The intra-party bargain over ministerial appointments: how party leader performance affects the ‘partyness’ of government

Abstract In parliamentary democracies, the logic of delegation from voters to government requires that political parties control government actions. Recruiting government personnel through the party organisation is the primary mechanism for parties to retain such a dominance over the government. Existing research has examined secular trends and cross-sectional variance in ministers’ party ties, mostly focussing on appointments of party members to government office. By contrast, this article centres on the appointment of members of the party elite as a yardstick for party control over government. It explores short-term variance in the ‘partyness’ of appointments, arguing that performance-related shifts in the intra-party power balance condition party elites’ access to ministerial office. Utilising data on ministerial appointments in Austria (1945–2017; n = 603), the article demonstrates that successful party leaders can relax party control by minimising appointments of party elite members, while relatively unsuccessful leaders have to compensate party elites with government jobs.

society (Katz and Mair 2018;Lawson 1980), we know surprisingly little about the factors driving the 'partyness' of ministerial appointments. Empirical studies are rare and have focussed predominantly on cross-sectional variance (Blondel 1985;De Winter 1991;Dowding and Dumont 2009;2014) and long-term trends (Andeweg 2000b;Costa Pinto et al. 2018;Strøm 2002). However, much of the variation in the 'partyness' of appointments is yet to be explored. For one, this phenomenon differs substantially in the short term between cabinets and parties. Secondly, while much of the theoretical literature has claimed that the appointment of members of the party elite to government will provide for reliable party agents in public office, most empirical studies focus on appointments of party members to ministerial office, which is a much more inclusive category, comprising all individuals with formal party membership.
By exploring the short-term drivers of party elite appointments to ministerial office, this article addresses both gaps in the existing literature, arguing that power dynamics within political parties are crucial to understanding change in the 'partyness' of ministerial appointments. Based on a comprehensive data set of ministerial appointments in Austria (1945-2017; n = 603), the article demonstrates that the power balance between the primary selectors in ministerial appointments (the party leader and the remainder of the national party elite) affects the composition of a party's team in the cabinet.
The article advances the following theoretical argument. While the party leader has incentives to appoint outsiders and low-level party members to increase their own manoeuvrability in office and to keep potential challengers at a distance, their choice in these processes is constrained by the preferences of party elites, who claim such attractive job opportunities for themselves in exchange for their support for the leader. Although the leader cannot ignore these demands at will, their ability to maintain intra-party support by delivering on the party's goals allows them to minimise appointments of party elite members, thereby effectively reducing party control over the government. By studying the relationship between intra-party power and ministerial appointments, this article adds to the literature on party government and party organisation, contributing new empirical insights into an understudied phenomenon with substantial implications for the legitimacy of party-based representation.

Parties and governments
In parliamentary democracies, political parties link government actions to the preferences of the electorate, thus establishing legitimacy through responsive governance. The concept of party government denotes parties' varying capacity to fulfil this task (Rose 1969). Where it applies, '[p]opular election gives the governing party […] the authority to claim popular compliance with what it says is right' (Rose 1974: 380) and government actions 'are influenced by values and policies derived from the institutions of party' (Rose 1974: 379). Besides several general conditions for party government -including electoral competition between parties with different policy proposals (Thomassen 1994), the formulation of these proposals by the party and the impact of voters' evaluation of them on vote choice -the recruitment of government ministers 'by and through parties' (Katz 1986) is crucial for parties' ability to 'translate possession of the highest formal offices of a regime into operational control of government' (Rose 1969: 413;Mair 2008).
Many studies assume that recruiting ministers from the pool of party members (insiders, as opposed to outsiders without ties to the nominating party) will ex-ante minimise the risk of agency loss (Müller 2000;Strøm 2000), granting the party control over the government via ministers' self-selection into the party organisation (Strøm 1990) and their subsequent political socialisation into party norms and ideology. Several of these studies have identified substantial increases in outsider appointments in various European countries (Alexiadou and Gunaydin 2019;Amorim Neto and Strøm 2006;Cotta 2018;Dowding and Dumont 2009;Strøm 2002), calling parties' grip on governments into question. While definitions of outsider-ministers still differ considerably throughout the literature (Bertsou and Caramani 2020;Cotta 2018;McDonnell and Valbruzzi 2014;Seixas and Costa 2021), several contributions have associated the apparent rise of outsiders in government with general tendencies of party decline (Blondel and Cotta 2000;Krouwel 2012;Strøm 2002). As parties lose their rooting in society and trust among voters, insiders increasingly give way to a specific type of outsiders, namely technocrats, whose political independence and expertise substitute for parties' waning ability to provide the basis for legitimate government (Bertsou and Caramani 2020; Costa Pinto et al. 2018;Habermas 1973). Technocrats' explicitly non-partisan attributes and their technical knowledge -which is also a popular trait among voters (Lavezzolo et al. 2021) -render their appointments particularly likely in times of political or economic crisis (Alexiadou and Gunaydin 2019;Centeno 1993;Fischer 1990;McDonnell and Valbruzzi 2014). Beyond these specific situations, however, outsider appointments are still clearly the exception, not the rule (Amorim Neto and Strøm 2006;Strøm 2002;Kaltenegger and Ennser-Jedenastik 2022), and secular increases in outsider appointments -while present in some countries -are hardly a universal pattern (Costa Pinto et al. 2018: 277;Dowding and Dumont 2009).
Much of the party government literature sets a higher threshold for the 'partyness' of appointments, implying that the relatively imprecise conceptual distinction between insiders and outsiders is an insufficient criterion to assess party control over government. Instead, these studies suggest that only a fraction of the insider group, namely members of the national party elite, will reliably act in the best interest of the party. After all, the wider pool of insiders (party members, activists, party officials, elected public officials) is a quite heterogeneous group in terms of policy preferences (Bäckersten 2022;Ceron 2015;Greene and Haber 2016;May 1973;Van Haute and Carty 2012;Van Holsteyn et al. 2017), goal orientations (Müller and Strøm 1999) and skills (Alexiadou 2016). The ex-ante criterion of self-selection and socialisation into a relatively broad set of party norms and values might therefore be too weak to contain the risk of agency loss effectively. The significantly smaller subgroup of the party elite, however, will have internalised the party's policy agenda to the highest degree (Cox and McCubbins 1993) and will prioritise its implementation in government due to direct involvement in the development of policy programmes (Rose 1974: 414). 1 In addition, while lower-level party members have ongoing non-political careers in tandem to their political ones, professional politicians in the party elite are particularly vulnerable to the threat of withholding office in the future, which provides the party a valuable tool for ex-post control (Andeweg 2000b;Cox and McCubbins 1993;Müller 2000). With some qualifications, the available empirical evidence on the subject supports this perspective, as high-ranking party officials are indeed the most effective ministers in terms of implementing the party agenda in government office (Alexiadou 2016). Rose's (1974) seminal account and others have accordingly assessed the 'partyness' of ministerial appointments based on the share of members of the party elite in the cabinet (Blondel and Cotta 2000;Helms 1993;Müller and Philipp 1987). The more cabinet positions they fill, relative to lower-level party members and outsiders, the stronger party control is over government actions. Applying this more rigid criterion, parties still appear to exercise meaningful control overall. According to De Winter (1991: 44), over 40% of all ministers in 13 European democracies held positions in their respective national party organisations. Müller and Philipp (1987) report similar numbers among Austrian ministers, and more recent evidence on Swedish, Portuguese and Belgian cabinets indicates that national party office continues to be an important stepping stone to government office in the 21st century (Dowding and Dumont 2009). Descriptively, these studies show substantial variation between countries and over time (Alexiadou 2016;Andeweg 2000b;Dowding and Dumont 2009). While some of the cross-sectional variation can be attributed to institutional differences between political systems (Andeweg 2000b;Blondel 2000;Cotta 2000), there is (again) no conclusive evidence of a secular downward trend in the 'partyness' of government appointments (Dowding and Dumont 2009). Strikingly, short-term national swings in party elite appointments, which are evident from the literature, are yet to be studied systematically.

The intra-party bargain over ministerial appointments
The theoretical starting point to understand short-term dynamics in the 'partyness' of appointments is to acknowledge that party-government relations are essentially an intra-party affair in parliamentary democracies (Blondel and Cotta 2000;Laver and Shepsle 1990). As parties cut across institutions, different components of the party (its membership organisation, the party central office, the parliamentary party and eventually the party in government) face different sets of opportunity structures, causing various tensions between different 'strata' of the party (Cotta 2000: 94;May 1973;Müller and Strøm 1999). When the party enters government and individual party nominees get government posts, a new component of the party, with its own set of constraints and opportunities, temporarily emerges (Cotta 2000). The party's government ministers then have access to administrative resources, can shape government policy, make appointments in the public service as well as the state sector, and distribute patronage. As their interests in these fields will systematically diverge from those of the party in central office due to different opportunity structures, they have incentives to break the chain of delegation (Müller 2000), either by gaining autonomy from party institutions or by even reversing the relationship, instrumentalising the party organisation (Andeweg 2000b;Blondel 2000).
In this way, systematic tensions between intra-party actors continuously 'threaten' the party government model and shape the party-internal processes of ministerial selection. The national party elite typically negotiates nominations for government office with the party leader, who will, in most instances, also preside over the party's team in the cabinet (Andeweg 2000b: 131). The party leader thus plays a double-role in these appointment processes. For one, they stand at the helm of the extra-parliamentary party organisation and will -as a nominator/selector -have a strong say in the selection of government personnel (Strøm et al. 2010). Secondly, however, they are a natural choice for the highest government office available to the party and thus a set nominee. As such, the party leader may steer the intra-elite bargain over the selection of their cabinet peers to increase their own (and the government's) autonomy in office (Andeweg 2000b).
When making these decisions, the selectors do not have complete freedom of choice. In addition to constitutional constraints (such as recruiting from the pool of MPs in Westminster systems or formal prerogatives of the head of state in the selection process), the final group of ministers on which the selectors agree has to fulfil certain criteria of representativeness (Andeweg 2000b: 126;Rose 1974: 363). Amongst other things, they have to arrive at a 'balanced package' in terms of gender (Goddard 2021), region and faction (Ceron 2014;Ennser-Jedenastik 2013;Mershon 2001). The competence of ministerial candidates, their age and their policy preferences (Bäck et al. 2016;Kam et al. 2010;Müller 2000;Rose 1974) are additional factors to take into account. To comply with these criteria, insiders are the standard recruitment pool for government office (Andeweg 2000b: 133), allowing the selectorate to draw on ample information about various characteristics of party nominees, while uncertainty is higher when appointing outsiders (Müller 2000;Strøm 2000). Selectors generally benefit from this informational advantage and the insider pool will usually be large enough to find suitable candidates. Even if the party leader occasionally wants to use outsider appointments to proactively shift the party position on specific issues (Alexiadou and Gunaydin 2019), or seeks to appeal to voters by appointing 'celebrities' (Street 2012), excessive deviations from the in-house recruitment norm will likely generate costly resistance throughout the party organisation (Strøm 1990).
In contrast, the decision over the integration of party elites into the cabinet should be divisive. Party elites will often put pressure on the party leader to recruit ministers from the top of the party hierarchy. Such appointments tie ministerial decisions closely to the party elite and reward elite members personally; after all, a ministerial post would be the highpoint of most political careers. Hence, it is likely that party elite members' office ambitions (Müller and Strøm 1999) are the catalyst for party control over government. 2 By contrast, for the party leader, appointing from the party elite carries substantial risks, which incentivises them to avoid this if possible. The leader will generally prefer candidates who choose loyalty to them over party loyalty (Rose 1974: 363) in order to consistently pursue the coalitional and electoral strategies they deem optimal for their political survival. While these strategic decisions will sometimes require the leader to carefully 'bend' the party line or to disregard certain factional interests (e.g. for electoral reasons), party elite ministers will constrain the leader's ability to do so. Not only are they the most effective party agents in government overall (Alexiadou 2016), they are also held accountable by various intra-party selectorates (see Online appendix, Section 2), which will at times lead their preferences to collide with those of the party leader. In these situations, the leader's options to keep party elite ministers on course will be limited due to the independent power base such heavy-weights have within the party organisation. As the party leader also directly depends on party elites' support in leadership elections to remain in office, members of the party elite might even utilise their intra-party standing (e.g. formal decision-making powers, mobilising potential) to pressure the leader into following their demands. What is more, by adding government office to their pre-existing standing in the party, elite members might evolve into more dangerous challengers for the party leadership in the future.
Given these divided preferences in ministerial selection, I expect the share of party elite members in the cabinet, and thus the level of party control over the government, to result from a bargain between the party leader and the other selectors from the party elite. Notwithstanding factional and regional intra-party divides (which will exist within and outside of the party elite), the stronger the leader's bargaining position relative to that of the party elite, the fewer candidates should be appointed from this group. Based on the literature, party performance should be the main determinant of the power balance between the selectors (Abou-Chadi and Orlowski 2016; Adams et al. 2004;Fagerholm 2016;Gauja 2016;Harmel and Janda 1994;Meyer 2013;Panebianco 1988;Schumacher et al. 2013). Party leaders who are unable to provide in terms of the party's policy, office and electoral goals are likely to see their intra-party support coalition crumble, and thereby run a heightened risk of being replaced (Andrews and Jackman 2008;Bueno de Mesquita et al. 2003;Bynander and t'Hart 2007;Ennser-Jedenastik and Müller 2015). To maintain intra-party support despite malperformance, party leaders have to compensate intra-party actors by granting them more influence in decision-making (Gauja 2016;Greene and Haber 2016;Kaltenegger et al. 2021) and access to public office (Strøm 1990). While better performance should thus strengthen the party leader's position in the bargain and decrease the 'partyness' of appointments, weaker performance will lead the party leader to 'buy back' the support of party elites by accepting their demands for government jobs. Party elite appointments should therefore be more likely when the party performs relatively poorly in competition.
For one, the party's electoral performance will trigger these dynamics. Electoral success is key to a party's access to office and policy influence (Müller and Strøm 1999). As such, it is instrumental for achieving the party organisation's common goals, as well as for elite members' ambitions for public office. Electoral success is thus a 'public good' in terms of the party organisation, but one which has important repercussions for the 'private goods' that a party leader may distribute among party elites. Party leaders with reduced access to these goods, on account of poor electoral performance, will have to compensate party elites by integrating them into the cabinet. H1: The less successful party leaders are electorally, the more ministers are drawn from the party elite.
Office achievement will have similar effects on the party leader's leeway in ministerial selection. While electoral success is of mere instrumental value to parties, government office has the intrinsic value of granting power, resources and prestige -in fact, probably the most generous set of spoils available in parliamentary democracies -to a relatively small number of appointees (Müller and Strøm 1999). In this way, government posts are important 'private goods' that a party leader may selectively allocate to consolidate their intra-party support coalition. As party elites are particularly office-oriented (Müller and Strøm 1999;Strøm 1990), they will react strongly to any perceived shortage in the supply of government jobs (Ennser-Jedenastik and Schumacher 2015; Kaltenegger et al. 2021). Moreover, executive office has the added instrumental value of giving the party the best opportunity to implement party policy. It is therefore also a 'public good' , which is appreciated, to varying degrees, by all echelons of the party. Failing to provide policy influence via government participation will thus weaken the party leader's backing throughout the party hierarchy.
Certainly, office is a constant when studying governing parties. All party leaders have succeeded in getting their party into government through electoral performance and/or coalition negotiations. By convention, even the party's specific role in the government (e.g. junior or senior coalition partner) is largely determined by its relative electoral strength. However, there is meaningful variation across governing parties in terms of the price the party leader is required (and eventually is willing) to pay for government participation. Specifically, when forming a coalition government, the party leader can be more or less successful in getting the most, as well as the most valuable government posts out of a coalition deal. Intra-party actors will scrutinise these deals based on the number and the value of the ministerial portfolios that the party attains; and they will demand, at least, an appropriate share of the spoils and sufficient policy influence for the party. Party leaders who, in the perception of party elites, make too many concessions to the coalition partner -giving up too many or too important ministries for the sake of government participation -will have to compensate party elites in the ministerial appointment process. Conversely, the leader's autonomy in this process, and therefore opportunities to appoint non-party-elites, will be greater when the outcome of coalition negotiations is more favourable to the party.
With regard to the intrinsic value of office achievement, each ministerial post, as well as the type of posts the party receives, directly affect party elites' access to spoils. The more ministries the party controls and the more important, prestigious and resourceful these ministries, the better the outlook for party elites. Hence, the less favourable the deal the leader has brokered, the more elites will demand compensation in the distribution of the available spoils.
H2: The less successful party leaders are in securing office spoils for the party, the more ministers are drawn from the party elite.
Likewise, the more leaders fail to provide instrumental ministries to implement party policy, the more they will have to consolidate their intra-party support coalition by following party elites' demands for government jobs.
H3: The less successful party leaders are in securing policy influence for the party, the more ministers are drawn from the party elite.
A plausible alternative explanation for correlations between party performance and party elite appointments would be that the party leader forces party elites into government in hard times, in order to show a disgruntled rank and file a united leadership front. This might be necessary if party elites, despite their natural desire for office, decline positions in the current cabinet to avoid association with an unsuccessful leader, which could harm their long-term career prospects. From this perspective, the driver of party elite appointments might be the party leader managing discontent among the rank and file. The hypotheses at hand at least allow the plausibility of the two arguments to be weighed against each other based on variance in goal orientations across the party hierarchy (Müller and Strøm 1999). Specifically, if the party leader, shielding themself from rank-and-file unrest, were the dominant mechanism, party elite appointments would be strongly influenced by the leader's success in securing the rank and file's prioritised goal: policy influence (H3). Similarly, since the party's access to spoils hardly affects rank-andfile members, support for Hypothesis 2 should be limited. Evidence in line with Hypothesis 2 and weak or no support for Hypothesis 3, by contrast, point to the bargaining mechanism due to party elites' office-orientation. While Hypothesis 1 is generally uninformative in terms of the specific mechanisms at work, equal support (or non-support) for Hypotheses 2 and 3 will also preclude inferences in this regard.

Empirical strategy
This study uses individual-level data on the appointments of Austrian ministers and junior ministers from 1945 to 2017, excluding only the non-elected provisional government of the immediate post-war months. As a textbook party democracy, traditionally having 'parties that play an important role in society and governments that have brought large sectors of the state under their control' (Andeweg 2000a: 48), Austria is a likely case to find patterns of party government. While its institutional characteristics generally resemble those of many other parliamentary democracies, specific features of its constitutional framework (e.g. the PR-electoral system) additionally favour the party government mode (Müller and Philipp 1987). 3 Given these characteristics of the Austrian case, findings indicating that successful leaders may effectively limit the 'partyness' of appointments should have high external validity. Conversely, generalisability should be more limited for results suggesting stable party influence over appointments.
Data on individual ministers have been collected via the official website of the Austrian Parliament, 4 Wikipedia 5 and the Munzinger biographical database. 6 The data set is complemented by additional biographical information from the AUTELITE (Müller et al. n.d.) and the Party Congress Politics projects (Müller and Kaltenegger forthcoming). Statistical analyses use a dichotomous dependent variable, indicating whether a member of the national party elite (vs. non-party-elite) is appointed. Ministers with membership in national party executive bodies (this group typically includes deputy party leaders, the leader of the parliamentary party, leaders of regional branches and functionally defined subgroups) before their initial appointment to government qualify as party elites (Blondel and Cotta 2000;Helms 1993;Müller and Philipp 1987;Rose 1974). Government personnel that do not fall into this category are coded as 'non-party-elite' , regardless of formal party membership, lower-level party office or public office. While the organisational structure of the parties studied is relatively stable over the observation period, the measure is sensitive to changes due to organisational reform.
The main independent variables are the party leader's record in the last general election and two variables measuring the spoils and policy influence components of office achievement. Since the party leader's standing will crucially depend on intra-party actors' perception of their performance (Ennser-Jedenastik and Schumacher 2021), rather than a strictly absolute assessment, the measures applied take the likely expectations of intra-party stakeholders into account. In this vein, I operationalise electoral performance as the change in a party's vote share between the last two general elections. Negative values represent vote loss and positive values indicate vote gains for the party. For appointments during a party leader's 'grace period' , where they have not yet competed in a general election -and are therefore not responsible for the party's previous electoral performance -the variable is set to the neutral value zero (Ennser-Jedenastik and Schumacher 2015).
Building on the literature on portfolio allocation, I operationalise the party leader's success in securing office spoils as the difference between the share of parliamentary seats the party contributes to the government's majority and the party's portfolio-weighted share of cabinet posts. This is because intra-party actors will adapt their expectations to the party's bargaining position in coalition negotiations. In these processes, proportionality of a party's share of ministerial portfolios to the share of parliamentary seats it provides is the main criterion to evaluate the fairness of a coalition deal. It is actively used as a bargaining convention in coalition negotiations (Bäck et al. 2009) and the final allocation of portfolios among coalition parties is typically very close to proportional (Gamson 1961;Warwick and Druckman 2001). Party elites, demanding an appropriate piece of the pie for their party (Browne and Franklin 1973;Ecker and Meyer 2019), will therefore judge the party leader's success in securing office based on the leader's ability to claim a fair share of cabinet posts in coalition negotiations. Party leaders who manage to exceed intra-party stakeholders' expectations by securing a surplus of government posts (relative to the party's seat contribution) will be perceived as winners, while those who receive less will be branded as losers (Ecker and Meyer 2019). Positive values on this variable again mark a surplus of government posts, while negative values indicate that the party is undercompensated. The value for parties receiving cabinet posts in perfect proportion to their seat contribution is zero, which is always the case for single-party governments and (analogous to the electoral performance variable) if the party leader took over the leadership position after the government was formed. To account for objective differences in spoils (e.g. relative power, resources and prestige) that the various ministerial portfolios offer, I use Druckman and Warwick's (2005) expert-survey based portfolio weights.
Finally, I operationalise party leaders' ability to provide instrumental ministries for the realisation of party policy, building on the work of Bäck et al. (2011;Ecker et al. 2015;Ecker and Meyer 2019). Policy areas are matched with ministerial portfolios based on their formal competencies. Using coded manifesto data provided by the Austrian National Election Study (AUTNES) (Dolezal et al. 2016;Müller et al. 2012), I then assess the relative importance of a ministry for a particular party at a specific point in time by calculating the share of manifesto core-sentences dealing with the respective policy areas. I use the sum of these salience scores across all portfolios the party receives in a cabinet to measure the party leader's performance in securing policy influence. The resulting variable reflects the total share of a party's policy intentions that is covered by its cabinet posts. For single-party governments, where the governing party controls all portfolios, the value is set to 100. Again, if the party leader did not yet participate in coalition negotiations, I assign a neutral value, which in this case is the mean value of policy influence across all cabinets. I use absolute levels of policy influence in the regression models, instead of adjusting the measure to the party's parliamentary seat contribution, because portfolio allocation does not follow the rules of a zero-sum game when portfolio payoffs are valued differently by the coalition parties (Bäck et al. 2009;Browne and Feste 1975;Ecker and Meyer 2019;Laver and Shepsle 1996;Warwick and Druckman 2001). As gains for one coalition party are not necessarily associated with losses for the other, intra-party actors cannot draw on the proportionality heuristic (Ecker and Meyer 2019) to evaluate the outcome of coalition negotiations. I therefore assume that intra-party actors will react to the overall potential to enact party policy.

Controls
Besides the main independent variables, the analysis controls for the party's specific role in government (single-party government, senior/junior coalition partner) to account for uneven distributions of the performance measures across these groups (Browne and Franklin 1973;Ecker and Meyer 2019) and to pick up potential independent effects on party elite appointments. Single-party government, in particular, should increase the attractiveness of ministerial office for party elites (Cotta 2000: 69), promoting fusion of the cabinet and party executive bodies (Andeweg 2000b;Blondel 2000). I further control for the absolute number of posts that a party holds in a particular cabinet as access to more ministries will enable the party leader to appoint the number of party elite members necessary to consolidate their (the leader's) intra-party backing, while giving them greater freedom of choice in the selection process for each additional post. A higher number of cabinet posts should therefore decrease the likelihood of party elite appointments (Andeweg 2000b: 137).
Since more valuable portfolios should offer more attractive career opportunities to party elites (Cotta 2000: 69), I account for each individual portfolio's spoils value (objective portfolio salience) (Druckman and Warwick 2005) as well as its value in terms of policy influence (subjective portfolio salience) (Bäck et al. 2011;Ecker et al. 2015), expecting that party elite appointments will be more likely the higher the payoffs the portfolio provides (Diodati and Verzichelli 2017;Rose 1974) and the more instrumental the portfolio.
Moreover, I control for selector characteristics that should affect the likelihood of party elite appointments (Andeweg 2000b). First, party leaders who have risen through the ranks of the party will have a larger network of trusted allies within the party elite than relative newcomers, and should therefore be more likely to support party elite appointments. Accordingly, I include a dummy variable indicating whether the party leader was themself a member of the national party elite before their appointment to government office. Secondly, the smaller the group of selectors, the easier it is for party leaders to buy themselves freedom of choice in ministerial selection by paying off a sufficient number of co-selectors. While larger selectorates will require party leaders to provide 'public goods' , benefitting wider strata of the party organisation, smaller selectorates will incentivise party leaders to offer individual selectors ministerial posts ('private goods'), thus fostering party elite appointments (Bueno de Mesquita et al. 2003). Drawing on information from party statutes, the variable shows the number of members in party bodies that were formally responsible for nominations to government office at a specific point in time. Since the selectorates in the cases studied are almost exclusively party executive bodies, this control variable also picks up variations in the supply of party elite candidates for ministerial office.
Furthermore, appointments in cabinet reshuffles might differ from the first nomination round (Blondel 2000;Krouwel 2012), due to smaller numbers of ministerial posts to re-assign, and because they are decoupled from previous elections and coalition formation. For instance, the party leader might be able to manage upcoming replacements in advance in these situations, thus minimising the involvement of the party elite. To address this issue, I include a dummy variable for appointments in cabinet reshuffles in Models 1-4 and omit these cases in Model 5, limiting the analysis to the initial round of appointments made in each cabinet. Party fixed effects are included in all regression models to control for residual inter-party variance stemming from party-organisational features as well as party-specific recruitment conventions. Finally, 20-year period fixed effects account for potential secular trends or periodic patterns. The Online appendix provides additional information on key variables, descriptive statistics, anecdotal evidence and robustness checks.

Analysis
In order to test the hypotheses outlined above, I run binary logistic regressions with individual appointments as the unit of analysis and the appointee's party elite status as the dependent variable. Random effects on the cabinet level and cabinet-clustered standard errors are included in all regression models to account for unobserved heterogeneity between cabinets and interdependence between appointments to the same cabinet. 7 The respective party leaders are excluded from the regression analysis as their role in the cabinet is, de facto, beyond debate.
Results indicate that a party leader's electoral and office achievement records indeed affect the 'partyness' of ministerial appointments (Table 1). Electoral performance has significant negative effects on party elite  appointments in Models 1, 4 and 5, thus supporting Hypothesis 1. With each percentage point of the vote the party gains, the probability that a cabinet member is drawn from the party elite decreases by 0.7% (Model 4) on average. Given the empirical range of the variable, this effect is substantial in size ( Figure 1). 8 In line with Hypothesis 2, party leader success in securing spoils also significantly decreases the likelihood of party elite appointments (Models 2, 4 and 5). A 1% surplus in spoils, relative to the party's seat contribution, renders party elite appointments 0.9% less likely (Model 4), again amounting to sizable differences across the variable's range (Figure 2). Regarding Hypothesis 3, regression results are inconclusive. While the party leader's ability to provide instrumental portfolios has the expected negative effect on party elite appointments  in Models 3, 4 and 5, coefficients do not reach conventional levels of statistical significance (Figure 3). Among controls, being the larger coalition party does not affect party elite appointments in a systematic way, while single-party government has a strong and significant positive effect, as expected. By tendency, the results also show a negative correlation between party elite appointments and the number of government posts a party controls. Surprisingly, there is no consistent evidence that portfolio value (objective or subjective) significantly affects party elite appointments, although the coefficients for portfolios' spoils value are positive in all models, as was expected. While the party leader's route to office has substantial and highly significant effects on party elite appointment in all models, the coefficients for selectorate size have the expected negative sign but are less pronounced overall. Cabinet reshuffles do not affect party elite appointments in a significant way (Models 1-4), nor do they appear to bias the regression results (Model 5; Online appendix, Section 5.3). Finally, party dummies suggest significant residual differences between parties and period dummies show periodic swings in the 'partyness' of ministerial appointments but do not indicate a secular decline.

Conclusion
This article advances a theoretical argument on how intra-party power dynamics condition ministerial selection and tests its implications using individual-level data on ministerial appointments in Austria . It presents a rare quantitative assessment of the factors driving party elite appointments to government office, which engages with theories on party government and party organisation. Statistical analyses support the core argument that a party leader's success in securing party goals is an important short-term driver of the 'partyness' of ministerial appointments. Party elite appointments are the more likely, the less successful party leaders are electorally and in terms of securing office spoils. Only the party leader's success in providing instrumental ministerial portfolios for the implementation of party policy does not reduce party elite appointments in a significant way.
These results are in line with the view that ministerial appointments stem from a consequential bargain between the party leader and the party elite, where party leaders seek to keep party elites out of the cabinet and are more able to do so when they are more successful in terms of party competition. The alternative explanation, that the party leader forces reluctant party elites into government to divert the blame for suboptimal performance, might still occasionally apply. However, results for the two office achievement variables suggest that the party leader conceding cabinet positions to office-seeking party elites is the more plausible explanation overall. A more thorough qualitative examination would still be required to disentangle the specific mechanisms at work. Moreover, despite high external validity of these findings due to Austria's characteristics as a likely case, there are also natural limits to any case study. Specific features of the case, such as little alternation in government, the convention that the party leader always assumes a government role or idiosyncrasies of Austrian parties' organisational structures affect the incentives of the actors involved in the bargain over ministerial appointments and might therefore impact the external validity of the findings. Comparative research is needed to confirm the inferences drawn.

Notes
1. Note that this does not imply complete homogeneity of the party elite group. 2. The individual preferences of party elite members will depend on the attractiveness of their current public and party offices relative to a potential government position (Cotta 2000: 69). In most instances, the latter will provide generous benefits. 3. Despite its semi-presidential constitutional framework, the Austrian political system can clearly be categorized as parliamentary. In government formation, the president's involvement is largely a formality. Also note that ministerial nominations are unconstrained by any formal requirements with regard to ministerial candidates' prior political mandates or their qualification. 4. https://www.parlament.gv.at/. 5. https://www.wikipedia.org/. 6. https://www.munzinger.de/.
7. This approach suits the hierarchical data structure and allows to control for potentially influential variables on the appointment level (e.g. portfolio value, cabinet reshuffle). 8. Section 4 of the Online appendix provides a discussion on potential 'mechanical' effects of electoral performance on party elite appointments.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by the Austrian Science Fund (FWF) under Grant number P33596-G; and the Austrian Academy of Sciences (ÖAW) under the DOC-Fellowship programme.