Residential Mobility, Voter Registration, and Electoral Participation in Canada
Brians, Craig Leonard, Political Research Quarterly
This paper tests two propositions advanced to explain reduced voter turnout among the residentially mobile: (1) when citizens have a personal responsibility to re-register following a move, this poses a participation barrier and reduces turnout; (2) the act of moving severs the social networks that normally provide citizens with the information and support to make political choices, thereby reducing turnout. These alternative explanations are evaluated using Canadian national election data. The analysis reinforces previous research asserting the importance of registration barriers in reducing the turnout of those who have recently moved. Additionally, I find that movers' social ties play an independent role in their turnout, with moving particularly attenuating unmarried citizens' turnout. These findings are extended to suggest that recent U.S. initiatives facilitating voter registration may produce less than previously predicted turnout gains among the mobile.
This article expands the consideration of the relationship between residential mobility and reduced political participation, with voter registration regarded as but one of several factors that may individually or cumulatively reduce movers' turnout. Most current mobility and voter turnout research focuses on the U.S. and almost solely considers the nexus between turnout and institutional (i.e., voter registration) provisions, giving little attention to the effects moving has on one's social ties (e.g., Squire, Wolfinger, and Glass 1987). Accurately specifying the relationship between residential mobility and voter turnout has theoretical implications as well as public policy ramifications, given that advocates of the institutional explanation frequently propose plans intended to remedy the supposed ills of U.S. states' voter registration systems (e.g., Wolfinger and Highton 1994; Piven and Cloward 1989).
I propose two alternative explanations linking residential mobility with decreased voter turnout: (1) voter registration costs reduce movers' turnout, and (2) broken social ties depress turnout among the mobile.l These alternatives are then tested in a country (Canada) where individually borne voter registration costs are quite low. This setting particularly facilitates a test of the social effects of moving on turnout, absent elevated personal voter registration responsibilities.
The institutional provisions requiring re-registration following a move may pose a threshold barrier to migrant citizens' voting. In the existing literature this explanation emphasizes American citizens' responsibility to re-register following any move, as a prerequisite to voting (Kelly, Ayres, and Bowen 1967; Rosenstone and Wolfinger 1978; Powell 1982). Only one other modern democracy-Jamaica-requires as much citizen initiative to register to vote as does the U.S. (Powell 1982: 114). Registration requirements are considered a significant factor attenuating the United States' voter turnout, which one would otherwise expect to be quite high given citizens' attitudes and demographic characteristics (Powell 1986; Glass, Squire, and Wolfinger 1984; Rosenstone and Wolfinger 1978). Although several studies empirically relate mobility to depressed turnout (Verba and Nie 1972: 139; Glass et al. 1984: 53, 55; Wolfinger and Rosenstone 1980: 50-54), a single article dominates current institutional explanations of mobility's dampening effects on U.S. turnout. In 1987 Squire, Wolfinger, and Glass reported that U.S. citizens who had moved within the last two years voted 17 percent less than did those who had not moved in over two years, Squire et al. concluded "that mobility has a substantial and statistically significant impact on turnout" due to registration requirements (1987: 52). Additionally, they found that once citizens overcome the threshold of registration they are likely to vote, regardless of their mobility However, this assertion fails the strongest empirical test to which Squire et al. …