“We don’t abandon our own people”: public rhetoric of Russia’s governors during the full-scale invasion of Ukraine

ABSTRACT

The 2022 invasion of Ukraine tests the resilience of the authoritarian regime in Russia. Most of the elites showed their commitment to the autocrat by publicly supporting the “special military operation”. However, the intensity of their support has varied across the ranks. By examining Russian governors, I propose that elites vary their communicative responses to the war strategically, depending on their personal considerations and the structural characteristics of their regions. I apply text-as-data methods to analyze gubernatorial posts on the social network VKontakte in 2022–2023. I find that governors up for election, outsiders, with weaker political standing, and from poorer regions communicate more support measures for combatants and their families, and humanitarian aid to the annexed territories, when discussing the war in Ukraine. Therefore, by co-opting war-affected groups and emphasizing performance legitimacy, governors minimize the threat of anti-regime mobilization without resorting to large-scale repression and contribute to authoritarian resilience.

KEYWORDS:

• Authoritarian resilience
• Russian invasion of Ukraine
• legitimation
• co-optation
• public communication
• social media

Introduction

Russia’s full-scale invasion of Ukraine, launched by President Vladimir Putin on 24 February 2022, without consultation with a wide range of elites (Gabuev 2022; Troianovski 2022), marked a new stage in the autocratization of the political regime in Russia. Before the full-scale war in Ukraine, the Kremlin had systematically limited the opposition and kept citizens politically demobilized to avoid discontent and preserve the stability of the regime (Alyukov 2022). However, during the war, the Kremlin had to mobilize some of the population to participate in hostilities while continuing to prevent threats to the regime. In this regard, repression increased under the pretext of protecting the Russian army from being discredited. At the same time, public officials continued to use non-violent forms of regime maintenance, by co-opting the population and legitimizing an authoritarian rule.

This study applies the principal – agent model to explain the behavior of an important part of the lower-tier political elites in Russia, regional governors, during the full-scale invasion of Ukraine. Acting as agents, governors have been assigned war-specific tasks from their principal, President Vladimir Putin and his administration. Among other things, regional heads started to distribute material benefits to the participants of the war and their families and provide humanitarian aid to the annexed territories of Ukraine (Busygina and Filippov 2024, 200). The fulfilment of war-specific tasks has been also reflected in the public communications of governors. I argue that by publicly supporting groups directly affected by military actions, governors have been co-opting the masses and maintaining the performance legitimacy of the regime, thus contributing to authoritarian resilience. At the same time, personal career considerations and regional structural characteristics have influenced the extent of governors’ efforts to co-opt the population and bolster regime legitimacy through publicizing support for soldiers, their relatives, and the annexed territories.

To test the theoretical expectations, I collected an original dataset of 64,621 governor posts on the most popular social network in Russia, Vkontakte (VK), from 24 February 2022 to 31 October 2023 and conducted text-as-data analysis using structural topic modeling. According to the empirical results, governors up for election, with weaker political standing, an outsider status, from less developed regions, and with higher military mobilization rates communicate significantly more performance of war-related tasks in their public communication, focusing on support measures for the war-affected groups. By revealing the strategic logic of elite rhetoric during the war, this study contributes to the literature on the resilience of political regimes (Dimitrov 2013; Holloway and Manwaring 2023; Merkel and Lührmann 2021; Ngoun 2022; Sinkkonen 2021), authoritarian stability (Gerschewski 2013; Magaloni 2008; Svolik 2012), and the public communication of autocratic leaders and elites (Baturo and Mikhaylov 2014; Boussalis, Dukalskis, and Gerschewski 2022; Maerz and Schneider 2020).

Public communication of elites and authoritarian resilience

The public rhetoric of elites is an essential element of their political behavior. It reveals important information about politicians, such as ideological orientations and latent personal traits (Rheault and Cochrane 2020), especially in opaque authoritarian contexts (Baturo and Mikhaylov 2013). The literature has shown how elite discourse is strategically employed to support democracy (Maerz and Schneider 2020), tackle policy positions of state leaders (Baturo and Mikhaylov 2014), and communicate authoritarian elite cohesion (Schedler and Hoffmann 2016). Public communication of autocrats is strategically tailored to control the population (Boussalis, Dukalskis, and Gerschewski 2022; Wedeen 2015) and legitimize the authoritarian regime (Dukalskis and Gerschewski 2017; Dukalskis and Patane 2019; Maerz 2019). Moreover, structural characteristics of political regimes such as the degree of autocratization and personalization can be traced through the discourse of politicians (Liu 2022; Maerz and Schneider 2021).

In the same manner as propaganda, elite public communication provides informational signals to the population and the autocrat. For example, state officials demonstrate public allegiance to the national leader (Baturo, Khokhlov, and Tolstrup 2023; Shih 2008) and reassure the population amidst crises (Windsor, Dowell, and Graesser 2014). Elite rhetoric emphasizes the performance legitimacy of the regime and spurs patriotism. At the same time, politicians engage in the intimidation of regime opponents and indoctrination of citizens, reproducing and reinforcing propaganda frames (Baturo and Tolstrup 2023). Finally, elites protect the popularity of the autocrat by publicly accepting the blame for unpopular decisions (Busygina and Klimovich 2022). Overall, the three pillars of authoritarian regime stability – repression, co-optation, and legitimation (Gerschewski 2013) – are reflected in the way political elites speak in public.

This paper focuses on the role of elite public communication in maintaining authoritarian resilience, defined as the capacity of the regime to immediately cope with challenges, adapt to changing circumstances, and persist over time (see Paustyan and Busygina forthcoming in this issue), in Russia during the first 20 months of the full-scale invasion of Ukraine. The concept of authoritarian resilience differs from authoritarian stability. While the latter refers to the ability of a regime to maintain a strong hold on power, authoritarian resilience is related to the overall robustness and adaptability of the autocracy to challenges. I focus on the discursive aspect of authoritarian resilience. In particular, I proceed from the fact that public officials are key actors in determining the perceptions of the crises and constructing a “myth of resilience” (Keck and Sakdapolrak 2013, 12).

As mentioned by Elina Sinkkonen (2021, 1178), “Authoritarian resilience depends on controlling society and maintaining a necessary level of elite coherence.” Repression helps autocrats to address political threats over a short-term period, but they are costly in the long run (Beetham 1991, 28). That is why modern autocrats combine repression with non-violent strategies to stay in power and act not only repressively, but also “responsively” to popular expectations (Baberowski and Martin 2022, 21). Specifically, authoritarian regimes supplement repression with co-optation and legitimation strategies. Co-optation is defined as “all measures by ruling elites with the aim of strengthening and/or widening the societal base of a regime by giving relevant individuals or societal groups a stake in the status quo” (Josua 2016, 35). Authoritarian rulers co-opt elites and masses by providing them with perks such as access to formal institutions, financial payments, and social status (Frantz and Kendall-Taylor 2014, 334–335). A straightforward way to gain popular support and silence criticism is through “co-optation via transfers,” when material resources are allocated to strategically important groups of people (Svolik 2012, 164).

Co-optation is connected to another non-repressive autocratic strategy – legitimation, or a process of elites to attain “the citizens’ acceptance of the incumbents’ claim to rule” (Josua 2016, 37). Even in personalist authoritarian regimes, the people remain the main source of legitimacy (Goode 2016, 420). While legitimation is based on regime discourses, co-optation supplements it with concrete incentives to guarantee the acceptance of authoritarian rule, either in a form of diffuse support, linked to ideational factors, or specific support, related to the fulfilment of particular popular demands (Easton 1975). Political leaders employ patriotism and nationalism to boost their legitimacy through diffuse support (Hale 2018; Laruelle 2019). In turn, obtaining specific support lies at the intersection of co-optation and legitimation. On the one hand, the regime distributes material benefits and creates concrete incentives to support it (Frye, John Reuter, and Szakonyi 2014). On the other, autocrats engage in performance legitimation, by publicly emphasizing the achievements of the regime and fulfilment of popular demands in written texts and speeches (Baturo and Tolstrup 2023). In this regard, electoral authoritarian regimes have an advantage over closed autocracies in understanding the preferences of citizens. By allowing multi-party elections, autocrats reduce the information deficit about the needs and mood of the society, observing the level of popular dissatisfaction and support for the programs of co-opted opposition parties (Cassani 2017, 353–354). Overall, regime delivery and its effects on diffuse support of autocrats, gained through the co-optation of masses and performance legitimation, influence both the stability of the autocracy and its resilience, by preventing large-scale anti-regime mobilization (Frantz and Kendall-Taylor 2014, 334–335).

The principal – agent model and public rhetoric of Russia’s governors during the full-scale invasion of Ukraine

All authoritarian rulers rely on elites to maintain their power (Tullock 1987; Wintrobe 1998). The personalist authoritarian regime in Russia under Putin is no exception (Golosov 2023). An autocrat in Russia has built his ruling coalition through a mix of repression, co-optation, and promotion of key elite groups, including oligarchs, law enforcement agencies (siloviki), and top federal bureaucracy (Snegovaya and Petrov 2022; Yakovlev 2021). By strengthening the “vertical of power” (Gel’man and Ryzhenkov 2011), making all its elements accountable to the president and personally interested in maintaining him in power (Smyth 2014), the Kremlin solved the problem of the authoritarian power-sharing (Svolik 2012).

In Russia’s political system, regional governors constitute an important part of the lower-tier elites (Waller 2023, 3). In the standard principal – agent model of federal relations, governors act as agents responsible to two principals: the central government and the regional population (Starodubtsev 2018, 146). The national principal, unable to deal alone with the complexity of governing a large and diverse country, delegates some of its authority and specific tasks, such as social policy, to regional agents. The agent is expected to follow the principal’s interests and ensure the fulfilment of her task. Since the agents possess more information about their areas of responsibility and might have different preferences compared to the interests of the principal, the decentralized system faces the challenge of agency loss, when the agents do not deliver the results. To gain credible commitment from agents, the principal has to organize a control system and provide incentives to regional administrators, which would prevent their opportunistic behavior.

Before Vladimir Putin consolidated his power, Russia’s governors had possessed significant independence in decision-making, being responsive to multiple principals, including the president, regional elites, oligarchs, and citizens (Sharafutdinova 2010, 676). Often, the preferences of principals had diverged from each other and created a “common agency” problem (Bernheim and Michael 1986), when an agent could strategically select which principal’s interests to pursue in the first place. This led to significant gubernatorial discretion and autonomy, threatening the territorial integrity of the Russian Federation (Treisman 1997) and Putin’s re-election prospects in 2004, as they depended on the mobilization of votes by regional political machines (Hale 2003). To eliminate threats from governors, the Kremlin implemented institutional reforms aimed at the centralization of power. With the growing personalization of the authoritarian regime, Russia’s governors received strong incentives to maintain the high popularity of Putin (Busygina and Filippov 2021). To achieve this, regional heads mobilize votes for the autocrat and the ruling party, maintain socio-economic development and political stability in their regions, and promote pro-regime discourses (Libman and Rochlitz 2019; Reuter and Robertson 2012; Sharafutdinova 2016).

Since governors are the main representatives of state power in the regions, they maintain the appearance of accountability to the population in the state media and on social networks (Renz and Sullivan 2013). Many governors started personal blogs during the 2008–2012 presidency of Dmitry Medvedev to demonstrate allegiance to the modernization agenda (Toepfl 2012). Ahead of the 2018 presidential elections, the Kremlin launched a program to monitor the response of regional authorities to citizen complaints on social networks. In December 2022, federal, regional, and local authorities were obliged to create official pages on VKontakte and Odnoklassniki and publish up-to-date information about their activities. By strengthening control of official online pages, the Kremlin has produced “the vertical of power on the Internet.”¹ Therefore, social networks have become an important tool in the menu of authoritarian strategies in Russia. Public officials at all levels of government are encouraged to cover their activities online, reinforcing state discourses and legitimizing the Russian political regime.

The 2022 full-scale invasion of Ukraine has been a profound shock for elites in Russia. In the same manner as during the COVID-19 pandemic, President Putin shifted responsibility for dealing with domestic repercussions of the war to governors. They have been assigned with distributing support measures to the soldiers and their families, organizing the military economy, mobilizing new conscripts, and “shaping public opinion in favour of the ‘people’s war’ with the West” (Busygina and Filippov 2024, 200). Governors had to adapt to the new reality and demonstrate their commitment to the war cause by engaging in public displays of loyalty. At the beginning of the full-scale war in Ukraine, governors publicly supported the “special military operation” (SMO). However, while some governors made short public statements on official regional websites, others have repeatedly declared their support for war on personal social media pages.² When the prospects for a short war collapsed and President Putin announced the partial mobilization of military reservists on 21 September 2022,³ governors were instructed to avoid direct discussion of military actions and instead emphasize support for the combatants and their families (Cancian 2022).

By delegating to governors the task of communicating war-time regime performance, the Kremlin aimed to co-opt the strategically important groups of the population and legitimize its rule. As shown in the empirical section below, when governors post on social media about the Russo-Ukrainian war, they indeed tend to discuss material aid and support measures for soldiers, their families, and the occupied regions of Ukraine. In this way, regional heads have engaged in performance legitimization of the regime and prevented the threat of mass anti-regime mobilization without resorting to mass repression,⁴ contributing to the resilience of personal autocracy in Russia. However, the extent to which governors followed the principal’s instructions has varied across the ranks and over time.

I argue that the degree to which individual governors engaged in legitimization and co-optation strategies by publicizing aid for soldiers and occupied territories on social media has been driven by their political considerations and regional structural characteristics. First and foremost, governors facing their de jure elections must guarantee the personal support of the autocrat, which is required de facto to retain office (Blakkisrud 2015). This is achieved largely by maintaining political stability in the region. The literature has demonstrated how authoritarian regimes sustain the welfare state by providing social benefits to broad groups of the population in exchange for their acceptance of political rule (Logvinenko 2020). During crises, proper policy responses lead to electoral returns for the incumbent authorities (Bechtel and Hainmueller 2011). Moreover, electoral authoritarian regimes engage in vote-buying ahead of elections to maximize their vote shares and turnout and minimize the need for ex-post electoral manipulations (Frye, John Reuter, and Szakonyi 2019; Luo and Rozenas 2018; Nichter 2008; Stokes 2005). When a governor’s term approaches an end, she has incentives to communicate more about themes that both appeal to the population and demonstrate control over the situation to the autocrat. During the ongoing war, one way to please the population and the autocrat is through greater communication of support measures to the combatants and their relatives and humanitarian aid to the annexed territories.⁵ In this way, governors co-opt strategically important groups of people and boost specific support through performance legitimation (Easton 1975, 437). Consequently, I formulate the first hypothesis:

H1:

Governors are more likely to post about war-related topics when they approach their elections.

In addition to elections, which make governors temporarily vulnerable, regional heads can vary their rhetorical strategies depending on their general personal political standing (Baturo, Khokhlov, and Tolstrup 2023). Politically stronger governors have robust patronage ties and support from top elites, are not involved in conflicts with powerful elite groups, govern more attractive and resource-rich regions, and maintain a high level of regional socio-economic and political stability (Baturo and Elkink 2016; Reuter and Turovsky 2022). As a result, they are less subject to elite repression. On the contrary, weaker governors are less competent and rooted in the patronage networks, rely more on the resources from the federal center and therefore have limited leeway vis-à-vis the autocrat. An agent’s strength is likely to determine her effort in performing crisis-specific tasks set by the principal (Hermalin 1992). On this basis, I hypothesize that weaker governors will have more incentives to engage in the co-optation of war-affected groups and performance legitimation of the regime. By putting more effort into maintaining regime stability during the war, weaker regional heads display excessive loyalism and send a signal to the autocrat that they are useful in office (Libman and Heckenthaler 2023):

H2:

Politically “weaker” governors are more likely to post about war-related topics.

Another prominent personal characteristic of regional heads in Russia is their status as an insider or an outsider to their region (Buckley et al. 2014; Libman and Rochlitz 2019). The outsider is a governor with no experience of living, studying, or working in the region before taking office; 56%, or 49 out of 88 regional heads in my sample, are outsiders. Replacement of local governors with outsiders intensified after 2016 when Sergey Kiriyenko assumed the post of first deputy chief of staff of the Presidential Administration (PA), responsible for regional politics. Ahead of the 2018 presidential elections, the PA initiated the replacement of 47 governors (Ivanov and Petrov 2021, 162), promoting members of the presidential personnel reserve and graduates of the School of Governors training program to the regions (Pertsev 2023). New regional heads have been considered as technocratic managers rather than political figures. Governor-outsiders tend to be more independent from the interests of local elites (Schultz and Libman 2015) and treat their position as another step on the career ladder (Kynev 2019). Therefore, I hypothesize that governors with no previous connections to their regions will engage more in the co-optation of the masses and legitimization of the regime, portraying themselves as effective managers, suitable for promotion, in the eyes of the autocrat and his administration:

H3:

Governors-outsiders are more likely to post about war-related topics.

The structural characteristics of regions are likely to provide constraints on the ability of governors to engage in the performance of the autocrat’s tasks. Russian regions differ significantly in terms of socio-economic development and resource endowment (Libman and Rochlitz 2019). At the same time, the performance legitimation of the regime relies on the factual delivery of benefits to the targeted groups (Cassani 2017; Logvinenko 2020). If regional heads communicate about performance significantly more than they deliver it, they may lose credibility and popular support over time (Knutsen and Rasmussen 2018). Therefore, Russia’s governors are limited in communicating support measures for the war-affected population by the resources available for actual distribution. Usually, heads of poor regions depend on discretionary transfers from the federal budget (Marques II, Nazrullaeva, and Yakovlev 2016; Sharafutdinova and Turovsky 2017) and they tend to express greater loyalty to the autocrat (Baturo, Khokhlov, and Tolstrup 2023). Therefore, they have incentives to retain the flow of transfers by engaging in the performance of important tasks of the autocrat. Conversely, the heads of wealthy regions are less financially dependent on federal transfers and thus less incentivized to display wartime performance on social media, since they can boost popular support by actual delivery of benefits. Therefore, there are theoretical reasons to hypothesize that the governors from less developed and more transfer-dependent regions will communicate more about support for strategically important war-affected groups of the population:

H4:

Governors of less developed regions and regions with more federal transfers are more likely to post about war-related topics.

Finally, contextual factors can alter the communications strategies of lower-tier political elites. Specifically, Russia’s governors were responsible for the partial mobilization of military recruits, announced on 21 September 2022 (Busygina and Filippov 2024, 200). However, regions varied greatly in the relative numbers of soldiers mobilized. The main impact fell on regions with poor populations and a higher share of ethnic minorities, both in terms of recruitment⁶ and death toll on the battlefields (Bessudnov 2023). The number of mobilized soldiers can be considered as a proxy for the size of the regional war-affected population. If more people from the regions are affected by the war, the governors will have stronger incentives to co-opt them with the support measures and publicize it in their public communication. Therefore, I hypothesize that the need to co-opt the war-affected groups and engage in performance legitimation will be greater in regions with higher rates of mobilized soldiers:

H5:

Governors of regions with more mobilized soldiers are more likely to post about war-related topics.

Data and methods

I focus on the public communication of Russia’s governors on the social media platform VKontakte (VK). Among 100 governors in Russia in 2019–2023, 97 have had either a personal page or a public group on VK. Therefore, VK posts provide a unique source of data on the public rhetoric of regional leaders in Russia. I web-scraped 64,621 posts of 88 of Russia’s governors from the start of the full-scale war in Ukraine on 24 February 2022 until 31 October 2023.

I employ an unsupervised text analysis method Structural Topic Modeling (STM), to estimate the topics of the texts based on co-occurrences of words (Roberts, Stewart, and Tingley 2019). The topics are defined as “distributions over a vocabulary of words that represent semantically interpretable ‘themes’” (Roberts et al. 2014, 1066). Instead of defining the themes a priori, as in the case of the supervised models, the STM infers the topics from the texts. Compared to alternative unsupervised text-as-data methods, such as the Latent Dirichlet Allocation (Blei et al. 2002), the STM includes metadata in calculations of topical content and proportions. In addition, it allows us to explicitly estimate correlations between topics. Overall, the STM provides more accurate estimates of topics and facilitates hypothesis testing (Lucas et al. 2015, 262). This method has been widely employed in studies of policy frames (Fabrizio, Shipan, and Wüest 2016), elite rhetoric (Shayegh et al. 2023), and ideological discourses (Karell and Freedman 2019).

I ran a set of the STM models from 10 to 20 topics and the set of covariates to estimate topic prevalence and test the hypotheses. Elections is a dummy variable, equal to 1 in case governors faced elections in 2022–2023, compiled from the Central Electoral Commission. To operationalize governors’ political standing, I used the Rating of Political Stability of Regional Heads by the Minchenko Consulting Group (Gov rating), one of the most reputable sources of data on Russia’s governors (Ivanov and Petrov 2021, 169–70). I used the integral scores of the governors’ political stability from four waves of the Minchenko Rating, corresponding to the time coverage of VK posts. The governor’s status as an outsider in the region (Outsider) was compiled from the dataset on governors’ biographies by the International Center for the Study of Institutions and Development (ICSID) and complemented by the author’s coding based on the official regional webpages.

To account for the level of regional socio-economic development, I used the logarithm of the gross regional product per capita (GRP pc) in 2021, before the economy transformed into a military one, from the Federal State Statistics Service. In turn, regional dependence on the federal center was measured as the logarithm of all federal transfers (subsidies, subventions, and other transfers) to the regional budgets per capita, in 2022 (Transfers pc). GRP per capita and federal transfers per capita are common proxies of regional wealth and resource endowment (Turovsky and Gaivoronsky 2017).

The regional share of mobilized soldiers (Mobilized soldiers %) was obtained from the Mediazona dataset. After an announcement of partial mobilization on 21 September 2021, enlisted soldiers were granted an opportunity to register their marriages immediately instead of applying one month in advance. Many conscripts used this opportunity to enable their new wives to receive social dividends as relatives of mobilized soldiers. A corresponding surge in weddings compared to the usual number made it possible to assess the scale of mobilization. Since the Mediazona data were calculated for the main wave of mobilization in September–October 2021, I used separate regression models Mobilized soldiers % as a predictor, limited to the sample of VK posts after 31 October 2022.

To control for other factors that might explain the prevalence of the SMO topic in governors’ posts, I included additional variables in the STM specifications. First, I used three dummy variables for governors’ professional backgrounds: federal bureaucracy (Backgr fed politics), private business (Backgr business), and law enforcement agencies (Backgr siloviki), based on the ICSID dataset and original data collection. Governors with experience in the federal bureaucracy are more technocratic, less connected to their regions, and more accountable to the autocrat (Buckley et al. 2014). As a result, such governors are expected to perform the autocrat’s tasks with greater care. In turn, governors with business backgrounds treat their political office as a rent-seeking opportunity (Gehlbach, Sonin, and Zhuravskaya 2010; Szakonyi 2020, 2021) while governors-siloviki tolerate less political openness and prefer a firm hand in managing the region (Rivera and Rivera 2019; Soldatov and Michael 2018). Governors-businessmen also have more outside career options and thus can be more indifferent to the principal’s task of highlighting regime performance during the war. At the same time, siloviki can be more naturally inclined to the discussion of war-related topics.

Secondly, I control for additional personal characteristics of governors. Gov conflict fed is a binary indicator for the conflicts between the governor and the federal elites, derived from the Minchenko Rating. Governors involved in conflict with the top elite are more likely to be threatened by siloviki (Yakovlev and Aisin 2019), so they might put more effort into getting protection from the autocrat, by performing his important wartime tasks. Furthermore, I control for the status of governors who were appointed as interims after the 2021 electoral cycle (Gov interim). Such governors are selected by the Kremlin and might have the more immediate support of the autocrat. Due to their newcomer status, they are limited in claiming credit for regional performance and therefore can engage less in war-related rhetoric. Gov pol experience is the number of years in the office of a governor and Gov age is his age. More experienced and older regional heads might be more prone to resignation and therefore try more to prolong the support of autocrats by engaging in war-related rhetoric.

Finally, I control for additional regional characteristics. Unemployment in 2021 is an alternative indicator of socio-economic development. Reg deputies UR share is the share of United Russia deputies in the regional legislature. This indicator proxies the control of governors over regional elites and overall political stability. Suppose the regional legislature is dominated by deputies from the ruling party. In that case, the governor faces fewer constraints on his behavior and can claim all the credit for regional performance during the war in public. In addition, the regression model controls for whether the region is located on the border with Ukraine (Border Ukraine), and therefore more prone to the effect of hostilities. Descriptive statistics, a detailed description of all variables and sources of data, are reported in the online Appendix.

Analysis

Figure 1 plots the identified topics and topic proportions based on the estimates of the STM model with 19 topics and a set of covariates, described in the previous section. I manually assigned topic labels based on the most frequent topic-specific words and the content of posts with the highest topic prevalence. The STM results indicate that Russia’s governors use their VK pages primarily to congratulate the population on holidays, highlight infrastructure projects such as roads and houses, and communicate regional governance activity. The next most frequent topics include industry development, relationships with the federal center, and management of the regional budget and social payments. Thus, even during the ongoing war, governors prioritize coverage of routine regional issues and portray themselves as effective managers (Sharafutdinova 2016, 382). At the same time, the theme of the “special military operation” (SMO) is covered to a lesser extent, on par with World War II (WWII).

Figure 1. Topic proportions of VK posts from Russian governors in 2022–2023.

The top words associated with the SMO topic are presented in Table 1 and capture two aspects of the war discourse: support measures for the participants of the special military operation and their families and humanitarian aid to the annexed Ukrainian territories. The former frame is associated with such top words as “family,” “wife,” “participant,” “mobilized,” and “parcel,” while the latter with top words “humanitarian,” “resident,” “LNR,” and “sponsored.” Many words, assigned to the SMO topic by the STM algorithm, are relevant for both frames: “aid,” “support,” “cargo,” “dispatch,” and “deliver.” Overall, the two frames capture the rhetoric related to the material co-optation of the war-affected groups and the performance legitimation of the authoritarian regime, in line with the theoretical expectations.

Table 1. Top words associated with the SMO topic.

Top 20 highest probability words	Top 20 frequency and exclusivity words
Aid (помощ-), family (сем-), fighter (бойц-), SMO (сво-), military (воен-), support (поддержк-), serviceman (военнослужа-), mobilized (мобилизова-), region (област-), help (помога-), participant (участник-), need (необходим-), humanitarian (гуманитарн-), guys (ребят-), deliver (переда-), special (специальн-), resident (жител-), house (дом-), zone (зон-), all (все-)	Mobilized (мобилизова-), parcel (посылк-), humanitarian (гуманитарн-), mobilization (мобилизац-), SMO (сво-), Orel (орловск-), goods (вещ-), cargo (груз-), serviceman (военнослужа-), volunteer (волонтер-), send (отправ-), fighter (бойц-), sponsored (подшефн-), wife (жен-), guys (парн-), dispatch (отправк-), volunteer (добровольц-), letter (письм-), enlistment office (военкомат-), LNR (лнр-)

On average, 5.6% of text per post covers the SMO topic. Nevertheless, there is a significant variation across time. Russia’s governors significantly increased their coverage of the support measures to the war-affected groups after President Putin announced the partial mobilization on 21 September 2022,⁷ as illustrated by Figure 2. Since governors were assigned the task of mobilizing soldiers, they strategically adjusted their rhetoric, emphasizing benefits to the recruits. Over time, when the main wave of mobilization subsided, the proportion of the SMO topic demonstrated a downward trend, although it remained higher than during the early period of the military conflict.

Figure 2. Percentage of the SMO topic over time. Note: the line displays smoothed conditional means, and the shaded area represents the standard errors.

An example of a governor’s post with high prevalence of the SMO topic comes from the head of the Ul’yanovsk region, Alexey Russkikh, on 27 July 2023:

A batch of humanitarian aid was delivered to the SMO zone from the Ul’yanovsk region: four … trucks with equipment, food, medicine … The guys say thank you to all the residents of the region. Any help from home is important for them.

Humanitarian aid to the population of annexed Ukrainian territories is covered in the post by the head of Perm’ Krai, Dmitry Makhonin, on 10 March 2022:

[We] sent the first humanitarian aid to the residents of the Donetsk and Lugansk republics … I am grateful to the caring residents of the Prikamye and the organizations that responded to the call for help. Things and food arrive at collection points every day, so I am sure that this humanitarian convoy will not be the last …

The STM calculated topic proportions for each post in the corpus. I used the percentage of the SMO topic per post as a dependent variable in the least squares dummy variable (LSDV) regression models with regional and governor-level predictors, introduced in the “Data and Methods” section, and binary variables for months. Figure 3 displays coefficients for a regression model with all independent and control variables, corresponding to Column 10 of Table 6A in the online Appendix.

Figure 3. Factors of SMO topic prevalence in governors’ posts. Note: coefficients and 95% confidence intervals from the LSDV model with the percentage of the SMO topic per post as a dependent variable.

As can be seen from Figure 3, governors who faced elections in 2022 or 2023 posted significantly more about the aid to the war-affected groups, compared to other regional heads. Thus, the empirical results provide support for H₁. When governors approach the end of their term, they tend to please the population and the autocrat more. In the case of the ongoing military conflict, one way of doing it is by emphasizing support measures for the soldiers, their families, and the annexed territories. In this way, regional heads fulfilled one of the main wartime tasks set by the principal-autocrat and increased the chances of his support, which is crucial for retaining office in the personalist authoritarian regime. At the same time, governors co-opt strategically important groups of the population and highlight performance legitimation of the regime, thus preventing the threat from the masses and sustaining authoritarian resilience.

Governors’ political stability is negatively associated with the prevalence of the SMO topic in their social media posts, supporting H₂. When governors are politically “strong” due to patronage from a member of the top elite, the absence of conflicts with other elite groups, or effective management of the region and related factors, he has fewer incentives to put extra effort into fulfilling the principal’s tasks and co-opting war-affected groups through promises of material allocation.

In turn, governors-outsiders, with no prior attachments to their regions, cover more of the SMO topic in their social media posts, in line with H₃. Outsiders, promoted by the PA, tend to be more technocratic governors, oriented towards the delivery of the key performance indicators set by their principal. In this regard, calming down the regional population by emphasizing support measures for the participants of the war in Ukraine is one of the effectiveness indicators. By engaging in performance legitimation of the regime on social networks, governors-outsiders demonstrate their effectiveness as managers, worthy of promotion in the vertical of power. In summary, empirical results support H₁–H₃ and provide solid grounds to conclude that the personal considerations of governors explain the variation in their war-related rhetoric.

Turning to H₄, the level of regional development alters the effort of governors in performing the autocrat’s tasks of pacifying the regional population. Interestingly, heads of more wealthy regions discuss support measures for the soldiers and their families, as well as humanitarian aid to the annexed territories less, despite the large resources available for distribution. On the contrary, the wealth of the region is negatively associated with the prevalence of the SMO topic in the social media posts of governors. At the same time, federal transfers do not affect the coverage of the aid to the war-affected population. A possible explanation for this result lies in regional disparities in the level of soldier mobilization. Poor regions were reported to provide more soldiers,⁸ and therefore their governors had more people who would need to be co-opted to prevent them from organizing anti-regime protests.

The results support H₅: governors from regions with a higher share of mobilized conscripts tended to communicate more aid to them and their families in public rhetoric. Figure 4 plots the corresponding marginal effect of the share of soldiers, mobilized from 21 September–31 October 2022, on the SMO topic share in governors’ posts in the subsequent months. The corresponding regression Table 7A is reported in the online Appendix.

Figure 4. Marginal effects of the percentage of mobilized soldiers on the percentage of the SMO topic in governors’ posts.

Figure 5 plots the top-15 governors in terms of the highest and lowest average share of the SMO topic per VK post. On average, 31% of the text per post by Andrey Klychkov of Orel Oblast is about the aid to war-affected groups of the population, significantly ahead of all other governors. Klychkov has consistently emphasized supplies to fellow-countrymen soldiers. For instance, on 3 May 2022, he posted about his trip to Donbas:

Figure 5. Top-15 governors with the highest and the lowest average share of the SMO topic per post.

… I visited three locations of our troops and handed over soldiers’ parcels collected by residents of the Orel region. Six tons of the necessary things were delivered to our … guys … It is indisputable that Russian soldiers are provided with everything they need, but I am sure that such trips to the locations of our troops are very necessary and important for everyone, both the serviceman and us …

Discussion

Despite the great economic and human costs of the war in Ukraine, the personalist authoritarian regime in Russia has shown considerable resilience. Political elites played a significant role in this, maintaining their own coherence during a crisis manufactured by the autocrat and a narrow circle of his advisors. Moreover, at the time of writing in March 2024, the elites have prevented mass anti-regime mobilization by the population. On the one hand, the regime increased repression by passing new laws against discrediting the Russian army, pushing the opposition out of the country, and imprisoning many anti-war politicians and public figures. On the other, the level of repression is still significantly below the level of the Soviet totalitarian regime. Maintaining a relatively low (from a historical perspective) level of repression is possible through the successful use of alternative, non-violent strategies of authoritarian rule (Gerschewski 2013; Wintrobe 1998). In particular, the Russian regime managed to maintain its stability by materially co-opting important segments of the population – those directly affected by the war – soldiers, their relatives, as well as residents of occupied territories located near combat zones. At the same time, the authoritarian ruler, acting as a principal, instructed his agents – regional governors – to emphasize the regime’s achievements (Busygina and Filippov 2024, 200) and thus use another non-violent strategy of authoritarian rule – performance legitimation (Cassani 2017).

The empirical analysis of governors’ posts on the most popular Russian social network VKontakte revealed that the regional heads carried out the autocrat’s task by emphasizing support measures for war-affected people in their public communications and thus contributing to the authoritarian regime’s resilience. However, the level of effort applied to the task varied among governors and followed a strategic logic. Governors who needed to demonstrate their effectiveness to gain the support of an autocrat – those who were nearing the end of their term, who were politically unstable, or who had no ties to the region before taking office – emphasized more the topic of helping the war-affected population on the VK social network. At the same time, governors of more developed regions communicated the performance legitimacy of the regime to a lesser extent, which can be explained by the fact that they sent fewer soldiers to the front. Indeed, the governors of regions with a larger share of mobilized soldiers in the fall of 2022 emphasized much more the topic of helping combatants and their families after the end of the main wave of mobilization.

In general, the system of incentives built for governors in Russia during Putin’s rule turned out to be quite effective in maintaining the resilience of a personalist autocracy to the challenge of a full-scale war, at least in the short term. Integration of governors into the vertical of power as lower-tier elites and turning them into agents accountable to only one principal – the autocrat – and personally interested in maintaining his popularity brought dividends to the Kremlin during the full-scale invasion of Ukraine in 2022–2023. Acting as agents of the principal, governors were involved in co-opting the masses and legitimizing the authoritarian regime in public communication. This has become an important part of the regional aspect of the regime’s high resilience. However, it remains questionable whether Putin’s personalist regime will be able to maintain the same resilience in the prospect of a long-term war. History shows examples of how war became extremely unpopular among the population and elites, and the regime’s ability to co-opt them and legitimize its rule ended due to limited resources (Kendall-Taylor and Frantz 2014, 41; Pipes 1990). Regional governors will likely remain loyal to the autocrat until members of the top-tier elite begin to betray Putin and initiate a cascade of defections (Hale and Colton 2017).

Conclusion

This paper demonstrates the strategic logic behind the communicative strategies of Russia’s governors during the first 20 months of the full-scale invasion of Ukraine. Elite rhetoric on the social network Vkontakte in 2022–2023 reflects the personal considerations of governors, constrained by regional structural characteristics. Governors up for their elections, with less power vis-à-vis the federal center, and with no prior connections to their regions tend to post significantly more about the Russo-Ukrainian war, focusing on support measures for the regional combatants and humanitarian aid to the annexed territories. At the same time, heads of wealthier regions covered the topic of the war less in their social posts. Finally, governors from regions with a higher share of mobilized soldiers posted significantly more about support measures to war-affected groups of the population. The results are suggestive that the efforts of agents in performing the principal’s task followed a strategic logic in the case of governors in Russia. All governors engaged in a public display of support for the soldiers and their families, as well as for the annexed territories of Ukraine. In this way, they co-opted strategically important groups of the population and highlighted performance legitimacy, preventing mass anti-regime mobilization and contributing to the resilience of Russia’s authoritarian regime. Governors and broader political elites managed to preserve their coherence despite the challenges of the full-scale war. No regional head openly criticized the special military operation in Ukraine and defected during the Wagner Group rebellion in June 2024. Instead, governors continued to serve the role of subordinate agents in the vertical of power, performing the autocrat’s tasks and keeping the population politically demobilized. Thus, this paper demonstrates that the authoritarian regime in Russia maintained high resilience during the first years of the full-scale war in Ukraine, including at the regional level. Further research can compare the communications strategies between lower- and higher-tier elites and how they relate to the propaganda on state media.

Acknowledgments

I am grateful to Ekaterina Paustyan, Irina Busygina, Guzel Garifullina, Stanislav Klimovich, Sasha de Vogel, Alexander Baturo, Timothy Frye, and participants at the Hanse-Wissenschaftskolleg workshop and the MPSA 2024 Conference for their helpful comments on earlier drafts.

Disclosure statement

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

Supplementary data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/1060586X.2024.2353006

Additional information

Funding

The work has been supported by the PhD Scholarship from the School of Law and Government, Dublin City University.

Notes

1. https://tass.ru/obschestvo/16469577, accessed 6 April 2024. https://www.pnp.ru/politics/chinovnikov-zatyanut-v-seti.html, accessed 6 April 2024. https://vecherka.su/articles/policy/181574/, accessed 6 April 2024.

2. https://semnasem.org/articles/2022/08/10/13-molchalivyh-gubernatorov, accessed 31 March 2024.

3. Partial mobilization has been an unpopular government decision, supported by 51% of the citizens in the first days after its announcement. The predominant feelings about this measure, reported by 47% of the population, were anxiety, fear, and horror. See https://github.com/dorussianswantwar/research1/blob/main/Глава_07/Хроники%207.0%20Аналитический%20отчет%20(Экспресс%201%20и%202%20волны).pdf; https://www.levada.ru/2022/09/29/konflikt-s-ukrainoj-sentyabr-2022-goda/, accessed 31 March 2024.

4. The mothers, wives, and other relatives of mobilized soldiers organized small-scale anti-war protests and petition campaigns, demanding better equipment for soldiers and their return from the front (Noble and Petrov 2023). However, the sporadic protests have not transformed into a major anti-regime campaign, which could be attributed to the effects of co-optation and legitimization efforts of the Kremlin during the Russo-Ukrainian war, described in the literature on authoritarian survival and collapse (Frantz and Kendall-Taylor 2014, 335; Kuran 1991).

5. On 30 September 2022, Vladimir Putin signed “accession treaties” with four occupied regions of Ukraine: Donetsk, Luhansk, Zaporizhzhia, and Kherson; see https://www.theguardian.com/world/2022/sep/30/putin-russia-war-annexes-ukraine-regions, accessed 31 March 2024. At the initiative of the president, governors were given the opportunity to take regional patronage (shefstvo) over parts of the occupied regions. By the time of signing the “accession treaties,” more than 40 cities and regions had already become “chiefs” over the occupied territories. The main responsibility of the chiefs is to financially support the restoration of facilities in the annexed regions; see https://novayagazeta.ru/articles/2023/09/29/deneg-uiba, accessed 31 March 2024.

6. https://www.economist.com/graphic-detail/2022/10/21/where-are-russias-newest-soldiers-coming-from, accessed 31 March 2024. https://zona.media/article/2023/03/02/wedding-season-2, accessed 31 March 2024.

7. https://www.themoscowtimes.com/2022/09/21/explainer-what-does-russias-partial-mobilization-mean-a78854, accessed 31 March 2024.

8. https://www.economist.com/graphic-detail/2022/10/21/where-are-russias-newest-soldiers-coming-from, accessed 31 March 2024. The Pearson correlation between the logarithm of GRP per capita and the share of mobilized soldiers in my sample is negative and equal to −0.12 (p = 0.3). However, it becomes significant when I use the logarithm of absolute GRP, which correlates at the 10% significance level with the share of the mobilized conscripts (R =−0.28, p = 0.1).