|
Article Excerpt
Summary. We consider a model in which parties that differ in perceived valence choose how to allocate electoral promises (money, pork-barrel projects) among voters. The party perceived to be less valent has a greater incentive to "sell out" to a favored minority and completely expropriate a fraction of the electorate. By reducing the difference in perceived valence, campaign-finance regulations may reduce the extent of the expropriation and achieve a more equitable political outcome. We analyze various instruments of campaign-finance regulation from this perspective.
Keywords and Phrases: Campaign spending regulation, Redistributive politics.
JEL Classification Numbers: D72, H2.
1 Introduction
This paper is concerned with the effects of campaign spending and its regulation within a model of redistributive politics. The model is one in which parties (or candidates) compete for votes by promising slices of a cake. This is a benchmark model in much of the literature in political economics and political science. Most treatments of this model select as solutions allocations that are not egalitarian--not all voters receive equal slices. The tendency towards unequal allocations was noted by Shubik [28] and is the focus, for example, of Myerson [25] and Laslier and Picard [19]. Inequality among voters is viewed as an undesirable property of the solution. In this paper, we solve for a variant of this benchmark model in which parties are asymmetric. We are interested in the inequality of the solution as a function of the asymmetry between parties.
In our model, parties (or their leaders) are perceived by voters as differing in valence. A party with greater valence is believed by voters to be able to generate more resources (a bigger cake) once in office. That party will therefore be able to make electoral promises under a less stringent budget constraint, and so has an advantage. (3) In this environment, we show that electoral imbalance translates into fiscal inequality. To see why, consider the position of a party that is perceived as less valent. In order to win a voter, that party must offer that voter at least as much as the competing party. However, the less valent party is subject to a more stringent budget constraint and, in equilibrium, chooses not to treat all voters equally. Instead, the less valent party ignores a fraction of the electorate in order to concentrate its resources on the remaining voters. In this sense, the disadvantaged party "sells out" to a favored sub-group of the electorate and completely expropriates the remaining fraction of the electorate. The greater the disadvantage at the electoral stage, the larger the fraction of voters that are expropriated in the disadvantaged party's platform, and the greater the degree of fiscal inequality. (4)
Having argued that the degree of fiscal inequality is related to the difference in the voters' perception of parties, we set out to evaluate the role that campaign spending regulation can play in reducing the asymmetry between parties. This exercise is similar in spirit to that of Myerson [25]. Myerson [25] asks a constitutional design question by comparing electoral systems in terms of the inequality that they induce in equilibrium. We take the electoral systems as given, and compare campaign spending regulations according to the same criterion.
In our model, the campaign spending stage precedes electoral competition. In the campaign spending stage parties start out with different perceived valences, and can add to their valence by spending campaign funds. In effect, parties in the campaign spending stage are engaged in a spending contest for advantage in the electoral stage. Let us first concentrate on the incentives for parties to spend campaign funds in the absence of regulation. A fundamental feature of the vote shares (as determined in equilibrium at the electoral stage of our model) is this: if campaign expenditure is increased by the same amount for both parties, then the underdog gains in vote share. (5) Thus, if a sufficiently large amount is added to both parties' expenditures, any initial differential in perceived valence becomes negligible in the electoral game, which makes sense since the ratio of perceived valences converges to 1. (6) This "catching-up effect" is a natural property if we take the position, as we do in this paper, that campaign expenditure increases (or is a substitute for) a party's perceived valence. Note that this property implies that the productivity of campaign funds, measured in terms of the electoral advantage afforded by a unit of expenditure, must be greater for the disadvantaged party. But then the incentives to spend campaign funds must be larger, caeteris paribus, for the disadvantaged party. In our model this effect will lead the disadvantaged party to outspend its opponent in the attempt to catch up. (7)
This argument suggests that, since in our model campaign spending reduces the gap in perceived valence, a laissez-faire approach to campaign spending might be beneficial. In fact, in the special case in which the effect of spending on perceived valence is linear, we show that under laissez faire the disadvantaged party completely catches up and a perfectly balanced outcome is achieved. In such circumstances, campaign finance regulation can only hinder the catching-up process. Similarly, in the more general case in which laissez faire does not achieve perfect balance we show that regulations that provide free in-kind transfers in equal amounts to both parties have anti-competitive effects, because they make it harder for the disadvantaged party to catch up. The effect of a regulation setting a cap on expenditures is less immediate; in equilibrium, imposing a cap decreases the expenditure of both parties due to a strategic effect on the expenditure of the ex ante favored party. Nevertheless, we are able to show that the impact of a cap is stronger on the spending of the ex ante disadvantaged party, and so imposing a cap reduces the equilibrium vote share of that party. Again contrary to conventional wisdom, spending caps are anti-competitive. The effects of per-seat reimbursements and of matching funds on campaign spending are more complicated. Under these types of schemes, financing becomes cheaper for both parties. Thus, both parties increase their campaign outlays, and the total effect on vote shares depends on which party increases its expenditure more. In our model, that depends on the productivity of spending; we identify tight conditions on the curvature of the function that transforms expenditures into perceived valence, under which per-vote reimbursements and matching funds are pro- or anti-competitive.
1.1 Related literature
Asymmetric electoral competition
Our analysis contributes to a recent literature on electoral competition between asymmetric candidates or parties (see e.g. Aragones and Palfrey [2] and Grose-close [16]). One point of difference with much of this literature is that we assume that parties maximize their vote share instead of the probability of winning. This assumption makes our model a closer representation of electoral systems with proportional representation.
Redistributive politics
The literature on redistributive models of electoral competition is surveyed by Dixit and Londregan [13]. Our model of redistributive politics extends Myerson [25] to the case of asymmetric parties. An important feature of the asymmetric model is that in equilibrium voters are treated differently by different parties, and the difference in the parties' electoral appeal translates into polarized fiscal policies. Laslier and Picard [19] analyze a version of Myerson's model in which there are a finite number of voters and two parties and show that when the number of voters converges to infinity the inequality is smaller than in Myerson's model.
We endogenize the electoral appeal of parties through the campaign financing activity, and so obtain a model in which the extent of fiscal inequality is endogenous. In this sense, our analysis complements Dixit and Londregan [13], who call attention to the importance of fiscal inequality. In their model, inequality in the treatment of voters results from two features. One is the responsiveness of different voters to electoral promises: groups of voters which are more responsive receive more generous electoral promises. The second feature is the presence of a concern for inequality into the payoff functions of agents (parties and voters). In contrast, in our stylized model voters are identical and no political actor has a concern for equity. In our model, fiscal inequality emerges in equilibrium as the strategy of the party with less electoral appeal.
Campaign spending
The formal literature on campaign spending is vast; here we discuss the work to which this paper is more closely related, and we refer to Morton and Cameron [23] for a general survey. Most papers concentrate on implications for the probability of winning of incumbent vs. challenger candidates (Levitt [20], Erikson and Palfrey [14]) or on the amount of campaign contributions (Baye, Kovenock, and DeVries [6], Che and Gale [9]). In order to focus on campaign spending, most of this literature models the behavior of voters through an exogenous "vote production function" whose inputs are campaign contributions. In contrast, because we explicitly model the stage of electoral competition, we are able to draw implications for policy outcomes.
The effect of campaign contributions on policy outcomes is explicitly modeled in the literature on lobbying. In this literature, inefficiencies result from politicians pandering to lobbies, which are groups with preferences different from the median in the population. In most of the lobbying literature electoral competition remains in the background as lobbying is directed at an incumbent politician; an exception is Baron [5], who studies a lobbying model in which two parties compete for office. One key difference between our model of campaign spending and lobbying models is in the policy implications. In these models, the inefficiency is a result of lobbying, and so the best policy is to limit campaign financing. Some recent papers have analyzed models in which campaign advertising has a social benefit. Prat [27] analyzes the trade-off between the informational benefits of advertising and its social costs. In his analysis, the competence of candidates is not perfectly known. An interest group, which is better informed than the voters, gives contributions to candidates. Since, the group offers more contributions to a candidate he believes to be more competent, and thus more likely to be elected, contributions play the role of a signal. Limiting contributions and advertising has ambiguous welfare effects: the losses in terms of information about competence are to be balanced with policy distortions. Coate [10,11] analyzes models with directly informative advertising. In Coate [11], voters are uninformed about candidates' ideologies. He shows that when contributions come from interest groups who try to get their favorite candidate elected, limits on advertising redistribute welfare from moderate voters to members of the interest groups. In Coate [10], voters are uninformed about the competence of candidates. He shows that when advertising is financed by contributions from interests groups in exchanges for favors, the informational value of advertising on competence is mitigated by the knowledge that advertising is linked with the inefficient distribution of favors. In this context, a ban on advertising may well be Pareto improving.
In our model, instead, we take no stand on the origin of funds. We focus on the consequences of advertising expenditures in the context of an asymmetric electoral competition in which one party is initially disadvantaged. Forbidding campaign advertising has a negative impact since it maximizes the difference in electoral appeal and hence the incentive of the disadvantaged party to run on a fiscally polarized platform.
1.2 Outline of the paper
Section 2 presents the model and the timing of events. Section 3 solves for the equilibrium of the electoral game. Section 4 analyzes the campaign spending stage. Section 5 discusses the effects of campaign spending regulations. We discuss the fiscal inequality generated by unbalanced elections in Section 6. Section 7 extends the analysis to three or more parties. We show how the asymmetry between parties translates...
|
