by Peter John Loewen, Clifton van der Linden, Yannick Dufresne and Gregory Eady

Every election is historic, but some are more historic than others. Last spring’s federal election perhaps deserves a capital H. The Conservatives, led by a man in turn brilliant and seemingly reckless, won a carefully crafted majority. It was their first majority government since 1988, and a return to unrivalled power that few would have anticipated when Stephen Harper returned from his Calgary sojourn to lead the Canadian Alliance nine years earlier.

The success of the New Democrats was no less impressive, and certainly more dramatic given the events since the election. The party was led by a man we now understand was fatally ill, but at the time appeared indefatigable, and almost above politics. On May 2, Jack Layton’s NDP finally secured both widespread success in Quebec and status as the Official Opposition.

If two parties gained in this election then two had to lose, perhaps dramatically, to zero the sum. The Liberals were the first. A party that can be counted among the four most successful political parties in the history of modern democracy was reduced to less than three dozen seats, shut out of entire regions. Its leader had left the highest echelons of the academy on the promise of the Prime Minister’s Office. Instead, Michael Ignatieff now finds himself back in the academy,1 with his party near broke and uncertain of its future. The Liberals’ one consolation is that they remain an officially recognized party in the House of Commons, unlike the Bloc Québécois, which managed to elect only four MPs.

As historic and remarkable as this election may have been, it was also seemingly Seinfeldian – an election about nothing. It was, to judge by appearances, not an election contested over the familiar terrain of Quebec’s role in the federation, or Canada’s role in the world. Rather than a debate over the proper role of government in the economy, voters were treated to attacks on personality. In the place of a discussion of Canada’s role in Afghanistan, we heard about absentee candidates and skipped debates. It was, despite the best efforts of many, not even an election about the Prime Minister’s treatment of Parliament.

How could such a historic result occur in an election about nothing? How could issues matter so little? The answer is that it couldn’t and didn’t. In reality, as we shall show, rather than an election only about personalities, this was an election that largely turned on issues – in two ways. First, issue preferences mattered for vote choice. How Canadians felt about a handful of issues – among others Afghanistan, the long gun registry, Quebec’s place in the constitution and the role of the private sector in health care – had measurable effects on their vote choice. How much these issues mattered varied markedly between constituencies, suggesting the election was not fought nationally in any principal sense. Instead, it was a collection of local races.

Second, the importance of issues and the views individuals held on these issues varied systematically between Quebec and the rest of Canada. The local importance of issues and the way this varied inside and outside Quebec speaks clearly to the strategic prowess of both Stephen Harper and Jack Layton and to the wrong-footed campaigns of Michael Ignatieff and Gilles Duceppe.

Charting the ground wars

Political scientist Ken Carty, channelling Stephen Leacock, has observed that Canadian politics are national battles acted out in local terms. Campaigns, as it turns out, are not only or even principally about leaders and personality. If we look deep enough, we can see that they are about issues. While there is undoubtedly a highly visible “air war” fought between the political parties, there are also 308 less visible “ground wars” fought among local campaign organizations. Local candidates don’t only knock on doors and hammer in campaign signs. They participate in local debates. They mail voters flyers stating their views and priorities. Their activities and positions are reported in their local papers. What local candidates say and what is said about them is the talk of coffee shops and conversations between friends and coworkers. Because each province is different, and because constituencies can vary so much, different issues matter in different ridings.

Keen observers can see that issues matter to many voters, and often in idiosyncratic ways. Travel to Buckhorn, Ontario, for example, and in nearly every pub and restaurant you’ll find a yellow ribbon box, and above it a poster of Nicholas Bulger, a young soldier killed in Afghanistan. Open the local paper and you will usually find one or more letters from soldiers in Afghanistan who received packages from Buckhorn’s residents as a memorial gesture for Bulger. Our political parties are divided over Canada’s role in Afghanistan. How could we expect voters in towns like Buckhorn not to notice such a difference?

If you are prone to eavesdropping, leaning in on any coffee shop conversation in North Bay, Ontario, or in Lower Sackville, Nova Scotia, in October 2010 you would have found patrons talking about the long gun registry. If you were an MP in one of these districts, you would have spent that month responding to hundreds of emails and calls from constituents on both sides of the issue. In such places, the registry of long guns is very much a political issue, one that comes up when politics is discussed.

Which issues mattered where?

What keen observers know, pollsters and journalists might have a difficult time finding. A poll of 1,000 Canadians will ask only a handful of Canadians in each constituency which issue is important to them. Likewise, reporters cannot cover anywhere near all local campaigns, and even the most voracious of readers cannot read every local paper. The focus, as a result, is not on the issues that matter to the voters, but on questions of leadership and the personalities of would-be prime ministers.

To truly appreciate the local importance of different issues, we need a large amount of individual-level data on a large number of issues in a large number of constituencies. In 2011, for the first time, such data became available, via the Vote Compass voter engagement application.2 In the federal election, it took the form of a website, hosted by the CBC and Radio-Canada, on which individuals could indicate their positions on 30 issues divided evenly between social and economic concerns. After indicating their positions, they were plotted on a two-dimensional plane and shown their position vis-à-vis the principal federal parties.3
During the federal election campaign, nearly two million people completed the Vote Compass application. More than 700,000 of these Canadians also told us where they lived and for whom they intended to vote. This is not a poll that randomly samples participants: Canadians self-select into Vote Compass. This complicates making generalized statements, but using different statistical methods we can still see which issues mattered in each constituency.

Our goal in this article is to empirically demonstrate two things. First, Conservative and New Democratic success was largely, though not nearly completely, attributable to the positions they took on a number of key issues. To show this, we estimate statistical models of how individuals’ views on 30 issues, as captured in Vote Compass, mattered for Conservative, Liberal and New Democratic vote choice. In doing so, we find that four issues had important positive effects for the Conservatives and New Democrats: Canada’s role in Afghanistan, the long gun registry, private sector health care and Quebec’s place in the constitution.

Our second goal is to demonstrate that while the importance of issues varied at the local level, there was another obvious level of division – between Quebec and the rest of  Canada. We show this division in two ways: first, how individual voters’  feelings on the issues differed between Quebec and the rest of Canada; and then, which issues actually mattered for vote choice.

Please refer to figures 1– 4 at end

Consider first the results for Afghanistan. Views on Canada’s role in Afghanistan mattered for Conservative vote choice in fully one third of constituencies (133 to be exact). Vote choice for the NDP also turned on respondents’ views on Canada’s role in Afghanistan, in nearly 90 constituencies. For the Liberal Party, Afghanistan mattered for vote choice in just one constituency. For all of Michael Ignatieff’s manoeuvring on the issue, and for all of his international (and internationalist) credentials, he could not win voters on the issue of Afghanistan. This was devastating for a party whose leader’s main attribute was his experience in the world.

Liberal fortunes did turn on the long gun registry, an issue over which Ignatieff took great pains to whip his caucus. But this was as much to his loss as it was his gain. Few issues divide urban and rural Canada as starkly as the question of whether potential gun owners should be subjected to background checks and the disclosure of personal information (which are not required, it should be noted, when one attempts to acquire and license a car or a boat). Few issues also divide Quebec and the rest of Canada as clearly. Ignatieff was able to make the long gun registry matter for vote choice where he should not have (outside Quebec, and especially in rural Canada) and not where he should have (Quebec).

For Layton, by contrast, this issue did not matter one stitch, however much commentators may have suggested he would pay a price for pressuring his rural members to cast a vote for the registry. The registry was, however, a clear winner for Harper, as it was positively related to casting a Conservative vote in 178 constituencies. Given that the Conservatives’ majority was built on their showing in Ontario and Alberta, we should add that views on the registry mattered for Conservative votes in three quarters of Alberta constituencies and in two thirds of Ontario constituencies. Is it any wonder that Harper has been consistently willing to expend political capital on scrapping this legislation?

The next issue we consider is the role of the private sector in health care. If we look at regional distributions we note the surprising fact that the province with the most consistent support for more private sector involvement in the provision of health care is Quebec. How then does the party of Tommy Douglas perform so well in Quebec? The answer is that Layton was able to make the role of the private sector in health care central to NDP vote choice outside Quebec, while playing no role for NDP vote choice in Quebec. For his part, Harper was able to make his party’s more welcoming position on the role of the private sector in health care play outside Quebec, while having no role in Conservative vote choice in Quebec. And what of Ignatieff? He could make the Liberals’ intermediate position on private sector involvement in health care matter in just four constituencies. Canadians chose one of two ideologically informed positions rather than a position that cut the middle between them.

Finally, consider the importance of the NDP’s position on the constitutional recognition of Quebec as a nation. Much has been made of Layton’s Quebec roots and his congeniality. He was, in common parlance, un bon Jack. But he also took political positions in the campaign that suggested his party was more sympathetic toward Quebec aspirations and national identity than the other federal leaders. In particular, he emphasized a soft line on the Clarity Act, a willingness to extend Bill 101 to federally incorporated companies and a desire to enshrine Quebec distinctiveness in the constitution. Did these positions matter for his success in Quebec? Our data suggest they did, but not overwhelmingly. In 11 of 75 constituencies in Quebec the NDP’s position on Quebec’s place in the constitution mattered for NDP vote share. But as importantly, it mattered negligibly (in only one constituency) outside Quebec. It would thus appear that Layton was able to reach out to voters in Quebec without paying a cost among voters in ROC. It’s equally important to note that the Liberals’ vote in Quebec did not turn on their strong federalist position, even on Montreal Island. What has been a winning issue for the Liberals in the past, and should have been again this time, counted not at all.

Not just a horse race

When we turn on our TVs at night, we can be forgiven for thinking that election campaigns are “horse races” fought and won by personalities and personal attacks. But there is another dimension to election campaigns that takes place when voters chat about issues and parties’ positions with friends and coworkers over coffee, when they are reminded that a fellow citizen died in Afghanistan and go on to reflect on the parties’ positions on Canada’s involvement in that country. Parties gain advantage or lose ground when citizens visit a hospital and then imagine how their care would change (for better or for worse) under a private provider. Our politics are acted out in the media, but they are realized on the ground.

This was a historic election. One man, literally on his last legs, carried his party to historic heights. Another man, exceeding perhaps all expectations but his own, returned his party to past glory. Still another saw a past glory break apart, perhaps irreparably.

Just as the results of the election may well produce a permanently changed political system, we can imagine a permanently changed country.

The Conservative Party has built a majority with a mere smattering of support in Quebec, relying instead on majorities in Ontario and the west. The NDP, by contrast, crafted its caucus by appealing to Quebecers on issues uniquely important to them, avoiding those issues on which they were badly positioned, and by dominating the Liberals on issues that were formerly the domain of the “Natural Governing Party.” But the result of this is a caucus of inexperienced MPs from Quebec who may disagree fundamentally with their colleagues on questions of sovereignty and Quebec’s place in Canada. For the moment this issue lies dormant, leaving room for possible growth. But how long that period will last is unclear.

What is clear to us is that success in this election was underwritten not by a single issue, or even a suite of issues regarded as equally important across the election. Instead, success was the product of parties’ capacity to tailor their positions to the varying preferences of voters in different regions. The winning parties could make some issues matter in some places, and others matter in others. Whether this is cynical manipulation or an effective return to the old “brokerage” politics, we leave to the reader to judge. What we insist on is that the campaign was fought and won not just on the personalities of leaders, but also on issues that mattered, and mattered differently among the districts and regions.

Figures 1 – 4

Figures 1 to 4 tell an important story about the difference in the importance of local issues across Canada.

They suggest that the Conservatives and the NDP were able to make their positions on key issues matter in different locales, while the Liberals were largely ineffectual.

KEY to reading the data

Each point in the figures represents a constituency, and constituencies are arrayed from east to west, left to right. The vertical axis measures the certainty of the statistical relationship between respondents’ views on an issue and their vote choice for the party in question.

The higher the dot, the more certain we can be that this issue mattered for vote choice in each constituency. We vary the appearance of the dots to indicate whether an issue did not matter at all (black squares), mattered using conventional levels of statistical significance (grey squares) or mattered after correcting for “multiple testing” (dark grey circles). We consider important for vote choice only those estimates that exceed this final, very conservative estimate.

The main reason for using this approach is to test against “false positives”: we would risk considering chance findings as significant if we did not use these conservative thresholds.

Notes

1 Albeit at a rather nice place, if we do say so ourselves.

2 Such applications have been used in election in recent decades in more than a dozen European countries.

3 Those interested can still try the tool in two different versions at www.votecompass.ca

4 While we focus on four issues here, there were 30 in the federal Vote Compass. Complete graphics for each of these issues, all of which display a Quebec-ROC cleavage to some degree, can be seen at www.votecompass.ca

Clifton van der Linden, Yannick Dufresne and Gregory Eady are PhD candidates in the Department of Political Science at the University of Toronto. All four authors are members of the core team at Vote Compass.