The Spatial Unevenness of Levels of Political
Competition in the Russian Federation
An investigation into the convenience of the subnational resource curse in explaining variation in politicalcompetitiveness Bachelor thesis Robin Verheij 10576045 30-01-2017 8443 words
Mentor: Ph.D. Seiki Tanaka Second reader: Dr. Julia Bader
Introduction
Contrary to what has been widely recognized by comparative political scientists, subnational variation in political competitiveness is relevant for both unitary and federal countries around the globe (Snyder, 2001). National-level theories are increasingly able to explain the determinants of authoritarianism and democratization. However, the determinants of subnational authoritarian stability and democratization are still nebulous: a large part cannot be easily explained with national-level theories.
Considering that is seems likely to observe within-nation variation in enormous countries like India or Argentina, taking one data point seems illogical. Not only its size makes Russia an interesting case to study; the fact that Russia’s regions vary tremendously on many levels, including their economy, geography, ethnicity and political competitiveness also plays a role.
This thesis addresses first and foremost the alleged ‘subnational resource curse’. The resource curse is widely researched for antidemocratic effects of mineral wealth (Ross 2001; Beblawi 1987; Mahdavy 1970). Some recent research points in the direction of a subnational resource curse (Saikkonen 2016; Diaz-Rioseco 2016; Vasilyeva & Libman 2016).
In order to contribute to the growing research effort on subnational democratization and authoritarian breakdown and stability, this thesis addresses the following question: ‘‘Is the resource curse a valuable explanation for subnational variance in political competition in the Russian Federation?’’. This question is important to investigate, as ‘citizens, regardless of their location, should enjoy the benefits of democratic institutions and liberties’ (McMann et al., 2016: 4). More knowledge about subnational political competition could help to relate these findings to the occurrence of, for example, violence and civil rights on the subnational level.
This thesis provides support for the idea that the subnational resource curse does not explain subnational variation in the Russian Federation by taking into account the mineral extraction tax that the Russian government introduced in 2002. This thesis shows that since 2002, mineral rents that previously remained in the regions are ‘taxed away’ and therefore do not account for variation in subnational political competition. This leaves room for other explanations for authoritarian breakdown and stability, like the gross regional product or the relative size of the urban population. In addition, this thesis suggests other potential determinants of subnational regime variation and argues that more data collection is needed to fully probe the topic.
This thesis does not intend to sketch a complete explanatory model of variance in subnational political competitiveness. Rather, it investigates the convenience of the subnational resource curse as an explanation and it tests some other determinants. Thus, this thesis should be considered as a more general preamble to the subject. Bearing in mind the limitations of this thesis, there is a whole lot that this thesis cannot discuss. Therefore, only a small part of the universe of subnational political competitiveness is elucidated. A range of other potentially interesting variables is mentioned in the discussion section.
To answer the research question, this thesis starts with a literature review that elucidates common national-level determinants of autocratic stability and breakdown and democratization. It then introduces the subnational resource curse as a possible explanation for subnational variation. In the theory section, it is argued that the subnational resource curse is not operating in Russia since the central government taxes away resource rents. Therefore, it is expected that regional incumbents cannot monopolize these rents anymore for authoritarian practices. The analysis contains a descriptive part about how subnational variance emerged after the collapse of
the Soviet Union. Then, the analysis provides evidence that supports the notion that mineral rents are taxed away and subsequently investigates some other possible determinants of subnational variation in political competitiveness. The discussion addresses the limitations of this research and suggests a list of determinants for further research.
The contribution of this thesis to the existing body of literature is that it incorporates the idea that countries are internally heterogeneous instead of homogenous. Furthermore, this thesis contradicts the claimed antidemocratic effects of the subnational resource curse in Russia and starts an investigation into the true determinants of subnational regime variation. Besides, this thesis uses the most recent data possible. The most frequently used indicator for subnational political competition in Russia stems from 2005 (see Lankina, Libman & Obydenkova, 2016; Saikkonen, 2016). Considering that the Russian political landscape has changes significantly since then, more recent indicators seem to better fit into this dynamic reality.
On authoritarian breakdown and stability
This section provides a short overview of the existing literature on authoritarianism, distinguishing between factors that account for authoritarian stability and factors that account for authoritarian breakdown. It then shows what is missing in the literature and how this thesis attempts to fill this gap.
Among the variables that explain authoritarian endurance, natural resources are the most prominent factor in the literature. Especially in the Middle East, authoritarian stability is often explained through natural endowments (see Gause 1994; Chaudhry 1997; Vandewalle 1998). In his ground-breaking work ‘Does oil hinder democracy?’, Ross (2001) discusses the antidemocratic effects of mineral wealth in countries worldwide. According to Ross, democracy is hampered through three causal mechanisms. Firstly the rentier effect, where governments use low tax rates and high expenditures on patronage and public goods to diminish demands for democracy. Secondly, the repression effect, where governments build up their internal security forces to fend off democratization forces. Thirdly, the modernization effect, ‘in which the failure of the population to move into industrial and service sector jobs renders them less likely to push for democracy’ (Ross, 2001: 356, 357). The theory and argument section goes deeper into the logic behind the resource curse.
Apart from the effects of mineral wealth, other factors that influence authoritarian stability are studied. For instance, Bak and Moon (2016) argue that foreign direct investment (FDI) reduces the probability of experiencing political defies from elites. An important factor in this reasoning is the degree to which the government that ‘hosts’ the FDI is able to monopolize the prospective economic benefits. If this degree is high, authoritarian incumbents can spend more on patronage and so buy off dissent. This logic is comparable to the rentier effect of the resource curse since both mechanisms focus on patronage and bribing the opposition. Another mechanism through which FDI’s promote authoritarian stability is by enlarging the elites’ dependence upon the revenues of FDI’s. When elites get more dependent, defection is expensive because they cannot benefit from the prospected revenues anymore. Therefore, elites are expected to cooperate with the regime and therewith increasing regime stability (Bak & Moon, 2016: 1999).
Additionally, there are the factors that indulge authoritarian breakdown and democratization. One of them is the level of education in a country. Sanborn and Thyne (2014) show how higher levels of education promoted by the Tunisian state led to large protests and ultimately the country’s transition to democracy in 2011. Considering that the ‘disaffected graduates of the Tunisian education system’ could not find employment, the ‘educated population living in poor economic conditions galvanized opposition to the regime’ (Sandborn & Thyne, 2013: 791, 792). It is argues that education promotes democratization because it fosters ‘individual modernity’ (see Sack, 1973). This includes a growing awareness of one’s embeddedness in society: (former) students consider the context of their behavior and are expected to be interested in the mass media and worldwide travels. This attitude has a positive impact on political participation, stances and effectiveness (Sanborn & Thyne, 2014: 790-792).
Democratization is also often approached in terms of leverage and linkage to the West. Unlike education, this approach addresses non-domestic factors. Levitsky and Way define leverage as ‘the degree to which governments are vulnerable to external democratizing pressure’ and linkage as ‘the density of economic, political, diplomatic, social and organizational ties and cross-border flows of trade, people and communication’ (2006: 379). The interplay between
leverage and linkage determines the degree to which a country can democratize. Only leverage, in the sense of democracy promotion, is not enough for a country to democratize. Linkage, in the sense of civil society relationships to the West, is needed to ‘raise the cost of autocratic behavior by heightening the salience of abuse, increasing the likelihood of a western response and creating a broad domestic constituency for international norm-abiding behavior’ (Levitsky & Way, 2006: 396).
As shown here, in the study of authoritarian regimes, most researchers have focused on country-level explanations. In the past two decades however, major global centrifugal trends have raised the attention for subnational analyses, according to Snyder (2001: 93, 94). These trends, propelled by the ‘Third Wave of Democratization’ (Huntington 1991), induced decentralization and increasing heterogeneity within countries. McMann et al (2016: 4) determine that within one country, it is not necessarily the case that the same regime type exists everywhere. In other words, the territorial spread of regime type is irregular. Therefore, research on the national level fails to fully explain subnational variation in regime type. We will now narrow down the general topic of authoritarian breakdown and stability to subnational Russia.
According to Obydenkova and Libman (2013: 460), in a state with unsuccessful national democratization like Russia, it is important to wonder what describes the survival of subnational democracy. In their article, Obydenkova and Libman (2013) focus on the path-dependent character of subnational regime type. Their conclusion is that the regions that have experienced some kind of subnational democracy right after the collapse of the Soviet Union still produce different -more democratic- electoral outcomes than the regions that remained less democratic, despite Putin’s efforts of centralization (Obydenkova & Libman, 2013: 464). The reason for this resilience is that because of learning effects from the democratic past, the ability of the central government to change the political attitude back towards authoritarianism in a region is reduced (Obydenkova & Libman, 2013: 463).
Considering the national resource curse as a prominent factor that accounts for national variation in autocracy, Saikkonen (2016) argues that ‘subnational political competitiveness is related to the structure of the economy and to natural resource rents, and to a lesser degree to the specific Russian federal context’ (Saikkonen, 2016: 437). Her central claim is that national-level theories about authoritarian stability also explain subnational authoritarianism. In line with this claim, she finds evidence for the existence of a ‘subnational resource curse’. The logic remains the same in the subnational version of the resource curse argument: resource rents can permit for lower levels of taxation, therewith reducing popular dissent, and higher spending on patronage and state coercion. The resource-rich regions in Russia are more prone to become authoritarian, since the subnational regime monopolizes the rents made out of minerals (Saikkonen, 2016: 444).
This literature review provided a short introduction into the determinants of autocratic breakdown and stability on the national as well as the subnational level. Due to the occurrence of subnational variation in regime type, national-level theories do not fully explain the determinants of subnational authoritarianism. Despite the growing popularity of the subject, researchers still have not fully explored the determinants of subnational variation. This thesis is an attempt to contribute to the knowledge on this subject by combining existent determinants of subnational variations and adding some new possibly influential variables.
This thesis addresses first and foremost the alleged subnational resource curse. Even though Saikkonen (2016) finds evidence for a subnational resource curse in Russia, this thesis
argues that this theory does not explain subnational variance in political competition. The main argument is that Saikkonen overlooks the fact that the Russian government introduced a mineral extraction tax in 2002 that ‘taxes away’ the profit made out of minerals. This makes it impossible for a subnational incumbent to monopolize resource rents in order to employ authoritarian practices. This thesis shows how natural resource rents are redistributed and how much is actually left for the province. This paves the way for other possible reasons for subnational variance except for the resource curse.
Besides, the analysis includes the most recent possible data, until 2011. The most frequently used indicator for subnational political competition is the democracy index of the Moscow Carnegie Center that only goes until 2005 (see Lankina, Libman & Obydenkova, 2016; Saikkonen, 2016). Considering that the effects of the in 2002 introduced mineral tax would take more than three years to show, Saikkonens results do not cover this and could therefore be skewed.
Ross (2015: 246) argues that it is likely that autocrats who wish to solidify their regime accumulate remaining mineral revenues in order to be more resilient in the case of potential revenue instabilities. This underlines the possible remaining long-term effects of resource revenues once appropriated. In other words, autocrats prepare for potential bad luck by saving up rents.
Furthermore, the Russian political environment changed significantly from 2005-2013 under Putin’s autocratic rule, as the analysis will clarify. Hence, this thesis contributes to the existing literature by investigating the most recent possible year, taking into account the dynamic Russian political landscape and the time it takes for a new tax to yield effect.
Theory and argument
This section addresses the logic behind the resource curse and then argues why this does not explain subnational variance in political competition in Russia.
As mentioned in the literature review, Ross (2001) examined the antidemocratic effects of mineral wealth in countries worldwide. The three mechanisms through which this happens are the rentier effect, the repression effect and the modernization effect (Ross, 2001: 356, 357). Ever since this influential article, many scholars have tried to tackle the resource curse. In 2015, Ross published a new article, called ‘What Have We Learned About the Resource Curse?’. In this article, he recapitulates the debate on the topic and discusses the puzzles that remain. This article is used to define what is considered as natural resources and the rents made out of them.
What are natural resources? According to Ross (2015: 241), most conceptualizations of natural resources contain three features. Firstly the type of resource: only petroleum is widely acknowledged as a stimulus for autocratization. Petroleum consists of oil and gas (‘‘Petroleum Through Time’’, 2015). Other minerals and agricultural products are therefore not regarded as natural resources. The second feature is the relevant value of the resource, which encompasses ‘the quantity of production, the value of production, the rents generated by production and the value of exports’ (Ross, 2015: 241). Thirdly, the way in which values are quantified: whether oil is regarded as a fraction of the gross domestic product (GDP), total export, government revenue, per capita and so on (Ross, 2015: 241, 242). The operationalization section goes deeper into the precise measurement used in this thesis.
To define mineral rents, this thesis uses the same definition as Dunning in his influential book ‘Crude Democracy’ (2008). Dunning understands rents as ‘super-normal (oligopoly or monopoly) profits, or the excess over the return of capital, land and labor when these factors of production are put to their next-best use (Monaldi, 2002a). Dunning (2008: 6) argues that these rents ‘tend to pour directly into the fiscal coffers of the state, providing the Fisc with something akin to an externally generated ‘‘windfall’’ and also displacing taxation as an important font of public revenue’ (see Beblawi, 1987; Mahdavy, 1970).
Saikkonen (2016), who is also already mentioned in the literature review, is one of the few scholars to have written about an alleged Russian subnational resource curse, in her case especially about the rentier effect. She argues that national-level explanations for authoritarian stability still hold at the subnational level. Since the differences between Russian regions in terms of resource endowments are immense and natural rents can provoke lower taxation and higher spending on state pressure, Saikkonen hypothesizes that resource-rich regions can become more autonomous and therefore procreate less competitive elections (Saikkonen, 2016: 444).
It can be argued that this logic of reasoning focuses on the structure of the regional economy and that natural resources produce power asymmetries between the region and the central government (see table 1). Nevertheless, Saikkonen draws her conclusion a bit too soon. This section discusses other aspects to take into account that could influence the presumed undemocratic effects of the subnational resource curse.
The main subject that Saikkonen neglects is the relationship between the province and the central government. More specifically, she assumes too quickly that every region can keep the possible rents from mineral wealth to itself. In her conclusion (2016: 450, 451), Saikkonen mentions shortly that control over natural resources diverges between and within nations, and that this affects the resources accessible to subnational incumbents. She then encourages further studies to fully probe this question.
Why is it important to take into account the fiscal relationship between a country’s central government and its regions? As Ross discussed in 2015, the effects of oil on democracy depend on ‘the ruler’s ability to capture the available resource rents’ (Ross, 2015: 245). So the question arises to what extent a local incumbent has access to the resource rents and therewith the option to monopolize these rents for his own profit. Alexeev and Chernyavskiy (2015) have addressed this question, though in a slightly different context: they examine the impact of natural resources on local economic growth since the introduction of the mineral extraction tax (MET). Despite the different dependent variable (local economic growth instead of local political competition), this article is still useful because it assesses the tax relation between recourse-rich provinces and the central government. Moreover, it addresses directly the subnational resource curse: as Alexeev and Chernyavskiy (2015: 318) exemplify, their primary focus is ‘the degree to which the federal government has been able to tax away regional natural resource rents, especially oil and gas rents, mainly due to the unique system of taxation of oil extraction. Ultimately, their findings argue that the subnational resource curse hypothesis does not hold in Russia.
Let us now look deeper into this system of taxation on minerals in order to shape our argument. Oil and gas occupy a large part of Russia’s federal budget: in 2014 for example, revenues from these minerals comprised 46 percent of the government’s income (Waller, 2016). The three main taxes imposed by the Russian central government are value added tax (VAT), income tax, the mineral extraction tax (MET), a subsurface use tax and an export duty. . Oil and gas companies pay the MET to the region, then the regions transfers this to the central government. The VAT and the corporate income tax are quite facile to determine: the VAT is 18 percent on goods, works and services, while the corporate income tax is 20 percent (Waller, 2016). Yet, the MET, the subsurface use tax and the export duty are much harder to calculate.
This thesis focuses only on the MET since the subsurface use tax and the export duty vary among companies and are therefore too entangled to take into account here. Likewise, the precise calculation of the MET is challenging. In the first place because the formula depends on a lot of factors like the type of mineral, the amount of minerals, the price of minerals, the exchange rate of the ruble, the value of a reference fuel, the complexity of the fuel, domestic and foreign gas prices and the cost of transportation (Waller, 2016). Moreover, the precise calculation changes over time, aligning with modifications in the tax system. For example, Putin altered the tax system in 2012 in order to ‘increase the rent burden on mineral mining’ for the regional governments (Susarov, 2012).
Summarily, the amount of MET an oil or gas company has to pay is dependent upon the amount of mineral wealth a region has: the higher the amount of extracted minerals, the higher the amount of tax the province has to pay (see table 2). In other words, the share of the tax collection of mineral rents that remains in the province, decreases.
Even though Alexeev and Chernyavskiy find evidence to object to the resource curse hypothesis, they do not claim that resource-rich regions do not benefit from their regional wealth. Instead, they argue that ‘the pre-2002 regional rents from natural resources have remained within the regions’ (Alexeev & Chernyavskiy, 2015: 320), which could possibly explain Saikkonens findings. As discussed before, this is probable because of the long-term effects of previously accrued resource rents for the stabilization of authoritarianism (Ross, 2015: 246).
In sum, the effects of the resource curse accounts for authoritarianism and stability in some countries in the world. As discussed before, due to the uneven spread of regime type, subnational variance should be investigated in itself. Saikkonen’s (2016) effort to uncover a
Russian subnational resource curse is a good start to explain subnational variance but fails to incorporate the mineral extraction tax that was introduced by the Russian government. It is therefore argued that it is not just mineral wealth that explains subnational variance in political competition.
This thesis hypothesizes that the subnational resource curse does not explain subnational variance in political competition if mineral rents are taxed away by the Russian government since 2002. If this hypothesis were true, we would expect no relationship between regional mineral wealth and political competitiveness. The reasoning behind this is that the MET disables subnational incumbents to monopolize resource rents for rentier effects. Nevertheless, the remaining pre-2002 rent reserves could still clarify why resource-rich regions are still more authoritarian today than resource-poor regions.
Table 1. Saikkonen’s (2016) argument: the effects of resource rents on subnational political competition.
Resource rents (province) Degree of subnational political competition
High Low subnational political competition
Low Relatively more subnational political competition
Table 2. Thesis argument: what happens to political competition when taking into consideration the mineral extraction tax.
Resource rents (province) Mineral extraction tax Thesis argument
High High Rents taxed away: no effect
on political competition expected
Low Low subnational political
competition
Low High No tax duty, so no impact
Research design
This thesis analyses the determinants of subnational variation in political competitiveness in Russia, addressing first and foremost convenience of the subnational resource curse to explain this variation. As discussed before, the Russian subnational universe is still vague. This section goes deeper into the question how to examine the research question.
The database used in this thesis is the social and economic database created by the International Center for the Study of Institutions and Development as part of the project ‘‘Institutions and Economic Development: the Role of Bureaucracy and Experiments and an Instrument for Reform Analysis and Evaluation’’ (2011-2013).
To examine the impact of the mineral tax on the effects subnational resource curse, we need to know how much of the mineral rents are left for the province and how much is taxed away by the central government in Moscow. As discussed before, the precise calculation of the mineral tax is extremely devious and changes from time to time. Since we will not be able to determine the precise amount of mineral tax paid to the central government, we use the following strategy. We use the table presented by Alexeev and Chernyavskiy (2015) about the allocation of mineral tax between federal and regional budgets. We used the twenty-four most oil- and gas-rich regions and calculated for 2002, 2005 and 2010 the oil and gas revenue separately in US dollars. We converted tons of oil to barrels and cubic meters of gas to gallons and used the average value of the dollar for each year. Table 4 (based on the appendix) shows how much of the share of the mineral tax that remains in the five biggest oil and gas producing regions has declined from 2002-2010. This relative decline is used to calculate how much of the earned mineral tax is left for every region. If mineral rents were taxed away, we would observe that after a while, (almost) nothing of the revenue from oil and gas remains in the province.
Besides the effect of the mineral tax on the subnational resource curse, the analysis tests some other relevant explanations for subnational political competition. We discuss these explanations here. We start with two variables already discussed in the literature review, being the level of education and the presence of foreign direct investment in the region. Foreign direct investment is expected to reduce subnational political competitiveness because profits are used to buy of dissent (Bak & Moon, 2016: 1999). Education on the contrary, is expected to nurture ‘individual modernity’ and thereby contribute to higher levels of subnational political competition (Sandborn & Thyne, 2013: 791). Both these variables account for authoritarian stability or breakdown at the national level and therefore are meaningful to test at the subnational level.
Furthermore, it is suggested that more control over economic resources leads to more patronage and thereby elite cohesion and the strengthening of authoritarian rule (Sidel, 2014: 177). This explanation leaves room for the impact of pre-2002 resource rents on political competition since these rents can be considered as part of the local economy and therefore part of the control over the local economy. Following Levitsky and Way (2012: 870), patronage is understood here as ‘the redistribution of the spoils of office’, indicating the ability of the incumbent to spur the devotion of partners within the regime. Besides, regional wealth could foster the regional governments’ authoritarian capacity since it allows the local government to employ means of repression (Saikkonen, 2016: 443). This idea can be linked to the repression effect of the resource curse, where security forces are used to create obedience from citizens and so hamper democratic forces (Ross, 2001: 356). One example of this is the presence of tanks in the region that imaginably afflicts popular dissent. If this argument were true, we would observe
lower degrees of subnational political competition in regions where the incumbent has tight control over the local economy.
As Saikkonen (2016: 443) argues, following Levitsky and Way (2010), ‘state bureaucracy such as tax or regulatory agencies is typically employed in low-intensity coercion against political opponents’. This implies that jobs in the regional bureaucracy are a vital source of patronage. Therefore, we could expect that the larger the regional bureaucracy, the greater the state control, the greater the authoritarian stability and the lower the subnational political competition.
Another factor that this thesis takes into account is the size of the urban population in relation to the total population. It is suggested that it seems simpler to exert power over an urbanized population than over a scattered, rural population (McMann, 2016: 17). In this sense, more urbanized areas could be less susceptible to authoritarianism. In addition, Zubarevich (2016) argues that the distribution of the population is an important factor influencing the path of modernization, regarding the ‘rural and apolitical periphery’ of Russia. Therefore, we expect to observe more political competition in more urbanized regions.
In the analysis, a regression is preceded by a descriptive analysis. The main reason for this combination is the limited amount of available quantitative data. Also, not all factors are analyzable in a regression analysis. The descriptive analysis addresses the Russian political environment since the collapse of the Soviet Union and the inherited path-dependent structures that explain subnational variance.
The regression is an ordinary least squares (OLS) regression, assuming that there exists a linear relationship between the independent variables and the dependent variable. The discussion addresses the assumptions that belong to this regression.
The unit of analysis is the regions of Russia. Therefore, the spatial coverage is the Russian Federation. The Duma, the legislative body, elections used as benchmarks, hence the time span researched here is 2003-2011. Since this thesis analyzes temporal variance, a year dummy for 2007 and 2011 is added to the regression.
Operationalization
This section operationalizes the dependent and independent variables. The dependent variable is the degree of subnational political competition. Existing measurements of local political competition, democracy or autocracy vary on the matter of measuring ‘just’ the electoral system, to more elaborate measurements including significant aspects of democracy like civil rights, media freedom and rule of law. Gervasoni (2010: 314) argues that those significant aspects cannot be fully incorporated into an operable measurement of democracy because of the unavailability of comparable data for all provinces (in his case in Argentina). Furthermore, it is difficult to measure authoritarian or non-democratic practices like the exclusion of oppositions from state jobs or vote buying because they often happen in the most sophisticated and manipulative ways.
Therefore, it is argued that it is more suitable to use objective measures of political competition than subjective measures, also taking into account that it is not realistic to assume the validity and reliability of data about such nondemocratic practices (Gervasoni, 2010: 315). This is the reason why, following Gervasoni, this thesis uses objective measures to operationalize subnational political competition. It is widely acknowledged in the literature that ‘no country in which incumbents win more than 70 percent of the votes or seats is likely to be a democracy’ (Gervasoni, 2010: 315; Prezworski 1991; Alvarez et al., 1996; Vanhanen, 2000; Saikkonen, 2016; Panov & Ross, 2013). This is the reason that the 70 percent cutoff is used in this thesis, too. Not only because of the abovementioned reasons, but also because of the very limited amount of recent subnational data available about Russia, subnational political competition is operationalized here as the percentage of votes for United Russia (UR), Putin’s party, in the Duma elections of the years 2003, 2007 and 2011. This decision is based on the desirability of objective and recent data over more elaborate but much less recent data and pretends in no way to be an exhaustive measurement of subnational political competition. The discussion will reflect on the potential implications of this decision.
One problem with this operationalization that we can address here is that the percentage of votes also possibly reflects the true popularity of the UR. In other words, high values could indicate both a lack of political competitiveness and citizens’ approval of the UR. In order to overcome this problem, a dummy variable is created and added in the second model. If the percentage of votes for UR is above 70 percent, this is labeled as ‘high’, which means that a region is considered non-competitive. It is important to note that the higher the value of ‘votes for UR’, the lower the level of subnational political competition. Therefore, negative beta-coefficients in the regression table indicate higher political competitiveness. Reversely, positive beta-coefficients indicate that political competitiveness decreases when the independent variable increases.
We now operationalize the independent variables, starting with natural resources. As Ross (2015: 242) describes, there are many ways to measure resource endowment. However, one of the most useful measures is also the most challenging to acquire, namely the precise government revenues from the extractive sector. According to Ross, (2015: 242), these revenues are collected and hidden in a variety of ways and therefore extremely hard to determine. By the lack of the precise government revenues, in our attempt to operationalize natural resources we follow Dunning (2008). Dunning looks at the ratio of resource rents to gross domestic product to define a country’s resource-dependency (Dunning, 2008: 16). Applying this logic to the subnational level, natural resources are operationalized as the total amount of oil and gas
extraction per region (oil_extraction, gas_extraction) as a part of the gross regional product (GRP). We use the values for the years 2003, 2007 and 2011 because the Duma elections took place in those years. In the dataset, oil is displayed in thousands of tons and gas in millions of cubic meters. We transferred these values into barrels of oil and gallons of gas, the most common measuring units. Then, we calculated the total amount of oil and/or gas in US Dollars for every region, using the average price of oil and gas separately for every year determined by the World Bank. We converted this amount to rubles to determine the net value of resources per region, using the average USD to ruble currency for every year. We used yearly averages for currencies and prices because oil and gas prices as well as currencies can fluctuate enormously throughout the year. Therefore, taking one fixed month could skew the results. Then, in order to calculate resource dependency, we accumulated oil and gas rents and divided this through the gross regional product. Since the distribution of oil and gas extraction is skewed, we logged the final variable. This has possible implications for the highest values of the original variable, since these are ‘artificially’ lowered in order to fit them into a linear regression.
We now operationalize the other independent variables. Foreign direct investment is operationalized as the total inflow of FDI in thousands of USD (reg_fdi_incl_total). It is assumed that the more efficient the executive authorities, the more control they can gain over economic resources. Therefore, control of economic resources is operationalized as the efficiency of executive power (efftotal). Regional wealth is operationalized as the Gross Regional Product (reg_grp). Education is operationalized as ‘general level of efficiency in the sphere’ (educ). Regional bureaucracy is operationalized as the number of employees in the legislative department of the regional bureaucracy (reg_psexec_fed). The size of the urban population is operationalized as share_urbanpop, a variable that combines the yearly average of the residential population in the region (reg_urbanpop), divided by the yearly average of the total population (reg_pop). Table 3 shows all previously discussed variables, their operationalization and their expected effect on the dependent variable.
Table 3. The independent variables discussed in this thesis and their expected effect on the dependent variable. Independent variables and
operationalization Expected effect on dependent variable (subnational political competition) Oil and gas dependence
(LOGoilgas_dependence) None, since rents taxed away Foreign direct investment (reg_fdi_incl_total) -
Education (educ) +
Control over economic resources (efftotal) -
Gross regional product (reg_grp) -
Bureaucracy (reg_psexec_fed) -
Analysis
As introduced in the research design section, this section contains firstly a descriptive analysis of how subnational variation in political competitiveness in Russia emerged and the context to which the phenomenon can be understood. Then, it shows a table that supports the hypothesis that the subnational resource curse is not convenient to explain subnational variance in Russia. Subsequently, a regression analysis is used to investigate some other potential determinants of subnational variation.
According to Hale (2013), path-dependent structures of the Soviet-era are still extant in Russia in the sense that Soviet legacies of (among others) political economy and ethnicity still influence regional ‘machine politics’ (Hale, 2013: 244, 258). These inherited structures appear robust despite formal institutional changes in the political environment. Besides this structural view, Hale stresses that local political ‘actors had a great deal of transformational power of their own (…)’, therewith elucidating the interplay between path-dependent structures and agency-based behavior.
The collapse of the Soviet Union in 1991 propelled the decentralization and democratization of politics and governance at the national as well as on the subnational level (Reisinger, 2013: 27). In the 1990’s, a national democracy with authoritarian features co-existed with decentralized subnational authoritarian regimes, the situation Gibson (2005) calls regime juxtaposition. Gel’man (2010: 23) describes the ways in which subnational authoritarianism could evolve after the breakdown of the Soviet Union and the subsequent period of decentralization as follows: ‘asymmetry among various territories of the country, depending on their administrative status, level of socioeconomic development, pattern of settlement, sectoral structure and a number of other historically evolved factors, gave rise to the effect of departmentalism and localism in local administration, which were to play an important role in the post-Soviet period’.
Besides decentralization, the evolution of subnational regimes was also fostered by Yeltsins policy of privatization, which created more autonomy for regional incumbents to increase their control over local political institutions and important local resources (Hale, 2013: 230). Summarily, the transition from decades of centralized authoritarian rule to decentralized and -in a way- democratizing rule inherently generated subnational regimes (Reisinger, 2013: 27; Hale, 2013: 230).
The 2000s were characterized by Putin’s strategies of recentralization to regain control over the regions (Reisinger, 2013: 1, 27). According to Golosov, during this time the federal center served as a ‘weak national superstructure built upon a variety of subnational authoritarian regimes’, revealing Russia’s failure to democratize after the Soviet Union and so Putin’s petite commitment to democracy (Golosov, 2008: 631, 626). Rather than nurturing democracy, Putin used recentralization policies in order to replace the regions under tighter control of the federal centre (Golosov, 2008: 626). One concession he made was to give regions relative autonomy in exchange for political loyalty and the deliverance of votes in national elections. Gubernatorial elections were eliminated in 2004, to be re-established in 2012. As stated by Saikkonen (2016: 451), up until now these regional elections have not threatened the central state’s authority in the Russian regions. Thus, as Golosov fairly states: ‘centre-periphery relations lie at the core of electoral authoritarianism in Russia’ (Golosov, 2008: 637-638).
In 2008, the managed democracy, a democracy with elections that does not provide any other element of democracy like rule of law, political accountability or bureaucratic integrity, lapsed into an electoral authoritarian state under Putin (Golosov, 2008: 623). According to
Levitsky and Way (2002), electoral authoritarian regimes ‘do not meet minimum standards of free and fair elections, but non-free elections nevertheless remain the principal source of regime legitimacy’. To guarantee its strength, such a regime resorts to practices of manipulation, fraud or more refined techniques like the appointment of local incumbents (Golosov, 2008: 623; Reisinger, 2013). The subsequent period can be characterized as the consolidation of authoritarianism (see Gelman, 2011: 35).
This examination of the emergence of subnational political regimes demonstrates that on the national as well as the subnational level, democratization is not a linear process and that transitions from autocracy to democracy cannot be considered as a two-dimensional ‘bridge’. Moreover, the previous section has shown that post-Soviet path-dependent structures together with individual agency of political actors co-create variation in degrees of democracy over time as well as in space.
This section addresses the supposed Russian subnational resource curse. Table 4, based on table 1 in the appendix, shows how much of the resource rents is left for each province. This gives us an insight into the extent to which subnational incumbents are able to monopolize resource rents. The table shows that the regional share of mineral rents decreases over time for both oil and gas: the region can keep 5 percent of their oil revenues for itself in 2005, while gas rents are already completely taxed away in 2005. In 2011, both oil and gas rents are transferred to the central government. Hence, nothing is left for the province. Consequently, to get back to our hypothesis, we find support for the idea that the subnational resource curse does not explain variance in political competition in Russia. It could still be possible that previously accrued rents remained at the region’s disposal. The following regression could settle this presumption.
Now we have shown that mineral rents are indeed taxed away we look at what else could explain subnational variance in political competition. Table 5 presents estimates from the OLS models on subnational political competition. The syntax is available upon request. The first model includes two dummies for the years 2007 and 2011 and the variable oil and gas dependence. The second model includes the dummy for popularity of United Russia (70% cutoff). In both the models, resource dependency is not significant. This offers support for the hypothesis that if resource rents are taxed away –what they are-, then there is no relationship between resource wealth and political competitiveness. Therefore, we can assume that the subnational resource curse does not explain subnational variance in Russia. These findings also reject the idea that remaining rents in the region still account for variance since resource-rich regions do not appear to be less competitive than other regions. However, it is interesting that the standardized beta-coefficient is negative (-.084), which implies that the higher the resource dependency, the lower the percentage of votes for United Russia and the higher political competition if the effect were significant.
Table 5. Some determinants of subnational regime variation. 1 2 3 4 5 Dummy2007 .000*** (.733) .001*** (.336) .260 (-1.134) .000*** (.588) .003* (.246) .000*** (.486) .263 (-.084) .000*** (.613) Dummy2011 .076+ (.145) .000** (.128) Dummy popularity .000*** (.429) .222 .000*** (.644) .010** .000*** (.900) .600 Log_oilgasdepen dence (-.098) .006** (-.434) .022* (-.540) .866 GRP Bureaucracy (.261) .148 (-.125) (.394) .157 (-.269) (-.059) .789 (.112) Share urbanpop FDI Education Efftotal .006** (-.206) .597 (-.063) .344 (.128) .437 (.153) .978 (-.009) .526 (.121) .644 (-.104) Adjusted R2 .412 .634 .696 .717 .720 N 79 79 75 38 21
Source: social and economic database created by the International Center for the Study of Institutions and Development. +p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001.
The third model includes GRP, bureaucracy and the share of the urban population. The model includes 75 of 83 regions. GRP and urban population are significant at the 0.01 level. The
standardized beta-coefficient of GRP is positive (.261), indicating that when GRP increases, the amount of votes for UR increases too thus the level of political competitiveness decreases. This offers support for the idea that regional wealth could foster the regime’s authoritarian capacity.
The share of the urban population relative to the total population shows a standardized beta-coefficient of -2.06. This suggests a positive relation between urban population and political competitiveness, since the amount of votes for UR decreases when the relative urban population increases. This relationship aligns with the predicted relationship and insinuates that more urbanized areas are less susceptible to authoritarianism. The adjusted R2 of the third model is
.696. This means that 69,6% of the variance in the dependent variable is explained by the independent variables. Such a high value could indicate that some variables (partly) measure each other. The discussion will address this concern.
The fourth model includes foreign direct investment. We see that the number of observations drops from 75 to 38, probably because of the limited amount of data for some regions. This has important implications for the validity of the model because such a small number of observations can probably not reflect a random sample of all 83 regions. In other words, the 38 regions that are left are probably not randomly divided and therefore the results are skewed. This could be the reason that resource dependency is significant in this model. The fifth model includes all variables. We observe that N has dropped even more, to 21 observations. This makes it nearly impossible to make inferences from this model. Therefore, we cannot say anything about the effects of FDI, education and control over economic resources on subnational political competition. The discussion will go deeper into this concern.
Concluding, the analysis of the share of mineral taxes and the regression provides preliminary support for the hypothesis that the subnational resource curse is not a convenient explanation for subnational variance in political competition.
Discussion
The previous section firstly elucidated how subnational variance in political competitiveness emerged in Russia. Path-dependent structures, together with the agency of subnational incumbents, made post-Soviet Russia susceptible to subnational regime variation. Then, the analysis addressed the subnational resource curse. The main finding of this thesis is that the Russian government is able to tax away regional resource rents, providing support for the notion that subnational incumbents cannot monopolize the resource rents. Moreover, the regression showed no significant relationship between mineral dependence and political competition. These findings align with the hypothesis that mineral wealth does not explain subnational authoritarianism in Russia. To get back to the research question, this thesis argues that the subnational resource curse is not a valuable explanation for subnational variance in Russia. Considering the other tested explanations, the gross regional product and the relative size of urban population seem to influence the degree of subnational political competition, too. Other noteworthy possibly influential factors are discussed later in this section.
It is important to note that the findings do not imply that the resource curse has no explanatory value: only in the context of Russia, where the central government succeeds in taxing away resource rents, it does not account for variation. Then to what extent are the findings in this thesis generalizable to other contexts? It seems reasonable to suppose that the mediating effect of tax relationships between the center and the regions influence the subnational resource curse in other countries as well. Diaz-Rioseco (2016) looks for example at the intervening role of fiscal institutions in rent-sharing and political competitiveness in the provinces of Argentina. This indicates that the findings are useful to consider not only within the context of resource-rich autocracies, but also in the context of resource-rich democracies. However, an important condition is that the country has a high level of decentralization. Hence, Russia can serve as a contextual example for decentralized, resource-rich countries.
Since this thesis does not pretend to deliver an exhausting explanatory model of variance in subnational political competition, it focuses only on a fraction of the possible indicators. Below, a range of other potentially noteworthy suggestions for further research is presented.
Ethnicity, a post-Soviet legacy discussed in the analysis, is also an interesting and potentially influential topic for further research. Due to the lack of useful indicators in the dataset, this topic could not be addressed in this thesis. It can be argued that non-Russian ethnic groups are more susceptible to bureaucratic pressure and demands from the local authorities because of their dependence on these local authorities. As Bader and Van Ham show, electoral fraud appears more often in ‘ethnic’ regions than in the more ‘Russian’ regions (Bader & Van Ham, 2015: 518).
Another interesting topic for further research is the level of linkage and leverage to the West, already discussed in the literature review. This contains particularly external factors. This could be researched qualitatively, for example by looking at the amount of NGO’s available in regions and the degree to which they are free to generate civil initiatives.
Within the same context of leverage and linkage, more geographically oriented explanations are worth investigating too. Lankina & Getachew (2008) for example, argue that post-communist regions that lie closer to the West are more prone to democratic diffusion. The reason for this is the exposure to democratic values through ‘assistance from Western agencies’.
Another fascinating dynamic is the use of internet in the Russian regions. According to Ruijgrok (2016), internet has facilitated the manifestation of demonstrations in authoritarian
regimes due to the vast spread of information. Bearing in mind the numerous discoveries of fake news in Russia and the protests that emerged in the past years, greater access to internet in Russia’s regions could change people’s attitudes towards the local as well as the central government.
Concerning the methodological part of this thesis, the little amount of available data of the Russian regions posed a major limitation. Nevertheless, this means that one has to make the best of the data that is available. This thesis has done its best effort to do this. Yet, the construct validity of the operationalization of the dependent variable cannot be guaranteed, which could have influenced the regression outcomes. Doubtful construct validity means that the operationalization of the concepts in the mechanisms does not cover precisely what it should cover. One reason why the choice for this dependent variable (percentage of votes for United Russia in the Duma elections) could have skewed the results is because the percentage of votes for United Russia is extremely high due to envisioned sanctions if regions do not provide enough votes (see Bader & Van Ham, 2015).
Also, it seems likely that the elections are manipulated in an electoral autocracy. A look into the election report of the Office for Democratic Institutions and Human Rights (OCSE) for the election of the state Duma in 2011 supports this notion. International observers report for example the following events: ‘ballot box stuffing, partiality of most media and the undue interference of state authorities on different levels’ (OCSE, 2012: 1). Yet, taking into consideration the reduced availability of recent data, the current operationalization of subnational political competitiveness is still preferred over more elaborate but much less recent operationalizations.
Unfortunately, the effects of foreign direct investment, education and bureaucracy could not be analyzed in the regression because the number of observations was too small to make inferences from it. In sum, the lack of recent subnational data provokes substantive limitations for research into this topic. This stresses the need for subnational data collection, not only in Russia but also in the rest of the world.
The high value of the adjusted R2 (.634 in the second model) in the regression suggests
that a large part of the variance in the dependent variable is explained by the independent variables. This is a serious indication that the operationalization of the independent variables is could benefit from more available data, too. It could namely be the case that the independent variables partly measure each other. The gross regional product for example is used in the calculation of resource dependence and as a control variable at the same time, which could partially explain the high value of the adjusted R2.
The ordinary least-squares regression presented in this thesis assumes the linearity of the included variables. This can be problematic because the linearity cannot be guaranteed. The findings are restricted in the sense that they do not imply causation because no elaborate robustness check is executed. Further research is needed to assure the robustness of the findings.
Another factor that could have skewed the results is the possibility of spatial autocorrelation in countries. A possible spillover effect from one region to another seems imaginable (Vidyattama, 2014: 378). In Russia specifically, where oil fields are not randomly divided over the country and trespass regional boundaries, surrounding regions could benefit from other region’s economies.
Conclusion
This thesis can be considered as a preliminary investigation into the nebulous universe of subnational political competition in resource-rich and decentralized countries in the world. It provides support for the idea that the subnational resource curse is not a convenient explanation for variance in subnational political competition in the Russian Federation because resource rents are taxed away by the central government. Therefore, subnational incumbents cannot monopolize these rents.
The gross regional product and the relative share of the urban population are suggested to account for subnational variation. Besides, a range of other potentially important explanations is addressed in this thesis. Among these indicators are ethnicity and the level of leverage and linkage to the West.
On a concluding note, this thesis and most of all its limitations, shows the importance of further research into the topic of subnational political competition and the factors that account for autocratization and democratization. A reasonable next step would be more subnational data collection in decentralized countries in the world.
References
Alexeev, M. & Chernyavskiy, A. (2015). Taxation of natural resources and economic growth in Russia’s regions. Economic Systems, 39: 317-338.
Alvarez, M., Jose, A., Cheibub, F.L., Prezworski, A. (1996). Classifying Political Regimes. Studies in Comparative Development, 31(2): 3-36.
Bader, M. & Van Ham, C. (2015). What explains regional variation in election fraud? Evidence from Russia: a research note. Post-Soviet Affairs, 31(6): 514-528.
Bak, D. & Moon, C. (2016). Foreign Direct Investment and Authoritarian Stability. Comparative Political Studies, 49(14): 1998-2037.
Beblawi, H. (1987). The Rentier State: Nation, State and the Integratio of the Arab World. London: Croom Helm.
Chaudhry, K.A. (1997). The Price of Wealth: Economies and Institutions in the Middle East. Ithaca, NY: Cornell University Press.
Diaz-Rioseco, D. (2016). Blessing and Curse: Oil and Subnational Politics in the Argentine Provinces. Comparative Political Studies, 49(14): 1930-1964.
Dunning, T. (2008). Crude Democracy: natural resource wealth and political regimes. New York: Cambridge University Press.
Gause, F.G. (1994). Oil Monarchies: Domestic Security Challenges in the Arab Gulf States. New York: Counc. Foreign Relations Press.
Gel’man, V. (2010). The Dynamics of Subnational Authoritarianism. Russian Politics and Law, 48(2): 7-26.
Gervasoni, C. (2010b). Measuring Variance in Subnational Regimes: Results from an Expert-Based Operationalization of Democracy in the Argentine Provinces. Journal of Politics in Latin America, 2(2): 13-52.
Gibson, E.L. (2005). Boundary Control: Subnational Authoritarianism in Democratic Countries. World Politics, 58(1): 101-132.
Gilley, B. (2010). Democratic enclaves in authoritarian regimes. Democratization, 17(3): 389-415. Golosov, G.V. (2011). The Regional Roots of Electoral Authoritarianism in Russia. Europe-Asia
Studies, 63(4): 623-639.
Hale, H.E. (2003). Explaining Machine Politics in Russia’s Regions: Economy, Ethnicity, and Legacy. Post-Soviet Affairs, 19(3): 228-263.
Huntington, S. (1991). Democracy’s Third Wave. Journal of Democracy, 2(2): 12-34.
Lankina, T.V. & Getachew, L. (2008). The Virtuous Circles of Western Exposure in Russian Regions: A Case for Micro-Polity Analysis of Democratic Change. Journal of Communist Studies and Transition Politics, 24(3): 338-364.
Levitsky, S. R. & Way, L.A. (2002). The Rise of Competitive Authoritarianism. Journal of Democracy, 13(2): 51-65.
Levitsky, S. R. & Way, L.A. (2006). Linkage versus Leverage: rethinking the International Dimension of Regime Change. Comparative Politics, 38(4): 379-400. mo
Levitsky, S. R. & Way, L.A. (2012). Beyond Patronage: Violent Struggle, Ruling Party Cohesion, and Authoritarian Durability. Perspectives on Politics, 10(4): 869-889.
Libman, A. (2013). Natural resources and national economic performance: Does sub-national democracy matter? Energy Economics, 37: 82–99.
Lieberman, E.S. (2005). Nested Analysis as a Mixed-Method Strategy for Comparative Research. American Political Science Review, 99(3): 435-452.
Mahdavy, H. (1970). Studies in Economic History of the Middle East. London: Oxford University Press.
McMann, K.M., Gerring, J., Maguire, M.,Coppedge, M. & Lindberg, S.I. (2016). Governing Countries: A Theory of Subnational Regime Variation. V-Dem institute: University of Gothenburg.
Monaldi, F. (2002a). Rent-Seeking, Institutions, and Commitment: The Political Economy of Foreign Investment in the Venezuelan Oil Industry. Ph.D. Dissertation, Department of Political Science, Stanford University.
Obydenkova, A. & Libman, A. (2013). National autocratization and the survival of sub-national democracy: Evidence from Russia’s parliamentary elections of 2011. Acta Politica, 48(4): 459-489.
Prezworski, A. (1991). Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. Cambridge: Cambridge University Press.
Reisinger, W.M. (2013). Russia’s Regions and Comparative Subnational Politics. London: Routledge. Ross, M.L. (2001). Does Oil Hinder Democracy? World Politics, 53(3): 325-361.
Ross, M.L. (2015). What have we learned about the resource curse? Annual Review of Political Science, 18: 239-259.
Russian Federation, State Duma Elections, 4 December 2011: Final Report. (2012). Organization for Security and Co-operation in Europe. Retrieved from
http://www.osce.org/odihr/86959?download=true
Ruijgrok, K. (2016). From the web to the streets: internet and protests under authoritatian regimes. Democratization: 1-23.
Saikkonen, I.A.L. (2016). Variation in subnational electoral authoritarianism: evidence from the Russian Federation. Democratization, 23(3), 437–458.
Sanborn, H. & Thyne, C.L. (2014). Learning Democracy: Education and the Fall of Authoritarian Regimes. British Journal of Political Science, 44(4): 773-797.
Sidel, J. T. (2014). Economic foundations of subnational authoritarianism: insights and evidence from qualitative and quantitative research. Democratization, 21(1): 161-184.
Snyder, R. (2001). Scaling Down: The Subnational Comparative Method. Studies in Comparative International Development, 36(1): 93-110.
Susarov, A. (2012). Mineral Extraction Tax (MET). Russian Survey: retrieved from http://www.russian-survey.com/brief-news/129-mineral-extraction-tax-met Vandewalle, D. (1998). Libya Since Independence: Oil and State-Building. Ithaca, NY: Cornell
University Press.
Vanhanen, T. (2000). A New Dataset for Measuring Democracy, 1810-1998. Journal of Peace Research, 37(2): 251-265.
Vasilyeva, O. & Libman, A. (2016). Resource Curse in Subnational Hegemonic and Competitive Authoritarian Regimes. Retrieved from https://ssrn.com/abstract=2895438 or
http://dx.doi.org/10.2139/ssrn.2895438
Vidyattama, Y. (2014). Issues in applying spatial autocorrelation on Indonesia’s provincial income growth analysis. Australasian Journal of Regional Studies, 20(2): 375-402.
Waller, Z. (2016). Tax Changes in Russian Oil and Gas. Enerpo Journal: retrieved from https://enerpojournal.com/2016/04/20/tax-changes-in-russian-oil-and-gas/
Petroleum Through Time: What is Petroleum? (2015). Retrieved from
http://www.aapg.org/about/petroleum-geology/petroleum-through-time/what-is-petroleum
Zubarevich, N. (2016). Four Russias: the new political reality. Open Democracy: retrieved from https://www.opendemocracy.net/od-russia/natalia-zubarevich/four-russias-new-political-reality
