Democracy, Authoritarianism, and Ministerial Selection in Russia: How Presidential Preferences Shape Technocratic Cabinets

Abstract

Abstract: In Russia, politicians routinely select technical, non-party cabinets whose political outlook is difficult to decipher. The article asks how variation in politicians' preferences along the democratic-authoritarian dimension shapes political skills sought in cabinets and how ministers sharing politicians' aims are selected. The account of ministerial selection is tested in a controlled, comparative analysis of ministerial appointments in Russia, using an original dataset covering all full ministers appointed by presidents Boris Yel'tsin and Vladimir Putin, 1992–2008. The results clarify several important debates about the political implications of changes to cabinet composition in Russia and contribute to a more nuanced general understanding of the politics of technical cabinets.

When a new president picks his cabinet, he gives observers … [a] first set of solid clues about the kind of president he intends to be (Polsby, 1978, p. 15).

Russia is part of a wider set of political systems in which politicians routinely appoint overwhelmingly non-party cabinets—that is, cabinets without formal party participation (Amorin Neto, 2006; Amorim Neto and Strøm, 2006; Schleiter and Morgan-Jones, 2009, 2010; Samuels and Shugart, 2010). Often the political outlook of such cabinets is difficult to decipher, precisely because they are not based on the participation of parties with clearly defined policy programs. As a result, the widely shared conviction that a president's cabinet choices hold important clues about the kind of president he or she intends to be (P olsby, 1978, p. 15) is matched by considerable debate about the politics of such cabinets among both comparative scholars and Russia experts.

Thus, when President Putin secured Prime Minister Fradkov's resignation, and appointed a new cabinet under Viktor Zubkov in September 2007, speculation was rife. What clues did this cabinet change give about the kind of political future he was planning for Russia? This was an important question at a time when Putin was nearing the end of his second term and could not stand for re-election. Yet, the government change raised more questions than political analysts could answer: Duma elections were due to be held in December, presidential elections in May the following year. Was the government changed before these elections—rather than in response to them—to consolidate Putin's power ahead of the political uncertainties the elections might generate while keeping potential successors off balance (Goldman, 2007; Tretyakov, 2007)? Zubkov's appointment could be interpreted in this way—it was completely unanticipated and Zubkov had formerly worked in financial intelligence and was one of the petertsy, a group of officials who had a shared career history with Putin in St. Petersburg (Tretyakov, 2007). Alternatively, was the promotion of Zubkov a signal that Putin wanted to put politics on a cleaner and more professional footing with a government headed by a prime minister who was independent from the oligarchs and had no involvement in past corruption (Butrin, Netreba, Vsloguzov and Sharipova, 2007)? Again, this interpretation was not implausible: Zubkov, unlike other ministers, was not personally wealthy and lacked connections to Russia's financial-industrial groups.

These divergent analyses proceeded from the shared understanding that the choice of the prime minister (and indeed other government ministers) was shaped by the strategic direction Putin was aiming to pursue. Yet the widely discrepant assessments of the September 2007 government change are also indicative of the fact that ministerial selection often remains poorly understood when governments are nonpartisan, even among highly informed observers.

As the next section makes clear, the impact of different governing strategies on the types of non-party ministers that politicians appoint remains deeply disputed among Russia specialists and comparativists alike. This article aims to advance the debate by developing an approach to ministerial selection in Russia that is embedded in a general theory of cabinet composition by democratic and authoritarian politicians. My central argument is that variation of a politician's preferences along the democratic-authoritarian dimension affects what political skills they seek in their ministers and how they ensure that ministers have incentives to serve their principal. I test this account in a controlled, comparative analysis of 314 ministerial appointments in Russia from 1992–2008, which demonstrates that politicians choose very different types of nonpartisans depending on the manner in which they plan to govern. As the preferences of Russia's presidents became progressively less democratic, they were less likely to choose ministers whose aims and incentives were shaped by party affiliation and who had political expertise to negotiate with the assembly. Instead, authoritarian preferences raised the odds of loyalty appointments, and the probability that presidents would recruit political expertise to manage coercive agencies and economic client groups. These results help to clarify a range of debates about the changing political nature of Russia's cabinets and contribute to the comparative literature a better and more nuanced understanding of cabinets that include a significant share of non-party ministers.

Extant Scholarship

The discrepant interpretations of Zubkov's appointment in 2007 reflect a wider debate within the rapidly growing literature that examines elites and cabinet appointments in Russia. Much of this work has noted the increased reliance of President Putin (2000–2008) on personal loyalty in his ministers, his promotion of colleagues from St. Petersburg (the petertsy), and the KGB, military and other coercive agencies (siloviki) into positions of political power, as well as the growth of business-government links (Petrov and Slider, 2003). A significant group of experts interpret these changes in government and elite composition as signs of the increasingly authoritarian nature of the regime under Putin (Illarionov, 2009; Huskey, 2010), its sovietization (Kryshtanovskaya and White, 2009), militarization, and the oligarchic “interpenetration of business and government” (Kryshtanovskaya and White, 2003, 2005, p. 295). Yet, this view is not universally shared. There remains a question to what extent Putin's ministers represent a qualitative departure from cabinet selection under Yel'tsin. For a number of analysts, the technocratic nature of Yel'tsin's cabinets already evidenced a lack of commitment to democratic norms (Colton and Skach, 2005, p. 121). Moreover, scholars have pointed to alternative interpretations of the trends observed under Putin. Rivera and Rivera, for instance, argue that the increasing presence of business elites in government represents an embourgeoisement of the state that is likely to restrain the discretion and authoritarian inclinations of Russia's political elites in the longer run (Rivera and Rivera, 2006). Finally, experts have drawn attention to the heterogeneous political orientations of ministers who share military and security backgrounds. Thus, Renz proposes that the influx into the elite of siloviki—individuals with professional backgrounds in the armed services, law enforcement bodies, and intelligence agencies—cannot be said to favor a more authoritarian policy direction because the siloviki differ in terms of institutional attachment, former rank, role, and political outlook, and do not form a coherent political group (Renz, 2006).

This work has contributed much to the understanding of Russia's cabinets, but a critical omission has limited the ability of scholars to resolve the debates outlined above. In theoretical terms, this literature has lacked an approach to ministerial selection that is embedded in a more general theory of cabinet composition by democratic and non-democratic politicians (Gel'man and Tarusina, 2000, p. 320). In part, this reflects the fact that the impact of these governing strategies on the choice between non-party ministers remains poorly understood, not just among country experts but also within political science as a discipline.

What factors shape the attributes that politicians regard as desirable in non-partisan ministers is an, as yet, under-researched question in comparative politics. Moreover, among comparativists as is also the case among experts of Russian politics, scholarly opinion regarding the political outlook represented by non-party cabinets is deeply divided. Some comparative work regards highly non-partisan governments as indicative of governing strategies that lean heavily on executive prerogative (Amorim Neto, 2006), erode the role of representative assemblies, and signal a non-democratic outlook (Centeno and Maxfield, 1992). Other political scientists, however, stress that the resort to non-party ministers is often conditioned by a combination of under-institutionalized party systems and complex policy challenges and does not necessarily reflect a non-democratic outlook. Indeed, at times, technocratized cabinets may even help to legitimize democracy by making possible the management of complex technical challenges (Bermeo, 2003, pp. 216, 222).2

Clearly, neither of these possibilities can be ruled out ex ante, but the debates outlined above cannot advance without a theory of how variation in the governing strategies of politicians along the democratic-authoritarian spectrum affects the cabinets they form. What scholars require, then, is a better understanding of how different political strategies and contextual conditions shape ministerial selection when presidents name non-partisans. The next section develops such an approach by synthesizing scholarship on the differences between authoritarian and democratic regimes and their cabinets within a principal-agent framework.

Non-Party Ministers and Governing Strategy

Non-party ministers can be defined as cabinet members who do not act on behalf of a party in government. They may have known party preferences and may be members of a party, but serve in cabinet in a personal capacity, and their appointment does not constitute party participation in government. Politicians typically rely heavily on non-party ministers because a range of factors reduce the value of negotiating party-based cabinets. Such factors include under-institutionalized party systems (often the result of recent transitions from authoritarian rule), complex technical policy challenges, or the existence of a president who is institutionally powerful enough to rule without much reference to assembly parties (Bermeo, 2003; Amorim Neto, 2006), and, as a result, ministerial selection is not mediated by parties. How, then, do politicians choose the skills and attributes of non-partisan ministers, and what role does a politician's governing strategy play in shaping these choices? A principal-agent approach offers an analytical framework to address this question.

Principal-agent theory conceives of popularly elected politicians as political principals who appoint ministers as their agents “in order to devise and oversee the implementation of policies” (Dowding and Dumont, 2009, p. 5). Principals thus delegate substantial authority over policy to their ministers (Laver and Shepsle, 1994), and whether this delegation of authority is likely to yield the outcomes a president desires depends to a significant extent on the appointment of ministers with the appropriate incentives and skills. Presidents can therefore be expected to select agents who share their aims and incentives, and who are competent to advance these aims in government.

Two political attributes enhance the desirability of a minister for politicians. First, incentive compatibility between the minister and the president is highly desirable because ministers whose political incentives are well aligned with those of their principal can be expected to share the principal's political aims and to be dependable once in charge of a portfolio (Indridason and Kam, 2008, p. 637). Second, because ministers play a key role in devising and overseeing the implementation of policy, presidents can be expected to name individuals with political competence, that is skills to broker compromises with key actors in the policy process (such as other legislators and parties) (Huber and Martinez-Gallardo, 2008, p. 170).

The central argument developed in the following sections is that a politician's preferences to govern in a more or less democratic manner can be expected to shape not just how they select incentive compatible ministers, but also what type of political competence they seek in their cabinet. Those who plan to govern democratically are likely to choose ministers who differ significantly from the cabinets preferred by politicians whose outlook is essentially non-democratic.

Incentive Compatibility

In selecting ministers who have incentives that are compatible with their own, presidents can rely on two different strategies. Those who see elections and parties as the main drivers of a politician's incentives are likely to select ministers—even those who serve in cabinet in a personal capacity—from among the pool of political talent controlled by parties that share the president's policy aims. Alternatively presidents can rely on personal loyalty and make appointments on the basis of a shared career history or patronage relationships. In practice, all presidents are likely to employ a mixture of these strategies (Polsby, 1978), but how they plan to govern can be expected to shape where they place the emphasis.

When politicians plan to govern democratically, they anticipate working in a context in which governmental policy derives its legitimacy from voter choices, and in which elections and political parties exert a powerful influence in shaping ministers' incentives to pursue particular policies, even in technical cabinets. As a result party political backgrounds are a critical attribute that politicians can use to identify ministers who share their political aims. Thus, political parties are an important pool from which ministers are recruited under all democracies (Bermeo, 2003).

Elections and political parties are often less relevant to ministerial selection when politicians plan to govern in a more authoritarian manner. Only in single-party authoritarian regimes do governing parties typically play a central role in policy choice and the recruitment of political elites (Geddes, 2003). But, under other forms of authoritarian rule (personalist, military, or monarchical), policy and access to political power are controlled by a single individual or the military as an institution (Geddes, 2003; Hadenius and Teorell, 2007). Because ruling parties in such contexts do not truly control or legitimate policy choices, party membership is not an attribute that politicians can use to gauge the policy preferences and incentives of ministers (Ezrow and Franz, 2011, pp. 75–76). For this reason, rulers who seek to recruit incentive compatible cabinets to govern in these contexts typically rely on personal loyalty based on shared career trajectories and patronage relationships in order to ensure that ministers share their aims.

Political Competence

The second political attribute politicians can be expected to seek in their cabinets is political competence—that is skills that enable ministers to broker policy compromises with key political actors. Again, how presidents plan to govern can be expected to influence how they choose these skills. A democratic cabinet translates the choices of the electorate into a specific package of policies to advance and implement, and in all democracies legislatures are critical actors in deciding the success and adoption of these policies (Cotta, 1991, pp. 176–177). Hence, politicians must build, cultivate, and shape legislative majorities, and an effective way to do so is to choose ministers with legislative experience (Dogan, 1979, p. 10). Presidents who plan to govern democratically can therefore be expected to recruit ministers who have the legislative experience to advance the cabinet's legislative agenda effectively (Amorim Neto, 2006). Indeed, with the transitions from authoritarianism to democracy in Portugal, Spain, Italy, and Greece the prominence of the legislature as a recruitment pool for ministers rose markedly—between 51 percent and 80 percent of the ministers who were recruited after the transitions had parliamentary experience (Bermeo, 2003, p. 209).

Politicians who do not aim to govern democratically require a different type of political competence in their cabinets. In authoritarian regimes, policy is made by and “for the elite and the privileged” (Acemoglu and Robinson, 2006, p. 18), it is not formed on the basis of democratic contestation, representation, and accountability. As a result, authoritarian regimes de-emphasize the role of representative assemblies (Bermeo, 2003). Instead, these regimes rest on the support of key client groups, typically economic elites, the military, and security services. Because these groups have the strongest capacity to depose the authoritarian regime, they become central political actors who must be managed and co-opted in the policy process (Wintrobe, 2007, p. 390; Ezrow and Franz, 2011, p. 37). Consequently, politicians who envisage a more authoritarian style of rule can be expected to name cabinets that have the political skills and experience to manage coercive agencies and economic client groups, rather than legislative parties. In authoritarian Portugal, Spain, and Greece, for instance, individuals with military backgrounds accounted for between a quarter and a third of all cabinet positions (Bermeo, 2003, p. 208).

Confounding Influences

Of course, ministerial choices are not determined by the democratic or authoritarian preferences of politicians alone and the empirical analysis that follows controls for a range of confounding influences which derive from the political and economic context that confronts politicians. The political context can be expected to shape ministerial choices in two main respects. First, the magnitude of presidential support in the legislature impacts on the size of the talent pool from which politicians can select ministers with party background and legislative expertise (Amorim Neto, 2006). Presidents who have greater legislative support can more readily recruit ministers with the appropriate party political and legislative experience into their cabinets. Second, legislative fragmentation can make cooperation with the assembly more difficult to achieve (Schleiter and Morgan-Jones, 2009). As a result presidents often seek to rely more extensively on executive prerogative rather than statutes, which reduces the value of recruiting ministers with political party and legislative experience into the cabinet.

In addition, the policy context can be expected to impact on ministerial choice. Technically complex policy challenges—such as the Latin American debt crises, European integration, or Russia's transitional recession during the 1990s—are known to raise the need for technical expertise within the cabinet. Confronted with complex challenges, technically competent ministers are likely to perform better than generalists in judging which of the policy choices open to the government are likely to prove successful (Bermeo, 2003, p. 216; Bäck et al., 2009). Because technical policy expertise is typically the result of career paths that are quite separate from a party political or legislative career, complex policy challenges can be a factor that attenuates party political and legislative expertise in the cabinet (Bermeo, 2003, p. 216; Amorim Neto and Strøm, 2006). For instance, Centeno and Maxfield observe that in Mexico, and across Latin America's Southern Cone, the economic crises and debt crisis after 1982 caused an influx of techno-bureaucratic ministers into cabinets that resulted in the breaking of cabinet links with traditional party constituencies (Centeno and Maxfield, 1992, p. 78).

The extant comparative literature, then, suggests that in evaluating the relationship between governing strategies and ministerial selection, it is important to take account of alternative, contextual, influences on the cabinet. However, much of the empirical work on elites and government membership in Russia has so far relied on descriptive statistics, and does not address questions of multiple causation and confounding influences (Gel'man and Tarusina, 2000, p. 320). The empirical approach employed below addresses this lacuna.

In sum, the approach developed in this section applies a principal-agent perspective which makes clear that the variation of governing strategies along the democratic-authoritarian dimension can be expected to have a pronounced effect on the ministerial choices of presidents. The analysis below probes that expectation.

Data and Dependent Variables

To examine the impact of a president's governing strategies on the attributes and political competence of non-party ministers, the article employs an original dataset covering all full ministers appointed by presidents Boris Yel'tsin and Vladimir Putin from independence (1992) to the end of Putin's presidency in 2008. This is a departure from the approach of previous work which compares snapshots of governmental elites during selected years. The problem in comparing snapshots is that specific cross-sections of ministers may not be comparable or representative. Using a complete dataset of all ministerial appointments overcomes this problem by making possible an analysis of the full universe of cabinet appointments.

During the 16 years covered by this study, Russia's presidents appointed entirely non-partisan cabinets—even if ministers had a party political background, they acted in a personal capacity rather than as representatives of their party in government. Government membership is recorded at the time of the government's appointment. A new government is counted with every constitutionally required government change (after presidential elections), change of prime minister, and turnover of at least 50 percent of the cabinet members (to capture major cabinet changes comparable to changes in the party composition of partisan governments). In all, the data comprise information on the career histories of 314 ministers, 229 appointed by Yel'tsin, 85 by Putin.

The dependent variable in each of the analyses that follow is an attribute or competence of a minister, recorded on the basis of a minister's career background, including loyalty links to the president, political party experience, legislative experience, a career background in the military or security structures, and a career background in business. These are indicator variables, recorded on the basis of the career history of a minister during the five years that precede each ministerial appointment or re-appointment. The exception to this rule is loyalty links with the president. Because such links might have been formed at any career stage, they were recorded without time limit (governments, data sources, and coding rules for all dependent, explanatory, and control variables are detailed in the Appendix).

Explanatory Variables

To gauge the changing preferences of these two presidents regarding their governing strategy, I rely on a quantitative content analysis of the Presidential Addresses to the Federal Assembly. In this respect, too, the article departs from previous work. By developing an explicit measure of presidential preferences it turns the question of what impact governing strategies have on ministerial choice into a testable empirical question.

Presidential Addresses to the Federal Assembly are high-profile speeches that Russian presidents use to communicate their main priorities and strategic aims for the coming year not just to legislators, but also to the wider political elite, key stakeholders within the regime, and the public. To assess the degree to which the presidents were prepared to commit publicly to a democratic governing strategy, I recorded the frequency and nature of references to keywords with roots that refer to democracy itself, and four of its constituent features—freedom, rights, representation, and accountability. Each reference to democracy or its constituent features was coded as positive or qualified/negative, depending on whether the content of the statement expressed approval or criticism. In total, this coding method captured 862 references to democracy and its constituent features across the presidential addresses from 1994 to 2004, which were overwhelmingly positive (86 percent). Negative and qualified references were subtracted from positive ones to create an overall measure of the president's public commitment to democracy in a given speech and year. This approach gives a fine-grained measure of the president's intentions for the year ahead that captures significant variation from year to year, across presidential terms, and presidents. Note that annual Presidential Addresses to the Federal Assembly were delivered from 1994 onward. Values for 1992 and 1993 were extrapolated based on the average for Yel'tsin's first term.

Table 1 summarizes the measure and shows that Yel'tsin, in particular during his first term, was keen to commit publicly to democracy (positive references to democracy per speech averaged 137.8); this enthusiasm for democracy declined somewhat during his second term in which he made on average 62.3 positive references to democracy. Putin's readiness to commit publicly to democracy, rights, freedoms, representation, and accountability was notably weaker, averaging just 22 references in his first-term speeches, and further declined to an average of just 12.5 references per speech in his second term.

Table 1 Presidential Commitment to Democracy.^a

	Year	Public Commitment to Democracy	Average by Presidential Term	Freedom House Score	Freedom House Classification
Yel'tsin 1st term	1992	138		3.5	partially free
	1993	138		3.5	partially free
	1994	144		3.5	partially free
	1995	165		3.5	partially free
	1996	104	137.8	3.5	partially free
Yel'tsin 2nd term	1997	49		3.5	partially free
	1998	43		4	partially free
	1999	95	62.3	4.5	partially free
Putin 1st term	2000	39		5	partially free
	2001	21		5	partially free
	2002	14		5	partially free
	2003	14	22	5	partially free
Putin 2nd term	2004	6		5.5	not free
	2005	33		5.5	not free
	2006	4		5.5	not free
	2007	7	12.5	5.5	not free
^a Source: Author's database. See the Appendix for detailed information about sources.

To evaluate the validity of this measure, Table 1 also displays the Freedom House scores and categorizations of the regime, which are recorded retrospectively for each year (the presidential addresses are given early in the year, typically between February and May). Note that smaller values indicate more democratic environments on the Freedom House scale. As the table makes clear, the two measures correlate closely. A president's preferences regarding his governing strategy as conveyed by the annual Address to the Federal Assembly—clearly correlate with the state of democracy in Russia as subsequently recorded for each year by Freedom House. The declining readiness of Russia's presidents to make public commitments to democracy foreshadowed the declining quality of democracy from year to year, and eventually the evaluation of the regime as “not free” from 2004 onward.

Since governing strategy is the key explanatory variable of interest, I probe the robustness of the empirical results by running all analyses using both, the content-based measure of democratic governing strategy as well as the alternative Freedom House measure. To the extent that both measures yield substantively the same result, we can have greater confidence in the findings.

Control Variables

The analysis controls for the confounding influence of contextual factors. To take account of the changing economic policy context that confronted Russia's presidents the years in which real GDP per capita fell are recorded as years of economic crisis (crisis = 1, otherwise = 0). Russia's transitional crisis went hand in hand with the major policy challenge of guiding the country through the transition from a planned economy to the market. The crisis was severe and prolonged, and subsided only in 1999, when the economy began to grow again. As noted above, the economic crisis likely raised the need for technical expertise within the government, which could have depressed the representation of ministers with party political experience or legislative backgrounds. Moreover, as the transition progressed, the structure of the economy changed, and the proportion of GDP accounted for by the private sector rose until Putin began to re-establish state control over strategic assets. Because the increasingly privatized structure of the economy can be expected to affect the representation of ministers with business links in government, this factor is controlled for in the analysis of the business expertise recruited into the cabinet.

The political context, too, changed significantly throughout this period. After the collapse of the USSR, Russia initially featured a nascent party system, organized around poorly disciplined blocs and factions rather than institutionalized political parties (Remington, 2001, pp. 144–145). While some consolidation of the party system occurred after the adoption of a new constitution and Duma elections in 1993, the number of legislative parties and the share of independent deputies within the legislature remained high throughout the 1990s (Remington, 2001). In the 2000s, the party system began to consolidate following “a series of reforms that weakened the most important pre-existing parties and corralled a majority of the most influential independent politicians into a new pro-Putin organisation called United Russia” (Hale, 2010, p. 81). These changes to the party system not only affected legislative fragmentation but also the level of presidential support in the legislature. Yel'tsin faced a hostile assembly for most of his time in office, except for a short period after the 1993 Duma elections, when Russia's Choice, a pro-presidential party, was the largest legislative force (Remington, 2001, p 178). Putin, by contrast, built a pro-presidential legislative coalition early in his first term, and the 2003 Duma elections handed the pro-presidential party United Russia 49.3 percent of the legislative seats and effective dominance of the assembly. These changes in the political context are captured by two variables, legislative fragmentation, measured by the Laakso-Taagepera index, and an indicator variable that takes the value 1 when the party supporting the president accounts for the largest Duma grouping. Legislative fragmentation can be expected to make cooperation with the assembly harder to achieve, which might reduce the willingness of a president to recruit ministers with party political and legislative expertise into the government. In contrast, the level of presidential legislative support can be expected to correlate positively with the recruitment of ministers who have both types of expertise because it raises the probability that the president will be able to cooperate with the assembly and broadens the talent pool from which presidents can make their choices.

Since the explanatory and control variables have different scales, they are standardized to have a mean of 0 and a standard deviation of 1, so that the magnitude of their effects can be directly compared in the analysis that follows.

Multivariate Logit Analysis

To examine to what extent the expectations outlined above receive empirical support, I adopt a multivariate regression approach that models the effect of changes in governing strategy on the qualities that Russia's presidents sought in their ministers, controlling for contextual factors. The choice of statistical model is driven by two considerations. First, given that ministerial skills and characteristics are recorded as indicator variables, I employ a logistic regression model. Second, the data structure is hierarchical in the sense that ministers are clustered in cabinets, which are in turn grouped by concurrent presidential/legislative terms and under different presidents. Factors operating at each of these levels can be expected to give rise to variation in ministerial selection. To account for this data structure, I explicitly model those factors that can be expected to give rise to variation in ministerial selection at the level of the president (president's governing strategy), and at the level of the presidential and legislative term (changing levels of legislative fragmentation and presidential legislative support). To further account for potential heterogeneity of ministerial choices at the cabinet level, I estimate robust standard errors clustered on cabinet since model specification tests show that the implementation of cabinet level random effects is not warranted by these data.3

Table 2 examines how far governing strategy affects the attributes through which presidents seek to ensure that ministers share their aims and objectives.4 As anticipated, a more democratic governing strategy raises the probability that presidents recruit ministers associated with political parties that support them (models 1 and 2). While the coefficient for democratic strategy in model 1 (based on the content analysis measure) falls short of statistical significance, it has the anticipated positive sign and the coefficient in model 2 (based on the alternative Freedom House measure), has a large and statistically significant positive effect on the probability the president will select ministers from a party that supports him, as expected. Further, as is consistent with the expectations outlined above, a more democratic governing strategy powerfully reduces the probability that presidents will select ministers on the basis of personal loyalty (models 3 and 4). Irrespective of whether the content-analysis measure is used or the Freedom House alternative, models 3 and 4 show that presidents are significantly less likely to appoint on the basis of loyalty, when they plan to govern in a more democratic manner. The effect is substantively large. In model 3, a one unit increase in democratic commitment (equivalent to one standard deviation in this case because the variable is normalized), is associated with a 71 percent reduction in the odds that a minister with loyalty links to the president is appointed.5 The models control for the influence of potentially confounding contextual factors. Those include the economic crisis, presidential legislative support, and legislative fragmentation. The economic crisis might have depressed the representation of ministers with party political backgrounds in the cabinet, but in fact had no consistent effect on the party political affiliation of ministers (models 1 and 2). As anticipated, presidential legislative support has a large positive effect on the recruitment of ministers from parties supportive of the president. The coefficient for assembly fragmentation has the anticipated negative sign, but falls short of statistical significance in models 1 and 2. Note, that none of these controls have a consistent effect on the recruitment of ministers with personal loyalty links to the president. Put differently, as might be anticipated, the only robust and statistically significant predictor of recruitment decisions that prioritize loyalty is the president's governing strategy.

Table 2 Recruiting Incentive-Compatible Ministers.^a

	(1)	(2)	(3)	(4)
	Pro-Presidential Party	Pro-Presidential Party	Loyalty	Loyalty
Democratic Strategy	0.431		− 1.237
	(0.243)		(0.592)*
Democracy (FH)		0.870		− 0.893
		(0.238)**		(0.327)**
Economic Crisis	0.363	0.159	− 0.378	− 0.347
	(0.185)*	(0.153)	(0.480)	(0.666)
Presidential Party Largest	0.732	1.208	0.710	0.224
	(0.172)**	(0.200)**	(0.204)**	(0.151)
Assembly Fragmentation	− 0.632	− 0.728	1.693	0.947
	(0.366)	(0.294)*	(0.884)	(0.433)*
Constant	− 2.257	− 2.323	− 2.233	− 2.119
	(0.202)**	(0.154)**	(0.497)**	(0.355)**
Observations	314	314	314	314
Robust standard errors in parentheses; * significant at 5 percent, ** significant at 1 percent.
^a Source: Author's database. See the Appendix for detailed information about sources.

These findings lend support to the conclusions of experts who have argued that Yel'tsin and Putin's flagging commitment to a democratic governing strategy was a critical cause of the diminishing reliance on party affiliation and the increasing use of personal loyalty by both presidents in ensuring that ministers shared their political aims. This trend became particularly evident once Putin's reforms of the party system secured United Russia a clear majority in the Duma following the 2003 elections, which a more democratically oriented president might have used to form a single-party majority government. “[Y]et the government that Putin appointed was no more based on United Russia than the previous one had been. The government, in fact, contained only one member of the party; in addition, as before, it contained some leading figures from United Russia (such as Boris Gryzlov and Sergey Shoygu), who were not formally party members. These, though, were in the minority and continued to owe their positions to Putin's patronage rather than their standing in the majority party. The new prime minister, Mikhail Fradkov, moreover, was a career civil servant, not a United Russia functionary” (Wilson, 2006, p. 337). This choice made clear that Putin had no interest in the representative functions of the pro-presidential party that he had constructed, or indeed of Russia's newly reformed party system more widely. As Russia became an increasingly personalist and authoritarian regime, the government remained divorced from voter choices and the ruling party had little influence on policy, was not used to train and screen elites, and lacked “the integrating functions for elites that parties in both democratic and one-party states can carry out” (Huskey, 2010, p. 367). Instead of party affiliation, personal loyalty thus became the dominant mechanism though which incentive compatibility between the president and his ministers was secured.

Tables 3 and 4 examine how governing strategy affects what political expertise presidents seek in their cabinets. As Table 3 shows, a more democratic governing strategy does, as anticipated, powerfully raise the probability that a president will recruit ministers with legislative experience. Thus, in model 1, a one standard deviation increase in democratic commitment is associated with a 90 percent increase in the odds that a minister with parliamentary experience will be appointed (significant at the 1 percent level). Conversely, as a president's willingness to commit publicly to democratic procedures declines, so too does his readiness to name ministers who have the political competence to negotiate legislative support for government proposals as part of the democratic policy process. The effect is robust, whether the content analysis (model 1) or Freedom House measures (model 2) of democratic strategy are used. The analysis takes account of possible alternative influences—the strength of presidential legislative support, assembly fragmentation, and economic crisis. As expected, the probability that ministers with legislative experience will be recruited rises when the party that supports the president accounts for the largest group in the legislature. Controlling for a president's legislative support, though, assembly fragmentation has no consistent effect. Surprisingly, perhaps, conditions of economic crisis correlate positively with the appointment of ministers who have legislative experience. But this result mainly flows from the fact that the crisis coincided with the recruitment of ministers with legislative backgrounds when Yel'tsin's commitment to democratic governance was strongest in his first term.

Table 3 Recruitment of Ministers with Political Competence to Work with the Assembly.^a

	(1)	(2)
	Legislative Experience	Legislative Experience
Democratic Strategy	0.640
	(0.164)**
Democracy (FH)		0.768
		(0.102)**
Economic Crisis	0.951	0.776
	(0.133)**	(0.141)**
Presidential Party Largest	0.521	0.799
	(0.076)**	(0.020)**
Assembly Fragmentation	− 0.328	0.005
	(0.131)*	(0.050)
Constant	− 1.766	− 1.736
	(0.091)**	(0.083)**
Observations	314	314
Robust standard errors in parentheses, * significant at 5 percent; ** significant at 1 percent.
^a Source: Author's database. See the Appendix for detailed information about sources.

Table 4 Recruitment of Ministers with Political Competence to Manage Repressive Agencies and Economic Actors.^a

	(1)	(2)	(3)	(4)	(5)	(6)
	Coercive Agencies	Coercive Agencies	Business (all)	Business (all)	Business (primary career)	Business (primary career)
Democratic Strategy	− 0.233		− 0.227		0.055
	(0.108)*		(0.071)**		(0.087)
Democracy (FH)		− 0.380		− 0.263		0.005
		(0.103)**		(0.079)**		(0.093)
Economic Crisis	− 0.132	− 0.118	− 0.033	0.030	0.265	0.271
	(0.118)	(0.130)	(0.069)	(0.083)	(0.077)**	(0.093)**
Presidential Party Largest	0.164	− 0.077	0.151	0.015	0.118	0.112
	(0.067)*	(0.082)	(0.047)**	(0.057)	(0.056)*	(0.069)
Proportion GDP Private			0.240	0.391	0.264	0.227
			(0.151)	(0.119)**	(0.159)	(0.128)
Constant	− 2.111	− 2.109	− 1.208	− 1.207	− 1.809	− 1.808
	(0.112)**	(0.099)**	(0.074)**	(0.074)**	(0.080)**	(0.080)**
Observations	314	314	314	314	314	314
Robust standard errors in parentheses, * significant at 5 percent; ** significant at 1 percent.
^a Source: Author's database. See the Appendix for detailed information about sources.

These results support the analysis of experts who have argued that the progressive depletion of legislative expertise in Russia's cabinets is indicative of a more authoritarian approach to governance. As Huskey puts it with reference to Putin's cabinets up to 2008 “[o]ne of the first conclusions to be drawn from the post-communist Russian elite is that it is almost entirely bereft of leaders whose career path includes elective office…. The Russian case, therefore, contrasts sharply with patterns of political recruitment found in the democratic world” (Huskey, 2010, pp. 363–364).

Table 4 examines how far the declining commitment to democracy increased the recruitment of ministers with political experience to manage actors who can be expected to become salient in a more authoritarian policy process—the coercive agencies and key economic client groups. Models 1–4 show that when presidents express a desire to govern democratically, they are significantly less likely to recruit ministers with a career background in the military or security agencies (models 1 and 2) and with links to key economic concerns (models 3 and 4). Both effects are of sizable magnitude: In models 1 and 3 a one standard deviation increase in democratic commitment reduces the odds that a minister with a coercive or business background is appointed by 21 percent and 20 percent respectively (significant at the 5 percent and 1 percent levels). These results are robust irrespective of whether the content analysis or the alternative Freedom House measure of governing strategy is used. Put differently, as a president's commitment to govern democratically declines, the probability that ministers will be recruited who can manage these central client groups of the regime rises. Both models control for contextual influences including the economic crisis, the size of the pro-presidential party and, in the case of regressions 3 and 4 the size of the private sector in the economy. None of these factors have consistent effects on ministerial recruitment. The central reason for the enhanced presence of ministers with coercive and economic backgrounds in the government, in other words, is the changing governing strategy of a president.

These results contribute to and clarify the debate about the trend toward the increasing “militarization” of Russia's national leadership and its implications (Kryshtanovskaya and White, 2003; Renz, 2006; Illarionov, 2009). While the siloviki are indeed drawn from a range of backgrounds (Bremmer and Charap, 2006–2007, p. 87), and hold a variety of political views (Renz, 2006), the crucial question is why they are being promoted to central positions in government. By early 2007, this group controlled more than 10 government agencies and had partial control over several more (Bremmer and Charap, 2006–2007). As the analysis above shows, the central factor predicting their recruitment into government is the eroding presidential commitment to democracy. Irrespective of the individual outlooks these ministers may have, their extensive presence in government puts at the disposal of the Russian president the political expertise to employ security and military agencies in the pursuit of a range of political aims. Indeed, the recruitment of ministers with the political expertise to manage the coercive agencies has been matched by the use of these agencies in a variety of policy areas, including most prominently, perhaps, the control of the North Caucasus region with the continued and extensive use of armed force (e.g., O'Loughlin et al., 2011), and the progressive restriction of the freedoms of assembly and association backed by overwhelming police responses to opposition protests.

Potentially more ambiguous is the increased presence of ministers with business links in government, and indeed, it has variously been interpreted as a strategy to manage and co-opt important client groups, or as signaling the embourgeoisement of the state given the transition from a planned economy to the market, and the growing share of the private sector in Russia's economy (Kryshtanovskaya and White, 2005; Rivera and Rivera, 2006, p. 131). Models 5 and 6 (Table 4) therefore examine the career backgrounds of ministers with business experience more closely: while a process of embourgeoisement of the government would suggest the influx of ministers into cabinet whose primary careers are in business and who genuinely represent business interests, the management of key economic client groups can be achieved through the appointment of ministers who are primarily government administrators with some experience of cross-posting onto the boards of major Russian companies. Models 5 and 6 focus on ministers whose primary careers are in industry or business, not government administration. Again, the models control for alternative explanations. The results suggest that governing strategy had no effect on the recruitment of ministers whose primary career was in business—irrespective of whether the content analysis or Freedom House measure is used. Put differently, the increased presence of ministers with business links in government appears to have been driven by the recruitment of government administrators who had been cross-posted to sit on the boards of Russian companies. Thus, as the commitment of Russia's presidents to democratic governance eroded, they named administrators who had the political expertise to manage key economic client groups in the interests of the president, a policy that was consistent with the progressive limitations that Putin imposed on a range of oligarchs and big business interests in the 2000s.

Again, these findings help to clarify a central debate about the political nature of Russia's cabinet. The changing patterns of ministerial selection do not appear to imply a progressive embourgeoisement of the Russian cabinet. Rather these results lend support to the analysis of experts who have argued that “besides the militarization of labor in officialdom, a notable feature of the Putin era has been the etatisation of leadership in industry” (Huskey, 2010, p. 367).

In sum, the results of the logistic regression models presented in Tables 2-4 jointly give a consistent picture which shows that the variation of presidential governing strategies along the democratic-authoritarian dimension has a statistically significant and substantively powerful effect on ministerial selection. As the commitment of Russia's presidents to a democratic governing strategy eroded and gave way to increasingly authoritarian preferences, presidents secured incentive compatibility with their ministers less by relying on party affiliation, and more by personal loyalty links. Moreover, they changed the type of political expertise they recruited into their cabinets, lowering the reliance on ministers who had the experience to work successfully with the legislature, and instead privileging ministers who had the political skills and expertise to manage the coercive agencies and key economic client groups.

Conclusion

This article has examined ministerial selection in non-party cabinets and probed what political outlook different selection patterns reflect. The central argument is that variation in the governing strategies of politicians along the democratic-authoritarian spectrum is a critical determinant of the attributes and competencies that politicians choose in non-party ministers. This argument is tested in a multivariate logit analysis of 314 ministerial appointments in Russia, from 1992 to 2008. The results lend broad support to the expectations derived in the article.

These findings contribute to two distinct literatures, comparative work on ministerial selection, and the literature on political and governmental elites in Russia. To the comparative literature on ministerial selection the article contributes a better and more nuanced understanding of cabinets with a significant share of non-party ministers. Approaching ministerial selection from a principal-agent perspective, the article shows that a politician's preferences to govern in a more or less democratic manner shape how they select incentive compatible ministers and what type of political competence they seek in their cabinet. This result speaks to the starkly polarized views in the debate about non-partisan ministers, who are regarded by some scholars as cronies of inherently non-democratic provenance while others argue that such ministers may bring technical competence to cabinets that can help to legitimize democracy by enabling the government to tackle technically complex policy challenges. The findings of this article help to resolve this debate by reformulating our understanding of the political nature of such cabinet appointments. The theory and evidence presented here make clear that that the appointment of non-partisan ministers per se does not allow for any inferences about the politics of a cabinet.

Instead it is the attributes and competences politicians seek in their ministers that convey information about their preferences and governing strategies: When politicians plan to govern democratically, they are more likely to choose ministers who have party political and legislative expertise, recruit less through loyalty, and have a much lower probability of seeking links of the cabinet to coercive agencies or key economic client groups than politicians whose outlook is essentially non-democratic.

Second, the findings of this article contribute to a range of debates about the interpretation of changes in Russia's governmental elites. While this article builds on previous work, it also offers a novel approach that improves on previous efforts in several respects: it is the first study to embed the analysis of cabinet composition in Russia in a more general theory of ministerial selection by more and less democratic politicians, the first to develop a content-analysis based measure to gauge the president's preferred governing strategy through the analysis of annual Presidential Addresses to the Federal Assembly, to employ a complete dataset of cabinet ministers from 1992 to 2008, and to use a multivariate regression approach instead of descriptive analysis. This approach makes it possible to study the impact of presidential governing strategy on ministerial choice as a testable empirical question. The results clarify several important debates among experts of Russian politics about changes to the governmental elite, and whether or not they signal the militarization, increasingly technocratic-authoritarian nature, or embourgeoisement of Russia's cabinets. The findings make clear that the increasingly authoritarian preferences of Russia's presidents powerfully predict the rising reliance on loyalty, and the recruitment of ministers with political skills to manage the military and coercive agencies as well as key economic client groups. The same erosion of the willingness to commit to democratic governance also markedly reduced the presence of ministers with party political expertise and the political skills to manage negotiations with the assembly. In short, by spelling out theoretically what patterns of ministerial selection we can expect to find in the cabinets of more and less authoritarian politicians and testing these expectations empirically, the article makes clear that the changes in the composition of Russia's cabinets do indeed mark an authoritarian turn.

Notes

¹ University Lecturer in Politics, St. Hilda's College and Department of Politics and International Relations, University of Oxford, [email protected]. I gratefully acknowledge comments on earlier versions of this article by Cecilia Martinez-Gallardo, Edward Morgan-Jones, Paul Chaisty, and participants of the Monday Seminar at St. Antony's College, University of Oxford. I would also like to thank Alisa Voznaya for superb research assistance.

² See Gill (2012) for a rather different view.

³ All of the models presented below were also run in random effects specifications. In none of them is the random effects coefficient statistically significant.

⁴ All statistical analyses are performed using STATA SE (version 11).

⁵ The odds ratios referred to in interpretations of the magnitude of the effects are simply exponentiated coefficients. They are therefore not reported separately here.

Appendix: Data Sources and Coding

Cabinets Formed, 1992–2008

Yel'tsin: Gaidar June 1992, Chernomyrdin December 1992, January 1994, August 1996, Kiriyenko March 1998, Primakov September 1998, Stepashin May 1999, Putin August 1999

Putin: Kasyanov May 2000, Fradkov March 2004, May 2004, Zubkov September 2007

Ministers Appointed to Each Government

Data sources: Presidential decrees, newspaper reports, Russian government website. Accessed via http://www.eastview.com, http://www.government.ru

Ministerial Attributes

Data source: Integrum biographical database http://www.integrumworld.com

Coding: Ministerial attributes are indicator variables (1 = attribute present, 0 otherwise). All attributes (except personal loyalty) are recorded for 5 years prior to the appointment (or re-appointment) of the minister. Personal loyalty is recorded without time limit.

Attribute	Definition
Pro-Presidential Party experience	Membership of a party that typically lends legislative support to the president
Loyalty	Post working with future president in same institution at same time prior to president's election
Legislative experience	Service as deputy in Congress of Peoples' Deputies or Federal Assembly
Coercive agencies	Appointment in KGB/FSB, military, or ministries related to defence or law enforcement
Business (all)	Appointment in state or private company including as non-executive board member
Business (primary)	Primary career in state or private company

Presidential Governing Strategy: Content Analysis Variable

Data Sources: Presidential Addresses to the Federal Assembly (Poslaniye Prezidenta Rossii Federal'nomu Sobraniyu RF), 1994–2007, http://archive.kremlin.ru/, http://www.intelros.ru/strategy/gos_rf/psl_prezident_ rf_old/.

Coding: References to keywords with roots that refer to democracy, freedom, rights, representation and accountability were recorded (demokrat-, svobod-, prav-, predstav-, otvetstven-). Where a term had multiple meanings, only references to the political meaning were coded. In each case it was recorded whether the reference was positive or negative/qualified.

Presidential Governing Strategy: Freedom House Variable

Data source: http://www.freedomhouse.org/.

Coding: Democratic governing strategy (1 = partially free, 0 = not free).

Political and Economic Context

Economic Crisis

Data sources: Angus Maddison, Statistics on World Population, GDP and Per Capita GDP, 1-2006 AD (update: March 2009) http://www.ggdc.net/maddison/, and OECD (2009).

Coding: 1 = no growth or negative growth, 0 otherwise.

Presidential Legislative Support

Data source: http://www.russiavotes.org/.

Coding: 1 = presidential party largest, 0 otherwise.

Effective Number of Legislative Parties

Data sources: http://www.russiavotes.org/, Remington (2001).

Coding: Laakso-Taagepera Index.

Proportion of GDP Accounted for by Private Sector

Data sources: World Bank, Russian Economic Trends, http://web.worldbank.org/, Rosstat (2003, 2009).