Responsive to whom? Political advising and elected careers in institutionalized democracies

Introduction

What do voters want from their elected officials? Professional politics require no official qualifications, nor do voters always reward the skills which are typically asserted as essential for high office (Murray, 2015). While political scientists have long been interested in the career backgrounds of elected officials, there is no consensus on whether “professionalized” or “amateur” representatives are better for democracy. On one hand, a broadly “amateur” elected chamber at the mercy of an experienced executive raises concerns about democratic accountability and the mechanisms citizens have to keep their governments in check (Savoie, 1999). On the other hand, academics have lamented the rise of “professional” politicians who do not come to politics concerned by a desire to fix problems in their local community, but via politically appointed jobs, which gives prospective candidates minimal “real world” experience (Cairney, 2007; Evans, 2012; King, 1981; Walter, 1984; Weller and Fraser, 1987; Wilson, 2015). The concern is that these professionals bring different skill sets to politics (Cairney, 2007), which increasingly distance them from the concerns of the citizenry and/or the development and implementation of long-term policy (Cairney, 2007; Evans, 2012; King, 1981; Walter, 1984; Weller and Fraser, 1987; Wilson, 2015). Moreover, political staffing jobs are often controlled by parties, potentially creating a socializing environment where partisan considerations shape political aspirants’ view of politics more than civic-minded notions of representation. Politicians have always cared about getting re-elected, but the ways in which an increasing number of prospective politicians are socialized into politics may be affecting how they view their representative duties.

Political staff are often viewed with suspicion because the partisan nature of their work means their aim is to serve the needs of elected officials rather than the public. As advisors, these individuals surround elected officials. They assist them in undertaking their parliamentary and ministerial activities, making the workload of elected elites feasible. However, they also to act as gatekeepers between elected officials and other agents such as civil servants, media, and citizens. Scholars have argued that politically appointed staff can even be a threat to the impartiality of the civil service and the broader democratic order (Dowding and Taflaga, 2020; Eichbaum and Shaw, 2008), because they can complicate lines of accountability and their presence can negatively alter the decisions of government actors. Indeed, at least some of the apprehension toward the role of political staff also relates to their motivations—specifically, the (mostly untested) normative claim that they perform their duties because they see their role as a stepping-stone into elected office, rather than because they strive to contribute to good governance (Tiernan, 2011). Given this, who political advisers are and what they believe about how to represent and legislate for the public, is important.

Should citizens (and/or party selectorates) be concerned about this pathway to elected office, or are these concerns merely the inaccurate perceptions of an overly-cynical public? In this study, we ask: do former political staffers approach representation differently than candidates with other forms of pre-legislative experience? Using a cross-national survey of political candidates, we find that former advisors see representation differently to other politicians. In particular, they are more willing to prioritize their party’s interests above their own convictions and to prioritize both over the constituents they represent. These findings suggest that former advisors approach their duties in a manner consistent with the “loyal partisan” archetype of representation.

This study makes three contributions to our understanding of representative government. First, the study empirically informs our understanding of political careers in advanced democracies by shedding light on an increasingly common pathway to elected office—the rise of the political advisor turned elected politician. We have every reason to believe this pipeline will become more prominent as time goes on, and so it is important to understand its implications. In particular, we argue that as the demographic and experiential makeup of legislatures change, scholars should devote increased attention to how these changes may affect the broader political system.

Second, this study contributes to our understanding of mass-elite linkages between voters and their elected representatives. The question of how representatives relate to their constituents is as old as representative government. The “mandate-independence controversy” asks whether representatives should be “delegates” of their voters’ preferences or “trustees” of their voters’ interests. This study sheds light on how representatives of different backgrounds—in particular, those who have previous experience as political staff/advisers—view their responsibilities, and in turn, how they are likely to behave while in office.

Finally, our results have implications for the increasing personalization of politics in modern democracies. While personalization can be either centralized (giving power to a single party leader or executive) or decentralized (giving power to individual legislators)—both at the expense of political parties—decentralized personalization has received much less attention in the literature (Balmas et al., 2014). Our study contributes to our understanding of why decentralized personalization appears to be largely absent from advanced democracies—leading to either the increasing prominence of parties or leaders at the expense of rank-and-file legislators.

The rise of the professional in politics

Political staff are relatively new and emergent actors in executive government, and their work is shaped by the institutional settings in which they operate. Furthermore, their work roles and responsibilities have evolved along different paths depending upon the jurisdiction (Shaw and Eichbaum, 2018). Thus, while a single unified definition of what constitutes a political advisor is hard to come by, there is a broad consensus that a politically appointed staff member is a “temporary public servant appointed to provide partisan advice to a member of the political executive who is exempt from the political impartiality requirements that apply to the standing bureaucracy” (Shaw and Eichbaum, 2018: 3).

As Connaughton (2010, 2015) has argued, politically appointed staff undertake four main types of functions: they are either experts, partisans, coordinators or minders. This typology reflects the fact that the work of staff encompasses a broad range of activities from policy work, partisan strategy to administrative, and emotional support roles (Maley, 2000). However, some staffing roles, such as media advising, are highly specialized and expertise in this field may require additional experience beyond institutional party politics. Indeed, as Fisher (2016) notes, media advisers, particularly when they have industry experience, can be treated skeptically by ministers and other political staff because they lack sufficient background in the institutional and partisan world of party politics. This reflects the fact that some advisor career paths may be more advantageous than others.

There is growing recognition that politically appointed staff are policy actors in their own right and have become institutionalized within executive governance (Craft, 2013, 2015; Eichbaum and Shaw, 2007; Gouglas et al., 2017; Maley, 2000, 2015; van Den Berg, 2017). As political staff have accrued more responsibilities that used to fall upon ministers, they have complicated lines of accountability and in some cases even crowded out senior bureaucrats (Christiansen et al., 2016). Political staff’s policy making roles include agenda setting; linking ideas, interest and opportunities; mobilizing; bargaining and; “delivering” (Maley, 2000). The roles that advisors play depend on the institutional context. In some jurisdictions, staff may be critical to developing policy ideas and proposals (Askim et al., 2017; Craft, 2016) while in others their function is to co-ordinate actors to achieve policy outcomes (Gouglas et al., 2017; Hustedt and Salomonsen, 2017). In both cases, political staff are central to initial policy making and deal-making within and across governments.

In each of these roles, staff have the opportunity to understand the governing process from the inside and gain experience in skills that are directly relevant to the work of a parliamentarian. This experience may give politically advisors who continue into elected politics an advantage over other political aspirants. There is evidence to suggest these mechanisms are also gendered. For example, in Australia, women make up more than 50% of staff that have ever served, but are far more likely to occupy lower status roles and be less proximate to the minister (Taflaga and Kerby, 2020). In Canada, women staff are more likely to perform administrative work, while men are more likely to perform policy work (Snagovsky and Kerby, 2019). Evidence from Denmark and Sweden suggests that women are more likely to occupy roles where they do not discuss high politics and strategic factors (Christiansen et al., 2016). However, the case appears to be the different in the Netherlands where gender is more balanced (van Den Berg, 2018). Thus, even though time as a politically appointed staff member may be an opportunity to gain valuable skills that are useful in a later political career, the capacity to capitalize upon them may not be equally distributed across advisor cohorts.

Political parties are central to the rise of staffers-turned-candidates. Parties are key gatekeepers controlling access to parliamentary careers. Not only are political parties responsible for selecting and supporting candidates to elected office, party actors control access to political advising jobs to begin with. As Eichbaum and Shaw (2011) and Maley (2011) have argued, the world of the ministerial office is hierarchical and exists as part of a wider link of partisan networks. Just like other professional domains, access to opportunities within political careers is not evenly distributed across all actors (Norris and Lovenduski, 1995; Verge and Claveria, 2016). In the case of politically appointed staff, it is not just skills related to the business of government that may give this group of actors an advantage—it may be the political networks that they develop while undertaking their duties.

In recent decades, political scientists have described an increase in the overall number of elected representatives with a previous background as political staff across Westminster regimes (Evans, 2012; King, 1981; Wilson, 2015). Similar trends were observed even earlier in Australia (Walter, 1984; Weller and Fraser, 1987). Studies of party leadership candidates in the UK suggests that the occupational profile of these most successful political actors has changed markedly in recent years—they appear to lack “life” experience, but have accrued very high levels of “political” experience (Barber, 2014; Cowley, 2012).

This trend also appears to be growing among elected political elites more generally. Evidence from the UK suggests that MPs with a background as politically appointed staff are different from other political candidates (Goplerud, 2015). Political advising is not a long-term career for most staff in the UK. In this environment, there are fewer special advisers, and staff typically serve for around 3 years. This appears to be sufficient time to build up relevant skills and networks (Yong and Hazell, 2014). Indeed, just over a quarter of all special advisers from the Thatcher and Major years went on to run for office, suggesting that perhaps staff are more ambitious than individuals from other potential candidate pools (Goplerud, 2015: 340).

Evidence from the UK suggest that former politically appointed staff also appear to have more success after their initial selection. This group is more likely to run in safer seats, which almost doubles their overall success rate (Goplerud, 2015). As Allen (2013, 2014) argues, political staff are more likely than people with “instrumental” backgrounds—that is, career backgrounds related to politics such as journalists, trade union officials and think tank workers—to progress further up the ministerial ladder and faster. Evidence from Germany reflects a similar pattern, suggesting staff are able to capitalize and exploit networks they have built up to both be selected and elected to higher office (Ohmura et al., 2017). However, recent research from both Germany (Ohmura et al., 2017) and Canada (Snagovsky and Kerby, 2019) finds that the capacity to participate fully in these networks is more favorable to men rather than women. Finally, there is evidence that experience as a political advisor can be helpful in securing employment after they leave politics. However, this “attractiveness in the job market is less a result of what they actually do while in office and more a question of the insider knowledge and networks they acquire” (Askim et al., 2020: 2). Thus, years spent advising politicians appears to not only give these political aspirants the skills and networks to get ahead, but also a different socialization into the world of politics when compared with other candidates.

Political advisors as representatives

Does it really matter whether former political advisors get elected to public office? Former advisors are likely to have more skills than other newly elected legislators, allowing them to hit the ground running when it comes to performing the duties of their office. In this sense, a voter might see electing a former advisor as the smart choice, since they are likely to be more effective than candidates without political experience. Are former advisors likely to use these skills and networks to further their constituents’ interests, however? Indeed, we have good reason to think that former political advisers bring more than their expertise in the process of politics, but may even approach their responsibilities differently in the first instance.

A central question each representative must confront is to whom they should be responsive. One difference between former advisors and those with other types of pre-legislative experience might be their views on the nature of mass-elite linkages in representative style (Converse and Pierce, 1986). A classic view of this question is the so-called “mandate-independence” controversy—that is, whether a representative should approach their responsibilities as a “delegate” or “trustee” of their constituents’ interests. The “delegate” mode of representation sees representatives as proxies for their electors: delegated to “represent their express preferences to the elected chamber, with relatively little room for discretionary decision making by the legislator” (Anderson and Goodyear-Grant, 2005: 1031). By contrast, the Burkean representative-as-trustee is committed to the broader national interest or their own judgment of the public good: “the representative may consult but has no strict obligations to act in accordance with the opinion of the constituency” (Anderson and Goodyear-Grant, 2005: 1032). Several arguments have sought to extend representation beyond the mandate-independence dichotomy. For example, Eulau et al. (1959) describe a third category of “politico”: a representative who alternates between the delegate and trustee interpretations of their responsibilities depending on the context. Converse and Pierce (1986: 494–9) describe a continuum between the two “extremes” of the delegate-trustee dichotomy where most legislators are likely to reside.

In parliamentary democracies, the idea of representatives with “personal mandates” is at odds with the incentives that political parties create to promote cohesion. As Rahat and Kenig write, “is not because legislators from the same party hold exactly the same opinions on every issue, but because legislators, at most times, accept the framework of the party, which obliges them to vote together” (2018: 224). This creates a third pole towards which legislators can gravitate—prioritizing the needs of their political party over both their personal preferences and constituents’ desires. This type of legislator is known as the “loyal partisan” (Andeweg and Thomassen, 2005: 508).

Each of these models of representation is associated with different conceptual outcomes. Under the delegate model of representation, governments and representatives should dynamically change their positions depending on what a majority of their constituents think. Under the trustee model, representatives should “stick to their guns” according to their own perceptions of the public good. Meanwhile, the loyal partisan model dictates that legislators should defer to their parties, since it is these actors who rightly determine legislative outcomes. These modes of representation describe a trade-off between policies that representatives themselves consider “good” versus those that are “responsive” to other actors—in this case constituents or political parties (Franklin and Riera, 2016). Moreover, these models lead to different levels of congruence between citizens and representatives, with a high number of “loyal partisans” in a system being associated with lower congruence between parties and their voters (Önnudóttir, 2014). Finally, these conflicts are not purely theoretical. For example, in their landmark study, Converse and Pierce (1986) find an overwhelming majority of French legislators have experienced conflicts between the various dimensions of their representative roles. In their study, approximately two-thirds of legislators reported conflicts between their personal convictions and their party’s position, and similar numbers have experienced conflict between their personal conviction and their constituents’ preferences.

The phenomenon of former advisors becoming elected officials also informs the personalization of politics. Personalization refers to a “process in which the political weight of the individual actor in the political process increases over time, while the centrality of the political group (i.e., political party) declines” (Rahat and Sheafer, 2007: 65, emphasis in original). Here, there is an important difference between centralized and decentralized personalization: the former “implies that power flows upwards from the group (e.g., political party, cabinet) to a single leader (e.g., party leader, prime minister, president)” and the latter implies “power flows downwards from the group to individual politicians who are not party or executive leaders (e.g., candidates, members of parliament, ministers)” (Balmas et al., 2014: 37).

Aside from the idea of a personal electoral mandate, personalization also concerns the legislative behavior of individual politicians. Here, individual politicians gain increased prominence at the expense of collective party unity (Rahat and Sheafer, 2007). Thus, decentralized behavioral personalization manifests when backbench legislators increasingly engage in private legislation or otherwise step back from the party and its discipline (Balmas et al., 2014). Individualized behavior can threaten party cohesion if interpreted as public dissent and “may lead to organizational tension, the amendment of government bills, electoral misfortune, or the replacement of one set of leaders with another” (Kam, 2009: 10). While these forms of dissent are rare, they indicate a “tip-of the iceberg phenomenon where for every MP who votes against the party or speaks out in the media, 10 more unhappy MPs are believed to exist” (ibid). Here, there should be a clear link to representative style—a greater number of MPs who prioritize their party’s views over their own or those of their constituents should make decentralized personalization less likely. In other words, more representatives who subscribe to the “loyal partisan” mode of representation should lead to less power for individual legislators, and more power for political parties or party leaders.

Political advisors acquire skills and develop partisan networks through their time in office, where through their work, they are socialized in the norms and procedures of their respective partisan worlds. Here, political advisors learn the material benefits of being a team player (Kam, 2009) and come to understand the incentive structure inherent to modern party-based democracies. In this way, staying responsive to a political party is a self-interested act and “an effective strategy to secure a seat high enough on the party list [or a safe-enough constituency] to have a chance of getting elected, and of staying in office” (Önnudóttir, 2014: 543). Given the skills, experience and networks that former advisors develop as the natural outcome of their jobs, they should therefore behave in a way that ultimately reduces decentralized behavioral personalization. That is, former adviser candidates should favor approaches to representation that place greater emphasis on party and leadership priorities, rather than their own policy priorities or those of their voters. Consistent with the “loyal partisan” view of representation, we expect:

H₁: Candidates with experience as political staff are more likely to prioritize their party’s position over both their own positions and their constituents’ positions.

In summary, political advisors are an increasingly common part of professional politics, and there is evidence that more and more individuals use political advising as a springboard to elected careers. Moreover, we should expect these individuals to approach their responsibilities differently to other candidates. In the next section, we describe how we test the role of political advising experience with reference to these theoretical expectations.

Data and methods

To understand the experiences of politicians who have experience as political advisors, we draw upon two waves of a cross-national survey of political candidates. These data are from modules 1 and 2 of the Comparative Candidate Survey (Lutz et al., 2016a; 2016b), and contain a total sample of approximately 27,000 candidates running for federal office in 21 countries over the course of 42 elections (see online appendix Tables A1 and A2 for more details). By using a common core questionnaire, the Comparative Candidate Survey (CCS) is able to measure candidates’ political backgrounds, policy positions and political attitudes across a wide range of countries. A list of variables considered in this analysis, along with summary statistics, is shown in online appendix Table A3.

Our study aims to assess the impact of prior political advisor experience on candidates’ attitudes towards representation. The main independent variable of interest is a binary variable, representing candidates’ prior paid experience as a political advisor for a legislator or political party. This measure corresponds to respondents’ answer to the survey questions, “Before becoming a candidate for the [national parliament], were you ever employed in a State or Federal MPs or minister’s office?” (Module I) and “Regarding your political experience, have you ever…worked as paid party/campaign worker or MP employee?” (Module II). This latter wording is broader, encompassing those who work in the central party as well as for an MP directly. To account for these differences, we also include a dummy variable to distinguish between the first and second waves of the CCS, as well as to account for other differences in question wording or sampling methodology (which are otherwise very similar).

We specify three models to examine the probability former staffers will prioritize various loci of representation (themselves, their constituents, their party). All three models have a binary dependent variable which corresponds to candidates’ opinions on how legislators should vote when their personal position, their constituents’ positions and their party’s position are in conflict. These variables correspond to a survey question which asks: ‘How should a Member of Parliament vote in parliament: (1) “if the voters in his/her constituency have one opinion and his/her party takes a different position?” (2) “if his/her own opinion on an issue does not correspond with the opinion of the voters in his/her constituency?” and (3) “if his/her own opinion on an issue does not correspond with his/her party’s position?”

Our study also controls for the potential impact of basic demographics, such as age, gender, and education. Overall, we expect younger, better-educated, and female candidates to have more party-centric views. This is because younger legislators have had less time than older legislators to solidify their policy convictions before entering politics, suggesting they may be more open to adopting the party line and/or they may be more susceptible to the socialization of party politics. Younger candidates also have a potentially longer political career ahead of them, strengthening the incentive to be a team player. With regards to education, better-educated candidates should be more aware of the career incentives of following the party line. By contrast, our expectations regarding gender are derived from previous literature, which suggest women legislators experience greater pressure to conform to party discipline than men legislators (Clayton and Zetterberg, 2021; Cowley and Childs, 2003). We are also able to account for the impact of specific political factors, such as whether or not the candidate was ultimately elected to office and whether their party formed government. All-else-equal, we expect successful candidates and candidates from the governing coalition to have a more party-centric outlook towards representation. Further, we control for system-level factors like the effective number of parliamentary parties, average district magnitude, and proportionality of the electoral system—each giving prospective candidates different incentives to either differentiate themselves from the party and their co-partisans or incentives to adopt the party brand as much as possible (Shugart, 2001).

Finally, we specify a series of mixed-effects logistic regression models at the country-year level to account for the shared unobservable characteristics that arise from the clustered nature of the data, such as country-specific factors which lead to some countries having more adviser/staffer-candidates than others, as well as election-specific variation.

Results and discussion

Figure 1 shows the number of candidates in CCS countries who identified their previous experience as political staff. The percentage of total candidates in the sample who have experience as advisors ranges from 6% in Finland (corresponding to 113 candidates) to almost one-third in Norway (corresponding to 338 candidates).OPEN IN VIEWER

In addition to the differences between countries, Figures 2–6 show interesting demographic differences between former staffers and non-staffers. Unsurprisingly, the average staffer-turned-candidate tends to be younger than candidates with other types of experience, and is an average 5 years younger (42.7 years compared with 47.4 years, respectively). A difference of means test confirms this difference is statistically significant at the 0.001 level. Former advisors are typically more educated than non-adviser candidates, with more having completed university (67.6% vs 63.2% for non-staffers) or started but not finished university (13.7% vs 10.7% for non-staffers). By contrast, non-staffers are much more likely to have completed a post-secondary trade or vocational qualification (12.4% vs 7.4% for staffers). When candidates do attend university, there are also interesting differences in what they study: former advisors are much more likely to have a social sciences background (41.6% vs 26.6% for non-staffers), while non-staffers are much more likely to have professional or applied science background (53.1% vs 44% for staffers). Finally, there are differences between staffers and non-staffers in terms of marital status: staffers are more likely to be single (25.4% vs 19.2% for non-staffers), while non-staffers are more likely to be married (70.2% vs 65.9% for staffers). A chi-2 test suggests each of these differences is statistically significant at the 0.001 level. Interestingly, one area where there is no difference is sex—women are distributed equally across both groups, even though they account for only a third of all candidates.OPEN IN VIEWER

OPEN IN VIEWER

Figure 3.

Candidate education in comparative candidate survey countries by former political advisor experience.

OPEN IN VIEWER

Figure 4.

Candidate university discipline in comparative candidate survey countries by former political advisor experience.

OPEN IN VIEWER

Figure 5.

Candidate marital status in comparative candidate survey countries by former political advisor experience.

OPEN IN VIEWER

Figure 6.

Candidate sex in comparative candidate survey countries by former political advisor experience.

Overall, the descriptive statistics reveal interesting demographic differences in the profiles of former staffers and non-staffers. Candidates with former advisor experience are more likely to be younger, better-educated, have a social-science background, and be unmarried compared to candidates with other types of experience. Moreover, it is possible that some of these factors may influence how former advisors approach representation. This underlies the importance of controlling for these factors in our multiple regression, so as to isolate the unique effects that political advising experience has on political attitudes.

Do former advisors approach representation differently to other candidates? Table 1 shows the results of the logistic regressions for loci of representation according to three theoretical typologies—delegate, trustee, and loyal partisan. Specification 1 reports the determinants of a candidate choosing their constituents over their party in the event of a conflict (delegate vs partisan). Specification 2 presents the results for a choice between their constituents’ positions compared to their own position (delegate vs trustee). Specification 3 presents the results for a choice between a candidate’s own position compared to their party’s position (delegate vs partisan). In all three models, coefficients are represented as odds ratios—a coefficient that is greater than 1 (even odds) means the first option in each pair is more likely to be chosen compared to the second option, while a coefficient of less than 1 means it is less likely to be chosen.

OPEN IN VIEWER

Table 1.

Attitudes toward representation among candidates in comparative candidate survey countries.

	(1)	(2)	(3)
	Voters over party	Voters over self	Self over party
Staffer	0.651*** (0.0419)	0.636*** (0.0441)	0.827** (0.0548
Elected	0.516*** (0.0361)	0.484*** (0.0368)	0.799** (0.0557)
Female	0.962 (0.0424)	1.390*** (0.0633)	0.809*** (0.0378)
Age	1.000 (0.00156)	1.007*** (0.00163)	0.994*** (0.00168)
Education	0.883*** (0.0156)	0.799*** (0.0139)	1.084*** (0.0199)
Candidate’s party formed government	0.856*** (0.0385)	0.838*** (0.0393)	0.787*** (0.0381)
Gallagher index	1.041 (0.0229)	1.036 (0.0320)	1.018 (0.0403)
Effective number of parliamentary or legislative parties	0.891 (0.0695)	0.809 (0.0884)	1.137 (0.159)
Average district magnitude	0.987*** (0.00349)	0.996 (0.00463)	0.993 (0.00575)
Comparative candidate survey Wave 2	0.944 (0.181)	1.159 (0.312)	0.692 (0.239)
N	10352	10730	10646
groups	20	20	20

The results provide strong support for our hypothesis: former advisors are more likely to view representation through the “loyal partisan” framework. Political staff-turned-candidates are much more likely to choose their party position over their constituents’ preferences. They are also more likely to prioritize their own positions over those of their constituents. Finally, they are also more likely to cede their own convictions should they conflict with their party. All three coefficients are statistically significant at the 0.01 or 0.001 level. Importantly, former advisors’ partisan views of representation occur even when controlling for a number of covariates, including who actually gets elected. This matters because elected candidates tend to follow the same pattern: candidates who are ultimately successful in their electoral bid are either already more likely to adopt the “loyal partisan” mode of representation when running (because they are incumbents, party activists or both, for example), or more likely to adopt that mindset shortly after taking office. These results suggest that the increasing number of former advisors who run for office is likely to lower the overall decentralized personalization of their political systems. However, it is unclear whether this trend will result in centralized personalization in a party leader, or depersonalization overall.

The same trends for loci of representation are evident for candidates from the party of government. More educated candidates are also less likely to prioritize their voters over their party or their own opinions, but these candidates, in contrast with political advisors, are more likely to say they would choose their own position over their party’s should they conflict. Women candidates are equally likely to say they would take their voters’ position over their party’s position if forced to choose. Further, women are less likely to prioritize their own position if it conflicted with their party, but more likely to choose their voters’ position over their own. ¹

Discussion and conclusion

Anecdotal evidence has long hinted that political advising is becoming an increasingly common pathway to elected office. However, concerns about the rise of a “careerist” political path from political adviser to politician have been subjected to limited empirical investigation. We extend the field’s understanding of a prominent pathway to public office by investigating how candidates’ prior experience as political advisors may shape their attitudes towards representation. To do so, we examined cross-national survey data of political candidates from a wide range of institutionalized democracies. We are particularly interested in how former political advisors view their representative responsibilities, which in turn indicates who they see themselves as responsive to in the fulfillment of their duties.

Does a candidate’s prior experience as a political advisor matter? We find evidence that it does. Specifically, we find former advisers/staffers make for a different kind of elected official—they are more likely to prioritize their party’s interests over their own personal convictions, and to give precedence to both over their constituents’ views. These preferences are strongly in keeping with the “loyal partisan” model of representation—exemplifying representatives who are more responsive to their party’s concerns than the priorities of their constituents (delegate) or themselves (trustees). Advising experience has an effect over-and-above other demographic and political variables, which themselves reveal interesting patterns: for example, we find candidates with a university education, those that come from the governing party, and those who ultimately get elected are all less likely to prioritize their voters’ concerns over their own concerns, or those of their party. Candidates who ultimately get elected, and those that are in government, are also less likely to stick to their guns, rather than toe the party line, in the event of a conflict, while those with a university education appear to be slightly more willing to stand up to their party if they have a strong personal preference. Gender does not appear to affect a candidate’s decision with regards to a choice between their constituents and their party, but women candidates indicated they were less likely to prioritize their own positions if they conflict with their voter’s preferences or party’s wishes.

Our study is subject to some limitations; for example, while the CCS is an invaluable tool for researchers seeking to understand how representatives think and act, some types of candidates are less likely to respond to the survey than others. While we control for as many of these attributes as we can in our model, we note that this is an imperfect solution. As a result, more research and data collection is needed to extend the scope of our analysis.

Aside from representative style and the nature of mass-elite linkages, this study also contributes to debates about the personalization of politics in modern democracies. The literature on the personalization of politics identifies several key explanations for this phenomenon. At the institutional level, Rahat and Kenig (2018) describe the mediatization of politics, the decline of mass-party linkages and the broader individualization of society as key causes of the personalization of politics. McAllister (2015) argues political parties, political leaders, voters themselves and the news media all have incentives to encourage centralized personalization, while local candidates have incentives to cultivate a personal vote through constituency service—an element of decentralized personalization. Other research shows that local candidate motivation is a critical predictor of whether decentralized personalization takes place at the local level—for example, candidates “who turn to “personalized” campaign do so because they indeed want to promote individual goals and not just party goals” (Eder et al., 2015: 325, emphasis added).

Our results contribute to this literature by adding a new explanation for why decentralized personalization appears to be increasingly absent from modern democracies. This new explanation concerns the changing nature of who goes into politics in the first place—that is, an increasingly party-centric group of candidates, whose pre-parliamentary experience has exposed them to both an alternative pathway into politics and socialized them into the “rules of the game” of their party and polity. This opportunity is distinct from candidates coming to politics via other career paths. Given former advisor’s ability to exploit their knowledge and networks, a situation arises where there are fewer actors pushing for the primacy of local concerns or personal convictions when faced with internal party pressures. In this environment, two outcomes become more likely: the depersonalization of politics—where the party brand becomes more important than any individual actor—or centralized personalization—where the power of both local candidates and political parties is increasingly vested in a single political leader. In both cases, backbench legislators have fewer incentives—or even a lack of desire—to pursue individual agendas through private legislation or to speak out against the party line. This is particularly strong where institutional settings (parliamentary rules, party culture, etc.) encourage party discipline.

Not only do these results shed light on an increasingly important pathway to elected office across many advanced democracies, they should give pause to those concerned about the quality of representation in these countries. On one hand, some commentators are rightly concerned about the dangers of an “amateur” elected chamber (Atkinson and Docherty, 1992; Docherty, 2011; i.e., Franks, 1987) and its dangers in empowering the executive at the expense of the legislature (Savoie, 1999). On the other hand, there are reasonable concerns about the rise of “careerist” politics—along with the accompanying fear that staffers on the track to elected office may be more preoccupied with their own career prospects than the public good. In these contexts, our research suggests the goal of electing more experienced legislators should not be pursued at all costs. While those with prior experience as advisors may have the skills to handle the complexities of their new jobs as elected officials and therefore more able to stand up to their party and executive, they may also be less willing to do so.

Finally, if the increase in political advisors running for elected office is at least in part the result of insider networks which give these candidates privileged access to information or resources, the success of these types of candidates is likely to further entrench and perpetuate these same networks. Ultimately, this reduces the diversity of representative bodies—favoring party operatives at the expense of those with other forms of life experience. These findings present a dilemma to party selectorates, which are part of the ecosystem contributing to changing ideas of representation amongst political aspirants and candidates. However, the dilemma of weighing different attitudes of representation amongst political insiders compared with more traditional/community-based candidates may prove insurmountable for party organizations because of the self-reinforcing incentive structures involved. In all cases above, it may be in citizens’ best interests to push back against this trend, and to demand legislators who stand up for their interests as opposed to those of their party.

Supplemental Material