Political Dynasties in Canada

Abstract

Using a unique dataset of legislators' electoral and biographical data in the Canadian provinces of Ontario, Quebec, New Brunswick, Nova Scotia and the federal parliament, this article analyses the extent to which family dynasties affected the career development of legislators since the mid-18th century. We find that the prevalence of dynasties was higher in provincial legislatures than it was in the federal parliament, that the number of dynasties in the Senate increased until the mid-20th century, and that the proportion of dynastic legislators at the subnational level was similar to the numbers seen in the United Kingdom during the early 19th century. Our results confirm the existence of a clear career benefit in terms of cabinet and senate appointments. In contrast to the American case and in line with the United Kingdom experience, we find no causal relationship between a legislator's tenure length and the presence of a dynasty.

Canada ranks among the most dynastic democracies in the world (Van Coppenolle and Smith 2022), with a substantial lineage of political families elected in both subnational (hereafter, provincial) and national (hereafter, federal) levels. This tradition spans from the mid-18th century to the present day, making it one of the longest continuous sequences of dynastic political representation within legislative assemblies. The enduring presence of these dynasties represents a long-lasting legacy of the country's colonial heritage. From the influential families of Trudeau, Mercier, Ford and Regan, with their respective prime-ministerial lineages and cabinet contributions going back multiple decades, the annals of Canadian politics are full of illustrative examples of political dynasties. These family connections raise important questions about the nature of democratic equality, especially if they provide unfair electoral advantages or career benefits.

The extent to which being part of a family dynasty is associated with certain types of privilege has been subject to increased scholarly attention in recent years (for a review of the literature, see Geys and Smith 2017; Van Coppenolle and Smith 2022). Indeed, researchers have shown that members of political dynasties receive some sort of electoral benefit from their family ties in legislative assemblies (Feinstein 2010; Fiva and Smith 2018; Laband and Lentz 1985), but the evidence is mixed concerning what causes a dynasty to emerge. Ernesto Dal Bó et al. (2009) and Pablo Querubin (2016) argue that there exists a causal relationship between dynasty formation and tenure in office in both the United States and the Philippines. However, no such effect is found in Norway or the United Kingdom (Fiva and Smith 2018; Van Coppenolle 2017).

In the comparative context, Canada represents an ideal case. Canadian political institutions were influenced both by the country's experience with British imperialism and in reaction to the development of the American Republic (Dawson 1954; Epstein 1964). On the one hand, Canada is a Westminster constitutional monarchy with an unelected upper chamber. On the other hand, Canada, like the United States, lacks a landed gentry and aristocratic class structure, but shares a history of frontier development and colonial democracy (Antoine et al. 2022). Taken together, these characteristics create a space to promote both social mobility and an open political system, while simultaneously allowing for the presence of an unelected chamber reminiscent of the House of Lords, which could potentially give rise to patronage and nepotism.

Canada's French population also represents a unique contribution to the development of political dynasties. Historically concentrated in Quebec, French-Canadian society was for a long time characterized by a quasi-feudal political system with relatively dense kinship networks (de Tocqueville 2003). This unique context introduces yet another factor that could have promoted family relations in politics (Lemieux 1971). Likewise, family ties in politics could also have existed elsewhere in Canada, particularly prior to Confederation in 1867, when British North American colonies were governed by their own separate legislative assemblies with local elites sharing strong family connections.

Considering that Canada has an electoral history going back to the mid-18th century, it is surprising to note that very little has been written on the development, presence and electoral benefits of being part of a political family dynasty in this country. While Daniel Smith (2018) and Brenda van Coppenolle and David Smith (2022) include this case as a comparative unit in their analysis of political dynasties across several countries, the work of Matthew Godwin (2018) represents the only systematic study of family relations in Canada to date. In this article, we add to the comparative literature on political dynasties by analysing the structure of legislative careers over the last 250 years. We leverage a unique dataset of legislator biographical information in the Canadian provinces of Ontario, Quebec, New Brunswick and Nova Scotia, the first four original provinces of the Dominion of Canada. These data go back to 1758 and also include the biographical information of federal legislators from both the federal House of Commons and the Senate after 1867. The inclusion of both federal and provincial data allows us to observe variations within this country, while controlling for overarching structural and cultural factors. Our results demonstrate that there are legislative career advantages associated with political dynasties in Canada. However, these effects are generally small and limited to the provincial arena. We also find evidence of other career benefits, notably that members of political dynasties are more likely to be appointed to the Senate and to the provincial and federal cabinets. In line with the study of dynasty formation in parliamentary democracies (Van Coppenolle 2017), and in contrast to presidential systems (Dal Bó et al. 2009), our analysis confirms that dynasty formation in Canada is not caused by tenure in office.

Political family dynasties

Given the large amount of attention on family successions in the business literature (e.g. Davis and Harveston 1998), it is surprising to find a limited amount of research that explains why political dynasties emerge in democratic contexts. Indeed, research on this topic has confirmed that offspring follow their parents into politics in larger numbers than in the general population, with the exception of farming and self-employed proprietors (Dal Bó et al. 2009; Laband and Lentz 1985). The implications suggest that dynasties provide potential legislators with a series of advantages ranging from skill transfer, voter loyalty, incumbency advantage and name recognition (Feinstein 2010; Fiva and Smith 2018; Geys 2017; Laband and Lentz 1985; Smith and Martin 2017; Van Coppenolle and Smith 2022). In the US, David Laband and Bernard Lentz (1985) argue that being part of a dynasty might make ‘politics’ a hereditary skill, providing the legislator with voter loyalty and a higher degree of incumbency advantage than would normally be enjoyed by legislators who are not part of a dynasty. Name recognition offers dynastic legislators an electoral advantage in open-House seats because voters rate these candidates ‘warmer’ on feeling thermometers, when compared to their non-dynasty counterparts (Feinstein 2010; Smith 2018). In Norway, Jon Fiva and Daniel Smith (2018) find that an electoral advantage exists for members of a dynasty as well. This is particularly noteworthy because Norway uses a closed-list proportional representation system, implying that dynasty linkages are present in systems where party elites are responsible for selecting candidates and, crucially, their ballot list order.

Beyond these electoral advantages, dynasty members also benefit from career advantages. Dynastic members are more likely to be appointed to cabinet, but only if their previous relatives have been appointed to cabinet as well (Smith and Martin 2017; Van Coppenolle 2017; Van Coppenolle and Smith 2022). We expect the dynasty–cabinet link to be a relevant factor in Canada, as it is in Ireland (Smith and Martin 2017) and the UK (Van Coppenolle 2017). Unlike presidential systems, the pool of candidates in parliamentary systems is drawn directly from the ranks of elected representatives. As such, familial ties and name recognition might induce an intra-parliament benefit to legislators who can advance their careers more easily than first-generation legislators.

The conclusions are less clear when it comes to the formation of such dynasties. In the most comprehensive study to date, Van Coppenolle and Smith (2022) looked at the role of political dynasties in a historical and comparative perspective by analysing the careers of politicians over time in 15 advanced industrialized democracies, including the federal Canadian parliament. Their findings confirm the general decline in hereditary politics, with the exception of Japan and Ireland, where parties are weak and elections are highly personalized (see also Smith 2018; Smith and Martin 2017). This study also confirms that the Canadian federal parliament saw a reduction in dynastic selection after Confederation; however, members of political dynasties showed a significant advantage in the selection to cabinet positions. The authors offer several hypotheses to explain these results, including political modernization, institutional changes and electoral system reforms.

While the analysis by Van Coppenolle and Smith (2022) provides us with an insightful understanding of the global trends in political dynasties, the work of Dal Bó et al. (2009) offers crucial insights into the possible role of tenure length in dynasty formation. This study confirms that a causal relationship exists in the US Congress between the time spent in office (tenure) and dynasty formation. By using a regression discontinuity design to compare those legislators who just barely lost/won their first re-election attempt, the authors find that dynasty formation (and continuation) is caused by tenure. These findings have also been confirmed in the Philippines (Querubin 2016). However, using a regression discontinuity design similar to Dal Bó et al., other studies have found that in cases where candidates narrowly missed re-election, their relatives were subsequently elected; this phenomenon has been observed both in Norway (Fiva and Smith 2018) and the UK (Van Coppenolle 2017). Evidently, tenure length is not a causal mechanism in all countries and appears to be limited to presidential systems.

The Canadian case

Colonial Canada was originally dominated by an elite network, known as the Family Compact in Ontario and the Château Clique in Quebec, who wielded control over the government and crown land (Tepperman 1972). These groups represented a privileged segment of the population composed of influential families that used their connections and power to dominate politics, especially in the executive council and the upper house, which operated independently from the legislative assembly and whose members were appointed by the governor (Dawson 1937 [1965], 1954). Although the influence of these elites was somewhat reduced after Confederation in 1867, it is possible that remnants of their power persisted throughout the 20th century, particularly in the unelected Canadian Senate.

Likewise, the presence of elected colonial assemblies in Canada, which predate the formation of legislative political parties, most likely contributed to promote the emergence of family dynasties. The first few elections in the 18th and 19th centuries concerned a small number of electors; corruption was rampant and family clans often found themselves at odds, not because of partisan differences but rather because of conflicts over land or other resources (Hughes 1943; Miner 1939). As a result, political power was highly personalized, with family names serving as labels to coordinate voting in exchange for work, government positions or gifts (Lemieux 1971; March 1974). Even though patronage reforms were implemented in the early 20th century at the federal level, the impact of family networks was felt longer in provincial politics, where local elites retained control over political appointments for a longer period (Lemieux 1975). This prolonged influence could also have created more opportunities for political dynasties to persist in the provinces.

These political conditions lead us to expect that family ties had a strong influence in Canadian politics during the 19th century. However, it is unlikely that this effect will have persisted beyond this point. Relatively lower franchise requirements after Confederation reduced the likelihood that a small and organized group of voters could promote the re-election of their family members to parliament. Although Canada's immigration came mostly from the British Isles until the 1900s, subsequent waves of newcomers increased the diversity of the population and altered the composition of parliament (Ogmundson and McLaughlin 1992). It is possible that these changes, in combination with the shifts towards mass electoral appeals, reduced the importance of family dynasties in politics during the 20th century. However, we suspect that the role of family connections could have continued to play a significant role in the selection of cabinet members.

Indeed, examples of family linkages in Canada abound. At the federal level, both Prime Minister Justin Trudeau's father (Pierre) and grandfather (James Sinclair) previously served in the House – with Pierre serving as a long-time prime minister and James Sinclair as a cabinet member. Provincially, in Quebec, Premier Honoré Mercier's son (Honoré Mercier II), grandsons (Gaspard Fauteux, Paul Gouin, Honoré Mercier III), son-in-law (Premier Lomer Gouin) and great-great-grandson (Thomas Mulcair) were elected at least once to the parliament of Quebec. The Ford family in Ontario is another example, with Rob Sr. being elected as a Progressive Conservative to the provincial legislative assembly in 1995. His son, Doug, is currently the premier, and his grandson, Michael, serves as a cabinet minister. Similarly, in Nova Scotia, Gerald Regan, premier from 1970 to 1978, had a father-in-law, John Harrison, who served in the federal parliament, a son, Geoff Regan, elected to the House of Commons and later Speaker, and a daughter-in-law, Kelly Regan, who was cabinet minister in the Nova Scotia Assembly.

Method

The goal of this study is to measure the advantages of family ties in politics by leveraging a large and consolidated dataset of legislators from Ontario, Quebec, Nova Scotia and New Brunswick since the 18th century. Information for these legislators includes age, year of death, gender, political affiliation (if applicable), occupation, place of birth, religion and family relations (Rivard et al. 2024). We employ Dal Bó et al.'s (2009: 116) definition of dynasties, as ‘those from a family that had previously placed a member’ in the legislature. Similarly, those who start a dynasty are also considered part of it, given that they had family members elected afterwards. We limit family linkages to those observed within chambers and not across chambers. For example, Justin Trudeau's federal-level connections – his father, Pierre, and his grandfather – are identified. However, Caroline Mulroney, a member of the Ontario legislature, is not identified as a dynast given that her father, former Prime Minister Brian Mulroney, did not sit in the Ontario legislature. Since we are interested in the effects of being part of a family dynasty, our unit of analysis is the individual legislator.

The biographical information of provincial legislators was primarily collected from their respective parliamentary biographies. We relied on various legislature-specific sources to aggregate this content. These include comprehensive biographies of legislators, including their kinship, for legislators in Quebec and Nova Scotia. For legislators in Ontario and New Brunswick, we relied on legislators' official biographies and the Dictionary of Canadian Biography to complete the missing information. For federal legislators, we used the family links information provided by the Parliament of Canada in its publicly available dataset.

The identified family links are: siblings, in-law relatives (e.g. brother, father, son, daughter, sister, grandfather, grandson, mother), cousins, father/mother/son/daughter, grandparent/grandchild, great-relations (e.g. great-grandfather, great-granddaughter/grandson, great-great-great grandson, great-great-granddaughter, great-nephew/uncle), uncles/aunts/nephews/nieces, godparent/godchild, half-sibling and step-relations (e.g. step-father, step-daughter).

There are, of course, methodological limitations to our approach. To begin, we have only limited information about members who are currently in office, and historical record-keeping may have overlooked their more recent relatives. Furthermore, we acknowledge that in some cases, other links might remain unaccounted for due to incomplete archives. Despite these issues, our biographical dataset remains the most comprehensive consolidation of available historical information to date. For a more detailed description of our data-collection process and its limitations, see Alex Rivard et al. (2024).

Biographical data were combined with electoral data from the Canadian Elections Database (Sayers 2017) and each province's election reporting body. Electoral data are available from 1867 onwards, as is information about whether the legislator was appointed to the cabinet or not. At the federal level, we rely on the Canadian parliament's biography of all elected members of the House and senators since 1867, the first year in which a federal election was held in Canada. Legislators' parliamentary biographies maintain the same information as does our provincial dataset. The federal data include members from all provinces and not just Ontario, Quebec, New Brunswick and Nova Scotia. We combine the parliamentary data with candidate-level election results (Sevi 2021).

We present a series of estimations that model the electoral benefits and career advantages of having past relatives in the legislative chamber. All models include province and parliamentary sitting fixed effects, unless otherwise noted. The inclusion of these fixed effects is crucial because they allow for the incorporation of unobserved heterogeneity specific to each province as well as the over-time trends that have occurred – particularly at the provincial level – since the 18th century.

Historical evolution of dynastic links

Figure 1 confirms that family dynasties at the provincial and federal levels have long been a constant presence in Canadian politics. The first panel in the figure reports the proportion of all sitting members in a legislature who have a family member elected prior to their own entrance (henceforth, pre-relative); the second panel shows those who have a family member elected to their respective legislature after their own initial election (henceforth, post-relative). We find that 11.8% of legislators in the Canadian provinces had a relative elected prior to their entrance and 12% had a relative elected after their entrance.

Figure 1. Share of Legislators with a Pre- and Post-relative Family Tie.

Note: Figure shows the proportion of legislators in each provincial legislature with at least one family member elected prior to their entrance (pre-relative, first panel) and with at least one family member elected after their initial election to their respective legislature (post-relative, second panel). The vertical line represents Confederation in 1867.

As Figure 1 shows, dynastic links are much more prevalent in Quebec than in the other provinces: 16.2% of legislators in Quebec had a pre-relative elected to the legislature whereas this is only true for 7.4% of Ontario legislators, 9.4% of New Brunswick legislators and 13.7% of Nova Scotia legislators. The share of Quebec legislators with a post-relative elected after initial entrance is likewise higher at 16.8%, compared to Nova Scotia at 13.3%, New Brunswick at 9.5% and Ontario at 7.7%.

Figure 1 also confirms that dynasties have been declining in the four Canadian provinces over time. The share of legislators with a pre-relative peaked at roughly 50% in both Nova Scotia and Quebec during the first half of the 19th century. This trend is notable since it effectively mirrors what is observed in the British House of Commons and the United States during the same period (Van Coppenolle and Smith 2022, see Figure 2). Note that the proportion of pre-relatives in these two chambers appears to surpass even the highest levels observed in both of these countries. It is also interesting to note that the share of legislators with a pre-relative in Ontario and Quebec begins to drop considerably around the middle of the 19th century. We also note that the share of legislators across all provinces with a pre-relative declined until Confederation in 1867 – the vertical line in Figure 1. Yet, the proportion of legislators with a prior elected relative soon began to rise again after the federal union of Ontario, Quebec, New Brunswick and Nova Scotia. It appears, then, that the new Canadian federal structure initially served as an opportunity to reshape the political landscape. However, these results also suggest that Confederation provided an opportunity for new dynasties to form.

Figure 2. Proportion of Legislators with a Pre- and Post-relative by Parliament and Chamber Type at the Federal Level.

The federal House of Commons and the Senate have likewise seen dynastic linkages in both chambers, but their presence has been more muted than in their provincial counterparts. Only 4.8% and 5% of House members have had a relative elected before or after them respectively. In contrast, Figure 2 shows that 9.1% and 6.9% of senators have had a family member either elected to the House or appointed to the Senate prior to their entrance in this chamber. We can see the evolution of these dynastic connections at the federal level in Figure 2, which replicates Figure 1. Each panel reports the proportion of MPs (first panel) and senators (second panel) with pre- and post-relative relations. These plots limit the sample to those who were only ever MPs and those who were only ever senators.

Among MPs in the federal parliament, the trend of legislators with a family link in the House has been one of decline. The case is more perplexing for senators. The second panel shows that, among those who were only ever senators, there was an increase in the proportion of members with a pre-relative (either in the Senate prior to their appointment or elected to the House) until the mid-1940s. As senators were appointed for life until 1965, there seems to have been fewer opportunities for new appointments and potential shifts in power dynamics in this chamber.

Nevertheless, that the Senate has had a greater share of members with a pre- and post-relative also indicates that the chamber shares characteristics closer to the pre-reform House of Lords in the United Kingdom (Von Coppenolle 2017). While it is true that the Canadian Senate was perceived to be an elitist institution, it was decidedly designed not to be a hereditary institution. And while it is true that the family linkages identified in the Senate are not inherited parent–child appointments, that there was clearly some form of family linkage within the Senate offers more evidence that appointments to this upper chamber were not truly made independent of hereditary/familial considerations (VandenBeukel et al. 2021).

Modelling dynastic links

Table 1 reports the results of several analyses aimed at determining whether having a pre-relative has an influence on having a post-relative. In other words, these models estimate whether being part of a dynasty (having a pre-relative) increases its likelihood of continuing in the future.

Table 1. Post-relative on Pre-relative

Notes: *p < 0.10; **p < 0.05; ***p < 0.01. Legislators born after 1950 and those who died in office have been removed. Models 3 and 4 include only those who were only ever elected to the Commons. OLS regression.

The outcome variable in Table 1 takes the value of 1 if the legislator has a post-relative and 0 otherwise. Pre-relative, our main predictor of interest, is a dichotomous variable that takes the value of 1 if the legislator had a family member elected before them and 0 otherwise. We also include a host of covariates measuring occupation as a series of dichotomous variables comparing those with a law background (lawyer, judge, notary, solicitor etc.), those with a business background and those who had an agriculture background.¹

We further include a control for whether the legislator was born outside the province they represent (outsider) and for women legislators. Models 1 and 3 are bivariate linear probability regression models without fixed effects and models 2 and 4 are the fully specified models with parliament and province fixed effects respectively. In Table 1, we restrict our analyses to those legislators born before 1950 as we believe that those born after this year have not yet had enough time to see the continuation of a dynasty; we also exclude those legislators that died in office (for a similar approach, see Dal Bó et al. 2009).

Across all model specifications, the pre-relative coefficient is statistically significant and positive. Note that in the fully specified models, the impact of having a pre-relative is substantially different. Among MPs born before 1950, having a pre-relative elected to the House of Commons makes one nearly 12% more likely to continue the dynasty. In contrast, having a pre-relative makes one 6% more likely to have a post-relative elected to the same provincial chamber. These results suggest that having a family elected prior to one's entrance significantly increases the likelihood of having a future family member elected to the legislator's respective chamber.

We now turn to the question of measuring the electoral advantage of having an ancestor in the legislature. Table 2 reports four different models to measure this effect in both provincial and federal legislatures by following the same structure as in Table 1. Some caution is warranted here in interpreting the results. Our dataset only includes individuals who were successfully elected to a legislature. We do not have data for candidates that were part of a family dynasty but that were never elected. The results in Table 2 therefore require some qualification: in this analysis, we limit the sample only to a legislator's first contested election by comparing those who had a pre-relative to those who do not in order to determine whether dynasts inherit an initial electoral advantage before other advantages (such as incumbency) occur.

Table 2. Vote Share on Pre-Relative

Notes: *p < 0.10; **p < 0.05; ***p < 0.01. OLS regression. Data begins in 1867 and the total sample used.

Table 2 indicates that there is an electoral benefit awarded to members of a political dynasty – but the reward is rather limited (only an additional 1.5% of the vote at first successful election) and only at the provincial level. Table A.1 in Appendix A in the Supplementary Material replicates Table 2 by looking only at the first successful election attempt. Table A.1 confirms the findings in Table 2 – there is a small electoral advantage for future legislators but only at the provincial level.

We note, however, that vote share only tells one part of the story related to career success. Appointments to the cabinet, at the discretion of the premier or prime minister, represent a considerable promotion in a legislator's career, a process that has never been purely meritocratic. Historically, selection to the federal cabinet reflected linguistic, regional, and religious cleavages to ensure that ministers broadly represented the electorate (Everitt and Lewis 2020). In the contemporary era, new characteristics, such as gender and ethnic identities, take a more prominent role (Kerby 2009; Kerby and Snagovsky 2021). Still historical biases and family linkages seem to persist. Indeed, this is precisely what Van Coppenolle (2017) finds in the UK and in their comparative analysis (Van Coppenolle and Smith 2022) – that members of dynasties are more likely to be appointed to cabinet.

There is descriptive evidence of a dynastic benefit at the executive level. Of Canada's 23 prime ministers, 13.6% are a member of a dynasty and 31.8% had a family member elected after them. Of the 120 premiers since 1867, 24.1% are part of a dynasty and 26.6% had a post-relative elected to their respective provincial legislature. These figures are more startling at the cabinet-level: 18.8% and 31.2% of cabinet members are part of a dynasty at the federal and provincial level respectively; whereas 23% and 30.5% have had a post-relative elected to office federally and provincially. Yet, in Canada, cabinet is not the only method for career advancement. Until Justin Trudeau introduced a reform for the non-partisan selection of senators in 2015, appointments to this chamber were usually seen as patronage opportunities for the governing party.

Table 3 estimates a model to determine if being part of a dynasty provides career advancement prospects in two different specifications. On the one hand, the dependent variable in models 1 to 4 represents whether a legislator was, at any time in their career, appointed to cabinet. On the other hand, the dependent variable in models 5 and 6 takes the form of 1 if the legislator was appointed to the Senate and 0 otherwise.

Table 3. Career Advancement on Pre-Relative

Notes: *p < 0.10; **p < 0.05; ***p < 0.01. Models 1 through 4 regress whether the legislator was ever a cabinet member on whether the member had a pre-relative elected to the same chamber. Models 5 and 6 regress whether a senator had a pre-relative elected to either the House of Commons or appointed to the Senate. Data are post-1867 in both arenas and those that died in office have been removed. Because we are interested in the effect of a pre-relation on cabinet selection, we include legislators born after 1950 in order to maximize the entire sample. OLS regression.

Being part of a dynasty appears to offer considerable career advantages for dynastic legislators. Members of a family dynasty are 6.6% more likely to be appointed to the cabinet provincially and 7.2% more likely to be appointed to the federal cabinet. Like cabinet selection, model 6 also confirms that dynasty members are 4.3% more likely to be appointed to the Senate, when compared to those without a previous relative in parliament. This raises further questions about the Senate as a non-hereditary body. Certainly, Senate appointments were not handed down within the family as they were under the peerage system in the UK, but dynastic linkage strongly suggests that Senate appointments are nevertheless associated with family dynamics. This last result echoes Van Coppenolle's (2022) findings regarding the political dynamics in the Netherlands, where changes to the Dutch lower chamber to increase electoral competition subsequently decreased the likelihood that elected legislators would promote their relatives to the upper chamber.

The results of Table 3 also suggest that being a member of a dynasty does not reward the legislators with an electoral advantage. Rather, it appears that the career advantages of family connections emerge once a legislator is elected to the legislature. From this perspective, family ties might not mean much to voters in elections, but they do seem to matter a great deal for those in a position of power who decide whom to promote in the legislative arena.

Finally, Table 4 presents the estimations of a series of regressions, modelling the probability of having a post-relative elected to office based on whether the legislator was a long-term legislator and a series of controls. Models 1 through 3 exclusively look at the provincial level but model 3 restricts the analysis to post-Confederation provincial elections as a more apt comparison to the federal level.

Table 4. Post-relative on Tenure

Notes: *p < 0.10; **p < 0.05; ***p < 0.01. All models are linear probability models that estimate the likelihood of having a post-relative elected in the same legislature. Sample limited to those born before 1950 and to those that did not die in office. Model 3 limits the provincial sample to elections since 1867. Models 4 and 5 are limited to those who only served as MPs. OLS regression.

The results here suggest that tenure has an effect on having a post-relative elected to office. Models 1 and 4 include those legislators with a pre-relative and shows that, in provincial legislatures, long-term legislators are roughly 2.5% more likely to have a post-relative in the assembly. Models 2 and 5 restrict the analysis to legislators without a previously elected relative, ostensibly measuring the start of a dynasty. The results in models 2 and 5 are similar to 1 and 3 – the effects are strong, significant and roughly equal in strength both at the provincial and federal election levels. Note, as well, that the effects are also confirmed in model 3 when provincial elections are constrained to the post-Confederation era.

Tenure as a cause of dynastic formation?

Dal Bó et al. (2009) raise a concern that the presence of future elected relatives might be the result of unobserved family characteristics. They address this problem by adopting a regression discontinuity design. The key assumption of this approach is that narrowly decided elections ‘randomly’ assign politicians to either the winner or the loser groups, irrespective of family background. In this context, winning with a narrow margin implies that both the winner and loser should share similar unobservable traits, such as political influence. Thus, for narrow margin winners, any disparity in the likelihood of family members later holding office should be caused by the election results. The analysis implies that a longer time in power – rather than shared family attributes – raises the chances of establishing a political dynasty by driving both lengthy tenures and future dynastic achievements.

Below, we use the same logic put forward by Dal Bó et al. (2009) and Van Coppenolle (2017) by replicating their instrumental variable regression discontinuity design to draw a comparison between legislators who marginally lost their first re-election and those who just managed to secure their first re-election victory. We first construct the ‘long-term’ marker indicating whether a legislator served at least two terms in office; in our data, this corresponds to 61.3% and 58.2% of MPs and provincial legislators born before 1950. As such, we observe the effects of dynasty formation on narrowly re-elected first-term incumbents (Dal Bó et al. 2009; Van Coppenolle 2017) rather than on narrowly elected first-time candidates (Fiva and Smith 2018; Querubin 2016).

Figure 3 plots the discontinuity at a 5% vote window in provincial (first panel) and federal (second panel) elections among legislators' first re-election attempt. Values to the left of the vertical line indicate the percentage by which the legislator lost, and values to the right indicate the percentage by which the legislator successfully won re-election. The y-axis is the proportion of those with a post-relative elected and the x-axis is the margin of loss/victory. In both arenas, there is an increasing trend among losers such that there is a higher proportion of post-relatives elected to office as the margin of loss approaches zero. But when the margin of victory approaches 5%, we find that the share of post-relatives decreases provincially and increases federally. Evidently, there is a considerable share of legislators who narrowly lost their re-election who still had family members elected after them.

Figure 3. Discontinuity at 5% Window for Legislators' First Re-election Attempt in Provincial and Federal Elections

Note: Died in office removed and limited to those born before 1950.

Table 5 estimates the causal effects using the regression discontinuity method. Like Van Coppenolle's and Dal Bó et al.'s approaches, we created a ‘WIN’ variable that takes the form of 1 if the legislator won their first re-election and 0 otherwise. The first stage, the top panel of Table 5, regresses our long-term variable on our win indictor across different specifications in both provincial and provincial elections. We limit our analysis to ±2.5% and ±5% vote windows, altering the presence of dynastic legislators. Across all specifications, the ‘WIN’ coefficient is positive and significant.

Table 5. Results of Instrumental Variable Regression Discontinuities across both Arenas and Differing Vote Windows

Notes: *p < 0.10; **p < 0.05; ***p < 0.01. ‘WIN’ serves as the instrument and is a dummy variable that takes the form of 1 if the legislator won their first re-election attempt and 0 otherwise. The first panel is the first stage regression and the second panel is the second stage. Controls (as identified in Tables 1–4) are included. Sample restricted to legislators born before 1950. Those who died in office have been removed.

The second panel in Table 5 regresses the post-relative dependent variable on long-term, which is instrumentalized through the ‘WIN’ variable from the first stage. Across all specifications, in both provincial and federal elections, the long-term variable is statistically insignificant. As such, similar to Van Coppenolle in the case of the UK, we find no causal relationship between a legislator's tenure and dynastic linkages in two ways. First, that the long-term coefficient is insignificant in models where pre-relatives are included indicates that legislator tenure does not induce the election of a post-relative, even if the legislator had a relatively long tenure in office. From this perspective, then, the descendants of a long-term legislator are just as likely to be elected as are those legislators without a long-term pre-relative. Second, and most importantly, that the long-term coefficient is insignificant for models where legislators with pre-relatives are excluded shows that dynasty formation is likewise not affected by tenure length. As a result, there is no intrinsic benefit associated with tenure length and the formation of a dynasty at either the provincial or federal level in Canada.

Conclusion

This study contributed to the literature on political dynasties by examining family influences in Canadian legislatures, both before and after Confederation. Our results confirmed that there is a small electoral benefit for dynastic legislators at the provincial level. We also found that being a dynastic legislator offers clear career advantages in terms of cabinet appointment and being nominated to the Canadian Senate. Finally, unlike the US case, but mirroring the UK experience, our last set of analyses suggest that there is no causal relationship between tenure length and the establishment of enduring political dynasties in Canada.

At the beginning of this article, we argued that the Canadian case could increase our understanding of how institutions shape the legacies of legislative careers by leveraging the institutional differences between the provinces and the federal government. In their most comprehensive study to date, Van Coppenolle and Smith (2022) observed a decline in political dynasties as advanced democracies matured. Exceptions to this trend were found in Japan and Ireland, where political dynasties persist due to the influence of weaker political parties and highly personalized elections.

Our results confirm, however, that the enduring presence of decentralized party systems and candidate-centred elections was not sufficient to maintain the influence of dynastic politics during the modern era in Canada (Cross 2007; Johnston 2017), even in the smaller legislatures of Nova Scotia and New Brunswick. It seems that the other factors identified in the literature to explain the weakening of political dynasties in developed democracies, such as political modernization, institutional changes and electoral reforms, are more likely to account for the decline observed in this case.

Unfortunately, the concept of modernization is difficult to define because it encompasses several related dimensions, which include not only demographic shifts, immigration patterns, the expansion of the franchise and the rise in education levels, but also urbanization, industrialization and the proliferation of modern mass media. When we concentrate on the first dimension, demographic shifts, it becomes apparent that nations such as Japan, Ireland or subnational units such as Quebec, which have been historically more homogeneous, could promote stronger familial ties in their political systems.

Another aspect related to elections extends to the selection of candidates. Benny Geys and Daniel Smith (2017) infer that decentralized candidate selection mechanisms might benefit individuals who are members of a dynasty. Given the ‘franchise’ nature of the Canadian political party, in which the local constituency is given primacy over candidate selection/election (Carty 2002, 2004), we should therefore consider the possibility that family influence now plays a role at different levels of the electoral process, such as during primary elections (where name recognition might be more important), or in fundraising and party contributions.

The fact that we found the most significant effects of political dynasties in the Senate and cabinet appointments highlights the importance of other institutional factors on the persistence of family influence in politics. However, to gain a more comprehensive understanding of the impact of political dynasties, future research should explore other aspects of legislative assemblies as well, including committee chairmanships, bill sponsorship and participation in debates (see Prihatini and Halimatusa'diyah 2022). Furthermore, researchers should explore the possibility that the influence of dynasties extends beyond the legislative arena, providing family members with greater access to civil service jobs and non-political career advantages.

Finally, our study offers additional evidence to confirm that political dynasties operate differently in parliamentary and presidential systems. In the US, an electoral advantage is observed at the legislative level. Conversely, in the UK and Canada, two parliamentary systems with appointed upper chambers, the influence of lineage appears to be much more prominent. Still, the American case warrants further investigation. Is it possible that family connections matter more to the appointment of cabinet members in the US as well? Likewise, Senate connections are more likely to matter in the US when compared to the House, but can this be explained by appointments made prior to the 17th Amendment, which introduced the direct election of senators? Perhaps these aspects could be further explored in US State and colonial Assemblies (Chaffey 1970; Polsby 1968; Squire 2017), where varying levels of professionalization and family dynamics could have played a significant role in maintaining political dynasties.

Ultimately, our study highlights the importance of political fairness in a more established democracy such as Canada. If family lineage promotes access to positions of power, does it reduce career opportunities for competent and deserving individuals who are not members of political dynasties? Our results confirm that this might be the case for cabinet and Senate appointments in this country, where family ties continue to be important contributors to structural inequalities.