Testing Rational Partisan Theory When Elections Are Endogenous Events: Some Empirical Evidence from the United Kingdom
Heckelman, Jac C., Quarterly Journal of Business and Economics
Rational partisan theory predicts macroeconomic fluctuations are triggered by possible changes in government policies due to elections. Empirical testing may fall prey to an endogeneity problem when incumbent governments determine the timing of an election and voters respond to current economic conditions in their choice of party support. Hausman tests suggest elections in Britain are endogenous to growth. Stronger support is found for rational partisan theory using an instrumental variable routine.
The connection between elections and macroeconomic fluctuations has spanned a lot of ground in the political economy literature. The more recent literature focuses primarily on partisan political policies and rational expectations. Under certain conditions, election outcomes may trigger temporary changes in the macroeconomy. In this paper, we argue that empirical tests need to account for the possibility that the economic environment may create elections and/or determine the election outcome. When this holds, instrumental variable estimation represents an improvement to the ordinary least squares (OLS) analysis commonly employed.
The next section highlights the development of the political business cycle literature.  Then we suggest a correction to the empirical approach to the most recent wave of electoral models, namely so-called rational partisan theory. Finally, we summarize the empirical results and offer suggestions for future research.
Development of Election Cycle Literature Political Business Cycles
Extensive research has been devoted to the notion of politically generated business cycles. Nordhaus (1975) has suggested incumbent politicians have an incentive to stimulate the economy to improve their reelection chances, even though any economic changes will be short-lived. A political business cycle is created due to a myopic electorate that displays retrospective voting behavior. Voters in the Nordhaus model employ adaptive expectations.
Two theoretical developments have been advanced in the literature to adjust the. Nordhaus model. First, in response to Kramer's (1971) popularity functions estimates, Stigler (1973) argued that rational voters should be aware economic changes are transitory and thus should not look to economic indicators in deciding how to vote. The implication for the political business cycle is that voters should not reward incumbents for planned manipulation of the economy. In addition, the rational expectations approach to macroeconomics suggests expected fiscal and monetary manipulations will not affect real output growth or unemployment, which would seemingly break any link between elections and the economy.
Political cycles can exist, however, under imperfect information. For example, Rogoff and Sibert (1988) develop a model of asymmetric information incorporating rational expectations in which government performance is affected by competency shocks. Incumbents signal their competence by altering fiscal policy. Because the government's competence is revealed to the public only after a lag, changes to inflation may be unexpected. The composition of the resulting political cycle depends on the type of shock that occurs and whether a separating equilibrium exists.
A second criticism of the political business cycle theory entails the notion that politicians in the Nordhaus model are pure office-seekers. Hibbs (1987) provides evidence that partisan influences play a role. Alesina and Sachs (1988) combine the notion of partisan political parties, rational expectations, and uncertainty. Although economic agents are aware of the parties' differing monetary policies, they are uncertain of the election outcome when setting their wage contracts. To minimize the variance of their forecast error, inflationary expectations before the election are based upon a weighted average of the parties' expected monetary policies. …