Public Policy

• 4 questions of public finance:

1. When should the government intervene?
2. How might the government intervene?
3. What is the effect of those interventions?
4. Why do governments intervene in the way they do?

Theoretical Tools

• Set of tools designed to understand the mechanics behind economic decision making
• How do individuals choose how much to consume or how hard to work?
Constrained utility maximization
• How do firms choose how much to produce?
Production theory
• What is the theoretical effect of a redistribution policy on economic efficiency end equity?
Equilibrium and social welfare

Consumer Choice (Positive Approach)

Preferences and indifference curves
• Preferences: asking what people prefer, not what they can afford
• Indifference curve (IC): graphical representation of all bundles of goods making an individual
equally well off (all bundles provide equal utility, individual is indifferent as to which to consume)
Consumers prefer higher ICs (non-satiation; more is better)
ICs are always downward sloping (trade-off)

Utility Mapping of Preferences

• Utility function represents an individual’s set of preferences: U = f(X1,X2,X3,…)
X1,X2,X3,… are quantities of the goods consumed
• Marginal utility = additional increment to utility obtained by consuming an additional unit of a good
• Diminishing marginal utility = consumption of each additional unit makes the individual less happy
than the consumption of the previous one
• Marginal rate of substitution = rate at which consumer is willing to trade one good for another
MRS = slope of the IC (MRS = -MUM/MUC)
PM/PC < MRS: more utility from more movies/ fewer CDs
PM/PC > MRS: more utility from more CDs /fewer movies
• Budget constraint represents all combinations of goods an individual can afford to buy with entire
income
Y = PCQC + PMQM
Slope of the budget constraint (-PM/PC) = rate at which one can trade one good for the
other in the marketplace
• Opportunity cost = cost of any purchase is the next best alternative use of that money or the
forgone opportunity
1
 Price Effects: Substitution and Income Effects
• Substitution effect: holding utility constant, a price rise of a good will always cause an individual to
choose less of that good
• Income effect: rise in the price of a good will typically cause an individual to choose less of all goods
because her income can purchase less that before

Example: Labor Supply Effects of Welfare Programs

• TANF: monthly support check to families with income below a threshold level that is set by each
state
• Suggested policy reform: cutting the TANF benefits
• Policy question: Labor supply effects for single mothers?
• Example of a single mother who spends all her earnings and benefits on food & prefers time at
home to time at work
• Setting:
Max. available hours of leisure per year = 2.000
Hourly wage = $10 (for simplicity PFood = $1)
TANF benefits = $5.000 which are reduced by 50% for every $ earned (benefit reduction
rate/implicit tax)
• Policy impact/ labor supply effects depend on:
Pre-reform labor supply
Relative preferences for food and leisure
The resulting income and the substitution effect

Equilibrium and Social Welfare (Normative Approach)

• Welfare economics = study of the determinants of welfare in society
Social efficiency = net benefits that consumers and producers receive because of their trade
of goods and services
• Market equilibrium = combination of price and quantity that satisfies demand & supply
(determined by interaction of supply and demand curves)

Demand Curves

• Elasticity of demand = % change in the quantity demanded of a good caused by each 1% change in
the price of that good
!"#$"%&'(" $*'%(" +% ,-'%&+&. /"0'%/"/ 12/2
𝜀= !"#$"%&'(" $*'%(" +% !#+$"
= 14/4
• Perfectly inelastic demand:
Elasticity of demand is zero
Quantity demanded doesn’t change when prices rise
Demand curve is vertical
2
• Perfectly elastic demand
Elasticity of demand is infinite
Quantity demanded changes infinitely for even a small change in price
Demand curve is horizontal

Supply Curves
• Outcome of profit maximization by firms
• Total production: production function: q = √𝐾 × 𝐿
• Marginal productivity = impact of a unit change in any input, holding other inputs constant, on the
5, 5,
firm’s output (56 ; 57 )
• Total cost: C = rK + wL
• Marginal cost = incremental cost to a firm to produce one more unit of q
• Profit = Revenue – Cost -> Profit = pq – rK – wL
Profit maximized when marginal revenue equals marginal cost

Equilibrium
• Market-level supply and demand curves interact to determine the competitive market equilibrium
-> maximizes social efficiency

Social Efficiency
• Net gains to society from all trades made in market / total social surplus
• Consumer surplus = benefit that consumers derive from consuming a good (demand curve =
willingness to pay)
Determined by market (equilibrium) price and elasticity of demand
• Producer surplus = benefit producers derive from selling a good (supply curve = marginal cost)
Determined by market (equilibrium) price and marginal costs
• Competitive equilibrium where supply = demand, maximizes social efficiency
• Deadweight loss = reduction in social efficiency from preventing trades for which benefits exceed
costs

Social welfare
• Level of well-being in society, determined by how much gets produced (efficiency) and how it’s
distributed (equity)
• Society can attain any socially efficient outcome by suitably redistributing resources among
individuals & then allowing them to freely trade
Difficult in practice
• Equity-efficiency trade-off = choice society must make between the total size of the economic pie
and its distribution among individuals
• Social Welfare Function (SWF): combines the utility functions of all individuals
Utilitarian social welfare function maximizes the sum of individual utility: SWFU = U1 + U2
+…+ UN
Rawlsian social welfare function maximizes the utility of the worst-off member of society:
SWFR = min(U1, U2,…,UN)
3
 • Social welfare functions reflect different possible equity criteria, including
Commodity egalitarianism: principle that society should ensure that individuals meet a set
of basic needs, but beyond that point income distribution is irrelevant
Equality of opportunity: principle that society should ensure that all individuals have equal
opportunities for success but not focus on the outcomes of choices made

Example: Labor Supply Effects of Welfare Programs

• Policy suggestion: cutting TANF benefits
Gain in social efficiency
• But
Citizens also care about equity
Social welfare programs may decrease social efficiency but increase social welfare

Empirical Tools

Causation vs. Correlation

• Correlation: two economic variables are correlated if they move together
• Causality: two economic variables are causally related if the movement of one causes movement of
the other
• Correlation between A and B can have 3 explanations:
A is causing B (causality)
B is causing A (reverse causality)
Some third factor C is causing both (omitted variables)
• Correlation alone does not imply causality

Randomized Control Trial (RCT)

• Random assignment rules out reverse causation and omitted variable
• Any difference between treatment and control group must be due to treatment
• RCTs are “gold standard” for determining causality

Observational Data
• Usually RCTs are not available/possible
• Data generated by individual behavior observed in real life
Challenge: bias
Solution: statistical methods allow us to approach the gold standard of randomized trials

Time Series Data

• = same entity at different points in time
• Time series analysis = analysis of the co-movement of two series over time
To identify & measure correlation -> often produces striking patterns
Excluded variables may be driving the results (esp. macro.- & wage-subsidy programs)

4
Cross Sectional Data
• = many entities at a given point in time
• Cross-sectional regression analysis = statistical method of the relationship between two or more
variables exhibited by many individuals at one point in time
• Regression analysis = finds best-fitting linear relationship between two variables
• Regression line = measures best linear approximation to the relationship between any two
variables
• 𝑦 = 𝛽8 + 𝛽9 𝑥 + 𝑢 (𝑦: dependent variable, 𝑥: independent variable, 𝑢: residual/error term)
∆𝑦 = 𝛽9 ∆𝑥 (if ∆𝑢 = 0): 𝛽9 = slope (change in 𝑦 if 𝑥 increases by 1), 𝛽8 = intercept
Assumption: Ε[𝑢|𝑥] = 0 (𝑢 and 𝑥 are independent/uncorrelated)
• OLS
• Control variables = are included in cross-sectional regression models to account for differences
between treatment and control groups that can lead to bias)
Challenges: unobserved confounding variables, immeasurable

Quasi-Experiments
• Panel data = many entities at different points in time
• Changes in economic environment that create nearly identical treatment and control groups for
studying the effect of that environmental change (allows us to take advantage of randomization
created by external forces)
• Difference-in-difference design = difference between changes in outcomes for the treatment group
that experiences an intervention and the control group that does not
• Dummy variable = variable indicating the belonging to a category
1 = belonging to a category, 0 = not belonging to a category
• 𝑂𝑢𝑡𝑐𝑜𝑚𝑒 = 𝛼 + 𝛽 𝑝𝑜𝑠𝑡 + 𝛾 𝑡𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡 𝑔𝑟𝑜𝑢𝑝 + 𝛿 𝑝𝑜𝑠𝑡 × 𝑡𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡 𝑔𝑟𝑜𝑢𝑝 + 𝜈
𝛼 = pre-mean in control region
𝛽 = Time difference in control region (trend)
𝛾 = Regional difference pre
𝛿 = Deviation from the trend (reform effect)
• Problems
Underlying assumption: certainty that we purged all bias from the treatment-control
comparison
Line of defense: intuitive approach = given the experiment, most of the bias has been
removed, statistical approach = use alternative or additional groups to confirm (placebo)
Interpretation: experiments give reduced form impact of some policy but don’t explain why
it works
Reduced form estimates = measures of impact of an independent variable on a dependent
variable
Structural estimates = estimates of the features that drive individual decisions such as
income and substitution effects or utility parameters

5
Behavioral Insights in Public Policy

• About our mind

System 1: fluent, spontaneous, fast = automatic
System 2: effortful, reflective, slow = deliberative
• About our emotions
Loss aversion: losses loom larger that gains
Example: a causal effect of negative shocks on stress levels
• About our behavior (jumping the gun: when causes trump probabilities)
representativeness
§ judgements of probability are masked by judgements based on similarity
§ happens due to insensitivity towards prior probability of outcomes (base-rate
frequency), insensitivity towards sample sizes (law of small numbers), misconception
of chance & validity, insensitivity to predictability
base rate ignorance
causal interpretations are more appealing (than statistics and facts)
• similar concepts:
anchoring: people begin process of estimation with whatever information (anchor) readily
appears in their minds
adjustment: people adjust initial answer away from anchor to get their final answer (but
the final conclusion is biased towards initial anchor)
priming with anchors: possible to influence your choice or answer by suggesting a starting
point (anchor)
availability heuristic: people make judgements based on how easily an example, instance or
case comes to mind
affect heuristic: represents a reliance on good or bad feelings experienced in relation to a
stimulus
• reasons for failing policies:
Corruption: Food stamps (example)
§ Moral hazard & free riders, Information asymmetry, trust-mistrust issues
Lack of institutional support: introducing computers in a school with frequent electricity
cuts or unskilled trainers
§ Imprecise estimates of the infrastructure, people may lie on their CVs
• Randomized evaluations/ RCTs:
+ takes out opportunity cost of blindly implementing a policy
+ learning about the benefits & shortcomings of a policy
- expensive to experiment at scale
- evaluations take time, esp. long run effects

6
Externalities

• Definition: externalities arise whenever the actions of one party make another party worse (better)
off, yet the first party doesn’t bear the costs (or receives the benefits)
They can arise from production or consumption; can be negative or positive
• Market failure is a problem that causes the economy to deliver an outcome that doesn’t maximize
efficiency -> government interventions are justified

Externality Theory
• Social-welfare-maximizing equilibrium: Social Marginal Costs = Social Marginal Benefits
• The market sets: Private Marginal Costs = Private Marginal Benefits
• In the absence of externalities (if Marginal Damage = 0): PMC = SMC and PMB = SMB the market
equilibrium is efficient
• In the presence of production/consumption externalities the market equilibrium is inefficient
• Negative production externality:
when a firm’s production reduces the well-being of others who are not compensated by the
firm
drives a wedge between private & social marginal cost
§ private marginal cost (PMC): direct cost to producers of producing an additional unit
of a good
§ marginal damage (MD): costs associated with the production of the good that are
imposed on others
§ social marginal cost (SMC) = PMC + M

• Negative consumption externality:

when an individual’s consumption reduces the well-being of others who aren’t
compensated by the individual
drives a wedge between private & social marginal benefit
§ private marginal benefit (PMB): direct benefit to consumers of consuming an
additional unit of a good
§ social marginal benefit (SMB) = PMB – MD

7
 • positive production externality:
when a firms production increases the well-being of others but isn’t compensated by those
• positive consumption externality:
when an individual’s consumption increases the well-being of others but isn’t
compensated
• quick hints:
Negative production externality: SMC curve lies above PMC curve
Negative consumption externality: SMC curve lies below PMC curve
positive production externality: SMB curve lies below PMB curve
positive consumption externality: SMB curve lies above PMB curve

Private Sector Solution

• externalities undermine efficiency
• Ronald Coase: Why won’t the market simply compensate the affected parties for externalities?
Solution: Internalizing the externality
• Coase theorem (Part I):
When there are well-defined property rights & costless bargaining, then negotiations
between the party creating the externality & the one that’s affected can bring about the
socially optimal market quantity
§ Government intervention to establish property rights
• Coase theorem (Part II):
Efficient solution to an externality doesn’t depend on which party is assigned property
rights
• Problems:
Assignment: who is to blame?
Holdout: shared ownership gives each owner veto power over the others
Free rider: when investment has personal cost but common benefit, individuals will
underinvest
Transaction costs & negotiating problems: hard to negotiate with too many individuals on
one or both sides
Coase theorem won’t help with large-scale, global externalities

Public-Sector Remedies
• Three types of remedies to resolve the problems associated with negative externalities
Corrective taxation to discourage usage
Subsidies to encourage usage
Regulation to directly change usage
• Corrective taxes (& subsidies):
Change private marginal cost (benefit) without affecting the social marginal cost (benefit)
Can be used to internalize the externality
Called “Pigouvian taxation”

8
 • Regulation:
In an ideal world Pigouvian taxation equals regulation
Regulation (quantity approach): traditional choice for addressing externalities
Taxes (price approach): more effective way for addressing externalities

Price vs. Quantity – Taxes vs. Regulation

• Goal: find most efficient path to meet the targets (focus only on efficiency)
• Taxes:
If price (tax) = MD, firms internalize MD & pollute until SMC = PMC
Taxation requires to only know MD
• Regulation:
Efficient solution is for SMB = SMC & SMC = PMC
Regulation requires knowing whole SMC curve
• Price regulation (taxes) vs. quantity regulation:
Taxes raise private marginal costs to social marginal costs & cause efficient production
Taxes give plants more flexibility in choosing their optimal amount of reduction
Quantity regulation ignores the fact that the plants have different marginal costs of
pollution reduction
• Policies for multiple plants with different production costs:
Quantity regulation: marginal cost of reducing pollution is equal to social marginal benefit
of that reduction (for each plant) -> inefficient & too expensive
Corrective tax: Pigouvian taxes cause efficient production by raising cost of input by size of
external damage
Quantity regulation with tradable permits: allows market to incorporate differences in cost
of pollution reduction across firms -> Coasian solution
• If costs are high, regulation could be expensive; taxation avoids this problem since firms will adjust
until cost of adjustment (tax) but if costs are uncertain, so is the amount of pollution reduction that
a tax achieves
Taxes lead to lower costs but less control over amount of pollution reduction
Choice of instrument of the government depends on whether the government wants to
reduce amount of pollution or wants to minimize costs
• Quantity regulation ensures the right amount regardless of costs
• Price regulation ensures most efficient, means getting quantity right isn’t crucial

9
Public Goods

Public Goods – A Taxonomy

• Pure public goods: non-rival in consumption & non-excludable
• Impure public goods: satisfy the two public good conditions to some extent but not fully

Optimal Provision of Public Goods

:;!" ? @ 4
• Optimality condition for the consumption of private goods: "
:;#!
= 𝑀𝑅𝑆<=,= = 𝑀𝑅𝑆<=,= = 4!
#!

• Private goods:
Optimality condition on the supply side: 𝑀𝐶<= = 𝑃<= (if 𝑃= = 1)
? @
Equilibrium: marginal cost = marginal benefit (𝑀𝐶<= = 𝑀𝑅𝑆<=,= 𝑜𝑟 𝑀𝑅𝑆<=,= )

• Public goods:
Non-rival: ones utility from research doesn’t reduce the others
? @
Social-efficiency-maximizing quantity solves 𝑀𝑅𝑆A"B"'#$* + 𝑀𝑅𝑆A"B"'#$* = 𝑀𝐶
Social efficiency is maximized when marginal cost is set equal to sum of MRSs

Private Provision of Public Goods

• Market doesn’t produce efficient amount of public goods due to the free rider problem
Examples: radio & tv program, Wikipedia
• Private markets can provide public goods when:
Private suppliers are given the ability to overcome the problem of non-excludability (e.g. by
charging user fees)
Some individuals care more than others
People are altruistic: individuals value benefits & costs to others in making their
consumption choices
There’s a high amount of social capital: value of altruistic & communal behavior in society
People might simply feel good about contributing

10
Challenges of Public Provision
• Government provision potentially solves the problem of non-contributors
• Difficulties:
Crowd-out: as government provides more, the private sector will provide less
Warm glow: people contribute even though government is already providing
• Modes of government provision:
Entire provision by public sector
Subsidized or mandated private provision (public incentives)
Contracting out (mixture of private & public provision): government retains responsibility
but hires private-sector firms -> problems: different incentives, no competitive bidding
procedure

Public Provision of Public Goods and its Difficulties

• Policy design: identification of optimal provision of public goods needs the knowledge of the MRS
of each individual
Costs, benefits, individual preferences

Cost-Benefit Analysis

• Government interventions have costs & benefits and any public investment requires a cost-benefit
analysis
Monetary & non-monetary costs
Direct & indirect benefits

Measuring Current Costs

• Cash-flow accounting: calculates costs solely by adding up what the government pays for inputs to
a project & calculates benefits solely by adding up income or government revenues generated by
the project
Doesn’t correspond to social marginal costs
• Cash costs arising to the government:
Opportunity costs in the case of imperfect markets
Rents

Measuring Future Costs

• Opportunity costs to society with a new wrinkle
Associate future stream of costs with the one-time costs associated with capital and
operation
Presented discounted value (PDV) using the social discount rate
§ PDV = a dollar next year is worth 1+r times less than a dollar now because it could
earn r% interest if invested

11
 Measuring Benefits
• Difficult because one must place value on intangible benefits
• Market-based measures: (use wages as a measure for the value of saved time)
Assumptions: individuals spend saved time at work; perfectly competitive labor market
with freely adjusting hours
Problems: work hours aren’t necessarily chosen freely, non-monetary aspects of the job
• contingent evaluation to measure value of time
asking individuals to value an option they aren’t choosing = hypothetical questions
problems: isolation of issues, order of issues matters, embedding effect shouldn’t matter
• revealed preferences to measure the value of time: (letting actions of individuals reveal their
valuation)
if people are willing to pay x for something, it’s at least worth x to them

Discounting Future Benefits

• benefits are mostly long-term compared to costs
• cost-benefit analysis must discount these future benefits
• choosing proper discount rate is difficult:
present value of future benefits is very sensitive to discount rate
benefits to current & future generation

Cost-Effectiveness Analysis
• for projects with immeasurable benefits: only count costs and choose most effective based on
them

Issues
• common counting mistakes:
counting secondary benefits
counting labor as benefit
double-counting benefits
• distributional concerns:
costs & benefits may not go to the same people; choice of social welfare function matters
• uncertainty
costs & benefits are often highly uncertain

12

Powered by TCPDF (www.tcpdf.org)