Reforming election dates in Canada: towards an explanatory framework.

Alcantara, Christopher ; Roy, Jason

Since 2001, ten governments in Canada (nine Canadian provincial governments (1) and the federal government) have adopted fixed election date legislation, setting all future general elections to occur every four years. (2) The first province to pass such legislation was British Columbia in 2001, followed by Newfoundland and Labrador in 2004, Ontario in 2005, New Brunswick in 2007, Saskatchewan, Prince Edward Island, and Manitoba in 2008, Alberta in 2011, and Quebec in 2013. The federal government passed fixed election date legislation in 2007. The only provincial government to not pass fixed election date legislation is Nova Scotia.

The emergence of fixed election date legislation is an important development not only for Canada, but also for students of Canadian public administration. Indeed, this legislation is one of the first visible reforms successfully passed by the federal government and a majority of provincial governments to address the so-called "democratic deficit," which some commentators have argued is one of the most serious problems facing this country (Aucoin, Jarvis, and Turnbull 2011; Tanguay 2009; White 2005). As a result, a body of literature has emerged that examines the impact of this legislation on our Westminster parliamentary system. One set of commentators has argued that this change has undermined our parliamentary system, while others have suggested that it has strengthened it (see Aucoin, Jarvis, and Turnbull 2011; Pepall 2010). What has yet to be explored, however, are the dynamics that generated this change.

So what explains these trends, and more specifically, the timing of the legislation? Surprisingly, there has been little research that has focused on these questions directly (although see Cross 2005) and even less effort towards advancing a theoretical explanation for why a Westminster parliamentary democracy might pass this type of legislation. The typical assumption in the literature, usually offered in passing, is that these governments are simply responding to public and scholarly concerns about the undemocratic nature of the election timing power (Dodek 2010; Leuprecht and McHugh 2008; Milner 2005). While we do not disagree with this reasoning for why fixed election date legislation was enacted, we do note that there is a void in the existing scholarship regarding the timing of fixed election date legislation across Canadian jurisdictions. What explains the twelve-year gap between British Columbia's 2001 transition to fixed election dates and Quebec's 2013 enactment of a similar policy? If fixed election date legislation is a policy response intended to address democratic malaise, why is there a temporal gap in the passage of this legislation? Furthermore, why hasn't every province adopted such a law?

In this article, we approach this puzzle deductively by relying on the theoretical insights of multiple streams theory, originally developed by John Kingdon (2003) and used by scholars to study policy change in a variety of countries (Blankenau 2001; Zahariadis 2007), including Canada (Henstra 2010). Using secondary literature, news sources, and Hansard transcriptions, we apply this theory and logic to the experiences of the federal and provincial governments of Canada. Our findings suggest that fixed election date legislation is the result of three particular "streams" converging at different moments in time and jurisdictions. More specifically, we conclude that it is the political stream that is crucial for explaining the timing of legislation across Canadian jurisdictions. The implications of these results offer further insight into the opportunities and challenges Canadian policymakers face in their effort to address public malaise within a Westminster institutional setting.

The problem and potential explanations

The problem this article addresses is two-fold. First, we consider why the Canadian federal government and nine provincial governments adopted fixed election date legislation. Our findings, which are consistent with the existing scholarship, suggest that the adoption of fixed election date legislation is a government response to the public's concerns regarding democratic performance (Dodek 2010; Leuprecht and McHugh 2008; Milner 2005). The willingness of a government to enact such a policy, however, gets to the core of this article's contribution. What explains the differences in the timing of the legislation being passed across jurisdictions?

One possible explanation is that fixed election date legislation and other democratic reforms like citizens' assemblies and juries are just convenient and harmless fads that have emerged at roughly the same time (Adams and Hess 2001). Others might argue that the reigning neoliberal economic paradigm in Canada has had a powerful influence on all recent government approaches to policy, focusing party and government promises mostly on consultative mechanisms, which are relatively cheap and open-ended, rather than substantive policy mechanisms, which can be expensive (Larner 2000; Phillips 2006). Although both are plausible, neither explains the timing of Canadian jurisdictions that have passed fixed election date legislation. In the case of the "convenience" argument, it fails to specify the factors that produced a twelve-year gap between the policy response in BC in 2001 and the one in Quebec in 2013. In the case of the neoliberal argument, one would have to assume that the sequencing of jurisdictions that adopted fixed election date legislation would reflect the precise timing in which neoliberalism took hold in each jurisdiction. However, at the federal level, neoliberalism has been the dominant paradigm since at least the early days of Jean Chretien's government in 1993. Yet it was a conservative government, under Stephen Harper, that passed fixed election date legislation in 2007. Another problem with the neoliberal argument is that it does not specify what causal mechanisms might connect the neoliberal ideas to the passage of the legislation in the ten Canadian jurisdictions over a twelve-year period.

Perhaps a more parsimonious explanation is that fixed election date legislation was simply the result of a significant change in the incentive structures facing Canadian political parties. A variety of jurisdictions in Canada have altered their campaign finance laws over the last ten years (Young and Jansen 2011), which in turn may have created powerful incentives for parties to want to regularize election timing to facilitate campaign planning and the efficient expenditure of campaign funds. While plausible, our analysis of the literature, media reports and Hansard found no evidence that policymakers were driven by these factors. Indeed, if this explanation was true, there should be a direct correlation between changes in electoral financing laws and the adoption of fixed election date legislation across jurisdictions. Such a correlation is not evident.

We contend that the most likely explanation for the timing of fixed election date legislation can be drawn from Kingdon's (2003) multiple streams model, which seeks to explain how certain issues end up on the agenda of government officials and how particular policy solutions come to receive serious consideration. Although his model was originally intended to explain agenda-setting, Kingdon did acknowledge that his theory could be used to explain policy choices, and scholars have done so in a variety of contexts and policy areas (Kingdon 2003; Henstra 2010; Mucciaroni 1992; Zahariadis 1992).

At the core of his theory is a conceptualization of the policy environment as consisting of three streams: the problem stream, the policy stream, and the political stream. Each stream tends to operate independently, but there are moments when they come together and it is at these critical junctures that there is the greatest potential for policy change (Kingdon 2003: 19). Policy windows, therefore, are "opportunities for action" when developments in one or more of the streams converge and make it more likely for policy change to occur. Policy entrepreneurs take advantage of these windows to pursue policy change (Kingdon 2003: 166, 179). As outlined below, this understanding of the policy process offers considerable leverage at addressing how ten different governments in Canada over a twelve-year period passed fixed election date legislation. We build this argument in the following section by briefly describing each of the streams (problem, policy, and political) and presenting evidence from our cases to show how each stream contributed to the change. As a point of departure, Table 1 provides a summary of information pertaining to the federal and provincial jurisdictions that have passed fixed election date legislation. We refer to the information presented in the table throughout the discussion that follows.

Fixed election date legislation: Towards an explanation The problem stream

Kingdon's problem stream focuses on the types of policy problems that typically capture the attention of policymakers. Problems find their way on to the agenda of governmental officials in a number of ways. First, government and non-government studies on specific issues or indicators such as consumer prices, cost of government programs, and disease rates, for example, can be used to show that something is amiss or deserves further attention (Kingdon 2003: 91). Second, focusing events, such as an airplane crash, might prompt action because such events are relatively rare and occur suddenly. Relatedly, personal experiences and symbols may serve to focus attention on an already existing problem. Medical research on a particular disease, for instance, is likely to resonate with law-makers who have had family members, or are themselves, afflicted by a similar condition (Kingdon 2003: 95-96). Finally, policy feedback on programs already in operation can spur policy action. Policy feedback can include data gathered through regular monitoring or through more informal channels such as public complaints to politicians. Bureaucrats who actually implement and oversee programs are also able to provide important data to policymakers. Feedback spurs policy action if it suggests that a policy has not been implemented according to the original legislative intent, the delivered results fall short of the targets set out at the beginning, costs are expanding too fast, or if there are negative or positive unintended consequences (Kingdon 2003: 101, 103).

Recognizing a problem may not always be enough to place it on the agenda, given the infinite number of issue-areas that exist. Successful agenda setting and policy change will often depend on how the problem is defined and linked to various policy solutions. Therefore, policy entrepreneurs usually devote considerable resources to problem definition and the coupling of problems with developments in the political and policy streams (Kingdon 2003: 115).

In regard to fixed election date legislation, it seems likely that this issue rose to the top of the policy agenda following numerous studies that found Canadians were becoming increasingly dissatisfied with their political institutions and representatives (Aucoin and Turnbull 2003; Blais and Gidengil 1991; Blais et al. 2002; Cross 2004; 2010; Gidengil et al. 2004; Gidengil et al. 2010; Milner 2005). Based on these trends, commentators and academics have concluded that Canada suffers from a significant and persistent democratic deficit. In essence, they argue that the locus of power in Canada has shifted away from citizens towards elites (Savoie 2010), and "that the traditional mechanisms of representative democracy--political parties, elections, and territorially based legislatures--are simply not up to the task of articulating or defending the interests of the vast majority of citizens in the current age" (Tanguay 2009: 223-224). As a result, many citizens have become disengaged from Canadian political life. Beginning with the 1988 federal election, for instance, voter turnout in Canada has declined from 75% to 59% in 2008 (Tanguay 2009: 223). While turnout increased marginally in the 2011 federal election, reaching 61%, participation remained low compared with participation levels from the 1960s and 1970s. In addition to electoral participation, survey data show that fewer Canadians are getting involved in election campaigns and political parties and that in general, they have also become less optimistic about the ability of Canadian political parties to successfully effect meaningful change (Cross 2010; Gidengil et al. 2010). As a result, "many Canadians are simply disengaged: they do not vote in elections, and they do not sign petitions, take part in demonstrations, or boycott products, either" (Gidengil et al. 2010: 96-97).

While no single explanation has been agreed upon to explain Canadian democratic malaise, many scholars point to dissatisfaction towards political leadership, specifically the prime ministers of Canada, past and present, and the provincial premiers. For example, Donald Savoie (2009), Herman Bakvis (2000), and others (Aucoin, Jarvis, and Turnball 2011; Bernier, Brownsey, and Howlett 2005) have convincingly shown that over time, power in the Canadian system has become concentrated in the centre, and specifically in the hands of the federal and provincial first ministers and their supporting officials, organizations and departments. As a result, and especially in majority government situations, Canadian first ministers have the ability to bend the political system to serve their interests and the interests of their parties (Savoie 2009: 125-126).

The discretion that first ministers have to control the timing of elections has received substantial scholarly attention. A number of scholars have documented the many instances when prime ministers and provincial premiers have used this power to call elections during periods that were most conducive to their party's re-election chances (Aucoin, Jarvis, and Turnball 2011: 130-132; Pepall 2010: 11; Roy and Alcantara 2012). Predictably, critics and opposition parties have objected "strenuously to the ability of prime ministers to call an election simply when it is politically advantageous" and have regularly called on governments to pass legislation to curtail this behaviour (Aucoin, Jarvis, and Turnball 2011: 131). For example, Ross Wiseman, an MLA in the Newfoundland and Labrador legislative assembly, emphasizes the potential abuse of this power for political gain:

If you look at the history, and most recent history, of elections in Newfoundland and Labrador, we saw, in a period of 1993, 1996, 1999, rapid successions and rapid calls of elections. Mr. Speaker, in each of those election calls there was a significant amount of staging, a significant amount of manipulation that took place in advance of those elections. One could say that it was a manipulation of the electorate (Wiseman 2004).

Similar sentiments were expressed in Ontario when Liz Sandals addressed fixed election date legislation in 2004. According to Sandals:

The significance of this [fixed election date legislation] is that politicians, the Premier, will no longer be able to play games with the election date. As we've seen over a number of years, Premiers do in fact play games with the public and with the political process in trying to choose an election date to their best advantage. We're fixing this election date so that the election will be at the convenience of the public, not at the convenience of the Premier, whoever that Premier might happen to be (Sandals 2004: 3215).

These quotes show that policy makers were aware of the public's growing democratic malaise and of the effect that an unchecked election timing power could have on deepening this malaise. Indeed, it may be that federal and provincial politicians and policymakers turned to fixed election date legislation not because the public demanded it per se, but because they saw it as a useful way to address an amorphous public unhappiness with democracy in the public realm (Dodek 2010; Milner 2005). With this general democratic malaise rising to the top of the policy agenda, the conditions were ripe for coupling this stream with an appropriate solution, given the political will to do so.

The policy stream

The policy stream determines which policy solutions are chosen to address the problems raised in the problem stream. Policy ideas are generated by a group of specialists in a specific policy area and can include "researchers, congressional staffers, people in planning and evaluation offices and in budget offices, academics, interest group analysts" (Kingdon 2003: 116). The people who make up these communities generally know each other's ideas and proposals fairly well and are generally united by their interest in the issues, although the degree of fragmentation varies across policy areas. Numerous ideas and proposals are churned out by this community but the actual subset that eventually receives serious attention is much smaller.

Kingdon emphasizes that the content of ideas is central to which policy solutions gain favour. Well-reasoned and detailed arguments can help in the adoption of particular policy options but do not guarantee success (Kingdon 2003: 127). More important is that the idea is "technically feasible, meaning it is likely to achieve what it is intended to accomplish. It must also be compatible with the dominant values of the policy community" (Henstra 2010: 244). Technical feasibility refers to whether policymakers think that a proposal can actually be implemented and whether they think it will deliver on its goals. Value acceptability refers to how closely the proposal fits in with the belief system of the participants, such as in terms of their liberal-conservative orientation, their attitudes towards big government versus small government, or their notions of equity and efficiency. If a proposal has too high a cost, if the public may be hostile to it, or if politicians are likely to reject it, then it will probably not receive serious attention, even within the policy community (Kingdon 2003: 138-139).

In the context of the problem and policy streams, fixed election date legislation fits these criteria very well, because it directly and effectively addresses the undemocratic use of the election timing power while at the same time respecting the values and goals of the affected policy elites. For some policy elites, such as those from marginalized political parties, fixed election date legislation is an effective and accomplishable step in the right direction towards addressing the democratic deficit (Green Party of Canada 2008; Green Party of Manitoba 2004; Green Party of Saskatchewan 2007). Similarly, the legislation is compatible with the interests of elites from the dominant political parties, who generally want to protect as much of the status quo in Parliament as possible, but also realize that the democratic deficit and the election timing power are problematic. Indeed, it is important to re-emphasize that fixed election date legislation in Canada allows policymakers to address the undemocratic nature of the election timing power on a public relations level while at the same time maintaining the limited discretion of first ministers as mandated by constitutional conventions. This finding is confirmed by the secondary literature and Hansard transcripts that show that policymakers believed that such legislation would be effective and congruent with their goals and values (Antle 2004: 5; Aucoin, Jarvis, and Turnball 2011: 131; Byrne 2004; Grimes 2004; Harris 2004; Sandals 2004: 3215, 3216; Sugimoto 2005; White 2005: 176-177; Wiseman 2004). Gilles Bisson (2005: 1607), NDP MP from Timmins-James Bay, for instance, has said during the debates in Ontario that "I don't think it's [fixed election dates] a bad idea. We recognize this is the British parliamentary system. Governments can still fall by the way of a non-confidence motion." John O'Toole (2004: 3221), PC MPP for Durham, initially opposed the legislation, arguing that it was "destroying the traditions of the Canadian parliamentary system." Yet, at the end of the debates, all of the parties supported the legislation, with 67 MPPs from all three parties voting in favour of the legislation and 6 voting against it. These results suggest that Ontario politicians from all three political parties saw this legislation as something that would not only address perceptions of a democratic deficit, but also maintain some of the fundamental principles inherent in Westminster parliamentary systems.

Indeed, a variety of research has shown that the push for electoral system reform in Canada has generally been unsuccessful because many of these elites oppose electoral system reform. They oppose this type of reform because their organizations and members benefit substantially from the single-member plurality system, while also making it difficult for new political parties to enter the system and form the government or official opposition (Courtney 2005: 154; Massicotte 2005: 90-91; Mendelsohn and Parkin 2005: 141; Pilon 2006). Fixed election date legislation, on the other hand, does not threaten the existing electoral system, and so the dominant political parties continue to reap the benefits conferred on them by the single-member plurality system. (4)

An examination of the legislation confirms that fixed election date legislation was the logical instrument to adopt in the context of the problem and political streams described above and below. All of the fixed election date legislation that has been passed addresses two goals. First, they set the general election date to occur on a specific day and month every four years, thus removing the election timing power from the hands of the first minister (British Columbia 2001: ch. 36, s. 1; Canada 2007: ch. 19, s. 2; Manitoba 2008: ch. 43, s. 6; New Brunswick 2007: ch. 57, s. 1; Newfoundland and Labrador 2004: ch. 44, s. 1; Ontario 2005: ch. 35, s. 1; Prince Edward Island 2008: ch. 9, s. 2; Quebec 2013: ch. 13; Saskatchewan 2008: ch. 6, s. 4). (5) Second, the legislation allows elections to occur outside of these dates, but only if the government loses the confidence of the House or if the Crown exercises its royal prerogative of dissolution; the latter scenario happened at the federal level in 2008 and in Quebec in 2014 (Alberta 2011: ch. 19, s. 2; Canada 2007: ch. 10, s. 1; Manitoba 2008: ch. 43, s. 6; New Brunswick 2007: ch. 57, s. 1; Newfoundland and Labrador 2004: ch. 44, s. 1; Ontario 2005: ch. 35, s. 1; Prince Edward Island 2008: ch. 9, s. 2; Quebec 2013: ch. 13; Saskatcehwan 2008: ch. 6, s. 4). (6) In essence, the principles of responsible government are maintained because the House of Commons can continue to hold the government responsible throughout the four-year term. In effect, the legislation addresses an important aspect of the democratic deficit problem without fundamentally altering the electoral system or the principles of responsible government. This point is important because electoral reform in general is very much affected by the willingness of the dominant political parties to support and implement it (Courtney 2005: 154; Massicotte 2005: 90-91; Mendelsohn and Parkin 2005: 141).

Political stream

The third, and in our view most relevant stream for understanding the timing of fixed election date legislation, refers to activities or events that "are electoral, partisan, or pressure group factors" (Kingdon 2003: 145); all of which exert pressure on policymakers to engage in policy change. Change can be brought about by "shifts of important participants (e.g., a change of administration or the influx of new legislators), or as a response to shifts in national mood or interest group configurations" (Kingdon 2003: 146). The national mood refers to public opinion, which government officials and policymakers take seriously when considering policy change. The turnover of key administrative, congressional, or bureaucratic actors can also lead to policy change by bringing in new issues, reconsidering dormant ones, or keeping some off the table. Kingdon (2003: 154) points out how under the Reagan administration issues like cuts in domestic spending and school prayer were prioritized while the subject of comprehensive national health insurance languished mainly as a result of turnover.

In terms of the passage of fixed election date legislation in Canada, the key elements in this stream are: i) the presence of an incumbent political party that has won consecutive mandates as the government (which in turn contributes to the notion that democratic alternation is not occurring regularly enough) (Massicotte 2005: 69); ii) an opposition party that adopts fixed election dates in its party platform prior to defeating the incumbent government in an election; and iii) once in power, the winning opposition party, now the government, introduces and passes the legislation in their respective jurisdictions. Although all three streams are important for explaining the passage of fixed election date legislation, it is these factors in the political stream that are crucial for explaining the timing of legislation across jurisdictions. Columns three, four and five of Table 1 summarize these factors as they relate to the various jurisdictions.

In British Columbia, for instance, the New Democratic Party formed the government for ten years, winning majority governments in 1991 and 1996 before being defeated by Gordon Campbell's Liberals in May 2001. The BC Liberals' platform for that election promised to "establish a fixed provincial election date under the BC Constitution Act, to ensure provincial elections must be held on a fixed date every fourth year, or immediately if any government loses a confidence vote in the Legislature" (BC Liberals 2001: 4). In Newfoundland and Labrador, the provincial Liberal party ruled for 14 years, forming majority governments in 1989, 1993, 1996, and 1999, until its defeat by Danny Williams and the Progressive Conservative Party in October 2003. Much like Gordon Campbell's Liberals, the Newfoundland and Labrador PCs also promised to pass fixed election date legislation and quickly did so in December 2004 (Canadian Press Writers 2004: A5). In Ontario, the Progressive Conservative party governed for eight years, winning majority governments in 1995 and 1999 before suffering defeat in October 2003 at the hands of the Ontario Liberal Party under the leadership of Dalton McGuinty. The 2003 Ontario Liberal platform promised to introduce "fixed dates for elections" (Ontario Liberals 2003: 4). At the federal level, the Liberal Party of Canada was able to win three majority governments and one minority government, dating back to 1993, before losing to the Conservative Party of Canada in January 2006. During the election campaign, Conservative leader Stephen Harper promised to introduce fixed election dates to "make sure that government is responsive and accountable to Canadians" and he did so in 2007 (Conservative Party of Canada 2004: 6; see also Conservative Party of Canada 2006: 44). In New Brunswick, the Progressive Conservative Party ruled for seven years, winning majority governments in 1999 and 2003, before losing to the Liberals in September 2006. During the election, Liberal leader Shawn Graham promised to "provide fixed dates for Provincial Elections [sic]" (New Brunswick Liberals 2006: 32). In Saskatchewan, the New Democratic Party won consecutive elections beginning in 1991, totalling sixteen years in power, before being defeated by the Saskatchewan Party in September 2007. As in the other provinces, the Saskatchewan Party promised to establish fixed election dates should they be elected to govern the province (Saskatchewan Party 2007: 39). Similarly, in Prince Edward Island, the Progressive Conservative Party governed for eleven years, capturing majority governments in 1996, 2000, and 2003, before suffering defeat at the hands of Robert Ghiz and the Liberal Party of PEI. During the election, the Liberals promised to pass fixed election date legislation after it had conducted full public consultations on the idea (Thibodeau 2007: A4). Finally, the Parti Quebecois' (PQ) 2012 victory in Quebec created the conditions for legislative change. Fixed election date legislation was part of the PQ's election platform leading up to its autumn 2012 electoral victory. Following three consecutive Liberal governments, the PQ came to power and, with this change in the political stream, the three streams were set to converge; in June of 2013, Quebec became the ninth Canadian province to adopt fixed election date legislation.

In sum, in all of these jurisdictions, the main opposition party promised to legislate fixed election dates prior to their electoral victory over the incumbent and multi-term governing party. Shortly after taking office, the winning opposition parties, now the government, then passed fixed election date legislation in accordance with their election promises. Thus, we argue that the timing of the passage of legislation across jurisdictions is best explained by factors in the political stream.

The two exceptions to the change in multi-term governments outlined above are Manitoba and Alberta. In both provinces, fixed election date legislation was passed under an incumbent government. However, these outliers may be explained, in part, by changes in key political actors. For example, in Alberta, the legislation was introduced following a change in party leadership. During the 2011 Alberta Progressive Conservative leadership race, Allison Redford included fixed election date legislation as one of her platform promises. Following her party leadership victory, she followed through on this promise, passing fixed election date legislation on 8 December 2011. In many ways, these events fit with our other cases, where the political stream converges with the problem and policy streams to produce fixed election date legislation. However, the political stream in Alberta does not flow from a change in governing party, rather a change in party leadership. In all other regards, Alberta fits the pattern established above. (7)

Manitoba is arguably the only exception to the multiple streams theory and the evidence we have marshaled. The Manitoba legislation was passed under the stewardship of Gary Doer, the longstanding leader of the Manitoba NDP. Doer introduced this legislation in 2008, nearly ten years after first coming to power in Manitoba. While the problem and the policy streams in this province resemble the other Canadian cases, the political stream deviates from the patterns we have observed elsewhere. The most likely explanation for this outlier reflects a policy actor, namely a backbench member from the opposition party, who introduced a motion to adopt fixed election date legislation that died on the order paper after Doer called the 2007 provincial election (CBC 2012). Following the NDP's return to a majority government, Doer reintroduced fixed election date legislation. Similar to arguments made in other jurisdictions, Doer presented this legislation as a means of improving democratic accountability, suggesting that the political stream, whether by design or desire, had converged with the problem and policy streams in Manitoba.

Discussion

Overall, our findings suggest that fixed election date legislation is likely to occur when three streams converge. In the problem stream, legislation is facilitated by a recognition among policymakers that the public is concerned about a democratic deficit and in particular, the potential misuse of the election timing power to unfairly advantage the incumbent party. The secondary literature, news reports, and Hansard transcripts indicate that policymakers were aware of these problems and that they needed to respond to them in some way. In the policy stream, fixed election date policies were chosen because policymakers thought that they would be effective for addressing concerns about the democratic deficit and the election timing power. They also chose these options because the ideas behind these policies were compatible with their goals and values. Finally, in the political stream, fixed election date legislation was facilitated by opposition parties adopting this policy idea and then implementing it once they defeated the longstanding incumbent governing party in a general election. This last stream, in particular, is essential for understanding why different jurisdictions in Canada passed fixed election laws at different moments in time.

Our findings also serve as an important reminder of the challenges governments face enacting policy. While political problems and potential solutions may be readily identifiable, without the political will to couple these streams, policy change is unlikely. In fact, the argument put forth above may help us to understand why similar policy change has yet to occur in the province of Nova Scotia, the only Canadian province without fixed election dates. While the problem and solution are arguably as prominent in this jurisdiction, sufficient political will has yet to emerge that would advance fixed election date legislation. This political will, however, may have increased as a result of the 2013 provincial election, which saw the Liberal party defeat the one-term NDP government. Prior to its electoral victory, the Liberal Party adopted fixed election date legislation as part of its election platform, thus fulfilling an important condition for legislative change (Liberal Party of Nova Scotia Election Platform, 2009: 31). We do not, however, anticipate fixed election date legislation in this jurisdiction. The one significant difference in Nova Scotia compared to the rest of Canada is that the Liberals did not come to power following a multi-term government and so although change is possible, it is still unlikely.

Conclusion

This article set out to address the timing of fixed election date legislation in Canada. We began by presenting a number of possible theories for explaining these trends. Based on our analysis of the evidence, we found that Kingdon's multiple streams theory provided the strongest analytical leverage for answering our research questions. Specifically, we found that the timing in which fixed election date legislation was passed in the Canadian provinces as well as the federal government reflects a convergence of three streams: the problem stream, where democratic malaise and dissatisfaction with first ministers' ability to time elections for political advantage had become a salient issue; the policy stream, where adopting fixed election date legislation provided a viable solution to address the problem; and the political stream, a change in government (or leadership in the Alberta case) that produced the political will to join the problem stream with the policy stream. It is this third stream that we have argued is most important in understanding the timing of fixed election date legislation in Canada.

While we believe this work makes an important contribution to our understanding of the conditions under which fixed election date legislation is adopted, there are a number of other questions surrounding this legislation that warrant further study. For example, the problem stream reflects the belief that citizens were unsatisfied with the practice of first ministers calling elections when the timing was optimal. A policy solution was put forth to correct this problem. Did it succeed? Are citizens in jurisdictions with fixed election date legislation less cynical of their political actors and institutions following this change? As we noted above, and as others have pointed out (see Desserud 2005), this legislation provides the governing party with an "escape clause" as per the constitution guiding Canada's system of responsible government. Put simply, it is still possible for elections to occur outside of the fixed election date. Indeed, this was the case when the minority Conservative federal government called the 2008 and the 2011 Canadian federal elections and the Quebec government called the 2014 provincial election. Given this option, the potential for citizens to remain skeptical is a very real possibility. It also reinforces the importance of political agency in Canada, and specifically the ability of first ministers in Westminster parliamentary systems to circumvent checks on their power passed in Parliament. We believe exploring the success or failure of this policy is an important area of future research that warrants further attention (see Dodek 2010 for some preliminary arguments about how the legislation has failed to achieve its intended goal of addressing political dissatisfaction).

Another avenue for future study would be to test our model by comparing two or three jurisdictions using a more inductive, historical method. It may be that individual political contexts, in the tradition of historical institutionalism, may temper the effects of the different streams and perhaps introduce new factors unanticipated by the multiple streams framework. This small-n comparative approach could be used to study jurisdictions within Canada, or across countries.

Finally, researchers might apply our model to jurisdictions outside of Canada. Do our findings hold beyond the Canadian border? While we expect our framework will travel well, applying this theory to other countries would advance the conclusions we draw here. Drawing upon evidence from other countries that have had this type of legislation for a longer tenure would also provide insight into future obstacles that may transpire in the Canadian case. For example, as five provinces (Saskatchewan, Manitoba, Ontario, Prince Edward Island, and Newfoundland and Labrador) and the federal government are set to go to the polls in the fall of 2015 under current fixed election date legislation, concern of voter fatigue has been raised. It may be that the solution aimed to re-engage a disillusioned citizenry turns out to do more harm than good. Indeed, as the Canadian experience with fixed election date legislation is still in its infancy, the full implications of this legislation are largely unknown.

References

Adams, David, and Michael Hess. 2001. "Community in public policy: Fad or foundation?" Australian Journal of Public Administration 60 (2): 13-23.

Alberta. 2011. Election Amendment Act. Statutes of Alberta, chapter 19.

Antle, Rob. 2004. "Fixed election dates legislation among 34 bill passed before House of Assembly closed." The Western Star 17 December, p. 5.

Aucoin, Peter, Mark D. Jarvis, and Lori Turnbull. 2011. Democratizing the Constitution: Reforming Responsible Government. Toronto: EMP.

Aucoin, Peter, and Lori Turnbull. 2003. "The democratic deficit: Paul Martin and parliamentary reform." Canadian Public Administration 46 (4): 427-49.

Bakvis, Herman. 2000. "Prime Minister and Cabinet in Canada: An autocracy in need of reform?" Journal of Canadian Studies 35 (4): 60-79.

BC Liberals. 2001. A New Era for British Columbia: A vision for Hope & Prosperity for the next decade and beyond. BC Liberal Platform.

Bernier, Luc, Keith Brownsey, and Michael Howlett, eds. 2005. Executive Styles in Canada: Cabinet Structures and Leadership Practices in Canadian Government. Toronto: University of Toronto Press.

Bisson, Gilles. 2005 (12 December). "Election Statue Law Amendment Act, 2005." Ontario, Legislative Assembly, Hansard. First Session, 38th Parliament.

Blais, Andre, and Elisabeth Gidengil. 1991. Making Representative Democracy Work: The Views of Canadians. Ottawa: Royal Commission on Electoral Reform and Party Financing and Dundurn Press.

Blais, Andre, Elisabeth Gidengil, Richard Nadeau, and Neil Nevitte. 2002. Anatomy of a Liberal Victory: Making Sense of the Vote in the 2000 Canadian Election. Peterborough, Ont.: Broadview Press.

Blankenau, Joe. 2001. "The fate of national health insurance in Canada and the United States: A multiple streams xplanation." Policy Studies Journal 29 (1): 38-55.

British Columbia. 1996. Constitution Act. Revised Statutes of British Columbia, chapter 66.

--. 2001. Constitution (Fixed Election Dates) Amendment Act, 2001. Statutes of British Columbia, chapter 36.

Byrne, Ed. 2004 (7 December). "Orders of the day." Newfoundland and Labrador, Legislative Assembly, Hansard. Vol. XLV No. 53. Available at: http://www.assembly.nl.ca/business/ hansard/ga45sessionl/04-12-07.htm

Canada. 2007. An Act to amend the Canada Elections Act. Statutes of Canada, chapter 10.

Canadian Press Writers. 2004. "Nfld voters soon to head to polls." Cape Breton Post 6 December, p. A5.

CBC. 2012. Fixed Election Dates. Available at: http://www.cbc.ca/news/interactives/mapfixed-election-dates/. Accessed 19 September 2012.

Conservative Party of Canada. 2004. Demanding Better: Conservative Party of Canada, Platform 2004.

--. 2006. Stand Up for Canada: Conservative Party of Canada Federal Election Platform 2006.

Courtney, John. 2005. "Is talk of electoral reform just whistling in the wind?" In Strengthening Canadian Democracy, edited by Paul Howe, Richard Johnston, and Andre Blais. IRPP: 149-58.

Cross, William. 2004. Political Parties, Canadian Democratic Audit. Vancouver: UBC Press.

--. 2005. "The rush to electoral reform in the Canadian provinces: Why now?" Representation 41 (2): 75-84.

Cross, William. Ed. 2010. Auditing Canadian Democracy. Vancouver: UBC Press.

Desserud, Don. 2005. "Fixed-date elections: Improvement or new poblems?" Electoral Insight 7 (1): 48-53.

Dodek, Adam. 2010. "The past, present, and future of fixed election dates in Canada." Journal of Parliamentary and Political Law 4: 215-38.

Gidengil, Elisabeth, Andre Blais, Neil Nevitte, and Richard Nadeau. 2004. Citizens, Canadian Democratic Audit. Vancouver: UBC Press.

Gidengil, Elizabeth, Richard Nadeau, Neil Nevitte, and Andre Blais. 2010. "Citizens." In Auditing Canadian Democracy, edited by William Cross. Vancouver: UBC Press, pp. 93-117.

Green Party of Canada, 2008. Un-/non-/ir-responsible government? Available at: http:// www.greenparty.ca/node/6713

Green Party of Manitoba. 2004. 2004 Policies. Available at: http://greenparty.mb.ca/policies/ policies.

Green Party of Saskatchewan, 2007. Green Party Platform. Available at: http:// www.greenpartysask.ca/policy/electoral_reform.

Grimes, Roger. 2004 (7 December). "Orders of the day." Newfoundland and Labrador, Legislative Assembly, Hansard. Vol. XLV No. 53. Available at: http://www.assemblynl.ca/ business/hansard/ga45session1/04-12-07.htm

Harris, Jack. 2004 (7 December). "Orders of the day." Newfoundland and Labrador, Legislative Assembly, Hansard. Vol. XLV No. 53. Available at: http://www.assembly.nl.ca/ business/hansard/ga45session1/04-12-07.htm

Henstra, Daniel. 2010. "Explaining local policy choices: A multiple streams analysis of municipal emergency management." Canadian Public Administration 53 (2): 241-258.

Kingdon, John. 2003. Agendas, Alternatives, and Public Policies. Toronto: Longman.

Lamer, Wendy. 2000. "Neo-liberalism: Policy, ideology, governmentality." Studies in Political Economy 63: 5-25.

Leuprecht, Christian, and James T. McHugh. 2008. "Fixed election cycles: A genuine alternative to responsible and responsive government?" Commonwealth and Comparative Politics 46 (4): 415-441.

Liberal Party of Nova Scotia. 2009. Transforming Nova Scotia.

Manitoba. 2008. The Lobbyists Registration Act and Amendments to The Elections Act, The Elections Finances Act, The Legislative Assembly Act and The Legislative Assembly Management Commission Act. Statutes of Manitoba, chapter 43.

Massicotte, Louis. 2005. "Changing the Canadian electoral system." In Strengthening Canadian Democracy, edited by Paul Howe, Richard Johnston, and Andre Blais. IRPP: 65-98.

Mendelsohn, Matthew, and Andre Parkin. 2005. Getting from Here to There: A Process for Electoral Reform in Canada." In Strengthening Canadian Democracy, edited by Paul Howe, Richard Johnston, and Andre Blais. IRPP: 137-48.

Milner, Henry. 2005. "Fixing Canada's unfixed election dates: a political season to reduce the democratic deficit." IRPP Policy Matters 6 (6): 1-44.

Mucciaroni, Gary. 1992. "The garbage can model and the study of policy making: A critique." Polity 24 (3): 459-82.

New Brunswick. 2007. An Act to Amend the Legislative Assembly Act. Statutes of New Brunswick, chapter 57.

New Brunswick Liberals. 2006. Charter for Change: Shawn Graham's plan for a better New Brunswick. September.

Newfoundland and Labrador. 2004. An Act to amend the House of Assembly Act and the Elections Act, 1991. Statutes of Newfoundland and Labrador, chapter 44.

Ontario. 2005. An Act to amend the Election Act, the Election Finances Act and the Legislative Assembly Act, to repeal the Representation Act, 1996 and to enact the Representation Act, 2005. Statutes of Ontario, chapter 35.

Ontario Liberals. 2003. Government that works for you: The Ontario Liberal plan for a more democratic Ontario. The Liberal Party of Ontario, www.leonarddomino.com/news/ platform-ontarioliberal2003.pdf.

O'Toole, Peter. 2004. Ontario, Legislative Assembly. Hansard. First Session, 38th Parliament. 23 June.

Pepall, John. 2010. Against Reform. Toronto: University of Toronto Press.

Phillips, Susan. 2006. "The intersection of governance and citizenship in Canada: Not quite the third way." IRPP Policy Matters 7 (4): 1-31.

Pilon, Dennis. 2006. "Explaining voting system reform in Canada, 1874 to 1960." Journal of Canadian Studies 40 (3): 135-61.

Prince Edward Island. 2008. An Act to Amend the Election Act. Statutes of Prince Edward Island, chapter 9.

Quebec. 2013. Loi modifiant la Loi electorale afin de prevoir des elections a date fixe. Statues of Quebec, chapter 13.

Roy, Jason, and Christopher Alcantara. 2012. "The election timing advantage: Empirical fact or fiction?" Electoral Studies 31 (4): 774-81,

Sandals, Liz. 2004 (23 June). Ontario, Legislative Assembly, Hansard. First Session, 38th Parliament.

Saskatchewan. 2008. The Legislative Assembly and Executive Council (Fixed Election Dates) Amendment Act, 2008. Statutes of Saskatchewan, chapter 6.

Saskatchewan Party. 2007. Securing the Future: New Ideas for Saskatchewan.

Savoie, Donald. 2009. "Power at the apex: Executive dominance." In Canadian Politics, edited by James Bickerton and Alain-G. Gagnon. Toronto: UTP Higher Education, pp. 115-31.

--. 2010. Power: Where is it? Montreal-Kingston: McGill-Queen's University Press.

Sugimoto, Dawn. 2005. "Fixed dates worth a look." The Lethbridge Herald Wednesday 30 November, p. a8.

Tanguay, Brian. 2009. "Reforming representative democracy: Taming the 'democratic deficit.'" In Canadian Politics, edited by James Bickerton and Alain-G. Gagnon, Toronto: UTP Higher Education, pp. 221-18.

Thibodeau, Wayne. 2007. "Ghiz unveils measures to make government more accountable." The Guardian 24 May, p. A4.

White, Graham. 2005. Cabinet and First Ministers. Vancouver: UBC Press.

Wiseman, Ross. 2004 (7 Dec.). "Orders of the day." Newfoundland and Labrador, Legislative Assembly, Hansard. Vol. XLV No. 53.

Young, Lisa, and Harold J. Jansen. 2011. Money, Politics, and Democracy: Canada's Party Finance Reforms. Vancouver: UBC Press.

Zahariadis, Nikolaos. 1992. "To sell or not to sell? Telecommunications policy in Britain and France." Journal of Public Policy 12 (4): 355-376.

--. 2007. "The multiple streams framework: Structure, limitations, prospects." In Theories of the Policy Process, edited by Paul A. Sabatier. Boulder, Colorado: Westview Press.

Notes

(1) Our manuscript was accepted prior to the ninth provincial government (Quebec) adopting fixed election date legislation (June 2013). In our original manuscript, we predicted that Quebec would soon pass this type of legislation because all of the factors that we had identified for the other jurisdictions were in place. We were happy to see our prediction come true and have updated our paper for publication accordingly.

(2) It is noteworthy that these laws also permit elections to occur outside of these dates, but only if the government loses the confidence of the House or if the Crown exercises its royal prerogative of dissolution.

(3) The tenure of the previous government is estimated by subtracting the year the party came to power from the year they were defeated.

(4) It should also be noted that fixed election date legislation was not the only possible solution to the problem stream here. Besides selecting from the broad range of possible democratic reforms, policymakers could have also chosen to transfer the election timing power to some sort of independent third party, or enacted a set of stringent rules that the Lieutenant-Governor or Governor-General had to follow (see Aucoin, Jarvis, and Turnbull 2011).

(5) The one exception is Alberta, whose legislative amendment specifies that a general election must occur sometime between March 1 and May 31 in the fourth calendar year following the last general election (Alberta 2011: ch. 19, s. 2).

(6) British Columbia's legislation leaves unaltered s. 23(1) of the province's Constitution Act of British Columbia, which reads: "The Lieutenant Governor may, by proclamation in Her Majesty's name, prorogue or dissolve the Legislative Assembly when the Lieutenant Governor sees fit" (British Columbia 1996: ch. 66, section 23(1)).

(7) We should note that long-term Alberta PC leader, Ralph Klein, was initially replaced by Ed Stelmach (2006-11). While this change in leadership provided the conditions for the political stream to fit our model, Stelmach opposed fixed date election legislation and as such, failed to achieve the political will necessary to pass this legislation.

Christopher Alcantara is associate professor of political science, Wilfrid Laurier University, Waterloo, Ontario. Jason Roy is associate professor of political science, Wilfrid Laurier University. The authors would like to thank the editor and anonymous reviewers of this Journal for their helpful comments and suggestions on an earlier draft of this paper.

Table 1. Canadian Fixed Election Date Legislation Summary (3)

 Date
 Legislation
Jurisdiction Passed Government Date Elected

British Columbia 2001/08/27 Liberals May 2001
Newfoundland 2004/12/16 PCs October 2003
 and Labrador
Ontario 2005/12/15 Liberals October 2003
Federal 2007/05/03 Conservatives January 2006
 Government
New Brunswick 2007/06/26 Liberals September 2006
Saskatchewan 2008/04/28 Saskatchewan September 2007
 Party
Prince Edward 2008/05/22 Liberals May 2007
 Island
Manitoba 2008/10/09 NDP September 1988
Alberta 2011/12/08 PC August 1971
Quebec 2013/06/14 PQ September 2012

 Previous
 Government
Jurisdiction and Tenure

British Columbia NDP--10 Years
 (2 terms)
Newfoundland Liberals--14 Years
 and Labrador (4 terms)
Ontario PC--8 Years
 (2 terms)
Federal Liberals--13 years
 Government (4 terms)
New Brunswick PC--7 years
 (2 terms)
Saskatchewan NDP--16 Years
 (4 terms)
Prince Edward PC--11 Years
 Island (3 terms)
Manitoba NDP--Government
 since 1988
Alberta PC--Government
 since 1971
Quebec Liberals--9 Years
 (3 terms)