Size and Local Democracy: Scale Effects in City Politics

As James Madison would not hesitate to tell us, the scale of a polity or jurisdiction is one of the most basic factors organizing political life. By scale, I refer to the number of inhabitants (or, alternately, constituents or voters) in a political unit, although geographic size may also shape political behavior. Large jurisdictional scale implies that candidates for office must campaign in larger constituencies, necessitating more use of paid media, more fundraising effort, and professional campaign advice. To residents of large-scale polities, government often seemsdistant, remote, and bureaucratic, and the intercession of interest groups, lobbyists, or organized protest activitymay bemore necessary to access or influence public officials. In these andotherways, the incentives, constraints, and opportunities facing politicians and citizens alike tend to differ systematically depending on a jurisdiction’s scale.1 A generation ago, Dahl and Tufte (1973) highlighted some of the ways in which the scale of a polity can affect political participation, political efficacy, and electoral competition, although they readily admitted that a paucity of suitable data made their exercise more suggestive than definitive. Today, although the scale of municipal jurisdictions in the United States varies considerably across the landscape, scale remains oddly under-analyzed as a variable thatmight shape the behavior of local politicians and voters.2 Among the major contemporary works in urban political behavior, only the research of Oliver (2000; 2001) takes considerable account of scale effects. Controlling for other relevant factors, Oliver finds an association between larger city population size and reduced levels of both voting and nonelectoral forms of political participation. He further shows that residents of larger cities are less interested in local (though not national) politics and are less likely to be politically mobilized. Oliver derived these findings from data not initially intended for the study of local politics—the national Citizen Participation Study—and then geocoded the place of residence of the respondents, matching them to local census data. Despite this somewhat herculean effort, an assessment that rests on a widely scattered sample of individualsmay not be an ideal way to examine the nuances of local electoral systems. Overall, little research on local political behavior in the United States approaches the scale issue in a self-conscious way.3 This oversightmay arise from the organization of American political science. There is not really a recognized subfield of “local politics,” per se, in which the size of the local unit is considered as one of its most basic political facts of life. “State and local politics” specialists sometimes treat the local as an afterthought, whereas “urban politics” experts typically limit their focus to large cities, with “large” defined somewhat arbitrarily (Danielson and Lewis 1996). But political boundaries— and the scale of a political unit—surely domore than carve out neutral containers for politics; rather, they help construct the politics that take placewithin them (Weiher 1991; Burns 1994). Recent work in urban politics has gone beyond individual case studies to examine larger-N samples of municipalities, but the cities composing these samples still usually adhere to some prescribed population-size threshold. The insignificance of smaller localities seems to be presumed rather than researched. A few urbanists explicitly examine the suburbs, but their focus has tended toward the features of suburban communities other than scale. Indeed, in an era inwhichmany suburban jurisdictions (particularly in the Sunbelt) have population sizes exceeding those of some traditional central cities, one ought not conflate “suburban”with “small.”4 The city/ suburbdichotomy—if such adichotomyexists—is distinct from, though related to, the question of jurisdictional scale. If scale is a key factor organizing local political life, then attempts to generalize about cities by focusing only on communities above some size threshold, such as 50,000 residents, result in truncated samples. The only reliable way to consider the effects of scale is to assemble and analyze relevant data for cities with a wide range of sizes. This is not to say that one would want to randomly sample all 19,000-plus municipalities in the United States, since most are tiny, with nearly half boasting fewer than 1,000 residents as of 2002 (Christensen and Hogen-Esch 2006, 87). One defensible approach to this challenge would be to sample municipalities of various size ranges in relation to their relative shares of the national population—although ultimately, one’s sampling strategy should be dictated primarily by the research questions posed. Ideally, researchers could use an appropriately constructed local elections data archive for three levels of analysis: (1) the voter, (2) the candidate or campaign, and (3) the political jurisdiction itself. Suchamultilevel approach—potentially incorporating empirical approaches that explicitly take account of the nestedcharacterof thedata—coulddeliver a troveof important findingsaboutpoliticalbehaviorandelectoral institutions.One can sketch out some preliminary hypotheses regarding how jurisdictional scale might matter at each level.