payments Deficit: The Strategic Options.’ The Times,
23 January 1974.
(with Francis Cripps) ‘Budget Deficit and Balance of Payments.’ The Times, Letters
to the Editor, 5 February 1974
‘A Budget to Forestall Recession.’ The Times, Letters to the Editor, 19 March 1974.
‘Policy Choice.’ The Times, Letters to the Editor, 29 March 1974.
‘The Choices Facing Mr Healy.’ The Guardian, 22 July 1974.
‘The Real State of Our Economy.’ The Sunday Times, 6 October 1974.
(with Adrian Wood) ‘Uses and Abuses of Stock Appreciation.’ The Times,
12 November 1974.
(As part of the Cambridge Economic Policy Group’) ‘Case for Import Controls.’
The Guardian, 17 February 1975.
‘Fallacy at the Core?.’ The Guardian, 25 February 1975.
(with Francis Cripps) ‘Profits, Stock Appreciation and the Sandilands Report.’ The
Times, 1 October 1975.
‘The Case for Import Controls.’ The Sunday Times, 28 March 1976.
‘The Money Supply and Inflation.’ The Times, Letters to the Editor, 14 July 1976.
‘Programme for Economic Stability.’ The Times, Letters to the Editor, 27 September
1976.
‘What Britain Needs Is Growth – But It Must Be the Right Kind.’ The Times,
1 November 1976.
‘The Economic Crisis: The Effectiveness of Devaluation.’ The Times, Letters to the
Editor, 6 November 1976.
‘Protection Only Way for Short-Term Cut in Jobless?.’ The Times, 10 November
1976.
‘Analysing the 1976 Devaluation.’ The Times, 16 November 1976.
266 Wynne Godley – A Bibliography

‘Public Spending Cuts.’ The Times, Letters to the Editor, 25 November 1976.
(with Francis Cripps) ‘Why the Chancellor Should Be Thinking of a £2 Billion Cut
In Taxes.’ The Times, 21 March 1977.
‘Money Supply and Inflation.’ The Times, Letters to the Editor, 18 April 1977.
(with Alister McFarquhar) ‘EEC Membership and Food Prices.’ The Times, Letters
to the Editor, 31 May 1977.
(with Alister McFarquhar) ‘EEC and Agriculture.’ The Times, Letters to the Editor,
13 June 1977.
(with Francis Cripps) ‘Towards a Fairer System of Local Authority Finance.’ The
Times, 4 July 1977.
‘Large Scale Devaluation Is Not the Answer.’ The Times, 18 July 1977.
‘Size of Budget Deficit.’ The Times, Letters to the Editor, 10 July 1977.
(with Richard Bacon) ‘Why the EEC Must Be Set on a Better Course.’ The Guardian,
Letters to the Editor, 10 February 1978.
(with Richard Bacon) ‘How to Count the Cost of British Membership of the EEC.’
The Guardian, Letters to the Editor, 16 February 1978.
(with Francis Cripps) ‘Expanding Out of Rising Prices.’ The Guardian, 29 March
1979.
‘The CAP and the EEC budget.’ New Society, 26 April 1979.
‘PSBR and Money Supply Targets.’ The Banker, September 1979.
‘A Response to Wage Demands Built on False Assumptions.’ The Guardian,
2 October 1979.
‘Problems of the CAP.’ The Financial Times, Letters to the Editor, 1 November 1979.
(with Richard Bacon) ‘Paying for the CAP.’ The Financial Times, 12 November 1979.
‘Wynne Godley Calls for General Import Controls.’ The London Review of Books,
24 January 1980.
‘Cost to Britain of Farm Proposals.’ The Times, Letters to the Editor, 21 February
1980.
(with Francis Cripps) ‘Only a U-Turn Will Steer Britain Clear.’ The Guardian, 24
March 1980.
(with Francis Cripps) ‘The Economic Outlook from Cambridge.’ The Financial
Times, Letters to the Editor, 14 April 1980.
‘If Britain Left the E.E.C..’ The Times, Letters to the Editor.’ 25 April 1980.
‘Testing Time for Monetarism.’ The Times, Letters to the Editor, 24 May 1980.
‘The Siege Has Begun.’ The Observer, 10 August 1980.
‘The Steep Drop at the End of the Tory Tunnel.’ The Guardian, 6 October 1980.
‘How Far Will the Government Let Unemployment Go?.’ The Times, 22 October
1980.
‘Causes of the Recession.’ The Financial Times, Letters to the Editor, 4 November
1980.
‘Seeking a Recipe for Recovery.’ The Times, Letters to the Editor, 11 November
1980.
‘The Causes of Recession.’ The Financial Times, Letters to the Editor, 12 November
1980.
‘Monetary Policies.’ The Times, Letters to the Editor, 19 November 1980.
‘The Causes of the Recession.’ The Financial Times, Letters to the Editor,
19 November 1980.
‘Time to Cut the Noose That Strangles Britain.’ The Guardian, 1 December 1980.
 Wynne Godley – A Bibliography 267

(with Robert Neild) ‘Monetarism’s Testing Time.’ The Times, Letters to the Editor,
9 January 1981.
‘Catastrophic Policies.’ The Financial Times, Letters to the Editor, 6 February 1981.
‘Need to Expand the Economy.’ The Times, Letters to the Editor, 19 February 1981.
(with Francis Cripps) ‘Inflation: Can the Patient Survive?.’ The Times, 9 March
1981.
(with Francis Cripps) ‘A Budget That Will Produce a Hyper-Slump Such As Britain
Has Not Seen Before.’ The Guardian, 16 March 1981.
‘Controls on Imports.’ The Financial Times, Letters to the Editor, 11 May 1981.
‘Cambridge Case on Economy.’ The Telegraph, Letters to the Editor, 24 July 1981.
‘Economic Forecasting.’ The Times, Letters to the Editor, 4 August, 1981.
‘Depth of Recession Was Forecast.’ The Times, Letters to the Editor, 5 August 1981.
‘Economic Forecasting.’ The Financial Times, Letters to the Editor, 13 August 1981.
‘An Urgent Case for Reflation.’ The Times, Letters to the Editor, 18 December 1981.
(with Francis Cripps and Terry Ward) ‘Mrs Thatcher on Course for over 4 million
Unemployed.’ The Guardian, 26 April 1982.
‘Getting Desperate.’ The Observer, 30 May 1982.
‘Deficit That Can Starve.’ The Observer, 29 August 1982.
‘Shore’s Plan: Flawed But Feasible.’ The Times, 24 November 1982.
‘Mr Shore’s Strategy.’ The Times, Letters to the Editor, 8 December 1982.
‘Told You So.’ The Times, Letters to the Editor, January 1983.
(with Francis Cripps) ‘No Oil to Pour on a Chancellor’s Troubled Waters.’ The
Guardian, 14 March 1983.
‘Cambridge Keynesianism Fights Back.’ The Financial Times, April 1984.
‘No Real Signs of an Economic Up-Turn.’ The Times, Letters to the Editor, 20 July
1983.
‘The Myth of the Consumer Boom.’ The Observer, 21 August 1983.
‘False Formulae for Rate-Capping.’ The Times, Letters to the Editor, 13 January
1984.
‘Tory Economic Record.’ The Times, Letters to the Editor, 8 June 1985.
‘Expansion and Unemployment.’ The Financial Times, Letters to the Editor, 30 July
1985.
‘The Too Optimistic View of the Peers.’ The Observer, 27 October 1985.
‘A Doomed Economy.’ New Society, 17 January 1986.
‘Manufacturing Industry.’ The Financial Times, Letters to the Editor, 31 January
1986.
(with Ken Coutts, Bob Rowthorn and Terry Ward) ‘The Downward Path.’ The
Guardian, 10 March 1986.’
‘Oh for Some Truly Radical Policies.’ The Observer, 2 November, 1986.
‘A Growth in the Heart of the Economy.’ The Observer, 7 August 1988.
‘Déjà vu Dogs the Consumer Boom.’ The Observer, 21 August 1988.
‘Why I Won’t Apologise.’ The Observer, 18 September 1988.
‘The Mirage of Lawson’s Supply-Side Miracle.’ The Observer, 2 April, 1989.
‘Exposed: Lawson’s Bogus Billions.’ The Observer, 9 April, 1989.
‘Why the Figures Tell Another Story.’ The Observer, 5 July, 1989.
‘Economic Disaster in Slow Motion.’ The Observer, 27 August, 1989.
‘Monetary Myths and Miracles.’ The Observer, 3 September, 1989.
‘On Track for a Major Recession.’ The Observer, 15 October, 1989.
268 Wynne Godley – A Bibliography

(with Ken J. Coutts and Gennaro Zezza) ‘Is Britain in Credit with the Rest of the
World?.’ The Guardian, 26 January 1990.
‘Where Macroeconomics Went Wrong (A review of A Market Theory of Money by
J.R. Hicks).’ Times Literary Supplement, 18–24 May 1990.
‘Recession Deep, Inflation High.’ The Observer, 19 August 1990.
‘Common Sense Route to a Common Europe.’ The Observer, 6 January 1991.
‘An Old Limousine.’ New Statesman and Society, 11 January 1991:18–21.
‘Out of the cul-de-sac.’ New Statesman and Society, 18 January 1991:18–20.
‘Not a Dirty Word.’ New Statesman and Society, 8 February 1991, 18–20.
‘Terminal Decay: There Is Virtually Nothing That the Chancellor Can Do to Avert
the Slump.’ New Statesman and Society, 15 March 1991:11–14.
‘Giving Up.’ New Statesman and Society, 29 March 1991:16–17.
‘New Consensus - Same Old Recession.’ The Observer, 12 May 1991.
‘Bottoms Out?.’ New Statesman and Society, 17 May 1991: 22–23.
‘A Long View.’ New Statesman and Society, 28 June 1991: 18–19.
(with Robert Rowthorn and Ken J. Coutts) ‘The Route Out of Recession.’ The
Observer, 5 January 1992.
‘Escape from the Infinite Recession.’ New Statesman and Society, 20 March 1992:
30–31.
‘A Severe Hangover.’ New Statesman and Society, 10 April 1992: 26–27.
(with Ken J. Coutts, Jonathan Michie and Robert Rowthorn) ‘Hands-Off Eco-
nomics Equals Stagnation.’ The Observer, 19 April 1992.
‘No Cause for Optimism.’ New Statesman and Society, 17 July 1992: 18–19.
‘Maastricht and All That.’ London Review of Books, 14 (19), 8 October 1992: 3–4.
‘Letting Things Rip (review of T. Congdon, Reflections on Monetarism).’ London
Review of Books, 15 (1), 7 January 1993, p. 9.
‘If in a Year’s Time a Chancellor.’ London Review of Books, 15 (7), 8 April
1993, p. 6.
‘Derailed.’ London Review of Books, 15 (16), 19 August 1993, p. 9.
‘Curried EMU: The Meal That Fails to Nourish.’ The Observer, 31 August 1997.
‘Why the World Could Still Catch Asian Flu.’ The Observer, 26 April 1998.
‘US Risks Stagnation If Net Lending Driving Growth Were to Fall.’ Financial Times,
Letter to the Editor, 25 May 1998.
‘Motor Starts to Sputter. After Six Years of Rapid Growth, the US Economy Faces
– At Best – A Period of Prolonged Stagnation.’ Financial Times, 10 July 1998.
‘Brake on US Growth.’ The Independent, Letters to the Editor, 20 July 1998.
‘Global Slowdown.’ New York Times, Letters to the Editor, 20 July 1998.
‘Don’t Let Gordon Off – He Is Not God.’ The Observer, 23 August 1998.
‘Policies Are Farcical.’ Financial Times, Letters to the Editor, 2 October 1998.
(with Bill Martin) ‘Big Spenders Head for Crisis.’ The Independent, 29 December
1998.
‘Determining Balance between Receipts and Outlays.’ Financial Times, Letters to
the Editor, 26 January 1999.
‘The US Economy: An Impossible Balancing Act. The US Economy Is Facing a
Wall: Personal Savings Are Declining Unsustainably. When This Stops, As It
Must, There Will Be a Recession.’ Financial Times, 19 February 1999.
‘Wrong about US Recessions.’ Financial Times, USA edition, Letters to the Editor,
1 March 1999.
‘Growth Fueled by Heavy Borrowing Hard to Sustain.’ USA Today, 17 March 1999.
 Wynne Godley – A Bibliography 269

(with L. Randall Wray) ‘Nation of Savers, or Just Spenders?.’ The New York Times,
Letters to the Editor, 10 May 1999.
‘Funding May Be Flightier Than It Looks.’ Financial Times, Letters to the Editor,
17 December 1999.
‘U.S. Expansion May Be More Fragile Than Supposed.’ Financial Times, Letters to
the Editor, 11 August 2000.
(with L. Randall Wray) ‘Monetary Independence Vital to Political Independence.’
Financial Times, Letters to the Editor, 12 January 2000.
‘Growing Deficit Not Sustainable For Ever.’ Financial Times, Letters to the Editor,
19 September 2000.
‘Bush Should Triple His Tax Cuts. The US President Is Right to Relax Fiscal Policy
But His Plans Are Not Aggressive Enough to Avoid Recession.’ Financial Times,
22 January 2001, p.15.
‘Saving Masud Khan.’ London Review of Books, 23(4), 22 February 2001:3–7.
‘Recession, USA.’ The Guardian, 23 October 2001.
‘US Fiscal Policy Is Not Neutral.’ Financial Times, Letters to the Editor, 18 December
2001.
‘Manufacturing Matters Very Much Indeed.’ Financial Times, Letters to the Editor,
18 January 2002.
‘Kick-Start Strategy Fails to Fire Spluttering US Economic Motor.’ The Guardian, 21
January 2002.
‘Comparison of Household Debt Burdens.’ Financial Times, Letters to the Editor,
9 April 2002.
‘Complacency about Debt.’ Financial Times, Letters to the Editor, 2 May 2002.
‘One-Trick MPC Could Not Halt the Crisis.’ The Observer, 26 May 2002.
‘Why the World Cannot Rely on the US to Drive Growth.’ Financial Times, Letters
to the Editor, 16 July 2002.
‘Huge Fiscal Expansion Shortened U.S. Recession.’ Financial Times, Letters to the
Editor, 8 August 2002.
(with Bill Martin) ‘America’s Years of Living Dangerously.’ The Observer, 1
September 2002.
‘CAP Is a Destructive Monstrosity for Britain.’ Financial Times, 30 September 2002.
‘The New Interest–Rate Orthodoxy Is as Flawed as the Old One.’ The Guardian, 11
November 2002.
‘Behind US’s Miraculous Performance.’ Financial Times, Letters to the Editor, 4
December 2002.
‘Dangerous Folly of Abiding by Brown’s Golden Rule.’ Financial Times, 3 January
2003.
‘Too Big a Trade Imbalance to Handle at Home.’ Financial Times FT.com site,27
January 2003.
‘Accounting for Acquisitions via Share Exchange.’ Financial Times, Letters to the
Editor, 19 March 2003.
‘One-Club Golf Is for Losers.’ The Guardian, 19 May 2003.
‘The Awful Warning of the Lawson Boom.’ Financial Times FT.com site, 6 August
2003.
‘When the Trust That You Have in Your Bank Is Lost.’ Financial Times, 6 December
2003.
‘How RBS Gave Advice to My Niece.’ Financial Times, Letters to the Editor, 20
December 2003.
270 Wynne Godley – A Bibliography

‘Forecasting Is Defunct as a Means of Shaping Monetary Policy.’ The Guardian, 11

October 2004.
(with Alex Izurieta) ‘Deficits That Need a Global Answer.’ Financial Times, 3
December 2004.
(with L. Randall Wray) ‘Fed Can Handle Reserves to Keep US Rates on Target.’
Financial Times , 21 September 2005.
‘Pensions £150bn Black Hole May Not Be So Deep.’ Financial Times, Letters to the
Editor, 11 January 2006.
(with L. Randall Wray) ‘The Balance of Trade, Not Payments, Is True Measure of a
Deficit’s Effects.’ Financial Times, Letters to the Editor, 15 February 2006.
(with L. Randall Wray) ‘Obscure Argument Not Easy To Follow.’ Financial Times,
Letters to the Editor, 17 February 2006.
‘US Economy and the Deficit Predicament.’ Financial Times, Letters to the editor,
30 May 2006.
‘New Balance of Payments Figures May Transform Strategic Outlook.’ Financial
Times, Letters to the Editor, 28 December 2007.
‘Have Oxford Trio Built New Theory?.’ Financial Times, Letters to the Editor, 1
May 2008.
‘Tackle Inflation with a Sensible Incomes Policy.’ Financial Times, Letters to the
Editor, 20 June 2008.
(with Graham Gudgin, Bill Martin and Barry Moore) ‘Bank Downplayed the
Downside Risk.’ Financial Times, Letters to the Editor, 20 November 2008.
‘Immediate Cuts to Budget Deficit Will Worsen Recession.’ Financial Times, Letters
to the Editor, 9 October 2009.
‘An Expansionary Route to Cut Deficit.’ Financial Times, Letters to the Editor, 20
April 2010.

Sources
1. Godley’s own CV, written in 1992 (Levy Institute of Economics);
2. Search results from Google Scholar (using “Publish or Perish”) and
Google;
3. University of Cambridge Library on-line catalogue;
4. Search results on the archives at Financial Times (www.ft.com),
The Guardian (www.guardian.co.uk), The New Statesman (www.
newstatesman.com), London Review of Books (ww.lrb.co.uk);
5. Search results from SCOPUS;
6. Lists previously compiled by Alex Izurieta and by Claudio Dos Santos;
7. A scrapbook of newspaper clips in the possession of Eve Taylor.

Note
We believe we have identified nearly all published papers and working papers. By
contrast, we have most likely left out several of Wynne Godley’s many articles in
newspapers. We are grateful for the help provided by Claudio Dos Santos, Alex
Izurieta, and Eve Taylor (Wynne Godley’s daughter).
Index

accounting capital accumulation, 129, 131–2,

consistent, 23–31 140–50
inflation, 21–38 capital assets, 124
social, 125–8 capital equipment, 75
stock-flow, 123, 150 capital gains, on equities, 33–4
accounting consistency, 15 capital stock, 75
adjusted fiscal ratio (AFC), 220–1 cash holdings, 97
adjusted trade ration (ATR), 222–3 central banks, 167–9, 178–9, 189–93,
Article 12 control, 240–1 203, 206
asset-demand functions, 136, 167–8, combined fiscal and trade ratio
171 (CFTR), 223–4
assets commercial banks, in RSFM, 51–3
capital, 124 commodity market, 65–6, 218
prices, 12–13 Congressional Budget Office (CBO)
forecasts, 230–2, 237–8
balance of payments, 191–2, 212–13, consistent accounting, 2–3
221–5, 230–6 of nominal stocks and flows, 23–7
balance sheet, 94 in real terms, 28–31
banks’, 105, 108–9 consistent monetary economics,
consistency, 24–5 160–1
balance sheet matrix, 7, 162 consumer spending, 32
bank deposits, 139, 143–4 consumption, 96–8, 124, 135, 137–8
bank loans, 51–3, 90–2, 98–100, 107, changes in propensity to consume,
113–14, 133–4 140–3
changes in interest rates on, 143–4 increases in, 147
bank money, 102 consumption demand, 67
banks consumption function, 8, 85–6, 135,
balance sheets, 105, 108–9 165
behaviour of, 98–100, 133–4 corporate profits, 73–4
commercial, 51–3 corporations, 46–50
in Kaleckian growth model, 133–4 see also firms
profits of, 98–100 cost-plus pricing, 4
role of, 24–5, 90–2 Cripps, Francis, 1–3, 5, 8, 13, 249
bond market, 65–73 currency depreciation, 240, 241–2
bond rate of interest, 114–16 current account deficit, 13
bonds, 5, 36–8
budget constraint, 123–4, 133 Davidson, Paul, 24
budget deficit, 194, 201, 203, 217 debt dynamics, 11–12, 201–2
budget surplus, 13, 185–6, 218, 231 debt identity, 13
debt-to-capital ratio, 130
Cambridge Economic Policy Group defensive asset ratio, 109–10
(CEPG), 8–9 demand-supply relationship, 41
capacity utilization, 130 disequilibrium hypothesis, 134–5

271
272 Index

disposable income bond issuance by, 75

personal, 26–9, 33, 34, 164–5, 227 borrowing by, 132
public, 30–1 budget constraint of, 123–4, 133
real, 85 capital equipment, 75
distributed dividends, 132–3 cash flow of, 148
dividends, distributed, 132–3 changes in parameters controlled
dollar devaluation, 240, 241–2 by, 147–50
decisions by, 129
economic policy, 9 expectations of, 96, 111–14
eurozone, 9, 159–87 investment by, 129–32
Economic Report of the President in Kaleckian growth model, 128–33
(ERO), 232, 234 profits of, 73–6, 95–6, 147
economic systems, inflation retention ratio of, 148–50
accounting of whole, 21–38 securities issues by, 148
Eichner, Alfred, 4 stock exchange value of, 152n11
endogenous money, 5, 81–9 fiscal policy, 4, 11, 13, 42
equilibrium eurozone, 159–87
general, 40 reaction function, 204–11
macroeconomic, 40 role of, 203–4
neoclassical, 41–2 RSFM and, 55–8
equities in SFC model, 194–214
capital gains on, 33–4 US, 216, 219–21
issuance of, 148 fixed coupon bonds, inflation and real
price of, 136–7 capital losses on, 36–8
propensity to hold, 144–6 flexible exchange rates, 159–86
euro, 174, 184–5 flow identities, 25–7
European Central Bank (ECB), 9–11, flow of funds account, 161
159, 172–3, 178–9, 185 flow of funds matrix, 70–1, 92–4
European Community (EC), 189–93 flow variables, 22
European government, 11 foreign assets, 235
eurozone economic policy, 9–10, foreign direct investment, 235
159–87 foreign sector, 211–13
exchange, theory of, 90–1 foreign trade, 221–4
exchange rates free trade, 9
flexible, 9–10, 159–86 Friedman, Milton, 204
management of, 13 full employment, 11, 13, 67, 194, 202,
exogenous growth rates, 129 210–11, 213–14
exogenous interest rates, 168–73 fundamental accounting identity, 8–9
exogenous money, 90
expectations, 90, 91, 96 General Agreement on Tariffs and
household, 135–7 Trade (GATT), 240
random, 111–14 general equilibrium, 40
exports, 165–7, 211–13, 240 global financial crisis, 11
external shocks, 173–8 government
behaviour of, 100–2
federalism, 193 budget constraint, 100–1
Federal Reserve, 172–3 government debt, 107, 108, 113, 114,
firms 116, 211
behaviour of, 94–6, 128–33 constraints on, 179
 Index 273

government debt – continued personal disposable, 26–9, 33, 34,

interest on, 29, 32 164–5, 227
real, 35–6 public disposable, 30–1
targets for, 194 real, 28–9
government expenditure, 32, 108–11, real disposable, 85
200–1, 205–6, 208–10, 218 real net, 66
Graziani, Augusto, 5, 6, 81–9 sectoral, 26–7
Greek crisis, 10 income shocks, 112
gross domestic product (GDP), 194, inflation, 42
200–1, 203, 220–1, 224, 239 on fixed coupon bonds, 36–8
addition to, 108–9 national income and, 54–5
debt relative to, 12, 13, 14, 176–8, on perpetuities, 36–8
183–5 rate of, 204–11, 213–14
steady-state flow of, 102–3 in RSFM, 53–5
growth, real rate of, 201–2 inflation accounting, 3
growth models of whole economic systems, 21–38
classical, 140 inflation neutrality, conditions for,
heterodox, 131 31–6
Kaleckian, 7–8, 123–56 inflation tax, 191
Marxian, 140 interest payments, 69, 71, 72, 74,
130–1
Haig-Simons measure of income, 3, interest rates, 69, 106, 110, 113, 194
165 bond rate of, 114–16
Hicks, John, 3, 23, 41, 43, 88, 90, 117 changes in, 143–4
historical time, 90–1 exogenous, 168–73
household debt, 13–14, 224–8 household portfolio responses to,
household income, 66–9, 72–3, 77, 115
134, 137–8 inflation and, 54–5
households money rate of, 116
allocation of interest bearing assets money supply and, 73
of, 111 nominal, 23, 29, 31–2, 202
behaviour of, 96–8, 134–7 real, 29, 33, 34, 201–2, 203
consumption by, 124, 135, 137–8, inventories, 25–7
224–8 levels of, 103–8
expectations, 135–7 response of, to sales shocks, 112
in Kaleckian growth model, 134–7 investment demand, 67
in RSFM, 50–1 investment finance constraint, 71
wealth of, 123–4, 152n19 investment functions, 129–32, 139,
household savings, 133–5, 224–8 148–9
housing prices, 35 IS-LM model, 90, 121n3
Italian circuit school, 5
imports, 18, 165–7, 211–13, 240–1
income Kaldor, Nicholas, 5, 8, 40, 43, 87, 90,
definition of, 23, 134 123–4, 138–9
Haig-Simons measure of, 3, 165 Kaleckian growth models, 7–8,
household, 66–9, 72–3, 77, 134, 123–56
137–8 banks in, 133–4
national, 26, 30–2, 54–5, 88, 161 behavioural relationships, 128–39
personal, 161 changes in interest rates and, 143–4
274 Index

Kaleckian growth models – continued money supply, 42, 65–7, 72–3, 77, 86,
changes in parameters controlled by 90, 91, 138
firms and, 147–50 Moore, Basil, 87, 131
changes in propensity to consume mortgage debt, 35
and, 140–3
national balance sheets, 24–5
changes in propensity to hold
national income, 26, 32, 161
equities and, 144–6
inflation and, 54–5
changes in real wages and, 146–7
real, 30–1, 88
experiments, 139–50
national income accounts, 22
firms in, 128–33
national income and product
households in, 134–7
accounts (NIPA), 161
social accounting and, 125–8
negative external shocks, 173–8
system-wide implications, 137–9
neo-charalists, 11
Keynesian economics, 40, 81–9, 90
neoclassical equilibrium, 41–2
labour demand, 55 neoclassical marcoeconomic model
labour market, 40, 65–6, 218 bond market and, 69–73
leverage ratio, 130 government in, 101–2
loans, 51–3, 90–2, 98–100, 107, inconsistency in, 65–80
113–14, 133–4 numerical simulation of, 77–9
standard, 68–9
Maastricht treaty, 10–11, 185, 189–93 neoclassical paradigm (NCP), 39–40,
MacDougall Report, 192 42
macroeconomic equilibrium, 40 neoclassical production function,
macroeconomic models, 5 40–1
inconsistency in neoclassical, 65–80 neoclassical theory, 4–5, 150
macroeconomic theory, 6, 22 neoclasssical synthesis (NCS), 90–1
neoclassical, 39–42 neo-Pasinetti model, 123–4, 135, 139
managed trade, 9 net export demand, 221–2
market price, 41 net foreign indebtedness, of US, 13
mark-up, 4, 128–9, 146–7 net lending, 12
Minsky, Hyman, 130–1 New Cambridge approach, 13, 242
monetary economy, 124, 138–9 nominal interest rates, 23, 29, 31–2,
monetary policy, 11, 194, 203–4, 217 202
monetary theory of production, 5 nominal stocks and flows, consistent
money accounting of, 23–7
bank, 102 nominal wages, 88–9
demand for, 88, 107, 138, 168 normal regime, 139, 140, 143–6
endogenous, 5, 81–9
open economies, 9
exogenous, 90
holdings, 124 Pasinetti, Luigi, 40, 43, 211
quantity theory of, 65 Patinkin, Don, 67, 73–6
real demand for, 73 perpetuities, inflation and real capital
real stock of, 229 losses on, 36–8
response of, to income shocks, 112 personal disposable income, 26–9,
money creation/destruction, 13 33–4, 164–5, 227
money deposits, 137 personal income, 161
money market, 65–6, 72 Phillips curve analysis, 204–5
money rate of interest, 116 post-Keynesianism, 11, 131, 150
 Index 275

prices real stock of money (M3), 229

of equities, 136–7 real wages, 88–9, 146–7
of goods, 40, 41 retained earnings, 132–3
market, 41 retention ratio, 148–50
mark-up and, 128–9 Robinson, Joan, 5, 117, 139
price stability, 11 RSFM, see real stock-flow monetary
private expenditure function, 242–7, model (RSFM)
251
sales shocks, 112
private savings, 12
savings
private sector
household, 133–5, 224–8
indebtedness of, 13
paradox of, 140, 141–3
net financial assets, 13
sectoral income, 26–7
net lending to, 12
social accounting matrices (SAM), 123,
production, monetary theory of, 5
125–8
production function, neoclassical,
sovereignty, 10–11, 190, 191
40–1, 81
steady-state growth, 133
profits, 26, 95–6
stock appreciation, 26
bank, 98–100
stockbuilding, 42
distribution of, 73–6
stock exchange value, 152n11
wages and, 147
stock-flow accounting, 123, 150
propensity to consume, 140–3
stock-flow coherent method, 1–2, 5–7
protectionism, 240–1
stock-flow-consistent (SFC) model
public debt, see government debt
accounting framework, 92–4
puzzling regime, 139, 141–6
analytical results, 200–1
quantity theory of money, 65 applied to three economies with
two currencies, 159–87
random expectations, 111–14 arithmetical results, 199–200
reaction function, 204–11 behavioural assumptions, 94–102
real capital losses development of, 14
on fixed coupon bonds, 36–8 fiscal policy in, 11, 194–214
on perpetuities, 36–8 with foreign sector, 211–13
real disposable income, 85 interest in, 14–15
real income, 3, 28–9 Kaleckian growth models in, 123–56
real interest rates, 29, 33, 34, 203 long-run properties, 102
real money stock, 12 outline of, 195–9
real national income, 30–1, 88 parameters of, 118–20
real net income, 66 research on, 15–16
real public disposable income, 30–1 simulations, 102–16
real stock-flow monetary model variables of, 118
(RSFM), 4, 43–60
stock-flow consistent accounting, 2–3
basic accountancy of, 43–6
stock-flow ratios, 14, 15
commercial banks in, 51–3
stock variables, 22
corporations in, 46–50
Stone, Richard, 123
features of, 58
subprime financial crisis, 12
fiscal policy and, 55–8
Sylos Labini, Paolo, 3, 39–41
full steady-state of, 55–8
household sector in, 50–1 tariffs, 241
inflation in, 53–5 taxation, 190–1, 218
276 Index

Taylor, Lance, 161 foreign trade and payments, 221–4

theory of the firm, 68 gross domestic product (GDP),
theory of the household, 68 220–1, 224
three-country model, 159–87 policy considerations for, 240–2
equations and experiments, 178–83 private saving, spending and
exogenous interest rates, 168–73 borrowing, 224–8
fiscal policies, 168–73 private sector behaviour scenarios,
negative external shocks and, 173–8 236–40
overview of, 161–8 strategic prospects, 229–40
time, 42, 90–1 unsustainable processes in, 12–13,
Tobin, James, 8, 14, 43, 56, 82, 90, 97, 213–14
102, 117, 123 US housing boom, 12, 13–14
total financial wealth, 24–5
trade valuation ratio, 123–4, 131, 139
free, 9 wages, 40, 55, 88–9, 146–7, 148
managed, 9 Walrasian model, 91–2
trade balance, 212–13, 221–4 Walras’s law, 5, 66, 67, 69–72
transactions-flow matrix, 7, 82–4, 126, wealth, 3, 124
161, 163 allocation of household, 113
two-country models, 160–1 bonds and money as shares of, 107
unemployment, 40, 217, 218, 236 components of, 104, 110
United States of households, 123–4
adjusted fiscal ratio (AFC), 220–1 wealth accumulation, 96–7, 165
balance of payments, 232–6 whole economic systems, inflation
budget deficit, 217 accounting of, 21–38
economy of, 216–18 Wood, Adrian, 4, 43, 48, 50, 133
eurozone and, 159–87 world economy, 161, 238–40
fiscal policy, 216, 219–21 World Trade Organization (WTO), 240
foreign indebtedness, 232–6 Wray, L. Randall, 87