A direct test of direct democracy: New England town meetings

Abstract

Representative democracies govern most locales in the US, making it difficult to compare performance relative to direct democracy. New England states, however, provide an opportunity to test both direct and representative democracy at the local level. This article uses revealed preference axioms to compare spending patterns in New England towns and cities against median voter hypothesis benchmarks. Contrary to previous evidence, we find no differences between direct and representative democracy. The results suggest that horizontal competition arising from local fragmentation minimize differences between direct and representative local government, providing support for wider applicability of median voter-based empirical models of local government behaviour in the US.

Keywords:

• median voter
• direct democracy
• representative democracy
• revealed preference

JEL Classification::

• H11
• H41: H71

Notes

¹ See Turnbull and Chang (1998), Fischel (2001, pp. 87–93), and Turnbull and Geon (2006) for summaries of relevant empirical evidence.

² See the Connecticut Secretary of State's State Register and Manual, the Maine Municipal Association, and the Massachusetts Secretary of the Commonwealth's Guide to Town Meetings for additional details.

³ See Holcombe (1989), Turnbull and Djoundourian (1994), Fischel (2001, pp. 87–93) and Mueller (2003, pp. 243–248) for surveys of this literature.

⁴ Another practical advantage of the nonparametric revealed preference method in this application is that it does not require the large number of observations needed in the econometric approaches. This is important because the numbers of observations for the individual states in our sample are too small to obtain reliable estimates using the standard expenditure demand models in the studies cited earlier.

⁵ The data in Table 1 represents the aggregated data from all towns and cities in each New England state, including those not represented in our sample.

⁶ The theory of structure-induced equilibrium provides one rationale for why multidimensional community decisions, like those associated with multiple tax bases, may replicate median voter equilibria based on the assumption of a single tax base (Shepsle and Weingast, 1981; Niemi 1983). Turnbull and Djoundourian (1994) offer empirical evidence supporting this view, showing that municipal governments that rely on multiple tax sources are no more likely to diverge from the median voter model outcome than those that rely more heavily on a single tax source.