The Emergence of Parties in the Canadian House of Commons (1867-1908)
Total Page:16
File Type:pdf, Size:1020Kb
The Emergence of Parties in the Canadian House of Commons (1867-1908). Jean-Fran¸coisGodbouty and Bjørn Høylandz y D´epartement de science polititque, Universit´ede Montr´eal zDepartment of Political Science, University of Oslo Conference on the Westminster Model of Democracy in Crisis? Comparative Perspectives on Origins, Development and Responses, May 13-14, 2013. Abstract This study analyzes legislative voting in the first ten Canadian Parliaments (1867-1908). The results demonstrate that party voting unity in the House of Commons dramati- cally increases over time. From the comparative literature on legislative organization, we identify three factors to explain this trend: partisan sorting; electoral incentives; and negative agenda control. Several different empirical analyses confirm that intra-party conflict is generally explained by the opposition between Anglo-Celtic/Protestants and French/Catholic Members of Parliament. Once members begin to sort into parties according to their religious affiliation, we observe a sharp increase in voting cohesion within the Liberal and Conservative parties. Ultimately, these finding highlight the importance of territorial and socio-cultural conflicts, as well as agenda control, in ex- plaining the emergence of parties as cohesive voting groups in the Canadian Parliament. This study explains the development of party unity in the Canadian House of Commons. We take advantage of the historical evolution of this legislature to analyze a complete set of recorded votes covering the first ten parliaments (1867-1908). This early period is of interest because it was during these years that the first national party system was established, the electoral franchise was limited, and the rules and procedures of the House were kept to a minimum. Since the Canadian legislature is an Anglo-American hybrid, it also affords the opportunity to explore the differences between the American and British parliamentary systems (Kam, 2009, 2001; Franks, 1987; Kornberg, 1967; Epstein, 1964). This form of government requires strong party discipline, while the sectional divisions coalesced by the federal structure, should lead us to expect a much looser American-style party organization (Epstein, 1964, p. 57). From a historical perspective, the Canadian House of Commons is also interesting because it integrates four former independent colonies|with their own partisan structure|into a new federal legislative system. Much like in the years after the creation of the European Union, we should find that over time, the regional ties between Members of Parliament (MPs) gradually weakened at the expense of partisanship (Hix, Noury and Roland, 2005; Haas, 1958). However, we should also find tensions between the unity of federal parties and the need for legislators from different regions to cater to the interests of their constituents (Lebo, McGlynn and Koger, 2007; Hix, 2002). We are primarily interested in determining what factors are most important in explaining legislative behaviour and the organization of the Canadian House of Commons. We begin by providing a brief historical overview of the political context in the years following Con- federation.1 We outline three groups of variables that we believe can explain the increase in party unity in the House of Commons: partisan sorting; electoral incentives; and negative agenda control. We also relate each of these factors to the literature on legislative organiza- tion more broadly. Although some researchers, such as Cornell (1963) and Hare (1973), have analyzed the composition of different legislative coalitions prior to Confederation to explain 1For a detailed account of elections and the political economy of Canada during this period, see Beck (1968) and Easterbrook and Aitken (1988). 1 the roots of the Liberal-Conservative party merger of 1854 (which later became the Conser- vative party), this study represents the first historical account of legislative voting and party unity in Canada.2 Our research also addresses the more general question of the development of legislative organization and party unity in Westminster style parliamentary systems. Since legislatures ultimately deal with the most important issues of the day, the historical analysis of parliamentary voting can provide researchers with a unique tool to effectively learn about the evolution of political conflicts. We believe that such an analysis is important in the Canadian context because contemporary parties tend to reflect the political cleavages of the periods in which they were first created (Liphjart, 1977; Lipset and Rokkan, 1967). Party Unity and Confederation Much like in the British case, it is generally assumed that party unity in the Canadian House of Commons was very weak in the years following Confederation (Carty, 1988). Indeed, the first few parliaments contained several Liberal or Conservative \loose fish” or \shaky fellows" who did not always support their own party (Kornberg and Mishler, 1976; Underhill, 1935; Reid, 1932). Figure 1 supports this assumption by displaying a conventional measure of party unity indicating the proportion of recorded divisions where a simple majority (.50) of the Conservative party voted against the Liberal party (solid line) in the first ten Canadian parliaments (1867-1908). The Figure also includes an additional measure of unity where at least .90 of one party voted against .90 of the other (dotted line). This last index identifies the proportion of highly partisan votes (see for example Brady, Cooper and Hurley 1979). The Figure illustrates an important surge in the number of partisan votes during the 4th Parliament while the proportion of majoritarian votes remains relatively constant over time. After the 4th term, the number of majoritarian and partisan votes in the Commons gradually converges. This is in sharp contrast with voting in the U.S. Congress during the same period. 2However, Eggleston (1988) wrote a Master's thesis to analyze some aspects of legislative voting in the first three parliaments. Massicotte (1989) also found that in the provincial legislature of Quebec, discipline was at its weakest in the years following Confederation. 2 Party Unity 1.0 0.8 0.6 0.4 Percentage of votes 0.2 Majority 50/50 Majority 90/90 0.0 1 2 3 4 5 6 7 8 9 10 Parliaments 1st-10th (1867-1908) Figure 1: The plots report two measures of party unity. The solid line represents the pro- portion of votes in a given parliament where .50 of one party voted against .50 of the other. The dotted line represents the proportion of votes where .90 of one party voted against .90 of the other. At its highest point in 1902, the House of Representative had a comparable level of highly partisan votes of .64 (Brady, Cooper and Hurley, 1979). Still, the level of voting unity found in the Canadian House of Commons at the turn of the century is much higher than what we observe in the British Parliament (Lowell, 1908, table 2, p.77). How can we account for this increase in voting unity in the Canadian House of Commons? The literature on legislative voting generally explains party unity in terms of internal and external factors. Internal factors are associated with leadership and the centralized structure of legislative decision-making (Stephens and Brady, 1976; Brady and Althoff, 1974). For example, when members of the government control the legislative agenda, they can often avoid splitting their caucus by preventing a divisive issue from coming to a vote (Cox and McCubbins, 2005, 1993). In addition, parties can use incentives such as patronage, committee assignments, or other legislative perks to promote member loyalty (Carey, 2009). External factors, such as who controls the selection of candidates as well as the importance of|and 3 access to|campaign resources for individual members, can also have an impact on legislative voting (Hix, 2004). Unfortunately, since both external and internal factors generally coexist simultaneously in a legislature, it is extremely difficult to disentangle between their individual effects. As Krehbiel (1993) points out, high levels of partisan voting could simply be explained by the cohesiveness of the preferences of members, independently of parties or the electoral process (see also Krehbiel, 2000). Some parties may even systematically oppose government proposals for strategic reasons, regardless of the preference of their members (Dewan and Spirling, 2011). In order to estimate the independent effect of these different factors on legislative voting, this study proposes an empirical strategy to determine the sequence of their occurrence in the House of Commons. We identify three types of events|ideological sorting, the modification of electoral laws, and the increasing control of the agenda by the governing party|to account for the rise in party voting unity over time. We are interested in determining whether the gradual economic integration of the Canadian provinces and the increasingly divergent views adopted by both parties on language and religion help explain the consolidation of the party system during this period. We also consider the possibility that electoral pressure, or the late adoption of the secret ballot and same day elections, reduced the number of independent minded legislators by increasing the benefits of being loyal to one of the two major parties. Finally, we are interested in finding out if, by increasing its control over the parliamentary process over time, the government could have changed the rules and procedures of the House in order to keep some of the most divisive issues from the legislative agenda. Since we use historical voting data in this analysis, we are able to establish precisely when these factors emerged and disentangle their independent effects on the development of parties as cohesive voting groups. The rest of the study is organized as follows. We begin by describing in more details the three different factors that can explain party unity in the Canadian House of Commons: ideological sorting; electoral incentives; and negative agenda control. Next, we present the empirical strategy, methodology, and data used to evaluate each of their contribution.