Site MenuAboriginal Electoral Participation in Canada
1. Literature Review
Our literature review poses and then offers answers to two questions. First, what are the factors that explain variations in electoral participation? In other words, what systematically causes turnout to increase, and what causes it to decrease? Second, what do we know about the sources of electoral participation among Aboriginal citizens? To undertake this review, we first present a simple, stylized model of voter turnout. This necessarily involves an abstraction of the complex decision of whether to vote. However, it does allow us to identify the "moving parts" of the decision to vote and to consider the effects of each individually. This review will highlight the factors thought to be important for the decision to vote. Afterward, we ask what existing work on Aboriginal participation tells us about the determinants of turnout in the Aboriginal context.
1.1 The Model
A simple model of the decision to vote is presented by Riker and Ordeshook (1968). In their abstraction, just four factors influence the decision to vote. The first is C, or the cost of voting. All else held equal, voters are less likely to vote as the cost of doing so increases. The second factor is B, the marginal benefits that an individual will receive if their preferred party wins the election. The third factor is P, the probability of casting a decisive vote. The final factor, D, is an individual's sense of civic duty or obligation.
This decision model takes the simple form of:
PB+D>C,
meaning that if the utility an individual derives from fulfilling their duty to vote plus the expected benefits of the election (that is, benefits discounted by the probability of being decisive) is greater than the cost of voting, then an individual will vote. If these benefits do not exceed the cost of voting, then an individual will not vote.
We note two things about this model before proceeding. First, the probability of casting the deciding vote in an election is very small, so the marginal benefits to be realized by one's preferred party winning government are steeply discounted (i.e. PB probably equals something close to 0). It becomes quickly apparent that duty is doing the most work on the left-hand side of the equation. Second, this is obviously a very crude simplification of the decision to participate. Nonetheless, it serves the purpose of organizing a discussion of the decision to vote, to which we now turn.
Benefits
Elections convey three types of benefits: first, benefits to the voter; second, benefits to other citizens or groups; and third, benefits to society as a whole.
To illustrate benefits to a voter, imagine that an elector who works in the public service is asked to vote in an election in which one party, Party A, is promising to maintain public service jobs while the other, Party B, is promising to eliminate the very type of job this elector holds. For this elector, the benefits of the election can be thought of as the difference in their material well-being if Party A wins over Party B. Such a conception can be extended with felicity to any number of other programmatic differences between parties, be they differences in tax policy, disagreements over levels of government spending, conflict over the distribution of government spending, etc. The important point is that, for many electors, one party is better for them materially than another party. Accordingly, the larger the difference in potential benefits if an elector's preferred party wins, the more likely an individual is to cast a ballot.
Existing evidence suggests that this is very much the case, with electors demonstrating a greater likelihood of voting in elections for more consequential offices (such as national elections) versus less consequential offices (such as local aldermanic races). The other implication is that electors who have more at stake in government policy may be more likely to vote. For example, those who pay more in taxes and are more integrated into the economy will be more likely to vote (Wolfinger and Rosenstone 1980). This has been extended as one explanation of why youth are less likely to vote than older citizens (Blais and Loewen 2009; Goerres 2007; Wattenberg 2007; Rubenson et al. 2004; Blais et al. 2004; Wass 2007; Blais et al. 2007).
Benefits to others work much the same way as benefits to oneself. The distinction is that some voters will cast ballots because they have "social preferences," and thus care about the outcome of an election not only for themselves but for others as well (Loewen 2010; Dawes, Loewen and Fowler, forthcoming; Fowler and Kam 2007; Uhlaner 1989a, 1989b). For example, individuals might be motivated to vote for a party advocating for tuition decreases because they have grandchildren in university. In a less focused manner, some might vote for Party A over Party B because they feel that this party is better for the "average person." Likewise, others might be motivated to vote because doing so confers a benefit on their preferred party (though evidence of this, in the Canadian context, is scant; see Loewen and Blais 2006, 2011). An obvious implication of these findings, for both Aboriginal turnout and the turnout of new Canadians, is that electors with social preferences are only more likely to vote if they feel integrated into the political community and feel that voting confers benefits on those for whom they are concerned (see White et al. 2006 for a review of the turnout behaviour of immigrants).
Regardless of whether individual electors are preoccupied by benefits to themselves or benefits to others, the implications are the same: turnout will increase the greater the differences between parties, and turnout will also increase the more electors feel that at least one party offers policies that will improve their own lot or the lots of others.
The third type of benefits are those to society as a whole. Such benefits do not refer exclusively or even principally to particular policy benefits. Instead, they refer to the general benefits accrued to citizens living in a democracy. To the extent that individuals believe democracy confers benefits on all citizens, they will be more likely to vote (see Blais 2000 for a review).
As a final addition, current research demonstrates that the time horizon of benefits also appears to matter, such that benefits expected in the more distant future are less likely to motivate electors than those which are immediate. Similarly, at an individual level, it has been shown that those who are more impulsive are less likely to participate in elections, since they more steeply discount any benefits derived from voting (Fowler and Kam 2006).2
Probability of being decisive
How likely is an individual to be decisive in an election? And how does the probability of being decisive affect the likelihood of an individual participating in an election? Empirical and theoretical work suggests that the probability of casting a decisive vote is approximately equal to 1/N, where N is the number of electors in a voter's district (Gelman, Katz and Bafumi 2004). In other words, the probability is exceedingly small. Put another way, it is very unlikely that a single person's vote will determine an election outcome.
Despite this, individuals vote. This suggests, perhaps, that the probability of being decisive does not matter in an absolute sense, but instead exercises an effect at the margin. The literature supports this view. First, as elections become more closely contested, we witness a marginal increase in the rate of participation (Blais 2006). However unlikely, individuals do appear to respond positively to an increased probability of being decisive.
Second, there is emerging evidence that the competitiveness of elections, especially early in an elector's political experience, has a long-term effect on the probability that an individual votes. Both Franklin (2004) and Johnston, Matthews and Bittner (2007) provide evidence that those whose first election is a closely fought affair are more likely to develop a habit of voting. Loewen and Rubenson (2011) also provide experimental evidence suggesting that democratic competition increases voter turnout. The logic here seems to be that in more closely fought elections voters are more likely to believe they can make a difference, and are thus more likely to vote. If voting is habit forming (Plutzer 2002), then early experiences of voting can inculcate a tendency toward regular participation. Contrarily, while Blais and Rubenson (forthcoming) do not dispute the relationship between competition and turnout, they find little supporting evidence that a change in the nature of elections is driving turnout decline.
Third, survey research has demonstrated a negative correlation between turnout and the belief that one's vote does not make a difference. While this may, for the most part, reflect a more general belief that party policies do not differ and thus there are no differences in expected benefits, it is also consonant with the view that electors are less likely to vote when they feel they are not likely to be decisive (Blais 2000).
Fourth, and finally, research in psychology has shown that: a) individuals are poor at calculating probabilities; and b) individuals routinely overestimate their own ability to be decisive in an event (Langer 1975).
While P may rarely matter in the strictest terms, electors do respond to changes in the probability of being decisive in a manner congruent with the model of voter turnout. The lesson to be drawn from this, in the most general sense, is that electoral systems which feature frequent close races are more likely to exhibit higher turnout than those with few close races and many disproportionate or lopsided results.
Duty
As Blais (2000) notes, the notion of voting as a duty encompasses several different but related ideas. For example, D may represent the psychological benefit one incurs from voting. It may also be a well-developed sense that voting is a moral obligation, a promise that one keeps to oneself. It may also be a sense of obligation to others, inculcated early in life via parents and school (Campbell 2006). Such a sense would rely on the belief that others were counting on your vote. Indeed, when electors are reminded of this, their probability of participation increases measurably (Rosenstone and Hansen 2003; Putnam 2000; Green, Gerber and Nickerson 2003).
Sense of duty may also be conceived as a personality trait, encapsulating not only a feeling of obligation to vote but a more general feeling of social obligation, including actions such as serving on a jury, giving money to charity, and paying taxes to help others (Loewen and Dawes, forthcoming; Fowler, Baker and Dawes 2008).
Finally, one may feel a sense of duty to honour democracy and the sacrifices made by forbearers to maintain democracy. Such a feeling does not rely on whether the individual's vote will actually maintain democracy, as Blais points out (2000, 93).
What these conceptions share is, first, that duty is non-instrumental. It does not rely on the benefits of the election. Instead, it is a stable belief that voting is important. Second, these conceptions assume that voters see voting as a good and right action within one's community. Duty takes on a moral dimension, whereby one votes merely because it is the right thing to do. Likewise, voters may cast ballots because they do not wish to be seen in a bad light by their fellow citizens (Gerber, Green and Larimer 2008).
Costs
The final component of this turnout model is the cost involved in voting. As cost increases, the probability of voting decreases. We can identify two broad types of costs. The first are those involved in making a decision about whether and for whom to vote. The second are those involved in the actual act of voting.
Beginning with the costs of making a decision, it is important to recognize that the political world is confusing for many citizens. Differences between parties are not always readily apparent. Federalism only increases this confusion. For citizens who are not politically attentive or informed, the psychological costs of choosing between parties can be marked. Accordingly, those who have lower levels of interest and attention will be less likely to cast ballots (Gidengil et al. 2003, 2004a; Pammett and LeDuc 2003; Blais and Loewen 2009). More generally, citizens who lack the resources that aid in making a political decision – whether in time, education or social networks – will be less likely to vote (Verba, Schlozman and Brady 1995).
The second set of costs is imposed in a more straightforward fashion. The more effort required in going to the polls, the less likely individuals are to cast a ballot. If registration is at the initiative of the elector, turnout will be lower (Highton 1997). If information about polling station locations is not easily obtained, individuals will be less likely to vote. The farther polling stations are from individuals, the less likely they will be to cast a ballot (Brady and McNulty 2007). Indeed, even in situations of increased but minor inconvenience, such as when it rains, individuals will be less likely to vote (Gomez, Hansford and Krause 2007). The converse is that when voting is made less costly, individuals will be more likely to cast ballots. For example, online and mail-in ballots can be expected to increase turnout (Southwell and Burchett 2000; Canada-Europe Transatlantic Dialogue 2010). Likewise, offering advance voting opportunities can have a positive if modest effect on voter turnout (Blais, Dobrzynska and Loewen 2007).
Finally, we note that in addition to costs depressing turnout, costs can also increase turnout if they are imposed on non-voters, namely through compulsory voting coupled with fines for abstention. While making voting compulsory does not likely improve the quality of citizen engagement and knowledge (Loewen, Milner and Hicks 2008; Milner, Loewen and Hicks 2007), it certainly does increase voter turnout (Blais 2000; Blais, Dobrzynska and Massicotte 2003).
Taken together, these four elements – benefits, probability of being decisive, duty and costs – give us a good understanding of the decision to vote. Indeed, they have been the foundation of decades of research into voter turnout. We next examine the current understanding of the determinants of Aboriginal turnout.
1.2 Findings on Aboriginal Turnout
Over the last two decades, concern over Aboriginal turnout in Canadian elections has emerged and increased. It has resulted in a small but burgeoning literature. A common observation throughout this work is that voter turnout in federal elections tends to be lower among Aboriginals than non-Aboriginals. This conclusion has been uncovered both by studies that tabulate official election results for polling stations located entirely or principally within First Nations reserves (Bedford and Pobihushchy 1996; Bedford 2003; Guérin 2003; Dalton 2007), and by studies that draw upon reported turnout in public opinion surveys (Dalton 2007; Harell, Panagos and Matthews 2009; Howe and Bedford 2009). The pattern, however, is not consistent across space. In some provinces and regions, Aboriginal turnout is markedly lower than non-Aboriginal turnout, while the gap is much smaller in other locations (Bedford and Pobihushchy 1996; Bedford 2003; Guérin 2003; Harell, Panagos and Matthews 2009). Time-wise, there are indications that Aboriginal turnout has declined in federal elections since the franchise was extended to all First Nations people in 1960, though again some places buck this trend (Bedford and Pobihushchy 1996; Bedford 2003).
Various potential explanations have been proposed to account for the weaker participation of Aboriginal electors in federal elections in Canada. The first explanations put forward were decidedly not Aboriginal-specific; they stressed factors that explain turnout among the general population. For instance, Bedford and Pobihushchy (1996) suggested that the socio-demographic conditions of Aboriginals could be the key. In fact, the 1991 Royal Commission on Electoral Reform and Party Financing had pointed to the lower socio-economic and educational levels along with the geographic dispersion of Aboriginals as the source of their lower participation rate. The younger age profile of Aboriginals could also be critical, as the youth vote substantially less (Guérin 2003). Another "conventional" approach is to insist on socio-psychological dispositions generally correlated to turnout, notably sense of civic duty, political efficacy and political apathy (Bedford and Pobihushchy 1996).
The first uniquely Aboriginal explanation was offered 15 years ago by Bedford and Pobihushchy (1996), and is often labelled the "nationalist" explanation (see also Schouls 1996). According to these authors, a change in consciousness is taking place among Aboriginal people. As Aboriginals go through a process of decolonization from a historically oppressive society, they come to see themselves decreasingly as Canadians and increasingly as members of a separate nation. Consequently, many refuse to participate in an electoral process that is not their own, that is fundamentally alien or irrelevant. As Aboriginal self-identification climbs, Aboriginal organizations become the central locus of authority and legitimacy. In a similar vein, turnout in federal elections can be considered as incompatible with the desire for autonomy and self-governance (Cairns 2003; Guérin 2003).
A second contending explanation is the "alienation" thesis. This alienation refers to a sense of social and political exclusion. Aboriginals feel excluded from the democratic and representation process in Canada (Cairns 2003; Ladner 2003; Dalton 2007). The existing institutions are seen as defending the interests of non-Aboriginal people and as the instruments of Aboriginal oppression. A major obstacle to Aboriginal participation would thus be lack of trust in the Canadian political system.
The final explanation deals with the failing political efforts and opportunities dedicated to Aboriginal electoral mobilization (Ladner 2003; Silver, Keeper and MacKenzie 2005). The argument is that dominant political forces simply do not strive to foster participation among Aboriginal communities. In this light, turnout might be higher if political parties gave more attention to Aboriginal concerns in their platforms, nominated more Aboriginal candidates, campaigned more on Aboriginal issues, and did more to energize Aboriginal electors.
The number of studies assessing the empirical validity of competing explanations, however, is extremely limited (Ladner and McCrossan 2007). Research relying on official voting results from First Nations reserves can provide some information about the aggregate level of turnout, but it cannot offer insight into the factors that push some Aboriginals to cast a ballot and others not to. Only individual survey data (or experimental designs) could deliver in this regard.
Yet only two studies of this type have ever been conducted in Canada, both sponsored by Elections Canada. One used the Equality, Security and Community survey, which contains an oversample in 2004 of about 600 Aboriginals from Alberta, Manitoba and Saskatchewan (Harell, Panagos and Matthews 2009). The other employed the large General Social Survey, which reached approximately 700 Aboriginals from across the country (Howe and Bedford 2009).
Both studies uncovered that several socio-demographic factors were correlated to reported Aboriginal turnout: age, education, income and urban versus rural residence. Electoral participation tended to be higher among older, more educated and more prosperous individuals, and people in rural areas. However, the relationships did not always remain statistically significant once other control variables were taken into account.
The two studies also found that attachment to Canada, trust in the federal government or Parliament, and confidence in Canadian institutions were all associated with higher voter turnout. The impact of these predictors is compatible with interpretations based on the political exclusion and alienation thesis.
As well, each study identified significant variables that the other study could not consider (because they were not present in the data). Harell, Panagos and Matthews (2009) confirm that participation in organizations – particularly Aboriginal organizations – stimulates voter turnout. Howe and Bedford (2009) conclude that participation is less widespread when respondents are residentially mobile, single (especially single parents) and lack social trust.
Some variables were found to be inconsequential as determinants of turnout once all controls were inserted into the model: living on or off a reserve, Aboriginal identity (of the nationalist hypothesis), satisfaction with the federal government's responses to Aboriginal issues (Harell, Panagos and Matthews 2009) and health status (Howe and Bedford 2009). In the case of news consumption, the evidence was mixed across the two surveys.
While these two studies have considerably improved our understanding of Aboriginal electoral participation, much remains to be learned. One study had to rely on data, for its main dependent variable, that referred to behaviour in a federal election held several years before; the other's Aboriginal sample spanned only three Prairie provinces; and both were necessarily limited to the questions asked in their respective surveys. This paper intends to extend the analysis to further increase the understanding of Aboriginal participation.
2 This work is part of a growing body of literature on the relationship between personality or individual differences and voter turnout (Fowler et al., forthcoming; Mondak et al. 2010; Fowler and Dawes 2008). This work does not integrate easily into the proposed framework and cannot be easily tested in an Aboriginal context, given the data that follow. Accordingly, we merely highlight it here.