Are Marginalized Communities Disenfranchised?

Voter Turnout and Representation in Post-merger Toronto

Bruce M. Hicks 9 novembre 2006

There is scant literature on municipal electoral politics in Canada, but what little exists points to the likelihood that in a larger city, voter turnout would decline and incumbency would become the major predictor of electoral success. These points have been repeatedly raised in debate over the merits of municipal merger and clearly warrant further study. In an examination of the three elections following merger of the City of Toronto, it is found that incumbents had a virtual lock on the new council positions. Significant for local councilors who might be contemplating future initiatives for merger is the finding that incumbency at the regional-level provided a distinct electoral advantage over membership in city and town councils. The most significant finding is that, while voter turnout surprisingly increased as opposed to decreased in the larger ëmegacityí, voter engagement was greatly influenced by socio-economic status, raising new questions about the implications of merger and about voter turnout in Canada in general. 

Introduction

On April 21, 1997, the Legislature of Ontario, after much heated debate, merged six city and borough governments with the government of the Municipality of Metropolitan Toronto, thereby creating a single municipal government to provide local services to over two million people (City of Toronto Act, 1997).1 The large size and entirely urban character of this new municipality, plus the rather vocal opposition to the idea of merger, 2 led to the new City of Toronto being dubbed ‘megacity’.

Since then, a number of other municipalities of varying sizes have been merged throughout Canada, including most of the municipalities in the Province of Quebec and several in the Halifax area of Nova Scotia. So heated was the debate over merger in Quebec that two years after the law came into effect and merger had been completed, the idea of merger was still an identifiable campaign issue in the provincial election of 2003 and a subsequent vote on demerger was held in 2004.

Whether or not the merger trend will continue to spread across Canada, or to other countries, or whether the move in the future will be to follow the example of several Quebec cities and demerge, it is clear that the debate is far from over. It is therefore important that the implications of merger/demerger are properly understood. One of the areas most in need of analysis is the representational question, as elected municipal councils are one of the institutions being merged.

With the last election held on November 10th 2003, Toronto has now had three elections for its megacity council 3, thereby allowing us to examine and draw some conclusions about incumbency and voter turnout in Canada’s largest city, the single-tiered Canadian m’egacity’ of Toronto.

There are two arguments that are relevant to the representational question. Verba and Nie (1972) have suggested that smaller municipal structures might be more effective because people are more likely to know one another and to know the candidate in small communities, thereby creating a direct link between political representation and the sense of community, something that will decline in large, complicated and impersonal cities. On the other side of the debate, Milbrath (1977) has suggested that voter involvement might be higher in large cities because of the excitement generated by larger election campaigns in a concentrated urban environment. Increased media, more resources and professional political communication strategies would combine to mobilize and connect voters to their elected representatives.

Since then, most of the evidence supports Verba and Niei’s contention that voters in the large municipalities will be less engaged. In particular, Kushner, Siegel and Stanwick (1997) found in the province of Ontario that voter turnout was lower, and incumbents had an increasing electoral advantage, the larger the municipality. In fact, they went so far as to suggest that the ‘lower voter turnout that accompanies increased municipal size supports a political argument against consolidation’ (542). It is therefore expected that voter turnout will be lower for the megacity of Toronto than it was for the former municipalities.

With a larger city and thus larger wards, voters are further removed from elected officials. It will be harder for voters to come into personal contact with their candidates during the campaign period or to obtain sufficient information about increasingly complex urban-governance issues so as to make informed decisions. This, in turn, will make voters less motivated to go to the polls.

It is also expected that incumbency will offer a great advantage in the megacity. Sometimes characterized as ‘preferring the devil you know to the devil you don’t’, incumbents benefit from name recognition, appreciation for public service and experience. Incumbency can offer voters a sense of stability and continuity, often there is personal appreciation for a specific service provided to a constituent and sometimes it is tied to vested local interests. Furthermore, the office itself has re-election built into it by allowing (at taxpayer expense) for constituent contact and for feedback on a regular basis in the name of informed representation.

Given the ethnic diversity of Toronto, election campaigns can be complex. Campaign literature often needs to be in multiple languages, canvassers may need to speak a variety of languages and sometimes cultural differences need to be addressed when navigating the minefield of local politics (where zoning and by-law changes often pit community against community). Incumbents, however, enter a campaign already organized to deal with linguistic and cultural community challenges, having had to function in this political environment on council, usually with the advice and support of the city bureaucracy and paid political advisors (not to mention such basic services as translation). They will often have previously established networks among community leaders.

It is therefore expected that the complexity and diversity of a megacity will give incumbents a strong advantage. Furthermore, it is expected that the larger and more diverse the city, the larger and more complex the wards, the greater the ‘disconnect’ will be from the voters. This should manifest itself in lower voter participation.

Research

There is surprisingly little academic research on municipal elections in Canada. Major metropolitan areas have been studied (Magnusson and Sancton, 1983) and a few specific subjects have been explored in detail such as gender (Gidengil and Vengroff, 1997) and ward versus at-large elections (Tennant, 1980). There have been two studies that addressed the determinants of electoral success at the councilor level in Ontario municipalities, using the years 1982, 1988 and 1994 (Kushner, Siegel and Stanwick, 1997), and for the first mayoral election in the merged city of Toronto (Stanwick, 1997).

This lack of research is disappointing. One would think that, with ward boundaries drawn around homogeneous communities and the lack of formal party identification and apparatus, municipal elections commend themselves to the study of such things as the impact of socio-economic factors and incumbency (research items that have been of great interest to Canadian political science at the federal and provincial levels).

The primary reason there has been so little study is because of the difficulty in collecting and analyzing municipal election information. For the most part, municipalities administer their own elections, so researchers must contact each municipality separately and rely on the good graces of city officials for data and information. With the numerous re-organizations of municipalities during the 1980s and 1990s, ward maps, council composition and methods of selection were frequently re-engineered, which is a further impediment to analysis. With merger, this has been complicated yet again since records from previous councils are not usually turned over to the successor municipality (since in law this is a new corporation).

To undertake this study, information pertaining to the candidates in the 1994, 1997, 2000 and 2003 elections was taken from newspapers and from whatever city records could be found.4,5 Election statistics and boundary maps were provided by the City of Toronto Clerk and the Director of Elections, and their respective staffs, as well as the Toronto Archives.6

Census data were collected to measure the following population characteristics: English mother tongue, total visible minorities, percentage of immigrants, mobility based on five years, percentage of rented private dwellings, less than grade 9 education, university education, unemployment rate, average household income and median household income, and percentage of low income incidence in economic families. These socio-economic data were derived from the 2000 census, which the City of Toronto had commissioned Statistics Canada to generate for each of the 44 Toronto ward boundaries.7 These census data were applied to both the 2000 and 2003 ward turnout in an effort to determine if, and how, socioeconomic factors impact on voter turnout.8 Voter turnout figures were determined based on the number of eligible electors identified per the voters list, plus revisions at the poll, contrasted with the number of people who actually voted.9

For the purposes of this study incumbency is defined as being a member of the council of either Metropolitan Toronto or one of the six city and borough councils, or of the new megacity council, as of the day of the election (Nov. 10th of that respective year). Councillors and mayors who had served before but were not on council at the time of the election are not treated as incumbents, nor is a person who was serving on any other elected body, such as a school board or provincial legislature, at the time of the election (unless they were simultaneously serving on one of the seven councils).

Voter Turnout

Turnout for municipal elections is almost always lower than for provincial or federal elections (Drummond, 1990: 250). And larger cities typically have lower turnout than smaller municipalities (Kushner, Siegel and Stanwick, 1997: 542). Nevertheless, voter turnout went up in the 1997 Toronto municipal election, the first for the newly merged megacity.

In 1994, the last election prior to merger, there was a turnout of 35.3 percent in what would become the megacity of Toronto. This is in keeping with the provincial average of 37 percent for municipalities with a population of 100,000 or more in that year (Kushner et al., 1997: 542). Turnout for the megacity of Toronto in 1997 was 48.68 percent (up 13.38 points or a 40 percent increase). And, as Table 1 shows, an increase in voter participation is true for each of the former six city and borough areas.10

The first merger election in 1997 was seen as a battle between the mayor of the old City of Toronto, Barbara Hall, largely identified with the New Democratic Party, and the mayor of the old City of North York, Mel Lastman, largely identified with the Progressive Conservative Party.11 Since both the cities of Toronto and North York contributed roughly equally to the population of the new megacity,12 even without partisan differences one could have expected regional loyalties to make this a tight race.

TABLE 1 Voter Turnout Before and After Merger

When there is a tight race between front-running candidates, voter turnout will be higher (Tindal and Tindal, 1995: 247). It will also be higher when there is a highly partisan contest (Karnig and Walter, 1983: 491505). There were also between five and seven referendum questions in 1997, and these have been shown to increase turnout (Tindal and Tindal, 1995: 247). Furthermore, this was the first election following merger, and merger had been strongly opposed, resulting in this vote being ‘the culmination of roughly one year of extraordinary high profile politics which never lost momentum’ (Stanwick, 2000: 556).

TABLE 2 Voter turnout within the Megacity of Toronto

If the 1997 merger election were an anomaly, we would expect voter turnout to decline dramatically for the 2000 and 2003 elections. Table 2 shows that voter turnout was, in fact, much lower in the two subsequent elections for each of the former municipalities.13

However, voter turnout has not dipped below the 1994 level of 35.3 percent. In 2000, voter turnout was 36.1 percent and in 2003 turnout was 40.2 percent.14

Table 3 shows that while the 2000 and 2003 elections witnessed a decline in voter participation over the original merger election in all of the former municipalities, there was in fact an increase over the turnout numbers experienced in the pre-merger election of 1994 in all but one municipality. What is more, there appears to be an increase in half of the city over the record highs for turnout in the former municipalities.

TABLE 3 Change in voter turnout

There are several caveats worth noting, aside from the obvious one that the numbers in Table 3 are more dramatic because they are the percentage of change as opposed to simply the number of points that have shifted. The cities in question were all in the category of large municipalities before merger, as defined by Kushner et al. (1997): that is to say they each had a population of 100,000 or more and were already experiencing lower turnouts. The previous record highs are, by definition, anomalous turnout results unique to each of those elections in which they were registered.

That being said, these figures would seem to indicate that voter turnout in a megacity might increase, in spite of expectations. Milbrath (1977) argued that voter turnout would be higher in urban environments because the political excitement would encourage participation and mobilize voters. While research has shown that large municipalities have lower voter turnout than small municipalities, perhaps there is a critical mass of urban size that creates enough excitement and awareness about the election to overcome the disconnect with the individual voter that is the expected side effect of larger city and ward size.

a) Socio-Economic Factors

After some initial evidence that was presented to the Royal Commission on Electoral Reform and Party Financing that socio-economic factors may have an influence on voter turnout in Canada (Bakvis 1991), the general consensus that has emerged among Canadian political scientists is that they are not having a significant impact on voter turnout, though they continue to be part of the debate over turnout in other developed democracies.16 Part of the reason for this is the lack of class-based politics and identity in Canada, though it is undoubtedly also a by-product of how electoral boundaries are drawn at the federal and provincial levels. Vastly different communities are often combined into a single constituency, creating an almost median constituency profile within regions or cities.17 The smaller wards at the municipal level, with boundaries more closely coinciding with homogeneous self-identified ‘communities’, should lend themselves to a closer examination of SES factors and how they might impact on voter turnout in Canada.

Table 4 reports Pearson correlation coefficients measuring the linear relationship among the socioeconomic characteristics of the wards and the voter turnout for each of the 2000 and 2003 elections (with their significance in parenthesis).

What we find is that turnout was significantly higher in wards with a greater proportion of high-income earners and a lower incidence of low-income households. Similarly, turnout was significantly negatively correlated with the unemployment rate in each ward. While the correlation with the percent of rental units was not statistically significant, it is in the direction consistent with the argument that in municipal politics (where taxes are collected on property) the property owners are more aware of municipal taxes and more concerned with the tax rate and the services the city provides. The fact that turnout was higher in areas where mobility was lower also lends credence to this conventional wisdom.

In wards with proportionately greater immigrant populations and visible minorities, turnout was significantly lower; and in wards with a greater proportion of people whose mother tongue was English, the turnout was higher. Language and culture create barriers for new Canadians that make it harder for them to fully participate.

TABLE 4 Relationship between Voter Turnout and Ward Demographics

Education is another item that one would expect to have an impact on voter turnout. Not surprisingly, in wards with proportionately more university-educated residents, turnout was significantly higher and in wards with proportionately more people who had not finished grade nine (among people 20 years of age or older), the turnout was lower. This is in keeping with the argument that information costs (i.e. the costs of obtaining, sorting and processing information) are higher for people with lesser education.

There is not only some degree of overlap and redundancy in some of these variables, but there can also be seen to be a causal direction. For example, greater education can lead to higher income and, since highly educated people earning higher incomes are able to afford more expensive homes, property taxes might increase the perceived importance of electoral participation for this group. Unfortunately, because of the relatively few number of cases available, it will not be possible to disentangle these variables. However, it should be possible to come up with a predictive model.

b) Predicting Voter Turnout

Up until this point we have concentrated on the independent effects of voter-turnout and socio-economic factors. A regression model will now be estimated to determine the relative impacts of socio-economic factors on voter turnout.

Table 5 shows a number of high correlations, such as visible minorities, immigrant ratio and English as a mother tongue, or unemployment and low-income households. To avoid the problem of multicollinearity, the regression model has been reduced to four factors that are representative of specific clusters. These are items expected to influence voter turnout and are quite distinct from one another.18 This model has been run for both the 2000 and 2003 elections.19 An income variable was not included because it was found to be too closely correlated with education, in particular, but also with mobility and visible minorities.

TABLE 5 Correlations among Socio Economic Factors for the wards in Toronto Municipal Elections

TABLE 6 Multiple Regression Model Predicting City of Toronto Voter Turnout

 

This model is quite robust. Within it, non-mobility is the most significant factor in explaining voter turnout when the other items have been controlled for, but each of the other items (except for number of candidates in 2003) remains statistically significant.

As mentioned above, homeowners receive a tax bill and are aware of the cost of municipal government (renters never see a tax bill since it is incorporated into their rent). What is more, the longer a person stays in a community the more likely it is that she or he will connect with that community. Both of these motivating factors for municipal electoral participation would be captured in the mobility variable.

c) Number of Candidates

The relevancy of the number of candidates is deserving of specific comment. In the bivariate analysis the number of candidates running was not statistically significant. What is more, it appeared to have less of an impact in the 2003 election than in the 2000 election. When included in the regression model, the number of candidates was significant in 2000 and appears to have had more of an influence on voter turnout than the other variables and certainly more than it did in 2003 (though with a higher standard error).

Most voting behaviour analysis works from the assumption that a ‘voter’ may be motivated by a number of factors, and researchers attempt to find a model that properly predicts this theoretical individual’s behavior. However, it is quite possible that different factors motivate different voters. If we postulate that there might be a segment of the population that will be mobilized by the excitement of politics – let us call it the ‘politically interested cohort’ – this might explain what is happening statistically with the ‘number of candidates’ variable.

While the number of candidates appears to have little impact on voter turnout on its own, when other mobilizing influences are controlled for (in this case visible minorities, education and non-movers) there appears to remain a segment of the population mobilized simply by the excitement of the campaign. This might also explain why this variable was less significant and of lower influence in the 2003 election in the regression model since that election had a hotly contested mayoral race. The ‘politically interested cohort’ would have been mobilized across the city in 2003, making the impact of multiple candidates in any one ward less significant.

At first glance, ‘politically interested cohort’ might seem to be a misnomer, since those persons genuinely interested in politics – persons the media and public discourse often refer to as ‘political junkies’ – will likely be mobilized by any election campaign. However, this segment of the population is likely composed of ‘soft voters’ who have no strong self-interest mobilizing them to vote, such as drives Downs’ (1957) homo economicus. These people would become aware of the election because of the increased excitement generated by the campaign (either because of a plethora of candidates in their ward or a high profile mayoral contest), and would get swept into the voting booth on election day by sheer momentum, generated by political interest, rather than economic self-interest (or a sense of civic duty).

d) Analysis

Clearly socio-economic characteristics, such as education/income and ethnicity, are having an impact in Toronto municipal politics. Part of this might be because of the electoral lists. Electoral lists are generated from tax rolls,20 so the more marginalized population that would include many immigrants, people who don’t speak English, visible minorities, unemployed, low-income families, renters and people who have moved recently, will be underrepresented on the list. While this does not prevent these people from voting, there is a mobilizing dimension to the electoral list whether it be generated through the personal contact of enumeration or simply receiving election literature (such as polling station information), properly addressed to the voter at her or his correct address (Black, 2003). These omissions have a ripple effect since canvassing by the respective campaigns usually is based on the electoral list and missing names will result in campaigns skipping that residence for telephone and door-to-door canvassing and, therefore, not including them on election day when ‘getting-out the vote’. Finally, the energy required from the voter to have her or his name added to the list at the polling station may be seen by some as too great a barrier to voting.

Why the correlations become stronger for the variables in 2003 (something that is also seen in Table 4 with the bivariate relationships) is unclear. Voter turnout was higher in this election and one would expect that in higher turnout elections socio-economic differences would be less visible (obviously with 100 percent turnout there would be no evidence of socio-economic differences). That being said, it is worth keeping in mind that ‘higher’ turnout at the municipal level still means that less than 50 percent of the people on the electoral list voted.

One explanation for the differences on how SES factors impacted on these two elections might be found in the events of the election itself. Mayor Lastman, who had won the election in 1997 over Barbara Hall, did not run again in 2003. ‘Right-leaning John Tory’ came from ‘single digits last winter. . . close to pulling off an upset win’ over ‘left-leaning’ councilor David Miller (Globe and Mail, Nov. 11, 2003: A14), which may mean that particular SES constituencies were mobilized specifically in that election.

This hotly contested and partisan competition at the level of mayor would explain the higher voter turnout in 2003 over 2000. It could also explain why the number of candidates at the councilor level was not significant in our voter turnout model for 2003, and yet it was for 2000 (as noted above).

Another possibility is that there is an increasing disconnect between certain segments of Toronto and municipal politics. Obviously, with only two elections to study a trend cannot yet be predicted, but this warrants further study.

In addition, the Milbrath (1977) mobilization theory that could explain the higher voter turnout in megacity Toronto over its component cities does not presume that everyone mobilizes equally. The larger city with more complex wards could act as a deterrent to specific segments of society. Clearly there is an imbalance in participation rates along socio-economic lines in both elections.

Incumbency

Most of the research into incumbency in Canada has been done at the federal and provincial levels, where it has been found that incumbents probably have an advantage because of name recognition, experience and better access to campaign financing (Krashinsky and Milne, 1986). This translates at the municipal level, where there is no formal party identification,21 to a strong advantage, something even more significant in larger municipalities (Kushner et. al., 1997).

One of the advantages attributed to incumbency is that ‘incumbents normally do not have the disadvantage of having to run against an incumbent themselves’ (Marland 1998: 34). However, the election of 1997, which combined the Toronto Metropolitan (or ‘regional’) government with the six local governments, reduced representation and pitted incumbent against incumbent. This makes the issue of incumbency in that election particularly deserving of examination.

a) Merger Election (1997)

In the process of merging the six city and borough governments with the government of the Municipality of Metropolitan Toronto, the number of elected council positions went from 101 down to 56 plus the mayor.22 In what can best be described as a political version of ‘musical chairs’, 83 incumbents divided themselves up relatively evenly and competed against one another for the remaining seats left at the municipal level.

TABLE 6 Comparative success rate between incumbents and non incumbents

With so many incumbents running against one another it is not surprising that so few non-incumbents were successful. Table 6 shows that with 193 people competing for 56 councillor positions in the merger election, non-incumbents had less than a two percent chance of success.

A word of caution should always be expressed when examining chances of electoral success at the municipal level. With few barriers to entry, ‘frivolous’ candidates will invariably distort the results.23 However, even allowing for candidates who are running for attention or the novelty, running on a single issue or running because of a mistaken belief that their ideas or candidacy have a following, there was no shortage of qualified candidates who sought election to the new megacity council.24

b) Regional v. local incumbency

The idea of merger arose in the 1990s where buzzwords like ‘eliminating duplicatioon’ had become part of the Canadian political lexicon and the commonly used reference to local and regional government as ‘twotiered’, which implied duplication, brought municipalities and their services under increasing public and government scrutiny. It is therefore worth exploring whether or not incumbency on either tier of government provided an electoral advantage.

TABLE 7 Comparative success rate between regional and local councillors

Prior to merger, there were 34 people who served on the Metropolitan Council for the greater Toronto area. These included the mayors of the six cities and boroughs, the councillors from the city of Toronto and councillors elected from the other cities specifically to serve on metro (elected either city-wide or for several combined local wards).

Table 7 shows that an incumbent coming from the metro or upper tier had better chances of electoral success in the merger election than did a local councillor. At face value this appears contrary to the old adage that ‘all politics is local’ and the belief that elected representatives in small communities have a stronger bond with their constituents. However, it is important to keep in mind that this is a reflection of the chances of electoral success in the new council with its larger wards, not a comment on respective connectivity between voters and their representatives at either of the two previous levels.

The new 1997 council had only 28 wards, with dual member representation.25 These were larger, more complex and diverse wards than all but a few candidates had campaigned in and represented before merger.

c) Megacity

Table 8 makes it apparent that incumbency continues to be an overwhelming determinant for electoral success in the megacity. While non-incumbents had a slightly better chance of success in the two subsequent elections than they did in the merger election, it is well below the 14 to 16 percent average experienced by non-incumbents in large municipalities across Ontario (Kushner et al 1997: Table 2), though the success rates of incumbents is not out of proportion.

TABLE 8 Success rate of incumbents subsequent to merger

The lower success rate for non-incumbents (without a proportional change in success rates for incumbents) may be an indication of a lower attrition rate in a megacity, possibly due to higher salaries and a perceived increase in status of megacouncil membership (making the position attractive as a career and less a stepping stone for higher political office).

The electoral success of incumbents was higher in the 2003 election. In part, this is because in the 2000 election the city of Toronto made the transition from 28 dual-member wards to 44 single-member wards, resulting (once again) in several incumbents competing against each other for seats on the new council.

However, it is possible that the increase in success rates for incumbents is due to the fact that elections in the megacity increasingly favour the incumbent as the city becomes more complicated and impersonal. As with voter turnout, however, it is not possible to forecast a trend with only two elections. What we can state definitively though is that in a megacity incumbency matters, providing a virtual lock on electoral success.

d) Analysis

In the 1997 merger election, incumbents running from the metro or regional level had a distinct advantage. It is possible that the voters considered the candidates who served on metro more qualified for the new positions. The delivery of services in Metropolitan Toronto had become more centralized over the previous decade and it would not be unreasonable to view merger as the absorption of the local governments into the larger entity of metro Toronto.26

It is also possible that this is simply a reflection of the increasing power of incumbency in the megacity of Toronto. Councillors running from metro would have higher name recognition and quite likely would be perceived as being more ‘senior’ by the voters. They would have larger election machines already in place having run in larger geographic areas than their local councillor cousins, familiarity with more of the communities contained within the very large and ethnically diverse megacity ridings, and easier access to resources (plus previous experience raising that sort of money).27

These advantages of incumbency would in all likelihood favour the metro councillor over the local councillor in a merger election and continue to help incumbents against outside challenge in the large wards of the megacity.

It would appear that voters may be relying on the shortcut offered by the slogan ‘re-elect’ to augment the lack of necessary information to make their selections in the large wards of a more complex megacity.

Policy Prescriptions

One must be cautious about extrapolating from this evidence and making unwarranted conclusions or rash recommendations. After all, these results technically apply only to one city and, even there, to three specific elections. They may not translate to other cities or levels of government, and may not hold over time. For example, it has already been noted that the little research that exists at the municipal level has found in the past that large cities have lower voter turnout, and yet voter turnout was higher in ‘megacity’ Toronto. We have advanced a plausible explanation for why that might have occurred, but the conflicting evidence also urges caution from both a comparative and temporal perspective,

It is usually an ecology fallacy to use data generated at the groupor aggregate-level and attempt to draw conclusions about individuals. For example, if neighborhoods with high rates of unemployment also have high crime rates, one cannot conclude that the unemployed people in neighborhoods are committing crime. However, it is accurate to state that there is a higher incidence of crime in these neighbourhoods (this is a simple observation of fact) and all we have suggested so far is that wards with a greater number of renters or mobility in residency, and wards with a high number of visible minorities, have lower turnout. As for the extrapolation that these groups might be the specific ones being disproportionately disenfranchised, this is probably a reasonable assumption since the correlates are high, the wards are somewhat homogeneous and there is strong theoretical and intuitive justification for what might be causing the fluctuations in voter turnout along socio-economic lines. Nevertheless, caution is also urged.

One must also be cautious about taking findings at the municipal level and extrapolating to the provincial and federal levels. We have earlier suggested that the larger federal and provincial ridings might be masking the impact of SES factors on voter turnout in Canada. While this is likely true (and there is much anecdotal evidence to support this statement), it is also true that lower voter turnout can accentuate disparities. In other words, at the municipal level, where voter turnout is only in the 30-40 percent range, it is more likely that upper income or higher education will have a more dramatic impact on the election (something picked up first in the voter turnout figures) than they will have in a federal election where turnout is fairly high and where persons from all classes and income groups become mobilized. While voter turnout at the federal level has dropped from the post World War II average of 75 percent, it is still at the cusp of 60 percent, which would tend to indicate that SES differences might be mitigated by higher turnout, though even in these cases it is likely that economic factors are disproportionately impacting on voter participation.

In spite of these cautions, the results of this study are particularly disturbing. Voting is a right. It is even enshrined in section 3 of the Canadian Charter of Rights and Freedoms. While it is accepted by many people that rights are subject to individual choice – in other words a person has the right to vote or not vote since an abstention is in some ways equally a vote (which is the common argument against introducing compulsory voting in Canada) – if that right is not freely accessible to all members of society, then the legitimacy of the government, the fairness of the system and the equality of the society come under suspicion.

Clearly there is evidence that persons of lower income and less education, people who are renters and frequent movers, and persons who are visible minorities, new Canadians and non-primary English speakers, are voting less, at least in Toronto elections. This has to be troubling for the municipal government collectively and for the councillors who, once elected, will want to lay claim to a mandate from ‘the people’. The relatively low voter turnout at the municipal level frequently poses challenges for local politicians to claim a mandate, but the standard political response to media questioning about their legitimacy is that the ‘voters had a choice’. If there is evidence that some voters are systemically disenfranchised because of race, income or education, then the validity of that claim becomes circumspect.

There are two steps that could be undertaken that would instantly increase, perhaps even double, turnout at the municipal level, and these would be to:

  • Make voting compulsory. In over 30 countries where some form of compulsory voting is in place, voting turnout is on average higher and significantly higher for secondary and tertiary elections.28 It should be noted that what is compulsory is not voting for a particular candidate but attendance at a polling station (voters are often given the option of selecting a box indicating ‘none of the above’ or alternately spoiling their ballot) and the sanctions tend to be small, ranging from having the names of recalcitrant voters published on a list to a nominal fine sporadically enforced. Nevertheless, some opponents of compulsory voting argue that this infringes on freedom of choice (i.e. the choice to not participate in a democracy). It can be equally argued that the opposite is true: that refusing to vote is a genuine act of expression when done in defiance of sanctions. An alternative could be to use financial incentives instead of sanctions such as a voters’ tax credit (see Hicks 2002).

  • Use provincial (or federal) political party labels. The names of political parties at the federal and provincial level provide an information shortcut for voters enabling them to identify candidates/parties that generally support policy perspectives similar to the ones they hold.29 Furthermore, political parties often have ready-made workers in the form of party members who will actively support candidates running under the party label. The result is that in countries where political parties are active in municipal politics voting turnout is significantly higher.

In addition, there are a number of steps that can be undertaken to address the specific problems arising from this new evidence:

  • Decouple the electoral list from the tax rolls. Homeowners should not have the advantage of being pre-registered to vote and personally informed by mailings on where to vote (when and how). Some steps have been taken in some provinces to augment the tax lists by mailing enumeration forms to property owners and asking them to provide the names of the people who occupy their property or properties.30 While this is designed to capture voting age children, tenants and elderly dependents, it still leaves the system in the hands of the wealthy landowners.

  • Take electoral list production out of the hands of property assessment departments. This is more than a decoupling of the list from the tax rolls. By having the department that manages tax assessments also manage the voters’ list, information related to a basic right – the right to vote – is being collected and managed by a government department the sine qua non of which is tax collection. This raises some serious privacy concerns and these, in turn, result in administrative challenges.31 This is unnecessary since each municipality has its own electoral officer and each province has an electoral officer, many of whom maintain permanent voters’ lists.

  • Ideally door-to-door enumeration should be done at election time. Since there has been a shift toward permanent electoral lists at the other levels of government, it is unlikely that there would be support for a return to wholesale enumeration. But targeted enumeration, in areas where there is a large number of renters/movers and in specific linguistic and ethnic communities, could address weaknesses in electoral lists and have the added benefit of encouraging voter participation in these lower-turnout areas.

  • Electoral lists should be shared between all levels of government. Increasingly federal and provincial governments have moved to permanent voters lists. The federal and provincial electoral commissions should share the information on these lists with the electoral officers of municipalities.32 In fact, since cost savings seem to be a preoccupation of all levels of government these days, the maintenance of these lists could be centralized with either the federal or provincial electoral commission to avoid duplication and wasted resources.

  • Renters should be given a tax receipt informing them of what portion of their rent goes to property tax. This was required of landlords in some provinces (e.g. Quebec in the 1980s) in order for lower-income renters to apply for provincial property tax credits, though later eliminated to save landlords an unnecessary paper burden.33 While not as effective in driving home the point that municipal politics has a related cost as the requirement of writing a cheque to the municipality (something landlords must regularly do), it would have the benefit of letting renters know that they are paying property taxes that are buried within their monthly rent bill.

  • Draw electoral boundaries around similarly situated groups. Similar turnout should be used to group voters. This will ensure equal voice among voters by preventing domination of one area by another. For example, if a community where voter turnout is only 20 percent is coupled with a community where turnout is 40 percent, candidates will focus only on the latter area; however, if the entire ward has roughly the same voter turnout (even if it is only 20 percent across the ward), then no one community will come to dominate and at the council level representation will be equalized.

  • Give people time off work to vote in municipal elections. Increasingly governments have moved toward setting voting hours to ensure that polls are open long enough in the evening so that employers do not have to give any time off work.34 The motivation for this change was not the fact that employers routinely violated this obligation to provide time off (though they did), but rather to save employers money. Yet even with this change, those who are least likely to be able to take advantage of the current voter times are those who work shift-work, temporary employees and those who work menial cash and part-time jobs – in other words, the most marginalized in society. The concern for the cost of employers fails to take into consideration that there are tangible costs associated with voting (transportation, babysitting, etc.) that will, by definition, impact disproportionately on lower-income Canadians. Time off work will partially compensate for these costs and create a sense of obligation (even guilt) compelling some to vote.

  • Since there is evidence that some voters will mobilize because of increased excitement over a campaign, then priming interest in a campaign might help to mobilize participation. The state can certainly advertise more, but a more effective approach would be to get the media to promote elections in a more serious way. This is admittedly a challenging idea since the media should not be subject to state interference, but there is no reason the city section needs to be the third part of the paper any more than the provincial should be second (especially at election time). It is possible that the media could aggressively promote the fact that elections are taking place without being co-opted by the state or by a candidate.35

  • Language translation and ethnic community outreach. At whatever level, the state needs to put more resources into ethnic community mobilization. It has been assumed by those who study federal electoral participation that ethnic communities may mobilize more easily than the majority population.36 However, this latest evidence is that the rank-and-file members of a community are not easily mobilized and that outreach may be essential to bridge the communication and cultural gap. This could take the form of active enumeration and a marketing/outreach campaign to mobilize voters.37

  • Renter outreach and mobilization. The lack of a sense of community among the most mobile of the population poses a particular challenge. Cities have increasingly moved to organize and fund community groups, either through grants or tax credits, including business improvement associations and community residency associations. If the corporate world is deserving of financial incentive in order to achieve a business goal of improving the marketability of a business district, it seems appropriate to provide similar incentives to renters both to engender a sense of community and to improve the physical environment of their neighbourhoods. At the building level, housing co-operatives have been successful ways to generate a sense of ownership among renters that has translated into tangible building improvements, and this idea could be expanded to the community level.

As pointed out earlier, the implications of this evidence make the lack of similar findings at the federal and provincial levels suspicious. We have noted that greater voter turnout and larger ward size might conspire to either mask or partially eliminate the impact of these factors on voter participation. If it is the former, this is particularly troubling, especially since its existence is likely known by political strategists and candidates and just not seen by the public, the media and academics.38 Therefore, more research needs to be done on how SES factors might impact at the strata of federal and provincial elections. An examination of voter turnout at the polling station level (as opposed to the larger riding level where multiple communities are consolidated) would quickly shed light on how social and economic differences might be impacting elections, and any impact that shows up in voter turnout would be a good litmus test of a larger impact on electoral influence.

Turning to the issue of merger politics and the broader issue of representation, there are a number of disturbing issues raised by the virtual stranglehold that incumbents have on municipal politics and that regional councillors have on merger elections. Any steps taken to obtain higher voter turnout overall (through something like compulsory voting) will undoubtedly change the dynamics of municipal elections. In addition, several steps can be undertaken to level the playing field, and thereby open up municipal politics to women and minorities in particular:

  • Use political party labels. While the use of party labels has been shown to boost voter turnout (as previously noted), it also provides a counterweight to incumbency. Incumbents are not always considered by the majority of electors to be the best choice for re-election, depending how they performed while in office, but the advantage of name recognition in a wide field of candidates can often be a prohibitive obstacle to change. The information shortcuts provided by party labels, even exclusively municipal political parties, permit voters to easily consider alternatives and help to identify relative positioning. As a result, political parties narrow the field and permit regular turnover as voter interest and priorities shift. Furthermore, provincial (and federal) political parties have over time developed mechanisms to aggressively recruit women and minorities to run as part of their teams.

  • Eliminate the label ‘re-elect’ in merger elections. The expression ‘re-elect’ on signage and campaign material is intended by politicians to be an information shortcut to illustrate to voters that they have held the position before, thereby implying a history of service, already established credentials and experience, and engender appreciation for work already done. In a merger election, several candidates can lay claim to being the incumbent and yet none of the candidates is, technically, running for re-election since the same positions no longer exist. Therefore, the label is misleading and an information shortcut should not be allowed to be substituted for the actual advantages coming from service that will be enjoyed by a truly hard-working councillor.

  • Consider ethnic communities in drawing ward boundaries. As noted in the previous list of recommendation, drawing ward boundaries along similarly situated communities in terms of voter turnout would equalize representation across wards. Similarly, ensuring that ethnic communities (whose voter turnout numbers appear to be lower) are not overshadowed within a ward will facilitate greater representation for these communities. If socio-economic differences are resulting in differences in voter behaviour, whether it is through turnout or vote choice, it is important that this not be masked. While masking it might create the impression of harmony in society, disparities beneath the surface go unaddressed and can lead to more serious divisions down the road. By giving groups their own representation, it will ensure that all people are represented on council regardless of race and ethnicity. It is through this mechanism or, alternately, a set-aside of seats for specific ethnic groups that many of the national legislatures in other countries have been able to remedy the under-representation of minority groups.

  • Create dual-member constituencies. This is the easiest and most efficient way to equalize representation along gender lines, thereby ensuring that one representative in each ward is male and another is not male. Dual member constituencies have existed in Canada since before Confederation, though the division was rarely along gender lines (for example, in PEI the division in voting was between landowners and non-landowners even after the vote was extended to women). The new territorial legislature of Nunavut, in Canada’s north, would have been the first legislature to create dual-member ridings divided by gender in 1999, but the referendum only obtained the support of 43 percent of the population. Ironically, many municipalities already use dual member wards, including the new city of Toronto between 1997 and 2000, so this would be a relatively simple change at the municipal level.

  • Language translation services. Candidates should be provided language translation services for election materials (flyers, websites etc.) to enable them to communicate their message to the voters and reduce the advantage that incumbents have through in-house translation services.39

  • Financial assistance for women and minority candidates. It is well established that women and minority candidates have greater difficulty raising funds in politics and this is truer at the municipal level where incumbency provides a natural advantage for fundraising. Providing financial assistance to candidates running from under-represented groups, perhaps tied to a minimum-vote threshold so as not to be seen to encourage ‘frivolous’ candidates, would help to correct this imbalance.

  • Offer provincial income tax receipts for municipal political donations. Tax receipts are rarely issued in municipal politics making donations the purview of the wealthy or of those interested in specific policy outcomes and by-law changes. In the few municipalities where tax receipts are issued for donations they are offered as a credit against municipal property taxes, which is a further advantage and incentive to landlords and developers. Offering provincial income tax receipts, with the requisite sliding scale and donation limits that have become de rigueur in Canadian politics, would encourage greater levels of financial contribution by a larger portion of the population. This, in turn, would help more candidates mount effective campaigns.

Conclusion

There does seem to be a ‘disconnect’ between the voters and their megacity government, though not exactly in the way predicted.

Contrary to expectations, voter turnout appears to have increased as a result of merger. This may be due to the increased attention that politics in a megacity generates, specifically the campaign for megamayor, which seems to have more influence on voter turnout than the excitement generated by local campaigns. The fact that the mayor of Toronto is directly elected by more voters than any other elected official in Canada and that most of Canada’s media is concentrated in Toronto, would certainly contribute to mobilization.

However, voter turnout is still below 50 percent. More importantly, voter participation at the municipal level is not uniform. Marginalized communities are being disproportionately disenfranchised in Toronto politics either because of administrative barriers like the voters’ lists, the campaigns of the candidates themselves or the complexity and size of the megacity, its wards and issues.

That higher income people might be somewhat engaged in municipal politics is expected given the way municipalities are funded and the decisions municipalities have the power to make. But property issues, such as taxes and zoning, are not the only responsibilities of municipalities. Some provinces have increasingly downloaded services to this lower level of government (one of the most significant being welfare in Ontario). An imbalance in participation along socio-economic lines means that certain groups are not having a full say on which people and policies can materially impact on their lives.

The sheer number of incumbents running in a merger election results in a virtual shutout of newcomers to municipal politics. And incumbency at the regional level seems to give a candidate a distinct electoral advantage in that initial merger election. What is more, incumbency continues to be an overwhelming advantage for electoral success in a megacity following merger. All this points to the likelihood that in the larger and more complex wards of a megacity the benefits of incumbency, such as name recognition, access to resources and the ability to mount a campaign in a large ethnically diverse urban setting, are formidable advantages.

The fact that incumbency is such an advantage to electoral success in Toronto may be of legitimate concern to women and minorities who are currently underrepresented at every political level. If new candidates are unable to break into municipal politics, it is unlikely that imbalances will be corrected any time soon, particularly if the early indication that attrition rates are lower in the megacity turns out to be a reality. This can have even larger implications given that municipal politics has frequently been the avenue by which women and minorities obtain the experience and credentials to break into higher office.

It is even possible that this might be contributing to socio-economic disparities in voter participation. If certain segments of society do not see themselves represented in government, they may feel less connected to that government, which in turn may make them less willing to vote, which might discourage others from seeking office, and so on.

A decade after merger it is clear that there are representational implications to the creation of a megacity. Whether or not there is a directional trend remains to be seen.

Yet in spite of the troubling undercurrent revealed through these findings, it is within government’s ability to take remedial steps. The changes suggested in this paper range from the administrative to the structural, but each are changes that could be helpful to municipalities beyond the merger context in order to further democratize the voting process.

The two most effective initiatives would be to introduce compulsory voting and to have political parties operate at the local level. Each would dramatically increase voter turnout – perhaps as much as double its current level – and they would create a more informed electorate that is better able to choose between alternative candidates.

If changes made in the name of cost savings and increased efficiency have created barriers to participation or even if natural systemic flaws are beginning to emerge, then it is incumbent on society to watch for these inadequacies and to take all necessary steps to correct problems. The ultimate goal for improving democracy should always be equality in participation and representation.

  1. In actual fact, municipalities are not merged, only their services: Municipalities exist in law as bodies corporate and politic, so what the provincial legislature does is dissolve several municipal corporations and create a new corporate entity.

  2. Including a court challenge and appeal that coincided with the 1997 Toronto municipal election.

  3. Municipal elections in Ontario are held once every three years.

  4. Increasing amounts of electoral and statistical information have been made available to the public via the web site of the City of Toronto. Little information on councilors, and no information on candidates, is published in this manner by the city.

  5. For example, membership on the former Metropolitan Toronto Council was taken from the minutes of Metro Council, and the names of councillors in North York was taken from an old City of North York phonebook made available through the Toronto Archives.

  6. Pauline McAdam, Administrative Assistant to the Director of Elections, and Linda Sellers, Election Coordinator, of the City of Toronto, and Mark Cuddy of the Toronto Archives, are deserving of special mention because of their generous help with numerous and time-consuming demands for information.

  7. Source: Urban Development Services, City Planning Policy and Research, City of Toronto (September 2003)

  8. Only one small change occurred between 2000 and 2003. A section of ward 28 (between Rosedale Valley Road, the Don River and Bloor St. E.) was transferred to ward 27.

  9. At the time of writing, revisions at the poll were not yet available for the most recent election, so voter turnout for the 2003 election is inflated by approximately 1.5 percent (i.e., the average number of revisions made on election day). Statistics were provided by the Director of Elections for the City of Toronto.

  10. This table redistributes turnout as though merger had not taken place, assigning city and borough populations and turnout statistics to the previous municipal boundaries.

  11. There are no formal party affiliations in municipal politics in Ontario and no party designations appear on the ballot, though frequently candidates will be identified in the media or will self-identify with a federal or provincial political party. 

  12. In 1994 Toronto had 442,980 registered voters and North York had 376,902.

  13. The results are projections since the old municipal boundaries no longer exit. Results from various wards were redistributed for these elections based on advice from Election Services of the City of Toronto.

  14. The 2003 figures do not include revisions at the poll and are likely to be approximately 1.5 percent inflated (supra note 9). 

  15. The record turnout for the previous city of East York was 36%, Etobicoke 35.9%, North York 31%, Scarborough 48%, Toronto 45% and York 38.5% (The Toronto Star Nov. 15, 1997: C2). 

  16. The one exception is age, which has increasingly become the focus of research (perhaps to the detrimental exclusion of other factors) as political scientists in Canada struggle to explain why voter turnout has been declining at the federal level over the last decade.

  17. For example, one of the poorest areas of Montreal, St. Henri, is combined with largely Anglophone and wealthy Westmount in a single federal electoral district (St-Henri–Westmount); and propertied Rockcliffe Park is combined with the lower income and largely Francophone communities of Vanier and Lowertown to make the federal riding of Ottawa–Vanier. This almost conscious attempt to reduce cleavages by combining communities is much different than the U.S. approach where racial gerrymandering has created a number of ridings composed of similarly situated ethnic and SES communities. 

  18. Collinearity tests for this model reported a variance-inflation factor for each variable between 1 and 1.5 and tolerances between .91 and .67.

  19. While the N for each of these elections is relatively low at 44, combining the two elections (since both use 2000 census data) would violate the principle that cases should not be related. Nevertheless, the data was combined to see what results would be generated, and results reflective of both elections were found. 

  20. The electors list is produced by the Municipal Property Assessment Corporation (MPAC) that was established by the Municipal Property Assessment Corporation Act and came into existence on December 31, 1998. Prior to that, responsibility for property assessment rested with the cities and boroughs, which similarly used the tax rolls to generate electoral lists. 

  21. Some cities in Canada (e.g. Montreal) recognize municipal parties. These municipal parties facilitate the election of council/mayoral slates, however national/provincial party labels of ëLiberalí, ëConservativeí, ëNDPí, etc., which would permit heuristics (i.e. information shortcuts that permit left-right identification), do not exist at the municipal-level in Canada.

  22. East York had 4 ward councillors, one metro councillor and a mayor; Etobicoke had 12 ward councillors, two metro councillors and a mayor; North York had 14 ward councillors, 7 metro councillors and a mayor; Scarborough had 14 ward councillors, 6 metro councillors and a mayor; Toronto had 25 ward councillors and a mayor; and York had 8 ward councillors, two metro councillors and a mayor. 

  23. A disproportionate number of ëfrivolousí candidates tend to run for the mayoral position, due to its high profile. This may in part explain why, in 1997, 20 people competed for the mayoral post and yet Lastman and Hall received almost all of the votes (97.9% between them).

  24. For example, 11 of the candidates the Toronto Star (Nov. 9, 1997: F2) endorsed in the 1997 election were neither mayors nor councillors. 

  25. Each voter was entitled to choose two candidates from among the list of candidates on the ballot and the candidates which received the first and second largest number of votes were each elected. 

  26. While legally the new city of Toronto is a distinct corporation, viewing it as the success to metro may have some legitimacy. Information on the former Metropolitan Council was available directly from officials at the new megacity of Toronto, whereas information on the city and borough governments for the most part needed to be obtained from the archives.

  27. Candidates for council were entitled to raise $3,500 and for mayor $5,500, plus fifty cents for each elector entitled to vote (Municipal Elections Act, 1996) meaning that councillors who represented larger areas prior to merger would likely have raised larger amounts of money. 

  28. The higher turnout for national elections under compulsory voting is on average 10 percent. However, at the state or local level the difference in turnout can be double or triple since these elections are, almost universally, more poorly attended than their federal counterparts.

  29. In some municipalities, such as Montreal, political party labels are used in municipal elections, but these are municipal parties created specifically for that purpose. The result is that they provide an information shortcut for the voter when choosing councillors and a mayor but do little beyond providing a slate of candidates. 

  30. Municipalities have increasingly turned to alternate sources to augment the list, such as the federal permanent voters registry, but this has been ad hoc and not entirely successful, and, since the genesis of the municipal lists continues to be the tax rolls, this gives property owners a disproportionate advantage.

  31. For example, in the 2006 municipal election campaign in Ontario, the London Free Press (August 26, 2006, p.A3) reported that more than 90,000 dead people were on the municipal voting lists, a situation that could not be remedied since privacy legislation limited the sharing of information between the provincial department that issues death certificates (Government Services) and the Municipal Property Assessment Corporation (MPAC). 

  32. Interestingly, the reason the London Free Press was able to determine that there were 90,000 deceased persons on the municipal electors list is because it was compared with the federal ëpermanent register of electorsí, but once again privacy restrictions prevented the sharing of the information between Elections Canada and the MPAC.

  33. Now renters simply complete a provincial tax-credit schedule to calculate their property tax, something taxpayers will not realize they have done if they use a tax-preparation company or software, assuming they complete the income tax form at all (which is not required of low income or non-income Canadians). 

  34. 34 Assuming a polling station is open from 10:00 a.m. to 8:00 p.m., with the requirement that an employee must have three consecutive hours to vote, an employee who works from 8:00 a.m. to 4:00 p.m. will have four consecutive hours between 4:00 p.m. and 8:00 p.m. and the employer does not have to provide time off with pay.

  35. 35 Historically, French and English journalists and editors in both North America and in Europe have each had different conceptions of the role of the media. The Anglo media has held the ideal of impartiality up as indisputable since the advent of freedom of the press, while the French media often refers to itself as ëagents of social changeí which is reflective of its instrumental role in the French revolutions. As the French Canadian media sees its role as more proactive, this may explain higher voter turnout numbers in Quebec. [At the very least it points to yet another area of inquiry.]

  36. This is probably because of media interest in how ethnic communities mobilize during party nominations.

  37. The Chief Electoral Officer of Canada (federal) has been trying initiatives to level the playing field by such things as door-todoor enumeration in low-turnout communities and aggressive advertising campaigns among low-turnout cohorts such as young persons (with admittedly varying degrees of success).
  38. One of the first pre-writ tasks assigned to campaign managers by the various political partiesí headquarters and campaign manuals is to colour code electoral maps for their riding, identifying which areas of the riding have low-voter turnout in the past and identifying which areas voted for their opponents. The coding informs their decisions on where to spend resources and, assuming these maps show differences in turnout and voter choice along socio-economic divisions, will impact on the importance these communities will have to their candidates. So while class differences may not be part of political speak in Canada, it is likely part of political planning, making it even more important that it be identified and properly studied. 

  39. While it might be argued that the sensitivity to translate material into the languages of the various communities in a ward is a test of skills for a candidate, it is also reflective of resources available. When one couples a disparity of resources with the differential of turnout, it makes a compelling case for state intervention. After all, the voting exercise is not, theoretically, about comparing the skills of the candidates (something made clearer through free competition), rather it is about voter choice. A fully informed electoral is essential for choice. 

Bakvis, Herman (ed.). 1991. Voter Turnout in Canada (Research Studies of the Royal Commission on Electoral Reform and Party Financing, 15). Toronto: Dundurn Press.

Black, Jerome H. 2003. ìFrom Enumeration to the National Register of Electors: An Account and an Evaluationî, IRPP Choices.

Black, Jerome H. and Lynda Erickson. Forthcoming. ìWomen Candidates and voter bias: do women politicians need to be better?î, Electoral Studies.

Canadian Charter of Rights and Freedoms. Schedule B, Part I of the Constitution Act, 1982. ìAppendicesî, Revised Statutes of Canada 1985 (Ottawa: Queenís Printer for Canada).

City of Toronto Act, 1997. Statutes of Ontario 1997, v.1., c.2. Toronto: Queenís Printer for Ontario.

Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper.

Drummond, Robert. 1990. ìVoting Behaviour: Counting the Changeî, in Graham White, ed., The Government and Politics of Ontario. Scarborough: Nelson Canada.

Eagles, Munroe. 1991. ìVoting and Non-Voting Canadian Federal Electionsî in Herman Bakvis, ed., Voter Turnout in Canada. Toronto: Dundurn Press.

Erickson, Lynda. 1993. ìMaking Her Way In: Women, Parties and Candidacies in Canadaî in Joni Lovenduski and Pippa Norris [eds.], Gender Party Politics. London: Sage Publications.

Gidengil, Elisabeth and Richard Vengroff. 1997. ìRepresentational Gains of Canadian Women or Token Growth? The Case of Quebecís Municipal Politicsî, Canadian Journal of Political Science 30.

Globe and Mail. (Nov.) 1994, 1997, 2000 and 2003. Toronto: Thomson Publishing.

Hicks, Bruce M. 2002. ìThe votersí tax creditî, Policy Options 23: 4.

Higgins, Donald J.H., Local and Urban Politics in Canada. Toronto: Gage.

Karnig, Richard C. and B. Oliver Walter. 1983. ìDecline in Municipal Turnout: A Function of Changing Structureî, American Politics Quarterly 11.

Krashinsky, Michael and William J. Milne. 1986. ìThe Effect of Incumbency in the 1984 Federal and 1985 Ontario Electionsî, Canadian Journal of Political Science 19: 337-43.

Kushner, Joseph, David Siegel and Hannah Stanwick. 1997. ìOntario Municipal Elections: Voting Trends and Determinants of Electoral Success in a Canadian Provinceî, Canadian Journal of Political Science 30.

Madnusson, Warren and Andrew Sancton [eds.]. 1983. City Politics in Canada. Toronto: University of Toronto Press. 

Marland, Alex. 1998. ìThe Electoral Benefits and Limitations of Incumbencyî, Canadian Parliamentary Review 21: 4

Milbrath, Lester. 1977. Political Participation: How and Why People Get Involved in Politics. Chicago: Rand McNally.

Sancton, Andrew. 2000. Merger Mania: The Assault on Local Government. Montreal: McGill-Queenís University Press.

Stanwick, Hannah. 2000. ìA Megamayor for All People? Voting Behaviour and Electoral Success in the 1997 Toronto Municipal Electionî, Canadian Journal of Political Science 30.

Rust-DíEye, G.H. 1989. ìTiers are Not Enough: The Exercise of Upper-Tier Planning Responsibilities in the Greater Toronto Areaî [unpublished].

Tennant, Paul. 1980. ìVancouver Politics and the Civic Party Systemî, in Dickerson et al., eds., Problems of Change in Urban Government. Waterloo: Wilfred Laurier University Press.

Tindal, Richard and Susan Nobes Tindal. 1995. Local Government in Canada. Toronto: McGraw Hill Ryerson. 

Toronto Star. (Nov.) 1994, 1997, 2000 and 2003. Toronto: Toronto Star Publishing.

Verba, Sidney and Norman H. Nie. 1972. Participation in America: Political Democracy and Social Equality. New York: Harper & Row.