Party Hierarchies and Women’s Representation at Different Career Stages

In most parliamentary democracies, politicians with progressive ambition must move through the ranks of their parties’ hierarchies in a systematic manner before reaching the top leadership positions (Müller Reference Müller2000). We can stylistically conceptualize this hierarchy at the national level as involving three stages: (1) initial nomination and list rank, (2) renomination and rank advancement, and (3) promotion to cabinet.

The first stage represents entry into national politics, which happens when an individual gains a spot on a national party list for legislative elections. Most new candidates receive a low (non-winnable) list position, so to continue pursuing a political career it is necessary to seek renomination even after an initial loss. In existing research on women’s representation at this stage, scholars have suggested that incoming men and women may receive differential treatment—for example, women may receive fewer donations, loans, or transfers from parties to fund their campaigns (Piscopo et al. Reference Piscopo, Hinojosa, Thomas and Siavelis2021). Institutionalized candidate selection, centralization, and inclusivity of decision-making can make parties’ nomination processes more open to women candidates (Bjarnegård and Zetterberg Reference Bjarnegård and Zetterberg2016; Hinojosa Reference Hinojosa2012), but informal networks, statistical discrimination, and informational inequalities might hamper women’s entry when a party selectorate is composed mainly of men (e.g., Crowder-Meyer Reference Crowder-Meyer2013; Folke, Rickne, and Smith Reference Folke, Rickne and Smith2021; Niven Reference Niven1998). In contexts where parties favor loyalty, women might also be less likely to be seen as having this qualification (Josefsson Reference Josefsson2020).

The second stage represents the time in a politician’s career after an initial debut on a national party list but before advancement (if it occurs) into the highest levels of leadership. For those who desire to stay in politics, the goal is to get renominated to progressively higher (i.e., electorally safer) ranks on the party’s list, culminating in the top-rank (list leader) position. In this stage, women may face gender bias insofar as equal quality or accreditation does not get them to the same ultimate outcome (Folke and Rickne Reference Folke and Rickne2016; Muriaas and Stavenes Reference Muriaas and Stavenes2024). In some cases, scholars have found that women who are incumbents are equally or more likely to be renominated (Black and Erickson Reference Black and Erickson2003; Shair-Rosenfield and Hinojosa Reference Shair-Rosenfield and Hinojosa2014), but this is not universal across comparative cases.Footnote ⁴

Finally, the third stage represents promotion into higher offices, such as ministerial positions in the cabinet. In parliamentary democracies, the cabinet is the apex of power, and ministerial appointments are thus highly sought after by individual MPs. Cabinet ministers have influence over policy and the legislative agenda, and also enjoy perquisites, such as higher salaries and staff resources (Martin Reference Martin2016). The visibility of a cabinet position might also make it easier for an MP to launch a lucrative postretirement career outside of politics. Most research finds that women are less likely to attain higher office, including cabinet positions (Krook and O’Brien Reference Krook and O’Brien2012; Verge and Claveria Reference Verge and Claveria2016). Women often come to power through opposition parties with declining seat shares (O’Brien Reference O’Brien2015), take longer to advance (Kroeber and Hüffelmann Reference Kroeber and Hüffelmann2022), and attain lower-prestige posts (Escobar-Lemmon and Taylor-Robinson Reference Escobar-Lemmon and Taylor-Robinson2009; Reynolds Reference Reynolds1999).Footnote ⁵

How Majoritarianism Creeps Into PR Systems

Research on gender and electoral systems differentiates between countries that use majoritarian rules to elect leaders and those that use PR. Yet while this dichotomous conceptualization might be reasonable for making broad comparisons across national legislatures, we highlight that aspects of majoritarianism are present even in PR systems. Indeed, all democracies have local executives, such as mayors and governors. Sometimes these executives are directly elected, in which case the electoral rules are inherently majoritarian, since only a single seat is at stake.

Even if local executives are indirectly elected by councils or legislatures, majoritarian competition still emerges within parties over key nominations. For example, in local assemblies in Norway, it is a strong empirical regularity that a top-listed candidate from the largest party list becomes mayor.Footnote ⁶ Since there can be only one person at the top of the list, competition to secure this position is also majoritarian in nature.

Existing research highlights the challenges that majoritarian offices pose to women, even when voters (and whatever biases they may hold) are removed from the selection process. First, such offices tend to place greater time demands on officeholders, making it harder to balance a political career with family obligations (Bernhard, Shames, and Teele Reference Bernhard, Shames and Teele2021; Iversen and Rosenbluth Reference Iversen and Rosenbluth2010). In addition, research in several contexts shows that women in politics face more violence and abusive public treatment than men, and this is exacerbated in majoritarian offices (Håkansson Reference Håkansson2021; Krook Reference Krook2020).Footnote ⁷

Finally, when faced with majoritarian competition, risk aversion or endogenous preferences may emerge and discourage women from running as candidates (Fox and Lawless Reference Fox and Lawless2011; Kanthak and Woon Reference Kanthak and Woon2015). These considerations may also apply to progressive ambition within party hierarchies, especially for high-profile positions like mayor or list leader. Indeed, a survey of Norwegian party youth members found that women expressed lower ambition than men for elected offices, and that these differences were more pronounced for executive offices (prime minister, cabinet minister, and mayor) compared to legislative offices (MP and councilor) (Kolltveit Reference Kolltveit2022).

We argue that gaps in women’s representation within PR systems might arise wherever positions are majoritarian (or single-occupancy) in nature. This includes both formal executive offices like mayor and prime minister, as well as important party positions like list leader, which are not strictly “offices” but nevertheless serve important and focal roles within party organizations.Footnote ⁸ The combination of majoritarian rules and coveted status means that, as in majoritarian electoral systems, gender disparities may be larger and harder to eradicate in these positions. Our core empirical expectation is that, even within PR systems, gaps in women’s representation will appear in these majoritarian positions.

Majoritarian Stepping Stones in Parties’ Seniority Systems

If gaps in women’s representation appear in single-occupant positions in early career stages, will these gaps produce negative cascade effects for women’s progression into higher-level offices? Here, it is important to return to the role of parties in structuring careers. We focus in particular on a common feature of party organization: seniority-based systems of promotion.

Seniority systems are present within parties and legislatures across various contexts, operating in some cases as a formal rule and in others as a strong informal norm (Cirone, Cox, and Fiva Reference Cirone, Cox and Fiva2021; de Winter Reference de Winter, Blondel and Thiébault1991; Hall and Shepsle Reference Hall and Shepsle2014; Heinsohn and Schiefer Reference Heinsohn and Schiefer2019; Ranney Reference Ranney, Butler, Penniman and Ranney1981). In either case, a seniority system features two basic regularities. First, any incumbent in good standing will (if they wish) be renominated or reappointed to their current positions. Such norms have been documented in US congressional committees, under the rubric of the “property rights” norm (Shepsle Reference Shepsle1978), and in candidate nominations in a wide range of democracies, from Japan (Cox and Rosenbluth Reference Cox and Rosenbluth1996) to Norway (Cirone, Cox, and Fiva Reference Cirone, Cox and Fiva2021). As Ranney (Reference Ranney, Butler, Penniman and Ranney1981, Reference Schwindt-Bayer99) notes, while parties aim to balance many considerations in candidate nominations, “the greatest single advantage an aspirant for a candidacy can have is to hold the office already.”

Second, seniority progression rules ensure that open nominations for offices (e.g., those with no incumbent claiming them) will generally be allocated preferentially to party members with the highest accrued seniority or those holding predefined feeder offices. For example, committee chairs in the US (Shepsle Reference Shepsle1978) and initial cabinet promotions in Japan (Epstein et al. Reference Epstein, Brady, Kawato and O’Halloran1997; Ono Reference Ono2012) have been awarded on the basis of accrued seniority, while winnable list spots in Norway also tend to be awarded based on previous experience (Cirone, Cox, and Fiva Reference Cirone, Cox and Fiva2021).

How might seniority systems affect women’s progression through the party hierarchy? If promotion occurs purely on the basis of seniority, then women’s advancement into higher levels of office will initially be delayed. In any seniority system that begins with significant gender disparities, it will not be possible to achieve gender parity across all levels quickly, unless seniority is waived at the expense of men in the pipeline. Gender quotas with placement mandates are a key tool to overcome this generational lag in list placement by forcibly displacing some men from the hierarchy (e.g., Besley et al. Reference Besley, Folke, Persson and Rickne2017; Dahlerup Reference Dahlerup2006; Krook Reference Krook2006; O’Brien and Rickne Reference O’Brien and Rickne2016).

Women’s advancement into the highest offices can also be stymied by seniority systems if women tend to enter politics at more advanced ages (e.g., due to motherhood and family care obligations) than their male colleagues, and therefore have less time to develop the requisite level of seniority for a given promotion before their retirement (e.g., Lazarus, Steigerwalt, and Clark Reference Lazarus, Steigerwalt and Clark2023). However, age of entry and career length appear to be less correlated with gender outside of majoritarian contexts (Joshi and Och Reference Joshi and Och2021; Muriaas and Stavenes Reference Muriaas and Stavenes2024).

Beyond these potential sources of friction in the seniority system, our focus is whether the presence of single-occupant positions in a party hierarchy—what we call majoritarian stepping stones—affects women’s career progression. If women face challenges in obtaining these positions, or avoid pursuing them in the first place, then the supply of women eligible for promotion to the next level in the party hierarchy will be reduced, and women’s advancement might be differentially slowed compared to men, creating negative cascade effects in the career pipeline.

We argue that high-level local offices like mayoralties are potential majoritarian stepping stones to national careers in PR systems.Footnote ⁹ This argument resonates with a large body of previous work on the impediments women face in local, entry-level politics (Brown et al. Reference Brown, Mansour, O’Connell and Reeves2025; Cipullo Reference Cipullo2021; Folke and Rickne Reference Folke and Rickne2016; Fulton et al. Reference Fulton, Maestas, Sandy Maisel and Stone2006; Franceschet and Piscopo Reference Franceschet and Piscopo2014; Holman Reference Holman2017; Martin and Conroy Reference Martin and Conroy2024; Payson, Fouirnaies, and Hall Reference Payson, Fouirnaies and Hall2023; Thomsen Reference Thomsen2015).

Similarly, top national list positions may be majoritarian stepping stones into cabinet. While many factors contribute to cabinet selection decisions, seniority is often privileged, and party leaders routinely tap members in top list positions for ministerial roles (Blondel and Thiébault Reference Blondel and Thiébault1991; Cox et al. Reference Cox, Fiva, Smith and Sørensen2021).Footnote ¹⁰ This means that women’s promotion to cabinet might be affected by their access to list-leader positions. If parties deny equally qualified women the top-rank positions, but favor this rank in cabinet promotion decisions, this would reduce women’s chances to advance into the executive.

Ultimately, whether majoritarian positions create negative cascade effects on women’s aggregate representation throughout the party hierarchy is an empirical question, and depends on three conditions: (1) these positions are indeed important stepping stones in the seniority system; (2) women are indeed underrepresented in them; and (3) parties strictly adhere to seniority rules that require passing through them to reach higher offices. Our empirical analysis will shed light on whether each of these conditions is met in the case of Norway.

Procedures and Priorities in Selection Decisions Across the Hierarchy

Norwegian parties feature common pathways of career progression, and a documented use of seniority systems in promotion (Cirone, Cox, and Fiva Reference Cirone, Cox and Fiva2021). At the local level, municipalities are run by local councils, each headed by a mayor.Footnote ¹⁵ Typically, mayors tend to be top-ranked candidates from the largest parties.Footnote ¹⁶ Local elections are conducted under a flexible-list PR electoral system, in which voters can cast preference votes for any candidate on any party list, but local parties control the composition of the list and can give candidates a “head start” worth up to 25% of the total votes. Previous research shows that parties tend to reward incumbents with these privileged (safe) rank positions (Fiva and Røhr Reference Fiva and Røhr2018).

Local office is an important stepping stone to national office in Norway, and a high proportion of first-time MPs have prior local-level experience (Cirone, Cox, and Fiva Reference Cirone, Cox and Fiva2021). From 1945 to 2021, around 80 percent of men and 70 percent of women MPs had prior local council experience.Footnote ¹⁷ In the first few decades of the same period, around 30 percent of male MPs previously held a mayoral post, but this dipped to around 20 percent since the 1970s. For women MPs, in contrast, mayoral experience was exceedingly rare until the early 2000s. Since then, women have only modestly trailed their male counterparts, with around 15 percent of women MPs having prior mayoral experience. While these patterns make clear that serving as a local mayor is not a sine qua non for nomination as a candidate in national elections, our empirical analysis will investigate the differences in probability of advancement from this position.

To elect 150–169 members to the Storting, the national parliament, Norway has used a PR electoral system since 1921.Footnote ¹⁸ Candidate nominations and rank positions on party lists are determined in a decentralized fashion in each district by dues-paying party delegates at nominating conventions (Valen Reference Valen, Gallagher and Marsh1988; Valen, Narud, and Skare Reference Valen, Narud, Skare, Narud, Pedersen and Valen2002), and very few party spots are contested, given the proposals of nomination committees (Cirone, Cox, and Fiva Reference Cirone, Cox and Fiva2021). Because of this, the nomination procedure is considered to be closed (Narud and Valen Reference Narud and Valen2007).

In constructing lists of candidates, local party nominating conventions aim to balance representation on several dimensions, including geography, gender, age, and organizational ties to important interest groups (Means Reference Means1972; Valen Reference Valen, Gallagher and Marsh1988; Valen, Narud, and Skare Reference Valen, Narud, Skare, Narud, Pedersen and Valen2002). Ticket-balancing across such dimensions helps to keep party activists and interest groups satisfied, and also enhances mobilization efforts (Fiva, Halse, and Smith Reference Fiva, Halse and Smith2021). But seniority is arguably the first criterion used in candidate selection. Prior to an election, incumbents are asked if they are interested in renomination, and those who answer affirmatively tend to be considered unless they have lost the confidence of local activist groups or networks in the party (Valen, Narud, and Skare Reference Valen, Narud, Skare, Narud, Pedersen and Valen2002, 202).

Norway is rare among parliamentary systems in that early dissolution of the Storting is not allowed. Members serve four-year terms (three-year terms prior to 1945). As in other parliamentary systems, however, cabinets are formed through party negotiations after elections, and can be forced to resign with a no confidence vote. Norway’s cabinets have featured a mix of single-party majority, single-party minority, and coalition governments (Narud and Strøm Reference Narud, Strøm, Bergman and Strøm2011). The prime minister’s discretion in selecting cabinet ministers has varied, depending on pressure from the party organization and constraints on personnel decisions involving coalition partners.

Many cabinet ministers have prior experience as MPs, but it is not a requirement, and ministers appointed from outside of parliament (including the party organization hierarchy) are common. Nonetheless, MPs tend to be promoted to the more important portfolios (such as prime minister and finance minister) and tend to have more extensive experience in parliament (Narud and Strøm Reference Narud, Strøm, Bergman and Strøm2011). If appointed, an MP must resign from the Storting and is replaced by a deputy (a lower-ranked candidate on the same party list). For the purposes of our analysis, we focus only on the sample of MPs who are appointed to cabinet following an election (i.e., we do not analyze ministers appointed from outside of parliament).

Data and Methods

Our analysis will make use of a century’s worth of detailed, micro-level data on Norwegian politicians, and this expansive coverage is one of our study’s significant contributions. We collect and link data sources on election outcomes for individual politicians across local and national politics to data on individual background characteristics (i.e., education) from 1921 onward. For comparison, prior research on gender in national elections typically focuses on recent decades, starting in the 1980s or 1990s (e.g., Piscopo et al. Reference Piscopo, Hinojosa, Thomas and Siavelis2021; Smrek Reference Smrek2022), or in the case of the US, the 1970s (e.g., Volden, Wiseman, and Wittmer Reference Volden, Wiseman and Wittmer2013). Similar research on local elections begins in the 1950s for the US (Ferreira and Gyourko Reference Ferreira and Gyourko2009) or the 1980s for Sweden (Folke and Rickne Reference Folke and Rickne2016). In addition, while prior research tends to focus on one stage of a political career, we consider women’s outcomes across the entire hierarchy of a typical career.

Our dataset includes national-level candidate data from Fiva and Smith (Reference Fiva and Smith2017) and local-level candidate data from Fiva, Sørensen, and Vøllo (Reference Fiva, Sørensen and Vøllo2024), updated with information on recent candidates, and linked with municipality-level information from Fiva, Halse, and Natvik (Reference Fiva, Halse and Natvik2023). Our primary dataset covers the universe of candidates in Norwegian parliamentary elections from 1921 to 2021 (N = 68,207), and includes individual-level information on party, list rank, party vote share, electoral outcomes, postelectoral cabinet appointments, and personal characteristics (such as gender, age, education, and prior political experience).Footnote ¹⁹ We include postelection cabinet appointments through June 2024. A secondary dataset on local-level offices covers all mayors elected from 1971 to 2023 (N = 5,987). Personal identifiers included in the datasets allow us to track individuals across political offices.Footnote ²⁰

For most of our research questions, we can gain considerable insight from a simple descriptive analysis of longitudinal patterns. For others, we pair the descriptive analysis with panel data regressions to investigate individual candidates’ career progression outcomes controlling for party, district, and year fixed effects.Footnote ²¹ We limit our data sample to candidates running for one of the seven main parties that have dominated Norwegian politics since the introduction of PR, unless stated otherwise.Footnote ²²

Moving from Local Politics to National Politics

Experience in local politics is a common feature of national-level candidates’ biographies (Cirone, Cox, and Fiva Reference Cirone, Cox and Fiva2021). As noted earlier, since 1945, around 80 percent of men and 70 percent of women who became MPs had local council experience, and as much as 30 percent of men MPs (but, until recently, not women MPs) had mayoral experience. As Figure 1 shows, there were virtually no women mayors prior to 1970, and women’s share of mayoralties exceeded 25 percent only after 2015, never rising above 40 percent.

These patterns already suggest that local offices are an important stepping stone (condition 1), but that parties do not strictly adhere to seniority rules that require passing through them—especially the mayoralty—to be nominated as a candidate on a national list (condition 3). Nevertheless, there may be differential treatment of men and women who do (and do not) pass through the majoritarian stepping stone of the mayoralty, which may either impede or enhance women’s representational outcomes throughout the party hierarchy.

Data on local candidates are not available for the entire time period covered by the national-level data. However, we can examine the effect of holding the mayoralty on men’s and women’s national political careers for the period between 2003 and 2011, for which we have complete data on all candidates for local election lists (Fiva, Sørensen, and Vøllo Reference Fiva, Sørensen and Vøllo2024). For these years, we are able to compare candidates’ progressive trajectories while controlling for party-district-year fixed effects.

We analyze progression from local to national politics separately for men and women using the following specification:

(1) $$ \begin{array}{rl}{Y}_{rpmdt}={\gamma}_{pdt}+{\beta}_r+{\lambda}_rWoma{n}_{rpmdt}+{u}_{rpmdt}& \end{array} $$

where $ {Y}_{rpmdt} $ represents a future electoral outcome for the candidate ranked r for party p in municipality m belonging to parliamentary election district d at time t (that is, we aggregate the data from municipal council elections to the larger election district to which the municipalities belong in national parliamentary elections).

We consider two outcome variables: (i) running for national office ( $ Run $ ) in the future and (ii) winning national office ( $ Win $ ) in the future. $ Woma{n}_{rpmdt} $ is a dummy for women candidates. Because our model controls for party-district-year fixed effects ( $ {\gamma}_{pdt} $ ) and rank fixed effects ( $ {\beta}_r $ ), the parameter $ {\lambda}_1 $ isolates the differential effect of becoming mayor by comparing outcomes for women and men running for the same party in the same parliamentary district in the same election year (but in a different municipality). This is captured by the coefficients $ {\lambda}_1,\dots, {\lambda}_{10} $ .Footnote ²⁷

In our baseline analysis, we consider only the local party lists where the first-ranked candidate ultimately became mayor.Footnote ²⁸ Figure 3 presents the results. There are two big takeaways. First, it is much more common for candidates who won the mayoral spot to run in a future national election than candidates who were ranked farther down the list. For local parties that held the mayoralty, only about 10 percent of candidates ranked second or third went on to run in a national election, compared to about 47 percent of first-ranked women and 35 percent of first-ranked men. Winning national office is uncommon, but virtually unheard of for candidates ranked below the top position. Outside of the first rank, less than 10 percent of candidates—both men and women—ran for national office, but very few won a seat in parliament (as measured up to the 2021 national election).

Figure 3. The Differential Effect of Local List Rank, and Becoming Mayor, on Running and Winning in Future National Elections, 2003–2011

Note: The top panels display averages of $ Run $ and $ Win $ for women and men, by list rank of the local candidate. The bottom panels provide estimates of $ {\lambda}_1,\dots, {\lambda}_{10+} $ based on Equation 1. The sample is restricted to local candidates for one of the seven main parties in the 2003–2011 period where the first-ranked candidate on the relevant list ultimately became mayor. List positions below 10 are in the “10+” category. We exclude candidates that previously ran for national office and municipalities with directly elected mayors. Standard errors are clustered at the party-district-year level. Table A.1 in the Supplementary Material provides the results in table format.

The second major takeaway from Figure 3 is that women who won the mayoralty were much more likely to run for national office, and also more likely to win, compared to men. Around 8 percent of women mayors won a seat, compared to approximately 3 percent of the men. The bottom panels of Figure 3 plot the $ {\lambda}_r $ coefficients, for $ r=1,\dots, 10+ $ , along with $ 95\% $ confidence intervals. As can be seen, the largest, and statistically significant, effects are those comparing (first-ranked) women mayors to (first-ranked) men mayors. Women in the second rank position (i.e., not mayors) were also slightly more likely to “catapult” into a national candidacy, but not more likely to win.

To understand these findings in terms of their substantive significance, we refer to Cirone, Cox, and Fiva (Reference Cirone, Cox and Fiva2021, 242) who note that, “in any given year, there are about 11,000 local councilors chasing only 169 seats in parliament.” The baseline probability of a parliamentary candidacy is $ 0.056 $ , and only $ 0.002 $ of a win. In this sense, the probability of transitioning from the mayoralty to a national run or win, when compared against these benchmarks, is impressive.

In sum, although women have been underrepresented in mayoral positions (Figure 2), those who do make it onto this local stepping stone have better chances of getting nominated to a national list, and winning, than their male counterparts (Figure 3). One limitation of this analysis is that we only have local-level candidate data to study differential outcomes for men and women since 2003, a time when men’s overrepresentation in mayoral positions was beginning to decline. For the elected mayors, however, our data go back to 1971, and show that women mayors have been more likely to appear on a future list, and win national office, in nearly every year observed.Footnote ²⁹ A second limitation is that we lack variables at the local level to explore the mechanisms behind women’s differential promotion outcomes. It could be that women mayors are promoted to national candidacies at higher rates because they are of higher quality than their male counterparts, or because they have served longer in local politics and have therefore accrued more seniority. We can, however, overcome these limitations in our analysis of national-level careers.

Moving from Parliament to Cabinet

Once a politician makes it onto a national list, the goal is to secure the top rank, as these are the safest spots and are associated with prestige and power within the party. Our data on national candidates indicate that women debut at a similar average list position as men, and progress upward through the ranks at the same rate, but that they have been much less likely to win the top spot for nearly all of the last century.Footnote ³⁰ These dynamics produce the gaps in Figure 2, where men are relatively overrepresented in the top list positions vis-à-vis their representation on the lists as a whole.

At the same time, Figure 1 reveals a different pattern for the cabinet, where the share of women among ministers has consistently led the share of women MPs and candidates (including list leaders) since the 1960s (despite the absence of firm quotas for women ministers). As in the preceding analysis for advancement from local politics to national politics, here we investigate whether there is differential treatment of men and women who do (and do not) pass through the majoritarian stepping stone of the list leader position when it comes to promotion into cabinet.

Figure 4 presents descriptive plots of the fraction of women and men who are promoted to cabinet positions conditional on their list positions. For both men and women who hold less prestigious list positions, there is a negligible chance of promotion to cabinet. Once the second spot is attained, however, the chance of getting into cabinet rises substantially. Notably, the probability that a woman gets promoted into cabinet is higher at both the second and first list positions. About 8 percent of women and 5 percent of men move from the second spot into a ministerial role, while 29 percent of women and 20 percent of men move up from the first spot. These gender differences are statistically distinguishable from zero ( $ p=0.054 $ ; $ p=0.008 $ ), as shown in the bottom panel of Figure 4.

Figure 4. The Differential Effect of National List Rank on Cabinet Promotion, 1945–2021

Note: This figure examines whether there are gender differences in cabinet promotions based on national list position. The top panel shows the raw data, and the bottom panel shows the coefficient estimates of promotion to cabinet by list rank and gender using data from the 1945–2021 period. We include cases where a cabinet appointment occurs after the election as well as when it continues over from the previous term (i.e., reappointment). The samples are limited to candidates running for parties that were part of any cabinet following the election. List positions below 10 are in the “10+” category. Table A.2 in the Supplementary Material provides the results for the bottom panel in table format.

These descriptive patterns reveal that top list spots, and not positions lower down, are important stepping stones for reaching the cabinet. But this analysis does not account for other features of candidates’ political careers or abilities that might be driving cabinet appointments. In contrast to the local-level data, our national-level data allow us to evaluate promotion rates for men and women list leaders conditional on seniority and higher education (one observable measure of quality).

We evaluate these dynamics with the following regression:

(2) $$ \begin{array}{rl}\begin{array}{l}Cabine{t}_{it}={\lambda}_{p[i]t}+{\alpha}_1Woma{n}_i+{\alpha}_2Seniorit{y}_{it}\\ {}\hskip5em +{\alpha}_3ListPosition{1}_{it}+{\alpha}_4Seniorit{y}_{it}\cdot Woma{n}_i\\ {}\hskip5em +{\alpha}_5ListPosition{1}_{it}\cdot Woma{n}_i& \\ {}\hskip5em +{\alpha}_6HighEducatio{n}_i\\ {}\hskip5em +{\alpha}_7HighEducatio{n}_i\cdot Woma{n}_i+{\epsilon}_{it},\end{array}& \end{array} $$

where $ Cabine{t}_{it} $ is a dummy variable that takes the value of 1 if candidate i is promoted to cabinet in election period t. As explanatory variables, we include a dummy variable for gender ( $ Woma{n}_i $ ), accrued seniority, a dummy variable for the list-leader position, and the interactions between these latter two variables and gender. In our full specification, we also include a dummy variable for higher education (i.e., postsecondary school). The regression also includes party-year fixed effects ( $ {\lambda}_{p[i]t} $ ) ensuring that all inferences are drawn from candidates running for the same party in the same election year. The measure of seniority ( $ Seniorit{y}_{it} $ ) is simply the number of times a candidate has previously run for a national office prior to the most recent election. A gender-neutral seniority system implies that $ {\alpha}_4=0 $ , while a gender-neutral effect of the top-rank position as a stepping stone into cabinet implies that $ {\alpha}_5=0 $ .

Table 1 presents the results of this regression, for five separate models based on all candidates (models 1 and 2) or elected MPs only (models 3–5). We consider all candidates (not only MPs) because unelected candidates can still be appointed to cabinet. The samples are limited to candidates and MPs from the parties that participated in any cabinet following the election. Because we only have complete biographical data for MPs, we can only include the education dummy for this sample (model 5). The results show that seniority increases the probability of cabinet promotion across all models, as does the top-rank position. There is also clear evidence that better-qualified MPs, as measured by education level, are more likely to get promoted.

Table 1. Effects of Gender, Seniority, and Top-Rank Position on Cabinet Appointment

Note: Seniority is measured as the number of previous candidacies. We include cases where a cabinet appointment occurs after the election as well as when it continues over from the previous term (i.e., reappointment). The samples are limited to candidates running for parties that are part of any cabinet following the election. Standard errors are clustered at the individual level and reported in parentheses. All specifications include party-year fixed effects. * denotes 10% statistical significance, ** 5% and *** 1%. Table A.3 in the Supplementary Material includes dummies for additional ranks.

The small and insignificant coefficients on the interaction term for $ Seniority\cdot Woman $ indicate that the slope of the relationship between seniority and cabinet appointment does not vary by gender. Meanwhile, the coefficients on the interaction term for $ ListPosition1\cdot Woman $ are larger, and statistically significant in the sample of all candidates. This reflects the differential rates of promotion seen in Figure 4. Table A.3 in the Supplementary Material shows a positive effect for the interaction of the second list position and gender, as well, in line with the descriptive pattern.

In sum, these analyses produce two insights about women’s progression into the cabinet. First, most cabinet appointments go to senior politicians in the top two list positions, and are not awarded to candidates or MPs elected from lower down on the list. Second, although women are underrepresented among list leaders throughout this entire period, they are more likely than men to be promoted from this position, and also more likely to catapult to the cabinet from the second rank (thereby bypassing the majoritarian stepping stone of list leader).

Additional Analyses

Our appendices consider two other potential factors that might affect the dynamics of women’s career progression: differences in candidate quality and age of entry for men and women, and differences in renomination and advancement in list rank (prior to the top-rank position).

First, if men and women have different characteristics—for example, if the men who seek national office are better educated, or are younger when they debut on a national list—this could be driving parties’ placement of these candidates (e.g., Joshi and Och Reference Joshi and Och2021; Lazarus, Steigerwalt, and Clark Reference Lazarus, Steigerwalt and Clark2023; Muriaas and Stavenes Reference Muriaas and Stavenes2024). In the Supplementary Material, we plot the average age of elected politicians and the fraction with higher education in each election year, and compare these variables by gender (see Figure A.3 in the Supplementary Material). After 1970, we see similar patterns for both men and women, suggesting that age and education are not driving gender gaps in representation.

Second, to test whether unequal renomination has created friction in women’s representation (e.g., Black and Erickson Reference Black and Erickson2003; Nowacki Reference Nowacki2025; Shair-Rosenfield and Hinojosa Reference Shair-Rosenfield and Hinojosa2014; Smrek Reference Smrek2022), we employ an RD design to isolate the causal effect of winning a seat in parliament on renomination to a winnable spot, separately for women and men. We estimate this equation using candidates who, for the first time, are either next in line to win a seat, or first in line to lose a seat, and measure the distance to the seat threshold using the metric proposed by Folke (Reference Folke2014), following Cirone, Cox, and Fiva (Reference Cirone, Cox and Fiva2021) and Fiva and Smith (Reference Fiva and Smith2018).Footnote ³¹ Figure A.9 in the Supplementary Material indicates that women have been just as likely to be renominated to a winnable spot on the party list as their male colleagues in these narrow-win scenarios. In sum, this analysis finds no evidence for gender bias in renominations, suggesting that the system is gender-neutral in this part of the career hierarchy.