Macro-Financial Policy Analysis

with
Dynamic Stochastic General
Equilibrium Models
Lecture 2

Oleksandr Faryna
National Bank of Ukraine
National University of Kyiv-Mohyla Academy
[email protected]
Recap: DSGE models

• D – for Dynamic – Economies evolve over time

• S – for Stochastic – Economic environment appear subject to
unexpected disturbances creating the business cycles
• GE – for General Equilibrium – interaction between different markets
and prices are determined jointly
Recap: The Canonical New Keynesian DSGE
• The biggest problem of applied modeling may be an overliteral
interpretation and extrapolation of the model’s result to the real
world
• A very good discussion of this can be found in Jon Faust, "DSGE Models: I
Smell a Rat (and It Smells Good)", IJCB 2012
Schedule
Date Lecture/seminar Topic
Sep 27 Lecture 1 Course introduction
Oct 4 Lecture 2 and 3 Real Business Cycle model
Oct 11 Lecture 4 and 5 New Keynesian model with sticky prices
Oct 18 Lecture 6 and 7 Small open economy model
Oct 25 Lecture 8 and 9 Calibration and Estimation. Other nominal and real
frictions. Test. Assignment intro
Nov 8, 15 Seminar 1, 2, 3 and 4 Introduction to software (Dynare). Model solution and
simulation.
Nov 22 Seminar 5 and 6 Assignment defense
Agenda for today

• Inspecting the mechanisms in simple models

• Solving constrained maximization problems – Lagrange multipliers
• Household’s consumption/saving decision
• Household’s labor/leisure decision
• Firm’s profit maximization – static environment

• Setting-up a Real Business Cycle (RBS) model

Lagrange multipliers

• Economic decisions usually modeled as an optimization problem

subject to constraints
• Households make consumption decisions constrained by their budget
• Households make labor supply decisions constrained by the fact that there is
24 hours a day
• Firms make production decisions constrained by limited production capacity
• Firms minimize costs constrained by the orders they have to fulfill
Constrained optimization problems

max 𝐹(𝑋1 , … , 𝑋𝑛 )
𝐺 𝑋1 , … , 𝑋𝑛 ≤ 𝑌

Objective function 𝐹(𝑋1 , … , 𝑋𝑛 ), where 𝑋1 , … , 𝑋𝑛 are instruments/

control variables subject to a constraint 𝐺 𝑋1 , … , 𝑋𝑛 ≤ 𝑌
• Usually solved by the Lagrange approach, but other methods also possible
• Transform constrained into unconstrained maximization problem
• Dynamic programming
Constrained optimization problems (2)

• Introduce a new variable, 𝜆, the Lagrange multiplier, and set up the

Lagrangian ℒ
ℒ = 𝐹 𝑋1 , … , 𝑋𝑛 + 𝜆 𝑌 − 𝐺 𝑋1 , … , 𝑋𝑛
• The necessary first-order conditions (assuming an interior solution)
for a maximum are
𝛿ℒ 𝛿𝐹 𝛿𝐺
= +𝜆 = 0, 𝑓𝑜𝑟 𝑗 = 1, … , 𝑛
𝛿𝑋𝑗 𝛿𝑋𝑗 𝛿𝑋𝑗
An example – two variable problem

• Maximize 𝐹 𝑋1 , 𝑋2 subject to 𝐺 𝑋1 , 𝑋2 = Y

ℒ 𝑋1 , 𝑋2 = 𝐹 𝑋1 , 𝑋2 + 𝜆(𝑌 − 𝐺 𝑋1 , 𝑋2 )
An example – two variable problem (2)
• Necessary conditions for a local maximum requires setting first-order conditions equal to
zero
𝛿ℒ
= 𝐹𝑋1 − 𝜆𝐺𝑋1 = 0
𝛿𝑋1
𝛿ℒ
= 𝐹𝑋2 − 𝜆𝐺𝑋2 = 0
𝛿𝑋2
𝛿ℒ
= 𝑌 − 𝐺(𝑋1 , 𝑋2 ) = 0
𝛿𝜆
𝜹𝑭
Note derivatives may be denoted as or 𝑭𝑿
𝜹𝑿

Solve for the optimal values 𝑋1∗ , 𝑋2∗ and 𝜆∗ three unknowns and three equations)
Interpreting the Lagrange multiplier
The optimal 𝑋1∗ , 𝑋2∗ and 𝜆∗ are functions of the constraint parameter 𝑌
(where 𝑌 is a fixed exogenous parameter), what happens to these
critical values when we change Y?
ℒ ∗ = 𝐹 𝑋1∗ , 𝑋2∗ + 𝜆∗ (𝑌 − 𝐺 𝑋1∗ , 𝑋2∗ )
Totally differentiating ℒ ∗ with respect to 𝑌 we get
𝛿𝐹(𝑋1∗ , 𝑋2∗ )
= λ∗
𝛿𝑌
The Lagrange multiplier tells the effect of a change in the constraint
(parameter 𝑌) on the optimal value of the objective function 𝐹
Examples from economics

• Cost minimization – two inputs capital, 𝐾, and labor, 𝑁

min 𝐶(𝐾, 𝑁) 𝑠𝑢𝑏𝑔𝑒𝑐𝑡 𝑡𝑜 𝑌 = 𝐾 1−𝛼 𝑁 𝛼
𝐾,𝑁
∗
𝛿𝐶
𝜆∗ =
𝛿𝑌
The Lagrange multiplier interpreted as marginal cost
Examples from economics (2)

• Utility maximization – two inputs, 𝑋 and Z with prices 𝑃𝑋 and 𝑃𝑧 ,

respectively, and income 𝑌
max 𝑈(𝑋, 𝑍) 𝑠𝑢𝑏𝑔𝑒𝑐𝑡 𝑡𝑜 𝑃𝑋 𝑋 + 𝑃𝑍 𝑍 = 𝑌
𝑋,𝑍
𝛿𝑈 ∗
𝜆∗ =
𝛿𝑌
• The Lagrange multiplier interpreted as marginal utility of income
Real-life example

green or red
Utility maximization – two goods
max 𝑈(𝑋, 𝑍) 𝑠𝑢𝑏𝑔𝑒𝑐𝑡 𝑡𝑜 𝑃𝑋 𝑋 + 𝑃𝑍 𝑍 = 𝑌
𝑋,𝑍
ℒ = 𝑈 𝑋, 𝑍 + 𝜆[𝑌 − 𝑃𝑋 𝑋 − 𝑃𝑍 𝑍]
𝛿ℒ
1 ∶ = 𝑈𝑋 − 𝜆𝑃𝑋 = 0
𝛿𝑋
𝛿ℒ
2 ∶ = 𝑈𝑍 − 𝜆𝑃𝑍 = 0
𝛿𝑍
𝛿ℒ
3 ∶ = 𝑌 − 𝑃𝑋 𝑋 − 𝑃𝑍 𝑍 = 0
𝛿𝜆
• Solve for 𝑋 ∗ , 𝑍 ∗ , and 𝜆∗ - three unknowns and three equations
Interpretation of FOCs
• Combine (1) and (2) to get
𝑈𝑋 𝑃𝑋
=
𝑈𝑍 𝑃𝑍

𝑈𝑋
• is (Intra-temporal) MRS – Marginal Rate of Substitution - the household’s willingness
𝑈𝑍
to substitute one good for another
• MRS X for Z, the amount of Z that the household is willing to give up to get an additional unit of X
to maintain the same level of utility
𝑃𝑋
• is MRE – Market’s Rate of Exchange – the market’s willingness to substitute one
𝑃𝑍
good for another, given by the price ratio
Solve for a specific utility function
• A common utility function in economics is the Constant Elasticity of
Substitution (CES) utility function
1 1
1− 1−
𝑋 −1 𝑍
𝜎 −1 𝜎
𝑈 𝑋, 𝑍 = +
1 1
1− 1−
𝜎 𝜎
• And if 𝜎 = 1
𝑈 𝑋, 𝑍 = 𝑙𝑛𝑋 + 𝑙𝑛𝑍
Solve for a specific utility function (2)

𝜎 equals the elasticity of substitution, which has the following meaning

𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 (𝑍 ∗ / 𝑋 ∗ )
𝜎=
𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 (𝑃𝑋 /𝑃𝑍 )
First order conditions with CES-utility
1 1
1− 1−
𝑋 −1 𝑍
𝜎 −1 𝜎
𝑈 𝑋, 𝑍 = +
1 1
1− 1−
𝜎 𝜎

1 1 1
1− −1 −
1
𝑈𝑋 = ∗ 1− ∗𝑋 𝜎 =𝑋 𝜎
1 𝜎
1−
𝜎
1
−𝜎
𝑈𝑍 = 𝑍
First order conditions with CES-utility

1
−
1 ∶ 𝑈𝑋 − 𝜆𝑃𝑋 = 0 → 𝑋 𝜎 − 𝜆𝑃𝑋 = 0
1
−
2 ∶ 𝑈𝑍 − 𝜆𝑃𝑍 = 0 → 𝑍 𝜎 − 𝜆𝑃𝑍 = 0
3 ∶ 𝑌 − 𝑃𝑋 𝑋 − 𝑃𝑍 𝑍 = 0
First order conditions with CES-utility

• Combine (1) and (2) to get

𝜎
𝑍 𝑃𝑋
=
𝑋 𝑃𝑍
𝑌 = 𝑃𝑋 𝑋 + 𝑃𝑍 𝑍
The optimal consumption levels
• Solving for optimal consumption levels gives
∗
1
𝑋 = 𝜎 𝑌
𝑃
𝑃𝑋 + 𝑃𝑍 𝑋
𝑃𝑍
∗
1
𝑍 = 𝜎 𝑌
𝑃
𝑃𝑍 + 𝑃𝑋 𝑍
𝑃𝑋
1
𝜎 𝜎
1 1 𝑃𝑋
𝜆∗ = 𝑃𝑋 + 𝑃𝑍
𝑌 𝑃𝑋 𝑃𝑍
Increasing 𝑌 “loosens” the budget constraint, 𝜆∗ decreases
Consumption/saving: one good – two periods
Real-life example

≠
Today Tomorrow
Real-life example

=𝜷
Today Tomorrow
Consumption/saving: one good – two periods

• Households live two periods

• 𝐶1 period 1 consumption and 𝐶2 period 2 consumption
• Income, 𝑌1 and 𝑌2 , exogenous in both periods
• Real interest rate, 𝑅, exogenous
• Perfect capital markets: households can borrow and lend any amount at a
given rate 𝑅
The budget constraint

First period: households choose to consume or save, 𝐵1 , out of period

1 income
1 ∶ 𝐶1 + 𝐵1 = 𝑌1
Second period: households receive income and principal plus interest
on its savings
2 ∶ 𝐶2 = 𝑌2 + 1 + 𝑅 𝐵1
The budget constraint (2)

Combine (1) and (2) to get the intertemporal budget constraint

𝐶2 𝑌2
𝐶1 + = 𝑌1 +
1+𝑅 1+𝑅
Present value (or discounted sum) of consumption equals the present
value of income
The household’s utility function

𝑈 𝐶1 , 𝐶2 = 𝑈 𝐶1 + 𝛽𝑈(𝐶2 )
where the parameter 𝛽 ∈ (0,1) is the subjective discount factor
1
(hence, the time preference 𝜌 = − 1)
𝛽
Household’s maximization problem
max 𝑈(𝐶1 , 𝐶2 ) = 𝑈 𝐶1 + 𝛽𝑈(𝐶2 )
𝐶1 ,𝐶2
𝐶2 𝑌2
𝑠𝑢𝑏𝑔𝑒𝑐𝑡 𝑡𝑜 𝐶1 + = 𝑌1 +
1+𝑅 1+𝑅
Trade-off
• Consuming early is better since future consumption gives less utility
𝛽 ∈ (0,1)
• However, postponing consumption through savings allows higher
consumption next period, since savings are rewarded at rate 𝑅
Step 1: The Lagrangian
𝑌2 𝐶2
ℒ = 𝑈 𝐶1 + 𝛽𝑈 𝐶2 + 𝜆 𝑌1 + − 𝐶1 −
1+𝑅 1+𝑅
FOCs with respect to 𝐶1 , 𝐶2 and 𝜆
𝛿ℒ
1 ∶ = 𝑈𝐶1 − 𝜆 = 0
𝛿𝐶1
𝛿ℒ 𝜆1
2 ∶ = 𝛽𝑈𝐶2 − =0
𝛿𝐶2 1+𝑅
𝛿ℒ 𝑌2 𝐶2
3 ∶ = 𝑌1 + − 𝐶1 − =0
𝛿𝜆 1+𝑅 1+𝑅
Step 1: The Lagrangian (2)
Equations (1) and (2) give
𝑈𝐶 1
4 ∶ =1+𝑅
𝛽𝑈𝐶2
𝑈𝐶 1
• is (Inter-temporal) Marginal Rate of Substitution – household’s
𝛽𝑈𝐶2
willingness to substitute (give up) consumption today for consumption next
period
• 1 + 𝑅 is Markets Rate of Exchange – the market’s willingness (or price) to
substitute (give up) consumption today for consumption next period
Step 1: The Lagrangian (3)

Another interpretation of the FOCs

𝑈𝐶1 = 𝛽𝑈𝐶2 (1 + 𝑅)
• Marginal cost, 𝑈𝐶1 , (MC) of less consumption in period 1 equals
marginal benefit, 𝛽𝑈𝐶2 (1 + 𝑅), (MB) of consumption in period 2
Step 2: Solve for the CES utility function

1 1
1− 1−𝜎
𝐶 𝜎 −1 𝐶2 − 1
𝑈 𝐶1 , 𝐶2 = +𝛽
1 1
1− 1−
𝜎 𝜎
• Now sigma measures the intertemporal elasticity of substitution, i.e.
the response of consumption today relative to next period to a
change in the interest rate
Step 2: Solve for the CES utility function (2)

• Given the FOC 𝑈𝐶1 = 𝛽𝑈𝐶2 (1 + 𝑅) and the CES utility function, we
can show that
𝐶2 𝜎
= 𝛽 1+𝑅
𝐶1
• The intertemporal elasticity of substitution determines the response
of relative consumption to an interest rate change
 1 1
1− 1−
𝜎
𝐶 𝜎 −1 𝐶2 − 1
𝑈𝐶1 = 𝛽𝑈𝐶2 (1 + 𝑅) 𝑈 𝐶1 , 𝐶2 = +𝛽
1 1
1− 1−
𝜎 𝜎

𝐶2 𝜎
= 𝛽 1+𝑅
𝐶1
Step 3: Solve for optimal consumption levels
𝐶2 𝜎
• Substitute = 𝛽 1+𝑅 into the budget constraint
𝐶
𝐶2 1 𝑌2
𝐶1 + = 𝑌1 + to yield the demand functions
1+𝑅 1+𝑅
1 𝑌2
𝐶1∗ = 𝜎−1 𝛽 𝜎
𝑌1 +
1+ 1+𝑅 1+𝑅
𝜎 𝛽𝜎
∗
1 + 𝑅 𝑌2
𝐶2 = 𝜎−1 𝜎
𝑌1 +
1+ 1+𝑅 𝛽 1+𝑅
• Substitution and income effects are related to the first term in brackets and
the wealth effect to the second term
The substitution effect

• When 𝑅 increases, price of second period consumption falls making

savings more attractive, hence 𝐶1 falls
The income effect

• Effect depends on the household is a lender or borrower

• Lender: when 𝑅 increases, income increases from saving, and consumption
rises in both periods
• Borrower: when 𝑅 increases, debt become more costly, hence, consumption
falls in both periods
The wealth effect

• When 𝑅 increases, the market discount factor falls, which reduces the
𝑌2
present value of income, 𝑌1 + , and hence consumption falls
1+𝑅
Evaluate substitution/income effects

• Case 1, 𝝈 > 𝟏 – The substitution effect dominates the income effect

• Case 2, 𝝈 < 𝟏 – The income effect dominates the substitution effect
• Case 3, 𝝈 = 𝟏 – The substitution and income effect cancel each other
The labor/leisure choice
The labor/leisure choice

• Intra-temporal optimality condition

• Inter-temporal optimality condition
• Frisch elasticity of labor supply
Household’s preferences

𝑈 𝐶1 , 𝐶2 , 𝐿1 , 𝐿2 = 𝑈 𝐶1 + 𝑉 𝐿1 + 𝛽 𝑈 𝐶2 + 𝑉 𝐿2
where 𝐿 denotes leisure and 𝑉(∗) denotes utility from leisure
Household’s budget constraint

• The one-period budget constraints

𝐶1 + 𝐵1 = 𝑊1 𝑁1
𝐶2 = 𝑊2 𝑁2 + 1 + 𝑅 𝐵1
where 𝑊 denotes wages, 𝑁 hours worked, and 𝐵 savings
Household’s budget constraint (2)

• Corresponding intertemporal budget constraint

𝐶2 𝑊2 𝑁2
𝐶1 + = 𝑊1 𝑁1 +
1+𝑅 1+𝑅
• Note 1: income is endogenous 𝑌 = 𝑊𝑁
• Note 2: Household’s time constraints are
• 1 = 𝐿1 + 𝑁1
• 1 = 𝐿2 + 𝑁2
The Lagrangian

𝑈 𝐶1 + 𝑉 𝐿1 + 𝛽 𝑈 𝐶2 + 𝑉 𝐿2 +
𝑊2 𝑁2 𝐶2
ℒ= max 𝜆 𝑊1 𝑁1 + − 𝐶1 − +
𝐶1 ,𝐶2 ,𝐿1 ,𝐿2 ,𝑁1 ,𝑁2 1+𝑅 1+𝑅
𝜇1 1 − 𝐿1 − 𝑁1 + 𝛽𝜇1 1 − 𝐿2 − 𝑁2
FOCs
𝛿ℒ
= 𝑈𝐶1 − 𝜆 = 0 ∶ 𝜆 = 𝑈𝐶1
𝛿𝐶1
𝛿ℒ 𝜆1
= 𝛽𝑈𝐶2 − = 0 ∶ 𝜆 = 𝛽 1 + 𝑅 𝑈𝐶2
𝛿𝐶2 1+𝑅
𝛿ℒ
= 𝑉𝐿1 − 𝜇1 = 0 ∶ 𝜇1 = 𝑉𝐿1
𝛿𝐿1
𝛿ℒ
= 𝛽𝑉𝐿2 − 𝛽𝜇2 = 0 ∶ 𝜇2 = 𝑉𝐿2
𝛿𝐿2
𝛿ℒ 𝜇1
= 𝜆𝑊1 − 𝜇1 = 0 ∶ 𝜆 =
𝛿𝑁1 𝑊1
𝛿ℒ 𝜆𝑊2 𝜇2
= − 𝛽𝜇2 = 0 ∶ 𝜆 = 𝛽(1 + 𝑅)
𝛿𝑁2 1 + 𝑅 𝑊2
Interpreting FOCs

𝑈𝐶1
= 1+𝑅
𝛽𝑈𝐶2
Marginal rate of substitution and Marginal rate of exchange - same as
before
Interpreting FOCs (2)

𝑉𝐿1 𝑉𝐿2
= 𝑊1 𝑎𝑛𝑑 = 𝑊2
𝑈𝐶1 𝑈𝐶2
• Marginal rate of substitution – the rate at which a household is willing
to substitute (give up) leisure for additional consumption
• Marginal rate of exchange – the market’s rate at which a household
can exchange leisure for consumption
• The wage rate reflects the opportunity cost of leisure
Interpreting FOCs (3)

Combining (1), (2), and (3) can be summarized as

𝑉𝐿1 𝑊1
=𝛽 1+𝑅
𝑉𝐿2 𝑊2
Interpreting FOCs (4)
1
• 𝑊1 = 𝑊2 and 𝛽 > : The interest rate is above the time preference (or
(1+𝑅)
household’s discount factor above market’s discount factor ⇒
• Income from labor supply today can be saved at a rate of return exceeding the
household’s rate of time preference – 𝑁1 ↑ and 𝑁2 ↓
1
• 𝑊1 = 𝑊2 and 𝛽 < : The interest rate is below the time preference ⇒
(1+𝑅)
𝑁1 ↓ and 𝑁2 ↑
1
•𝛽= : Relative labor supply depends on relative wages
(1+𝑅)

• Labor supply is greater in periods with a high wage rate because the utility cost of
producing a given amount of income is lower
Frisch elasticity

• Frisch elasticity measures the substitution effect of labor supply with

respect to the wage, keeping marginal utility of wealth constant
𝛿(𝑁1 / 𝑁2 )
𝐹𝑟𝑖𝑠𝑐ℎ 𝑒𝑙𝑎𝑠𝑡𝑖𝑠𝑖𝑡𝑦 =
𝛿(𝑊1 / 𝑊2 )
• The higher the Frisch elasticity, the more willing households are to
work today if the wage increases
• The Frisch elasticity depends on the utility function
Solve for a specific utility function

• A common utility function in economics is the Constant Elasticity of

Substitution (CES) utility function
1
1+𝜂
𝑁
𝑈 𝐶, 𝑁 = 𝑙𝑛𝐶 − Ψ
1
1+
𝜂
• 𝜂 is Frisch elasticity – elasticity of substitution between labor supply
and wages
Firms decisions

• Two ways to describe the firm’s decision problem

• Profit maximization
• Cost minimization

• Simple case – static environment

• Perfect competition – no monopolistic (market) power
• No market rigidities – prices are fully flexible
Profit maximization
• Assumptions
• Y total production of output, number of output produced, with price P

• Two inputs
• K capital input, total amount of machinery, that must be paid a rental rate R
• N labor input, total hour of work (or measure of employment), that must be
paid wage W

• Perfect competition
• The firm takes prices P,R and W as given
Profit maximization (2)

• Firm maximizes profits Π (static problem)

Π = PY − RK − WN
• By choosing 𝑌, 𝐾, 𝑎𝑛𝑑 𝑁
max Π 𝑠𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 𝐹 𝐾, 𝑁 ≥ 𝑌
𝑌,𝐾,𝑁

• The Lagrangian
ℒ 𝑌, 𝐾, 𝑁, 𝜆 = 𝑃𝑌 − 𝑅𝐾 − 𝑊𝑁 + 𝜆 𝐹 𝑅, 𝑁 − 𝑌
Profit maximization (3)
• First order conditions (FOCs)
𝛿ℒ
=𝑃−𝜆 =0
𝛿𝑌
𝛿ℒ
= 𝜆𝐹𝐾 − 𝑅 = 0
𝛿𝐾
𝛿ℒ
= 𝜆𝐹𝑁 − 𝑊 = 0
𝛿𝑁
𝛿ℒ
= 𝐹 𝐾, 𝑁 − 𝑌 = 0
𝛿𝜆
Profit maximization (4)

• Can be simplified to
• 𝑃𝐹𝐾 = 𝑅
• 𝑃𝐹𝑁 = 𝑊

• The firm should keep hiring capital until it’s marginal revenue product
falls to the rental rate
• The firm should keep hiring labor until its marginal revenue product
(equal to P times the marginal product) falls to level of the wage
Combining the FOC:s
𝐹𝐾 𝑅
=
𝐹𝑁 𝑊
𝐹𝐾
• - 𝑀𝑅𝑇𝑆𝑁,𝐾 – Marginal Rate of Technical Substitution between capital and
𝐹𝑁
labor – the rate at which capital can be replaced by labor, keeping output
constant
• How many units of labor must be added to replace one fewer unit of capital, keeping output
constant
𝑅
• - 𝑀𝑅𝐸𝑁,𝐾 – Markets Rate of Exchange between capital and labor – the rate at
𝑊
which labor can be replaced by capital while maintaining the same cost
• The market’s "willingness" to substitute one factor for another
Setting-up the RBC model

• The basics of dynamic general equilibrium modeling

• Setting up the model
• Solving for the equilibrium conditions
• Solving for the steady state
• Solving for the dynamics
Households
• Decentralized market allocations
• Get utility from consumption and leisure
• Supply labor, earning a wage rate
• Own the capital stock and rent it to firms
• Earning a rental rate of capital each period
• Alternatively, firms own the capital stock
• In this model this yields the same solution, but not always the case, though
The households’ maximization problem
∞

max 𝐸𝑡 ෍ 𝛽𝑡 𝑈(𝐶𝑡 , 𝐿𝑡 )
𝐶𝑡 ,𝐿𝑡 ,𝑁𝑡 ,𝐾𝑡 ,𝐼𝑡
𝑡=𝑜
1 𝐶𝑡 + 𝐼𝑡 = 𝑅𝑡 𝐾𝑡−1 + 𝑊𝑡 𝑁𝑡
2 𝐿𝑡 + 𝑁𝑡 = 1
3 𝐾𝑡 = 𝐼𝑡 + 1 − 𝛿 𝐾𝑡−1
Substitute out investments to get one constraint less
1 + 3 ⇒ 𝐶𝑡 + 𝐾𝑡 − 1 − 𝛿 𝐾𝑡−1 = 𝑊𝑡 𝑁𝑡 + 𝑅𝑡 𝐾𝑡−1
The Lagranian

ℒ = 𝐸𝑡 ෍ 𝛽𝑡 [𝑈 𝐶𝑡 , 𝐿𝑡 +
𝑡=𝑜

+𝜆𝑡 𝑅𝑡 𝐾𝑡−1 + 𝑊𝑡 𝑁𝑡 − 𝐶𝑡 − 𝐾𝑡 + 1 − 𝛿 𝐾𝑡−1 +

+ 𝜇𝑡 (1 − 𝐿𝑡 − 𝑁𝑡 )]
The Lagranian (2)
∞

max 𝐸𝑡 ෍ 𝛽 𝑡 𝑈(𝐶𝑡 , 𝐿𝑡 )
𝐶𝑡 ,𝐿𝑡 ,𝑁𝑡 ,𝐾𝑡 ,𝐼𝑡
𝑡=𝑜

ℒ = 𝑈 𝐶𝑡 , 𝐿𝑡 +
+𝜆𝑡 𝑅𝑡 𝐾𝑡−1 + 𝑊𝑡 𝑁𝑡 − 𝐶𝑡 − 𝐾𝑡 + 1 − 𝛿 𝐾𝑡−1 +
+ 𝜇𝑡 1 − 𝐿𝑡 − 𝑁𝑡 +

+𝛽𝑈 𝐶𝑡+1 , 𝐿𝑡+1 +

+𝛽𝜆𝑡+1 𝑅𝑡+1 𝐾𝑡 + 𝑊𝑡+1 𝑁𝑡+1 − 𝐶𝑡+1 − 𝐾𝑡+1 + 1 − 𝛿 𝐾𝑡 +
+ 𝛽𝜇𝑡+1 1 − 𝐿𝑡+1 − 𝑁𝑡+1 + ⋯
Equilibrium conditions
𝛿ℒ
1 ∶ 𝑈𝐶𝑡 − 𝜆𝑡 = 0
𝛿𝐶𝑡
𝛿ℒ
2 ∶ 𝑈𝐿𝑡 − 𝜇𝑡 = 0
𝛿𝐿𝑡
𝛿ℒ
3 ∶ 𝜆𝑊𝑡 − 𝜇𝑡 = 0
𝛿𝑁𝑡
𝛿ℒ
4 ∶ −𝜆𝑡 + 𝛽𝜆𝑡+1 (𝑅𝑡+1 + 1 − 𝛿) = 0
𝛿𝐾𝑡
𝛿ℒ
5 ∶ 𝑊𝑡 𝑁𝑡 + 𝑅𝑡 𝐾𝑡−1 − 𝐶𝑡 − 𝐾𝑡 + 1 − 𝛿 𝐾𝑡−1 = 0
𝛿𝜆𝑡
𝛿ℒ
6 ∶ 1 − 𝐿𝑡 − 𝑁𝑡 = 0
𝛿𝜇𝑡
Substitute out the Lagrange multipliers
𝑈 𝐿𝑡
1 ∶ = 𝑊𝑡
𝑈𝐶𝑡
𝑈𝐶𝑡
2 ∶ = 𝑅𝑡+1 + 1 − 𝛿
𝛽𝑈𝐶𝑡+1

3 ∶ 𝑊𝑡 𝑁𝑡 + 𝑅𝑡 𝐾𝑡−1 = 𝐶𝑡 + 𝐾𝑡 − 1 − 𝛿 𝐾𝑡−1
4 ∶ 𝐿𝑡 + 𝑁𝑡 = 1
• Note equations (1) and (2) are the same conditions as in the two period
model
• 4 equations and 4 unknowns 𝐶𝑡 , 𝐿𝑡 , 𝑁𝑡 , 𝐾𝑡
Firms

• Firms are price takers and choose capital and labor to maximize
profits
• Production takes place within a large number of competitive firms and no
scale effects, can then be modeled as taking place in one representative
price-taking firm

• Firms rent capital from households at the rental rate of capital

The firms’ maximization problem
max Πt = 𝑌𝑡 − 𝑅𝑡 𝐾𝑡−1 − 𝑊𝑡 𝑁𝑡
𝐾𝑡−1 ,𝑁𝑡

Subject to
𝑌𝑡 = 𝐴𝑡 𝐹 𝐾𝑡−1 , 𝑁𝑡
The first order conditions are given by
𝐴𝑡 𝐹𝐾𝑡−1 = 𝑅𝑡
𝐴𝑡 𝐹𝑁𝑡 = 𝑊𝑡
2 equations and 2 unknowns 𝐾𝑡 , 𝑁𝑡
General equilibrium

8 equations and 8 unknowns 𝐶𝑡 , 𝐿𝑡 , 𝐾𝑡 , 𝑁𝑡 , 𝑌𝑡 , 𝐼𝑡 , 𝑅𝑡 , 𝑊𝑡

𝑈 𝐿𝑡
1 ∶ = 𝑊𝑡
𝑈𝐶𝑡 6 ∶ 𝐴𝑡 𝐹𝑁𝑡 = 𝑊𝑡
𝑈𝐶𝑡
2 ∶ = 𝑅𝑡+1 + 1 − 𝛿 7 ∶ 𝑌𝑡 = 𝐴𝑡 𝐹 𝐾𝑡−1 , 𝑁𝑡
𝛽𝑈𝐶𝑡+1

3 ∶ 𝑊𝑡 𝑁𝑡 + 𝑅𝑡 𝐾𝑡−1 = 𝐶𝑡 + 𝐼𝑡 8 ∶ 𝐾𝑡 = 𝐼𝑡 + 1 − 𝛿 𝐾𝑡−1
4 ∶ 𝐿𝑡 + 𝑁𝑡 = 1
5 ∶ 𝐴𝑡 𝐹𝐾𝑡−1 = 𝑅𝑡
The processes for shocks

• One more variable still remains undefined – 𝐴𝑡

• A single shock in our case – technology shock
• Need low of motion for the shock
𝑙𝑛𝐴𝑡 = 𝜌𝑙𝑛𝐴𝑡−1 + 𝜖𝑡
Solving for the steady state
• Solve for the steady state in which

• 𝐴𝑡 = 1 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑡

• … 𝑋𝑡−1 = 𝑋𝑡 = 𝑋𝑡+1 = ⋯ = 𝑋෠
where the bar denotes the steady-state value of a variable
• Need to assume specific utility and production functions
1
1+𝜂
𝑁𝑡
𝑈 𝐶𝑡 , 𝑁𝑡 = 𝑙𝑛𝐶𝑡 − Ψ
1 + 1/𝜂
𝛼
𝐴𝑡 𝐹 𝐾𝑡−1 , 𝑁𝑡 = 𝐴𝑡 𝐾𝑡−1 𝑁𝑡1−𝛼
The steady state

Substitute out leisure and note 𝐴መ = 1

෡ 1/𝜂 𝐶መ = 𝑊
1 ∶ Ψ𝑁 ෡ 5 ∶ 𝐶መ + 𝐼መ = 𝑊
෡𝑁෡ + 𝑅෠ 𝐾
෡

2 ∶
1
= 𝑅෠ + 1 − 𝛿 6 ∶ 𝑌෠ = 𝐴መ 𝐾
෡𝛼𝑁
෡ 1−𝛼
𝛽
7 ∶ 𝐼መ = 𝛿 𝐾
෡
𝑌෠
3 ∶ 𝛼෡ = 𝑅෠
𝐾
𝑌෠
4 ∶ (1 − 𝛼) ෡ ෡
=𝑊
𝑁
The steady state

• The steady state consists of a nonlinear system of 7 equations and 7

unknowns 𝐶,መ 𝑁,
෡ 𝐾,
෡ 𝑌,
෠ 𝐼,መ 𝑅,
෠ 𝑊෡

• To solve the nonlinear system, numerical methods are necessary in

basically all models of interest
Solving for the dynamics

• How the system reacts to shocks

• Need to transform nonlinear model to a linear format
• Log-linearization around a steady state
• Software does it for you
• We will skip the technical part for this course
Dynamic solution: Response of the system to
a 1 percent technology shock