The visible minority, immigrant, and bilingual composition of ridings and party support in the Canadian federal election of 2004 (1).

ABSTRACT/RESUME

Analysis of results among Canada's 308 electoral districts reveals that for the federal election of 2004, the visible minority, immigrant, and bilingual composition of ridings had substantial effects on levels of support for the Liberal, Conservative, and New Democratic Parties, as well as for the Bloc Quebecois. Regression analyses at national and regional levels reveal that riding characteristics--percent immigrant, percent visible minority, and percent bilingual--have impacts on party support that persist when social class composition, measured by average family income and percent with university degrees, is taken into account. The implications of these findings for the 2006 federal election are also discussed.

L'analyse des resultats des elections federales 2004, selectionnes parmi les 308 circonscriptions electorales canadiennes, demontre que la composition ethnique et immigrante des circonscriptions ont eu un effet significatif sur le soutien offert aux libtraux, aux conservateurs, aux nouveaux democrates et au Bloc Quebecois. Selon l'analyse de rtgression aux niveaux national et regional, les caracteristiques des circonscriptions, soit le pourcentage de la population bilingue, des immigrants et des minoritts visibles, influencent le soutien offert aux partis, effet qui persiste lorsque la composition des classes sociales, determinees par le revenu median par famille et le pourcentage d'universitaires, entre en ligne de compte.

INTRODUCTION

In the aftermath of the 1995 referendum on sovereignty, which the "yes" side lost by a razor-thin margin (49.4% to 50.6%), Quebec Premier Jacques Parizeau blamed the loss on wealth and the minority vote. He implied that these alien elements were thwarting the wishes of the real Quebecois (the pur laine), a majority of whom said "yes" to sovereignty. Blaming the loss on "money and the ethnic vote"--at a time when the Parti Quebecois was attempting to be more inclusive--lead, in short order, to the resignation of the premier. Parizeau's fate notwithstanding, the effects of money and the ethnic vote on party support are the subject of this paper.

Over the past twenty years since their iron grip on Quebec was broken by Mulroney's Conservatives (1984) and the Bloc Quebecois (1993), the Liberals have become increasingly dependent for their electoral success on Ontario in general and the multicultural ridings of Canada's major metropolitan areas in particular. This pattern has long been recognized by newspaper columnists and academics alike: "Liberals depend on multicultural votes" (Simpson 2005); in her Hamilton riding, Sheila Copps "had strong ties to the large immigrant community, which traditionally supported the Liberal Party" (Campbell and Christian 1999, 103); recent immigrants tend to support the Liberals (Pelletier 1991, 145). Such scattered observations notwithstanding, Stasiulis points to "a dearth of research in Canada focused specifically on the electoral experiences of immigrants and ethnocultural minorities in electoral politics and the political process at various levels" (cited in Siemiatycki and Saloojee 2002, 242). Those conducting research in the area deal mainly with the extent to which ethnocultural and visible minority communities field candidates and win elections at federal, provincial, or municipal levels or, more concisely, access to the political process through minority representation (Pelletier 1991; Simard 1991; Stasiulis and Abu-Laban 1991). Of course, electoral success by minority candidates requires the mobilization of a supportive ethnocultural vote and substantial voter turnout. Discussion of electoral success by minority candidates rarely makes the Liberal/minority link, since they are selected and elected in all of the major political parties (Dhillon 2005; Siemiatycki and Matheson 2005). Bird (2005, 82) notes that the Liberals won in each of the nineteen most diverse ridings of the Greater Toronto Area in 2004--despite the fact that they ran only four minority candidates there, "while the Conservatives and the New Democratic Party (NDP) each ran six." (2)

Less attention has been paid to the party preferences or voting patterns of immigrant or visible minority individuals and communities. The link between immigrant or visible minority status and support for the Liberal Party is frequently reported, but is seldom the specific focus of the research. In a recent and notable exception, Andre Blais (2005, 823, 834), collapses all Canadian election studies (since 1965) to examine the relationship between religion or ethnicity and voting Liberal. He finds that "religion and ethnicity are important determinants of voting behaviour, and that Catholics and Canadians of African, Asian, or Latino origin are strong supporters of the Liberal Party." The Liberals won most of the elections from 1968 to 2004 "thanks to the strong support of Catholics and Canadians of non-European origin." In effect, "the Liberal Party would not dominate in Ontario in the absence of the strong support it enjoys among Catholics and visible minorities."

The link between ethnicity or race and social class, either as a vertical mosaic (Porter 1965) or as a rainbow class structure (Frideres 2005), has been well established in the literature. Using the rainbow analogy, Frideres points out that class differentiation once based on ethnicity is being reconstituted on the basis of race or colour as recent immigrants (largely visible minorities) face discrimination, blocked aspirations, and a downward spiral. Less well-defined is the relationship between social class and voting behaviour. It has been generally accepted that Canada is low on class voting because of multiple social cleavages based on region, language, and religion, but Lambert and Curtis (1993) argue that class voting appears when you let voters define the class orientation of the parties. Gerber (1986, 127-28) shows clear class effects on party preference in an aggregate- or riding-level analysis that avoids regional cleavages by dealing with Ontario alone: that analysis reveals greater NDP support in highly unionized ridings and greater Liberal support in more affluent and well-educated tidings. In Quebec as well, the wealthy prefer the Liberals. "The Bloc Quebecois is also at a disadvantage compared to the Liberal Party, because it is not the political party that is preferred by the wealthy" (Bernard 2001, 143).

To complicate matters, there are other assumptions about class voting that are not supported by the evidence. One is that union members are inclined to vote for the NDP, considered to be Canada's only working class or left party. Whereas "union members are somewhat more likely to vote for the NDP than nonmembers," it is not due to working-class consciousness as they tend to identify with the middle class. In fact, in "the 1997 federal election, union members were more likely to vote for the Reform Party (20 percent) than for the NDP (15 percent)." Similarly, "only 8 percent of manual workers voted NDP while fully 29 percent voted Reform" (Gidengil 2002, 282-83). Imagine union members and manual workers voting for the far right!

Two studies of class voting that are particularly relevant to this examination turn another expectation upside down. Fletcher and Forbes (1990) examined the impacts of education and occupation on the NDP vote. As one would expect, they found that people of lower occupational status vote NDP more than those of higher occupational status. Education, it seemed, had a weak effect but, contrary to expectations, the most educated respondents were the ones most likely to vote NDP. Observing the same pattern, Nakhaie and Arnold (1996, 187) introduce the concept of a "New Class" to account for the tendency of the more highly educated to vote NDP. These people have university degrees (often in the social sciences and humanities), work in social-scientific, artistic, and literary fields, and are "more interested in social rather than economic radicalism."

Carty and Eagles argue that "politics is local," and that riding characteristics (including percent immigrant) provide the context within which constituency battles take place. Their "bottom-up perspective on Canadian politics" recognizes the fact that "local diversity creates a richly varied field for the practice of democratic politics" (2005, 172). If immigrants and visible minorities are to affect election outcomes and elect minority politicians, they need to be concentrated in specific ridings (Dhillon 2005; Siemiatycki and Matheson 2005) which, indeed, they are. Describing the situation in the late 1980s, Stasiulis and Abu-Laban point out that "the striking feature of the majority of 'ethnic ridings' is their location in urban centres, with the Metropolitan Toronto area having the largest number" (1991, 18). Newcomers to Canada, drawn to specific locations by good economic conditions and concentrations of compatriots, end up in the …