Fiscal Federalism and Electoral Accountability

• Published date: 01 February 2017
• DOI: http://doi.org/10.1111/jpet.12179
• Date: 01 February 2017
• Author: TOKE S. AIDT,JAYASRI DUTTA

FISCAL FEDERALISM AND ELECTORAL ACCOUNTABILITY

TOKE S. AIDT

University of Cambridge

JAYASRI DUTTA

University of Birmingham

Abstract

We evaluate how governance uncertainty—exemplified by turnout

uncertainty—affects the trade-off between internalization of externali-

ties and political accountability in the design of the fiscal state. We show

that centralization only weakens political accountability in the presence

of negative externalities. Unlike positive externalities, negative exter-

nalities allow federal politicians to extract higher rents. This yields two

new insights. First, decentralization can only Pareto dominate central-

ization in economies with negative externalities. Second, centralization

may not be Pareto efficient in economies with positive externalities de-

spite the fact that policy can be tailored to regional taste differences

and centralization internalizes the positive externality.

1. Introduction

Should a society opt for a centralized fiscal system where spending decisions are made

by a central authority and financed from general tax revenues or should it opt for a

decentralized system where fiscal choices are made by local authorities and financed

by local taxes? Oates (1972) answered this question in his Decentralization Theorem

which states that decentralization is desirable if externalities are weak and regional dif-

ferences in taste are large.1The design of the fiscal state also entails many political econ-

omy trade-offs.2This paper contributes to the political economy of fiscal federalism by

exploring how governance uncertainty affects the trade-off between internalization of

externalities and the perceived benefits of electoral accountability under regionalism.

1Harstad (2007) provides a rationale for why it might be politically optimal to select uniform federal

policies.

2We discuss this literature in Section 2.

Toke S. Aidt, Faculty of Economics, University of Cambridge, Austin Robinson Building, Sidgwick Av-

enue, Cambridge CB3 9DD (Toke.Aidt@econ.cam.ac.uk). Jayasri Dutta, Department of Economics,

University of Birmingham, Birmingham, U.K. (J.Dutta@bham.ac.uk).

We thank Robert Evans, Paola Profeta, Dalibor Eterovic, and Lars Feld for useful comments and

suggestions. This research was supported by a grant from the Barcelona Institute of Economics (IEB)

under its research program in Fiscal Federalism.

Received June 29, 2014; Accepted July 23, 2015.

C2016 Wiley Periodicals, Inc.

Journal of Public Economic Theory, 19 (1), 2017, pp. 38–58.

38

Fiscal Federalism 39

Governance uncertainty arises when there is uncertainty about which region in a feder-

ation will determine the outcome of federal elections.

The general framework of our analysis is the common agency model with gover-

nance uncertainty studied by Aidt and Dutta (2004). This model portrays a society pop-

ulated by heterogenous groups of voters (e.g., living in different regions) with conflict-

ing policy preferences. The groups of voters (the principals) use elections to hold an

opportunistic politician (the agent) accountable for his policy choices while in office.

They do so by voting retrospectively in an infinite sequence of elections, as in Ferejohn

(1986), Persson, Roland, and Tabellini (1997), Coate and Morris (1999), Aidt and Ma-

gris (2006), or Aidt and Dutta (2007). The critical new feature of the analysis is that

before each election the politician is uncertain about which group will be pivotal in

deciding the outcome of the election. We call this governance uncertainty. Governance

uncertainty has many different sources. To be concrete, however, we relate it to ran-

domness in the electoral turnout of voters in different groups.3These random turnout

shocks introduce uncertainty from the point of view of the politician as to which group

holds the majority among those voters who actually turn out to vote in any given elec-

tion. An example is random fluctuations in weather conditions. Such fluctuations can

affect the turnout rate in some geographical locations and not in others or keep certain

types of voters, such as the poor, at home (Roemer 1998; Gomez, Hansford, and Krause

2007). Art´

es (2014), for example, reports that election day rainfall decreases turnout in

Spanish general elections and that high turnout harms the conservatives and benefits

the smaller parties. Lind (2014) also reports evidence that rain on election day affects

voter turnout and that this, in turn, affects the party composition of the elected coun-

cils in Norwegian local elections. Along similar lines, Arnold and Freier (2015) show

that rain shocks affect turnout in the German state of North-Rhine Westphalia and that

lower turnout benefits the conservatives and harms the social democrats. These exam-

ples illustrate that random fluctuations in turnout, induced by rain shocks, can affect

election outcomes.

Based on this general framework, we develop a specific political economy model of

fiscal federalism in a country with two regions. Provision of local public goods in one

region has spillover effects on the other region. Centralization encourages internaliza-

tion, but introduces governance uncertainty because the federal politician cannot know

for sure ex ante which of the two regions will decide his reelection. We use this model

to study when provision of local public goods should be centralized. The model also

provides insights into the forces that stabilize and destabilize federal fiscal structures.

The main result of the analysis is an asymmetry between positive and negative ex-

ternalities with regard to how much rent politicians in a federation can extract. With

negative externalities, a federal politician can extract more rents from his citizens than

the collective of regional politicians. With positive externalities, this is not the case. In-

tuitively, the asymmetry arises because it is more expensive for the federal politician to

keep all regions satisfied when public goods provided in one region exhibit a negative

externality than when the externality is positive. To avoid being “ignored” when fiscal

externalities are negative, voters in the two regions must accept that the federal politi-

cian collects higher rents than with positive externalities. In the presence of negative

externalities, the classical result that federalism is better for citizens than regionalism

can even be turned on its head. Since federal politicians can extract more rents, fed-

eralism can be worse than regionalism for all citizens, despite the fact that regionalism

3Dhillon and Peralta (2002) survey the literature on voter turnout. Aldashev (2014) studies the link

between political rents and voter turnout.