Has the Median Voter Found a Ballot Box That He Can Control?

Abstract

This paper examines the appropriateness of the median voter hypothesis within an agenda control framework where the alternative to a proposed expenditure level is the existing level. Evidence from New York State school districts suggests that within this institutional setting the actual level of public service provision does not differ significantly from the median voter's preferred level. The evidence also suggests that demand function parameter estimates for educational expenditures are not sensitive to the assumption that the public service output level corresponds to the median voter's private equilibrium.

Does restricting choice in referenda enable governments to spend more?

“…The empirical evidence on referendums mostly suggests that outcomes are close to the median voter's preference. Holcombe (1980) looked at millage referendums in Michigan and found spending close to that preferred by the median voter, as did Munley (1982Munley ( , 1984 in New York. In an interesting test, McEachern (1978) found no difference in debt levels between majority rule referendums and no referendums, and concluded that without a referendum government officials tended to choose debt levels corresponding to the median voter's preference.…”

“…These are listed in Table 1. The list includes virtually all if not all of the states whose referendums have been studied in the literature: Louisiana (Pecquet et al 1996), Michigan (Holcombe 1980), New Jersey (Megdal 1983), New York (Munley 1982(Munley , 1984, Oklahoma (Dunne et al 1997), and Oregon (Romer and Rosenthal 1978). In these 12 states, the number of elections per year in the typical school district ranges from 0.364 in West Virginia to 2.782 in Oklahoma.…”

Evidence on voter preferences from unrestricted choice referendums

“…Florida's referendums provide direct evidence on the heterogeneity of preferences in each county. Prior research (Holcombe, 1980;Munley, 1982Munley, , 1984 has inferred the distribution from the outcomes of multiple referendums in the same jurisdiction under various distributional assumptions. 3 Florida's referendum system was earlier analyzed by Holcombe (1977), but without the detailed referendum results that were obtained for this paper.…”

Political markets, property tax referenda, and local school spending

“…Since Downs, the median voter model has been the workhorse for empirical assessments of the demand for local public services (Barr and Davis, 1966;Borcherding and Deacon, 1972;Bergstrom and Goodman, 1973;Romer and Rosenthal, 1979;Holcombe, 1980;Munley, 1984;Rubinfeld, 1987). This approach relies on assumptions of majority rule, single-peaked preferences and political competition to make the local median voter decisive.…”

“…Three other states have also had significant histories of property tax-based funding of public schools. Analyses of tax referenda have been reported for Michigan (Rubinfeld, 1977;Rubinfeld, Shapiro, and Roberts, 1987), Florida (Holcombe, 1980), and New York (Munley, 1984;Romer, Rosenthal, and Munley, 1992). 2.…”