Barriers to LGBTQ+ Representation

As more LGBTQ+ candidates have run in recent years, scholars have turned their attention to identifying barriers these candidates face. One strand of work considers barriers posed by voters. Scholars suggest that negative public opinion (Everitt and Camp Reference Everitt and Camp2014), public scrutiny (Wagner Reference Wagner2021), and/or stereotyping (Wagner Reference Wagner2019) can deter LGBTQ+ people from running. Other scholars identify voter bias as a barrier to at least some LGBTQ+ candidates (see Golebiowska Reference Golebiowska2003; Herrick and Thomas Reference Herrick, Thomas, Ellen and Tadlock1999, Reference Herrick and Thomas2001 for negative reactions to out lesbian and gay candidates; but see Everitt and Horvath Reference Everitt and Horvath2021; Loepp and Redman Reference Loepp and Redman2022; Magni and Reynolds Reference Magni and Reynolds2018, Reference Magni and Reynolds2021 for the lessening of penalties facing some LGBTQ+ candidates in some contexts; see Haider-Markel et al. Reference Haider-Markel, Miller, Flores, Lewis, Tadlock and Taylor2017; Jones et al. Reference Jones, Brewer, Young, Lambe and Hoffman2018; Jones and Brewer Reference Jones and Brewer2019; Magni and Reynolds Reference Magni and Reynolds2021 for voter biases against trans candidates).

A second strand of work explores barriers posed by parties. Ashe (Reference Ashe, Tremblay and Everitt2020) examines whether party gatekeepers prefer straight cis candidates but finds this is not the case. There is evidence, however, that parties pose barriers to LGBTQ+ candidates by running them in districts they are unlikely to win (Baisley and Albaugh Reference Baisley and Albaugh2025; Everitt and Tremblay Reference Everitt and Tremblay2023; Lapointe, Ferland, and Turgeon Reference Lapointe, Ferland and Turgeon2024; but for contrary evidence outside of Canada, see Grahn Reference Grahn2024; Haider-Markel Reference Haider-Markel2010). Incumbency could be another party barrier to LGBTQ+ candidates given the low number of out LGBTQ+ elected officials and party insiders’ preference to renominate incumbents (Albaugh Reference Albaugh2022).

We add to research on barriers facing LGBTQ+ candidates by examining another potential barrier: political finance. If LGBTQ+ candidates receive less money than their straight cis counterparts, this could help explain why they are less likely to win in some countries. Although there is some evidence that LGBTQ+ candidates in the US are more successful fundraisers (Magni and Reynolds Reference Magni and Reynolds2022), the same pattern might not hold in other contexts, especially in countries like Canada where an increase in the number of LGBTQ+ candidates running has not translated into more LGBTQ+ elected officials. In asking whether LGBTQ+ candidates receive less money, we take inspiration from work on campaign finance and women candidates.

Identity and Campaign Finance

Much of the research about identity and campaign finance has focused on women. Early work on gender and campaign finance explored whether financing was a barrier for women candidates. For instance, Uhlaner and Schlozman (Reference Uhlaner and Schlozman1986) analyzed whether major-party women candidates for the United States House of Representatives in the 1980 election raised less money than men candidates. The authors found that women raised less money than men but that this was due to factors — such as incumbency or committee chairmanships — that donors valued but that were more common among men. As Buckley and Mariana (2023) explain, in many contexts women continue to be disadvantaged when it comes to factors associated with raising money.

Over time, research began to show that fundraising and spending differences between candidates did not fall along gender lines, at least in some contexts (for evidence from the US, see Burrell Reference Burrell1985, Reference Burrell2014; Herrnson, Lay, and Stokes Reference Herrnson, Lay and Stokes2003; Hogan Reference Hogan2007; Thomsen and Swers Reference Thomsen and Swers2017; for evidence from Canada, see Garnett et al. Reference Garnett, Pruysers, Young and Cross2022; Tolley, Besco, and Sevi Reference Tolley, Besco and Sevi2022; but for evidence of women facing fundraising disadvantages in other contexts, see Cigane and Ohman Reference Cigane and Ohman2014; Piscopo et al. Reference Piscopo, Hinojosa, Thomas and Siavelis2022; Sidhu and Meena Reference Sidhu and Meena2007; Tovar Reference Tovar2007). Yet scholars largely agree that the playing field is still not even for women when it comes to political finance. Instead, researchers shifted their attention to other gendered differences in financing. They note, for example, that the costs of running for office can deter women from running in the first place (Burrell Reference Burrell1985; Murray Reference Murray2023). There is some evidence from the US that women have to spend more to achieve the same results as men (Herrnson, Hunt, and Kettler Reference Herrnson, Hunt and Kettler2022). Others show that even if women candidates raise the same amount or more than men, they do so by devoting greater effort to fundraising (Burrell Reference Burrell2003) and undertaking different fundraising activities (Crespin and Deitz Reference Crespin and Deitz2010). Women candidates, for example, are more likely to receive money from women, who are less likely to donate and donate in smaller amounts than men (McMahon, Sayers, and Alcantara Reference McMahon, Sayers and Alcantara2023; Tolley, Besco, and Sevi Reference Tolley, Besco and Sevi2022). In some contexts, women candidates make up for this by getting other funds. In the US, women’s donor networks, such as EMILY’s List and the WISH List, help women candidates and sometimes even provide them a fundraising advantage over men (Burrell Reference Burrell2003, Reference Burrell2014). In the Canadian context, Currie-Wood and Pruysers (Reference Currie-Wood and Pruysers2023) show that central parties can and sometimes do support women candidates with additional campaign funds but this has not been a consistent priority.

There has also been some research on race and campaign finance. Cho (Reference Cho2002), for example, finds that Asian Americans tend to donate to Asian American candidates. In the US, Grumbach and Sahn (Reference Grumbach and Sahn2020) show that the presence of an Asian, Black, or Latino nominee increases the proportion of Asian, Black, or Latino contributions in the general election. In Canada, Besco and Tolley (Reference Besco and Tolley2022) find that ethnic minorities are less likely to donate than other Canadians with the exception of South Asian Canadians who donate at very high rates.

Our work draws on and contributes to work on identity and campaign finance in three ways. First, most research on identity and campaign finance has focused on gender and race, and at times the intersection of gender and race (Grumbach, Sahn, and Staszak Reference Grumbach, Sahn and Staszak2022; Sanbonmatsu Reference Sanbonmatsu2025; Sorensen and Chen Reference Sorensen and Chen2022) or gender and class (Murray Reference Murray2023). We consider other identities — LGBTQ+ identities — and look at a different intersection: the intersection of gender and sexuality. However, in doing so, we take inspiration from early work on gender and campaign finance by examining whether there are financing gaps (in this case between straight cis and LGBTQ+ candidates). Like that early work, we consider whether these gaps persist after controlling for factors associated with fundraising (Uhlaner and Schlozman Reference Uhlaner and Schlozman1986).

Second, we expand work about gender and campaign finance by considering TNB candidates. This more comprehensive analysis of gender has become possible and more important in recent years as more of these candidates run. Our findings suggest that disaggregated intersectional analyses are important because straight cis women likely drive findings about the lack of gender financing gaps.

Finally, much of the literature on campaign finance has focused on the US. Yet the US is an outlier in many respects, including in how candidate-centered its elections are, the prevalence of private forms of political financing, and the existence of political action committees (PACs) (Feo, Fiorelli, and Piccio Reference Feo, Fiorelli and Piccio2023). As a result, scholars have emphasized the need for more work from other contexts, particularly from countries with different political systems and institutions (Besco and Tolley Reference Besco and Tolley2022; Feo, Fiorelli, and Piccio Reference Feo, Fiorelli and Piccio2023). We take up that call in this paper by focusing on Canada.

Why Might LGBTQ+ Candidates Receive Less Money?

In this section, we discuss why LGBTQ+ candidates might receive less money. In doing so, we focus on inequalities that might arise during fundraising for general elections. Given that, we do not focus directly on inequalities that may arise in political recruitment, but we discuss potential downstream consequences of inequalities in political recruitment as they relate to financing campaigns.

The simplest reason to expect LGBTQ+ candidates to receive less money is that they often run for parties that raise less money. Typically, right-of-center candidates have a fundraising advantage as they draw support from wealthy individual donors and/or corporate donors (Cross, Pruysers, and Currie-Wood Reference Cross, Pruysers and Currie-Wood2022; Feo, Fiorelli, and Piccio Reference Feo, Fiorelli and Piccio2023). However, LGBTQ+ candidates, especially from intersectionally marginalized subgroups, tend to run for left parties, particularly those that have historically taken more pro-LGBTQ+ policy positions (Everitt and Camp Reference Everitt and Camp2014; Everitt and Tremblay Reference Everitt and Tremblay2023; Everitt, Tremblay, and Wagner Reference Everitt, Tremblay and Wagner2019; Lapointe, Ferland, and Turgeon Reference Lapointe, Ferland and Turgeon2024). As a result, we expect LGBTQ+ candidates, especially those from intersectionally marginalized subgroups, to receive less money than their straight cis counterparts.

Party gatekeepers and donors may direct less money to LGBTQ+ candidates for reasons other than their gender and sexuality. These actors often want to ensure that contributions matter for election outcomes by helping the party win seats. As a result, they may funnel more money to those they think have a better chance of winning based on incumbency, district competitiveness, and past political experience (Buckley and Mariani Reference Buckley and Mariani2023; Feo, Fiorelli, and Piccio Reference Feo, Fiorelli and Piccio2023; Krasno, Green, and Cowden Reference Krasno, Green and Cowden1994; Uhlaner and Schlozman Reference Uhlaner and Schlozman1986). Yet we know that women and LGBTQ+ people are less likely to be incumbents, more likely to run in seats the party is unlikely to win, and less likely to have past political experience (Baisley and Albaugh Reference Baisley and AlbaughForthcoming; Baisley and Albaugh Reference Baisley and Albaugh2025; Buckley and Mariani Reference Buckley and Mariani2023; Everitt, Tremblay, and Wagner Reference Everitt, Tremblay and Wagner2019; Lapointe, Ferland, and Turgeon Reference Lapointe, Ferland and Turgeon2024; Uhlaner and Schlozman Reference Uhlaner and Schlozman1986). Intersectionally marginalized subgroups of LGBTQ+ people, such as queer women or TNB people, are even less likely to be incumbents or run in “winnable” districts (Everitt and Tremblay Reference Everitt and Tremblay2023). As a result, we examine whether financing gaps between straight cis and LGBTQ+ candidates and across LGBTQ+ subgroups persist after adjusting for other variables on which these candidates are disadvantaged. If these financing gaps go away after controlling for these factors, it does not mean that LGBTQ+ candidates (or particular subgroups) are not disadvantaged. Instead, it suggests that addressing other factors, such as incumbency, district competitiveness, or past political experience, would help address financing gaps.

Beyond these inequalities in factors typically associated with fundraising, we see reasons to expect LGBTQ+ candidates to be disadvantaged. First, as scholars increasingly argue, politics is shaped by cisheteronormativity — or the idea that cisgender people who are heterosexual are the norm and that their relationships are considered natural and superior (Ashe Reference Ashe, Tremblay and Everitt2020; Everitt and Tremblay Reference Everitt and Tremblay2023; Raney Reference Raney, Tremblay and Everitt2020; Schotel and Mügge Reference Schotel and Mügge2024). As these scholars note, LGBTQ+ candidates (especially TNB candidates) have long been excluded from politics. Even as these candidates have become more visible in formal politics in recent years, scholars have often described them as “space invaders” whose presence highlights and challenges the norms and boundaries of politics (Everitt and Tremblay Reference Everitt and Tremblay2023; Schotel and Mügge Reference Schotel and Mügge2024). Due to cisheteronormativity, party gatekeepers and donors may see LGBTQ+ candidates as less conventional and electable and therefore engage in “strategic discrimination” when donating or allocating funds (for more on strategic discrimination, see Bateson Reference Bateson2020). Party gatekeepers and donors may also have concerns about LGBTQ+ candidates being “single-issue” candidates and therefore offer less financial support (for more on fears about single-issue candidates, see Tremblay Reference Tremblay2022).

LGBTQ+ candidates may also be disadvantaged in fundraising due to wage gaps. Research on other marginalized groups finds affinity effects in campaign finance. More specifically, scholars note that women are more likely to support women candidates and that racial minorities are more likely to support racial minority candidates (see, e.g., Besco and Tolley Reference Besco and Tolley2022; Grumbach and Sahn Reference Grumbach and Sahn2020; Tolley, Besco, and Sevi Reference Tolley, Besco and Sevi2022). This can put women and racial minority candidates at a disadvantage because of wage gaps, with women and racial minorities having less money to donate due to earning less than men and white people, respectively (Grumbach and Sahn Reference Grumbach and Sahn2020; Tolley, Besco, and Sevi Reference Tolley, Besco and Sevi2022). One might expect affinity effects with LGBTQ+ donors being more likely to support LGBTQ+ candidates. This may leave LGBTQ+ candidates at a disadvantage because LGBTQ+ people earn less than straight cis people (Appiah et al. Reference Appiah, Brennan, Halpenny, Pakula and Waite2021; Carpenter et al. Reference Carpenter, Feir, Pendakur and Warman2024; Carpenter, Goodman, and Lee Reference Carpenter, Goodman and Lee2024; for a broader discussion of employment outcomes, see Waite, Pajovic, and Denier Reference Waite, Pajovic and Denier2020). These wage gaps are larger for some subgroups than others. For example, Appiah et al. (Reference Appiah, Brennan, Halpenny, Pakula and Waite2021) find that heterosexual men earned the most, followed by gay men, lesbian women, bisexual men, and bisexual women. Research that includes TNB people tends to find very large wage gaps (Carpenter et al. Reference Carpenter, Feir, Pendakur and Warman2024; Carpenter, Goodman, and Lee Reference Carpenter, Goodman and Lee2024). In addition, we note that young candidates often face biases in politics (Gélix and Chassé Reference Gélix and Chassé2024; Stockemer, Gallant, and Tolley Reference Stockemer, Gallant and Tolley2024). Given that TNB people tend to be young (Statistics Canada 2022), they may also be disadvantaged in financing due to their age.

Due to broader systems of marginalization and research on wage gaps, we expect differences in financing across LGBTQ+ subgroups. We focus our discussion on four categories we can examine with our data:

• • Queer cis men: Among LGBTQ+ subgroups, we expect queer cis men to be the most advantaged, in part because they are men. Although they are still disadvantaged by cisheteronormativity in politics, they were often the first of LGBTQ+ subgroups to enter formal politics (Everitt and Camp Reference Everitt and Camp2014; Rayside Reference Rayside1998). As a result, even if they are not the most “conventional” or “traditional” candidates, party gatekeepers and donors may have fewer concerns at this point about their electability. However, gay and bisexual cis men tend to earn less than straight cis men (Appiah et al. Reference Appiah, Brennan, Halpenny, Pakula and Waite2021).

• • Queer cis women: Among LGBTQ+ subgroups, we expect queer cis women to be the next most (dis)advantaged. Gender biases in politics likely disadvantage queer cis women compared to queer cis men. In addition, queer cis women typically entered formal politics later than queer cis men, and there are still fewer of them in elected office (Everitt and Tremblay Reference Everitt and Tremblay2023; Reynolds Reference Reynolds2013). As a result, we expect party gatekeepers and donors to see queer women as less “conventional” or “traditional” as candidates. Wage gaps pose more problems for queer cis women than queer cis men (Appiah et al. Reference Appiah, Brennan, Halpenny, Pakula and Waite2021), which also shapes our expectation that queer cis women will be more disadvantaged than queer cis men.

• • TNB candidates: We expect TNB candidates to be even more disadvantaged than queer cis men and queer cis women due to cisnormativity. In addition, TNB candidates typically entered formal politics even later than queer cis men and queer cis women, marking them as even less conventional candidates (Schotel and Mügge Reference Schotel and Mügge2024). There are still many contexts, such as federal politics in Canada, where there has yet to be an out trans or nonbinary elected representative (Everitt and Tremblay Reference Everitt and Tremblay2023). TNB people face significant wage gaps (Carpenter et al. Reference Carpenter, Feir, Pendakur and Warman2024; Carpenter, Goodman, and Lee Reference Carpenter, Goodman and Lee2024), shaping our expectation that TNB candidates will be disadvantaged in fundraising. In addition, past work suggests that TNB candidates face voter biases (Haider-Markel et al. Reference Haider-Markel, Miller, Flores, Lewis, Tadlock and Taylor2017; Jones et al. Reference Jones, Brewer, Young, Lambe and Hoffman2018; Jones and Brewer Reference Jones and Brewer2019; Magni and Reynolds Reference Magni and Reynolds2021). As a result, party gatekeepers and donors may see TNB candidates as particularly unelectable and therefore not direct funds to them.

• • Two Spirit candidates: We have few expectations about Two Spirit candidates given limited research about Indigenous candidates and financing and inequalities facing Two Spirit individuals in general (for a discussion of data limitations regarding Two Spirit people in wage gaps, see Appiah et al. Reference Appiah, Brennan, Halpenny, Pakula and Waite2021). However, they may be disadvantaged due to being both Indigenous and part of the LGBTQ+ community.

We also see reasons why LGBTQ+ candidates, including those from intersectionally marginalized subgroups, may be disadvantaged within certain parties, particularly right-of-center parties. In many countries, right-of-center parties have anti-LGBTQ+ voters as part of their base, but they may need to appeal to voters who are not anti-LGBTQ+ to win elections. Although right-of-center parties may nominate LGBTQ+ candidates strategically to signal to voters that they are not anti-LGBTQ+ (for a similar logic for women, see Weeks et al. Reference Weeks, Meguid, Kittilson and Coffé2023), there are a few reasons to expect party gatekeepers and donors in these parties not to back LGBTQ+ candidates financially. First, if these party gatekeepers and donors are themselves socially conservative or anti-LGBTQ+, they may be unwilling to finance LGBTQ+ candidates. Second, right-of-center parties may be especially concerned that their own voters may be unwilling to vote for LGBTQ+ candidates and not view LGBTQ+ candidates as electorally competitive enough to be worth backing financially. Finally, right-of-center parties are typically resistant to equity policies within their own organizations, which makes them (particularly) unlikely to identify, mitigate, or counter biases facing LGBTQ+ candidates, including in finance. All these factors lead us to suspect that LGBTQ+ candidates are likely to be disadvantaged within right-of-center parties, such as the Conservative Party of Canada.

The Canadian Case

To compare the political financing of straight cis versus LGBTQ+ candidates, we turn to the Canadian case. This case is useful for such a study given that many LGBTQ+ candidates have run in recent Canadian federal elections. In 2019, Canada experienced a “rainbow wave,” with a record number of LGBTQ+ candidates running for office (Lenti Reference Lenti2019). Many LGBTQ+ candidates ran again in 2021. The “rainbow wave” also ushered in more diversity in terms of LGBTQ+ candidates, Baisley and Albaugh Reference Baisley and AlbaughForthcoming CJPS, which presents an opportunity to examine differences in political financing among differently situated LGBTQ+ candidates (for example, cis versus TNB candidates, queer cis men versus queer cis women, etc.).

We note a few major institutional features of the Canadian case that may shape political financing. First, Canada has a Westminster parliamentary system with very high party discipline, making it unlikely individual candidates’ ideological positions matter for political financing. As a result, ideology is likely to play a role in shaping which parties donors support, as opposed to which individual candidates (for the importance of party to political donations in Canada, see Garnett et al. Reference Garnett, Pruysers, Young and Cross2022; but for the importance of ideology in the US, see Barber, Canes-Wrone, and Thrower Reference Barber, Canes-Wrone and Thrower2017; Barber Reference Barber2016). Second, Canada’s House of Commons uses a single-member plurality electoral system, similar to the US House of Representatives or the UK House of Commons, making district-level electoral contexts highly salient for party gatekeepers and donors. Third, Canada has a multiparty electoral system in which five parties win seats: the Liberals, the Conservatives, the NDP, the Bloc québecois (BQ), and the Greens. Prior to the 2015 election, the Conservatives formed a majority government and the NDP formed the Official Opposition federally for the first time in Canadian history. After the 2015 election, however, the Liberals formed a majority government with the Conservatives as the Official Opposition. The Liberals were reduced to minority government status as a result of the 2019 election, and the Conservatives remained as the Official Opposition. The Liberals secured a second minority after a snap election in 2021.

Comparatively speaking, Canada’s campaign finance regime is fairly restrictive (International Institute for Democracy and Electoral Assistance 2025). As Garnett et al. (Reference Garnett, Pruysers, Young and Cross2022) explain, “[t]he only legal sources of income for Canadian political parties and candidates are publicly funded election expense reimbursements and contributions from individual donors” (422). Since 2003, union and corporate donations have been prohibited in Canada, making individual donations central to political finance in Canada. However, there are limits on individual donations. In 2015 (the start of the period under study), individuals could donate a maximum of $1,500 per calendar year to each of the registered parties and another $1,500 per calendar year — in aggregate — to the registered electoral district associations (EDAs), nomination contestants, and candidates of each of the registered political parties.Footnote ⁴ The $1,500 limits increase by $25 each year and include both monetary and non-monetary (or “in-kind” contributions) (Elections Canada n.d.). There are also spending limits, which are based on the year, the length of the campaign, and the population of the district. Given that spending limits are based in part on the length of the campaign period, it is worth noting that there was some variation over these elections. By Canadian standards, the 2015 campaign (of 11 weeks) was long. The campaigns in 2019 (five and a half weeks) and 2021 (five weeks) were more typical. Third-party spending is also tightly regulated, and Canada does not have PACs. Many of these features of Canada’s political finance regime are common among post-industrial democracies, though some post-industrial countries do not have the same restrictiveness or give more public funding to parties.

The political finance system in Canada is quite party-centric. Although individuals can contribute directly to local candidates, most donations are made to registered parties (central or national party organizations) or to EDAs (local party organizations). Central and local party organizations can then transfer funds to local candidates (sometimes through complex pathways, such as from an individual contributor to the registered party to the EDA to the local candidate). In our data, for example, party transfers from either the central or local organizations account for the lion’s share of funding that local candidates receive: the median candidate who receives any money obtains 89% of their total fundraising from transfers from registered parties or EDAs, and 88% of candidates who receive any money obtain at least 50% of their total fundraising from transfers from registered parties or EDAs. Canadian parties, therefore, are important intermediaries for campaign finance, which presents an opportunity to examine how they allocate resources (see also Currie-Wood and Pruysers Reference Currie-Wood and Pruysers2023). In this way, Canada is more similar than the US to many other countries: work on countries other than the US has emphasized the important role of political parties in political finance, including in Chile (Piscopo et al. Reference Piscopo, Hinojosa, Thomas and Siavelis2022), Ireland (Buckley and Mariani Reference Buckley and Mariani2023), South Korea (Shin and Kwon Reference Shin and Kwon2023), Cabo Verde and Ghana (Wang, Muriaas, and Bauer Reference Wang, Muriaas and Bauer2023).

Canada has two well-documented patterns that play important roles in shaping LGBTQ+ representation. First, less electorally successful left parties, such as the NDP and the Greens, tend to nominate LGBTQ+ candidates in much larger numbers than the two major parties, the Liberals and the Conservatives. This pattern contrasts with other countries where major parties tend to nominate LGBTQ+ candidates who then have good chances of winning their seats, such as the US (Democrats) and the UK (Labour and Conservatives). Second, parties nominate LGBTQ+ candidates in less winnable districts, unlike the Democrats in the US (Haider-Markel Reference Haider-Markel2010) and Swedish parties (Grahn Reference Grahn2024). When we examine financing gaps without controls, we expect these factors to put LGBTQ+ candidates at a greater disadvantage in Canada than in some other countries.

Many Canadian parties have adopted equity policies, particularly for candidate searches and nominations. Several Canadian political parties set targets for the representation of candidates from certain groups. These policies were often set for women and, in some cases, have been expanded to include other equity-seeking groups, including LGBTQ+ people. In recent years, the Liberals have aimed for 30% of their candidates to be women (Lapointe, Ferland, and Turgeon Reference Lapointe, Ferland and Turgeon2024, 3); the NDP “aims to have women candidates running for election in at least 50 per cent of all electoral districts and to have women candidates in at least 60 per cent of the electoral districts where it has a chance of winning” (Ashe Reference Ashe, Tremblay and Everitt2020, 306); and the Greens set a target for 50% women (Lapointe, Ferland, and Turgeon Reference Lapointe, Ferland and Turgeon2024, 3). Since 2019, the NDP’s gender target explicitly includes TNB individuals. The NDP also has an additional target of nominating 30% of candidates from equity-seeking groups in winnable districts. In this policy, “equity-seeking groups” includes LGBTQ+ people, racialized persons, Indigenous people, persons living with disabilities, and youth (New Democratic Party of Canada 2019).

The Liberals, NDP, and Greens all have diversity search committees, which can pressure local organizations to meet these targets. These committees can assist with candidate searching and recruitment and even freeze nominations until there has been a “thorough” search for candidates from equity-seeking groups. Following the lead of the NDP, the Liberals and Greens now also include LGBTQ+ candidates in their policies requiring a “thorough” search for candidates from equity-seeking groups (Ashe Reference Ashe, Tremblay and Everitt2020; Everitt Reference Everitt2015; Everitt and Tremblay Reference Everitt and Tremblay2023; Green Party of Canada 2021; Liberal Party of Canada 2020).

Some Canadian political parties have also considered candidate diversity and equity when it comes to political finance. In 1983, the NDP established the Agnes McPhail fund to support women in nomination contests and general elections, and in 1984 the Liberals established the Judy LaMarsh Fund to support women candidates (Bashevkin Reference Bashevkin1993). Funds for LGBTQ+ candidates were established much later and have been less widespread. Although the NDP currently has Victory Funds for racial minorities, women, LGBTQ+ people, persons living with disabilities, and Indigenous peoples, the earliest documentation we can find of a Victory Fund for LGBTQ+ candidates is from June 11, 2022 — that is, after the period under study in this paper.Footnote ⁵

Data

We examine inequalities in political finance by combining two data sources. First, we draw on Baisley and Albaugh’s (Reference Baisley and AlbaughForthcoming) dataset of candidates for the 2015–21 Canadian federal elections, which provides information on LGBTQ+, TNB, and Two Spirit candidates. Although this dataset includes the five parties that won seats in the Canadian House of Commons, we exclude the BQ because it had only one out LGBTQ+ candidate during this time period. Second, we draw on finance data from Elections Canada, Canada’s federal electoral management body. We use Part 2f (“Summary of Contributions, Loans, Transfers, and Other Cash Inflows”) of reports of financial returns filed by candidates after each election. When possible, we use the financial returns amended by Elections Canada. When amended returns are not available, we use the original returns filed by candidates. We present an overview of the coding of all variables in the dataset and their sources in Online Appendix A.

Our explanatory variables are candidate identities. First, we examine a gap between straight cis and LGBTQ+ candidates. Then, given that there may be additional barriers facing particular subgroups, we examine six different non-overlapping groups: (1) straight cis men, (2) straight cis women, (3) queer cis men, (4) queer cis women, (5) TNB people, and (6) Two-Spirit people.Footnote ⁶

Figure 1 displays the number of out LGBTQ+ candidates by party in each year. Figure 1 shows the “rainbow wave” — the sharp increase in LGBTQ+ candidates from 2015 to 2019. It also highlights a partisan divide in nominating LGBTQ+ candidates. Unsurprisingly, the NDP, the party most tied to the LGBTQ+ movement (DeGagne Reference DeGagne2019), nominates the lion’s share of the LGBTQ+ candidates, followed by the Green Party, which is also ideologically left of the Liberal and Conservative parties.

Figure 1. Number of LGBTQ+ candidates, by party and year.

Figure 2 breaks down LGBTQ+ candidates into subgroups.Footnote ⁷ In 2015, LGBTQ+ candidates across parties were rare and typically queer cis men, with a few queer cis women. In 2019 and 2021, LGBTQ+ candidates became more diverse, including substantial numbers of queer cis women and TNB candidates, along with a few Two Spirit candidates. Figure 2 highlights that, when we conduct more intersectional or disaggregated analyses, we need to be cautious about small cell sizes in our regression models, particularly for Two Spirit candidates across parties and subgroups within the Liberal and Conservative parties, which nominate few LGBTQ+ candidates.

Figure 2. Number of candidates, by subgroup of LGBTQ+ candidates, party, and year.

Our main outcome of interest is the grand total contributions — the total amount of money raised as part of a candidate’s general election campaign. We use the reported grand total contributions for each candidate where possible. However, the Elections Canada data include some tabulation errors in grand total contributions. For example, some candidates report 0 Canadian dollars (CAD) in grand total contributions even when they have received money. We calculate the grand total contributions manually to correct these errors. In cases where we found candidates reporting campaign spending but not contributions, which are potential compliance errors, we replace the missing contributions data with overall campaign spending data. Our results are substantively similar if we focus on party transfers (the sum of monetary and non-monetary transfers from registered parties and EDAs to local candidates) or campaign spending (the total amount spent by a candidate in a general election campaign). We show these results in Online Appendices G and H.

Our dataset also includes a number of potential control variables, including party, year, gender, incumbency, challenging an incumbent, candidate race, past political experience at other levels of government, district competitiveness (the party’s win or loss margin in the previous election), the spending limit in each electoral district in CAD (calculated by Elections Canada based on population, in thousands CAD), district urbanity (pulled from Armstrong, Lucas, and Taylor Reference Armstrong, Lucas and Taylor2022 and rescaled 0–1), district median household income (in thousands of dollars), district median age, share of the district with a university degree, and the share of the district that is South Asian. We discuss the rationales for these control variables when we discuss our regression models in the Analysis section below.

We present descriptive statistics for our main outcome variable along with breakdowns for our explanatory variables and selected control variables in Table 1. We include five statistics that help illustrate the distributions: the (1) percentage of candidates who reported raising no money, (2) median, (3) mean, (4) standard deviation, and (5) range. We include the percentage of zeroes because political finance data often include substantial numbers of candidates who receive and/or spend zero money. In our case, zeroes are relatively rare. Only 7% of candidates received no money, and the overwhelming majority (83%) of the zeroes are Greens. These results suggest a pressing need to account for the large number of zeroes when analyzing Green candidates.

Table 1. Summary statistics for grand total contributions, overall and by categorical covariates

Table 1 shows some initial inequalities to explain in our regression analyses. LGBTQ+ candidates receive substantially less money than straight cis candidates. Among our disaggregated subgroups, we find that straight cis men receive the most, followed by Two Spirit candidates, queer cis men, straight cis women, queer cis women, and TNB candidates. The results for Two Spirit candidates may be surprising given that they may face marginalization both for being Indigenous and for being sexual/gender minorities, but the four Two Spirit candidates include Blake Desjarlais, the first Two Spirit Member of Parliament (MP) and one of the few NDP candidates who won a seat previously held by another party in 2021. Otherwise, these data typically show gender marginalization: men generally receive more than women, and cis candidates receive more than TNB candidates.

Although these bivariate analyses show some inequalities to be explained, the remainder of the table shows some other variables that may explain the bivariate inequalities. Partisan differences are likely to account for these bivariate inequalities. Conservative candidates receive the most, followed closely by the Liberals. However, NDP and Green candidates receive substantially less money overall. Given that LGBTQ+ candidates tend to run disproportionately for the NDP and, to a lesser extent, the Greens, the unequal fundraising across parties is likely to work against LGBTQ+ candidates. Fundraising goes down over time, for two reasons. First, in 2015, the Conservatives called a longer campaign period, which allowed candidates to spend more money in that election. Second, in 2021, the snap election left less time for parties and candidates to raise money. Given that LGBTQ+ candidates were more likely to be on the ballot in 2019 and 2021 than in 2015, we need to account for time in our regression models. Incumbents all received at least some money, often substantially more than candidates running in open seats or as challengers (running against incumbents). Candidates who have past political experience at other levels of government are more likely to receive money and receive substantially more than those who do not. Given that LGBTQ+ candidates are unlikely to be incumbents or have past political experience, these inequalities may account for the lower levels of fundraising among LGBTQ+ candidates.

Analysis

We account for the inequalities in political finance facing LGBTQ+ candidates using a series of regression models. We run two families of models. In all models, grand total contributions is the outcome. Model 1 uses the LGBTQ+ dummy as the main explanatory variable, and Model 2 uses disaggregated gender and sexual identity subgroups. We run separate models for each party because parties play a major role in distributing funds and may make different strategic decisions about, for example, how much to defend against other parties in areas they previously won or targeting new districts to win. However, we include a model with all parties as a baseline.

We estimate negative binomial models with Huber-White robust standard errors, adjusting for variables that may account for these gaps (for a recent example, see Piscopo et al. Reference Piscopo, Hinojosa, Thomas and Siavelis2022). These models address that the data do not contain negative values, which can bias ordinary least squares estimates. Negative binomial models rely on a log link. This log link addresses issues of heteroskedasticity and diminishing returns (a contribution of $1,000 means something different when a campaign has $10,000 than when it has $100,000). The distributions of the outcome variables visually correspond to negative binomial distributions rather than Poisson distributions, which also satisfy the first two criteria.

Given the relatively low share of zeroes in our data, we do not run zero-inflated negative binomial regressions, which are commonly used in other analyses of finance data. In our data, only the Green Party has a clearly inflated number of zeroes (see Table 1). In addition, there are potential estimation problems when using zero-inflated models, which account for “excess” zeroes beyond those predicted by a regular negative binomial model by mixing the negative binomial model with a binomial model, such as logit or probit. However, given some variables, such as incumbency, perfectly predict having more than zero contributions, the binomial portion of the model can run into estimation challenges. Given what we know about our data, we handle zeroes in two ways. In the main text, we exclude zeroes entirely, then we estimate zero-censored negative binomial regression models. These models are the most comparable across all four parties. In Online Appendix D, we supplement these results by running negative binomial models with the zeroes for the Liberals, Conservatives, and the NDP, then estimating a penalized logit model of whether Green candidates receive any money (Firth Reference Firth1993). Penalized logit accounts for perfect prediction (Rainey Reference Rainey2016).

Given that our descriptive statistics in Figure 2 show some small cell sizes, we note two exclusions from our models. First, we exclude Two Spirit candidates from our regression models. Second, we do not run models for the Liberals and Conservatives using the disaggregated gender and sexual identity subgroups. These exclusions follow our general rule not to run a regression analysis when we are dealing with cell sizes of under five.

Table 2 shows an overview of the models run in the main manuscript, including the parties included, the specification of our main explanatory variable, and our control variables. In Models 1 and 2, we adjust for variables well-attested to correlate with political finance outcomes. First, we include variables that capture the district electoral context for the party, including the party’s win/loss margin in the previous election, incumbency, and running as a challenger (against an incumbent). Second, we account for the substantial variation in fundraising across elections by including both year dummies and the district spending limit in each election (in thousands CAD). Third, we add district demographics known to predict differences in political donations geographically in Canada, including urbanity (rescaled 0–1), median household income (in thousands CAD), median age, university-educated population share, and South Asian population share (Garnett Reference Garnett2024). Finally, we add two sets of interactions, including a three-way interaction between the win/loss margin, incumbency, and year and a two-way interaction between challenger and year. Although three-way interactions are difficult to interpret, the three-way interaction between incumbency, win margin, and year is well-supported in scatterplots of the data (see Online Appendix B). When we run models that pool across parties, we also include party dummies and include them in our preexisting interaction terms.

Table 2. Overview of models