A large literature in American politics suggests that political leaders often respond to the possibility of a runaway bureaucracy by assuring that key decision-making positions are staffed by individuals whose policy priorities are similar to their own. As long as politicians can secure the most important administrative positions for bureaucrats with common interests, they can also keep opposing policies off the agenda and effectively bias agency decision making in their favor.Footnote 1 Conversely, if an agency were to be captured by bureaucrats with conflicting interests, they would risk seeing their policies go unimplemented as expertise and resources are diverted for other ends. This ‘ally principle’ drives political leaders to appoint, promote and support individuals who affirm their partisan ambitions, and to replace, marginalize and subvert those who deny them.Footnote 2
In this research note, we bring the literature on agency personnel management from the US system of separate powers to the parliamentary democracies of Europe. While scholars of American bureaucracy have conveniently couched their propositions and findings in general terms, from a comparative perspective US federal agencies often appear remarkably politicized, with incoming presidents having both thousands of central agency positions to appoint ‘at-will’ upon entering office and ample opportunities to instruct agency policy.Footnote 3 In Western Europe, by contrast, incoming political leaders must typically contend with both meritocratic recruitment regulations and stringent labor laws in order to gain access to comparable positions.Footnote 4 In addition, the individuals who occupy those positions commonly operate in the shadow of civil service codes that emphasize impartiality and political independence. This combination has led many observers to conclude that European bureaucrats – unlike their American counterparts – can typically prepare, formulate and implement the law of the land without concern for their ideological relations with the parties in power. The Organisation for Economic Co-operation and Development (OECD), for example, categorically states that, in Scandinavia and the United Kingdom, not even agency managers turn over with a change in government.Footnote 5 Given the substantial differences in political and administrative institutions, can anything be learned about agency personnel management in Europe by studying analogous staffing processes in the United States?
We argue that policy conflicts can be just as relevant for understanding the composition of agency personnel in European bureaucracies as they are for understanding the composition of the US federal workforce. To buttress this conjecture, we have assembled a novel event history dataset tracing the careers of all agency heads in the Swedish executive bureaucracy between 1960 and 2014 – in total, covering 14,839 yearly interval observations of 2,355 unique appointees with 2,147 observed exits. We show that, despite serving on fixed terms and under a constitution that explicitly promotes bureaucratic autonomy, impartial implementation and meritocratic recruitment, Swedish agency heads leave their posts at considerably greater rates following partisan shifts in government. In line with reigning theories of delegation in American politics, these findings indicate that even in institutional systems seemingly designed to insulate bureaucratic expertise from political control, partisan politics can drive bureaucratic turnover. Future studies of the relationship between politicians and bureaucrats in European parliamentary democracies should thus take this possibility more seriously.
THE POLITICAL DYNAMICS OF BUREAUCRATIC TURNOVER
Our analysis is grounded in the observation that the composition of agency personnel can depend on both political screening and bureaucratic selection. By ‘political screening’ we mean the strategies that political leaders use to sort among potential agents, for example, by performing background checks prior to employment or setting up contractual terms that only some types of agents find valuable. By ‘bureaucratic selection’ we mean the strategies that bureaucrats use to sort among potential principals, for example, by applying to specific kinds of jobs or acquiring traits that only some types of principals find valuable. Our basic contention is that, as long as either political screening or bureaucratic selection is driven by ideological concerns, partisan shifts in government can induce turnover among agency personnel.
Scholars of public bureaucracy have of course long emphasized the ideological aspects of political screening, particularly in the United States.Footnote 6 Moe, for example, argues that presidents commonly reserve the positions earning the highest pay and greatest authority for individuals who share their policy views.Footnote 7 Similarly, Lewis argues that, precisely because presidents are prone to fill central agency positions on the basis of ideology, loyalty or programmatic support, incoming presidents are also particularly inclined to replace the bureaucrats who were authorized by their political opponents.Footnote 8 In Europe, Ennser-Jedenastik provides evidence that central bank governors are appointed and dismissed under similarly politicized conditions as agency leaders in the US federal bureaucracy.Footnote 9 Though bureaucrats often do strike out on their own, their motivations can also often be traced to the political leaders who granted them the opportunity to do so in the first place.
More recently, scholars have also started investigating the ideological aspects of bureaucratic selection.Footnote 10 Gailmard and Patty, for example, argue that many people enter public service specifically for the opportunity to influence policy outcomes, and that policy discretion can therefore be understood as part of the compensation schedule for agency personnel.Footnote 11 Similarly, Doherty, Lewis and Limbocker argue that bureaucrats are ‘political animals’ who value their influence over policy, understand the implications of electoral turnover, and take potential policy conflicts into account when appraising their employment situations.Footnote 12 If an agency has its powers and resources opportunistically compromised by the government, the bureaucrats who work there may well view it as a cut in their pay. And if the cut in their pay were large enough, they may no longer find it worthwhile to stay in their job. Because political decisions about bureaucratic structure and process can affect how bureaucrats perceive the value of public service relative to other career opportunities, political leaders need not necessarily participate directly in hiring and firing decisions in order to have influence over agency employment outcomes.
For the purposes of this study, we do not treat political screening and bureaucratic selection as rival explanations, but rather as parts of the same general turnover dynamic. In the screening case, a change in government policy can promote bureaucratic turnover because the parties in power may see holdover bureaucrats as representing interests opposed to their own. In the selection case, a change in government policy can promote bureaucratic turnover because holdover bureaucrats may see the parties in power as representing interests opposed to their own. In both cases, however, common interests between politics and the administration should reduce employment tensions and increase the stability of bureaucratic tenure, while conflicting interests should increase employment tensions and reduce the stability of bureaucratic tenure.
The importance of agency personnel management has not eluded political leaders. As Richard NixonFootnote 13 put it to his cabinet:
I urged the new cabinet members to move quickly to replace holdover bureaucrats with people who believed in what we were trying to do. […] [I warned that] if we don’t get rid of those people, they will either sabotage us from within, or they’ll just sit back on their well-paid asses and wait for the next election to bring back their old bosses.
While the Nixon administration may be thought of as an extreme example of administrative politicization, similar tendencies can also be found in quite different settings. When the Swedish Social Democratic Party lost control of the executive cabinet in late 1976, for example, the new state secretary in the Ministry of Education, Bert Levin of the Liberal People’s Party, likened their inaugural experience to stepping into ‘a forest of red needles’, with many of the senior bureaucrats being deeply rooted in the social democratic labor movement.Footnote 14 Levin describes the difficulties of relying on bureaucrats who, even when seemingly trying their best to act as impartial civil servants, had a hard time grasping the political objectives of the new government. Yet if bureaucrats with conflicting interests can be either forcibly dismissed or convinced to leave the agency, while bureaucrats with common interests can be either directly appointed or convinced to enter the agency, then a hostile agency can also be turned into a friendly one.
We propose that, by providing incentives for politicians and bureaucrats to reconsider past employment decisions, situations such as those faced by Richard Nixon and Bert Levin should be particularly conducive to bureaucratic turnover. Of course, the composition of public personnel will not be perfectly or immediately adjusted to reflect the new policy regime. Modern bureaucracies typically have a variety of personnel management restrictions that can limit politicians’ ability to reshuffle agency personnel, for example, and downturns in private sector labor markets can limit bureaucrats’ job opportunities and make them less inclined to exit voluntarily. If the costs of reneging on past employment decisions are substantial for both sides, they may both find it preferable to endure their disagreements. However, the sort of institutional problems often encountered by political leaders in their quest for control over the policy process need not necessarily limit bureaucrats’ ability to opt in and out of public employment for ideological reasons. And the sort of labor market problems often faced by bureaucrats need not limit the ability of enterprising political leaders to purge agencies of ideological opponents. Though context certainly matters, on the margins we expect policy conflict to be a key driver of agency personnel turnover.
DATA, COVARIATES AND METHODS
To explore the political dynamics of bureaucratic turnover empirically, we have assembled a novel event history dataset with detailed information on the biographies and positions of all agency heads in the Swedish executive bureaucracy between 1960 and 2014. By ‘agency head’ we mean the highest-ranking official in a given public organization, such as bureau directors, commission chairs, county governors, superintendents or vice-chancellors. By ‘executive bureaucracy’ we mean all public organizations whose heads are appointed by the government, save for the courts. In Sweden, this includes organizations such as universities, county administrative boards, museums, government owned businesses, police authorities, regional archives, regulatory commissions and general public service agencies, and excludes a small number of organizations accountable only to parliament.Footnote 15
For our purposes, the Swedish case has several notable qualities. First, in contrast to the US system of the separatation of powers, the Swedish constitution defines an almost ideal-typical singular chain of delegation and accountability, running from the people, acting as the ultimate principal, through the parliament, executive cabinet and bureaucracy.Footnote 16 Historically, the main contenders for government formation have been the Social Democratic Party, which dominated Swedish politics for most of the post-war period, and a liberal-conservative multiparty coalition, which first emerged as a proper alternative in the late 1970s. The winning legislative coalition appoints the prime minister, who, in turn, is free to create cabinet departments and appoint individual ministers as he or she sees fit. The caveat is that all decisions in the cabinet, such as legislative bills, executive orders and the appointment of agency heads, must be taken collectively by the ministers at formal cabinet meetings. While the political institutions thus ensure that the Swedish bureaucracy is formally accountable only to a single political principal, collective action problems are still possible due to the cabinet’s decision-making rules.
Secondly, while the Swedish parliamentary chain of delegation grants the winning legislative coalition considerable agenda control, Swedish governments also operate under a set of administrative institutions that, on paper, severely constrain their ability to manage agency personnel.Footnote 17 As in many other Western European bureaucracies, the only positions that are open to political appointment are those at the very top of an agency’s hierarchy, such as heads and management boards. The official stance is that heads should normally be appointed on six-year contracts with a possibility for three-year extensions, but not for longer than twelve years in total.Footnote 18 Once appointed, the fixed terms ensure that heads generally must be fired ‘for cause’ – that is, for some form of serious misconduct. To further complicate matters, even the upper layers of agency positions are part of a constitutionally protected meritocracy. The Swedish Instrument of GovernmentFootnote 19 explicitly stipulates that all agency personnel must be appointed based on merit, and that all agencies enjoy formal independence in the sense that only the agencies themselves are allowed to dictate how they should implement the law or rule in specific cases.Footnote 20 By constitutional design, Swedish governments are thus to check the bureaucracy primarily through structure and process, and not by directly managing its personnel and day-to-day operations.
Thirdly, although constrained, Swedish governments still have enough room for maneuver for it to be meaningful to study the political dynamics of bureaucratic turnover. In cases of only modest policy conflict or limited salience, the parties in power can of course simply choose to wait until a given head’s term is up for renewal and hire someone else. In more pressing cases, however, they can also exploit their formidable control over the legislative process to impose their will through structure and process. In 1987, for example, the social democratic Carlsson cabinet granted itself the formal right to transfer agency heads to other (not necessarily permanent) positions, if deemed necessary for ‘organizational reasons’ or ‘the greater good of the agency’.Footnote 21 Since the implementation of the transfer reform, the Government Offices has typically held around ten former agency heads per year who have been allowed to keep their titles and pecuniary compensation, but with only minor, if any, formal responsibilities. Moreover, in Sweden, the position of agency head does not require internal advancement within the bureaucracy itself. Many heads are highly educated and have backgrounds in the political and private sectors, and are thus likely to have ample opportunities for alternative careers. Though neither political leaders nor agency heads have perfect control over their own employment situations, if our argument has merit, policy conflict should provide both sides with incentives to reconsider past employment decisions.
Data
We primarily used two sources when constructing the dataset: Sveriges statskalender (the Swedish State Calendar), which is a compendium of operative public organizations and employees published annually by the Swedish Government since 1812, and Vem är det?(Who’s Who), which is a Swedish reference publication with biographical information on most agency heads (similar to the British ‘Who’s Who’). In order to assure the reliability of the data, we have also further cross-checked and complemented these sources using a number of additional government publications from various years (Statsliggaren, Statliga myndigheter, Fakta om folkvalda and Enkammarriksdagen) and with CVs acquired through personal communications with agency staff.
To transform our sources into analyzable data, we have documented each individual head as a counting process that runs from the year of appointment until the year of exit, with one observation for each calendar year that the head remained in the post. The first year of service is coded as the appointment year, while the last year of service is coded as the exit year. For example, if a head entered service in 1990 and exited in 1993, he or she has four observations. In total, the sample includes 14,839 yearly interval observations on 2,355 unique appointments with 2,147 observed exits. The median tenure is six years, with the shortest appointment lasting for one year and the longest for thirty-four years. The dataset ends after 2014, leaving 9 per cent of the heads still in their positions.
In Figure 1, we show the total number of exits by observed years in a post, with the tenure times of all continuing heads included for reference. As can be seen from the event count, most of the action is located on the left-hand side of the distribution, with the vast majority of heads leaving their posts within twelve years. The patterns contrast notably with the US case, where almost one-fourth of program managers leave their posts by the end of their first year of service.Footnote 22 Swedish cabinets thus do not appear to suffer from the same sort of attrition problem that US presidents do. However, for a staffing process in which the baseline is supposed to be a six-year fixed term, we also find considerable variation. There is thus good reason to suspect that the Swedish civil service may not be quite as rigid as the official staffing policy implies.
Fig. 1 Frequency of exits by years on post
In Figure 2, we show the proportion of exits by calendar year, with the ideology of the government included for reference.Footnote 23 The first seventeen years of observation cover the second half of the Social Democratic Party’s ‘one-party state’.Footnote 24 During this period, 9 per cent of all agency heads left their post in each year, on average. The sharp increase in exits that followed the social democrats’ eventual defeat corresponds to almost 20 per cent of all heads in that year, and about three times the number observed in the preceding year. When the social democrats returned to power in the 1980s, the turnover rates did not revert to their post-war levels. Instead, the modern agency head appears to have turned into a rather more fragile position, with an all-time high of eighty-three exits, or 28 per cent of the population, in 2008.
Fig. 2 Proportion of exits by calendar year
Estimation
To get a closer look at the political dynamics of bureaucratic turnover, we model agency head tenures as a function of policy conflict between the appointing and sitting governments. To construct measures of policy conflict, we first match each agency head to attributes of the political parties that were in power at the time of their appointment. Then, for each subsequent year that a head remains in service, we examine how those baseline attributes compare to the attributes of the governing parties during that year. We then use a Cox regression to estimate the effects of policy conflict between governments on the turnover rates of agency heads.Footnote 25
We rely on two measures to account for changes in the government’s policy objectives. First, we construct a binary covariate, Cabinet Turnover, which denotes whether the cabinet responsible for appointing a given head has a different ideological affiliation than the current cabinet, with the appointer and incumbent being either a social democratic or liberal–conservative coalition, respectively. We assume (rather than demonstrate) that when political leaders appoint agency heads, they do so strategically to advance their own policy objectives. As long as this assumption holds, government turnover since the time of the appointment should increase the level of policy conflict between politics and the administration, because both sides will see the other as representing interests opposed to their own.Footnote 26 We expect that agency heads will be more likely to exit following partisan shifts in government.
Secondly, we use data from the Comparative Manifestos Project to construct a continuous measure of policy conflict, Policy Incongruence, to indicate the policy distance between the appointing and sitting governments.Footnote 27 If there is substantial over-time variance in policy preferences within the blocs, cabinet turnover may be a poor indicator of the actual level of policy conflict. The manifesto data offer a complementary perspective by allowing governments with the same ideological affiliation to have different policy priorities, but with the caveat that party manifestos may be more indicative of issue salience than of true ideology. The covariate is based on the policy positions of the prime minister’s party and captures the absolute distance between the left–right positions of the appointing and sitting cabinets. For readability, we have rescaled the original policy measure, which ranged from −100 to 100, into a measure ranging from −10 to 10. Analogously to the cabinet turnover covariate, we expect that agency heads will become more likely to exit as the policy incongruence between the appointing and sitting cabinet increases.
In addition to the indicators of policy conflict, we also include a number of time-constant and time-varying regression controls in our models. The range of covariates that could plausibly cause both ideological shifts in government and turnover among agency heads is limited. However, since unmeasured risk factors can bias hazard models even if they are uncorrelated with the observed covariates, we also consider variation that extends beyond potential confounders. To adjust for heterogeneity among agency heads, we include measures of whether they have a PhD (Academic Sector), whether they are associated with a political party (Political Sector), whether they have CEO experience from the private sector (Private Sector), whether they have CEO experience from the public sector (Public Sector) and their sex (Sex). To adjust for heterogeneity in strategic contexts, we include measures of economic growth (Economic Growth), whether it is an election year (Election Year), incumbent ideologyFootnote 28 (Liberal–Conservative), whether it is the first year of a new electoral term (New Term), and whether it is before or after the above-mentioned 1987 transfer reform (Pre-1987). To adjust for unobserved, time-constant factors at the agency level, we estimate models with agency fixed effects and random coefficients. Finally, because we have been unable to acquire relevant information on the background characteristics of some of our agency heads, we also use multiple imputation by chained equations to fill in missing values (ranging from 13 to 17 per cent of the observations). We briefly review the results of our modeling efforts below.Footnote 29
RESULTS
We present the results from six sets of Cox regressions. Models 1 to 3 focus on cabinet turnover, while Models 4 to 6 focus on policy incongruence. We begin by examining the impact of cabinet turnover along with the regression controls in Model 1 and then add the agency effects to the mix in Models 2 and 3. Models 4 to 6 follow the same procedure, except with policy incongruence in place of cabinet turnover. Scaled Schoenfeld residuals reveal no significant violations of the proportional hazards assumption using either linear, logarithmic, rank or Kaplan–Meier transformations of tenure times. Graphs of the Cox–Snell and efficient score residuals suggest that the models generally fit the data well with no influential outliers. As is often the case with hazard models, however, the fit is worse in the right-hand tail due to subjects dropping out over time. With this caveat, we report hazard ratios and standard errors clustered by agency for all models in Table 1. A hazard ratio above one implies that a one-unit increase in the relevant covariate is associated with an increase in the rate of exit; a ratio lower than one indicates that exit rates are decreasing when the relevant covariate is increasing.
Table 1 Partial Likelihood Estimates of Bureaucratic Hazards in Sweden, 1960–2014
Note: hazard ratios for listed covariates with standard errors clustered by agency in parenthesis (conditional on random-effect in Models 3 and 6). Efron method for tied events. **p<0.01, *p<0.05 (two-tailed)
Consistent with theoretical expectations, we find evidence that changes in a government’s policy priorities since the time of appointment can affect the length of tenures of agency heads. In all six models, the hazard ratios for cabinet turnover and policy incongruence show that agency heads leave their posts at higher rates when they are accountable to an ideological opponent of the appointing government than when they are accountable to an ideological ally of the appointing government. In all six models, moreover, the differences are statistically significant at conventional levels using a two-tailed test of significance. These results indicate that policy conflict is an important part of the explanation for why some agencies lose their heads while others do not.
The point estimates are also of substantial size. As can be seen in Model 1, where we include cabinet turnover along with the regression controls, party change in government is initially associated with a 54 per cent increase in the rate of exit. The agency effects change the results only slightly: Models 2 and 3 both produce point estimates well within one standard error of the estimate in Model 1. The random-coefficient model shifts the point estimate slightly upwards to a 58 per cent increase in Model 3, while the fixed-effects model produces the overall most conservative estimate, bringing the hazard ratio down to a 50 per cent increase in Model 2. As we show in the Appendix, the results are largely driven by the contextual covariates, while the background characteristics of the agency heads barely register on our focal relationship. If we drop all controls and agency effects, cabinet turnover is associated with a 73 per cent increase in the rate of exit. Adding only the background characteristics reduces the point estimate to a 69 per cent increase, while adding only the contextual covariates reduces the point estimate to a 54 per cent increase – the same result as in Model 1 (and with minimal change in standard errors).
The models focusing on policy incongruence follow the same patterns as those related to cabinet turnover. The policy scale ranges from −10 to 10. In our sample, the maximum observed distance is 8.5 units, with a standard deviation of 1.8 units. In Model 4, which includes policy incongruence along with the regression controls, the marginal effect of shifting the policy positions of either the appointing or the sitting cabinet away from the other is 4 per cent per unit. Adding the agency effects, as we do in Models 5 and 6, makes essentially no difference for the results (though the point estimates are rounded upwards in the models with agency effects). The marginal effect of shifting the policy positions of either the appointing or the sitting cabinet two standard deviations away from the other, for example, can then be read as a 14.4 per cent increase in the rate of exit. As with cabinet turnover, adjusting for the background characteristics of the agency heads barely registers on the focal relationship, while the contextual covariates have a more substantial impact. If we drop all controls and agency effects, policy incongruence is associated with an 11 per cent increase in the rate of exit per unit. Adding only the background characteristics rounds up to the same point estimate as the empty model, while adding only the contextual covariates reduces the point estimate to a 4 per cent increase per unit.
Though the statistical models are useful for describing general trends, Swedish political history is also rife with substantive examples of political disagreements over agency staffing. The social democratic Persson cabinet (1996–2006), for example, made several publically disputed appointments during its time office, which also led to a number of equally disputed dismissals by the liberal-conservative Reinfeldt cabinet (2006–14). Persson was repeatedly called by the parliamentary Committee on the Constitution to justify how, given the constitutional demands for meritocratic recruitment, the most prestigious administrative posts nonetheless tended to end up in the hands of people with strong connections to the Social Democratic Party. One of the most prominent cases concerned Bo Bylund, appointed by Persson as head of the Swedish Public Employment Service in 2005 and eventually dismissed by Reinfeldt in 2008. Employment policy was among the most salient issues for the Reinfeldt cabinet, and Bylund had a background in the labor union and as state secretary under a previous social democratic government. The liberal-conservative parties had already signaled their disapproval of Bylund’s appointment in the run-up to the election, with Reinfeldt openly referring to Bylund as a social democratic campaign aide. Following the social democratic defeat in the 2006 general election, Bylund’s contract was abruptly terminated three years early; the dismissal was justified as a natural part of organizational reform and nothing out of the ordinary.
Another example of how a change in government policy can affect bureaucratic turnover concerns the former head of the Swedish Migration Agency, Anders Danielsson. The Reinfeldt cabinet entered office with a permissive stance on migration and Danielsson, appointed in 2012, followed suit by taking a public stand for both the moral and instrumental merits of the free movement of people, particularly asylum seekers. When the social democrats returned to power with the aid of the Green Party in late 2014 (the Löfven cabinet, 2014–), they promoted a similarly permissive migration policy, giving little reason to suspect a conflict of interest between the new government and Danielsson. In autumn 2015, however, the situation changed drastically. Pressured by the largest per capita inflow of migrants in Europe, the Löfven cabinet announced a complete reversal of its initial stance on migration and declared that Sweden would henceforth only accept the absolute minimum numbers of immigrants required by European Union law. On 4 January 2016, Sweden introduced border controls, and immigration dropped dramatically. The new migration policy stood in sharp contrast to the policy that Danielsson had vigorously defended throughout his tenure, and on 16 September 2016 he announced his resignation and subsequent reassignment to the Swedish Red Cross. Our models suggest that cases like those involving Bylund and Danielsson may not be such rare events, but rather a general feature of partisan politics.
DISCUSSION
Research on the political dynamics of bureaucratic turnover in European parliamentary democracies should not end here. In this Note, we have focused on the average effects of policy conflict on the exit rates of agency heads. Yet such estimates may hide considerable variation, primarily for two reasons. First, our exit category does not discriminate between different types of departures. As a consequence, we implicitly treat, for example, dismissals, retirements and career changes as equivalent. Secondly, our measures of policy conflict are crude and do not consider the possibility that the parties in power may well agree on some issues yet disagree vehemently on others. In the Swedish case, for example, the social democrats sometimes appoint conservatives as agency heads, and vice versa. If our argument has merit, this should be unremarkable – but only as long as those appointments are made in exactly the agencies that are working on issues on which there is little or no policy conflict between the parties. More detailed information about both the situations surrounding each departure and the policy preferences of politicians and bureaucrats would allow us to estimate more nuanced models.
Furthermore, our research design precludes a proper assessment of the effects of different institutional arrangements. Though we provide evidence that standard ‘civil service’ restrictions such as tenure protection and meritocratic recruitment are insufficient, on their own, to keep partisan politics out of the staffing process, different combinations of political and administrative institutions may make it more or less feasible for political leaders to influence the composition of agency personnel. As Huber points out, the extent to which institutions create a ‘multiple principal’ problem should be of particular importance, as incentives to politicize the bureaucracy are directly affected by political competition.Footnote 30 With its emphasis on one principal for each agent, Sweden’s relatively straightforward parliamentary chain of delegation should be less affected by this problem than, for example, Germany or France.Footnote 31 At the same time, however, Sweden also has weaker restrictions on the partisan affiliations of top-level civil servants than countries such as Denmark or the UK, which may give Swedish political leaders a comparative advantage regarding opportunities for administrative politicization.Footnote 32 There are good theoretical reasons to believe that different institutions can interact and produce system-specific effects, but efforts to assess such claims empirically remain sparse.Footnote 33 An important next step for future studies is to pursue research designs capable of explicitly addressing how different political and administrative institutions can affect both screening and selection strategies.
We have also side-stepped two other factors commonly highlighted in the literature: expertise and patronage. As many have emphasized over the past few decades, the price of political control is the loss of bureaucratic expertise.Footnote 34 In some cases, political leaders may find this an acceptable trade-off, but in other cases they may also value bureaucratic inputs higher than their own. Since a government’s valuation of control and expertise can in principle vary both across issues and over time, properly accounting for this trade-off is difficult, particularly with observational data. At the other end of the spectrum, however, many have also noted how political leaders sometimes forego expertise entirely in order to use agency positions as rewards for loyal supporters.Footnote 35 Though we have emphasized the impact of policy conflict, the fact that turnover among bureaucrats tends to follow turnover among politicians is explained equally well by a staffing process swamped with patronage. Incorporating concerns about policy, expertise and patronage into a single empirical model would require measures capable of distinguishing between agencies that are electorally salient, technically demanding and those that are merely acting as turkey farms for the parties in power.
Finally, while we believe that much work remains to be done on the political dynamics of bureaucratic turnover in Europe, our results are also relevant for students of American bureaucracy. Congress frequently creates agencies with fixed terms and other restrictions in order to shield executive bureaucrats from presidential opportunism. And ever since Woodrow Wilson’s seminal call for a ‘science of public administration’, meritocratic civil service institutions have been seen as key in protecting bureaucratic expertise from political control.Footnote 36 However, if such constraints can be at least partly circumvented – as appears to be the case in Sweden – then the actual level of political influence over the composition of American public personnel may have been underestimated. Further investigations of the relationship between partisan politics and the career trajectories of civil servants (as opposed to political appointees) would contribute to a better understanding of the development and functions of the administrative state.Footnote 37
