WORKING CLASS PROTEST IN HYBRID REGIMES

Industrial conflict and strike patterns have received considerable attention across the social sciences. Economists have developed sophisticated models that build on Hicks's insight that, since both parties to a dispute would gain from avoiding work stoppages, asymmetric information and signaling must be key elements of strikes (1932).¹

However, these theories about organization and strikes all assume that workers' organizations are independent and representative; but what does working class protest look like if workers do not have organizations that represent them? The problem comes in two versions: workers who are in hierarchical unions intended for control rather than representation, and workers who are not organized at all. I consider these in turn, though in practice they are likely to coexist.

It is often recognized that the interests of union leadership and the interests of workers might not always be aligned (Golden 1997; Michels 1911; Murillo 2001). However, in many contexts in today's world, particularly outside long-standing democracies, problems of representation are much more severe. It is not simply that officials have a tendency to develop institutional priorities that differ from those of the rank and file. Instead, in much of the world, the unions that supposedly represent workers are not intended for that purpose at all but instead are instruments of control. In this context, the problem of analyzing union strategies and protest takes on a quite different character.

The extreme case is “authoritarian corporatism” in which unions are compulsory, noncompetitive, hierarchically ordered, and functionally differentiated with the purpose of ensuring state control over labor (Schmitter 1974). Examples are the Leninist “transmission belt” labor unions of the USSR and China, and authoritarian corporatist unions, such as the Organizacion Sindical Española in Franco's Spain. Such unions have little or no independence from the authoritarian regimes in question and were created with a view to imposing hierarchical control over workers' organization.

In this paper I focus on the less extreme case of regimes that have given up on overt authoritarian corporatism, but where democratization of labor unions lags behind liberalization in other spheres of the polity. Despite the introduction of competitive elections for national office, in hybrid regimes political machines or corporatist forms of labor incorporation can survive relatively intact. However, with the departure of an explicitly authoritarian political system, official unions no longer have a monopoly but have to compete for survival, relying either on elites and the state or reaching out to their membership base. This creates the possibility for now openly competing elites to use the existing union apparatus to influence the extent to which the public is mobilized around a political issue. Examples of such unions include Russia's largest union confederation, the Federation of Independent Trade Unions of Russia (FNPR), and Mexico's Confederacion de Trabajadores de Mexico (CTM; Bensusán 2000).

A slightly different version common in hybrid regimes is where employers, usually with the blessing of the state, take the initiative in organizing company unions designed to ensure labor discipline and prevent the emergence of representative and potentially troublesome unions. This is the case with the official unions in Malaysia's electronics sector, and with the so-called sindicatos de protección in Mexico. In both instances workers struggle to overcome substantial obstacles to collective action as they face the combined weight and coercive potential of employers, the state and, often, organized crime.

Organizational and institutional conditions in this type of regime are likely to give rise to patterns of strikes that are quite different from those in long-standing democracies. Elite politics plays a major role in determining who strikes and who does not. Consequently, to understand strike patterns in hybrid regimes we need a theory of elite conflict and the interaction of that conflict with top-down unions. Strikes patterns, I argue, depend on how bargaining over the distribution of resources is conducted. Under certain conditions, namely, when regional elites are relatively unified, strike patterns will depend on the political and economic resources of the region. Regions that have poor political relations with the center, and few other bargaining resources, will seek to mobilize workers to pressure the center for transfers. In these regions, the level of strikes will be high. Other regions will generally prefer “quiet” bargaining strategies, and strike intensity will be much lower.

The other case of nonrepresentation is where workers are not organized at all. This too is common, especially in service industries and in the informal economy, and is often associated with neoliberal economic policies (Kurtz 2004). There is relatively little theoretical work on strikes by unorganized workers. Shorter and Tilly (1971) show that unorganized workers rarely strike, and when they do, strikes tend to be longer and smaller. More directly, there is a small literature on so-called “wildcat strikes,” that is, strikes organized either against officially recognized unions or in their absence. In his seminal work on wildcat strikes, Gouldner emphasizes the moral and emotional dimension of wildcat strikes (1954, 53–64). From the workers' perspective, such strikes are a largely spontaneous response to situations in which conditions deteriorate to the point where workers sense that a moral, and not just an economic, boundary has been crossed. In a rather different vein, Brett and Goldberg (1979) find that patterns of wildcat strikes in coal mines are not a function of such things as working conditions and age profiles of the workforce; rather they depend on management style and the atmosphere of relationships between workers and management. Zetka (1992) finds that it is the specifics not just of industry but of labor processes and their effects on group solidarity that matter. Looking at the post-World War II U.S. auto industry, a context in which strikes were outlawed and heavily repressed, Zetka finds strikes were more likely where the work process itself required workers to coordinate their moment-by-moment activities. The likelihood of wildcat strikes, therefore, depends heavily on the details of work organization at the shop-floor level.

In what follows, I lack sufficiently detailed data to test theories that depend on shop-floor level variation. However, I am able to get some insight into the leverage offered by Gouldner's (1954) account in which wildcat strikes are more likely to occur when workers experience unusual levels of economic hardship, combined with a sense that the hardship is morally unjust. Such arguments recall an older line of thinking about protest grounded in theories of deprivation (Gurr 1970). Although such thinking has largely been superseded in the strike literature by institutional and bargaining power approaches, where institutions do little to reflect the interests of workers, or where workers lack organizations to bargain on their behalf, relative or unjust hardship is likely to be strongly associated with success in overcoming obstacles to collective action.

RESEARCH DESIGN: RUSSIAN LESSONS AND HYBRID REGIMES

In the remainder of the paper I focus on one hybrid regime, Russia, where both forms of labor “non-representation” are prevalent. I use an analysis of Russian regions as a plausibility probe to develop and test a theory of labor mobilization in which authoritarian corporatist legacies interact with open elite political competition. Russian experience is particularly interesting when thinking about hybrid regimes because Russia offers a clear example of strong institutional continuities to the former regime coexisting with liberalization of political competition. Though the “transmission belt” role of labor unions is no longer espoused officially and central control at the national level is much weaker, at the regional level there remain strong elements of authoritarian corporatism. This combines with open competition between regional elites and the center, and among different regional elites. The result has been tremendous variation in protest levels across different regions.

Looking at variation at the regional level in a large and diverse state like Russia is also useful in that it provides empirical leverage, controlling for such things as cultural repertoires or past national experience (Ekiert and Kubik 1998), while still having variation on important dimensions such as economic performance, the degree of political competition, ethnic polarization and so forth.

There has been quite a cottage industry of studies looking at protest patterns and politics in post-Communism generally, and in Russia in particular. Although workers played a major role in the Soviet collapse (Crowley 1997), the academic conventional wisdom has been that there has been no protest of note in post-Communist Russia. Mandel asks “Why is there no revolt?” (2000). Ashwin's analyzes “The Anatomy of Patience” (1999), Kubicek examines the “worker passivity in the face of severe economic crisis” (2002, 618), and Richter note “the absence of any sustained protest movement” (2006, 134).

However, the conventional wisdom is misleading. In fact, Russians at times and in certain places have been very highly mobilized. Even according to conservative official statistics, Russia had 111 working days per 1,000 workers lost to strikes in 1997, and 56.1 in 1998. This compares with 13.4 and 11.3, respectively, in Mexico, and 3 and 4, respectively, in what Ekiert and Kubik (2001) call “rebellious Poland” (ILO). Mobilization, as it usually does, has come in waves, with periods of high mobilization alternating with periods of passivity. Comparing peaks of mobilization cycles, Russia experienced 13,213 “protest days” (number of events × event duration) in 1997, and 16,652 in 1998, compared to 3,529 in 1992 and 4,382 in 1993 at the peak of protest in Poland (Ekiert and Kubik, 113).²

Working Days Lost to Strikes in the Russian Federation 1997–2000

The empirical analysis in this paper uses data from 1997 to 2000, covering most of Boris Yeltsin's second term and the period in which the new Putin administration was establishing itself. This is a key time-frame to look at because these years were the period of greatest political competition in the post-Communist era in Russia. As such, these years provide a good environment for looking at the interaction of open political competition with authoritarian labor institutions. I consider the implications of the time period used to test the theory in more detail later.

The conventional wisdom is not only misleading in terms of understanding Russia but also theoretically constricting, for it blinds us to variation that can be useful in the development and testing of theory. Part of the reason that scholars have missed the extent of protest in post-Communist Russia is that the national picture presented in Figure 1 masks enormous Russia regional variation. Table 1 shows the variation in the extent of one subset of protest: strikes. At the top end, we find seven regions, spread from Russia's far east, across Siberia, through the Ural mountains, to southern European Russia, with more than 500,000 working days lost to strikes over 4 years. In the Republic of Khakasiya, 809 working days per 1,000 employees were lost in 1997 and 890 in 1998. Primorski Krai lost 384 working days per 1,000 employees in 1998, and Kemerovo Oblast lost 959 in 1997. But these regions are unusual. Nearly half of the regions (37/88) reported-fewer than 10,000 working days lost in total over the 4 years.³

EXPLAINING POST-COMMUNIST PROTEST

What explains variation in strike patterns in post-Communist Russia? I first consider the existing literature on protest politics in post-Communism. Though this literature pays little attention to variation, it does provide two key insights: that strikes are more common where leadership is provided by trade unions, and that the unions tend to follow instructions from above in organizing protest. I then analyze some of the key elements of relationships among workers, unions, and political power, and show a number of direct and indirect ways in which regional governors are able to influence protest levels.

A range of cultural and ideological theories, from an aging population to the lack of an ideological alternative to the market, have been offered to explain the supposed lack of protest (Christensen 1999; Greskovits 1998). However, because the factors these theories emphasize vary little across regions, or over time, they are unhelpful in understanding variation. For this task, theories that focus on institutional and informational barriers to collective action are more helpful.

Javeline (2003) finds that, although workers in Russia in this period generally had great difficulty in allocating blame for their problems, those who were more successful in ascribing blame to someone were more likely to participate in protest actions. Success in blame attribution is in part a function of clarity of enterprise ownership, but also a matter of involvement with political organizations. Moreover, she shows that those who had been solicited by a trade union or other organization to participate in a protest action were more likely to join in than those who had not. In other words, individuals are more likely to participate in a protest action if one has been organized for them to go to, than if they have somehow to organize it themselves. Yet this does not tell us why an event is organized in the first place. To understand this means looking at the institutional and organizational level to discover why events are more often organized in some places than in others. This is turn depends on the nature of Russian labor unions.

With the exception of a few specialized branches, such as miners, dockers, and air-traffic controllers, independent unions have failed to establish a strong presence in Russia, and by 2000 at most 5% of union members were in alternative unions (Ashwin and Clarke 2003, 1). By contrast, the former official unions, renamed the Federation of Independent Unions of Russia (FNPR), have had considerable success in continuing to dominate the representation of workers. The lack of legitimacy and the inadequacies of labor unions in Russia as representative institutions are well documented (Ashwin; Clarke; Davis 2001). Nevertheless, these unions have a presence in every workplace that existed at the time of the collapse of the USSR, some 38 million members, and a 1% check-off from the salaries of members.⁴

Workers' subordination makes the unions a valuable partner for the state. To capitalize on this relationship, a set of institutions known by the general term “social partnership” has been created to preserve the role of former official unions and to enhance their near-monopoly of representation of workers. Social partnership consists of a huge bureaucratic effort to produce thousands of largely meaningless agreements among unions, employers, and the state, preserving a modified version of Soviet labor institutions in which unions are the junior partner in managing enterprises.⁵

However, if workers are in a subordinate position with respect to the unions, so the unions find themselves in a relationship of dependency with regional authorities. Unions are financially weak and depend on employers or the regional government for administrative resources and for a role in continuing to fulfill social welfare functions inherited from the Soviet era, and, due to their very limited legitimacy at the grassroots, for protecting the unions from encroachment by upstart independent unions. As a result, there is little protest without official sanction. As one union official put it, “In general all protests are held on the initiative of the powers that be: regional governors, directors of enterprises” (Ashwin 1999, 14).

The converse of union weakness is the political strength of regional governors. In most regions, gubernatorial influence extends to the disbursement of state funds, the ownership and profitability of banks and other enterprises, and a vast array of regulatory and administrative powers, creating the potential for powerful political machines (Hale 2003). The influence of governors on patterns of protest is felt in many ways. In most cases influence over protest levels is indirect, less a matter of directly organizing protests, so much as one of deciding when to permit them and when to prevent them. To give one example, during the 1997 railroad blockade launched by submarine repair workers in the Far Eastern town of Bol'shoi Kamen', then Primorskii Krai governor Yevgenii Nazdratenko provided buses to transport protesters the considerable distance to the railroad, and ensured that the police did not act to prevent the illegal blockade. Instead, police provided security for protestors. In the opinion of a highly respected regional commentator, there were no real civil society organizations capable of putting an event like this together, only the labor union in association with the regional administration could have organized it.⁶

In some cases, the use of administrative resources is more direct. In another interview in Vladivostok, one journalist suggested that in many cases “participants in protest actions were actors involved to create the impression of mass participation.” These “actors” were paid by the regional administration. Considerable effort is also put into controlling media coverage of protests and Nazdratenko reportedly paid Moscow-based television companies to cover protest events in his region. Although this may (or may not) be unusual, Russian labor law sets out very complex procedures for strikes (Maksimov 2004, 123), creating an opportunity for strikers to be punished with legal action. Without independent courts, such decisions are subject to close political control on the part of regional officials, ensuring that most strikes that take place are politically acceptable to the regional political leadership.

Given the weakness of the unions, the strength of the regional governors, and the elaborate institutional apparatus for coordination between them, the key players here are regional governors. Consequently, we would expect protest patterns to reflect their interests, taking place where protest is rational from the governor's perspective and not where not.

AN ELITE BARGAINING THEORY OF PROTEST

When would this be? The point of departure is to think about distributive politics and the potential bargaining strategies available to governors in a state divided into different regions or territorial political units that lobby the capital for money and other resources. Broadly, governors have a choice between quiet and noisy strategies, analogous to “loyalty” and “voice” in Hirschman 1970. Quiet strategies are defined as intraelite bargaining, whether through formal federal institutions or more informal political bargaining. Noisy strategies are defined as those that involve public political pressure, usually in the form of protest actions such as demonstrations, pickets, and strikes. All things being equal, governors prefer quiet strategies because the relative lack of public attention gives them greater flexibility in disbursing monies that are transferred from the center. By contrast, noisy strategies widen the circle of players, bringing into the picture mass actors who are likely to be in a stronger position to demand a share of the resources generated. Moreover, quiet strategies are less risky than noisy ones, in terms of intraregional politics. Noisy strategies create potential for unforeseen political or social instability and train participants in collective action that may make them more independent later.

However, not all governors will be successful in obtaining resources from the center using only quiet strategies. In fact, governors who have bad political relations with the center and who are weak in bargaining will systematically resort more often to public pressure, including the organizing of strikes and protests.

The reason for this turns on the nature of bargaining over transfers. In any bargaining situation between regions and the center, the key issue faced by the center is to prevent the regions from coordinating their strategies. Without coordination, the center can play the regions off against each other, buying off the most powerful regions, neglecting the others. Whether this is possible depends on the nature of the issue. On the question of transfers and payments, preventing coordination is relatively straightforward. Assuming some sort of budget constraint, central funds are clearly rival in consumption; what is transferred to one region cannot also be transferred to another. This is different from issues such as sovereignty and autonomy where a grant to one region is likely to increase the chances of concessions to another. In the autonomy case we are likely to see regional bandwagons. In the case of fiscal transfers bandwagoning is unlikely. Without a credible threat of interregional coordination, the center will be primarily concerned with the needs of the most powerful governors, leaving weak governors to exert what pressure they can through the use of noisy strategies. To implement this set of strategies governors will, of course, rely on the cooperation of other important players in regional politics; convincing these actors to go along will depend both on the interests of the players themselves and on the political strength of the governor. I first discuss the interests of each of the relevant players independent of the governor, and then consider factors that affect a governor's capacity to influence them.

The relevant players are workers, unions, and employers. Workers can be expected to participate in protests given that participation has few costs, particularly when action is backed by the regional administration and when there is the potential that noisy pressure might bring some concrete benefits. These incentives for workers to participate will be stronger the greater the economic hardship they face, especially where that hardship is considered to be unjust. Moreover, the greater the hardship, or the perception of injustice, the more likely there is to be wildcat action, that is protest activity beyond the framework of the official unions. Where deprivation is worse, official efforts to prevent protest are less likely to be successful. Hence we would expect workers to cooperate with governors promoting strikes, but to cooperate less when economic hardship is high and governors want to demobilize strikes.

As far as unions are concerned, I have outlined earlier the nature of the unions' dependency on the political support of governors. Consequently, union support will not tend to be particularly problematic. However, it does depend in part on the extent that independent unions have influence in the region. In cases where independent unions represent a real threat to successor unions, experience elsewhere suggests that successor unions will tend be more militant (Murillo 2001; Robertson 2004). This means more willingness to cooperate in noisy strategies and less in demobilization.

What about employers? In the case of demobilizing protest, the interest of employers seems clear, but why would they be willing to participate in noisy strategies involving strikes? In brief, of course, not all will. In particular, employers in small enterprises and new post-Communist start-ups are not likely to play along. However, given that these types of companies are both far more likely to be economically profitable, and almost entirely nonunionized, such employers are not likely to be called on. Outside of small enterprises and new start-ups, however, the distinction between employers and the state is often hard to make in practice. In some cases, teachers and healthcare workers prominent among them, the employer is the state. In other cases, formal legal privatization may have taken place, but especially for larger enterprises that were formerly (or still currently) state owned or that relied in part or whole on state orders, the economic environment in the 1990s was at most semimarketized. The combination of a political interest in employment preservation with massive nonpayment of obligations among enterprises and between enterprises and the state helped to maintain the soft budget constraints of the Soviet era. Hence to the extent that regions look for subventions from the center, rely on state employment, and/or face a weakly marketized economy, employers are likely to be more willing participants in regional bargaining coalitions.

Of course, governors have important levers of influence over these three sets of actors, as noted previously, and cooperating with governors will seem more reasonable when the governor is politically dominant in her region. Political dominance will depend on several factors, but most importantly on players' expectations about the likelihood that the governor will stay in power to see the deal through. When elections are far off and/or when governors do not face a strong challenger in their bid to gain reelection, governors are likely to have greater capacity to ensure cooperation from workers and regional elites.

ELITE BARGAINING AND PROTEST: HYPOTHESES AND MEASURES

ORGANIZATION, BUSINESS CYCLES, AND INFORMATION: HYPOTHESES AND MEASURES

I now turn to alternative explanations. Existing theories of strike patterns are primarily concerned with factors that influence relative bargaining power (independent organization and business cycles) or the information environment. In this section I briefly set out some hypotheses designed to test these theories.

The essential insight of bargaining power theories is that we should expect to see more mobilization in regions where workers' capacity to self-organize is high (Cohn and Eaton 1989; Snyder 1977), due both to the effect of independent organizing itself and to the galvanizing effect on Communist successor unions. Capturing capacity for self-organization of workers is tremendously difficult. Standard approaches using union density are useless where workers are passive members of former official unions. I have made a first cut at solving this problem by constructing a list of regions in which one of the leading independent trade unions, Sotsprof, has managed to establish a real organizational presence. Sotsprof is one of the largest alternative union confederations, with many small independent union affiliates in a range of sectors. Sotsprof tends to confederate existing local alternative unions rather than organize itself, so it provides a reasonable guide to regions where independent unions are present. I used official lists of Sotsprof branches and, on the basis of interviews with Sotsprof officials in Moscow, eliminated regions where organizations existed only on paper.

Bargaining power, of course, is also related to business cycles. In advanced industrial economies the expectation is that the best time to strike is when labor markets are tightening; that is unemployment is falling, and so workers are in shorter supply, and so are relatively stronger (Ashenfelter and Johnson 1969).

Turning to theories about factors that are likely to affect the information environment, uncertainty over the amount of money available for transfers (and so uncertainty over the center's capacity to make concessions) ought to increase the level of strikes. Hence, to the extent that the budgetary position of the central government is unclear, mistakes are more likely, and strikes should be more frequent. To proxy difficulties parties might have in estimating what the upcoming budgetary situation might be, I use the coefficient of variation in oil prices over three months. Table 2 summarizes hypotheses and measures.

STRIKE DATA

Strike data in this paper are drawn from a new protest dataset covering 88 regions of the Russian Federation for 4 years, from January 1997 to December 2000. The data were compiled by the author from unpublished daily text reports of the Interior Ministry. The reports are prepared at MVD offices in each settlement in the Russian Federation, and sent to Moscow to be compiled into summary reports, or svodki. Each daily report lists new, continuing, and ending strikes; the number of participants on a given day; the location; economic sector of the strike; and the demands made.

The new data are a considerable improvement on published strikes statistics that rely on self-reporting by enterprises. Moreover, only strikes that are legal and officially endorsed by the unions are recorded in official statistics. Both of these factors lead to significant undercounting. In an interview with the author, the President of Sotsprof, one of the new “independent” unions, estimated that 80% of the strikes organized by his union were ruled illegal. By contrast, the MVD reports are primarily used for national security purposes, and so officials are responsible for actively monitoring levels of actual, as opposed to only de jure, disruption in the areas under their supervision. Comparisons between the MVD data and official sources show the same macro-trends over time while suggesting higher levels of activity than published sources. The MVD reports are also an improvement on data available from nationally representative surveys. Although, as I have shown, strikes are concentrated in a relatively small number of regions, nationally representative samples that do not take this nonrandom variation explicitly into account are not likely to give an accurate picture of strike levels.

Obviously the new data do not solve all of the problems with the compilation of statistics on strikes and protests. Many of these problems, ranging from the social and technical definition of a strike, through to establishing the number of workers involved and the amount of time lost, arise every time labor statistics are discussed (Knowles 1952). Moreover, the MVD data are also subject to concerns about the political incentives and bureaucratic habits of that organization. It is a well-known regularity that police estimates of the numbers of participants in protest events are almost always conservative and are certainly below the numbers estimated by protest organizers, and usually below those of media observers too. Predictably, a comparison of the MVD police reports with opposition and scholarly reports of particular incidents (the Vyborg Cellulose Plant conflict and Astrakhan Gazprom blockade) suggest that the MVD is slow to report the beginning of the most conflictual events and understates participation when events are underway. On the other hand, under-reporting is less of a problem with this source than official sources or media sources. Nevertheless, in terms of absolute numbers, it is clear that this source should be treated as a conservative guide to the underlying phenomena.

Access to the underlying source for the data was, given its nature, limited to 1997 to 2000. Nevertheless, this is a particularly appropriate period for looking at the theoretical questions of interest. This period provides significant variation on the dependent variable, as well as on economic independent variables. Strikes and protests increased through 1998 when the balance of unpaid wages to Russian workers reached a peak of some R56 billion, and then declined through 1999 and remained low in 2000. The period also saw the August 1998 crash of the Russian stock market and currency, the most serious economic crisis that Russia had seen since the stabilization of 1992 to 1993. Following the crash and the concomitant currency devaluation, the Russian economy began to experience its strongest economic performance since the mid- 1980s.

There is also great variation in the political context: 1997 through 1999 was a period of acute elite conflict in the run up to the 1999 parliamentary elections and 2000 presidential elections, when the question of the succession to Yeltsin was being decided. By December 1999, it was clear that a new leader, Vladimir Putin was in place, and elites scrambled to pledge allegiance as a more vigorous regime established itself in the Kremlin. This provides an excellent opportunity to assess the tools and tactics of center-region competition and of competition for supremacy among local elites in a period in which that competition was most open. The principal limitation of the data is that we have a relatively limited view of the Putin era, only 10 months, and so we have less insight into patterns of behavior when elites are more afraid of openly criticizing the center. This limits the extent to which the analysis applies to strongly authoritarian contexts. On the other hand, we do begin to see in these 10 months the effect of the political consolidation of the Putin era as the authoritarian elements of the Russian political system reestablished themselves over the competitive elements.

Appendix 1 presents descriptive statistics. I use data from all sectors except mining. Miners' strikes are concentrated in coal producing regions. Moreover, the miners' movement is seen as quite different from the rest of labor politics in Russia (Borisov 1997; Maksimov 2004).

There are various ways of measuring strikes. I follow the classic treatment (Knowles 1952) in making the dependent variable the “'composition of the strike movement” or the “severity” of strikes (total number of working days lost): given by magnitude (workers per strike), duration (working days lost per worker), and frequency (number of strikes). An alternative to working days lost would have been to measure “strikes” as a count of events. This approach is inferior in this case for a number of reasons. Most importantly there are major conceptual difficulties in determining whether a number of events in different places are one strike or many. To do so presupposes an underlying theory of the organizational process behind the events: are they on the same issue, are they organized together, is there more than one legal establishment involved, and so forth. None of these issues are resolvable on the basis of the MVD reports.