Strengthening Canadian Democracy

Canadian Elections at the Millenium

Richard Johnston September 1, 2000

Since 1993, only the Liberal party seems poised to govern. Its popular base is weak and geographically unrepresentative, however, and its strength in Parliament is mainly a reflection of the fragmentation of the opposition. This fragmentation effectively excludes some parties from the game even as the opposition parties most able to achieve representation are limited in their ability or willingness to grow. The system thus delivers sharply disproportionate outcomes but no longer facilitates real alternation in power. There is a strong suggestion that voters are turning their backs on the system.

If some of the causes of fragmentation and withdrawal are universal, the particular form they take in Canada at the Millennium requires reference to history and geography, as the latter are processed by the country’s First Past the Post electoral system. The system encourages small parties to emphasize sectional appeals, but requires large parties to try and bridge regional differences. Large parties have usually failed to bridge all the gaps, especially in relation to the West. The most spectacularly successful bridging of gaps, by the Conservative party, created radically incoherent coalitions. The extreme case was the coalition assembled by Brian Mulroney in 1984, which managed to join nationalists within Quebec to the elements outside Quebec most opposed to that province’s claims. Before 1984, the Conservative party grew by absorbing much of the system’s earlier third-party impulse. The rest of that impulse was absorbed by the NDP, as the latter came to look more and more like a major party. The Conservative coalition’s explosion is the biggest part of the 1993 story.

The prospects for future consolidation of the right seem bleak. The Quebec nationalist part of the old coalition was not obviously an expression of right-wing politics to begin with. Only the Conservative party itself would have any realistic prospect of attracting this support back, and success in doing so would simply confirm that party’s unacceptability to the rest of the right outside Quebec. Reform, now the Alliance, is limited in its growth potential for mirror-image reasons. This leaves the Liberal party in a strategically powerful “interior” position, the only feasible party of national governance. The Liberals may be broadly acceptable for a long period, sometimes as a tepidly regarded secondbest. But it means that alternation of single-party governments, the life-blood of democratic politics in Westminster-type systems, is now highly unlikely. Thus, the time may finally have come for serious consideration of proposals for change in the electoral system, in particular for Proportional Representation.

Editors’ Note

André Blais, Paul Howe and Richard Johnston

In Canada and other industrialized countries, critical assessments of the quality of democratic life are commonplace nowadays. The refrain encompasses a wide variety of perspectives and concerns. Some worry that citizens no longer take seriously the duties and responsibilities of citizenship and hope to see traditional values re-instilled. Others are disturbed by declining participation in the organizations and associations that make up civil society and warn of the corrosive effect on social trust and political involvement. For others, the problem lies principally with outdated political institutions that do a poor job of facilitating popular participation in democratic life.

On all of these themes, there is a rich body of literature by scholars worldwide that can be mined for insights into the Canadian situation. Of particular relevance in the Canadian case is the important role that institutions play in structuring political participation and attitudes to the democratic system. For in many ways our political institutions, mainly inherited from the British and unchanged in their essence ever since, are illsuited to the demands and expectations of a modern polity. The House of Commons remains an assembly of MPs from small, geographically defined ridings that do not correspond to the constituencies relevant to modern policymaking. Parties, as ever, are essentially electoral machines, causing citizens to turn to interest groups and other organizations to engage in more meaningful political activism. Governments continually use opinion polls to make sure they are in step with the public’s views, but are reluctant to go the next step and make referendums a staple of the policymaking process. It may not be wise to overhaul all facets of the system, but a thorough examination of Canada’s main political institutions is in order.

In this first Choices paper in the Strengthening Canadian Democracy series, Richard Johnston looks at one institutional mechanism, the first past the post electoral system, which he believes is central to understanding what is probably the most salient characteristic of contemporary Canadian democracy, the fragmentation of the party system. For all the current maneuvering and jockeying taking place in opposition ranks, the Liberal party looks set to retain power for the foreseeable future. The four opposition parties, driven apart by disparate policies and confined to regional strongholds, are unable to bridge their differences — which is not undesirable in and of itself, but for the fact that it leaves the party system without an effective challenger to the government of the day. If our electoral system no longer excels at its supposed forte — clarifying lines of accountability and providing for decisive, if intermittent, alternation in government — the time has probably come to consider alternatives. Proportional representation (PR) would at least permit a more equitable expression of the votes cast on election day, while not necessarily underperfoming the current system in other respects.

In the short response piece we asked Lisa Young to prepare, Johnston is challenged on two points. First, Young questions whether the electoral system is chiefly responsible for the emergence of a five-party system in the 1990s. There are important social forces — the rise of Quebec nationalism, the growth of political cynicism, the newfound assertiveness of previously marginalized groups – that have made Canadian society more complex and help explain the fragmentation of the party system. Some of these dynamics, moreover, are common to most of the industrialized democracies, so that the made-in-Canada explanation Johnston advances may be off the mark.

Stephen Harper, in his reply, largely concurs with Johnston’s analysis of the past, but questions the scenarios and prescriptions he lays out for the future. If, for example, recent events highlight the deep fragmentation of the Canadian polity, we may need more than electoral reform to set matters right. Greater decentralization to the provinces is another obvious mechanism for accommodating social diversity. Furthermore, effective governance under a PR-system might require a more independent executive that could rule effectively despite division in the legislative branch. In the absence of comprehensive institutional reform, dissolution of the country becomes a plausible scenario.

Future Choices papers in our co-edited series will assess other elements of Canadian democracy, with a particular focus on institutional prognoses and remedies (though as Young reminds us, we cannot entirely set aside societal accounts of democratic change). Some papers will adopt a comparative perspective, considering, for example, the lessons for Canada of direct democracy as it is practiced in other countries and the role the media plays in political life elsewhere. Other contributions will be more Canada-specific, including a comparative examination of democratic practices in the Canadian provinces, and an assessment of the permanent voters list used for the first time in the 1997 federal election.

In another stream of the Strengthening Canadian Democracy initiative, IRPP will be releasing a series of working papers in its new Policy Matters series. The Institute has also been sponsoring public forums in different parts of Canada that are designed to enlarge the circle of contact with our project. If part of IRPP’s challenge is to generate insightful accounts and compelling ideas for democratic reform, strengthening Canadian democracy also means engaging Canadians in debate and discussion that spreads awareness and creates positive momentum for reform. The Institute hopes to achieve both sets of goals with its work in this important area.

Electoral democracy in Canada is sick. Whatever its faults, the old system at least delivered true competition for office and intermittent circulation of parties in government. Now, only the Liberal party seems poised to govern, but its base for governance is weak. Its traditional competitors, the Conservatives and, to an extent, the NDP, behave like fading great powers, unwilling to admit that their sun is setting. Its new competitors, Reform (now the Canadian Alliance)1 and the Bloc Québécois, harbour only an illusion of major-party status. The Bloc cannot seek to grow, for that would defy its nature. Yet its very existence may compromise the expression of Quebec’s interests. The Alliance does seek to grow, but risks repeating the cycle that led to the old system’s collapse. Voters, meanwhile, seem more and more reluctant to play the game at all.

How did we get here? Where, if anywhere, can we go? The paper begins by outlining where is here: a Liberal government weak in the country but strong in the House; an opposition too fragmented to mount a serious challenge to the Liberals; complacency on the surface of politics, but reasons for apprehension. Then follows an account of how we got here. It acknowledges that certain features of Canada’s predicament are local manifestations of global trends. It also acknowledges that the old system was, in a way, racing against time, as the underlying fundamentals of Quebec’s relationship with the rest of Canada evolved.

But, in keeping with the theme of the Strengthening Canadian Democracy project, my emphasis is on structural factors, on the electoral system in particular. Special attention is given to strategic choices by the Conservative party, choices which responded fully to the electoral system’s logic and which brought that party striking success, but which ultimately brought it low. As the Conservative party evolved, so did third parties. Some thirdparty change was not so much evolution as backwash from Tory initiatives. But some also reflected true third-party initiative, notably on the left. The implication in my account is that even the modest competitiveness of the old system was something of an illusion, that deep down only the Liberals could govern on a continuing basis. All along, the system may have been badly aligned with Canadian society. What the 1990s did was bring that fact out in sharp relief.

Finally, the paper looks downstream, to consider where the system might go from here. One possibility is that it already embodies natural, self-correcting tendencies, that true contestation for power will be restored. It seems more likely, though, that the system really is broken. This time the call for institutional redesign may not be a false alarm.

Where Are We Now?

In seats and votes, the political pattern of the 1990s is a mixture of opposites. The Liberals, the party of government, seem weak in the country but strong in the House of Commons. This disjunction represents the culmination of several trends, which together make the government’s true position more problematic than it seems on the surface. On the opposition side, the Alliance, the principal alternative to the Liberals, also has a weak vote base, weaker by far than any earlier pretender to government. In absolute number of seats, the Alliance is also weak. Yet it is not the smallest Official Opposition in history. Indeed Reform/Alliance after 1997 is stronger than the Liberals were after the 1984 election. These observations also hold, roughly, for the other main party of opposition, the Bloc Québécois. For all three parties, the secret to this disjunction lies in Canada’s First Past the Post (FPP) electoral formula.

Equally important is who is missing from the House, at least in force. The Conservatives, one of the traditional parties of government, may have been fatally weakened. So may the New Democratic Party, another pillar of the pre-1993 order. There is no mistaking how far these parties fell in popular esteem. But their standing in the House both exaggerates their plight and may yet worsen it, by a self-fulfilling process. If the Liberals, Reform, and the Bloc are beneficiaries of FPP, the Conservatives and NDP are now its victims.

Popular Weakness and Parliamentary Strength

The popular basis of the Chrétien government is the weakest of any majority government in Canadian history. Figure 1 indicates one facet of this weakness, by plotting government and opposition vote shares since 1878.2 For visual clarity, the figure groups elections into structurally distinct periods. The “First Party System,” 1878-1917, captures an era of close two-party competition. The next period, 1917-35, is a transition over which the first system died, but slowly and in a manner that only partly foreshadowed the next system. That system, the “Second Party System,” 1935-57, was absolutely dominated by the Liberals. The years 1957-63 brought a second transition. As with the first system, the first election in the transition ended the old system but barely revealed the next one. The “Third Party System,” 1963-84, saw the return of Liberal dominance, this time less overwhelming than in 1935-57. The last two periods mark the Mulroney-Campbell and Chrétien governments respectively. If the Conservatives’ nine years in power hardly qualify as a system in their own right, a government so powerful and so long lived does not seem like a mere transition. The pattern established in 1993 may last indefinitely and so qualify as a “Fourth” system, but it is too soon to say. In that pattern, the Liberals’ average share is almost two points lower than in their last protracted stay in power, 1963-84. Only one other majority government, elected in 1945, had a vote share as weak as the Liberals’ 1993 one, 41 percent. And at 38.5 percent in 1997, the Liberals reached a new majority-government low. Indeed, of all minority governments only two had smaller votes.3 The 1997 share was even smaller than a handful of earlier opposition shares.

Yet the Liberal pattern of the 1990s could be said to extend a trend originating in the 1920s. Since 1921, vote majorities have been rare and the trend has been generally downward, although the Mulroney Conservatives broke the trend temporarily. The downtrend is especially clear for the Liberals, the usual party of government.

The Chrétien government’s base is also narrow geographically, and this too continues a trend. Figure 2 plots one measure of geographic concentration, the standard deviation of the governing party’s vote share across provinces. The higher the standard deviation the more concentrated in one or a few provinces the party is. In one sense, the 1990s are not unusual at all, as sectional narrowness characterizes almost all governments since 1917. But the 1990s extend a disturbing pattern. When Conservative forces rally to form a government, they produce geographically quite inclusive electoral coalitions. But the rally never lasts long. Thus, the typical pattern has been Liberal government with a base that is both narrow and eroding.

For all that, the Chrétien government has been strong in the House, as Figure 3 reveals. It returned consecutive majorities, the first Liberal government to do so since 1953, and the majorities have not been weak by this century’s standards. The 1997 majority is thin, to be sure, but the 1993 margin was very comfortable, larger than any returned by Pierre Trudeau. Set against more than a century of elections, the parliamentary Liberals are currently neither peculiarly strong nor peculiarly weak. Three periods in Figure 3 exhibit stronger governments and three exhibit weaker ones.

Although the Chrétien government’s vote reproduces, even exaggerates, the regionally confined pattern of the Trudeau era, the party is more inclusive in seats. In 1993, the Liberals returned MPs from all ten provinces. The retreat in 1997 still left them with beachheads in nine. It is even tempting to propose that representation of regions improved in 1997 relative to 1993, in that more than one province was vital to maintaining the thin Liberal majority. Where in 1993, only Ontario MPs were a large enough block to deny their co-partisans a majority by unilateral withdrawal, in 1997, this was true outright of four provinces and almost true of two others.4 This stands in contrast to the effective exclusion of Trudeau’s last government from all of Western Canada.

The Liberal party is also the nearly universal second choice, according to the Canadian Election Study data in Table 1. Party groups are identified by self-reported vote in each survey’s postelection wave. Preference orders are inferred from party ratings on a 100-point scale. The Liberal tug was especially strong in 1993, when supporters of the four other parties rated the Liberal party far more highly than any other but their own. Where non-Liberal voters typically rated their own party about 65 on the scale, they rated the Liberals in the mid-to-high 50s. Party differences in Liberal rating were small. New Democrats rated the Liberals slightly higher than their own party, and even Bloc Québécois supporters rated the Liberals fairly highly. Tempers had clearly frayed by 1997, as the ratings gap between the Liberals and voters’ own parties widened considerably. In 1997, non-Liberal supporters rated their own party a little higher than in 1993, about 69 points, and rated the Liberals much lower, in the mid-40s. Gaps among voting groups were also wider, as Conservatives and New Democrats still rated the Liberals (just) over 50 while Blocistes and Reformers rated them around 40. Even so, only Bloc supporters typically named a party other than the Liberals (the Conservatives, in this case) as their second preference.

Similarly, the Chrétien government has enjoyed remarkable support between elections. Soon after each election, support in most polls — measured by respondents’ willingness to vote for the Liberal party if an election were held that day — soared. Support surges — “honeymoons” — have been the norm after majority victories, but in the 1990s honeymoons last entire Parliaments.5 If the Liberal crash in the 1997 campaign strongly suggests that these elevated support levels are illusory, they nonetheless seem to colour media coverage of the government and, perhaps, to embolden the government itself. Certainly, the Chrétien governments do not seem as much on the defensive as their predecessors were.

Fragmented Opposition

It is clear, then, that the disjuncture between the Chrétien government’s popular weakness and parliamentary strength has little to do with the Liberal party itself. The obvious place to look is across the aisle, to the opposition. And in the 1990s the outstanding fact about the opposition is its fragmentation.

First, consider the simplest facts, as revealed by Figures 1 and 3. The first lesson of Figure 1 is that opposition forces have fragmented even more than government ones. Where over the 20th century governments lost about 10 percentage points in the popular vote, the Official Opposition lost over 25 points. Fifteen of these 25 points were lost before the emergence of Reform and the Bloc. Of course, other opposition parties picked up the slack, as voters alienated from the governing party still must go somewhere. But that is the very point: they became less and less likely to cluster around a single pole of opposition.

This accounts for the disjunction, already hinted at, between the top line in Figure 1 and its equivalent in Figure 3: where governments’ vote shares have mainly shrunk, their seat shares have waxed and waned with no real trend. Governments in the transitional period of 1917-35 were weaker than in the old two-party framework that prevailed before 1917.6 But the Liberal governments of the 1935-57 period were arguably the most powerful in Canadian history, even though their popular base was smaller than in the pre-1917 systems. In the 1957-84 period, governments were weakened again. John Diefenbaker’s stunning 1958 victory was offset by the thinness of his 1957 and 1962 pluralities. Renewal of the Liberals after 1963 left them barely able to control House majorities. The Liberal grip on the House did tighten after 1968, but their majorities remained slim. Then, in 1984, the tide turned back toward governing parties. The Conservatives’ parliamentary strength under Brian Mulroney, in his first House at least, rivaled that for the 1935-57 Liberals, and the current Liberal government also seems comfortably situated.

Clearly important to the ebb and flow of Government House strength in Figure 3 is the ebb and flow of opposition fractionalization indicated by the bottom lines of Figure 1. Although the opposition vote trend is downward, it is not unbroken. Relative to earlier periods, the Official Opposition strengthened in 1957 and remained strong until 1984. The initial strengthening, 1957-63, was a bit artifactual, as it represented the temporary exile of the highly resilient Liberal party. More consequential is the fact that the Conservatives were so much stronger in opposition in the 1963-84 period than in the 1935-57 one. The Conservatives’ longterm vote gain (Figure 1) around 1960 was modest, about five points. But the seat gains (Figure 3) this brought were spectacular: a near doubling of the typical Conservative seat share, from under 20 percent to over 35 percent. This surge is a major theme below, for in it lie the seeds of the Tory débâcle of 1993.

With the Liberals’ defeat in 1984, opposition fragmentation resumed. Some of this reflected the thinning of the Liberals’ own base, whose continuance into the 1990s makes that party weak in the country if not in the House. Some of this 1984 fragmentation reflected further strengthening of the NDP, of which more below. Then came the 1990s and the emergence of Reform as the main pole of opposition.7 But Reform is weak, far weaker in popular vote than even the pathetic Conservative and Liberal oppositions of 1935-57 and 1984-88.

In seats, however, those particular oppositions were no stronger than Reform. Indeed, for most years between 1935 and 1957, Conservative oppositions were slightly weaker than Reform is now and the 1984-88 Liberals, with only 14 percent of seats in the House, were the weakest in history. Reform clearly squeezed many more seats per vote out of the system than did earlier weak oppositions. The secret of Reform’s success in seats is the party’s very narrowness in votes. Reform dominates only two provinces but there it is very dominant. Indeed, “too many” of its votes, strategically speaking, are concentrated in Alberta. Reform’s most “efficient” province, from this perspective, is British Columbia.

If Reform and, even more, the Bloc are somewhat more efficient than some earlier oppositions, they are stunningly more efficient than the Conservatives and the NDP are now. Each of these parties is only weakly differentiated by region, and neither currently dominates any single province. Only in Atlantic Canada do the Conservatives remain competitive. In Quebec and Ontario the Conservatives serve mainly to split the vote, among federalists in Quebec and on the centre- right in Ontario. The NDP is most competitive in its ancient heartland, Saskatchewan, but in no sense does it dominate that province any more. It remains to be seen whether the NDP’s 1997 gains in Atlantic Canada are sustainable.

Both parties used to be major players, of course, but in 1993 each lost two thirds of its former vote. For both, 1993 was an all-time low and 1997 was only marginally better. The Tories’ previous low came in a bitter election, 1945, in which the whole landscape trembled. Fortunately for the Tories, this was also a bad year for the Liberals.8 And the 1945 low was still eleven points above the Tories’ 1993 result and eight points above the 1997 one. For the NDP, the 1993 share was smaller even than the CCF share in this party’s first outing, in 1935. Each party took a geographic profile that sustained a large party under the logic of FPP and saw it forced downward across the board. Each party survives because of residual organization, but also because of strength at the provincial level. The

Worst of Both Worlds?

Canadians may thus be paying a price for our FPP electoral system without getting a commensurate benefit. The price is distortion of voters’ party preferences. The benefit is supposed to be the simplification of alternatives, which promotes, in turn, decisiveness and the clarification of lines of accountability. Table 2 conveys the total scope of distortion as well as the extent of simplification, election by election and with median values for selected periods.

Distortion is indicated by the leftmost column, labeled “disproportionality.” Its values indicate, roughly, the total deviation of seat shares from vote shares.9 Most values in Table 2 are fairly typical of FPP systems. On average, Canadian results are more distorted than those for US House elections, much less distorted than in India, and about the same as in the United Kingdom and New Zealand.10

“Simplification” is indicated by the two rightmost columns in Table 2, labeled the “effective number of parties.” The “effective” number attempts to make precise an intuition about the “real” number of parties in a system. Consider first a system with two parties. If each party is exactly the same size, then the number must be two. If one is a little larger than the other we may still feel comfortable calling the system a two-party one. What if one party is twice the size of the other? At the other extreme, what about a party too small to be a factor in the real competition? The “effective number” measure translates these concerns into a continuous index by taking a notion of fractionalization — a system is more fractionalized the larger the number of parties and the more equal their shares — and turning it upside down.11 Table 2 applies the “effective number” measure to both vote and seat shares.

The 1993 and 1997 elections score high, that is, above the all-elections median, on both distortion and fragmentation. On distortion, neither election is an extreme case. But every other election with above-average disproportionality was a landslide, a radically “simplified” result. The 1990s elections, in contrast, produced highly fractionalized Houses, with the 1997 result the most fragmented in history.

So here too the rules have changed. In the past, the more distorted the result, the smaller the effective number of parliamentary parties. The 1935- 88 relationship was:

Effective Number = 3.2 – 0.07 * Disproportionality. 12

By this relationship, a result as disproportionate as the 1993 one should produce a House of 2.0 parties, not the 2.4 that actually emerged. In 1997 the House “should” have the equivalent of 2.3 parties, not 3.0.

So many parties appeared in the House because so many appeared in the electorate. Both 1990s elections featured a record number of electoral parties, with only 1945 coming close among earlier results. The number of electoral parties testifies to the long-run failure of the old system at simplifying the alternatives. The argument for FPP in terms of consolidation and simplification has two parts, but only one part seems to hold in Canada. The first part, the one that works, is called the “mechanical” effect, the reduction in the effective number of parties as votes are translated into seats. Table 2 shows that, without fail, the number of parliamentary parties is smaller than the number of electoral parties, sometimes spectacularly so. A “psychological” effect is supposed to follow from this, but here the argument breaks down. Mechanical distortion on this scale should create powerful disincentives against fractionalization of the vote. Some voters sense that their votes will be wasted, and move strategically to a party that is both acceptable and viable. Parties pull candidates back where they are not viable and potential donors and volunteers restrain their largesse. Truly infeasible parties should thus fade and die, leaving two large parties, each competitive for power and each willing to accept underrepresentation when it loses as the price of controlling outright majorities when it wins. In the long run, this should mitigate, although of course not eliminate, the appearance of disproportionality.13

In the second system, 1935-57, the typical outcome was dramatically distorted. The typical disproportionality value, 19.7, was larger even than the actual value for 1993. The typical mechanical effect was to strip one whole party from the system (compare the electoral and the parliamentary medians for the system). The second-system pattern, then, was to take a somewhat fragmented vote and, through massive distortion, produce a onesided House, a House with fewer than two effective parties. By rights, this “should” have produced a downtrend in the effective number of electoral parties, as insurgent voters and elites were discouraged. Instead, the number of electoral parties stayed close to three, and then, after 1960 and the transition to the third system, kicked up over three. In the third system, the typical pattern was to take a quite fragmented vote and turn it into a modestly less fragmented House. The 1984 and 1988 elections scaled the fragmentation back a bit, but did not depart from the basic disproportionality-fractionalization relationship of earlier elections.

Now we have returned to distortion on the scale of the second system, 1935-57, but with parliamentary fragmentation exceeding that of the third system, 1963-84. In the second system oppositions were weak. Even if they occasionally embarrassed the government, they never came close to threatening its existence.14 In the third system, it was governments that were weak. At the moment, the pattern rather resembles the second system, in that governments are strong in the House out of proportion to their strength in the country. But will the Liberal government hold its majority past the next election? We may be on the verge of a situation in which both government and opposition are weak.

So we may have the worst of both worlds. It is a given that FPP produces more disproportionality than does the main alternative, a Proportional Representation (PR) formula.15 To the extent that disproportionality is mitigated, it is because parties which validly represent bodies of opinion are suffocated. FPP has other pathologies too, which this paper also chronicles. These things are all bad in themselves. But their flip side should be the creation of simple choices, effective governments and credible oppositions. The Canadian system has done a reasonable job in empowering government, in that most of the time single-party seat majorities appear. On the opposition side, the Canadian system has done a poor job for years, as oppositions have often been weak. And now the system may be on the verge of failing to empower governments. A multi-party legislature where no single party holds a majority is not a bad thing in itself. But if such a legislature becomes the norm, we must then ask whether there is any more point in accepting distorted results.


For the time being, the pattern may be messy but tolerable. The Liberal government enjoys the full measure of majority power. This may or may not be a good thing in itself but good or bad, the power of House majorities is not an issue peculiar to this decade. The government’s majority is more “manufactured” than most, but manufacturing is, again, not a new issue. The government’s programme has not provoked widespread opposition. And its standing between elections is eerily high.

“Eerie” does seem the right word, for the suspicion lingers that the absence of sharp debate and popular unrest makes for an inadequate test of the government’s true standing.16 In the 1997 election, the government’s support dropped sharply over the campaign, and its seat majority seems to have been saved only at the end.

The situation at present carries some echoes of the mid-1950s. Accounts of that period suggest popular distress at a government unresponsive to forces on either right or left, and arrogant toward the House of Commons. But none of the great parliamentary battles of the mid-1950s registered in downtrends, not even blips, in poll readings of government approval. What was missing, we must infer, was someone able to articulate the discontent as leader of a plausible alternative government. Only in 1957 could John Diefenbaker become the tribune for expressing the discontent. 17 Two things about Diefenbaker are relevant. One was his undoubted political skill. The other was that he inherited a party that, although weak, was still a consolidated, pan-Canadian alternative. Current discontent with the Liberal government does not seem on the scale of that in 1957. But in the absence of a credible alternative to the Liberals, how would Canadians express such discontent?

One answer might be, by exiting the arena entirely. And voters may already be sensing the impasse as, according to Figure 4, turnout in both 1993 and 1997 was below the postwar average.18 From 1945 to 1988, turnout as conventionally calculated, the number voting as a percentage of the number registered (the middle line in Figure 4), averaged 75.4 percent, with no trend. The corresponding figures for 1993 and 1997, respectively, are 69.6 and 67.0. The 1993 number is probably an underestimate as, outside Quebec, the voters list was carried over from the 1992 referendum and almost certainly had a surplus of the deceased and the recently moved. But most of the drop was real,19 and there is no explaining away lower turnout in 1997.

Indeed, this indicator understates the problem. As a gauge of overall eligibility to vote, the number of registered voters has deteriorated, with a particularly dramatic drop in 1997. As a percentage of the voting age population the registration rate used to fluctuate between 90 and 94 percent, and did so right down to and including 1993, according to the top line of Figure 5.20 But in 1997, the registration rate dropped to just over 85 percent. Why it dropped so dramatically is beyond the scope of this paper. It could be that the 1990s brought so large an immigration flow that the meaning of “age-eligibility” is different for the 1990s as compared with earlier decades. It could be that administrative changes account for the drop, as Elections Canada embarked on a transition to a permanent voters list in 1997.21

Whatever the reason for the drop in registration, turnout relative to age-eligibility in 1997 looks bad. From 1945 to 1988, there may have been a net downtrend in turnout calculated this way, but the trend was very modest. Turnout so measured averaged over 70 percent until 1965, then began a gradual drop. Only once since 1968 has the rate exceeded 70 percent. The 1993 rate, just below 64 percent, was a new postwar low. But 1993 pales in comparison to 1997, at 57 percent.

This puts Canada near the bottom of the industrialized- world turnout league tables. Measured by this standard, Canada comes off only a little better than the US. In the 1996 US House election, for instance, turnout measured this way was 49 percent, a very disturbing figure but not that much below Canada’s. No other G-7 country besides the US has turnout as low as Canada’s and only Japan and France have comparable rates. In round numbers, Canada is now 10 points below the United Kingdom, 15 points below Germany, and 30 points below Italy. Canada has never had peculiarly high turnout, but the gradual decline from the 1960s to the 1980s, followed by the precipitate drop in the 1990s, has taken us from the lower middle of the pack to very near the back.22

How Did We Get Here?

A Global Problem?

Does the shattering of old patterns reflect a global process of cultural and technological change? An argument along these lines has considerable currency and deserves to be taken seriously. This section outlines some arguments and reviews cross-national evidence. The review unveils commonalties of politics at the millenium, but it also suggests that differences among countries are as striking as convergences. Some, at least, of our problems are homemade. In particular, some originate in the strategic logic embedded in the electoral system.

That the western world has undergone a culture shift in the last thirty years seems undeniable. The strongest statement along these lines lies in the work of Ronald Inglehart,23 who claims that the ubiquitous prosperity of the postwar industrial world produced a generation with no experience of material want. This generation is thus largely unmoved by older controversies over shares of the material pie. The new sensibility resists old, consolidated forms of political action: big parties, big unions, the state generally. Moreover, newer, higher-order “postmaterial” drives are not easily shed, even if economic conditions worsen. The shift is thus permanent and is differentiated by period of birth. Canada, unsurprisingly, is not exempt from this culture shift.24

The claims are hotly contested, but the broad sweep of the thesis seems intuitively right. If the Inglehart thesis is hampered by weak measures and loose conceptualization,25 it nonetheless captures the culture shift in gross outline. The political agenda has either shifted or expanded, as claims rooted in identity and in secular individualism (admittedly somewhat contradictory bases) have gained ground. That many groups and political actors resist these claims is part of the point. How parties respond to this broadening and reorientation of debate is the concern of the most influential interpretations of the evolution of both the right and the left in Europe. Herbert Kitschelt, for instance, shows how for all of Western Europe the old capitalist/socialist dimension has been supplemented, although not supplanted, by an authoritarian/libertarian one. As well, the balance of electoral power has shifted to the right. The axis of competition in most countries combines the old and the new dimensions.26

If the form of the challenge is ubiquitous, response to it by voters and parties is not. Some old parties, constrained by internal politics, failed to adapt to the modified agenda, and suffered accordingly. Others were able to absorb the new lines of conflict. The difference is political, and, notwithstanding evolutionary cultural universals, the centrality of politics makes each country rather unique. This pattern — the importance of each country’s actual political history and the weakness of continent-wide trends — runs through most recent systematic investigations of European electoral patterns. For example, the psychological claim of parties has, on average, diminished, in the sense that fewer voters claim to identify with a party and, perhaps most critically, fewer claim a strong attachment. But in some countries partisanship so measured has actually intensified and in others, little change is visible. In general the differences among countries are more striking than the generality of the downtrend.27 Similarly, there appears to be no continent-wide trend toward party-system fragmentation or toward systematically higher levels of volatility.28 Trust in political institutions shows no Europe-wide trend, although individual countries exhibit movement. As with partisanship, trust is affected by political factors, and here a specific pattern sticks out: the more frequent is governmental turnover, the lower is the average level of trust.29

Also showing no Europe-wide trend, but rather a collection of country-specific stories, is electoral turnout.30 It is true that the countries most immediately comparable to Canada, the big, rich countries of the G-7, exhibit declining turnout. Of these, only Italy shows no trend. Calculating turnout the same way as before in this paper, as a percentage of the age-eligible population, and plotting trends, other countries have seen turnout drop in the postwar period from five (UK) to twelve (France) points. With a fifty-year postwar downtrend of 9.5 points, Canada’s drop is a bit greater than the G-7 average. Timing varies considerably across the G- 7. In the US, the UK, and France, turnout has been dropping in small steps for many years. In Germany, the downtrend seems to start in the 1980s. Only Japan is like Canada in seeing virtually the entire drop postponed to the 1990s. There may, then, be a story about secular change in turnout that is almost G-7-wide. But as with the European examples, differences across countries are at least as striking as the generalities.

One element in the “postmaterial” pattern does have real resonance in Canada, the rise of identity politics. The identities in question may be ancient, but their expression was inhibited by material want and global military insecurity. More recently, however, the very success of the West, economically and militarily, has loosened restraints on identity demands. In Europe, the consolidation of the Union has empowered national minorities in relation to their central governments, as, ironically, the cost of ethnic assertion has dropped.

This is a theme already familiar to Canadians. The emergence of the Bloc Québécois represents a certain electoral coming-of-age for Canada’s largest national minority, as forces already strong in Quebec’s provincial politics are now projected onto the federal scene. Here too, however, it does not suffice merely to allude to global patterns. The oddity of this universal trend is that it represents the reemergence of the particular, as arguments are couched increasingly in terms of history, memory, and place. This too directs us back to the indigenous elements in the country’s partisan history. The place to start is the electoral system.

The Centrality of the Electoral System

Clearly the electoral system is implicated in much of the earlier discussion. It was responsible for the vote-seat distortions that mark the 1990s. More generally, it affected the parliamentary power of winning parties, somewhat independently of their own vote-drawing power. A system with this much power constitutes a powerful set of strategic incentives in its own right. It does not just translate behaviour, it is an active force in shaping it.

To see how this is so, begin by reconsidering the seat-vote disjunctions discussed above. One indicator of disjunction is the ratio between the two quantities — seats and votes — for a given party. Winners under FPP almost always have ratios greater than 1.0, they win a larger percentage of seats than of votes. Losers, including the Official Opposition, commonly have ratios under 1.0, they win a smaller share of seats than of votes. The Liberals’ 1993 ratio, 1.47, is the fourth highest in history, after the 1935 and 1949 Liberal landslides and the 1984 Conservative one. It is slightly larger than the ratio for the 1958 Conservative landslide. Of course, the Liberal victory in 1993 was no landslide, just a comfortable majority. What makes the ratio so high is that this comfortable majority was generated by so weak a popular vote share. At 1.32, the 1997 ratio is lower than in 1993 but still well above the median for majority governments.

Does this mean that the Chrétien government has somehow changed the informal vote-to-seat translation rules? Far from it. Indeed, the Liberals won fewer seats in the 1990s than they “should” have, given the old rules. What has changed, as already outlined, is the fragmentation of the opposition. Consider this simple representation of the how Liberal and Conservative votes traditionally combined to produce Liberal seats: Liberal Seat % = 3.6 + 1.86 * Liberal Vote % – 0.90 * Conservative Vote %.31

What this means is that a Liberal vote gain of one percentage point would increase the party’s seat share by 1.86 points or, depending on the size of the House, five or six seats. Stated this way, the relationship assumes that the Liberal shift is not accompanied by a Conservative shift. If the Liberal gain comes at the Tories’ expense, that is, the Conservative vote share also drops a percentage point, then the Liberals would gain another two or three seats. Of course, these relationships also work in reverse, for Liberal losses and Tory gains. In the typical election, Liberal gains were in fact complemented by Tory losses, and vice versa. But not entirely. A striking long-run pattern in Figure 1, where government and opposition are compared, is asymmetry. Notwithstanding consolidation of opposition in and after John Diefenbaker’s period at the Tory helm, the underlying story has been a widening gap between government and opposition. The Tories’ collapse in the 1990s may be the final chapter in that story. It was certainly a gift to the Liberals.

Were the Liberals able to cash the gift in completely? Not quite, it turns out. Just cranking the 1993 and 1997 Liberal and Conservative votes through the equation above would yield Liberal seat shares of 65.5 and 58.3 percent, respectively. In fact, Liberal shares were rather lower each time, 60 percent in 1993 and 50.8 percent in 1997. So even though the seat/vote ratios seem high, they were not high at all given the weakness of the ancient foe. Why then did the Liberals not extract even more seats from the system? The answer has two parts.

First, the Liberals were too concentrated geographically for their own good. In certain provinces, they “worked” the map just right. British Columbia and Quebec may be examples: weak shares province-wide, but sufficiently concentrated (too concentrated in Quebec, perhaps) in the metropolitan centres to guarantee a decent representation. But in Ontario and, in 1993, Atlantic Canada, the Liberals wasted votes. The Liberal shares were so huge in those places that no amount of opposition consolidation would have turned more than a handful of seats, if any. No consolidation occurred, of course, and so the Liberals could have won about as many seats with a rather smaller vote, such that the “surplus” could have been put to better use in other regions.

Second, although the opposition was fragmented overall, it was not equally fragmented everywhere. Of the four serious opposition parties, two — the Conservatives and NDP — were geographically dispersed and two — Reform and the Bloc — were concentrated. The concentration of the latter two limited the ability of the Liberals to capture seats. Put another way, Reform and the Bloc captured so much of the old Tory vote in BC, Alberta, and Quebec that they did not so much fragment that vote as replace it.

In this narrative lies a generalization about the First Past the Post system. Success under FPP is not just a matter of accumulating votes. Also important is how those votes are distributed across the landscape, as parties are encouraged to think geographically. What it means for a party to think geographically is contingent, however, on how many votes it starts with: the fewer the votes, the more concentrated in one or a few places they should be; conversely, the more votes a party controls, the more spread out they should be.32 Opposition parties in the 1990s illustrate both sides of the first proposition: Reform and the Bloc benefit from concentration; Conservatives and New Democrats suffer from dispersion.

FPP in Canada is often criticized for sowing division, for encouraging parties to adopt regionally divisive strategies. The logic of the previous paragraph suggests that the criticism cannot be made unconditionally. Unquestionably, FPP encourages small parties to make narrow appeals and, to the extent that small parties flourish, the House may become a cacophony of sectionalism. This does seem to be a story of the 1990s. But large parties are encouraged to broaden their appeal, to craft geographically inclusive coalitions and to look for issues that cut through region.33 And most Canadian governments have, in fact, succeeded in pursuing an inclusive strategy.

First consider the place of Quebec, arguably the historic pivot for government. Figure 5 plots the place of Quebec and the other regions in the formation of governments. For each region, the smooth top line is its total number of seats in the House, a gauge of potential influence. The jagged line below is the number of seats from the region won by the Canada-wide winner, a gauge of the region’s actual influence. In all but three elections from 1935 to 1988, the winner also won a majority of Quebec seats. In over half the cases, withdrawal of Quebec seats would have deprived the winner of a House majority. In four of those nine, withdrawal would also have denied the winner a plurality, also the case with the three minority governments in which Quebec was well represented. The three elections in which the winner secured fewer than 20 Quebec seats were all Conservative victories, all minority governments, and all short-lived. Over most of this period, Quebec adhered to the Liberals and made them the only feasible party of government. That government might be weak, indeed a minority for critical periods, but installation of an alternative, Conservative government required an extraordinary and unsustainable concatenation of forces. All that seemed to change with Quebeckers’ willingness to support the Mulroney Conservatives. This apparently heralded the return of true two-party competitive politics, as the 1984 swing toward the Tories mirrored the earlier 1887-96 swing away from them. Quebec thus played the pivot in two equally vital ways: it arbitrated two great realignments; between these realignments, it made the Liberals virtually the sole feasible party of government.

Some of this power reflects mere size, for Quebec has always held a large body of seats, 65 at the beginning and 75 now. At present Quebec controls slightly under 25 percent of the House, and in earlier decades the share was larger. But Quebec’s importance goes beyond mere numbers. The province’s voters also used their voting power effectively, by consolidating support for a party that enjoyed significant support elsewhere and, when switching, doing so nearly wholesale. Comparisons with Ontario (Panel B in Figure 5) are instructive. Over this period, Quebec’s total number of seats always lagged Ontario’s, and the gap has widened. But each major party’s range of outcomes was 10 seats greater in Quebec than in Ontario. In Quebec, the Liberal minimum was 12 seats and the maximum, 73, for a range of 61. The corresponding Conservative numbers were 1 and 63, for a range of 62.34 In Ontario, the Liberal minimum and maximum were 14 and 64, for a range of 50. The Conservative numbers were 17 and 67, also a range of 50. The difference in seat pattern follows one in popular vote, as the range of vote shares was much greater in Quebec than in Ontario. This vote flux in Quebec is then translated into seat flux with remarkable efficiency, as spatial variance in vote shares within Quebec is low.35 Thanks to this consolidation of seats, Quebec representation in government almost always bulked about as large as Ontario’s. Indeed, in nine of the seventeen 1935-88 parliaments, Quebec members on the government side outnumbered Ontario ones.

On one hand, this is a story of sectionalism. Quebeckers appear to act on a sense of provincial interest, and major parties respond. On the other hand, Quebec seats never suffice by themselves to constitute a government. To win a majority, a party must, at the very least, match the Quebec total with seats from other regions. Usually, the requirement for seats outside Quebec has been greater.36

The other big lump of seats is Ontario, of course, but before 1993 Ontario did not play its weight. Over the years since 1935, the province’s total number of seats grew, but until 1993 its average number of government seats shrank. In one sense, then, the province became less, not more of a counter in the government game. It is true that when the whole country swung one way, so would Ontario. Thus in the 1958, 1968, and 1984 sweeps, the winner’s Ontario share was 60 to 70 seats. But these shares were not sustainable and three times the winner captured fewer than 40 Ontario seats. In contrast to the Quebec case, two of these three cases involved Liberal winners, in 1945 and 1972. Each time, the Liberals’ weakness in Ontario was a setback to the government, but only a temporary one. The essential fact about Ontario in comparison with Quebec, according to comparison of panels A and B, is the greater fragmentation in its seat shares.

But Ontario is still bound to be important simply for its sheer bulk: currently it holds roughly one-third of all House seats. In the typical pre-1993 House, about 50 government seats were from Ontario, smaller than the typical Quebec share but still more than one-third of the requirement for a majority.

Atlantic Canada also usually found a place at the table. It was next to impossible for this region to contribute more seats than Ontario to the winner. By the end of the period the whole region controlled fewer seats than British Columbia, not to mention Ontario or Quebec. Even so, from the 1930s to the early 1960s the region rivaled Ontario as a source of government seats. Over the whole period, Atlantic Canada contributed about half as many seats as Ontario to governing parties. Twice, 1945 and 1962, Ontario largely abandoned the governing party, so that the Atlantic region held as pretend many government seats as the biggest province.

So far, then, the story is one of inclusion. The very biggest place is well represented but not absolutely dominant. The smallest region is almost always present in some kind of force and occasionally rivals the biggest region. Quebec plays a key role by consolidating internally and, if an acceptable alternative presents itself, by swinging en masse. This is worth underlining. Whether or not one considers Canada a binational state, few doubt that the biggest challenge is the continued integration into the pan-Canadian scheme of the one jurisdiction with a francophone majority. The FPP electoral formula has historically facilitated — and may even have forced — a linguistically inclusive strategy on prospective governing parties.37 Three of four regions routinely win serious representation at the centre. Presumably this reflects active solicitation by the major parties.

The problem, of course, is the West. For most of this century the region has dwarfed the Atlantic provinces, and in recent years has pulled away from Quebec as well.38 British Columbia alone now has as many seats as Atlantic Canada. But this has not been cashed in on power at the centre. Over half the time from 1935 to 1988, the government party returned more seats from Atlantic Canada than from the West. Only five times did the region contribute seats in proportion to its intrinsic strength, and two of these (1962 and 1979) were short-lived minority governments.

It might have been tempting to see the Mulroney years as the region’s political coming of age. Not only was the West the springboard for the Mulroney government’s majority, the majority in question lasted two parliaments, the first consecutive majorities since 1953, the first for the Tories since 1891. The Conservatives’ strength in the West, when joined to a breakthrough in Quebec, seemed to form the basis for the first real two-party politics since the 1890s. It was not to be, of course. The key element in the 1993 election was the evaporation of Tory strength, most dramatically in these very regions. To understand that evaporation we have to go back to how that strength was built in the first place.

The other party that evaporated in 1993 was the NDP. Its pre-1993 history, in many ways, complemented that of the Tories. The NDP exists, arguably, because the Diefenbaker Conservatives preempted some of the old CCF base. But by reconstituting itself in a particular way, the NDP also changed the character of third-party activity. The explosion of 1993 was, in part, a recrudescence of forces that, between them, the post-Diefenbaker Tories and the NDP redirected.

The 1993 cataclysm was not all old West wine in new Reform bottles. The official opposition in 1993, after all, was the Bloc Québécois. For the West, 1993 arguably just restored an old pattern. The real novelty was the shrinkage of Quebec’s presence at the centre of power. Figure 5A indicates that the years since 1993 are the only ones in living memory in which Quebec had, for more than a few months, fewer than 30 seats on the government side of the House. This may signal a fundamental change to the informal rules of government formation. It certainly signals a fundamental change in Quebec’s own orientation to the system.

First Moves: Conservatives to the West, NDP to the East

Between the 1950s and the 1980s the Conservative Party made itself a progressively more formidable electoral force. Some of this it accomplished just by growing. Equally important, however, was the party’s efficient response, whether intentional or accidental, to the electoral map. Figure 6A, “Tories move out of Centre,” illustrates the first step, accomplished mainly under John Diefenbaker’s leadership. The light line gives the 1935-53 pattern, not a pretty picture for a party pretending to alternative-government status. The party’s places of relative strength were places of absolute weakness. The Conservatives averaged about 40 percent of the vote in Ontario and the Atlantic provinces, some ten points behind the typical Liberal share. In the rest of the country their typical share was simply pathetic: just over 20 percent in Quebec and around 20 percent in the four Western provinces. The Western reading is true for the whole period, but understates how much worse things were in that region than in Quebec. Two elections infused with wartime passion, 1940 and 1945, polarized French and English Canada and drove the Quebec Tory share under 20 percent. In the other three elections the Tories’ Quebec vote hovered around 30 percent. This fluctuation mattered hardly at all in seats, as the Tories won few in Quebec throughout the period. Countrywide, the Tories received an average vote share of 30 percent and a seat share of 19 percent.39 In this period, the Conservatives were barely plausible as an opposition party much less as a governing one.

In the West, Conservative weakness was not matched by Liberal strength. Both old parties were weak in the region, so Western politics was thirdparty politics. The West was the primary site for Canada’s first great electoral insurgency, the Progressive movement of 1921. Just as Alberta was the epicentre of the 1921 earthquake, so was it the main site for third-party activity in the 1935-57 system, where, according to Figure 6b, parties out of the mainstream averaged over 50 percent of the vote. British Columbia and Saskatchewan were not far behind, however, as third parties routinely pulled in over 40 percent, and the gap between Manitoba and all provinces to its east was large. The Western parties numbered two: Social Credit, which dominated Alberta and gained importance in BC; and the Cooperative Commonwealth Federation (CCF), with special strength in BC, Saskatchewan, and Manitoba.

In this period, then, Canada harboured two geographically segregated party systems. In the East, the two old parties prevailed, only weakly challenged by insurgents. In Quebec the Conservatives were terminally weak, but were complemented less by third-party activity than by hegemonic Liberal strength.40 In the West, third parties were strong, in some places the prevailing force.

Between the 1950s and the 1960s the Conservative Party gained considerable ground, less by augmenting its vote than by increasing its geographic efficiency. Where the party’s average vote share moved up only 4 points, to 34 percent, its average seat share doubled, to 38 percent. The modesty of the net vote gain reflects the offsetting of massive gains in medium-sized provinces by modest losses in large provinces, as indicated by Figure 6A. Where before the Conservatives thrice squandered 30 percent of the Quebec vote on virtually no seats, now they wasted fewer votes.41 Their decline in Ontario left them weak in the province, but their seat position there was already weak. In the West, conversely, they moved up dramatically, such that they tended to win outright majorities in Alberta and solid pluralities in Saskatchewan and Manitoba. The figure understates how well the party did in BC. Where in the 1960s the party reverted to its abject shares of the earlier system, in the 1970s it marched toward a plurality position, comparable to those in Saskatchewan and Manitoba. The party initially moved up in the Atlantic provinces, then fell back. In the 1962-80 elections, then, the Conservatives were a party initially of all outlying regions and then quite distinctively a party of the West.

As the Tories became more a party of the West, the West became less distinctively the home of third parties, as indicated by Figure 6B. In every Western province the third-party total fell, while third-party voting became a fact of life elsewhere. In BC and Alberta, Social Credit just evaporated. In Saskatchewan and Manitoba, the party losing ground was the successor to the CCF, the NDP. Within the West, BC displaced Alberta as the province with the largest ongoing third-party presence. Ontario moved up to a level comparable to the Prairie West. Most notable, however, was the rise of third-party voting in Quebec. In 1962 an incarnation of Social Credit burst on that province’s scene and lasted for nearly the entire period, weakening only at the end.

If all this still left Canada with a regionally differentiated system, the pattern of differentiation was sharply modified from before. The West remained distinctive from the East, but only in the strength of parties, no longer in their identity. The dominant party in the region, the Conservatives, had a significant presence everywhere else except Quebec. The West gave the Liberal party a weak share, but Liberals still remained in the region’s game. The West was distinctive in that its secondplace (in some provinces, first-place) party, the NDP, was overall the system’s third party, but the NDP was now more competitive for the biggest prize of all, Ontario. The most distinctive region was now Quebec, the only province clearly dominated by the Liberals yet also the new home of Social Credit.

As third-party voting spread, it became, as it were, domesticated. This is the point of Figure 7. Between 1921 and 1988, the total third-party vote became progressively less episodic, less susceptible to short-term flux. In part, this reflects the emergence of the CCF, later the NDP. The original Western impulse, in 1921 toward a loose grouping called the Progressives, embodied at least two tendencies. One was protest, which produced surge and decline, according to the state of the wheat economy. The other was programmatic, and this impulse finds its clearest expression in the CCF-NDP, whose series is dominated not by short-term flux but by gradual growth. Between the second and third party systems, short-term flux lost importance generally, although it did not disappear. Third-party activity increasingly came to mean NDP activity. Although the total third-party vote, and thus the non-NDP vote, surged in 1962, this did not signal a classic Western insurgency. Rather, it was the appearance of Social Credit in Quebec. Even this surge only temporarily delayed the NDP’s rise to monopoly of the third force.

Second Move: Conservatives to Quebec

The 1962-80 Conservative pattern — strong and growing in the West, consolidation elsewhere except Quebec — left the party powerful in opposition but still infeasible as a party of government. After 1962, they never won more than 36 percent of the total vote, and only twice, more than 40 percent of the seats. Their high point is telling. In 1979, on 36 percent of the vote they received 48 percent of seats, a remarkable showing for so modest a vote. But they won hardly any seats in Quebec and were lucky to win as many as they did outside that province.

As argued above, seats in Quebec tend to be available en bloc. For the Conservatives to win an outright seat majority, then, the obvious move was to invade Quebec, and this they did in 1984. Their 1980-84 transition looked simple: under the leadership of Brian Mulroney, they just added Quebec to the pre-existing coalition. Figure 8A indicates that almost all their gain over the 1962- 80 system came in Quebec.42 This gave them a smashing victory in 1984 and helped ensure that their 1988 majority was still comfortable. About half the Quebec gain came at the expense of the Liberal party.

The other half came from Social Credit. Figure 8B records that the third-party total in Quebec dropped by half, and this drop in turn is about half the total Conservative gain. After 1984, Quebec exhibited one of the weakest third-party shares, and most of this accrued to the NDP. The disappearance of Social Credit in Quebec completed the “domestication” of third parties. Figure 7 indicates that 1984 brought the smallest non-CCF/NDP component ever in total third-party voting. The 1988 pattern was essentially the same, although the slight widening of the NDP/“total” gap in that year was a warning sign: it reflects Reform’s first candidacies in Alberta.

Of course, the Conservative move into Quebec was not as simple as all that. The Tories exploited Quebec’s province-focused nationalism, which, for all its growth and prominence in provincial politics, was blocked in federal politics. The Liberals had been mobilized to fight that nationalism even as they asserted another, rights-based version of the francophone project. By their openness to Quebec nationalism, the Tories, perhaps inadvertently, expanded the agenda of federal politics.43

In one sense, the 1984 and 1988 elections indicate the resilience of Canada’s old parties. The very oldest, the Liberals and the Conservatives, continued to dominate the system. Indeed, they appeared to have restored much of the system’s pre-1921 competitive balance. Although they could not banish third-party voting, both of them were competitive for national office. If the Liberals lost ground after 1984, they were still imaginable as a party of government. Most critical is that they were no longer the only imaginable party of government. Even the “third” party, the NDP, could at least dream of power (and actually experience it in certain provinces). Its typical share by the 1980s was close to two-thirds the share returned by the next larger party. The NDP may have been perched on the brink of the inner circle. To the older parties, especially the Liberals, this might be an upsetting fact, an indication that their hegemony was still incomplete. But the NDP threatened neither a strictly partisan conception of politics nor the Westminster system. And in many ways, the NDP was already in the inner circle, whether of parliamentary procedure, of media access, or of finance.

The 1988-93 Transition

The resilience must have been more apparent than real, as the system collapsed in 1993. What voters did to bring the collapse about is shown in Table 3, which depicts the flow of voting and nonvoting between 1988 and 1993.44 Rows denote behaviour in 1988 and columns, behaviour in 1993. “Behaviour” includes non-voting (“Abstain”), being too young in 1988 but coming of age by 1993 (the row labelled “Entering”), and dying between 1988 and 1993 (the column labelled “Leaving”). Percentages sum to 100 across rows, that is, within 1988 groups. Take, for example, the top left cell, which indicates that of 1988 Conservative voters, 23.1 percent stayed with the Conservatives. The next cell along indicates that 19.2 percent of 1988 Tories shifted to the Liberals. On the 1988 “Abstain” row, the largest concentration of 1993 partisans, 19.4 percent, lies in the Liberal column, and among 1988 abstainers a Liberal vote was nearly three times as likely in 1993 as any other vote. But most likely of all, at 53.9 percent, was to abstain again. And so on. Note finally that percentages also appear along the bottom and right margins. The right, or row, margin gives the distribution across all alternatives for 1988. The bottom, or column, margin does the same for 1993. Including percentages for non-voting categories makes party shares seem smaller than we are accustomed to.45

Start with the party that commonly gets lost in discussion of 1993, the Bloc. The Bloc conjoins two rather different electoral streams. One is sovereignist and this stream flows mainly from outside the federal party system entirely. The other is disgruntled Tories, some of whom, presumably, are also sovereignist.

Of newly eligible voters, over 16 percent supported the Bloc, almost twice the Bloc share in the full electorate. Of voters eligible in 1988 but who abstained that year, almost seven percent chose the Bloc, the only party that attracted such voters at almost the same rate as in the whole electorate.46 Together, these new voters constitute over 40 percent of all Bloc supporters.47 We must suppose, however, that much of this mobilization is to federal elections rather than to elections as such. Many new Bloc voters are old political hands, willing to turn out in provincial politics where sovereignty has been a voting option since the 1960s but unable to assert the option in federal politics until this year.48

This interpretation is reinforced by two other pieces in the puzzle, voters inscribed for 1988 as New Democrats or as “others.” About five percent of 1988 New Democrats defected to the Bloc. As relatively few 1988 New Democrats lived in Quebec and thus were even eligible to shift to the Bloc, this statistic suggests that virtually the entire 1988 Quebec NDP vote switched to the Bloc. Similarly, two-thirds of 1988 “other” voters shifted to the Bloc. “Other” party voters were not numerous, of course, but most were in Quebec, and most supported the Parti Rhinocéros. The near-unanimity of this defection strongly implies that both the Rhinos and the Quebec NDP were parking spots for sovereignists.

Among the two big, old parties only the Conservative party was a truly significant source of Bloc support. A 1988 Tory was over four times as likely as a Liberal to vote Bloc, such that former Conservatives supplied over one-third of the total Bloc vote. This exchange was strictly one way, of course. Had these voters stayed with the Conservative party, the Tory share would have been about three points larger in Table 3’s accounting, or over 20 percent among active voters, rather than the 16 percent the party actually received.

Now to Reform. If the party did make net gains from demographic turnover and from exchange with abstention, as a new party it could hardly do otherwise. But just over five percent of newly eligible voters supported Reform, a much smaller share than Reform’s share in the full electorate. About seven percent of 1988 abstainers voted Reform, also under the whole-electorate share. Altogether, new voters accounted for only about one-sixth the Reform total.

The prime source of Reform support was the Conservative party. A Conservative was twice as likely as any other 1988 partisan to move to Reform. In part, this reflects the fact that 1988 Conservatives were highly likely to defect, somewhere, anywhere. But 1988 New Democrats were even more likely to leave their party, yet their rate of movement to Reform in particular was nowhere near as great. It might seem significant that New Democrats were slightly more likely than Liberals to go to Reform, but this truly is an artifact. The critical thing is that New Democrats, like Tories, were very likely to defect. Among defectors, former New Democrats were the least likely to vote Reform: 14 percent of defecting New Democrats went to the new party, as compared with 20 percent from the Liberals and 30 percent from the Conservatives.49 Ideological proximity was clearly the dominant fact, then, and proximity — more properly, distance — kept the overwhelming majority of New Democrats out of the Reform camp. This is all the more striking in light of the fact that NDP voters, disproportionately concentrated in the west, were more likely to encounter viable Reform candidates than almost any other kind of 1988 partisan was. Reformers, thus, are mainly former Conservatives.50

As the Tory vote shattered, the old system shrank. The same is true on a smaller scale for the NDP vote. But a counter-trend also burns through Table 3. Some of the fragmentation of the old Tory and NDP vote contributed to consolidation of the old system’s remains. A 1988 Tory was almost as likely to switch to the Liberals as to Reform, and twice as likely to go Liberal as to go Bloc.51 For 1988 New Democrats the Liberals were by far the biggest draw, such that almost as many voted Liberal as stayed with the NDP.52 That so many Liberals are former Conservatives and New Democrats must be kept in mind when discussion turns, below, to the future of the system.

Fragmentation, Consolidation, and Policy

From Table 3 it should be obvious why the “United Alternative” has so much strategic appeal on the political right. As most Reformers are former Tories, it seems natural to patch up the quarrel. But Table 1 gives one hint of the barriers to this. Notwithstanding old affinities, present-day Reformers dislike the Conservative party intensely, even more than they do the Liberals. The Conservative-Liberal disliking gap may have narrowed between 1993 and 1997, but the Reform- Conservative gap did not. And the feeling is mutual. Conservative voters found Reform much less acceptable than they did the Liberals, and between 1993 and 1997 the gap only widened. In 1997, Tories even preferred the NDP to Reform. This is not a landscape that favours consolidation.

Figure 9 stylizes the policy differences that underpin this enmity. The figure is based on questions in the 1988 and 1997 Canadian Election Studies that ask respondents first what they see as ideal policy in a domain, and then where they think each party stands. For example, respondents were asked how much they thought should be done for Quebec, with alternatives ranging from “much more” to “much less.” They then were asked how much each party wants to do, with the same response alternatives. Response to the Quebec questions defines the horizontal axis in each panel.53 The vertical axis captures each year’s central economic issue. For 1988, this is the closeness of Canada-US ties, prompted by the Canada-US Free Trade Agreement. In 1997, the issue is taxing and spending, welfare-state spending in particular. Points on the graph indicate average positions for respondents in each region (squares) and for their perceptions of parties (crosses), simultaneously on both dimensions. The panels differ subtly, so that each can bring out a different point. In Panel A, regional differences outside Quebec are glossed over. Instead respondents in and out of Quebec are distinguished by party, the better to dramatize the old system’s incoherence and, thus, its explosive potential. In Panel B, respondents are distinguished not by party, but more finely by region to make a point about Reform, in particular.

For 1988, the horizontal axis captures a central fact of the history that produced that decade’s party system: on Quebec, all three old parties adopted essentially the same position. This is how respondents saw the parties, and this is how the parties actually behaved, notably as all three endorsed the Meech Lake Accord. Given that the Accord had the express purpose of making an earlier constitutional settlement acceptable to Quebec, it makes sense that respondents saw all three parties as leaning mildly toward that province.54 For the Conservatives, commitment to the Accord was the final step in the process captured by Figures 6 and 8. Roughly the same was true of the NDP, for whom a credible commitment to Quebec was essential to breaking into the inner circle. But this consensus left the system ripe for invasion, with Reform already on the horizon.

Most vulnerable to invasion was the Conservative party, for the horizontal distance between the party and its own supporters could hardly have been greater. Figure 9 reflects the peculiar incoherence of the electoral coalition assembled by Brian Mulroney in 1984. But then, all parties contained some of this incoherence. Certainly, the NDP was almost as vulnerable as the Conservatives were. The Liberals too were divided along this line, although the distance between its camps was relatively small.

But the 1988 election did not turn on the Quebec- Canada dimension. Reform might have primed the issue but the party was still very new and, in any case, gaps among the old parties were huge on the other, unquestionably important issue, Canada-US relations. Party by party, Quebec and non-Quebec supporters were almost indistinguishable, and for the Conservatives and NDP, closely coincident with their own party’s position. By staking out this pole, the Conservative Party made the cost of right-wing defection prohibitive. The NDP might have done the same but for the ability of John Turner’s Liberals to close on them.55 For the parties as a system, the centrality of free trade arguably delayed a reckoning.

The reckoning arrived in 1993. The Bloc’s emergence on the far pro-Quebec side of the spectrum is, obviously, a critical part of the story. It burst onto the scene precisely as a reflection of the collapse of the Meech Lake Accord. Its mere existence, tying down close to half the Quebec electorate, narrows the field for pan-Canadian coalition building. Reform, of course, anchors the other pole on the Quebec dimension, as it also does for the spending dimension. On the Quebec issue, Reform in 1997 sits right where the typical Tory outside Quebec sat in 1988. On the same issue, Reform in 1997 may also be closer to the typical 1988 New Democrat than the NDP itself was that year.56 Indeed, Reform seems ideally located with respect to all respondents outside Quebec in 1997. On the horizontal dimension of Panel B, Reform is right in the thick of the various regions, with only the NDP as a serious rival.

But concentration on the Quebec issue misstates a key part of Reform’s appeal. Panel B makes clear that the West, Reform’s geographic base, is not peculiarly impatient with Quebec. British Columbia is indistinguishable from Ontario, as only the limitations of graphical presentation force the two apart. The Prairie provinces are the polar region, to be sure, but among the Prairie provinces, Alberta is not the polar one; rather it is the most like BC. Yet Alberta and BC are the heartland of Reform. Although within each province (Ontario no less than the West), Reform supporters are clearly distinct from Liberals and Tories, Western voters with a given orientation to Quebec are much more likely to vote Reform than a voter in Ontario, not to mention in Atlantic Canada, with the same orientation. Part of Reform’s appeal, we must conclude, is almost purely sectional. To many voters its appeal is as the party of the West, the West that “wants in.”57 This is another way in which Reform is — or rather, was — the Bloc’s mirror image. Whether or not the Bloc is irrevocably committed to full sovereignty for Quebec, its mere existence represents a repudiation of pan-Canadian schemes of linguistic accommodation. Quebec’s peculiar claim may still be as the francophone heartland, but the fact now remains that what once may have been linguistic demands are now couched in the rhetoric of jurisdiction and place. And if the Bloc is the party of one region, Reform, in the minds of many of its supporters, was the party of another region. For some Reform supporters, this may have been the essential thing.

Where Do We Go From Here?

Within the existing rules of the game, we need to consider two paths for the system’s further evolution. One is consolidation of the Right, the objective of Reform’s self-immolation earlier this year and Phoenix-like reemergence as the Alliance. We should not dismiss the possibility that pursuit of this alternative might only fragment the Right even more. The other path concerns the Left. As the popular vote on this side of the spectrum coalesced only in the 1990s, it is reasonable to ask if this consolidation can be maintained. If the system cannot move back toward effective two-party politics, the precondition for the distinctive virtues of the FPP electoral formula, then the question of structural reform is squarely on the table.

Consolidation of the Right?

The explosion of the Conservative party’s electoral coalition did not alter the geographic logic that preoccupied Tory strategists into the 1980s. Figure 9 makes clear that such a strategy does not require abandonment of basic conservative economic principles. On some economic questions, indeed, a non-centrist strategy is still probably the winning one.

The problem is that economic ideology was not the only thing, not even the most important thing, driving the votes of the Conservative party’s erstwhile base. Bear in mind, first, that some of the old Tory coalition fell to earth as the Bloc. Sovereignists mobilized into federal politics by the Bloc probably have little interest in renovating the rest of the federal party system. Quebec fugitives from the Conservative party were not attracted to that party in the first place by its economic ideology. Nor are they likely to be attracted by a decentralist conception of the federation that does not also involve recognition of Quebec’s special place. Right off the top, then, a consolidation strategy bearing the earmarks of Reform must write off a significant fraction of the original Conservative base.

That said, Reform’s base is one obvious place to start rebuilding the Right. Reform’s great strength was its geographic focus, and this bastion yields a significant parliamentary presence. At the same time, Preston Manning’s bold move to dissolve Reform into the Alliance recognizes that Reform’s weakness is the flip side of its strength: it is seen as too regional. Unfortunately, in attempting to dissolve his own party and yet lead the new, more inclusive one, Mr Manning turns the lesson of history upside down. History shows that parties have effected dramatic interregional moves. Figures 6 and 8 document two in the Conservative party’s recent history. A third example also stands out from the historical record, the move by the Liberal party into Quebec in the late 19th century. Each move involved an existing party dramatically raising its credibility in a region of erstwhile weakness by choosing a new leader from the target region, Laurier in 1887, Diefenbaker in 1956, and Mulroney in 1983. This is precisely the opposite action to Mr Manning’s.

If not Preston Manning, then who? It should be someone quite unlike Preston Manning, preferably from outside the West. As a distinctive feature of Reform was its hyper-concentration on the person of the leader, to the exclusion even of rivalries from provincial wings, it cannot realistically be someone else from the former Reform party. An obvious place to look is the Conservative party, possibly a provincial wing. The person would have to be a Conservative willing to gamble on leaving the ancestral party, at least until it is so beaten down that its remains seek union on post-Reform terms. This logic precisely underpinned the candidacy of Tom Long from Ontario. But for the analogy to be complete, Westerners would have to have been critical and visible in choosing the non- Westerner. Mere capture of the formerly Western party by outside forces would defeat the point. As it turns out, Stockwell Day’s leadership victory has foreclosed this scenario.

The Conservative party faces a subtly different imperative than does Reform. Although the Conservatives do need to diversify the geography of their appeal, their appeal is already quite diverse, and this fact remains the Tories’ great strategic promise. To realise the promise, they need not so much to widen the appeal as deepen it. But what does this mean? It might mean a more fullblooded economic and fiscal conservatism, the sort telegraphed by Jean Charest at the start of the 1997 campaign, then seemingly abandoned.

But such appeals might alienate one of the key potential pillars of a renewed Conservative party, soft nationalists in Quebec, whose presence in the Tory caucus, 1984-93, pulled it to the left. This potential cannot be written off lightly. We have noted already that had the Conservative party been able to hold these voters in 1993, it would probably have outpolled Reform. Few seats would accrue, but the votes would still be important in the field of moral claims. No less important, these votes, even as potential, figure in the larger credibility game. Quebec may not be the pivot for government that it once was, but its seats remain critical. Weak though they are, the Conservatives still are the only party other than the Liberals with “binational” potential. This must be part of why, according to Table 1, the Tories are most Liberals’ second choice. Of course, many of those Liberals were Tories.

Inside Quebec, though, the Tories are boxed in between hard federalists and hard sovereignists, each with a geographic bastion. Quebec Tories mimic the party’s strategic predicament for the country as a whole: appeal that is wide but not deep. Deepening the appeal might require a quite full-blooded nationalist position on Quebec- Canada relations. Such an appeal is unlikely to wash outside Quebec, however, precisely as so many former non-Quebec Tories who all along repudiated national-unity politics still have a party close to them on this issue.

And neither a Quebec-focussed appeal nor an artfully pan-regional one is likely to attract those westerners for whom Reform’s peculiar appeal was sectional. This is bound to be a problem for the Alliance. To an important degree, Reform was latter-day Social Credit, and a post-Reform rump born in reaction to the Alliance, need not govern to survive. Social Credit was never more than an Alberta, later Alberta-plus-interior-BC, party, yet it was stronger as a parliamentary presence than the CCF right down to 1958. This testifies to a logic that could sustain a post-Reform rump. If Social Credit, or the CCF/NDP for that matter, could think of themselves as affecting policy by their mere presence, why could a Reform survivor not do the same? Social Credit did disappear, of course, but the reasons are telling. Although its former supporters were absorbed into the regionally inclusive Conservative party, it is not obvious that they joined because the Conservative party was regionally inclusive. Rather they joined a party that, in the person of John Diefenbaker, crafted a self-consciously western appeal, and the final collapse of western Social Credit into the larger party occurred after the Tories lost power. That John Diefenbaker bequeathed his successors a party poised closer to the brink of power than it had been for years may have made its western supporters feel good about that party. But the final compromises necessary actually to deliver power were too readily portrayed as a betrayal.

In sum, the landscape on the Right seems fraught with incompatibilities. A strategy to deepen the Tories’ appeal in Quebec may founder in that province even as it alienates still more voters elsewhere. The Alliance’s broadening out from the Reform base still has no resonance in Quebec and yet risks undermining whatever the Alliance inherits of Reform’s specifically sectional appeal. The sectionalism of the Bloc and Reform should remind us that the landscape in question, the former Tory vote, is not just something that can be summed up as the “Right,” differentiated only by gradations of conservatism. Meanwhile, contenders for the turf seem likely to carry on like chain stores or gasoline companies engaging in “price wars,” except in this case the local chains may have surprisingly deep pockets.

Fragmentation of the Left?

As Table 3 shows, a significant fraction of the old NDP base enlisted in Liberal ranks. Why this happens remains to be analysed. Some of the story probably involves the discrediting of NDP claims to ideological distinctiveness by the bitter experience of power in the recession- and deficit-ridden early 1990s. Some probably was strategic, motivated by a desire to smite the Tories. The two motives were probably complementary. But in each motive lie the seeds of possible deconsolidation.

If the point of consolidation was to gain power for a broadly acceptable centre-left formation, just how far left the Liberals lean should matter to their supporters on the NDP flank. Indeed, the 1997 election was a warning signal from this very place. In 1993, such signalling would have seemed an undisciplined self-indulgence, given the need to drive the Conservatives from power. Is it now necessary to consolidate to keep them, or some other incarnation of the Right, from power? To the extent that Stockwell Day or someone else succeeds in uniting the Right, the Liberals can still stake a claim from strategic necessity. If, as I suspect, union of the Right is next to impossible, then union of the Left is not so manifestly necessary.

As always, much depends on geography. If fragmentation mainly transfers seats from the Liberals to the NDP, then Liberal majority governments will simply give way to Liberal minority ones. There is no reason to think that coalition government will become the norm, as the Westminster system perfectly embodies a logic that makes all players prefer single-party minorities to multiparty majorities.58 Each party can reasonably hope that the next election will decisively improve its fortunes; that is the promise embodied in Table 2. At the same time, leaders who know their electoral history — littered with the wreckage of parties that entered formal coalitions — rightly fear too close an entanglement with other parties. If minority government becomes the norm, the Liberals need not be utterly dependent on the NDP for a working majority. So long as the opposition remains fragmented, a Liberal government will usually be able to choose its allies issue by issue.

The geography of fragmentation will not necessarily be so favourable. Far from boosting NDP numbers, fragmentation of the left vote may only convey seats to the Right, especially if the Right succeeds in consolidating its own vote. But then, the more credible the threat of a united Right, the more united the Left is likely to remain. In a sense, the Left has a luxury the Right does not. The Left does seem more coherent than the Right, in that the landscape spanned by the Liberals and NDP admits continuous, fine-grained distinctions. Boundaries on the dimension can move according to circumstances. NDP supporters remain, as Mackenzie King described their CCF predecessors, “Liberals in a hurry.”

More generally, the Canadian system seems to have reached a point where only the Liberals are truly feasible as a single-party government. The historical record suggests that that has been true for decades, perhaps since 1896. This makes Brian Mulroney’s feat of winning two majority victories and governing as if he meant it all the more stunning. But it was an act of hubris, and the punishment in the end was correspondingly brutal. To say that only the Liberals are feasible is not to say that a majority of Canadians always feels disfranchised. In fact, the Liberals exemplify the “Condorcet winner,” the option that defeats all comers in straight fights. Table 1 makes clear that the Liberals are the near-universal second choice, where they are not the first choice. If each alternative takes the Liberals on alone, it is bound to lose. In this sense, Liberal victories are not accidents of the electoral system. The system has distorting effects, but it does not as a rule yield the “wrong” winner.59

This fact about the Liberals also helps explain the fragmentation of the rest of the party system. For Canadian voters, unlike opposition party elites, the stakes in elections may have gone down. The rest of the system is not so fragmented that only Liberal candidates can win. But the fact that the Liberals are now the only party that can win means that most of the time the government will be broadly acceptable. Greater opposition vote fractionalization is now also acceptable, so long as its geography permits some translation of opposition votes into seats. The Liberals now stand revealed more starkly as what they probably were all along, Canada’s equivalent of India’s Congress Party, at least the Congress of the first 40 years after Indian independence.60

Parties like this can be defeated only by ends-against- the-middle strategies. Such strategies are intrinsically hard to carry off and so are rare. When they succeed, the resulting governments are unstable or their electoral bases, unsustainable. This explains the longevity of Liberal governments and the chequered aftermath of the occasional Tory bout in power. Now such bouts seem even more remote.

Structural Reform?

A Liberal government, even a minority one, indefinitely in power is an unhealthy prospect in a democracy. Given that the Liberals are broadly acceptable to many voters — at least as a second choice — we are unlikely to see riots in the streets soon. Passion may still erupt on the left and right of the system, as with Reform in 1993. But the government can probably absorb most initiatives in the realm of policy, where those initiatives have real backing or where they are connected to shoring up the margins of the Liberal coalition. In this sense, Reform has already had an impact in the palpable rightward shift of the whole spectrum in 1993. The 1997 election, aided by the end of deficit politics, nudged the government, if not the whole system, leftward.

But the lowering of the stakes and the infeasibility of alternatives, opposite sides of the same coin, may be a key reason for the apparent withdrawal from electoral politics signaled in Figure 4.61 As symptom or as the thing itself, demobilization of a significant fraction of the electorate should not be passed over lightly. Under the existing structure, the prospect for renewal in participation and enthusiasm seems as remote as the prospect for party alternation in power.

But that is the point. If for most groups, indirect impact on policy is better than nothing, actually holding office is still critical. If office did not matter, no party would form a government as a minority. If office matters, offices should circulate, and circulation is a missing element in party politics at the millennium. We may finally have reached the point where structural reform can no longer be staved off. If the content of these reforms is the subject for other papers, two things are strongly implied by the analysis in this one. First, proposed changes should promote circulation of parties through office, not necessarily as single-party governments. The weakness of the existing system, manifest long before the 1990s, was its peculiar manner of circulation: none for extended periods, then powerful retribution on the usual governing party, the Liberals, retribution which in the long run only made things worse for the shortrun beneficiary, the Conservative party. The system should probably be a more sensitive register of opinion change even as it reduces the likelihood of cataclysm. All this is to say that the system probably must become more proportional. How it does so is the second implication of this paper. Most of the time, the parliamentary parties will have no obvious stake in structural reform. Incumbents are usually, by definition, beneficiaries of the existing rules. Proposals for reform thus must be opportunistic. Some of this may be in timing, in searching for moments when some parties already enjoying footholds can move up under new rules. We can imagine, for example, an ends-against- the-middle strategy in electoral reform, where parties that can improve under new rules extract electoral reform as the price of supporting a minority government. Or vote-seat translations under FPP may become just too unpredictable even for Canada’s historically risk-acceptant parties. 62 Sometimes the old game can be sustained no longer, and the only way for old players to survive is to change the game.

Comment on Johnston, “Canadian Elections at the Millennium”*

Lisa Young

In “Canadian Elections at the Millennium,” Richard Johnston provides an accurate description of contemporary Canadian electoral politics. The trends he identifies — a Liberal Party strong in the House of Commons but weak (and apparently declining) in the electorate, a fragmented opposition, and declining voter participation rates — are indeed highly salient characteristics of the current state of affairs. Moreover, Johnston is correct to identify these trends as reason for concern. There is little or no prospect of the Liberal party being removed from government in the foreseeable future, and the highly fragmented opposition cannot provide the alternative to the government of the day that is required for healthy democratic competition. Moreover, signs of malaise in the electorate suggest a generalized discontent with the democratic system. One might add to this list of concerns the development of a highly regionalized pattern of political competition which has transformed election campaigns from national contests with national issues into a series of regional competitions, each with its own issues.1 That said, an argument could be made that the proliferation of parties and the greater ideological and organizational diversity of these parties has made the party system more reflective of the wide range of interests and opinions in the Canadian electorate.

The central task Johnston sets himself in this paper is to explain how the current situation arose. His explanation focuses on the role played by institutions — in particular the single member plurality electoral system – and the incentives it provides to parties. Of particular importance, according to Johnston, are the strategic choices that the Progressive Conservative Party made in the 1980s, which set the stage for the electoral earthquake in 1993. In arguing that the current system is a direct result of the Canadian electoral system and the strategic choices made by Conservatives, Johnston also asserts that the current state of affairs is uniquely Canadian and thus largely unrelated to changes in values and political behaviour observed in other industrialized countries.

While I accept the logic of Johnston’s argument regarding the incentives embedded in the electoral system, this does not provide a complete explanation for recent developments in Canadian politics. The Canadian electoral system has not changed since Confederation, but the outcome of the 1993 general election was unprecedented in Canadian history. What has changed is Canadian society and the structure of the party system that mediates between society and government. The societal inputs being filtered through the electoral system and shaping the party system have changed profoundly in recent years, and these changes explain at least a portion of the current situation. The breakdown of the pre-1993 party system was as much a function of the party system’s inability to accommodate a range of social groups and interests as it was a consequence of the electoral system and the flawed Mulroney coalition. In this respect, it represents a period of transition between party systems. As was the case with earlier transitions, the multiplication of cleavages within the Canadian electorate created pressure on the old party system, which eventually ruptured in 1993.

This assessment flows from the historical model of party system change set out by R.K. Carty and David E. Smith.2 Their account of Canadian party systems considers the constellation of parties, but also party organization, representational focus, and modes of political communication within each system. Carty argues that there have been three party systems in Canada to date. The first, which extended from Confederation to 1917, was highly localist in character and depended on patronage to hold parties together. Parties were based in the parliamentary caucus, which retained the power to select the leader. The second system, which spanned the years from 1921 to 1957, was characterized by a politics of brokerage among regions. Extra-parliamentary parties became more significant, but strong regionally based ministers exercised power within the party. The third party system, from 1963 on, was the era of pan-Canadian politics. It was, above all else, characterized by its pan-Canadian ethos. This was a politics “dedicated to creating a Canadian community, and it became the task of parties in the [old] party system to define a national agenda and to mobilize Canadians, as individual participating citizens, in support of their competing visions for the country.”3 Extra-parliamentary parties became more significant, but the advent of television and the professionalization of politics ensured that the party leader was the central figure in party affairs.

Unlike Johnston’s analysis, which focuses on electoral alignments and realignments, the Carty/Smith model concentrates on parties as organizations which act as intermediaries between society and the state. Changes to the party system, by this view, are largely a consequence of social change. Between the first and second, and the second and third party systems were periods of transition. Carty notes that these periods followed “considerable social and demographic changes in the basic structure of the electorate. Both…helped break old electoral alignments and patterns of political organization, making it easier for new systems of partisan mobilization to emerge.”4 In Rebuilding Canadian Party Politics, Carty, Cross and I argue that the pre-1993 party system was unsustainable in part because it failed to accommodate significant cleavages within Canadian society. To understand the breakdown of the old party system, we need to look beyond the institutional logic of the electoral system to take into account the social changes that created pressure on political parties in the years leading up to the 1993 election. The two most significant sources of pressure were the multiplication of cleavages and the rising cynicism of the electorate.

Multiplication of Cleavages

In contemporary liberal democracies, political parties play a crucial role as intermediaries between society and the state. This role requires that parties be responsive to the views of the electorate, that the full range of societal interests be given a voice in the political arena, and that salient political identities be represented. In the Canadian experience, this has been complicated by the parties’ pattern of trying to straddle fundamental cleavages, rather than taking opposing sides.5 Each period of transition in the Canadian party system has been preceded by a multiplication of cleavages, as new interests have emerged (through Western expansion and immigration, for example) and existing groups (such as farmers or women) have become politicized. Parties have tried to accommodate these newly mobilized groups without alienating other supporters, but have for the most part failed to fully achieve this ambitious objective. This has contributed to the breakdown of coalitions within parties, usually resulting in the formation of new parties which, in turn, spark transition in the party system.

The period from 1970 until the mid-1990s was one of profound social change in Canada as in other advanced industrialized democracies. The rise of new social movements — like the women’s movement and the nationalist movement in Quebec — created important new pressures on the political system. These pressures took the form of calls for inclusion, but also introduced new political forms and tactics. Social protest, demonstrations and grassroots social movement activism became more prominent and widely accepted within the Canadian political arena. Interest groups became an even more important form of political organization. The social change that took place was not universally welcomed. These new social movements engendered negative reactions, which also placed representational demands on the political system.

The Quiet Revolution in Quebec had profound consequences for the politics of both Quebec and Canada. This period gave rise to a distinctive brand of federalism in Quebec, and equally produced the sovereigntist movement in opposition to it. With the ascendance of the sovereigntist movement’s partisan manifestation, the Parti Québécois, and its 1976 election victory , Quebec politics were transformed. Even though the federalist/sovereigntist cleavage has shaped Quebec politics for three decades, it was only with the formation of the Bloc Québécois in 1991 that sovereigntist voters were offered a political vehicle in the federal arena.6

Important repercussions were felt outside Quebec too. Decades of conflict over the Constitution and Quebec’s role in the federation drove all the major parties to adopt essentially similar positions on the national/constitutional question in an attempt to deflate the electoral power of the bicultural cleavage. This left substantial segments of the electorate without representation within the party system: Canadians whose primary political orientations were regional, neo-conservatives who argued for a substantial devolution of power from the national government, English-Canadians whose belief in the equality of the provinces caused them to reject any sort of recognition of Quebec’s distinctiveness — all found themselves without a voice in Parliament. The primary beneficiaries of this was the Reform Party.

Quebec’s Quiet Revolution was not the only social change of significance for the Canadian party system. From the 1970s on, the representational demands on the party system multiplied to include calls for representation from women and recent communities of immigrants.7 In the face of these calls for inclusion, all three parties took similar approaches: they tried to appeal to new groups of voters without alienating traditional supporters of the party. This followed the distinctive approach taken by Canadian parties in the face of profound ethno-linguistic and religious cleavages: to try to straddle the cleavages rather than taking one side.

This accommodative strategy was ultimately unsuccessful. Neither feminist nor minority groups were fully satisfied with the representative gains they won. At the same time, Canadians who opposed the idea of representational quotas, liberal feminist policy stances and government promotion of multiculturalism were left without a partisan home by the mid-1980s. In this sense, the three parties’ efforts to accommodate significant internal cleavages effectively drove activists at either end of the political spectrum out of the partisan arena. Again, the primary beneficiary in English Canada was Reform.

In essence, by the early 1990s, the three parties that comprised the party system found themselves unable to perform their crucial role as intermediaries between society and the state. They were too similar in their approaches to representation and in their stances on the national question. The electorate’s discontent erupted both during the 1992 Referendum on the Charlottetown Accord and in the 1993 federal election.

The Politics of Discontent

Whether it is a uniquely Canadian phenomenon, or part of a crossnational trend, cynicism regarding politics must be taken into account when coming to terms with contemporary party politics in Canada. As Carty et al. note, “One of the characteristics of the earlier periods of party system transition was a change in the norms of party democracy. In both of the earlier periods, voters expressed displeasure with the existing party system in terms of both the responsiveness of the parties to the concerns of voters and the participatory opportunities afforded voters in intraparty decision-making.”8

There is considerable evidence that the Canadian populace has grown dissatisfied with the political process in recent decades. One element of this is a resurgence of populism which has once again become an important element of Canadian political culture, and which extends beyond Western Canada. In a 1990 study of Canadians’ attitudes toward the electoral and political process, Blais and Gidengil found that a majority of Canadians expressed attitudes that could be considered populist.9 Data from the 1997 Canadian Election Study demonstrate that little has changed in the interim: a majority of respondents demonstrated anti-intellectual attitudes, and two-thirds believed that national problems could be solved if political decisions were brought to the grassroots.10 These populist sentiments are not concentrated in Western Canada; rather, they are shared by a majority of Canadians in all regions.

Coincident with the rise of populism has been growing discontent with government in general and political parties and Parliament in particular. The decline of public confidence in political parties is particularly striking: In 1979, 30 percent of Canadians polled by Gallup had a great deal or quite a lot of confidence in political parties; by 1989 the figure had dropped to 18 percent and by 1994 it had further plummeted to a mere nine percent.11 Politicians fare even worse than political parties in public opinion: Data from the 1997 CES show that 35 percent of respondents believed that political parties didn’t care what people thought, and 41 percent believed that MPs did not know what people thought. Over 80 percent of respondents agreed that MPs lose touch with their constituents after being elected.12 In their analysis of voting behaviour in 1997, Nevitte et al. found that cynicism13 was the most powerful dimension structuring the orientation of citizens outside Quebec, and that “nearly all voters are distributed toward the right [i.e. cynical] end of the scale, indicating that most voters were quite cynical about the political classes and the responsiveness of government.”14

The growth of cynicism appears to be part of a cross-national trend in established democracies. There is a discernible trend in these countries in recent decades as citizens have become increasingly critical of the major institutions of representative government.15 Although its shape and magnitude vary (as Johnston points out), the basic trend is undeniable, and is particularly evident in the US.16 In this sense, the current state of affairs appears less uniquely Canadian than Johnston’s characterization suggests.

The evidence that a substantial proportion of the Canadian (and particularly English Canadian) population holds views that may be termed cynical is convincing. The next question, then, is how these attitudes are translated into action: do voters exit the system (by not voting) or do they seek a voice for their views within the political arena? In the English Canadian case, it appears that the latter is by far the more common behavioural manifestation of political cynicism. According to Nevitte et al., cynicism accounts for only a small portion of the phenomenon of nonvoting in 1997.17

Rather than exercising the option of exiting the system, cynical voters have opted to express their discontent within the confines of the electoral arena. In both the elections of 1993 and 1997, political cynicism frequently translated into a vote for the Reform party. As Nevitte et al. note in their analysis of the determinants of the vote in 1997, Reform is “a powerful lightening rod for both anti- Quebec sentiment and generalized political disaffection.” 18 While Canadian voters in general became somewhat less cynical between 1993 and 1997, the level of cynicism expressed by Reform party voters remained constant.19 Although cynicism was not the only attitudinal characteristic that strongly predicted a Reform vote, it was certainly significant.

It is noteworthy that the predominant behavioural manifestation of cynicism was to support a new party, rather than to opt out of the system. As Carty et al. note,

…it is a curious fact that when Canadians get really angry about national politics and the accommodations it demands, dissatisfied with public policy, or disillusioned with their governments, and decide to do something about it, their instinctive response is to start by attacking the party system.…For Canadians… relegitimating the national community or reshaping their social contract means rebuilding national political parties.

For all that there is cause for concern regarding the shape of current electoral politics, this at least is a reason for some optimism. It suggests that the party system is sufficiently flexible that it can respond to the demands of the electorate, and that political cynicism has not produced a widespread disavowal of partisan democratic politics.

The Current State of Affairs

This, finally, brings us to the question of whether we should be particularly concerned about the current state of party politics in Canada. Johnston’s claim, clearly, is that we should. As he notes, there is no credible alternative to the Liberals, whose parliamentary majority is based on a preposterously small plurality of the popular vote. Without contesting the validity of this assertion, it is worth noting that there are also reasons for optimism. First, it is noteworthy that political discontent has resulted not in widespread exit from the democratic process, but rather in support for change within the party system. Second, it is equally significant that the electoral and party system proved sufficiently flexible to accommodate the social and political pressures that led to the formation of the BQ and Reform Party. Third, it is possible to argue that the breakdown of the old party system and the emergence of the new has reinvigorated Canadian partisan politics. For the first time in decades, federal political parties are taking profoundly different stances on the fundamental national question. In contrast to the pan-Canadian consensus of the third party system, the parties that comprise the emerging system take vastly different views on the question of centralization versus decentralization and the appropriate role of Quebec within Confederation. All of a sudden, the stakes are much higher in national electoral politics. If we take the view that one of the fundamental roles parties should play is to reflect societal concerns, then the parties that constitute the current system are, arguably, doing a good job.

Where the current system falls short is with respect to the aggregative/accommodative function of politics. If the most important aggregative function of political parties in Canada is integration of various regions, then none of the current parties is fully succeeding, although some are certainly trying. Moreover, as Johnston rightly points out, the very logic of the Canadian electoral system pushes parties that aspire to govern in the direction of regional accommodation. In this respect, we are witnessing a transformation of the Reform Party / Canadian Alliance away from the representative function discussed above toward the accommodative function as it seeks to broaden its base of support and form a government.  There are, of course, hurdles in the way. Most notable among them is a complete lack of support for the Canadian Alliance east of the Ottawa River. In this respect, Johnston is entirely correct to warn us that we lack a party that can credibly claim to offer an alternative to the governing Liberals. Would a different electoral system alleviate this difficulty? Perhaps. Electoral systems based on proportional representation have considerable merit in themselves, and would probably ease regional distortions in electoral support. That said, a PR system would do little to limit the fragmentation of the Canadian party system. Whether some variant of PR is adopted or not, the current constellation of parties almost certainly means that coalitions among parties will have to be created to form future governments. In short, PR may not (whatever its other merits) be a panacea for the ailments that plague the Canadian body politic. If we accept that the breakdown of the old party system was due, at least in part, to the parties’ inability to accommodate and reflect the diversity of Canadian society, and that rising cynicism among segments of the electorate played a crucial role in this breakdown, then proposed reforms should be evaluated with a view to determining the extent to which they strengthen parties’ responsiveness and accommodative capacities.


* Portions of the following rely heavily on R. Kenneth Carty, William Cross and Lisa Young, Rebuilding Canadian Party Politics (Vancouver: UBC Press, 2000). Carty and Cross are not responsible for the analysis presented here, but deserve much of the credit for the analysis of the emerging party system.

  1. See Carty, Cross and Young, Rebuilding Canadian Party Politics.
  2. David Smith, “Party Government, Representation and National Integration in Canada,” in P. Aucoin (ed.), Party Government and Regional Representation in Canada (Toronto: University of Toronto Press, 1985); R.K. Carty, “Three Canadian Party Systems: An Interpretation of the Development of National Politics,” in H. Thorburn (ed.) Party Politics in Canada, 7th ed. (Scarborough: Prentice Hall, 1996).
  3. Carty et al., Rebuilding Canadian Party Politics, p. 21.
  4. Carty et al., Rebuilding Canadian Party Politics, p. 142.
  5. David Elkins, “Parties as National Institutions: A Comparative Study,” in Herman Bakvis (ed.), Representation, Integration and Political Parties in Canada (Toronto: Dundurn, 1991), pp. 12-13.
  6. A possible exception to this statement would be the nationalist Social Credit party which elected several MPs in the 1970s.
  7. For a more extensive discussion of this, see Carty et al., Rebuilding Canadian Party Politics, pp. 88-91.
  8. R.K. Carty, William Cross and Lisa Young, “The Fourth Canadian Party System,” in William Cross (ed.) Canadian Democracy at Century’s End (Toronto: Oxford University Press, forthcoming).
  9. André Blais and Elisabeth Gidengil, Making Representative Democracy Work: The Views of Canadians (Toronto: Dundurn/ RCERPF, 1991), p. 19. They reported that 65 percent of respondents exhibited anti-intellectualism, and 74 percent favoured bringing decisions closer to the grassroots.
  10. Lisa Young, “Value Clash: Parliament and Citizens after 150 Years of Responsible Government,” in F. Leslie Seidle and Louis Massicotte (eds.), Taking Stock of 150 Years of Responsible Government in Canada (Ottawa: Canadian Study of Parliament Group, 1999).
  11. Young, “Value Clash,” p. 119.
  12. Young, “Value Clash,” p. 119-20.
  13. Nevitte et al. employ the following items in their measure of political cynicism:
    • Politicians are willing to say anything to get elected
    • Politicians are ready to lie to get elected
    • Do political parties keep their election promises most of the time, some of the time, or hardly ever?
    • I don’t think the government cares much what people like me think
    • Those elected to Parliament soon lose touch with the people.
  14. Neil Nevitte, André Blais, Elisabeth Gidengil and Richard Nadeau, Unsteady State: The 1997 Canadian Federal Election (Toronto: Oxford University Press, 2000), p. 48. The authors indicate there is some evidence that cynicism declined somewhat between 1993 and 1997 (p. 54).
  15. Pippa Norris, “Introduction: The Growth of Critical Citizens,” in Pippa Norris (ed.), Critical Citizens: Global Support for Democratic Governance (Oxford: Oxford University Press, 1999), p. 26.
  16. Russell J. Dalton, “Political Support in Advanced Industrial Democracies,” in Pippa Norris (ed.) Critical Citizens: Global Support for Democratic Governance (Oxford: Oxford University Press, 1999).
  17. Nevitte et al., Unsteady State, p. 64. This analysis of nonvoting finds that least likely to vote are individuals with low incomes and low levels of education, as well as young people. This raises the issue of the class dimension of nonvoting. It is perhaps not coincidence that non-voting has increased at roughly the same time as income disparities have also increased.
  18. Nevitte et al., Unsteady State, p. 92.
  19. Nevitte et al., Unsteady State.

Comment on Johnston, “Canadian Elections at the Millennium”

Stephen J. Harper

If post-1993 Canada is marked by a weak government entrenched in power by a fragmented opposition, then this is undeniably a problem worth solving in any democracy. In this regard, Richard Johnston’s analysis of federal party politics corresponds with my own writings on the party system over the past few years and I cannot disagree substantially with him. On the other hand, Johnston’s speculations on the future and his range of solutions is rather incomplete. What is the underlying nature of the Canadian party system and how does it fit into Canada’s model of government? Today’s federal multi-partyism is a serious dysfunction within Canada’s modified Westminster parliamentary model. This requires a strong national government challenged at all times by a united opposition. For the proper functioning of this system, either a two-party balance must evolve, or some type of electoral, institutional or even constitutional measures will have to be considered. There are four scenarios that might be envisaged.

Scenario #1: “Self-Correcting Tendencies”

According to one view, there is nothing peculiar to Canada that will prevent the party system from evolving of its own accord back towards stable and sustainable two-party competition in reasonably short order. The normal course of competitive partisanship will reveal what Johnston refers to as “self-correcting tendencies.” The current fragmentation of Canadian politics is thus a transitory phenomenon, originating from some combination of postindustrial, post-Cold War shifts in the left-right axis, and the personalities and issues of the 1988- 1993 period.

While Johnston is clearly dubious about this perspective, it is the view shared by most of the political parties themselves. Reform, through the Canadian Alliance, is attempting to quickly assemble a singular competitor to the “left-wing Liberals.” The NDP continues its perpetual quest to polarize debate against the “right-wing Liberals.” And the Joe Clark PCs seek restoration of themselves as a centrist alternative to the “incompetent Liberals.” Of course, the Liberals themselves see no particular problem that requires systemic change. Only the Bloc Québécois fundamentally rejects this genre of solution.

While all the federalist opposition parties cannot be right, one of them clearly could be, and the one clearly having the most success to date is the Canadian Alliance. Even since Johnston wrote his paper the CA has achieved some of the prerequisites he outlined for opposition consolidation. Through the United Alternative process, the CA managed to steal a significant chunk of the Progressive Conservative party out from under Joe Clark. That portion came largely from the provincial “right wing” parties, including the electorally critical PC party of Ontario. It includes a Quebec element, drawn largely from the Action Démocratique. And the PC party of Alberta has furnished the CA with its new leader, Stockwell Day.

“Stock” is both more dynamic and more conventional a political personality than Preston Manning, yet clearly acceptable to (and largely elected by) the base of the Reform movement. While, unlike failed contender Tom Long, he is not from Ontario, this should not automatically preclude Day making gains in the target region of the Alliance. Pierre Trudeau proved to be the right choice for the target voters of the Liberals in 1968 even though he came from one of the party’s electoral strongholds. And today no other federal leader is from Ontario anyway.

Johnston would no doubt point out that the Alliance under Day barely dents the francophone part of Canada’s “binational” political equation. Frankly, this is not fatal. The CA may not be competitive in Quebec, but between the party’s marginally extended Quebec presence and the leader’s passable bilingualism, it may prove to be at least legitimate in Quebec. This is all that really matters to the critical voters in Ontario. The realpolitik is that Quebec is not arithmetically necessary to forming government in Canada — today less so than ever — and has taken itself out of the government game by voting strongly for the Bloc.

For the Alliance, a strongly competitive party west of the Ottawa River is as viable a government coalition as the Liberal one extending from Westmount to Winnipeg. Besides, the very presence of a second party eligible for power could shake the Québécois away from their strategically dubious BQ voting and toward governing choices, thereby further consolidating a two-party system.

The real problem with the scenario, however, is that we are in fact nowhere near a bipolar political system in Canada today. Opinion polls still put the Alliance roughly 20 points behind the Liberals. It is a measure of Liberal dominance that such a distance is now considered a sign of competitiveness. Even optimistically the Alliance is far closer to dead last than it is to the government. The best we can say is the Alliance may be consolidating the Right, and this may portend a two-party race two elections hence.

Furthermore, there is a fairly obvious explanation for the perennial optimism of the opposition parties — survival demands it. Within the context of the first-past-the-post electoral system, belief in a two-party system is almost an institutional necessity. Not only does FPTP tend toward twoparty races, but it produces powerful punishments for any party appearing to be a “vote splitter.” The maximum of combative partisanship is almost always the necessity of the day and any talk of electoral reform or political coalition is simply perceived as weakness in the face of FPTP’s “winner take all” logic. Ironically, this is why Reform’s United Alternative worked to the extent it did. It offered carefully selected PCs an opportunity for honourable abandonment of their party, rather than promising the official PC leadership a coalition enterprise.

A deeper problem with the opposition view is its ahistorical nature. Whatever differing interpretations Tories and Reformers put on the current fragmentation of the opposition, they agree that it was due to events of a decade ago. As Johnston notes, however, history strongly suggests that the current situation is recurrent and durable. His statement — “only the Liberals are truly feasible as a single-party government. The historical record suggests that that has been true for decades, perhaps since 1896” — probably comes closest to the truth. It is indisputable that the Liberals have governed the country for an extraordinary part of the past century by the measure of any liberal democracy. Even more significant is that all the alternative majorities — of Borden, Bennett, Diefenbaker and Mulroney — were not only comparatively short-lived, but ended in a meltdown with long-term implications.

This recurrence of partisan trends goes beyond government and official opposition to the very nature of Canadian multipartyism itself. The “third parties” of the 1990s — the New Democratic Party, the Reform Party, and the Bloc Québécois — bear more than accidental resemblance to the Cooperative Commonwealth Federation, the Social Credit Party, and the Créditistes or Bloc Populaire of earlier decades. Such patterns can be traced as far back as the disintegration of the country’s founding Tory government and entities such as the Patrons of Industry, the Equal Rights Association, and the Parti National.

Scenario #2: Electoral Reform

If Canadian multipartyism is recurrent rather than temporary, then it may reflect the genuine nature of our party system, obscured by elements of a constitutional regime that demands two-party competition. To accommodate such tendencies in the Canadian electorate some, including Johnston apparently, make a strong case for a new electoral system. After all, multipartyism is not, in and of itself, a liability except under a parliamentary government with first-past-the-post election rules.

I need not repeat all the arguments against FPTP here, since Johnston has largely made them in his paper, but I would like to summarize some observations. A simple plurality system like Canada’s does not necessarily produce a two-party system and penalize small parties, as is often suggested. It only tends to force a two-party system at the constituency level and punish small parties without a territorial base. Thus territorially-based small parties actually gain from FPTP. This means that FPTP can accommodate a multiparty system, but only if it can be translated into a series of regional two-party systems. If fragmented sufficiently, these can create a “one-party-plus” system like the one we have today.

The result, Johnston shows, is that a plurality system may do more than simply manufacture an artificial majority, which is arguably desirable. FPTP can also distort the distribution of seats in ways that can become large and party-specific. The effect can be legislatures that defy not merely proportionality, but ordinality as well. Johnston actually understates the case when he says that “the system has distorting effects, but it does not as a rule yield the ‘wrong’ winner.” The current British Columbia and Quebec governments are just such cases. Reform under Preston Manning actually had a well-considered strategy to win a majority government with a minimal popular vote — a tactic which has always been the separatists’ Quebec strategy.

Johnston is probably correct to surmise that electoral reform could happen quickly under the right circumstances although current conditions do not favour it. These would take shape in a highly divided minority Parliament. The Liberals would have to become convinced that their narrowing popular and geographic bases mean they will require coalition partners to govern in the future. Then, and only then, will the Liberals desire a voting system that makes coalition a reasonable option for parties like the PCs and NDP who, given their current struggles, are likely to embrace it. The Alliance also has much to gain from a more rational electoral system given its national aspirations. The CA of the future is more likely to be attracted to the advantages of a new system than to Manning’s brinkmanship strategies — strategies that fit so well into FPTP. Only the BQ has no obvious interest in reform under any scenario, and it is also the least desirable coalition partner for any federalist party.

We should not, however, presuppose the model which might result from a reform of the voting system in Canada. There are many different options but two broad types of alternative — proportional and preferential. The proportional systems, to which Johnston refers almost exclusively, provide for a guaranteed rationality to parliamentary representation, but usually do not manufacture a legislative majority. Preferential systems are more likely to provide such a majority at the expense of minority representation, but only because more input is given to voters and larger parties must attract the preferences of smaller blocs. Preferential systems have particular advantages in the Canadian context because of the country’s immense geography and history of single-member representation. They also represent a less radical departure from the mechanics of a plurality system than any proportional model.

The problem is that Johnston neglects to ask the central question about electoral reform. Obviously, a new voting system would end the worst pathologies of FPTP, but would Canada end up with strong governments and strong alternatives in the opposition? It is true that other electoral systems work well in most parts of the democratic world, but detractors rightly point out spectacular failures like Italy, Israel, or fourth republic France.

It must be remembered that, while FPTP voting is a characteristic of the Westminster model, the requirement for a strong government and a strong opposition is part of any type of parliamentary system. The continental model may not demand a strictly two-party system, but it does function best with a two-bloc alignment. Places like Sweden, Australia, Germany and other stable parliamentary democracies have such a pattern. Such multi-party systems usually break into two loosely defined blocs with two leading parties, plus a limited number of swing parties offering reasonably predictable governmental coalitions.

Any party system of this nature implies that political conflicts reduce to a key dimension given to gradations of compromise and accommodation. In unstable systems however, party conflict can rarely be reduced to a single axis or, if it can, its sides are not easily bridged. The marks of unstable systems are not just large extremist or anti-system parties, but also sectarian ones based on region, religion, ethnicity and personality. Such deep, cross-cutting cleavages in the population dramatically reduce the possibilities of brokerage and the assembling of viable coalitions.

Is it possible that this is the nature of Canada’s party system? There is more than passing evidence of this character. Canada’s parties, no matter what their philosophical differences, have always had unique regional, religious, and ethnic support bases. Johnston’s own analysis of the 1988 and 1993 elections indicate the potential for regional/communal divisions upsetting coalitions built around more manageable economic issues.

How, then, would a new electoral system actually affect popular voting patterns in Canada? It has been postulated as far back as Cairns that the electoral system has reinforced the territorial nature of partisan voting and that proportional representation would make parties more national in character. But is the opposite not also possible — that the genuinely sectarian nature of the Canadian electorate produces territorially-based parties, a tendency that could become even more pronounced with a new electoral system, especially in the absence of any majority party?

These risks would be present even in electoral systems that provided for very few parties. Suppose, for example, that Canada adopted a model frequently advocated, the German mixed-member proportional system. It tends toward a small number of large or medium-sized parties, with few splinter parties. Could not the resulting system consist principally of a large (but minority) Liberal party and a large (but minority) Alliance party, elected mainly from votes in English Canada, and a third party, the Bloc Québécois, elected exclusively from Quebec? In such a system, one option would be for one of the parties to take the BQ as its coalition partner, providing the other with a potentially devastating issue for the next national election. The other would be a “grand coalition,” depriving the English majority of any real opposition or any significant electoral choice. The resulting system would be perpetually unstable.

Scenario #3: Institutional Overhaul

There is thus a third and more complicated possibility — that the underlying nature of Canada’s politics is not merely multi-party, but also highly regional, communal and sectarian. In this case, electoral reform by itself will not bring about viable, alternating, governing coalitions (although it could be desirable for more limited reasons). In fact, these characteristics may explain why Canada’s two-party model has long been more myth than reality.

If the Canadian polity is indeed of this nature, then significant institutional change, including constitutional reform, will have to be introduced to create a more “normal” partisan politics. Canada is already a federal state, a significant departure from the pure Westminster model. The national state could be further decentralized into more governable component communities. Belgium is clearly on such a trajectory, where Switzerland has always been. This would reduce the possibilities for intractable partisan cleavage at the federal level.

Another possibility might be to address the limits of sectarian, multi-party coalition-building by creating a more powerful and independent executive, not dependent on the short-term whims of the legislative parties. Departing from a parliamentary system has worked in Switzerland and the United States. It also proved to be the solution for France in 1958 and may be where Israel is headed today. Direct democracy has also been proposed as a method of working around our fractious parties.

Finally, Senate reform is a hybrid strategy, where regional or other communal representation is enhanced in the central government, rendering extensive coalition-building more rather than less necessary as a way of addressing actual and potential divisions in the country. Of course, various combinations of these strategies are possible.

Those who advocate more extensive institutional reform in Canada have a big problem to address however — the improbability of bringing it about. Significant third party movements in Canada have always advanced variations of these proposals for systemic change. Yet virtually no institutional evolution has responded to them. Indeed, attempts to open such debates have often themselves become victims of the kinds of cleavages they were supposed to resolve. Arguably this is what occurred to the last Progressive Conservative government. The fragments into which it descended were clearly regional/communal, and the immediate cause was Brian Mulroney’s efforts to achieve constitutional reform to accommodate them (whatever the merits or otherwise of those proposals).

Scenario #4: The Dissolution of the Canadian Dominion

This brings me to the final possible solution — the dissolution of Canada as we know it. While this is currently beyond the frame of reference of most Canadian voters, it is the position of one significant federal party, the Bloc Québécois. In fact, the explicit position of the Quebec sovereignty movement is that no amount of conventional coalition-building, electoral reform or institutional restructuring will lead to a permanent, healthy party system at the federal level. Canada is simply too diverse and dysfunctional an aggregate to justify national political parties or national political agendas. As such, political transaction costs are too high to justify anything other than separate sovereignties.

In fairness, some evidence for this view is embedded in the radically divergent interpretations of the Canadian state of the current major political parties. The dominant Liberal view is that Canada is a single nation with a binational character expressed through central institutions. For the Canadian Alliance, Canada is fundamentally a federation of 10 provinces with important characteristics expressed only through lower levels of government. And, for the Bloc Québécois, Canada is simply an artificial construct of what are genuinely “two nations.” Not surprisingly, the strongest bastions of each party are in the bilingual corridors, unilingual anglophone and unilingual francophone regions of the country respectively.

Furthermore, the deep regional divisions of Canada go beyond Quebec and the current era. The country’s core economic policies were developed in the early years of Confederation, when the West was still essentially a colony. The development of national political coalitions around these policies necessitated the ongoing exclusion of the West from government that Johnston observes. Western protest has been almost as powerful a source of political schism in Canada as Quebec nationalism.

Even supporters of free trade must admit that, since 1988, these regional divisions appear to have deepened rather than healed. Canada’s new, formal positioning within a continental free-trade area of a global economy responded to the demands of both Quebec nationalists and western populists, but has not served to integrate either into the Canadian state. Canada’s various regions, which have always had markedly different levels of development, are now clearly evolving such that the linkages of one to another grow increasingly weak.

It is thus hardly surprising that the great insurgencies of Canadian history — Quebec nationalism and Western protest — are stronger today than ever before. For notwithstanding the redefinition of Confederation in terms of “duality” in its second century, the bilingual regions of Canada continue to shrink and bicultural Canadians become increasingly rare. Quebec and the “rest of Canada,” rather than a French Canada and an English Canada, have become the real cultural entities. And the West, although still an economic minority outside of power, is more independent of the priorities of the central Canadian economy than ever before.

What should be done about these developments? If Canada is increasingly two, distinctive, cultural nations, and a number of increasingly independent economic ones, what are to be the emphases and preferences of national political parties? If it isn’t obvious that there can be such a national political agenda, why bother to try to create one or sustain the framework necessary to build it? Perhaps Canada could adopt institutional changes that would ensure more broadly-based governments than in the past, but haven’t the broadly-based governments been the ones most likely to fail?

In some ways, the “natural governing” Liberal Party betrays evidence of Canada’s existential struggle. Like all centre parties beseiged by a large but fragmented opposition in a large but fragmented country, it has defined itself not merely as the party of government, but as the party of the state. Having recreated its own logo as the country’s national flag and entrenched its centralized, bicultural nature as the country’s definition, it links its re-election to the very institutional continuity of the country. It also reflexively resists any significant reform championed by elements of the mutually incompatible opposition.

Convinced Liberals and federalists generally can be reassured however. Though evidence for this deconstructionist view of Canada (and of many other traditional nation-states within the global system) abounds, the political scene only superficially reflects it. The Reform/Alliance phenomenon has always been more about reforming the Canadian state than dismantling it. If it were otherwise, then its concentration on the federal arena would be truly nonsensical. And for all the long-term progress of the Quebec sovereignty movement, it currently appears hopelessly stalled. Independence can only be achieved either by tricking the Quebec electorate into a unilateral declaration, or maneouvering the rest of Canada into underwriting the project. With the Clarity Bill the order of the day, neither scenario is likely, with leading sovereigntists saying so privately and sometimes publicly.

In other words, and more than a little ironically, if any one thing today unites Canadians of different regional backgrounds and partisan stripes, it is a desire not to open major constitutional controversies.

I am somewhat loathe to conclude whether the problems in our party system are temporary, recurrent, permanent or fatal to the Canadian Dominion. Likewise, I am reluctant to proclaim that problems in our party system can be solved by mere evolution, or by institutional, constitutional or existential change. My own tentative assessment is the following:

  1. The Liberal Party is more likely to become a minority party in the near future than any of its opponents are to create majority coalitions.
  2. With no majority party in the country, the parties will probably begin to develop special relationships leading to electoral reform. Any such new electoral system will likely work better than what we have now.
  3. Electoral reform will open up prospects for more general institutional reconstruction in Canada. It is almost impossible to predict where such changes might lead, except that they will go beyond what the Liberals fear and fall short of what Quebec sovereignists desire.
  4. The Canadian Dominion is itself safe unless a “have” province (Ontario, British Columbia or Alberta) were ever to demand full autonomy, or Quebec were ever to become a “have” province.

* This paper could not have been written without the support, direct and indirect, of Canadian Election Studies investigators from 1988 to 1997, André Blais, Henry E. Brady, Elisabeth Gidengil, Richard Nadeau, and Neil Nevitte. They share in the genesis of many thoughts represented in this paper as my own. I am also grateful for earlier comments by Lisa Young, Stephen Harper, and Paul Howe. In the end, I am responsible for all errors of fact and interpretation, not to mention offences against the English language.

  1. Since this paper focuses mainly on the past, references are usually to Reform.
  2. A detailed justification of the specific periods to 1984 can be found in Richard Johnston, André Blais, Henry E. Brady and Jean Crête, Letting the People Decide: Dynamics of a Canadian Election (Montreal: McGill-Queen’s University Press, 1992), chap. two.
  3. The two minority governments with smaller bases were formed in 1962 (Diefenbaker) and 1979 (Clark).
  4. Outright: Ontario (101 seats), Quebec (26), Manitoba (6), and British Columbia (6 in 1997, now 7); almost: Newfoundland and Prince Edward Island (4 each).
  5. For the first identification of the honeymoon, see Richard Nadeau, “L’effet lune de miel dans une contexte parlementaire: le cas canadien,” Revue canadienne de science politique, Vol. 23, no. 3 (September 1990), pp. 483-97. For a more general analysis of government popularity, with evidence from the 1970s to the 1990s, see Richard Johnston, “Business Cycles, Political Cycles, and the Popularity of Canadian Government, 1974-1998,” Canadian Journal of Political Science, Vol. 32, no. 3 (September 1999), pp. 499-520.
  6. The 1917 election appears to the naked eye as the last of the old two-party elections, but it is equally well viewed as foreshadowing the fragmentation to come. The very formation of the Unionist coalition and the split in Liberal ranks reflected the inability of Liberals and Conservatives to contain all the divisions of Canadian life.
  7. Figures 1 and 3 take a small liberty with the historical record by treating Reform as the opposition in 1993 as well as in 1997. Reform enjoyed the second largest vote share throughout this period and all along has been the most serious alternative to the Liberals as a party of government.
  8. So much so that, even though 1945 gave the Conservatives their smallest vote share for the whole 1935-57 period, it also gave them their largest seat share.
  9. The index used here was first proposed in Michael Gallagher, “Proportionality, Disproportionality and Electoral Systems,” Electoral Studies, Vol. 10, no. 1 (March 1991), pp. 33-51. Arend Lijphart, Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies, 1945-1990 (Oxford: Oxford University Press, 1994), chap. 3, argues convincingly that this is the best overall indicator of disproportionality. Lijphart refers to this index as “Lsq.”
  10. New Zealand no longer uses FPP, so that comparison is now of strictly historical interest. For evidence on the FPP systems as well as many Proportional Representation ones, see Lijphart, Electoral Systems and Party Systems, Appendix B.
  11. The measure was first proposed in Markku Laakso and Rein Taagepera, “‘Effective’ Number of Parties: A Measure with Application to West Europe,” Comparative Political Studies, Vol. 12, no. 1 (April 1979), pp. 3-27.
  12. Estimation is by ordinary least squares regression (OLS). The standard error of the slope is 0.005 and the adjusted R2 = 0.92.
  13. The best treatment of these propositions is Gary W. Cox, Making Votes Count: Strategic Coordination in the World’s Electoral Systems (Cambridge: Cambridge University Press, 1997). See especially Part II (“Strategic Voting”) and Part III (“Strategic Entry”).
  14. The exception, which may only prove the second-system rule, was the period just before 1945. Poll evidence and provincial results between 1943 and 1945 frightened the government and stimulated considerable policy innovation, and its 1945 vote was weak. But much of what ultimately happened electorally was only further fragmentation of the opposition.
  15. For quantitative estimates of the difference, see Lijphart, Electoral Systems and Party Systems.
  16. Johnston, “Business Cycles,” compares governments’ inter-election popularity, as indicated by the Gallup poll for the 1974-98 period, and argues that the fragmentation of opposition in 1993 changed the basic dynamics of popularity to the government’s apparent advantage.
  17. See the discussion in Richard Johnston, Public Opinion and Public Policy: Questions of Confidence (Toronto: University of Toronto Press, 1986), chap. two, and in Richard Johnston, “Canada” in Byron E. Shafer (ed.), Postwar Politics in the G-7 (Madison: University of Wisconsin Press, 1996).
  18. The data for Figure 5 come from the International Institute for Democracy and Electoral Assistance (IDEA) website on “Global Voter Turnout” (https://www. i n t – As the website has only preliminary figures for 1997, these have been corrected from the computer file version of Chief Electoral Officer for Canada, Thirty-Sixth General Election 1997: Official Voting Results, Table 3.
  19. The absolute number ostensibly registered grew between 1988 and 1993 by almost exactly the same value as the growth in the estimated voting age population. This implies that every newly eligible voter was registered, or the arithmetic equivalent thereof. If we discount the 1993 registration figure by the average pre-1993 ratio of registration to eligibility, turnout for 1993 creeps up just over 70 percent.
  20. A significant decline registers even before the 1990s, but this seems to be an artifact of the 1949 and 1953 elections, in which registration was strikingly high. With those two elections removed from the series, no 1945-93 downtrend in the ratio of registered to age-eligible voters appears.
  21. If the drop is administratively driven, this could reflect either teething difficulties in the transition or a more basic flaw in the logic of the new system. That the change does reflect administrative practice rather than outside demographic forces is not, in any case, in contention here. My concern is simply to flag the issue.
  22. Data for other countries comes from the IDEA website. See above, note 18.
  23. The most recent statement and a useful source of references to his earlier work and to its major critiques is Ronald Inglehart, Modernization and Postmodernization: Cultural, Economic, and Political Change in 43 Societies (Princeton: Princeton University Press, 1997).
  24. Neil Nevitte, The Decline of Deference (Peterborough: Broadview, 1996) applies the Inglehart thesis to Canada.
  25. Harold D. Clarke and Nitish Dutt, “Measuring Value Change in Western Industrialized Societies: The Impact of Unemployment,” American Political Science Review, Vol. 85, no. 3 (September 1991), pp. 905-20, and James A. Davis, “Review Essay: Paul R. Abramson and Ronald Inglehart, Value Change in Global Perspective,” Public Opinion Quarterly, Vol. 60, no. 2 (Summer 1996), pp. 322-31 are especially telling critiques.
  26. Kitschelt’s account of the left is The Transformation of European Social Democracy (Cambridge: Cambridge University Press, 1994), and of the right, The Radical Right in Western Europe: A Comparative Analysis (Ann Arbor: University of Michigan Press, 1995).
  27. Hermann Schmitt and Sören Holmberg, “Political Parties in Decline?” in Hans-Dieter Klingemann and Dieter Fuchs (eds.), Citizens and the State (Oxford: Oxford University Press, 1995), Volume One in the “Beliefs in Government” Series (Max Kaase, Kenneth Newton, and Elinor Scarbrough, series editors), chap. four.
  28. Stefano Bartolini and Peter Mair, Identity, Competition and Electoral Availability: The Stabilisation of European Electorates 1885-1985 (Cambridge: Cambridge University Press, 1990)
  29. Ola Listhaug and Matti Wiberg, “Confidence in Political and Private Institutions,” in Klingemann and Fuchs (eds.), Citizens and the State, chap. ten. See especially Figures 10.1, 10.3, and 10.4. We cannot reject the possibility that high governmental turnover is as much symptom as cause of political distrust, but the Listhaug-Wiberg graphics lean to the opposite interpretation. Particularly striking is the variance in political trust by turnover within European subcultures. For instance, measured trust levels in Denmark, a country with high government turnover, are closer to those in Italy than to those in Norway, an historically stable system.
  30. Richard Topf, “Electoral Participation,” in Klingemann and Fuchs (eds.), Citizens and the State, chap. two.
  31. Estimates are based on OLS regression analysis for elections from 1935 to 1988 inclusive. Adjusted R2 for the equation is 0.86. Standard error for the Liberal vote coefficient is 0.17, for the Conservative coefficient, 0.11. Each coefficient has less than one chance in 10,000 of being the product of random covariation. Each coefficient indicates the effect on seat share of a unit shift in one party’s vote share, holding the other party’s vote constant. In fact, the two vote shares vary inversely with each other, as one would expect of two parties historically in such close competition. The regression coefficient of Tory on Liberal vote share is -0.94, an essentially one-to-one relationship. Note, however, that the adjusted R2 = 0.60, indicating plenty of slack in the relationship, the substance of which is discussed in the body of the text below. The conversion of percentage-point shifts into seat shifts is just crude arithmetic. In a House of 300 seats, a percentagepoint shift is three seats. Since 1935 the House has grown from 245 to 301, whence my rounding in multiples of 2-3 seats.
  32. For illustrations of the small-party logic see Alan C. Cairns, “The Electoral System and the Party System in Canada, 1921-1965,” Canadian Journal of Political Science, Vol. 1, no. 1 (March 1968), pp. 55-80. On the logic over the full range see Richard Johnston and Janet Ballantyne, “Geography and the Electoral System,” Canadian Journal of Political Science, Vol. 10, no. 4 (December 1977), pp. 857-66). Both articles illustrate arguments first made in Seymour Martin Lipset, “Party Systems and the Representation of Social Groups,” Archives européenes de sociologie/ European Journal of Sociology, Vol. 1, no. 1 (1960), pp. 50-85.
  33. The system unquestionably has another unfortunate geographic effect: it exaggerates regional differences in strength, making a party look stronger where it is strong and weaker where it is weak. See Cairns, “The Electoral System and the Party System.”
  34. The tilt to the Liberals reflects their stranglehold on nonfrancophone seats. Note that Figure 5 masks some of the party-specific variation just described, as the figure focuses on the government share, regardless of the party. So the massive swing from Liberal to Conservative in 1984, for example, instead appears as continuity, with Quebec’s share of government seats dropping slightly.
  35. On this see David Sankoff and Koula Mellos, “La régionalisation électorale et l’amplification des proportions,” Revue canadienne de science politique, Vol. 6, no. 3 (September 1972), pp. 380-99, who compare Quebec provincial elections with Ontario ones as well as with Canadian, British, New Zealand, and US ones. See especially Tableau II, “Propriété des distributions des comtés.”
  36. From 1935-88, the median number of Quebec seats won by the government was 58 and the median number of seats required for a majority was 134, hence a majority winner typically needed about 76 seats from other regions.
  37. Charles Boix, “Setting the Rules of the Game: The Choice of Electoral Systems in Advanced Democracies,” American Political Science Review, Vol. 93 no. 3 (September 1999), pp. 609-24, argues that because of their primary concentration in Quebec, Canadian francophones are one of the few ethnic minorities in the world that is better off under FPP than under PR.
  38. The westward shift should not be exaggerated. Recent change is still dwarfed by that from 1896 to 1911, which in effect called the West into existence as a political region. By 1911 Saskatchewan was the third largest province, a position it held until the 1951 census. Much of the postwar change has been redistribution within the region, and the West’s overall share actually shrank from the 1930s to the 1970s. Only since the 1970s has the region’s overall share grown.
  39. This overstates the true Conservative seat position, as it includes the artifactually high 27 percent of seats won in 1945 on the period’s lowest vote share. The Liberal vote share dropped precipitately that year, on a temporary fractionalization of the total vote. In the other four elections the Conservatives averaged under 17 percent of seats.
  40. In part, Liberal strength reflected the accession of Louis St. Laurent to Liberal leadership, as, before 1993, having a leader from Quebec was worth 13 points in Liberal share. See Richard Nadeau and André Blais, “Explaining Election Outcomes in Canada: Economy and Politics,” Canadian Journal of Political Science, Vol. 26, no. 4 (December 1993), pp. 775-90.
  41. If we accept that the normal pre-1957 Conservative vote in Quebec was closer to 30 percent than to 20 percent, Figure 5 understates the real drop in the Quebec Conservative share.
  42. The gains in Alberta and BC were more appearance than reality. Tory shares after 1984 were squarely in line with Conservative strength in those provinces at the end of the 1962-80 period. The whole-period line is dragged down by weaker (in the case of BC, much weaker) shares of the 1960s.
  43. This paragraph begs two qualifications. First, Quebec nationalism did have an outlet before 1984, Social Credit, and the absorption of the Social Credit vote into the Conservative coalition in 1984 was part of the latter’s pre-emption of Quebec nationalism. Social Credit was not a vehicle for the forward-looking variant of Quebec nationalism, however. Second, the Tories did not emit obviously nationalist signals in the run up to 1984. Indeed, one of Brian Mulroney’s first tests as leader was to align his party with the Liberal party’s rights-based, Official Languages agenda in relation to Manitoba.
  44. Respondents’ 1988 behaviour is based on their recall questions in the 1993 wave of the 1992-3 Canadian Referendum and Election Study. For more detail on construction of the table, see André Blais and Elisabeth Gidengil, “Constructing a Flow of the Vote Table” (Montréal: Université de Montréal, mimeo, 1995).
  45. It is tempting to say that shares here are of the whole electorate, but that is not quite true. Note that voters now dead are part of the 1993 total, just as voters not yet of age are part of the 1988 total. If it seems odd to count this way, doing so is necessary to make the accounting system consistent.
  46. For comparison, look at the percentage voting Liberal, 19.4. Although an absolutely larger percentage, former abstainers are much less likely to vote Liberal than theelectorate as a whole.
  47. The basis of this claim is as follows. New voters are 13 percent of the 1993 total, and the Bloc’s 16.2 percent of that 13 percent makes new-voter Blocistes 2.1 percent of all 1993 voters. Former abstainers constitute 21.7 percent of all 1993 categories and the 6.9 percent of these who vote Bloc thus are 1.5 percent of the 1993 total. The sum, 2.1 plus 1.5, is 3.6 percent of all categories. Dividing this sum by the total Bloc percentage, 8.8, yields 3.6/8.8 = 0.41, orabout 40 percent of the Bloc vote.
  48. Table 3 indicates that turnout dropped from 1988 to 1993, but it masks the fact that in Quebec, turnout grew, reflecting this mobilization of sovereignists.
  49. These percentages are calculated by omitting from the denominator all respondents who stayed with their 1988 party.
  50. Given the power of ideological distance, should we be amazed that any New Democrats moved to Reform? In fact, the NDP has always exchanged voters with the most distant party on the left-right scale, which is to say that party turnover in Canada is routinely intransitive, reflecting the fact that it is controlled by at least two policy dimensions, on one of which the NDP and Conservatives or Reform are not polar opposites.
  51. This defection took place both inside and outside Quebec, such that within each sub-electorate, the Liberals were not so peculiarly important.
  52. This pattern is not just a reflection of the sheer bulk of Ontario, where NDP switches to the Liberals might have been especially numerous. Even in the West, where Reform is most viable, the Liberal party was by far the main destination for ex-New Democrats. Only in Alberta did NDP defectors to Reform outnumber those to the Liberals. In Alberta, Reform was so overpoweringly attractive that any other pattern was next to impossible. Even so, relative to the respective drawing power of Reform and the Liberals, Alberta NDP movement was still disproportionately to the Liberals. In any case, the number of 1988 New Democrats in Alberta was minuscule. The big Western NDP battalions resided in BC, Saskatchewan, and Manitoba and in those provinces the Liberals were as differentially attractive to former New Democrats as they were in Ontario.
  53. In the 1988 study, the object was labelled “French Canada” but in 1997, “Quebec.” Evidence from 1993, when both versions were employed as a test, suggests little difference in the response evoked.
  54. Outside Quebec, respondents saw the lean as more than mild; inside Quebec, they did not see it as a lean at all. The mean perception splits the difference.
  55. For speculation on why it made sense for the Liberal party to stray so far from its supporters’ location, see Johnston, et al., Letting the People Decide.
  56. These observations are only suggestive. They require that the Reform location be projected from Panel B to Panel A, perhaps a questionable practice, and then horizontal distances compared.
  57. A few further comments on the basis of Reform support need to be made. Elisabeth Gidengil, André Blais, Richard Nadeau and Neil Nevitte “Making Sense of Regional Voting in the 1997 Federal Election: Liberal and Reform Support Outside Quebec,” Canadian Journal of Political Science, Vol. 32, no. 2 (June 1999), pp. 247-72, argue that, although Westerners have roughly the same mean position on Quebec-Canada relations, opinion on this issue makes a bigger difference to the Reform vote in the West. That is hard to square with my claim that the vote differences are about the same within these regions. The discrepancy may be between the bivariate graphical account that underpins my statement and the multivariate one that underpins theirs. Or the issue may be the year, 1993 (mine) versus 1997 (theirs). Nevitte, Blais, Gidengil and Nadeau, Unsteady State: The 1997 Canadian Federal Election (Don Mills: Oxford University Press, 2000), Table C.4, confirms the importance of West-focused regional alienation, however. Does Reform represent a more generalized alienation from party politics? The answer is Yes. For 1993, Reform was only weakly distinguished this way, much less than the Bloc, according to Richard Johnston, André Blais, Henry E. Brady, Elisabeth Gidengil, and Neil Nevitte, “The 1993 Canadian Election: Realignment, Dealignment, or Something Else?” presented to the 1996 Annual Meeting of the American Political Science Association. In 1997, Reform became more distinct, as cynicism about politics evidently diminished in the other party groups, according to Blais et al., Unsteady State, Figure 4-4 . There is, in sum, much work still to do in understanding post-1993 party politics, but most of the work lies off the path of this paper.
  58. On this see Kaare Strom, “Minority Governments in Parliamentary Democracies: The Rationality of Non-Winning Cabinet Solutions,” Comparative Political Studies, Vol. 17, no. 2 (July 1984), pp. 199-227.
  59. It can yield the “wrong” winner, to be sure. This was probably the case in British Columbia in 1996 or Ontario in 1990 and probably describes many Tory federal victories in the 20th century.
  60. The comparison is not idle. Canada and India have alwaysstood as the great challenges to the apparent logic of FPP. In the Canadian case, it traditionally sufficed to refer to local two-party consolidation, with different parties standing as front-runners in different places. India was harder to solve. William H. Riker, “The Number of Political Parties: A Re-Examination of Duverger’s Law,” Comparative Politics, Vol. 9, no. 1 (October 1976), pp. 93-106 addressed both the Canadian and Indian cases. For Canada, he concentrated on geography. For India he proposed that the key was Congress’ centrist position, which made it the Condorcet winner. Now we see that his argument also applies to Canada, and may have all along.
  61. Nevitte et al., Unsteady State, attribute two percentage points of the turnout difference between 1997 and the pre- 1993 average to competitive factors. Theirs is probably a lower-bound estimate.
  62. Boix, “Setting the Rules of the Game” argues that incumbent parties’ calculations of risk are historically the central factor in electoral reform.