Campaign Contributions as a Commitment Device€¦ · *Eco nomics PhD student, University of Califor ia Los Angeles. 1. 1. INTRODUCTION Why do campaign contributions by special interests

Campaign Contributions as a Commitment Device.

Zacharias Maniadis*

February 2006

Abstract: The paper explores the idea that the influence of special interest groups -

especially firms - on elections may have positive economic effects. This is the case when

contributions tend to restraint the scope of opportunism by the governing party that may,

in general, renege on its promises for economic stability and choose excessively leftist

policies. This is prevented if the private sector in the political game gets to move after the

policy is chosen, contributing to the governing party or to its rivals. Anticipating this, the

governing party will choose not to follow policies that will harm strong corporate groups.

This might be beneficial for the economy as a whole if the investment decisions of firms

determine economic growth and employment.

Keywords: Campaign Contributions, Probabilistic Voting, Policy Commitment.

*Economics PhD student, University of California Los Angeles.

1

1. INTRODUCTION

Why do campaign contributions by special interests exist? Do they serve any social

purpose or they just serve the interests of powerful groups that impose them on a society

that dislikes them? If this institution contributes to social efficiency by helping enforce

policy rules and promises then its existence can be economically rationalized. This paper

attempts to examine a novel channel through which contributions may affect economic

efficiency. Contributions may serve as a commitment devise that helps keep control over

the expectations of the private sector about economic policy, especially with respect to

important macroeconomic indices. The basic argument of this paper is that society as a

whole may benefit from this institution if it helps solve dynamic inconsistency problems

and induce investments

In the recent years the role of campaign contributions has been extensively discussed in

the United States and many types of campaign finance reform have been proposed.

Political scientists, economists and other social scientists have been examining the

economic and social effects of campaign contributions. For example, Levitt (1995) refers

to three main criticisms of the system of congressional campaign finance in the United

States at that time. Firstly, fundraising is an important activity for candidates that requires

2

too many resources, especially in terms of the time constraints of politicians, hence they

may not been able to carry out their more important tasks. Second, it is argued that the

system of contributions and fundraising may be biased towards incumbents. Thirdly, an

important consideration is whether organized interest groups exert excessive influence on

politics. To these arguments one may add that the system is may be biased towards right-

wing candidates, since the majority of special interest groups are thought to relate to the

corporate sector.

Intimately connected with these issues are considerations about the relationship between

regulated private campaign contributions, as an institution, and economic efficiency. For

example, an additional criticism asserts that increasing campaign money, after some level

of expenditure has been made, has no important effect on social welfare; therefore there

is a waste of resources. The literature on the efficiency of the institution of private

contributions has been increasing because of the large pressures for legislative change in

the US, informed by the idea that money plays an excessively large role in American

politics. In addition to the many objections to private campaign contributions, some

advantages have been proposed. In particular, it has been argued that contributions may

inform voters about the quality of candidates or their exact positions in the political

spectrum. This can be done with two ways: either political advertising is directly

informative of the qualities of politicians1, or it signals a hidden ability of a candidate that

the voters do not observe but the interest groups do2. This is welfare-increasing if the

choice of the better candidate is induced. But the arguments of the first type do not the

1 The papers by Austen-Smith (1987) and Coate (2001) are representative of this literature. 2 See Prat (1999).

3

basic question namely, why society tolerates special interest contributions, because the

perceived benefits of information come from campaigning in general. Hence, it may be

argued that generous public funding of campaigning is the optimal solution, since this

will allow for the benefits of informational advertising without the perceived negative

effect of shifting policies in favor of the interest groups that finance theses campaigns.

However, arguments of the signaling literature, and in particular the campaign

advertisement-signaling model developed by Prat, do offer a rationale for the existence of

private contributions.

In part two the concept of “mechanisms of commitment” is introduced and the

relationship between our model and the existing literature is discussed. In part three the

setting of the formal model is introduced. The equilibrium of the model is discussed in

part four. In part five the equilibrium is compared to the one in the model is extended

with an explicit commitment possibility. Part six discusses the assumptions and the

results of the model and its extensions. Part seven concludes.

2. MECHANISMS OF COMMITMENT AND RELATED LITERATURE.

In their famous paper, “Rules Rather than Discretion: The Inconsistency of Optimal

Plans” (1977) Kydland and Prescott underline the importance of policy rules that are

unalienable except under very extreme conditions. This importance stems from the well-

known problem of time inconsistency that occurs if policy-making is a dynamic process.

If the optimal current decisions of agents depend on future policy choices of the

4

policymaker then the so-called consistent policy may not be the optimal one. (The

consistent policy maximizes expected social welfare at any point in time taking into

account the future effects of policy). If this idea is true in real economic policy, then

policy rules and commitment are important. Hence, one may be interested examining

those specific institutions that ensure that policy rules are enforced3. One very important

example of such an institution is the independent central bank with a “conservative

director”. If monetary policy is out of the control of politicians then this may reduce

expected inflation because the conservative director earns no benefit by causing

unexpected inflation.

In this paper attention will be focused mainly on policies affecting the returns to capital

and possible commitments referring to these policies. In particular, the objective is to

examine an institution that may ensure that policy rules (for example, pertaining to

capital taxation) will not be violated or circumvented by opportunistic policymakers. This

institution is a legal framework allowing for corporate campaign contributions. In the

existence of this institution, if policymakers choose opportunistic policies they are

penalized by significantly undermining his party’s reelection prospects4. Furthermore, it

is important that society, as a whole, including people who earn mostly labor income,

may benefit from the establishment of this institution. This is despite the fact that the

resulting corporate influence in elections may militate against the choice of a labor-

3 It is worth noting that commitment may not necessarily require the existence of an institution. Reputation-building might be enough to ensure that politicians will not behave opportunistically. The literature on reputation is large. See, for example, Person and Tabellini, pp. 314-315. 4 This is of course based on a simplistic idea of special interest politics. What does good or bad policy mean for a specific firm? We assume that certain economic policies, such as those that affect macroeconomic stability and capital taxation more or less affect all firms similarly.

5

friendly policy. The reason is that in the long run they will benefit enough from the

higher level of investments in the economy, which is attained by enforcing the

commitment to a more capital-friendly policy.

This paradox resembles the one of strategic delegation, which is discussed in Person

and Tabellini (1994). This notion refers to the electoral support by some voters of

candidates that may not share their preferences about policy. This can be the case when,

for example, the elections cannot be won by candidates that share the preferences of these

voters. A related model with similarities with our model is presented in Person and

Tabellini (1994) pp. 318-323. This is a model with Citizen-Candidates, that is, where

candidates themselves are affected by the policy they choose. Here, middle-income

voters may vote for candidates that would protect the profitability of capital more than

they themselves would like to. This is because after elections take place, capital

accumulation decisions are made on the basis of predictions about future policy, enacted

in stage three. These decisions affect the welfare of all, as in our model. Thus, Person and

Tabellini also view this seemingly paradoxical mechanism as enforcing the society’s

commitment on policies that induce capital investments.

The notion that wage earners may like an institution that protects the rights of capital has

been examined in the political science literature. This is closely related with the idea of

“structural dependence” of democratic governments on capital. This view claims that the

policy-making of a modern democratic state is structurally constrained; this is because

investment decisions of wealth holders affect the future economic conditions for the

6

economy as a whole5. Therefore governments have to take into account of the effect of

their policies on investments and growth and voters realize this. Przeworsky and

Wallerstein (1988) introduce and test the idea of structural dependence using a formal

model. They show that without effective government intervention the wage earners are

constrained in their demands. They also show how a tax on consumption of profit-earners

can relax this constraint, invalidating “structural dependence”. Yet, in the dynamic setting,

where expectations matter, governments are constrained for the usual reasons of

credibility of promises for capital accumulation. This last conclusion is very similar to

the results of “equilibrium with no commitment” here. However, the authors do not

discuss any particular institutions that may result from the problem of “structural

dependence”.

The literature on campaign contributions is also large. In terms of its structure, our model

has similarities with the model of Snyder and Ting (2005). They also use a model with

voters (a representative voter) interest groups and parties to examine the importance of

elections as a means to control the performance of politicians. However, their focus is

mainly in comparing the incentive to control performance versus the incentive to elect

good types of politicians. It is interesting to note that they use an alternative assumption

regarding the effect of a contribution or “a bribe”. A “bribe” from the interest group

directly increases the utility of the party, whereas in our model it only affects the

probabilities of reelection.

5 See Przeworsky and Wallerstein, (1988).

7

It has already been said that the welfare effects of the institution of campaign

contributions have been examined in the literature. One important strand of the literature

examines informative advertising. These papers assume that the money spent on

campaigning may promote advertising that in fact increases the knowledge of voters

about the candidates’ abilities or positions. Austen-Smith (1987) is the first do tries to

explain the existence of campaign contributions assuming informative advertising.

A different approach has been to consider campaign advertising as a signal. Pratt (1999)

develops a formal model that assumes that the valence of candidates is more known to

the special interest groups than to voters. Accordingly, these groups are more likely to

contribute to the better candidate since they know that theses candidates are more likely

to win. Allowing campaign contributions may be efficient if the benefits of this

information for society exceed the costs of distorting the political promises of candidates.

Hence, this model derives a rationalization of the institution of campaign contributions by

special interest groups, unlike the directly informative advertising models. Our model

also offers a natural explanation for this institution in terms of efficiency.

3. THE SETTING OF THE MODEL

The main ideas are analyzed in a simple model with investment decisions, policy choices

and elections. We show that if corporate campaign contributions are not institutionalized

– in which case they are illegal and we assume they do not exist - then the time

8

inconsistency problem makes the incumbent party choose a labor-friendly policy. This is

because a very large portion of the constituency belongs to the group of wage earners and

prefers such a policy. Anticipating this, firms do not invest and all are worse off. If the

government could commit to the capital-friendly policy this would improve social

welfare but without legal campaign contributions the enforcement of this commitment is

not possible. Consequently, voters accept the existence of this institution because it

makes them better off.

The Players and the Pure Strategy Spaces

There are two parties, an incumbent party ( I ) and a challenger party ( ). There are two

groups of voters, the middle class (

C

M ) and the rich class ( R ), each of which has a

continuum of voters. Finally, there is the firm sector ( ). PlayerC , the challenger party,

never gets to move in our model but is used for expositional reasons.

F

This is a simple four-stage model. In stage zero, voters vote whether to accept or reject

the existence of the institution of private campaign contributions. We shall explain later

what this institution exactly does. The pure strategy space of voter j in stage zero is ,a r ,

where denotes accepting campaign contributions and rejecting them. In stage one the

firm sector decides whether to choose a high or low level of investment. Denote with

a r

x the level of investment, where x h= means that the investment is high and x l= means

that the investment is low. The firm sector at that point knows the choice of voters at

stage zero, so that the pure strategy space of in stage one isF , , ,hh hl lh ll , where for

9

example, is the strategy “choose low investment” if was the majority decision in

stage zero and “choose high investment” if was decided by the majority.

lh a

r

In stage two the incumbent party I decides whether to implement labor-friendly policy or

employer-friendly policy. Denote by s the policy choice, where s L= means labor-

friendly policy and means employer-friendly policy. Since the party knows both the

voters’ choice and the firms’ choice at the previous stages, the pure-strategy space for the

incumbent party is the set of all mappings of the form

s E=

: , , ,f a r h l L E× → . In stage

three, elections take place and voters decide if they vote for the incumbent or the

challenger party. Notice that as will be explained later, voters differ within each group

and among groups. So any pure strategy equilibrium must specify a pure strategy for each

voter of the two groups. Accordingly, the pure strategy space of voter j is the space of all

mappings of the form : , , , ,I Cq a r h l L E V V→× × . All equilibria we will find are

pure strategy equilibria.

The Payoff Functions

Firms: the payoffs of firms are their profits, realized in stage two. They depend on

whether they invested or not and on the policy choice of the incumbent party. Let

),( xsπ be the profit function of firms. The critical property of this function is the

following:

Assumption 1: ),(),( hLlL ππ > and ),(),( lEhE ππ >

10

This says that if the policy is labor-friendly the firm sector is better off having invested

low and if the policy is capital-friendly the opposite is true. This seems reasonable given

that investments entail some fixed costs and increase productive capacity. The function

π incorporates these costs here. If the variable costs of production are high enough, then

the optimal choice of the firm sector is not to produce a large quantity. Thus, fixed costs

cannot be covered and the firm is making loses. It is a logical assumption that variable

costs depend on minimum wages, insurance regulations, capital taxation and more

parameters that are incorporated in the policies E and . L

Parties: the payoff of the two parties is a fixed amount Ω that they get if they are elected

in stage three. They get zero if they are not elected. We assume that the utility from

choosing any level of policy in stage two is zero. In other words, parties have no

preference for any particular policy. This assumption is not necessary for the results.

Voters: the payoffs for voters are different between groups as well as within groups. They

are additive and they depend on which of the two parties gets elected.

For agent j in the middle-class group, the utility function is:

( , , , ) ( , ) ( ) ( )M MjM jM jMu s x v s x p s g xσ δ σ= + + + δ+ , if I wins.

( , , , ) ( , )MjM jMu s x v s xσ δ = , if C wins.

For agent k in the rich group, the utility function is:

( , , , ) ( , ) ( ) ( )R RkR kR kRu s x v s x p s g xσ δ σ= + + + δ+ , if I wins.

( , , , ) ( , )RkR kRu s x v s xσ δ = , if C wins.

11

To briefly present these terms, is the payoff function of a voter in group

over policies and the investment level in the economy (this payoff is realized in stage

two). The random parameter

( , ), ,iv s x i M R=

i

δ is the general popularity of party I relative to partyC ,

which cannot be discovered before the elections, and jMσ and kRσ show the ideological

preferences of agents j and k respectively. The functions ( ), ,ip s i M R= represent

satisfaction of voters with respect to the incumbent coming from their individual

economic condition, and the function represents the satisfaction coming from the

general condition of the economy

( )g x

6.

Let’s now explain in detail what these different terms mean and their important properties.

First of all, the utility for the middle class voters depends on current policy and

investment according to the payoff function and the corresponding function of

rich class voters is . The important thing here is that this term does not depend on

who gets elected: it is simply the realized payoff in stage two. This term therefore does

not affect the elections but it does affect the welfare properties of equilibria.

( , )Mv s x

( , )Rv s x

Assumption 2: for any . ( , )iv s h > ( , )iv s l s

This says that that for all voters the situation where investment is high is preferable that

the one where investment is low regardless of the policy chosen. This means that even

middle income voters would be better of in a thriving economy with capital-friendly

policy than in a shrinking economy with a labor-friendly policy. This is reasonable if we

6 This is justified if we consider the effect of the general economic conditions in the popularity of incumbents according to the models of retrospective voting. There is much evidence that shows that voters punish the incumbent party both for bad macroeconomic performance and individual low income in a retrospective manner: see Kramer (1971) and the surveys by Monroe (1979) and by Kiewiet and Rivers (1984).

12

consider the evidence in many countries where excessively leftist policies have the

average worker worse off by discouraging investments.

The functions ( ), ,ip s i M R= capture the fact that voters seek to discipline the incumbent

party for policies that affect their individual economic condition. For simplicity we

assume that:

( )Mp E = −z and ( ) 0Mp L z= >

( )Rp L = −w and ( ) 0Rp E w= >

This simply says that the middle group voters have an incentive to reward the incumbent

for choosing labor-friendly policy that improves their economic well-being, and the rich

group voters to reward the opposite policy. The symmetry assumption is only for

convenience. The function reflects the tendency of all voters to reward the

incumbent for achieving general prosperity in the economy. Again for simplicity, we

assume that

( )g x

( )g l µ= − and ( ) 0g h µ= > .

The importance of jMσ , kRσ is in capturing individual heterogeneity. Voter has a

specific individual preference for one of the two parties that does not depend on the

expected policy of the two parties. This might be due to ideological preference or due to

preference over a policy of the two parties that is fixed. This preference is represented by

the individual parameter

j

jMσ for voter in groupj M , and kRσ for voter in groupk R .

The distribution of these preference parameters in a specific group is uniform. The

distributions in the two groups differ, however. That is, jMσ follows the uniform

13

distribution in 2

1,2

1MM ϕϕ

− and kRσ follows the uniform distribution in 2

1,2

1RR ϕϕ

− .

The reason for this difference is that one social group may be more ideologically

homogeneous that the other, and jMσ , kRσ are the ideological homogeneity parameter.

The use of these parameters helps smooth the results and to understand the importance of

ideological homogeneity. Finally, the random parameterδ (remember that it represents

the general popularity of the incumbent party relative to the challenger party) follows the

uniform distribution in 1 1,2 2θ θ

⎧−⎨⎩ ⎭

⎫⎬

7. The realization of δ can be affected by random

elements of the political process, such as performance the final debate between the

political leaders.

The Institution of Campaign Contributions

To capture this institution without complicating the analysis too much, assume that the

firm sector can finance the campaign of the incumbent or the challenger party – a

decision that depends on the policy choice of the incumbent party. What the firm sector

wants is the choice of policy in stage two. Thus, it is safe to assume that the firms

convey the message to the incumbent party that if it chooses a policy they will

contribute to party and if it chooses policy they will contribute to

E

L

C E I . To provide this

kind of incentives is a weakly dominant strategy for , but we shall not discuss the F

7 This parameter increases the utility of all voters if the specific candidate is elected, so it appears in the utility function.

14

choice of the optimum contribution scheme here8. For simplicity, assume the aggregate

popularity of the incumbent increases in a well-defined way with contribution money and

this is the same for both groups of voters.

In particular, campaign contributions means a fixed monetary amount to the campaign

of the party to which they are given, and they have the psychological effect of adding a

fixed amount to the utility of all voters if the supported party gets elected. This is,

for example, because they are used for persuasive television advertising, and thus they

create a positive impression for this party.

c

( )e c

Let the parameters Ma , denote the fraction of the total voting population that belongs

to the groups

Ra

, ,M R respectively.

Assumption 3: M RM Ra z a wφ φ≥

Intuitively, this assumption is plausible if Ma is large relative to and if is greater

than or, at least, the values of and w do not differ much. It is also important that the

ideological homogeneity parameters

Ra z

w z

Rφ and Mϕ do not differ substantially in favor of the

rich group.

Assumption 4: ( )M R

M RM R

M R

a z ae ca a

wφ φφ φ

−≥

+

This assumption does not necessarily mean that the monetary amount is very large.

What is required is that the contributions have a substantial effect on the utility of voters,

c

8 An important question is: why would the interest group honor its promise and pay the contribution after the policy has been chosen? An answer can be given if we interpret the group F as a long-run player, who is interested for reputation building, and the politicians as short-run players, for example because each politician is elected for only two terms.

15

meaning, for example, that advertising is very persuasive. We will discuss the plausibility

of this assumption later.

Assumption 5: R

RM

MR

RM

M

aawaza

φφφφ

µ++−

≥

This requires that the effects of a positive economic performance on the incumbent

party’s popularity are large relative to the weighted effects on the incumbent’s popularity

stemming from the choice of an employer-friendly policy.

4. THE COMMITMENT OUTCOME

Illustrating the main ideas, we shall first briefly consider the game in stages one to three

only, without the possibility of campaign contributions. This is in order to show that that

the theoretical argument about policy rules is valid in this case, but its enforcement is not

trivial. Assume that commitment to a certain policy is costless. We want to examine

whether, in this game, the incumbent party would be better off in the equilibrium with

discretion or committing about the policy it will follow in stage two before the

investment decision in stage one would improve its position.

Claim: under assumptions one, three and five, if the incumbent party commits to follow

the employer-friendly policy in stage two it improves its position relative to the case

where it does not commit. If assumption two additionally holds, then everybody is better

of in this commitment equilibrium.

16

Proof: Proposition one in section five describes the equilibrium without commitment in

this game. Using its results, we know that the equilibrium payoffs with discretion are the

equilibrium probability of the incumbent party winning, that is:

=( )P LR

RM

MR

RM

M

aawaza

φφφφ

θθµ+−

+−21

If however, this party could commit before the investment decision, to choose employer

friendly policy in stage two, then assumption one ensures that hx = and therefore the

probability of its victory would be:

=( )P ER

RM

MR

RM

M

aawaza

φφφφ

θθµ++−

++21

So, the commitment outcome is preferable if

RR

MM

RR

MM

aawaza

φφφφ

θθµ++−

++21 ⇒≥

+−

−+− 021

RR

MM

RR

MM

aawaza

φφφφ

θθµ

022 ≥++−

+R

RM

MR

RM

M

aawaza

φφφφ

θθµ , which holds by assumption five.

This is a typical result that affirms that rules are better that discretion, especially when it

comes to capital taxation. The important issue here is how to achieve this result, or at

least approximate it with some cost, when a direct contract stipulating the commitment

arrangement is prohibited by law and any agreement is likely to bear the negative

suspicions of corruption. As the theorem shows, campaign contributions are likely to

approach this outcome, without avoiding the later suspicions though.

17

5. EQUILIBRIUM WHEN DIRECT COMMITMENT IS NOT POSSIBLE

Theorem: In the unique subgame perfect equilibrium of the game all voters vote to allow

for the institution of campaign contributions. The equilibrium strategies for all players

are:

1. The choice of from all voters is in stage zero. a

2. The strategy of in stage one is hl (firms invest only if the institution has

been approved).

F

3. The policy function of I in stage two is the following: ( , ) , ( , )f a h E f a l E= =

( , ) , ( , )f r h L f r l L= =

To prove this result we shall prove two propositions about the equilibria in the two

subgames that start at stage one.

Proposition 1: In the subgame where the voters reject the institution of campaign

contribution at stage zero, and under assumptions one and three, there exists a unique

subgame perfect equilibrium and the following pure strategies are equilibrium strategies

for ,F I :

• for at stage one, l F

• for the incumbent party at stage two (the incumbent chooses a labor-friendly

policy no matter the investment choice of the firms).

LL

18

Proof: Backward induction will be used. We shall start by considering the voting

behavior at stage three. The problem of voter j in group M is trivial. She votes for I if

the utility from doing so is greater that the utility from voting forC .

( , ) ( ) ( )M MjMv s x p s g x σ δ+ + + + > ⇒ ( , )Mv s x

[ ( ) ( ) ]MjM p s g x( ) ( ) 0M

jMp s g x σ δ+ + + > ⇒ σ δ> − + +

This holds for all voters in group M . Therefore, given , there will be a swing voter

who is indifferent between voting for

s

I and voting for . The ideology parameter for

this voter shall be

C

* [ ( ) (g x) ]MjM p sσ δ= − + + . Clearly, all voters of group M with

ideological parameter more than *jMσ vote for the incumbent party. Accordingly, the

share of the group M voters that vote for I is *12 M jMφ σ− .

Similarly, for the rich group, the swing voter has ideological parameter

* [ ( ) ( )RkR p s g x ]σ δ= − + + . The share of voters in this group that vote for the incumbent

is *12 R kRφ σ− .

Then, the share of votes over the whole population that party I gets is

*1( )2

MM jMa φ σΠ = − *

1( )2

RR kRa φ σ+ − =

1 [ ( ) ( ) ]2

M MMa p s g xφ δ+ + + + 1 [ ( ) ( ) ]

2R R

R s g xφ δa p+ + + =

1 [ ( ) ( ) ]2

M MMa p s g xφ δ+ + + + [ ( ) ( )R R

Ra p s g x ]φ δ+ + =

1 [ ( ) ( )]2

M MMa p s g xφ+ + + [ ( ) ( )]x ( )R M

R Ma aR RRa p s gφ + φ φ δ+ +

19

Since δ is still random, what the incumbent wants is to maximize his probability of

winning. This is equal to the probability that his share of total votes Π exceeds 12

.

1Pr( )2

Π ≥ = Pr [ ( ) ( )]M MMa p s g xφ + + [ ( ) ( )]x ( )R M

R Ma aφ φ δR RRa p s gφ + 0+ + ≥ =

Pr [ ( ) ( )]/( )M M R MM R M [ ( ) ( )a p s g x a aδ φ φ φ≥ − + + − R R

Ra p s gφ + ]x /( )R MR Ma aφ φ+ =

But given the distribution of the parameterδ , the probability that it exceeds a given

number c is 12

cθ− . Finally, the probability of the incumbent winning given already

chosen is the following:

,s x

P = 1 [ ( ) ( )] [ ( ) ( )] 2

M M R RM R

M RM R

a g x p s a g x p sa a

φ φθφ φ

+ + ++

+

Now, at stage two, the incumbent party anticipates this behavior of voters and chooses

the policy that maximizes maximize its utility. Since its utility depends only on the result

of the elections, it simply seeks to make its probability of being elected. It is worth noting

that a this stage the investment decision has already been taken and therefore the

incumbent party knows it cannot affect it. If it chooses policy , its probability of

winning is

L

=( )P L 1 [ ( ) ] [ ( ) ] 2

M RM R

M RM R

a g x z a g x wa a

φ φθφ φ

+ + −+

+

If it chooses policy E , its probability of winning is

=( )P E 1 [ ( ) ] [ ( ) ] 2

M RM R

M RM R

a g x z a g x wa a

φ φθφ φ

− + ++

+

20

Therefore, the incumbent party chooses the labor-friendly policy if

( )P L - , which implies that ( ) 0P E ≥

[ ( ) ] [ ( ) ]M RM Ra g x z a g x w [ ( ) ] [ ( ) ]M R

M Ra g x z a g x wφ φ≥ − + +φ φ+ + − ⇒

M RM Ra z a w M R

M Ra z a wφ φ≥ − + ⇒ φ φ−

0M RM Ra z a wφ φ− ≥ .

By assumption three, this holds. Furthermore, note that the policy choice does not depend

on the investment level x . When the policy is considered, investment decisions are

already made, and although they can make the incumbent party more popular, they

cannot affect its optimal decision. We conclude that the optimal strategy for I is , that

is, choosing a labor-friendly policy no matter what.

LL

This result is intuitive: the labor-friendly policy politically benefits the incumbent party if

the political clout of the middle-income group, net of the possibility of contributions, is

higher that the respective political clout of the rich group. This is what assumption two

maintains. The political clout of a group is determined by the intensity of the desire of

voters in this group to punish capital-friendly or labor-friendly policy, (for the middle and

the rich group respectively), weighted by the size and the ideological homogeneity of that

group.

The firms rationally anticipate this so they invest low in stage one, since assumption one

implies that they would reduce their profits if they invested high. So the optimal strategy

for is l . In a similar argument like in Kynland and Prescott, the government, in the F

21

absence of commitment, loses any control over the expectations of the firm sector, which

are no matter what. es L=

That completes the proof of proposition 1.

Proposition 2: under assumptions one, three and four there is a unique subgame perfect

equilibrium in the subgame where the institution of legal contributions is approved at

stage zero, and the following pure strategies are equilibrium strategies for ,F I :

• For at stage one, F x h=

• For the incumbent party at stage two, EE

Proof:

Again, backward induction is used. In this setting, the preferences of voters in stage three

depend on the policy for one additional reason: choosing s E= implies that contribute to

the campaign of the incumbent, and

F

s L= implies that contributes to the campaign of

the challenger party. So, in this case, if

F

s E= , for agent j in the middle-class group, the

utility function is:

( , , , ) ( , ) ( ) ( ) ( )M MjM jM jMu s x v s x p s g x e cσ δ σ δ= + + + + + , if I wins.

( , , , ) ( , )MjM jMu s x v s xσ δ = , if C wins.

For agent k in the rich group, the utility function is:

( , , , ) ( , ) ( ) ( ) ( )R RkR kR kRu s x v s x p s g x e cσ δ σ δ= + + + + + , if I wins.

( , , , ) ( , )RkR kRu s x v s xσ δ = , if C wins.

It is readily verifiable that now the probability of I winning is

22

'( )P E = 1 [ ( ) ( )] [ ( ) ( )] 2

M RM R

M RM R

a g x z e c a g x w e ca a

φ φθφ φ

− + + + ++

+=

1 (2

M RM R

M RM R

a z a w g s e ca a

φ φθφ φ

− ++ +

+) ( )+

If , then for agent j in the middle-class group, the utility function is: s L=

( , , , ) ( , ) ( ) ( )M MjM jM jMu s x v s x p s g xσ δ σ= + + + δ+ , if I wins.

( , , , ) ( , ) ( )MjM jMu s x v s x e cσ δ = + , if C wins.

For agent k in the rich group, the utility function is:

( , , , ) ( , ) ( ) ( )R RkR kR kRu s x v s x p s g xσ δ σ= + + + δ+ , if I wins.

( , , , ) ( , ) ( )RkR kRu s x v s x e cσ δ = + , if C wins.

It is readily verifiable that now the probability of I winning is

'( )P L = 1 [ ( ) ( )] [ ( ) ( )] 2

M RM R

M RM R

a g x z e c a g x w e ca a

φ φθφ φ

+ − + − −+

+=

1 (2

M RM R

M RM R

a z a w g s e ca a

φ φθφ φ

−+ +

+) ( )−

Again, the incumbent party, anticipating the behavior of voters in stage three and hence

these probabilities, will follow the employer friendly policy if

'( )P E '( ) 0P L− ≥ , which implies that

( )M R

M RM R

M R

a z a w e ca a

φ φφ φ

− ++ −

+( ) 0

M RM R

M RM R

a z a w e ca a

φ φφ φ

−+ ≥ ⇒

+

2 ( )e c − 2 0M R

M RM R

M R

a z a wa a

φ φφ φ

−≥

+

23

This holds by assumption four, which assumed that the effect of campaign contributions

is important for persuading voters (this important assumption is discussed later). Thus,

the incumbent party, under this contribution schedule of the firm sector, maximizes its

reelection probabilities if it chooses the employer-friendly policy in stage two. Once

again, the optimal strategy of party I does not depend on whether investments took place

or not in stage one. We conclude that under legal campaign contributions the optimal

strategy for the incumbent party in stage two is . Finally, rationally anticipating this,

the firm sector shall invest in stage one. This completes the proof of proposition 2.

EE

The last two parts of the theorem have already been proven. To prove part one, notice

that the equilibrium payoffs of all voters in stage two in the subgame with contributions

is and in the subgame without contributions it is 9),( hEvi ),( lLvi . From assumption

two, so all voters are better off if they approve the institution of

campaign contributions in stage zero. Assumption two therefore is the most important

assumption for campaign contributions to be welfare improving. As we shall discuss, this

is not as stringent an assumption as it may seem.

≥),( hEvi ),( lLvi

This proves the theorem.

Readers have probably noticed that the term does not play any role in equilibrium.

This emphasizes the point in this paper: that even if everybody, including the incumbent

)(xg

9 Under the reasonable assumption that the utility stemming from the election result, so all other components of the utility functions, are small relative to the realized economic payoff of period two. These are of course important for determining the electorate preferences of voters, but not a significant component of overall utility.

24

party, were better if this party committed to follow the employer-friendly policy, this

would simply not be possible without the presence of a guaranteeing institution.

6. DISCUSSION

The fact that the model does not have repeated interaction between the players, such as an

infinite horizon model, brings important limitations. First of all, it cannot incorporate the

incentives of voters to select good types of politicians but only the incentive to discipline

the performance of candidates. Secondly, it cannot examine the importance of reputation

building for achieving the commitment outcome. In this sense, the existence of the

institution of contributions can be viewed as an efficient condition for attaining this

outcome, not a necessary one, since an extension of the model may reveal other

commitment equilibria.

It is important to examine under which conditions the results of the model are relevant for

an economy. Assumption three requires that the “political clout” of the middle group in

the absence of campaign contributions is greater than the “political clout” of the rich

group. Whereas the assumption of a greater size for the middle-income group is hardly

disputable (so that is a realistic assumption), many political scientists assume

that the middle-income group is less ideologically homogeneous that the rich group, and

as a result

≥Ma Ra

RM φφ ≤ . See for example Person and Tabellini (2000) pp. 52-57. If this is true,

assumption three becomes less plausible.

25

Assumption four should also be discussed. There is an important debate in political

science regarding the importance of money in politics. Many authors like and Snyder,

Ansolabehere and Figueiredo (2003) argue that money is not that important in politics.

Their basic argument is that the money spent on campaign contributions are dwarfed by

the money that is at stake when economic policy is decided. At the same time, the legal

constraints on maximum contributions are not even binding. If money buys such

influence in policymaking decisions as it is argued, then the later fact is inconceivable,

given what is to be gained. They conclude that money cannot buy that much influence. If

the claim that money does not have a strong influence in politics is true, then cannot

be very large unless is unrealistically large.

)(ce

c

However, it should be noted that the idea of Snyder, Ansolabehere and Figueiredo is

contrary to conventional wisdom, which is the notion that money buys important

influence. This conventional wisdom is so strong that Gary Becker, in his influential

work (1983), did not include voting at all. He justifies this by the following:

“I too claim to have presented a theory of rational political behavior, yet have hardly mentioned voting.

This neglect is not accidental because I believe that voter preferences are frequently not a crucial,

independent force in political behavior. These ‘preferences’ can be manipulated and created through the

information and misinformation provided by interested pressure groups[…]”.(The emphasis is by the

author). This is just an example of the conviction that most people and scholars share,

that is, that interest groups have very strong effects on voting. In addition to all these we

should consider the debate about whether campaign advertisement is informative of

persuasive. Assumption four seems particularly plausible if the second view is true since

it seems to reinforce Becker’s argument about the manipulation of voters’ preferences.

26

Assumption two is not important for the argument that campaign contributions can affect

the economic policy, pushing in the direction of employer-friendly policy. However, the

argument that the institution of legal campaign contributions is beneficial for society as a

whole depends on this assumption. In this sense, the whole argument of this paper that

campaign contributions can serve a positive economic role under some circumstances is

based on this assumption. This assumption does not require that all people have the same

utility with respect to the policies chosen and the investment level. It may well be the

case that rich voters prefer the employer friendly policy and middle-income voters prefer

the labor-friendly policy. In other words, for every),(),( xEvxLv MM > x and

for every),(),( xLvxEv RR > x is compatible with assumption two.

The assumption of non-partisan politicians is not very important here. The results of the

model would not change if you had a left-wing candidate and a right-wing candidate. In

such a case, the two candidates have a strong incentive to follow their preferences no

matter what they have promised. We briefly sketch a model with partisan politicians but

with slightly changed timing. In this setting, like in the previous one, the firm sector

decides whether to invest or not in stage one. However, elections take place in stage two

and the winner implements the policy in stage three. The two parties here have different

preferences: they do not care about the result of the election per se, but for the policy

chosen. Party I prefers the labor-friendly policy, whereas party C prefers the employer-

friendly policy. More formally, and U . These preferences )()( EULU II > )(L)( UE CC >

27

are common knowledge. Clearly, parties do care about being elected since this will allow

them to implement their policy vs. the other.

The equilibrium without campaign contributions will be driven by the fact that each party

has a dominant strategy to choose its favorite policy after it gets elected. Knowing this,

the voters will anticipate policy from party L I and policy E from party . The greater

political clout of the middle group, by assumption three, will therefore lead to the election

of party

C

I , so policy will be chosen. Anticipating this, the firm sector does not invest

in stage one.

L

However, if before elections the firm sector can contribute to its favorite candidate, it is

clear that it will contribute to the conservative party knowing that upon victory this party

will choose policy E . With contributions, the right-wing candidate’s popularity would

increase and he would win the elections by assumption four. Thus, the firm sector would

invest in stage one. Again, all will be better off in this equilibrium relative to the

equilibrium without campaign contributions. This is a case where strategic delegation

makes sense. Person and Tabellini (1994) convincingly argue that voters would like to

convince firms that they will vote for the right-wing candidate, but this statement is not

credible, since at the moment of voting the investment decision already has been made. In

this sense our model explains how “indirect strategic delegation” works. Voters would

like to commit on voting the right-wing candidate but they cannot do so. Campaign

contributions help this strategic delegation be achieved.

28

A major issue is for what types of democracies the results in this paper are likely to apply,

and thus the insights from our model instructive. The existence of problems of time

inconsistency, especially with respect to capital taxation and macroeconomic stability,

seems to warrant commitment solutions. Hence, countries with strong leftist parties and a

tradition of populist governance are more likely to use devices such as the one described

in this paper for promoting investments. It may seem that the institution of private

campaign contribution can serve as tool that promotes growth in such cases. This does

not mean that countries should blindly accept unlimited special interest influence (see

next part). Finally, it is worth noting that our model has something to say even for

countries that do not seem to have a current policy credibility problem, such as the United

States. In particular, it may explain why this institution emerged in the first place. The

historic circumstances where this happened may well be similar with the contemporary

conditions in countries that need commitment.

7. CONCLUSIONS

We used a model to examine how the institutionalization of legal corporate campaign

contributions can ameliorate the credibility problem in economic policy and achieve

something close to the commitment outcome (at some cost). We concluded that this

could be achieved under some assumptions regarding the relative political strength of

voter groups and the importance of campaign contributions for shaping political

preferences.

29

The model could be tested with a data sample for several countries, which would include

information about the “institutional regimes” of countries with respect to campaign

contributions and their economic performances, especially with respect to investments. If

a positive and statistically significant relationship were found between the existences of

institutions where overt campaign contributions are permitted and economic efficiency,

this would support the model.

The practical significance of the results of this paper is that the existence of such a strong

influence of the corporate sector in many countries can be understood under the view of

economic efficiency attained with this institution. It must be emphasized that this analysis

does not imply that any society should permit unlimited electoral influence of corporate

interests. It just gives one argument for the possible economic efficiency of an institution

that allows for some influence. The criticisms mentioned in the introduction may well be

valid and, depending on the value system of a society, they may weight much more

heavily that economic efficiency. Political equality and transparency are two principles

that have great importance in their own merit, which should not be judged by their

economic consequences.

30

References:

Avinash Dixit, “The Making of Economic Policy, a Transaction Cost Politics

Perspective”. The MIT Press, Cambridge, Massachusetts, 1996.

Stephen Haber, Armando Razo, Noel Maurer, “The Politics of Property Rights”,

Cambridge University Press, Cambridge, 2003.

Adam Winkler, “"Other People's Money": Corporations, Agency Costs, and

Campaign Finance Law”. Georgetown Law Journal, June 2004.

Torsten Person, Guido Tabellini, “Political Economics: Explaining Economic Policy”.

MIT Press, Massachusetts, 2000.

Adam Przeworsky, Michael Wallerstein, “Structural Dependence of the State on

Capital”. American Political Science Review, March 1988, Vol. 82, pp. 11-29.

Alberto Alesina, “Credibility and Policy Convergence in Two-Party System with

Rational Voters”. American Economic Review, Vol. 8, No. 4, September 1988, pp.786-

805.

31

Avinash Dixit, Gene M. Grossman, Elhanan Helpman, “Common Agency and

Coordination: General Theory and Application to Government Policy Making”. The

Journal of Political Economy, Vol. 105, No. 4, August 1997, pp. 752-769.

Alberto Alesina, “Macroeconomic Policy in a Two-Party System as a Repeated

Game”. The Quarterly Journal of Economics, Vol. 102, No. 3, August 1987, pp. 651-678.

Gene Grossman, Elhanan Helpman, “Special Interest Politics”. MIT Press,

Massachusetts, 2001.

Finn Kynland, Edward Prescott, “Rules Rather than Discretion: The Inconsistency of

Optimal Plans”. The Journal of Political Economy, Vol. 85, June 1977, pp. 473-492.

James Snyder, Michael Ting, “Interest Groups and the Electoral Control of

Politicians”. Working paper, February 2005.

Stephen Ansolabehere, Erik Snowberg , James Snyder, “Television and the

Incumbency Advantage in US Elections”. Working paper, July 2004.

Stephen Ansolabehere, John M. de Figueiredo, James Snyder “Why is There so Little

Money in US Politics?”. The Journal of Economic Perspectives, Vol.17, No. 1, Winter

2003, pp. 105-130.

32

Stephen Levitt, “Policy Watch: Congressional Campaign Finance Reform”. The

Journal of Economic Perspectives, Vol. 9, No.1, Winter 1995, pp. 183-193.

James Kau, Donald Keenan, Paul Rubin, “A General Equilibrium Model of

Congressional Voting”. The Quarterly Journal of Economics, Vol. 97, No.2, May 1982,

pp. 271-293.

Gary Becker, “A Theory of Competition among Pressure Groups for Political

Influence”. The Quarterly Journal of Economics, Vol. 98, No.3, Aug. 1983, pp. 371-400.

Stephen Coate, “Political Competition with Campaign Contributions and

Informative Advertising”. Working paper, December 2001.

Stephen Coate, “Pareto Improving Campaign Finance Policy”, American Economic

Review, Vol. 94, No.3, June 2004, pp. 628-655.

Andrea Prat, “An Economic Analysis of Campaign Finance”. Working paper, October

1999.

Christian Schultz, “Strategic Campaigns and Redistributive Politics”. Working Paper,

September 2003.

33

Roderick Kiewiet, Douglas Rivers, “A Retrospective on Retrospective Voting”.

Political Behavior, Vol. 6, No. 4, 1984.

Kirsten Monroe, “Econometric Analyses of Electoral Behavior: A Critical Review”.

Political Behavior, Vol. 1, No. 2, 1979.

Gerard H. Kramer, “Short-Term Fluctuations in U.S. Voting Behavior, 1896-1964”.

The American Political Science Review, Vol.65, No.1, March 1971.

34