Elections and aggregation: Interpreting econometric analyses of local governments

Conclusion

This paper analyzes the provision of local government service under the assumption that fiscal policy is set by local officials in order to maximize the expected number of votes in the next election. The model includes both the widely-used median voter model and the “average voter” model as special cases. Further, it is a convenient framework for empirical analyses.

Numerical simulations indicate that the vote-maximization model is capable of producing levels of government spending either greater or less than those predicted by the median voter model, while exhibiting similar responses to changes in incomes, grants, prices, etc.. Thus, the model captures the advantages of the median voter model without its objectionable features.

Many avenues for further work remain, I will mention briefly only a few of the most obvious. The first is to make housing prices endogenously determined within the model. In this way, one could examine the role of both fiscal incentives (capitalization) and political incentives (votes) in the determination of optimal policies.

A second extension is to introduce explicitly the possibility of a mobile population and examine, in the spirit of Tiebout (1956), the effects of “voting with your feet” within the vote-maximization framework. The possibilities include both setting policies that de facto concede exit of some fraction of the current residents and/or setting policies to attract new residents with particular characteristics.

Another extension is to look at decisions over time. Despite widespread agreement upon the usefulness of the life-cycle model of private consumption, the median voter is essentially static. It seems a minimal extension to include both government and private consumption in an intertemporal framework. Unfortunately, the median voter may change over time, making implementation of the median voter model problematic at best.

In practice, intertemporal issues are important, emphasizing the absence of an acceptable framework. It is often suggested that elected officials will choose inappropriate intertemporal fiscal policies in response to the short-run focus of the electoral process. For example, it appears plausible that officials will borrow in order to finance higher levels of spending (and/or lower taxes) in the present. Because future residents are not present to vote against these policies, the electoral process produces incorrect incentives. In the same way, this line of reasoning suggests that officials will undertake too little infrastructure investments because the benefits accrue to future residents. One countervailing force is the effects of property capitalization. To the extent that the local housing market correctly capitalizes future taxes and services into housing values, current residents will be affected by inefficient policies and may use the electoral process to provide correct signals to officials.

There is another possibility as well. To the extent that current officials recognize that inefficient current policies reduce the probability of future re-election, they have an incentive — even in the absence of perfect capitalization — to adopt intertemporally appropriate tax and spending policies. Intuitively, the official must care about re-election. What if the particular individual chooses not to run for office again? Then, to the extent that political parties or machines cause the individual to value the election of his successor; i.e., to the extent that political “legacies” are important, the adopted policies may satisfy the conditions for intertemporal efficiency. The vote-maximization approach provides a framework in which to analyze these issues.

Lastly, one could extend the model to situations in which officials care not only about election, but also about the policies themselves. This provides not only a different tradeoff between the tastes of the public and private sector, but also a natural incentive for officials to seek re-election.