NPTEL-Economics-Public Economics

Module 5
Lecture 22
Topics
5.8 Complication of Public Provision: Crowd Out
5.9 Optimal Second Best Provision of Public Goods
5.10 The Problem of Crowd-Out
5.11 Bergstrom, Blume, and Varian (1986) Model of CrowdOut
5.12 Private Responses to Public Provision: Partial Crowd-Out
5.13 Public Goods with Endogenous Private Provision
5.14 Empirical Evidence on Crowd-Out
5.15 Kingma 1989

5.8 COMPLICATION OF PUBLIC PROVISION: CROWD

OUT
In principle, the government could solve the optimal public goods provision
problem and then either provide the good directly or mandate individuals to
provide it.
In practice, three problems emerge:
Crowd-out.
Measuring costs and benefits.

Measuring future costs & benefits is even more difficult.

Determining the publics preferences.

1
Indian Institute Of Technology, Kanpur

NPTEL-Economics-Public Economics

In some cases, the private market may already be providing a socially inefficient
level of the private good.

5.9 OPTIMAL SECOND BEST PROVISION OF PUBLIC

GOODS
Suppose govt has decided to levy a tax and provide public goods based on some
rule
Two complications arise when trying to get to Samuelson First Best level:
Interactions with private sector provision

Public provision may crowd-out some of the private provisionas

the government provides more of the public good, the private
sector provides less.

Government cannot finance public goods through lump sum taxation

Must modify Samuelson rule to account for distortionary taxation.

5.10 THE PROBLEM OF CROWD-OUT

Full Crowd out: Conditions
People care only about the total number of goods provided: there is no
warm glow from giving
Provision of public goods will be financed by charging equal amounts to
each of them.
The government provides no more public goods than were being provided
privately beforehand.
Then each unit of public provision will crowd out private provision one-for-one.

5.11 PUBLIC GOODS WITH ENDOGENOUS PRIVATE

PROVISION
Interest in crowd-out began with Roberts (1984)
Expansion of govt services for the poor since Great Depression accompanied by
comparable decline in charitable giving for the poor.

2
Indian Institute Of Technology, Kanpur

NPTEL-Economics-Public Economics

Inference government has grown tremendously without having any net impact on
poverty or welfare
Evidence mainly based on time series impressions.
But theory underlying this claim very sensible, as subsequent work showed

5.12 BERGSTROM, BLUME, AND VARIAN (1986) MODEL

OF CROWD-OUT
Individual h solves:

s.t.
FOC:
Nash Equilibrium exists & is unique.
G such that all individuals optimize given others behaviour.
Let

denote private equilibrium outcome.

Now suppose government introduces lump sum tax of

on each individual h.

Revenue used to finance expenditure on public good

Individual optimization problem is now.

.
is private provision of public good.
s.t.
Let

Total contribution of individual h.

Then we can rewrite the above problem as

Max.
s.t.
This is isomorphic to the original problem

3
Indian Institute Of Technology, Kanpur

NPTEL-Economics-Public Economics

Total public good provision unchanged!

Each person simply reduces voluntary contribution by

5.13 PRIVATE RESPONSES TO PUBLIC PROVISION:

PARTIAL CROWD-OUT
Full crowd-out is rare, though partial crowd-out is much more common and can
occur when:
People who dont contribute to the public good are taxed to finance its
provision.
Contributors effective income levels are higher than before.
As a result of this income effect, contributors buy more if the public good is a
normal good, offsetting the crowd-out to some extent.
Or when individuals derive utility from their individual contributions as
well as the total amount of the public good provided.
Warm glow model offsets crowd out effect
Suppose there are 2 individuals A&B. A
contributor, B
non contributor.
Initially A contributes 10 unit, B nothing. Government imposed
lump sum tax of 5 units each & provides 10 units. Now A gets 10
at the cost of 5. This is an effective increase in As wealth.

5.14 EMPIRICAL EVIDENCE ON CROWD-OUT

Two empirical questions motivated by theory
How large is the degree of crowd-out in practice?
What are the income and price effects on charitable giving?
Two strands of empirical literature
Field evidence (observational studies)
How individual contributions respond to government spending: shows small
crowd out. But they suffer from endogeneity problems.
Lab experiments
4
Indian Institute Of Technology, Kanpur

NPTEL-Economics-Public Economics

Lab experiments suggest crowd out is large.

Lab experiments may not capture important motives for giving: warm glow,
prestige
Traditionally, lab experiments have been more influential but recent field studies
may change this

5.15 KINGMA 1989

Studies individual contributions to public radio stations
Cross-sectional survey of individuals who listened to public radio. 3,500
individuals and 63 different radio stations.
Research Design: OLS regression of individual contributions on government
support

Di = individual contribution
Gi = government support
Xi = set of controls: individual income, individual education, age, price (tax
bracket).
SUMMARY STATISTICS FOR PUBLIC RADIO LISTENERS
CONTRIBUTION
NONCONTRIBUTION
MEAN
STANDARD
MEAN
STANDARD
DEVIATION
DEVIATION
CONTRIBUTION($)
45
52
0
0
AGE(YEARS)
46
14
45
16
INCOME($)
48.074
27.586
37.582
24.818
EDUCATION(YEARS) 16.4
2.4
15.3
2.8
LISTENING(HOURS
10.75
11.25
6.00
8.00
PER WEEK)
D = -65.036 + .539(INCOME) - .010(R) - .015(G) 15.014(PRICE)
(3.38)
(6.49)
(.94)
(3.02)
(.74)
+10.018(EDUCTN) + .288(AGE)
(8.01)
(3.26)
Source: Kingma 1989

Table 22.1

5
Indian Institute Of Technology, Kanpur