Over 40 years ago, Leon Epstein (1964) published a comparative study of Canadian political parties in the American Political Science Review that still merits a careful read. His aim in that essay was to juxtapose the Canadian experience against the American in order to learn something of the latter, but in doing so, he provided a succinct analysis of important dynamics structuring the Canadian party system. His account reminds us that, like the United States, Canada is “a diverse nation in a large land area,” has “developed a structural federalism,” has an “American-style social and economic class structure,” and uses “a single-member, simple-plurality election system.” Perhaps reflecting these underlying similarities, Canadian parties' (like their American counterpart's) “extra-parliamentary organizations are loose and non-doctrinal at every level” and structured federally.

That said, Epstein went on to identify two significant ways in which the Canadian party system was fundamentally different from that across the border. First, at least since the introduction of a universal franchise in the aftermath of World War I, Canada has always had more than two political parties, although the nature of its multi-party system has morphed over time. Second, the legislative parties are cohesive and disciplined. Both these features, he argued, were a consequence of the imperatives of a British-style parliamentary system (54). That argument, reminiscent of André Siegfried's 1906 observation that Canadian party organization and electoral campaign practice was a curious mixture of “old British forms and new American free-and-easy practices” (123), provides a starting point for any account of the contemporary system.

The recent general election of January 2006 saw the electorate split its support among several options: 5 parties got at least 4.5% of the vote—in practice that amounted to 3.8 effective parties, down from 4.3 a decade earlier in the 1997 election. The single-member plurality system had its usual effect of reducing the effective number of parties in the House of Commons—to 3.2, but that was up from 3.0 in 1997! The result was the second minority government in a row, and while minorities have not been uncommon in Canada (4 of the 11 elections since 1970 have produced one), there had not been successive minorities since the early 1960s. This was a reminder that the party system was still in considerable flux and the volatility that has marked it in recent years has not subsided. A second surprise of the 2006 election was that it was “won” by a party that did not exist a decade earlier.

In 1993 the existing national party system imploded (Carty et al. 2000). The social democratic New Democratic Party was reduced to about one-third of its previous strength while the then-governing Progressive Conservative Party broke into three distinct pieces—traditional Conservative supporters, populists in western Canada (organized as the Reform—later Alliance—Party) and French Canadian nationalists in the Bloc Quebecois dedicated to achieving political independence for Quebec, the country's second most populous province.

That highly regionalized fragmentation allowed the Liberal Party to win office with 41% of the vote, and the impact of the electoral system left the separatist Bloc as the country's official opposition during that Parliament. With their opponents divided, the Liberals were easily able to dominate national politics for the decade between 1993 and 2004. Right-of-center activists spent much of the time trying to repair the split, and after doing so, in early 2004 a partially restored Conservative Party managed to deprive the Liberals of a majority (see Figure 1). And it was that rejuvenated Conservative Party that took office in February 2006 with just 36% of the vote and 40% of the seats in the House of Commons. Given that only about 60% of those eligible to vote cast a ballot, the system delivered up a government that could claim to speak for little more than a fifth of the electorate.

Party Vote Shares: 1988–2006

These brief paragraphs hardly capture the shifting complexities of the Canadian party system, either among the electorate or in Parliament, but they do point to several of its important and recurring features. The first is what Epstein called the “vital division between French-speaking and English-speaking Canadians,” a division he thought “bulk[ed] larger than any division south of the border” (46). For most of the country's history, the major parties have sought to accommodate that division within themselves although, in practice, only the Liberals have managed to do so with any consistency since the First World War. The electoral earthquake of 1993 (which recorded the highest level of democratic electoral volatility anywhere in the 20^th century) broke that mould and left a large proportion of French Canadians supporting the Bloc. Given the regional concentration of its support, the electoral system over-represents the Bloc with about 50 of Parliament's total of just over 300 seats. This now makes it difficult for either of the two major parties to command enough support in both of the linguistic communities to fashion a national majority. In this way the Bloc now stalemates Canada's traditional practices of British-style majority parliamentary government.

The electoral earthquake of 1993 reshaped the party system, but not for the first time. Twice before in the century (in the early 1920s and again in the 1960s) the system had been broken and each time it had been rebuilt under old party labels that disguised just how much had changed. Each of those party system collapses had involved a major realignment of the electorate—largely in regional terms (see Johnston 2005 for the best brief account of these), but each had also been marked by major organizational and political communication transformations that redefined how Canadians did party politics (Carty 1988; 1995; 1997). They occurred when the diversity and shifting geography of the growing Canadian state stretched to the breaking point the capacity of the party system to respond to the latent demands of the political system. In effect, each was the consequence of the awkward fit between the pressures of a dynamic, open American-style society with the traditional, closed institutions of British-style parliamentary government.

Electoral politics in Canada is regional politics; it is also volatile politics. Blake's (1991, 255) measurements for the four decades from the Second World War indicate that electoral volatility was, on average, 1.5 times larger in Canada than in Britain and 2.6 times larger than in the United States. Table 1 illustrates the great swings in party seat totals over the past six elections.

Party Seats in Parliament: 1988–2006

The result of this volatility, aggravated and exaggerated by multi-party contests in many single-member districts, is high levels of turnover in the House of Commons. In some recent elections it has been greater than 50%, and over 80% of the electoral districts in the country were represented by members from at least two different parties during the decade from the mid-1980s to the mid-1990s. This party system produces a Parliament full of amateurs with limited capacity to effectively control the government as the constitutional principles of responsible parliamentary government require (Docherty 1997).

Any portrait of the Canadian party system that emphasizes its instability—either in terms of its restructuring every few decades, its regular multi-party volatility, or its parliamentary turnover—provides only half the story. The other reality is that for the last century the system has been dominated by the Liberal Party, whose record of easy victories (in office for about 70 years of the 20^th century) compares favorably with dominant parties like Ireland's Fianna Fáil, Japan's (post-war) Liberal Democrats, or the Swedish Social Democrats. Often described as the country's natural governing party, the Liberals have managed to survive both the periodic collapses of the party system and the ongoing electoral turbulence (the most recent example of its ability to ride out major party system disruptions can be seen in Figure 1).

But this long-standing Liberal party success itself obscures the reality that its base has been changing. Over recent decades the party has seen its overall level of support slowly shrink so that to win a parliamentary majority it now must depend on both the bonus provided by the electoral system and a fragmented opposition. At the same time, the party's regional base has been slowly narrowing, and in the last two decades its caucus has come to depend increasingly on support from the large central province of Ontario (Carty 2006). With the Conservatives disproportionately dependent upon the west, and the Bloc on Quebec, the party system is now more regionally fragmented than at any time in history.

Quite why the Liberals have been so dominant is a puzzle, for, as Epstein noted, the parties are federally structured and the organizations are “loose and non-doctrinal at every level.” The conventional wisdom has been that the Conservatives alienated themselves from French-speaking Canada at the time of the First World War and so allowed the Liberals an easy run. Given the geographic structure of the population, and the impact of the single-member district electoral system, that was undoubtedly true. The Liberals have not managed to carry Quebec since 1980, however, so their continuing success has a more complex explanation. André Blais (2005) argues that (with region) the key electorally-salient social cleavage in Canada remains religion, and it is Catholic support that accounts for Liberal predominance. But that itself leads to another puzzle, for as Blais notes, Catholics do not differ from other Canadians on the issues, are not particularly attracted by Catholic party leaders or candidates, nor is there a local contextual effect structuring their vote. An inability to account for why Catholics still disproportionately vote Liberal (by 18 percentage points in central and eastern Canada) has led Blais to “propose the creation of a special prize for the individual or team that solves the mystery” (830).

Referring to Canadian parties as non-doctrinal is a polite way of saying that they are opportunistic organizations more concerned with winning votes than defending or advancing any clear ideological positions. The major parties have always been prepared to steal policies from each other: in the 1980s they quite dramatically (and unapologetically) traded places on the question of free trade with the U.S., an issue that had divided them for most of a century and over which great electoral battles had been fought by earlier generations. This fluidity has been rationalized as freeing the parties to act as political brokers, able to accommodate the shifting and competing demands of a diverse electorate. And the need to practice brokerage politics as a way of holding a poorly integrated country together is held to account for the distinctive cast of the large parties' organizations.

These brokerage parties are dominated by their parliamentary leaders who have maximum freedom to define party policy and electoral strategy and organization. One hundred years ago Siegfried (1906, 136) wrote that the leader's “mere name is a programme in itself,” and that “the central organization of each party is reduced to a minimum,” namely the leaders themselves (118)—perhaps the earliest account of the “presidentialization” of politics in any parliamentary system. The parties that support these leaders, traditionally chosen in conventions, have been described as modern versions of 19^th-century cadre parties whose basic structure is that of a franchise organization (Carty 2002a; 2002b). Their essential organizational bargain is one of autonomy for the local constituency association (to manage its own affairs including nominating and campaigning for its preferred candidate) in exchange for tight centrally-imposed party discipline for members of the parliamentary caucus. Although that is a structure conducive to maximizing the parties' capacity to respond to the imperatives of a large number of distinctive communities (on which see Carty and Eagles 2005), it provides party leadership that is once both strong and fragile and so susceptible to system break-up when the society's demands stretch intra-party bargains too far.

From time to time, those who reject the politics of opportunistic brokerage have sought to build mass-style disciplined, ideological parties of the sort common in European parliamentary systems. The most persistent of these have been social democrats on the left of the political spectrum, but they have never managed to break the hold of the big brokerage parties on national government. Their New Democratic Party was badly shaken by the electoral earthquake of 1993 (see both Figure 1 and Table 1) and has only slowly re-established itself. In doing so, the party has emerged with a new face. Once a caucus of “agrarian socialists” (see Lipset 1968), the New Democrats have not won a single seat in their historic prairie heartland of Saskatchewan in the last two elections. Locked in an uncertain and electorally inefficient alliance with organized (English-speaking) labor, it is increasingly a party of urban Canada, able to exert some influence only when the Liberals are in a minority situation (as in 2004–06).

The federal structure of the country has led to a considerable diversity in the relationships between federal and provincial-level party organizations. In some provinces, party membership and activity is fairly closely linked, while in others there seems to be little connection between parties of the same name. The New Democrats have the most integrated organizations, the Conservatives the least. Indeed, the now governing Conservative Party has no presence in three provinces (which together have 40% of the population), while in the others Conservative supporters are still organized under the distinctive Progressive Conservative label. The Liberals' organization is quite varied with integrated membership in some provinces but quite separate organizations and structures in others.

The growing separation of federal (national) and provincial political worlds has led to a separation in their respective party systems so that few provincial and federal (electoral or parliamentary) systems are symmetrical, even within individual provinces. By-in-large the provincial systems are simpler, more stable, and often reflect a salient local social cleavage in the province—language, religion, and class being the most common (Stewart and Carty 2006). The large gap between the systems can be seen in the ease with which individuals cross lines when moving from one arena to the other: the current Liberal leader (and provincial premier) in Quebec was previously the national leader of the Progressive Conservatives, while former New Democratic Party provincial leaders and premiers are now senior figures in the national Liberal Party.

In most other democratic systems, political boundaries are less permeable and partisan organizations have a sharper definition. Canadian parties have adopted their distinctive characteristics, and the party systems their unique shapes, in response to the need to accommodate a fluid and dynamic federal society with the disciplining imperatives of British-style parliamentarianism. The institutional trade-offs necessary to provide for that balance have never proved to be easy nor enduring. This is what makes a study of the Canadian party systems so fascinating.

Epstein's instinct that there is much to be learned from a comparative analysis of Canadian party politics holds as true today as when he made it. Students of federal party organization will find that Canadian parties organize and behave differently than those in other federations—be they the U.S., Australia, or Germany. Scholars interested in understanding how parties operating in parliamentary regimes or under single-member plurality electoral rules will find Canadian parties do not look quite like those in similar systems. And those concerned with political parties' capacity to organize and articulate the diverse and changing continental societies in North America continue to have much to learn from systematic comparison.