Formal power in informal networks. Distribution of power resources in personalized bureaucracies: the case of Russia’s subnational elites

ABSTRACT The instrumental use of patronage for political appointments and career advancement can be found across all regime types and historical periods. Despite a pronounced academic interest in the political and economic effects of patron-client relationships, little is known about the nature of interplay between informal networks and formal hierarchies. How are formal powers distributed in personalized bureaucracies? The article addresses this question through a cross-temporal case study of subnational government in one of Russia’s regions – Sverdlovsk Oblast. Based on network analysis and negative binomial regression, the article shows that subnational leaders and their clienteles seek to monopolize those formal powers that allow administrative control over other executive agencies, while other coercive, financial, and normative powers are distributed relatively evenly. This pattern persists regardless of institutional context, degree of subnational autonomy, or the shape of informal networks, which signifies the importance of controlling functions that the core of informal networks in personalized bureaucracies performs.


Introduction. Patronage and informal networks: structure and effects
Patronage 1 is a multifunctional political phenomenon that may determine appointment decisions, structure career advancement, foster elite cohesion, or direct rent distribution.Although its functionality may be restricted by ideological and institutional constraints, patronage remains in demand in various political settings.A conventional view considers patronage an essential part of the authoritarian regimes' toolkit, 2 yet even liberal democracies are far from eradicating it.Both parliamentary and presidential systems often allow parties and presidents to reward their political allies with posts in executive agencies. 3Political dynasties as a distinct form of incumbency advantage have been largely documented in various democratic systems. 4Generally, with a high variation in its scope, forms, and functions, patronage can be found across all political systems and historical periods.
Taken in their complexity, patron-client relationships can be seen not simply as an asymmetric dyadic relationship of personal dependency concerned with reciprocal exchange of favors 5 but also as the basis for the specific social and power structures, often referred to as informal (or patronage) networks.Although vertical patronal relations are a linchpin of informal power networks, the latter can accommodate a variety of vertical and horizontal personalized ties such as nepotism, favouritism, or blat.
The particular structure of informal networks may vary and produce different micro-and macro-effects.For instance, the position in an elite network has been shown to influence the chances of achieving a Politburo membership for the Chinese Communist Party functionaries. 6Network position can also be a predictor for taking part in collective action.An analysis of the elite network on the eve of the 1991 Haitian coup d'état shows that families with higher centrality were more likely to be accused of mounting the coup. 7Analogously, as Padgett and Ansell's seminal study shows, the rise of Medici in fifteenth-century Florence can be explained not by the wealth or number of direct connections with other Florentine families through marriage or trade alliances, but rather by the strategically advantageous position they had in the network of noble families. 8The idea of elite structures' impact on political dynamic has found fruitful application for explaining the strikingly divergent pathways of political regimes with shared historical background, such as post-Soviet states.As Hale shows, 9 the presence of competing patronage pyramids and an inability for each pyramid to mobilize resources was one of the main sources for presidential ousters in Ukraine, Georgia, and Kyrgyzstan.At the same time, single-pyramid structures such as those in Belarus, Russia and Azerbaijan have allowed their leaders to freeze the political dynamic and avoid the Achilles' heel of patronal presidentialism -"lame duck syndrome."These findings buttress the theoretical assumption expressed at the outset of clientelism studies by Scott, who noted that patronage structures could vary in: a) the degree of monopoly over local resources by a single patron, b) the degree of monopoly over links to other structures by a single patron, c) the density of patron-client linkages in the population, and d) the extent of differentiation between different pyramids and clusters. 10erefore, the structural properties of elite networks are crucial for the outcomes of political cooperation and conflicts in the polities where political pursuits are predominantly organized through the chains of actual acquaintances.
The dominating academic stance towards patronage and informal networks considers both phenomena dysfunctional for public administration and economic development.Patronage is linked to corruption and rent-seeking, 11 bureaucratic inefficiency and professional incompetence, 12 and democratic backsliding. 13Patronage is also considered responsible for creating perverted incentives for political mobility, 14 hampering meritocracy, 15 and undermining international institution-building projects. 16he opposite thread of research underscores functionally positive traits of patronage.It can offer an informal solution to the principal-agent problem inherent in the bureaucratic hierarchy, especially when incentives to generate economic growth combine with rigid formal rules and tight political control.As Jiang shows, city leaders in China can achieve higher economic performance if a provincial-level patron backs them. 17Patron-client relations may form long-term alliances that guarantee rewards credibility, help to circumvent ineffective formal constraints and insert direct pressure on clients to support their patrons' political and economic priorities.These informal mechanisms might offer an insightful clue for understanding the reported effectiveness of Chinese bureaucracy in the lack of democratic institutions.This view resonates with research on political foundations of the South Korean rapid economic growth known as the Miracle on the Han River.Personal, often marital, relationships between political and business elites deterred the former from the unilateral revision of property rightsthe main obstacle for long-term investments in authoritarian regimes.Personal relationships also gave сhaebols' owners access to the decision-making process and cheaper loans, although it obliged them to contribute financially to the South Korean ruling party campaigns.This mode of relations fostered what Kang called the "mutual hostages" model 18 the equilibrium of interests within political and economic elites that prevented one another from violating informal rules, thereby reducing transaction costs and promoting growth.Informal networks can also play a key role in state-building.As Easter shows, the Bolsheviks managed to spread their rule across the territory of the former Russian Empire in 1920s using personal ties that they struck up when working in the pre-revolutionary underground and fighting together on the battlefronts of the civil war. 19Given the radical change in formal rules and the devastated political, financial, transport, and even communication infrastructures that accompanied the Soviet state-building project, its story also shows that informal institutions may serve as a "safety net" when formal institutions are disrupted.
Yet another strand of research does not make particular assumptions regarding functional and dysfunctional implications of patronage but instead seeks to estimate its impact on elite mobility.The study of careers of the regional party secretaries in the post-Stalin USSR shows that the recipe of political advancement combined rational-technical competencies (such as expertise and performance) and patronage links with allunion elites. 20A number of studies examine the combination of patronage and performance criteria for an elite advancement in another one-party system -China.While some studies underscore the equally important role of performance and factionalism, 21 others argue for a predominant role of political connections. 22Keller brings an innovative insight into these debates.By utilizing network analysis, she demonstrates that it is not even necessary to know a patron's identity to predict the chances of being promoted to Politburo, as an aspirant's place in an informal network is often more revealing. 23etwork analysis can even help identify patronsthey usually hold intermediary positions.The importance of network intermediaries is also shown by the research on the Mexican elites proving that network brokers and gatekeepers were the primary allies for building power coalitions during consolidated one-party rule. 24he extensive scholarship on political patronage and informal networks is unified by its focus on the external effects of both phenomena, while the structure of informal networks remains a neglected subject, which is especially true when applied to patronage's natural habitatexecutive agencies.If personal connections infiltrate bureaucracy, how does its formal hierarchy correlate with informal networks?Do the higher formal positions coincide with the centre of a network?If bureaucracies are not homogenous themselves and accommodate different types of power resources, how are these resources distributed within informal networks?What kinds of positions constitute the most wanted prize for competing informal groups?Formal hierarchies and informal networks function based on different, often mutually exclusive, principles.Nevertheless, despite inevitable tensions, personalized bureaucracies accommodate both types of structures along with the rules and principles that each carries.However promising, the questions of interplay between them remain an enigma for the existing scholarship.Part of the problem stems from the lack of reliable methodological tools for modelling and measuring both parts of the equationformal powers and informal networks.
The outlined questions are particularly important for analysing political systems exhibiting the institutional dominance of the executive as they give the top officeholders a considerable and often disproportionate share of power.Post-soviet states represent a promising analytical arena in this regard.They combine a long-standing tradition of using personal connections in political and economic interactions 25 and considerable volume of power their bureaucracies wield.Russia represents a highprofile case of such a combination.Its significant state economy empowers Russia's bureaucracy with substantial controlling, redistributive and regulatory functions. 26eparation of powers at national and subnational levels proves to be highly skewed towards the executive. 27At the same time, loyalty and personal trust established before the appointment remain the dominating principles for providing access to public office and career mobility, which is valid for both executive elites and streetlevel bureaucracy. 28The widely-acknowledged dominant role of administrative elites in Russia is surprisingly combined with rather limited knowledge about the distribution of power within its state apparatus.Wielding a high volume of power resources, the executive in Russia is far from being a unitary actor.The bureaucratic conflicts often spill over to other arenas, such as the parliament floor, 29 which signifies a high degree of intraexecutive factionalism and the analytical necessity to assess the relative power of the executives vis-à-vis each other.The existing scholarship provides a good starting point for such a task.Inferring power of an office-holder in personalized bureaucracy should take into account both the formal powers of their post and their place in an informal network.This article addresses how these two dimensions combine in the context of subnational government in Russia by examining the case of Sverdlovsk oblast, a region of Russia.The empirical analysis is based on two unique databases.The first comprises structured biographical data, which serves as the basis for modelling an informal network of regional officials.The second database consists of budgetary, legislative, and other archival data allowing research to offer a novel approach to infer the volume of power resources a particular post gives its holder.With the help of network analysis and negative binomial regression, the findings suggest that a specific combination of these powers follows a distinguishable pattern.Those powers that allow for monitoring the bureaucracy are always monopolized by the governor's clientele, while other formal powers are distributed within the networks relatively evenly.

Subnational executive elites in Russia: formal powers and informal networks
The role and influence of subnational executive elites in Russia have not been constant throughout the country's post-Soviet history.From a neo-institutional point of view, their political standing can be seen as the function of two related factors: formal powers and informal networks.The former provides regional elites with different types of administrative resources, while the latter ensures their interpersonal cohesiveness.Together, formal and informal resources determine the elites' ability to negotiate and compete with federal authorities and regional contenders.The volume of these resources has been the main object of centre-regional relations in Russia and varied from what can be called a "chaotic decentralization" of the 1990s to the solid "power vertical" put up during the last 20 years. 30

Formal powers
The formal institutional powers of the regional executives embrace two dimensions.The vertical one defines their relations vis-à-vis the federal centre, while the horizontal one relates to the strength against regional legislatures.
Although the first decade following the collapse of the USSR differentiated Russia's regions in their formal standings towards the federal centre, with republics having a superior formal status compared to all other regions, it generally brought all subnational units substantial political and institutional autonomy. 31Executive elites in all regions gained the power to appoint the heads of key federal ministries in their regions, such as police heads and prosecutors.The share of regional budgets in the consolidated Russian budget had risen to 60% by 1998.Importantly, heads of regional administrations had a seat in the Federal Council, the upper chamber of the Russian parliament, which gave them parliamentary immunity.As a result, extensive formal powers made the regional executive elites genuinely independent political actors able to challenge the federal centre.To illustrate, more than 20 governors signed a proclamation for Yeltsin's impeachment in 1999; even greater number of governors were members of the leading opposition party -KPRF.
Shortly after Vladimir Putin came to power, regional autonomy started to change, and governors lost their appointment discretion regarding prosecutors and police heads.The president secured a right to dismiss a governor found liable for the violation of federal law.In 2005, the direct gubernatorial elections in Russia were eliminated.The president received power to appoint regional leaders who then had to be confirmed by subnational legislatures.Although the direct elections of governors were reinstated in 2012 as a response to mass protests, the new law stipulated the so-called "municipal filter," which effectively blocked independent candidates from running for office.Since then, there were only four out of 150 cases where a Kremlin-backed candidate lost gubernatorial elections to an opposition candidate.Thus, since 2004, the federal centre has remained in complete control over the appointment process of regional head executives.At the same time, by 2005, almost all power-sharing treaties between the federal centre and subnational units were abolished, with Tatarstan being the only region that managed to sign a new one.When the term of the treaty came to an end in 2017, the Kremlin refused to prolong it, which ultimately made the formal powers of Russian regions fairly homogeneous and solely defined by the federal laws.
While losing their autonomy against the federal centre, subnational executive elites strengthened their position vis-à-vis regional legislatures.The 1993 Constitution gave regions freedom to define their institutional design.By 1998, almost all regions had made their institutional choice, which differed across Russia, although subnational assembles turned out to be generally more potent in republics than in other regions.
Putin's centralization policy led to the ubiquitous redefinition of the regional constitutions and statutes.Governors managed to embrace the opportunity and strengthened their powers akin to presidential standing against the federal parliament.Golosov and Konstantinova applied the Shugart and Carey presidential powers index to juxtapose the volume of gubernatorial powers in 2002 and 2014 and testified to two underlying trends.By 2014, executive elites acquired the decisive volume of formal powers to control the political arena within regions, while the cross-regional variation of those powers almost disappeared. 32

Informal networks
Unlike formal powers, informal networks of subnational elites in Russia remain a neglected subject. 33A groundbreaking study of elites in Tatarstan and Bashkortostan shows that a particular network structure may be highly consequential for the outcomes of political bargains with the federal centre. 34By 2010, two neighbouring regions faced a succession problem, yet only Tatarstan's elites managed to keep control of the region.Tatarstan's elite structure turned out to be cohesive enough to produce a successor and deter the federal centre from appointing an outsider.
Sverdlovsk Oblast often appears in the literature as another example of a region with strong and cohesive executive elites that buttressed the region's ambitions in political bargains with the federal centre during the 1990s.Sverdlovsk Oblast attempted to upgrade its formal status to the republic level in 1993.The region became the site of the first direct gubernatorial elections in Russia following a presidential moratorium introduced in 1994.It was the first non-republic region in Russia to sign a power-sharing treaty with the federal centre in 1996, which granted the region increased institutional and financial autonomy.The region's first governor -Eduard Rossel'stayed in power for almost 15 years and is always considered among the so-called "regional heavy-hitters," i.e. subnational leaders who managed to create powerful political machines in their territories.At the same time, this region swiftly overcame enormous economic challenges caused by the collapse of the USSR.The breakdown of central economic planning was especially detrimental for the regions whose economies were linked with the defense sector and heavy industry, with the state as their primary customer.Sverdlovsk Oblast, being among these regions, managed to successfully diversify its economy and quickly became a donor region.Qualitative accounts of Sverdlovsk Oblast's political and economic success during the 1990s underscore the importance of the elite network Rossel' managed to consolidate. 35he complexity of data needed to address an interplay between formal powers and informal networks empirically makes a single case study the most suitable research strategy.Studying Sverdlovsk Oblast gives us two theoretical advantages.First, it speaks to already established research interest in the region; therefore, the results of this study can be triangulated and supplemented by existing qualitative accounts.Second, it may partially alleviate the generalization problem that inevitably arises as a major limitation of a single case study.Limits of generalization are a valid concern, even for Russia itself.Although existing scholarship testifies to a solid unification trend, Russia's regions' diverse demographic and economic structure may still generate some residual heterogeneity in their formal and informal autonomy.Part of this heterogeneity may be taken into account by a cross-temporal perspective if we identify a region that has undergone a significant downward transformation of its autonomy.Sverdlovsk Oblast is such a case.
The era of Rossel's leadership ended after his 15 years in power when then-president Dmitry Medvedev did not offer him a new term.He instead nominated Alexander Misharin, Russia's former Deputy Transportation Minister and member of the Board of Directors of Russian Railways.Having little roots in the region, Misharin removed all the previous elite from executive positions and eliminated regional political institutions that ran counter to federal legislation, such as bicameral assembly and staggered elections.The current governor -Yevgeny Kuivashev, is also an outsider to the region and considered a part of the mighty Tyumen patronage pyramid led by Sergey Sobianinmayor of Moscow.Thus, apart from institutional recentralization, Sverdlovsk Oblast underwent a significant reduction in informal autonomy, concerned with dismantling cohesive elite networks and administrative intervention of the federal patronage pyramids.Therefore, this case provides the necessary cross-temporal variation in the formal and informal subnational autonomy.Despite a gradual decrease in autonomy, Sverdlovsk Oblast, particularly its capital city Yekaterinburg, remains one of the most opposition-minded localities in Russia.In national elections since 2004, the region consistently provided less support to the ruling party and its candidates compared to the nationwide average.In 2013, the governor's protégé lost the mayoral election in Yekaterinburg to a local civic activist, Yevgeny Roizman, which garnered nationwide attention.Five years later, the regional legislature abolished direct mayoral elections in the city.Nevertheless, the region's diverse political and economic landscape hinder the federal authorities from easily replacing its leadership.Over the past 25 years, only five Russian regions have experienced less turnover in their governor positions than Sverdlovsk Oblast. 36

Data and method
Formal powers and informal networks are dynamic phenomena.Governor rotations often entail reconfiguring the personal corpus of regional ministers and redefining formal powers between them.Therefore, before embarking upon the analysis, it is necessary to decide on the periods under consideration.Since the first gubernatorial elections in Sverdlovsk Oblast in 1995, three governors have succeeded each other -Rossel' (1995-2009), Misharin (2009-2012), and Kuivashev (2012-present).The logic for choosing particular time intervals for each governor's administration involved identifying the most stable periods when their apparatuses functioned without significant structure and personal composition changes.Thus, I excluded the first years in office for each governor, as well as periods right after the preterm government (or its chairman) resignations (2007, 2012) or major administrative reforms that altered the regional government's structure (1997, 2014, and 2016 reforms).Ceteris paribus, later periods were preferred due to better data availability.As a result, the analysis provides three temporal snapshots of the regional executive elites: 2004-2005, 2011-2012,  and 2019-2020.

Formal powers
An accurate estimation of formal powers is often needed for testing various theoretical assumptions.Gauging the importance of ministerial portfolios, however, remains a non-trivial task.Apart from remarkable country-specific exceptions, 37 the existing approaches to resolving this task rely heavily on expert surveys. 38While expert surveys have benefits, such as country-specificity, cross-national comparison, and interval ratings, relying solely on them to gauge portfolio salience has limitations.The first is time invariance.Unless the survey is longitudinal, it hits the boundaries of human memoryestimates of a ministry's importance for a period even slightly remote from the survey might be misleading if at all possible.Moreover, even concordant expert estimates may reflect not a ministry's importance but rather its public image.Even more critical is that the very concept of "importance" is not detailed enough to address the multifaceted nature of power.Executive posts might be strikingly different regarding what kind of power they give their holders, how inclusive these powers are, and how autonomously they can be exercised.To understand how patronage-based bureaucracies are organized, it is crucial to consider this diversity.A single case study allows for such a more complicated picture.
To this end, the study proposes a new method to infer the formal powers of the officeholders, which rests on the concept of "power resources."It is conventional to distinguish between three basic forms of social power. 39The first is coercive power, i.e. the ability to exert physical force, which is grounded in possession of special facilities, such as weapons and prisons, and inequality in potencies.The second is economic power, i.e. the ability to distribute scarce products.The third is normative power, i.e. the ability to impact an individual's or group's behaviour by promoting certain ideas and establishing them as mandatory. 40ureaucracy as an embodiment of state authority possesses all three types of power; therefore, this tripartite distinction provides a reasonable starting point for operationalizing formal power.Speaking of Russia's regional bureaucracy, it is important to note that coercive power in its pure sense rests solely in hands of the federal centre, which manage all agencies that embody state monopoly over violence (i.e.army, police, special services).However, regional agencies hold responsibility for control and oversight within their policy jurisdictions, which includes the power to administer penalties and suspend or terminate the activities of non-compliant entities, which is the essence of coercive power in a broader sense.Therefore, when speaking of regional executives, this study defines coercive powers as an ability to control and oversee the everyday activity of organizations and individuals and penalize them for breaching the rules.
These coercive powers can be derived from the regional bylaw and ministries' charters for further empirical examination.The database with all relevant coercive powers contains 319 records, sorted by the relevant officials and time periods.What complicates further comparison is that not all coercive powers are equal.For example, the finance minister's ability to oversee budgetary compliance and hold other agencies criminally responsible is far more consequential than the oversight over seed production executed by the minister of ecology.Therefore, it is reasonable to distinguish between types of controlled actors.At least four groups can be discerned in the Russian context: 1) NGOs, unitary enterprises, and budgetary institutions (e.g.hospitals, schools, theatres); 2) municipal administrations and other local bodies; 3) businesses; and 4) regional governmental bodies.It is plausible to assume that these actors possess their own power resources in ascending order.Thus, the ability to control a particular NGO activity would bring a score of "1" while oversight over regional ministries would bring the score of "4."The second distinction between coercive powers is their degree of inclusiveness.To illustrate, the minister of education can revoke the license to carry out educational activities, which may affect private companies working in the educational market.In contrast, the minister of economy and labour is empowered to ensure labour rights compliance by all the companies working in the region.If the range of the controlled actors is clearly delimited, as in the case of the educational licenses, the score is divided by two.
Financial power has several facets.Budget is essential resource ministries seek to maximize.Theoretically, the more budgetary funds a ministry enjoys, the more impact in policy-making it has.Therefore, the share of the regional budget allocated to a particular agency might serve as the first proxy for measuring financial power.Secondly, maintaining financial growth and avoiding budget cuts demonstrates a minister's capacity to garner support from executive and legislative actors.In this study, the difference between a ministry's initial budget share, envisaged by the annual budget bill, and the executed funds at the financial year's end is used to measure its budgetary stability.However well-funded a ministry is, it is only part of its financial power.Examining a particular official's role in the budget process is also essential.Some officials can redistribute the budget funds within their policy area.For example, the culture minister is entitled to reallocate funds among subordinate theatres or museums, which brings her a score of "1."A select group of officials can allocate means from the reserve fund, allowing them to subside particular agencies' activity discretionally (score of "2").Furthermore, an even smaller circle can initiate or approve the redistribution of budgetary funds between policy areas, granting them substantial leverage over agencies' fiscal capacities (score of "3").The cumulative score of these three measures represents the composite financial power of an office.
Normative power reflects how active a particular office is in redefining regional legislation.The ability to influence the legal landscape allows for reconfiguring the rules of the game for different regional actors, advancing the agency's policy-making agenda and expanding its sphere of interest.Three basic types of legal documents exist on a subnational level in Russiaregional law, government executive order, and governor's decree.Since the first two are issued by collective bodies, it may seem challenging to ascribe a particular piece of legislation to a specific official.However, the agendas of the government's and assembly's sittings (N = 163) provide information on who was reporting on a particular bill or order, clearly showing who was responsible for their drafts.The share of such pieces calculated based on the available archival data quantifies the normative power.
An additional form of power within bureaucratic hierarchies not captured by the three basic types discussed above is the power of appointment.Appointment grants access to formal positions and accompanying power resources.The appointment relations among all pairs of regional officials can be visualized as a matrix (Online Appendix A1).The extent of an official's appointment power can differ in two aspects.Firstly, the ability to select powerful officials is much more valuable than appointing those less powerful.Secondly, the specific role in the appointment process may vary from exclusive power not bound by consent of other actors (such as the regional legislature or federal ministries) to indirect influence, such as participation in the nomination preparation.In the Sverdlovsk case, five gradations of such roles can be discerned (Online Appendix A1).Therefore, appointment power represents the product of the appointee's relative influence and the official's role in the appointment.
These four types of formal power are conceptually independent and can be modelled separately.However, they have a critical dimension that divides them into two broader classes.Each type and subtype of formal power has a subject over which it can be exercised.One subset of powers targets executive agencies themselves.Appointments, oversight over other agencies, and redistributive functions in the budget process jointly define how much power a particular official has over the entire regional bureaucracy.We can denote this particular combination as a supervisory power.Contrarily, significant budget share, financial stability, budgetary discretion within the policy jurisdiction, and high regulatory activity, accompanied by the supervisory powers over non-governmental actors, provide a minister with an influence on the overall socio-economic process in the region.We can denote this class as a substantial power.The positions of health or education ministers, for instance, bear little supervisory yet much substantial power.Contrarily, vice-governors' functions mainly involve coordination and supervision of the regional apparatus yet often lack any direct leverage on particular public economy sectors.Importantly, all elements of formal power divide between these two categories without any overlap and residuals.It is fruitful, then, to estimate how these broader power combinations are distributed within informal networks.
For comparability, each primary type of power is normalized to a scale of 100, representing the proportion of power held by the official.The maximum score for supervisory and substantial powers is set at 300, as each encompasses three power subtypes.The operationalization scheme is summarized in Online Appendix A.

Informal networks
Modelling interpersonal networks is a challenging task, especially when it comes to political elites.Their personal connections are hardly observable and can be inferred indirectly at best.It is possible to distinguish between exploratory and structured approaches to this task. 41While the former relies on qualitative assessments of the insider sources regarding friendship and support, the latter takes advantage of publicly available data and infers personal connections through certain shared characteristics.These characteristics are contextual and vary based on the structural qualities of societies of interest.If marriage ties structured Haitian elite networks in 1991, 42 coworker ties represent the most relevant source of personal connections in the current Chinese context. 43Shared educational and professional background unanimously appears in both academic literature 44 and insider accounts 45 as the main source of patronage connections in the Russian context.In line with this assumption, the existing accounts of elite networks in Russia and the USSR often use biographical analysis to infer patronage connections. 46The presented study adheres to the same approach as the first step.It requires collecting a biographical database that contains information on career paths of Sverdlovsk Oblast's highest officials.The local encyclopedia 47 served as the leading source of such biographical data.The resulting database consists of 108 individuals and 885 person-period observations.Each includes years, cities, and organizations these individuals worked or studied.The R-script (written by the author) searched this database for interceptions in career paths that preceded entering regional government.
Although based on open sources and allowing for replicability, this approach may generate its own bias.First, shared educational or work experience does not necessarily mean positive and loyal interpersonal relations.Ex-colleagues might have hostile or apathetic attitudes or even remain unfamiliar with each other if the organization is too large.Second, personal connections can be fostered outside the shared workplace or university.Common hobbies, neighbourhoods, business interests of the families, or charity initiatives can be equally solid sources of personal trust and loyalty.A pure biographical approach would mistakenly bring the former into the model and miss the latter.In order to mitigate these difficulties, the resulted ties were validated with expert surveys (n = 19) conducted in May-October 2020.The expert pool encompassed local journalists, deputies, analysts, and political scientists with knowledge of local politics.Independently from each other, the experts were asked to identify stable and lasting relationships of trust and friendship between the actors and amend a sociogram resulting from the biographical analysis.If two or more experts made the same amendment (either removed or added a tie), this change was brought into the network model. 48etwork analysis provides various measures to assess an individual's network position.Among numerous centrality measures, three are considered fundamental and capturing distinct network power forms.The first is degree centrality, which simply counts the number of connections a node (actor) has.This measure is relevant for analysing elite networks as it jointly captures the number of clients, patrons, and ties with peers, all of which contribute to an actor's power base, protection, and alliance-building.The second measure is closeness centrality, which indicates how close an actor is to others in a network.Closeness reflects an actor's ability to spread and receive information or resources and can also show her popularity as a coalition partner. 49The third metric is betweenness centrality, which measures how often a node lies on the shortest paths between all pairs of actors.High betweenness centrality provides an actor with a unique type of network capitalbrokerage.The ability to connect otherwise disconnected parts of elite networks, or occupy structural holes, was shown to be the primary sources of the rising power of Medici in Renaissance-era Florence and its political centralization. 50part from centrality measures, it is reasonable to consider measures of shortest paths when describing an actor's network position.It might be possible that bureaucrats positioned remotely from the governors enjoy less formal power resources than those connected directly or in just one handshake.Including a discrete measure of geodesic distance into the model would allow testing this. 51At the same time, it might be possible that the number of steps to the head executive is less important than binary differentiation between those directly and indirectly (or by no means) connected.To sum up, three centrality and two shortest path measures may serve as independent variables for further analysis.

Hypotheses
The idea of this study is to identify whether any pattern exists in how informal networks distribute formal power resources within themselves.The exploratory nature of this task predetermined the operationalization scheme in such a way as to grasp different forms and facets of formal power and network position.The multitude of chosen measures for both phenomena allows for testing at least 35 dyadic associations between them.This study focuses only on the two primary hypotheses for closer empirical examination. 52s discussed above, supervisory and substantial powers grant officials distinct forms of political capital.Supervisory roles concentrate legal control over the bureaucracy, while positions with substantial power allow direct interventions in specific policy domains.Therefore, these roles significantly differ in required skills, knowledge, experience, and level of public scrutiny.It can be expected that posts with more substantial power prioritize niche knowledge and local expertise, while informal connections with other elite members play a lesser role or may even be undesirable.In contrast, supervisory positions, being at the top of the formal hierarchy, often serve as veto players as all crucial decisions and communication flow through their desks.With their superior status and higher protection from direct criticism, these posts are potentially more suitable for appointing personal clientele and rewarding key brokers.Whether these strategies constitute a persistent pattern can be assessed through the two hypotheses: H1: The greater the supervisory power of the office, the more centrally positioned its holder is within the informal network.H2: The greater the substantial power of the office, the less centrally positioned its holder is within the informal network.

Method
Both measures of formal power chosen as dependent variables share several features that prevent using Ordinary Least Squares tests to model them.The scores cannot take negative values and, in substance, represent count data.Moreover, all scores are distributed in a non-normal manner with the presence of overdispersion, which makes a negative binomial regression (NB) preferable over a Poisson model. 53Zero inflation might be a concern since many actors often lack formal powers of a particular type.The number of zero scores across periods and types of power varies between 22% and 56%; therefore, zero-inflated negative binomial (ZINB) regression might be an acceptable alternative.However, the theory behind ZINB regression assumes that two different processes should generate zeros: one produces excess zeros, and the other generates count, or true, zeros.There are no visible reasons to assume that such a two-step process occurs in the distribution of formal powers.Therefore, using ZINB regression would probably violate its theoretical expectations.At the same time, the resulting NB models fit zeros pretty well, underfitting just one or two of them.In light of these considerations, NB regression appears as the most appropriate statistical test.
It is also natural to expect that measures of centrality and shortest paths could lead to multicollinearity.To address the issue, further analysis focuses on betweenness as the primary independent predictor and uses degree and closeness centralities as robustness checks.Betweenness captures the unique network capital of brokerage, which has shown significant effects in previous elite network studies.The logic of elite formation in Russia also predetermines the importance of brokerage.The subnational elites are usually structured around several communities, as governors must coopt personal clientele and local elite fractions into their political teams.Additionally, since the continuous and dichotomous measures of geodesic distances can still be correlated with betweenness and among themselves, I will assess the remaining predictors' variance inflation factor (VIF) and reduce models stepwise by excluding network measures that demonstrate VIF above the threshold of 3.3. 54

Results and discussion
The results reveal that the networks' shape changes with each new governor (Figures 1-3).The 2004-2005 network exhibits a cohesive and relatively centralized pattern, revolving around two key figures: the governor and the prime minister.They substantially surpass others in the number of connections while sharing many of those.The 2012-2012 network is dispersed, with many structural holes and actors outside the largest component.The 2019-2020 network has three distinct communities with several intermediaries between them: the governor's clientele from another region and two groups of local elites.This changing pattern reveals that the centralization process, initiated under Vladimir Putin, not only limited the formal powers of subnational elites but also disrupted the cohesive personalized bureaucracies, a cornerstone of subnational political machines.In Sverdlovsk's case, no first network member had retained her portfolio by 2011-2012.Unable to build his network of comparable cohesion, the appointed governor got embroiled in a cascade of elite conflicts and resigned less than three years after assuming office.Soothing the intra-regional elite rivalry became the primary task of his successor.By keeping part of the previous network, providing positions for competing pyramids (such as the clientele of the mayor of Yekaterinburg), and even co-opting some of the first governor's clients, the new appointee generally succeeded in this task.
How are formal powers distributed in such different informal networks?Table 1 represents the results of NB regression models for all periods under consideration separately.The generalized linear models have a link function that complicates a plain interpretation of its coefficients.For ease of interpretation, the regression results table reports Incident Rate Ratios (e β ), which indicates how many times the value of the dependent variable changes with a one-unit increase of an independent variable.The regression models suggest that the distribution of supervisory power follows a discernable pattern.A governor and her clientele enjoy a substantial and statistically significant advantage in this type of power across all periods under consideration.The governor's inner circle (variable "connected_to_leader") exceeds the rest of the networks in their power to monitor the bureaucracy by 6.4 times  in 2004-2005, 17.6 times in 2011-2012, and 7.1 times in 2019-2020.Moreover, it is visually evident (Figures 1-3) that all officials with any significant authority over other agencies are the governors and their direct clients.Notably, the continuous measure of the shortest path was the first choice of elimination when checking for multicollinearity.The simple differentiation between those included in the personal clientele of the governor and those further than one step away explains the variation in all types of formal power more parsimoniously than an exact number of such steps.
Contrary to the supervisory powers, the substantial ones are distributed within the networks relatively evenly.Neither a high betweenness nor a direct connection with a governor impacts the volume of substantial powers, which is valid for all periods under consideration.A closer look at these positions reveals two insightful details.First is a special place for two portfoliosa finance minister and a governor.These positions combine a high share of both supervisory and substantial powers.Unsurprisingly, all finance ministers were the governors' direct clients and worked with them well before entering the regional government.The other positions with the highest substantial powers lack any considerable supervisory authority.The same portfolios appear in this category across all periodsministers of health, education, social care, state property, and housing and utilities.Interestingly, only two out of fifteen such ministers were the governors' direct clients.Outsider governors are especially prone to abstain from appointing their inner circle to such sectorial positions and usually gladly hand them over to local elites.This strategy seems to be perfectly rational for outsiders.Healthcare, education, social care, and housing, more than any other policy area, affect the everyday life of an entire region and provide jobs for a vast number of public sector employees.These portfolios thus require deep local knowledge and skills, which limits options for reshuffling.Almost without exceptions, the appointed governors selected local specialists with prior experience in these sectors to fill the respective positions.What seems counter-intuitive is that a strong machine governor also abstained from setting his clients to these posts, which reveals the second feature of sectorial portfolios.They are the most resourceful regarding budget funds and the number of employees.Since citizens interact with these areas regularly, the officials in charge face significant political responsibility and are prone to public criticism, particularly during times of crisis like the COVID-19 pandemic and elections, regardless of their competitiveness.All governors distanced themselves from this direct responsibility, regardless of the political and institutional context they operated in.
The monopolization of supervisory power by a clientele, a relatively even dispersion of substantial power, and the avoidance of personal association with holders of the top sectorial portfolios have consistently been strategies employed by all governors.These results are robust to including all actors in a single model, controlling for periods, and excluding governors from the models.Using other centrality measures as independent predictors does not substantially change the results.However, in the 2004-2005 network, the association between supervisory power and a direct connection with the governor falls below the conventional significance level (Online Appendix C).This can be attributed to the size of the first governor's clientele, which accounted for half of the network, while the number of available supervisory positions is inherently limited.Therefore, being part of the clientele was necessary but insufficient to acquire top supervisory positions.In contrast, outsider governors had fewer clients, and nearly all of those were appointed to these positions.As an alternative to NB regression, all possible dyadic associations between formal powers and informal networks' measures were also tested by Kendall's rank correlation, which generally confirms the findings (Online Appendix B).
Considering the consistency of the uncovered strategies, let us briefly revisit the changing political context to illustrate the diverse conditions and incentives governors had to consider when building their power coalitions.
Sverdlovsk Oblast did not differ in its institutional and political dynamic from the general trends outlined in the second section of this article.The first governor assumed office through fairly competitive elections, securing his victory only in the second round.The next governor was appointed by the president with the consent of the regional legislature following the elimination of direct gubernatorial elections.The third one was named acting governor by the president before the election and won it in a fairly sterile electoral environment, arose from the requirement for independent candidates to collect signatures of municipal deputies in order to appear on the ballot.The change of rules for assuming office effectively shifted the selectorate from the region's population to the presidential administration.Such a sharp change in accountability could not but change the priorities of the governors towards satisfying  federal needs and instructions in the first place, which, in turn, could significantly impact how governors organize their power networks.Other aspects of centreregions relations also transformed within the considered period, including the autonomy of the governor to exercise his appointment power.In 2004-2005, there was practically no interference from the federal centre in appointing regional officials.Later, however, the governors began to lose this autonomy.In 2011, the governor of Sverdlovsk Oblast could no longer name a minister of health without the consent of the federal health ministry.In 2019, the pool of portfolios that required federal approval also included the finance and education ministers.The increasing federal interference in regional appointments has been a widespread trend, not limited to Sverdlovsk Oblast.The pool of positions subject to federal vetting varied across regions based on ministerial charters.The practice has been fully institutionalized with the enactment of the Federal Law "On the General Principles of Organization of Public Authority in the Subjects of the Russian Federation" in 2021.As a result, candidates for key positions, including health, education, finance, and communal infrastructure ministers, now require approval from federal ministries in all regions.
While losing autonomy from the federal centre, the governor of Sverdlovsk Oblast has substantially upgraded their institutional standing within the region.In 2004 and 2011, their powers in cabinet formation were mediated by both the assembly and the premier.The former had the right to confirm or decline the premier's candidacy, while the latter proposed the candidates for ministerial portfolios to the governor.By 2019, the premier post was eliminated, and the governor became entitled to discretionally name all ministers except three that remained subject to the assembly's investiture.The same is valid for cabinet dismissal.While the governor could not dismiss the ministers that the premier proposed without their consent in the first two periods, they could do that at will in the third.In 2004, the regional legislature had unconditional power of censure against the governor, premier, and cabinet.Since 2011, the assembly could remove only the premier and three ministers.Although a no-confidence vote against the governor remained in the region's statute, the final decision on the governor's dismissal was passed to the president.On the contrary, the governor received a right to dismiss the assembly in 2007 due to changes in federal legislation to strengthen governors' standing against regional legislatures across the country.
The institutional landscape has undergone substantial transformation within the analysed timeframe, as did the political regime in Russia in general.All major indices measuring the level of democracy testify to an authoritarian reversal Russia's political regime underwent from 2004 to 2019.The political role of the governors echoed this process.If the first governor succeeded in fighting with the federal centre to upgrade the region's formal status and formed the powerful regional party, the subsequent governors were members of the Kremlin party "United Russia" and their institutional reforms aimed at eliminating any provisions that run counter to federal legislation.
Degree of subnational autonomy, rules for assuming a gubernatorial office, level of electoral contestation, the relative strength of legislature, and the very shape of informal networks varied across time.However, the pattern of building power networks remained the same.While delegating sectorial portfolios to other parts of local elites, the governors concentrate powers to oversight, redistribute budgetary funds, and appoint and dismiss other regional officials in the hands of their clientele.This tendency signifies the importance of controlling functions that the core of informal networks in personalized bureaucracies perform.
Systematic studies of how administrative apparatuses form in other Russian regions are scarce, making it difficult to compare the uncovered patterns with reliable quantitative aggregations.However, qualitative descriptions of how governors assemble their administrative teams suggest that these strategies are a rule rather than an exception.For example, when Mikhail Reshetnikov became governor of the neighbouring Perm Krai, he filled most supervisory positions with former colleagues from the Moscow city government, the federal Ministry of Economic Development, and even fellow students from his faculty.Meanwhile, his inner circle occupied only two out of 27 sectorial portfolios.It is common for appointed governor's inner circle to follow them from one position to another, forming tight-knit groups aptly called "wandering" elite cliques. 55ne notable example is Oleg Kozhemiako, whose team accompanied him during his governorship in four regions.Of course, exceptions and variations exist since team formation is a matter of political strategy constrained by local elite group interests, diverse career backgrounds, and the demands of the federal centre.For instance, Boris Dubrovsky, the former governor of Chelyabinsk Oblast, spent an entire career in Magnitogorsk Iron and Steel Works and did not forge a diverse clientele to fill even primary supervisory positions.
As mentioned earlier, sectorial portfolios, particularly in the healthcare sector, often face intense public criticism and are filled by local specialists who lack personal connections with the governor.This assumption was put to the test during the COVID-19 pandemic.When president Putin granted regions more policy freedom in combating the virus and warned governors of career consequences for poor performance, the regional healthcare ministers became the final chain of blame-shifting.From March 2020 to March 2022, healthcare ministers were fired at least once in 54 regions. 56In 13 regions, there were two reshuffles, and four regions witnessed three ministerial dismissals.The biographical analysis reveals that only 14 out of 160 ministers (8.75%) were directly connected to the governors, and only three of them experienced downward career mobility when leaving their positions.Others either retained their jobs or even rose through the ranks.
Although this research does not attempt to establish robust causality (formal powers may depend not only on a position in an informal network but on the skills, education, or experience in specific policy areas), it is possible to theorize on the direction of a possible causal link.Do the formal powers of a portfolio define whether a governor appoints a loyal client, or does the proximity to the governor allow a client to increase their formal power?The causal arrow is most likely double-sided.Such posts as prime minister, finance minister, or the head of the governor's administration are a priori strong in their formal status.Assuring that a rival patron-client pyramid does not control these posts makes appointing personal clientele the first-choice strategy for a governor.At the same time, once the governor's clients assume office, they often seek to increase their formal status.For example, a deputy governor in the 2019-2020 network gained the right to distribute the reserve fund, and the finance minister received a supplementary status of the deputy governor with attendant controlling functions.When both clients resigned, a governor stripped both positions of these privileges.These examples deter from assuming a one-way causality, although they provide additional evidence of interaction between informal networks and formal hierarchies.

Concluding remarks
Patronage is a ubiquitous phenomenon found across different regime types and institutional settings.As a recent review demonstrates, patronage appointments may have different nature of trust that underlies them (partisan vs. non-partisan) and types of skills that are expected from appointees (professional vs. political). 57Personalized bureaucracies, including the Russian one, are substantially autonomous from the parties' interests, making personal considerations the basis for patronage appointments.A poorly institutionalized party system combined with the lack of institutional trust and syncretism of political and economic interests makes patron-client networks the main actors in political contestation.Once succeeded, they infiltrate the bureaucracy, and once defeatedthey abandon it.Studying elite networks appears as an adequate approach to uncovering decision-making, elite advancement, and regime dynamic in personalized political systems.Network analysts advocate for moving beyond the dyadic patron-client relationships and suggest that the network position may reflect an actor's power much better.This proposition, however, rests implicitly on the assumption that being well-connected somehow provides an actor with certain power resources.This study aimed to translate this implicit assumption into an explicit hypothesis.In the context of Russia's subnational bureaucracy, it turned out that a particular combination of formal power resources indeed follows a relatively stable pattern of distribution.Legal instruments of monitoring the bureaucracy are wholly concentrated in the centre of an informal network, i.e. in the hands of the governor's clientele, while other formal powers are dispersed within networks relatively evenly.In the Sverdlovsk case, the controlling function of the network's core persists throughout its entire post-Soviet history, despite changing institutional and political background.How universal the uncovered pattern is an agenda for further research.Despite the acknowledged contextuality of patronage, informal networks infiltrate formal hierarchies in many political settings.Whether the regularities in their combination appear in other personalized bureaucracies is thus of high interest.

Figure 1 .
Figure 1.Formal power in informal network 2004-2005.Note: Rossel' is a governor, and Vorob'ev is a prime minister.For figures 1-3: implemented in Gephi with layout algorithm Force Atlas 2; The size of each node represents the number of its еdges while saturation corresponds to the volume of supervisory power.

Figure 2 .
Figure 2. Formal power in informal network 2011-2012.Note: Misharin is a governor, and Gredin is a prime minister.

Figure 3 .
Figure 3. Formal power in informal network 2019-2020.Note: Kuivashev is a governor, and Orlov is the first vice governor.The office of prime minister was abolished in 2016.

Table 1 .
Formal Power in Informal Networks: Negative Binomial Regression Model Results.