Has Leviathan Been Bound? A Theory of Imperfectly Constrained Government with Evidence from the States.

• Author: Caplan, Bryan

Bryan Caplan [*]

This paper develops a formal theory that combines power-maximizing "Leviathan" political parties with well-defined imperfections in the political process. The model implies that both parties tend to make government larger as their likelihood of electoral victory increases. Empirical tests on state-level data confirm this prediction. Racing the Leviathan hypothesis against alternative theories of party motivation indicates that both the Leviathan and the "contrasting ideologies" views have some degree of validity.

1. Introduction

   The "Leviathan" motive of politicians to maximize their power is a central assumption of the public choice approach to political economy. [1] This approach has been strongly attacked for underestimating the importance of competitive checks on politicians; Wittman (1989, 1995) in particular broadly argues that electoral competition is an effective solution to whatever principal-agent problem might exist between politicians and voters. Yet recent research in political economy (Dixit and Londregan 1995, 1996, 1998; Grossman and Helpman 1996) casts renewed doubt on the efficiency of political markets. Formal models incorporating imperfections in the electoral process show that politicians have the latitude to deviate somewhat from citizen interests. Although politicians face constraints, they retain some measure of monopoly power, a conclusion consistent with the public choice literature.

   The main difference between recent political economy and the public choice tradition arises from their assumptions about politicians' objective functions. The public choice approach tends to view all political parties as power maximizers, whereas others see parties as either vote maximizers (Dixit and Londregan 1995, 1996; Grossman and Helpman 1996) or promoters of conflicting ideologies (Alesina and Rosenthal 1995; Dixit and Londregan 1998). The present paper briefly develops a theoretical model that pits Leviathan--in the form of two power-maximizing parties--against the electoral constraints of modern political economy. The main implication of the model--an implication empirically tested against several alternatives later in the paper--is that both imperfectly constrained power-maximizing parties will make government larger as their likelihood of electoral victory increases. Intuitively, the results are consistent with Peltzman's (1992) informal analysis of government growth in the face of voters' fiscal conservatism:

   The larger question is how political agents can persistently dissipate voter wealth; that is, why has government grown so much (and why is it fiscally progressive)?... One [possibility] is that the political costs of growing budgets are too weak to compel much restraint.... These exercises suggest that incumbents can indulge in nontrivial spending growth before they risk a close call next election day. (pp. 358-9)

   The next section discusses the theoretical and empirical literature on politicians' objective functions and imperfections in the political sector. The third section presents the model of power-maximizing parties facing electoral constraints. Initially, politicians operate in a certain environment; this assumption is then relaxed to yield implications that are both more plausible and more readily tested. The fourth section tests the Leviathan model and its competitors on state-level political and economic data during the 1950-1989 period and analyzes the empirical results. Because section four finds that political parties differ less about the level of spending than one might expect, the fifth section examines whether parties matter for the composition of taxation and spending. The sixth section concludes the paper.

2. Related Literature

   The model developed here builds on the formal theory of recent political economy, the less formal analysis of government found in the public choice literature, and the broader debate about the extent to which politicians find themselves constrained to efficiently satisfy consumer preferences.

   The conclusions of my model, like most recent work in theoretical political economy, depend critically on the existence of an imperfection in the electoral process: Voters treat political parties as differentiated products. A literature beginning with Lindbeck and Weibull (1987) differentiates parties by assuming that they may easily alter their positions on some issues (such as the budget) but must hold their stance on other issues (such as abortion) fixed. Recent advances made on this foundation include those of Dixit and Londregan (1995, 1996, 1998) and Grossman and Helpman (1996). This assumption is not ad hoc: As Dixit and Londregan (1998) note, voters genuinely care about ideology, and even in the face of repeated electoral defeats, parties find it difficult to make genuine ideological shifts. Fixed ideological positions plus voter ideology adds up to imperfect political competition. [2]

   Although a growing literature builds on Lindbeck and Weibull's approach, there is much less consensus about parties' objective functions. Wittman (1983) provides a general discussion of the implications of politicians' preferences over policies as well as electoral outcomes. Some, such as Grossman and Helpman (1996) and Dixit and Londregan (1995, 1996), simply model parties as vote maximizers. Others, such as Baron (1994), assume parties maximize their probability of victory. In Alesina and Rosenthal (1995), one party has a progoverument ideology while the other has an antigovernment ideology; Dixit and Londregan (1998) have parties maximize a weighted average of their vote share and an ideological social welfare function. Electoral competition forces both parties to compromise their ideologies to some extent.

   Vote-maximizing and conflicting ideologies theories of political motivation differ from the power-maximizing or "Leviathan" parties often posited within the public choice literature. Brennan and Buchanan (1980), for example, assume that both Democrats and Republicans want to make government revenue and spending as large as possible. They only refrain from increasing it even more because of political and economic constraints, not because they do not want to. An analogous perspective is perhaps most dominant in the literature on the political economy of protectionism (Grossman and Helpman 1995). A common conclusion in this literature is that both political parties incline toward excessive protectionism.

3. The Theory of Electorally Constrained Power-Maximizing Parties [3]

   The players are two federal parties {i, j} and a continuum of citizens whose measure is normalized to 1. Play is simultaneous, and in any Nash equilibrium (i) all citizens must vote for their most preferred federal party and (ii) both political parties must offer platforms that maximize their expected utility.

   Citizen Preferences and Constraints

   Citizen utility depends on not only consumption of private goods [P.sub.c] and public goods G, but also on the political environment:

   [u.sub.c] = u([P.sub.c], G) + [[psi].sub.c][I.sub.i]. (1)

   [[psi].sub.c] [sim] U[-0.5 + [psi], 0.5 + [psi]] is c's relative taste for party i versus party j. [[psi].sub.c] is the amount of utility (positive or negative) the individual would be willing to give up in order to be ruled by i rather than j; [psi] may be interpreted as both the average and the median value of [[psi].sub.c]. [I.sub.i] is an indicator variable, which is 1 if party i is in power and 0 otherwise; [I.sub.j] = 1 - [I.sub.i].

   Citizens' corresponding indirect utility function [z.sub.c] is assumed to have functional form:

   [z.sub.c] = Z-[beta][(G-[[G.sup.*].sub.c]).sup.2] + [[psi].sub.c] [I.sub.i] (2)

   There are T distinct types of citizens, each with its own most desired level of government services, [[G.sup.*].sub.t] which is uncorrelated with [[psi].sub.c] Each of these types constitutes a fraction [[chi].sub.t] of the population, so

   [[[sigma].sup.T].sub.t=1] [[chi].sub.t] = 1 (3)

   Parties' Preferences and Constraints

   The parties compete in an election held each period. [G.sub.i] is the political platform offered by party i, and [G.sub.j] is the platform of j. The competing parties are both "power maximizers" who want government to be as big as possible assuming they are in office. [4] Formally,

   [u.sub.i] = [I.sub.[i.sup.*]][U.sub.+]([G.sub.i]) (4)

   [u.sub.j] = [I.sub.[j.sup.*]][U.sub.+]([G.sub.j]). (5)

   All properties of standard utility functions hold. It is further assumed that U(O) = 0: Controlling a government with no resources gives the same utility as being out of power.

   Political Equilibrium with Certainty about Political Advantage

   Because citizens vote to maximize their own utility, a citizen votes for federal party i if

   -[beta][([G.sub.i] - [[G.sup.*].sub.t]).sup.2] + [[psi].sub.c] [greater than or equal to] -[beta][([G.sub.j] - [[G.sup.*].sub.t]).sup.2] (6)

   and for party j otherwise.

   Suppose that [psi] can be observed without error by the political parties, and (for simplicity) that a tie goes to i if [psi] [greater than or equal to] 0, and to j if P [less than] 0. Given majority rule, it will then not be an equilibrium for both political parties to offer the median preference. If [psi] [greater than or equal to] 0, party i wins with certainty if it plays the median value of G; due to its advantaged position, it can definitely win even if its platform offers a somewhat larger public sector. Party i will want to keep increasing the offered level of government until it drives the percentage of votes it receives down to 0.5. Similarly, if [psi] [less than] 0, j wins with certainty if it exactly satisfies the median preference; due to its advantaged position, it can afford to offer a larger government. [5]

   Consider first the case where [psi] [greater than] 0. Assuming no corner solutions, [6] party i can drag its share of the vote down to .5 if it pushes its offered platform up until it satisfies

   [[[sigma].sup.T].sub.t=1]...