Environmental Challenges and Political Regime Transition: The Role of Historical Legacies and the European Union in Eurasia

ABSTRACT

This paper addresses the impact of political regimes on carbon emissions (CO2) from a historical perspective, covering the period of 1970–2015. Our sample consists of 153 countries that we follow for 20 years before the transition (1970–1990) and 26 years after it (1990–2015). During transition, post-Communist states experienced a decline in CO2, when controlling for economic variables. Transition had a positive impact on the environment only in cases when it was associated with democratization. When transition has led to consolidation of autocratic regimes, CO2 has increased.

Introduction

The studies on political regimes and their implications for environmental challenges and policies are generally inconclusive and point in different directions (e.g., Fredriksson and Neumayer 2013; Fredriksson and Wollscheid 2007; Wagner and Schneider 2006).¹ Some studies have argued that democracies are generally more efficient in ratification of international environmental agreements, in decentralization policies that allow for more efficient governance of commons, and in their accountability to the public, among other issues (Libman and Obydenkova 2014; Neumayer 2002; Obydenkova and Salahodjaev 2016, 2017; Ostrom 1990, 2012; Sanwal and Zheng 2018; Wagner and Schneider 2006). In the words of Eric Neumayer (2002): “Strong evidence is found that democracies sign and ratify more multilateral environmental agreements, participate in more environmental intergovernmental organizations, comply better with reporting requirements under the Convention on International Trade in Endangered Species of Fauna and Flora, put a greater percentage of their land area under protections status, are more likely to have a National Council on Sustainable Development in their country and have more environmentally relevant information available than non-democracies.”

However, there is still limited evidence on how political regime transition affects environmental changes in general and carbon emissions in particular. The modernization theory of democracy posits that democratization is strongly associated with urbanization and economic growth, among other factors (Przeworski et al. 2000; Przeworski and Limongi 1993; Rodrik and Wacziarg 2005). Arguably, urbanization, economic growth and population are the three most important factors accountable for the increase in CO2 (Jiang et al. 2016; Khondaker et al. 2015; Roth et al. 2019). This paper addresses, from a historical perspective, how political regimes and transition affected carbon emissions in post-Communist states, covering the long period of 1970–2015.² Our sample consists of 153 countries that we follow for 46 years—20 years before the start of the transition, from 1970, and 26 years after it, until 2015. The project uncovers the historical dimension of CO2 emissions during the period of communism and the bipolar world of 1970–1980 and depicts a trend of steady increase in carbon emissions across Communist states throughout the 1970s and 1980s.

The analysis presented in this paper aspires to provide a bigger picture of the implications of communism and post-Communist political regime transition for one of the main environmental challenges. The paper aims to provide the necessary wider context on the nexus of (post-)Communist states and the implications of their political regimes for the environment both chronologically and geographically. This analysis paves the way to a better understanding of follow-up discussions in this special issue on the triangle of democratization, environmental challenges, and the role of international organizations (democracies-composed and those led by autocracies) in post-Communist states.

Our findings indicate that the post-Communist democratization process decreases carbon emissions, after controlling for changes in income that capture economic decline along with other controls. An increase in the level of democracy was still associated with a decrease in CO2 despite increasing urbanization indicative of a higher quality of life and despite increases in economic growth and population after the end of the economic decline in the 1980s–1990s. The paper also confirms the logic of the environmental Kuznets curve, which is highly important for understanding the environmental challenges of post-Communist transition paralleled by transition to a market economy and an increase in income. According to the Kuznets curve, an initial increase in per capita income increases carbon emissions per capita until the per capita income reaches a certain point, after which further economic growth is associated with a decrease in CO2 emissions per capita.

The analysis controls for a number of economic, socio-cultural, geographic, and political factors and additionally builds on studies of the “democracy stock.” More importantly, within this special issue, the paper provides for a larger historical and geographic overview of the post-Communist states in terms of their dynamic in CO2 emissions over the decades and within a larger global context. When proceeding to more nuanced analysis of specific international organizations as actors in environmental politics and their implications for post-Communist states,³ it is important to have a clear picture of the nexus of political regimes and environmental challenges in Eurasia. This paper aspires to provide a wider global and historical context, within which further mapping of specific international organizations and their implications for sustainable development may be better understood. To this end, the paper proceeds as follows. The next section outlines the main theoretical issues in studies of democratization, sustainable development, and post-communism. The third presents data, empirical models, and descriptive statistics. The fourth section depicts the main empirical results and the final section concludes.

Post-Communist Democratization and the Challenge of Carbon Emissions

Debates on the role and effects of political factors for sustainable development have been growing fast since the collapse of communism (Andonova and Sun 2019; Bernauer and Koubi 2009; Bhattarai and Hammig 2001; Buitenzorgy and Mol 2011). However, the findings remain contradictory. On the one hand, within non-democratic regions, the ruling elite controls mass media and manipulates public opinion. The latter leads to public ignorance of global environmental concerns and poor environmental performance (if any). As the result of the lack of accountability of non-democracies to national and international publics, they may limit themselves to merely mimicking climate rhetoric (Ambrosio, Hall, and Obydenkova 2021; Hall, Lenz, and Obydenkova 2021; Ward, Cao, and Mukherjee 2014; Poberezhskaya 2015; Poberezhskaya 2018). Within the context of a bipolar world and Cold War competition for dominance, economic development was a clear priority for the USSR-dominated Communist states (on historical legacies, see Beissinger and Kotkin 2014; Beissinger 2002; Pop-Eleches and Tucker 2017). Therefore, our first hypothesis is the following:

Hypothesis 1: During the period of communism (Cold War) 1970–1990, CO2 emissions per capita were increasing across all post-Communist states converging to the level of consolidated democracies.

In contrast to a malicious image of autocracies as actors harmful for the environment, a number of studies on democratization have pointed out positive implications of transition in terms of promotion of environmental policies, adjusting state-level priorities, and increasing public awareness of the environmental agenda (Andersen 2002; Bragina et al. 2015; Kolk and van der Weij 1998). Indeed, the so-called “fourth wave” of democratization was influenced by international formal institutions and specifically the EU (Morlino 2011; Levitsky and Way 2010; Obydenkova 2012; Mišić and Obydenkova 2021). In contrast, some studies have argued that state capacity in autocracies is associated with higher pollution (Ward, Cao, and Mukherjee 2014); that autocracies tend to ignore public opinion on environmental conflicts (Demchuk et al. 2021); and that cooperation with an autocratic donor may undermine the priority of sustainable development (Obydenkova et al. 2021). To sum up, it is only within a climate of political openness, transparency, independent mass media, and governmental accountability to the public, is the collaboration of national and international actors efficient in terms of advancing the goals of sustainable development. All of these aspects of democratization are linked directly to more efficient management of climate policies and increased accountability of the government on these and other related matters. Within post-Communist states, a number of the EU-related leverages were applied to push environmental issues forward on the governmental agenda (e.g., Environmental Action Programme for Central and Eastern Europe; New Environmental Policy Instruments) (Kolk and van der Weij 1998; Andonova and Tuta 2014; Tews, Busch, and Helge 2003). Based on these numerous studies, we formulate our next two hypotheses as follows:

Hypothesis 2: Starting from the regime transition (from 1990s onwards) and along the transition path, the post-Communist states decreased CO2 emission per capita relative to the consolidated democracies.

Hypothesis 3: Post-Communist democratization has directly improved the environment by reducing CO2 emissions per capita (accounting for increased urbanization, population growth, and quality of life) and indirectly via EU membership. EU membership had a separate pathway to a reduction in CO2 emission per capita than democratization.

In general, and at the global level, democracies are indeed associated with political freedoms, civil liberties, freedom of information, pro-environmental civic awareness and pro-environmental legislation, and the rule of law—factors that all together are likely to improve environmental policy and decrease environmental degradation (e.g., Neumayer 2002; Ostrom 1990; Obydenkova and Salahodjaev 2016, 2017; Ward et al. 2014). The positive effect of democracy and quality of political institutions on a variety of environmental concerns was been shown to hold in the developing world and at the global level in Asia, South America, and Africa (see Aşıcı 2013; Bhattarai and Hammig 2001; Donohoe 2003).

On the other hand, there is some space for potential theoretical contradictions between democratization and specifically CO2 emissions. The modernization theory of democratization links political transition to urbanization, education, improving living standards, accessibility of goods, increase in travel and tourism—that is, higher quality of life. The higher quality of life has multiple indicators, many of which, if not all, relate to urban life style, increase in public transportation, ownership of individual vehicles, more frequent use of air flights and their accessibility, and increase in consumption associated with a market economy and better income, to name a few (e.g., Jiang et al. 2016; Saikawa 2013). Studies of the modernization theory of democratization, in fact, posit that democratization and modernization (proxied by urbanization, economic growth, and education) are highly interconnected and develop hand-in-hand during political regime transition and marketization (Aixala and Fabro 2009; Doucouliagos and Ulubasoglu 2008; Papaioannou and Siourouins 2008; Przeworski et al. 2000; Przeworski and Limongi 1993; Rodrik and Wacziarg 2005). While these studies debate the nexus of democratization, urbanization, and economic development, they still agree that economic growth and development are strongly associated with democratization and democracy as a political regime.

However, a different set of studies on the main drivers of CO2 emissions, outlined above, clearly indicate that economic factors (economic growth and development) are among the main causes associated with high CO2 emissions. Indeed, urbanization, rising income, and economic growth are among the main drivers of carbon emissions as multiple studies demonstrate (e.g., Khondaker et al. 2015; Roth et al. 2019; Sanwal and Zheng 2018). Moreover, the post-Communist regime transition was accompanied by a deep economic crisis throughout the 1980s and 1990s. It led to the closure of many industries in the region, which may also be associated with the decrease in CO2 emissions. Testing this potential contradiction between democratization and environmental challenges within the context of communism and post-communism over the larger period of analysis of 1970– 2015 aspires to provide better understanding of the nexus of political regimes, transition, and implications for the environment. While the impact of democracy as such has received some attention in the literature on sustainable development, the role of regime transition remains on the agenda.

Within these contradictory debates, one may expect to find a direct relation between the level of democracy and environmental indicators. Following Mikael Skou Andersen (2002), weak democracies cannot impose environmental policies in the early stages of democratization. Nonetheless, as the level of democratization increases, the marginal social benefit of environmental policies increases enforcement of such policies by society. In sum, based on the previous literature, we shall proceed to operationalize and test our hypotheses.

Data and Methods

In this section, we first discuss the primary data sources and descriptive statistics for the key variables used in the analysis. Then, we turn our attention to the measure of democracy and explore the main outcome variable’s trends for the post-Communist states against the consolidated democracies. Finally, we elaborate on the empirical modeling based on the patterns observed in the initial descriptive analysis.

Table 1 provides information on the key dependent and independent variables used in the analysis and the data sources. Our sample consists of 153 countries that we follow for 46 years—20 years before the start of the transition, from 1970, and 26 years after, up to 2015. We compile our sample using a rich set of sources: the main outcome variable, CO2 emission per capita, is from the World Bank Indicators; the independent variables are from various sources: Polity IV Project, World Religion Project, GeoDist Database, Barro-Lee Educational Attainment Dataset, and Ethnic Factorization Data (see Table 1). Descriptive statistics for the key variables can be found in Appendix 1.

Table 1. Datasets Used in the Analysis of Quality of Life

Data source name	Years	Variables
World development indicators	1970–2015	CO2 emission per capita, Urbanization rate, GDP per capita
Polity IV project	1970–2015	Polity 2 democracy Index, Institutionalized democracy score
World religion project	1990–2015	Christian majority indicator, Muslim majority indicator, % of Catholics
GeoDist database	1990	Area, Indicator of landlocked, Latitude, Longitude, Indicators of British, French, Russian, Ottoman colonies, Distance to Equator, Distance to Greenwich
Barro-Lee education attainment dataset	1990	Education (in years)
Ethnic fractionalization data	2003	Ethnic fractionalization

Democracy Measure

In this study we employ a measure of democracy widely utilized in the literature, the Polity 2 index. The measure is made available by the Center for Systematic Peace as part of the Polity IV project, which codes all countries’ democratic characteristics on a longitudinal basis. The cumulative index measures competitive and open executive recruitment, the existence of institutionalized constraints on chief executives, the guarantees of civil liberties, and the political participation of citizens. The index is computed by differencing the institutionalized autocracy score, A_t, from the institutionalized democracy score, D_t:

(1) $Polity{2}_{\mathrm{t}}=D_t-A_t.$(1)

Both institutionalized scores are constructed in additive eleven-point scales (0–10). Thus, the Polity 2 index ranges from −10 (hereditary monarchy) to +10 (consolidated democracy). The final score for any particular country represents three political regimes: “autocracies” with the score ranging from −10 to −6, “anocracies” with the score ranging from −5 to +5, and “democracies” with the score ranging from +6 to +10. For example, in 2007, Russia had an institutionalized democracy score of 6, and an institutionalized autocracy score of 1 bringing Russia’s overall democracy index to five, which categorized Russia’s political regime as “anocracy.”

There are five essential elements of the Polity 2 index. The first element measures the competitiveness of executive recruitment, which reveals how the existing modes of advancement create equal opportunities for an ordinary citizen to advance to a super-ordinate position. The second element assesses the country’s openness to executive recruitment. If the executive recruitment is “open,” then any politically active citizen has an opportunity to clinch the chief executive position through the lawfully determined process. The third element measures the extent to which executive constraints limit the decision-making power of the state’s chief executive. This element assesses the existence of any checks and balances embedded in the various parts of the chief executive’s decision-making. The fourth element is the competitiveness of participation, which measures evidence of the embedding of alternative policy preferences and leadership in the political system. The final element assesses the extent of regulation of citizens’ political activities.

We first analyze the institutionalized democratic score changes over time for the post-Communist and the counterpart countries in Figures 1 and 2. The average institutionalized democracy score patterns are similar to the patterns observed for the average democracy index. The counterpart, non–post-Communist states scored higher in terms of democratization in the earlier years before the transition. The gap converged at the time of transition, and then the gap narrowed by 2000. After 2000, the gap is stable with an increasing democracy score for both types of countries.

Figure 1. Average democracy index and institutionalized democracy score, polity 2, in post-communist states against the rest of the world.

The above comparisons miss the vast heterogeneity in the democracy index and institutionalized democracy score across the non–post-Communist states. It should be evident that European or North American states have paths to democracy that diverge from those for Sub-Saharan or Middle Eastern regions. To account for this heterogeneity, in Figure 2 we break down the non–post-Communist countries into seven regions: East Asia (EA), Europe (EU), Latin America & Caribbean (LA), Middle East & North Africa (MENA), North America (NA), South Asia (SA), and Sub-Saharan Africa (SSA). In the 1970s, the Communist states were considered among the least democratic societies, ranked last among eight types of countries. At the time of transition, in 1990, the post-Communist states’ position improved substantially. They were ranked sixth, surpassing the countries representing Sub-Saharan Africa and the Middle East & North African regions. Along the transition path, the post-Communist countries surpassed the countries representing the the East and South Asian regions, lagging behind only Latin American, European, and North American democracies. The same patterns, we can observe for the alternative measure: average institutionalized democracy score.

Figure 2. Polity 2 democracy index and institutionalize democracy score by region: 1970–2015 (East Asia (EA), Europe (EU), Latin America & Caribbean (LA), Middle East & North Africa (MENA), North America (NA), South Asia (SA), and sub-Saharan Africa (SSA).

Figure 2 suggests that the political liberalization processes that the post-Communist states experienced in the early 1990s had implications for other regions. First, for the countries of the Sub-Saharan and East Asian regions, we observe significant improvements in the democratization index and score. The USSR had a substantial influence on most countries’ political regimes, and the loosened grip in the 1990s created a jump-start for democratization in these regions. Latin American countries had also experienced substantial improvements in democracy, but the liberalization processes in the Latin American region started much earlier than in the countries of the Warsaw Bloc. Finally, the Middle Eastern and North African countries, regions where the USSR had substantial military and economic influence in the pre-transition period, experienced only slight improvements in the average democratization score or index after 1990. And as seen in the figure, the USSR’s dissolution had almost no impact on political rights and civil liberties in this region. Finally, this visual analysis identifies the list of countries we include in the control group in the difference-in-differences analysis. North American and European countries can be viewed as countries with consolidated democracies and further considered in our study as control group countries (Appendix 1A).

Democracy in Post-Communist States

Previous analyses of the institutionalized democracy score do not consider the significant variation in this score across the post-Communist states in the period of transition. In the 1980s, only the former republics of Yugoslavia had a democracy score different from zero. In the electoral processes of these states, certain democratic institutions were present before the USSR’s dissolution. Specifically, the chief executives faced some constraints on their executive powers. These states had between “slight to moderate limitations” and “substantial limitations” on the executive authority’s power. In 1990, the project’s experts downgraded the democratic score in these states to 0 because of persistent ethnic turmoil following the first years of political liberalization at the time of transition. The institutionalized democracy score significantly increased in other post-Communist states. For example, Hungary immediately reached the highest possible rating of 10 points; the Czech and Slovak republics representing a single country at that time, Czechoslovakia reached the rating of 8 points; and all the republics of USSR had the rating of 3 points. The average score was 3.3 points for the average post-Communist state, equivalent to the democratic environment in the Bloc’s largest country, the USSR. The variability in the institutionalized democracy score was 1.9 points, which is the difference between the upper and lower bounds of the 95 percent confidence interval.

Along the transition path, some countries stayed autocratic with an absence of any democratic processes (Uzbekistan, Turkmenistan), and some countries advanced immediately to the rank of consolidated democracies (Hungary). Therefore, it is important to distinguish between “political regime transition” and “democratization” (terms often used interchangeably in studies during the 1990s and early 2000s). Indeed, in some cases, the transition that took place resulted in consolidation of autocracy or a so-called hybrid regime (Levitsky and Way 2010). The variability in the democracy score was stable throughout the transition path, 2.6–3.0 points, while the average value increased from 3.3 points in 1990 to 6.5 points in 2015. The snapshot of the democracy stock in 2010 (Appendix 2A) shows that at least six countries had no signs of democratic institutions being embedded into their political regimes; these countries continued to be “autocracies”: Azerbaijan, Belarus, Bosnia, Kazakhstan, Turkmenistan, Uzbekistan.⁴ The same number of countries reached the rank of consolidated democracies: Hungary, Lithuania, Mongolia, Poland, Slovak Republic, and Slovenia. The average democracy score was 6.4 units, somewhere between the scores enjoyed by Ukraine and Georgia. The 95 percent confidence interval of the democracy score lay between 5 and 8, the scores experienced by Armenia and Latvia, respectively.

Democracy Stock

Several studies in the literature explore the impact of democratization on various political and economic outcomes using the stock of democracy (Wigley and Akkoyunlu-Wigley 2011; Besley and Kudamatsu 2006; Nazarov and Obydenkova 2020, 2021). This measure’s main advantage is that it explores the long-run impact of democratization, since democratic processes may not change society’s behavior and outcomes immediately. The previous analysis confirms the substantial variation in the democracy score across the post-Communist states after 1990. We construct the democracy stock starting from this year to the final year of the study period, 2015.

(2) $S_t=\sum _{i=1990}^{2015}{D}_i.$(2)

Our approach in the construction of democracy stock is somewhat different from the previously suggested measures in the literature. Instead of using the whole Polity 2 democracy index, we use the institutionalized democracy score of this index. Our previous longitudinal analysis shows that the Polity 2 index and institutionalized democracy score are highly correlated series, and since our main interest is to understand the impact of democratization on the various outcomes of the post-Communist countries, our analysis should be more accurate with the use of the democracy score instead of the whole index.

Previous studies document that even after three decades of political regime transition, Communist historical legacies have survived in a number of regions and substantial variation in the political regimes remains across the post-Communist countries (Beissinger and Kotkin 2014; Gel’man 2013, 2015, 2016; Giuliano 2006; Libman and Obydenkova 2019, 2020; Pop-Eleches and Tucker 2017). As of 2015, after consistent influence of the EU, Hungary, Lithuania, Slovenia, the Czech Republic, Poland, and the Slovak Republic turned into newly established consolidated democracies, with the maximum or near-maximum democracy stocks of 250 points accumulated during the study period (notwithstanding the recent wave of populism). In contrast, a few states “outside the EU’s reach”—Turkmenistan, Bosnia, and Uzbekistan—have accumulated zero stock during the same period. Other post-Soviet states exhibit a combination of a number of external autocracy-consolidation factors and/or the impact of natural resources, that prevented them from consolidating democratic institutions. For example, Kazakhstan, Azerbaijan, Tajikistan, and Belarus have some decorative and rhetorical democratic institutions embedded into their political systems, but they lag substantially behind the first group of post-Communist countries. The goal is to look at the consequences of transition for carbon emissions and to advance the understanding of the implications of fluctuation in democracy for the environment (CO2).

Descriptive Statistics

The average CO2 emission per capita in the study period for the average country is 4.6 tons; for the average post-Communist country it is 6.47 tons, and for the average consolidated democracy it is 9.58 tons. A further comparison of the outcome variable over time between Communist countries and consolidated democracies is presented in Figure 3 (Appendix 2 provides the list of countries comprising consolidated democracies).

Figure 3. CO2 emission per capita in post-communist and consolidated democracy.

Figure 3 demonstrates that the carbon emission in the consolidated democracies has been dissimilar to the post-Communist states during the last five decades. In 1970, CO2 emission per capita in the consolidated democracies was 3.3 tons higher. Because of the process of ecological degradation that the post-Communist states experienced in the pre-transition period, the two series converged before the start of transition at the level of 8.8 tons per capita. In the next decade, per capita CO2 in the consolidated democracies increased to 9.7 tons while it decreased in the post-Communist countries, leading to the most massive gap between two types of countries: 4.3 tons. In the final period of transition, the consolidated democracies experienced a 23 percent decline in CO2 emission per capita, or a 2.3 percent decline on an annual basis, reaching a low of 7.7 tons per capita in 2014. For the same period, the post-Communist states experienced an increase in per capita carbon gas emissions. Thus, if these trends continue, a new convergence of CO2 emission per capita should be anticipated between the two types of countries in the immediate future.

The previous descriptive analysis doesn’t reveal the differences in CO2 emissions across the post-Communist states. In Figure 4, we depict these differences; it shows that as of 2014, the highest level of CO2 emissions in the post-Communist states was in Estonia. The level of carbon emission per capita in this country was 14.9 tons. The lowest level was in Tajikistan, with the level of CO2 emission per capita close to 0.6 tons. Thus, Estonia had a 25 times higher level of air pollution than Tajikistan. The average level of carbon emission per capita across all post-Communist countries was 5.6 tons, which is close to the level experienced by the Slovak Republic.

Figure 4. Rankings of CO2 emission per capita among the post-communist states in 2014.

Empirical Models

Based on the varying differences in CO2 emission per capita depicted in Figure 3, we conduct a difference-in-differences analysis. This analysis enables us to investigate whether the post-Communist countries experienced different trends in CO2 emission before and during the period of transition relative to the consolidated states, after controlling for a rich set of socio-economic, institutional, and geographic factors. This analysis sets the path to test our first two hypotheses.

We run the difference-in-differences analysis by estimating the following regression equation:

(3) $Y_{it}={\theta }_1{T}_{1985-1989}+{\theta }_2{T}_{1990-1994}+{\theta }_3{T}_{1995-1999}+{\theta }_4{T}_{2000-2004}+{\theta }_5{T}_{2005-2009}+{\theta }_6{T}_{2010-2014}+X{ }_{it}^{\prime }\beta +I{ }_i^{\prime }\phi +G{ }_i^{\prime }\alpha +{\tau }_t+{\mu }_i+{\epsilon }_{it}$(3)

In the above model, ${\theta }_1-{\theta }_6$ measures the differences in the CO2 emissions per capita, $Y_{it}it$, in the post-Communist states relative to the consolidated democracies in five-year time intervals. θ₁ measures the difference in the five-year period prior to the start of transition, 1990; while θ₂−θ₆ measures the differences in the subsequent five periods of transition and what the size of these parameters suggests about differing trends in CO2 emissions in the post-Communist states relative to consolidated democracies. In the model, vector X_it controls for socio-economic differences across the states; it includes GDP per capita, its squared term, average years of education, and the population’s urbanization and growth rates. Vector I_i includes the institutional differences represented by the percentage of Catholics, an indicator of a Christian majority, and the ethnic factorization index. Vector G_i includes geographic controls, such as whether the country is landlocked and the distance to the equator and Greenwich, and land area in square kilometers. μ and τ are the post-Communist and year fixed effects.

Figure 5 suggests that the difference-in-differences analysis would only shed light on the indirect relationship between democracy and environment. The post-Communist countries have experienced dissimilar paths to democracy. Thus, an analysis that reveals heterogeneity in democratization is warranted to learn whether democracy itself is associated with CO2 emissions in the post-Communist countries.

Figure 5. CO2 emission per capita and democracy in post-communist states in 2014.

Taking this literature into consideration, we estimate the following regression model to test the third hypothesis.

(4) $Y_{it}=X{ }_{it}^{\prime }\beta +I{ }_i^{\prime }\phi +G{ }_i^{\prime }\alpha +{\gamma }_{1}D{S}_{it}+{\tau }_t+{\epsilon }_{it}$(4)

In this model, DS_it is the democracy stock for country i in year t and τ is the year fixed effects. The primary interest is the parameter γ₁. If its value is less than zero this is an evidence that democracy reduces CO2 emission. We perform two different robustness analyses in our study. First, we estimate Equation (4)(4) $Y_{it}=X{ }_{it}^{\prime }\beta +I{ }_i^{\prime }\phi +G{ }_i^{\prime }\alpha +{\gamma }_{1}D{S}_{it}+{\tau }_t+{\epsilon }_{it}$(4) with the additional control, EU membership and parameter γ₂ associated with it. In light of Hypothesis 3, this analysis enables us to quantify the mediating impact of EU membership on the democracy/environment gradient. Second, we estimate the model with the country-fixed effects, instead of vectors of institutional, I_i and geographic, G_i controls, to account for unobserved country-level heterogeneity not included in the given vectors.⁵ As shown in Figure 5, the relationship between democracy stock and CO2 emissions per capita for the post-Communist republics is barely negative. The autocratic states with zero accumulated democratic stock after 25 years of transition are on both ends of the spectrum. For example, autocratic Turkmenistan and Kazakhstan, in 2014, had the highest levels of carbon emissions among the transition economies, 12.5 and 14.4 tons respectively, while Uzbekistan and Azerbaijan are among the medium polluted countries with per capita emission of 3.42 and 3.92 tons, respectively.

The least polluted countries are Tajikistan and Moldova, 0.62 and 1.39 tons per capita. Despite their dissimilar levels of democratic development as of 2014, they exhibit significant differences in terms of democracy stock. That is, Tajikistan has democratic stock of only 11. In contrast, Moldova can be viewed more as a newly established democracy with a stock of 188 (the maximum possible value in 25 years of transition is 250).

Variations in the democratic stock cannot singly explain the differences in per capita CO2 emission across the post-Communist states. The impact of democracy on CO2 emissions may be confounded by various socio-economic, institutional, and geographic factors similar to Equation (3)(3) $Y_{it}={\theta }_1{T}_{1985-1989}+{\theta }_2{T}_{1990-1994}+{\theta }_3{T}_{1995-1999}+{\theta }_4{T}_{2000-2004}+{\theta }_5{T}_{2005-2009}+{\theta }_6{T}_{2010-2014}+X{ }_{it}^{\prime }\beta +I{ }_i^{\prime }\phi +G{ }_i^{\prime }\alpha +{\tau }_t+{\mu }_i+{\epsilon }_{it}$(3) . These factors may simultaneously influence the level of carbon emissions in the country on the one hand and its democratic stock after 25 years of transition, on the other hand.

The endogeneity problem due to the failure to control for all unobserved differences across post-Communist countries is a primary concern for drawing any definitive conclusions, especially based on the second analysis. The extensive controls would not allow one to account for all the possible differences that affect simultaneously CO2 emissions and processes of democratization. Thus, γ can be biased because of the presence of unobserved country-level factors. To address this endogeneity concern, in the model we also control for both time (μ) and country (π) fixed effects, omitting institutional and geographic factors from the model. The impacts of the latter factors on CO2 emission should be captured by the country fixed effects. Most importantly, the impacts of any unaccounted factors by institutional and geographic factors would be captured by country-level fixed effects, bringing the estimate of γ closer to its true value.

Results

Difference-in-Differences Analysis

In Figure 6, we depict the parameters from the difference-in-differences regression. First, in the pre-transition phase, 1985–1989, the average post-Communist state relative to the similar democratic state had a 1.53-ton higher CO2 emission per capita. This finding supports our first hypothesis. At the time of transition, 1990–1994, the difference decreased to 0.35 tons of CO2 with the higher emission per capita in the post-Communist states. We can further observe the reversal of the gap between the two types of states. In 2000–2004, the gap reaches a 1.75-ton per capita emission, which is less for the post-Communist states. In 2010–2014, the gap declined slightly, to 0.97 tons. The combination of these findings supports our second hypothesis.

Figure 6. Differences in CO2 emission between post-communist countries and consolidated democracies at transition and along the transition path (difference-in-differences approach).

The main implication from this analysis is that the post-Communist states experienced a higher level of ecological degradation before the transition. Democratic processes may have catalyzed the adoption of more environment-friendly policies during the period of transition. An alternative explanation could be that liberalization processes significantly affected the production of goods and services, inadvertently reducing CO2 emissions to the atmosphere. In the final period, we observe a slight increase in CO2 emissions in the post-Communist states relative to the counterpart states, probably, because of intensified economic growth experienced by many post-Communist states in the later period of transition; but still, the post-Communist states pollute less per capita than the consolidated democracies.

Democracy Stock Analysis

Main Results

In the preceding analysis, we provide evidence that the post-Communist states have experienced improvement in the environmental outcome relative to the consolidated democracies. The CO2 emissions per capita dropped substantially along the transition path, even though the post-Communist states had higher rates in the pre-transition period.

To account for a potential economic explanation of changes in CO2, in Table 2, in the first model, we control only for the year fixed effects and find that the relationship between the two variables is slightly positive. However, after adding the urbanization rate, population growth, income, and another proxy from the modernization theory of democratization—the education variables—the relationship switched to the negative sign. After adding environmental and geographic factors, the relationship decreased in magnitude but preserved its statistical significance at the 5 percent level. In Model 4, our findings indicate that a unit increase in democracy stock reduces CO2 emissions by 0.016 tons per citizen; or, one standard deviation increase in democracy stock (σ = 86) decreases the emission by 1.4 tons per citizen This gradient’s elasticity is 0.19 percent; thus, a one percent increase in democracy stock leads to a 0.19 percent decrease in carbon emissions. Suppose Turkmenistan, with zero democracy stock, has CO2 emissions per capita of 12.5 and turns into a fully democratic society, like Hungary, with a democracy stock of 250. In that case, the CO2 emissions per capita in Turkmenistan should drop to 8.5 tons, the level experienced by Poland and the Czech Republic. Thus, based on this series of regression models, we can conclude that democratization processes are associated with a decrease in carbon emissions in the post-Communist states. In contrast, from the historical perspective, before the start of transition, these countries were steadily heading toward environmental degradation in the period of 1970–1990; that trend was reversed by economic and political liberalization processes.

Table 2. Regression Results with Democracy Stock (Dependent Variable: CO2 Emissions Per Capita (Post-Communist States – 1990–2015))

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
Democracy Stock^a	0.006	−0.027***	−0.020**	−0.016**	−0.013**
	(0.011)	(0.011)	(0.008)	(0.007)	(0.007)
Income		0.641***	0.839***	0.792***	0.295**
		(0.232)	(0.193)	(0.185)	(0.116)
Income Sq.		−0.012	−0.020**	−0.019**	−0.007
		(0.009)	(0.008)	(0.007)	(0.005)
Urban population (% of total)		0.175***	0.188***	0.131***	0.162*
		(0.036)	(0.031)	(0.043)	(0.086)
Population growth, %		0.950***	0.749**	0.653**	0.391*
		(0.303)	(0.275)	(0.293)	(0.201)
Education			0.121	−0.031
			(0.278)	(0.275)
% of Catholics			−1.340	−1.403
			(1.939)	(1.800)
Christian majority			−1.950**	−1.810**
			(0.790)	(0.796)
Ethnic fractalization			2.713	1.457
			(3.576)	(3.160)
Distance from equator				1.078
				(1.053)
Landlocked				0.209
				(0.850)
Area in sq. kms				0.002**
				(0.001)
Constant	7.897***	−3.299*	−5.045	−5.921	−3.784
	(1.039)	(1.745)	(3.178)	(4.304)	(3.995)
Year dummies	Yes	Yes	Yes	Yes	Yes
Country dummies	No	No	No	No	Yes
General controls	No	Yes	Yes	Yes	Yes
Institutional controls	No	No	Yes	Yes	No
Geographical controls	No	No	No	Yes	No
Observations	663	662	662	662	662
R-squared	0.033	0.539	0.624	0.667	0.937

^aIn hypothesis 3, we test γ₁ < 0, we report the p-value for the one-tailed t-test for the coefficient associated with the democracy stock.

Robust standard errors in parentheses, *** p < .01, **p < .05, *p < .1.

One of our interests lies in quantifying the association of urbanization and population growth with carbon emissions. In all three models, Model 2–Model 4, we can observe strong and positive associations. In Model 4, the most preferable specification, a one percentage point increase in the urbanization rate increases the CO2 emission per capita by 0.13 tons and a one percentage point increase in population growth increases the emission of CO2 gases by 0.65 tons. Both results confirm the earlier conceptualization that higher rates of urbanization and population growth have detrimental impacts on the environment.

Our analysis of transition economies also confirms the existence of the environmental Kuznets curve, a concept widely discussed in the recent environmental literature. The inverted Kuznets curve theory argues that economic development increases environmental degradation up to a certain point, after which economic development improves environmental indicators (e.g., Yasin, Ahmad, and Chaudhary 2020). Based on Model 4, an increase in per capita income increases carbon emissions per capita until the per capita income reaches $22,000 per capita. Similar to the Kuznets approach, we find that a further increase in income reduces air pollution, thus decreasing environmental degradation with an increased living standard.

Only one country, Slovenia is on the other end of the Kuznets curve, with a negative relationship between income and CO2 emissions; the Czech Republic and Estonia have almost approached the flat part of the curve, while the rest of the post-Communist countries are in the rising part of the Kuznets curve. These countries would experience an increase in CO2 emissions per capita with an improved living standard. In fact, this historical perspective explains why some of the post-Communist new EU member-states challenged the EU’s Green Pact and the EU’s suggestions for a “Green exit from the pandemic.” Among the states arguing for not prioritizing the goals of sustainable development over short-run economic benefits were the Czech Republic and Poland.

Following the literature on historical legacies, we account for whether the state has a Christian majority (Pop-Eleches and Tucker 2017; Beissinger and Kotkin 2014; Beissinger 2002). Religion, as a proxy for historical legacy, seems to be another contributing factor related to environmental indicators. Even after controlling for economic and other factors, a Christian majority is associated with significantly lower levels of CO2. That is, a country with a Christian majority compared with a counterpart country has 1.8 ton less of carbon emission per capita, ceteris paribus. Suppose we compare Poland, a country with an 88 percent Catholic population, and thus a Christian majority, and Turkmenistan, a counterpart with no Catholic population. The difference in CO2 emissions per capita between the two countries is 5 tons per capita, and this institutional factor can explain about 1.8 tons, ceteris paribus.

The state’s size in terms of land area measured in square kilometers has a positive implication for environmental degradation. The larger states like Russia or Kazakhstan are associated with higher CO2 per capita, 11.85 and 14.36 per capita, respectively. Holding all else constant, any additional 1 m. square kilometers is associated with an additional 0.2 tons of CO2 emissions per capita.

Finally, despite the absence of statistical significance, the substantial economic significance makes the relationship between ethnic fracture and CO2 emission worth discussing. The results show that a country with no ethnic heterogeneity in the population will experience 1.4 tons more of carbon emissions per capita than a country with a 100 percent chance of two randomly drawn citizens being representatives of two different ethnic groups.

Robustness Check: EU Membership and Environment

EU membership may partially explain the results of the preceding analysis. The existing EU norms require more comprehensive ecological policies. Thus, the negative association between democracy stock and CO emissions can be partially explained by the positive association of EU membership with democratization (Appendix Table 4 demonstrates that the two are strongly correlated with r-0.7) and negative association of EU membership with CO2 emissions. While the democracy stock is at the center of our attention, it is still important to distill the potential impact of EU membership on combatting climate change and its mitigating impact on the relationship of interest. In Table 3, we report the results for the regression analysis where the main independent variable along with the democracy stock is the country’s EU membership status in the given year. The Eastern European states in our sample became EU members in different periods (see Table 5). Out of 29 states in our sample, by the end of the study period, only 11 states were EU members. Eight countries became members as early as in 2003. Romania and Bulgaria were invited in 2006, and Croatia entered the EU in 2012. The heterogeneity in the time of entry along with the presence of countries that have never entered the EU allows us to identify the effect of EU membership on our environmental indicator. First, regardless of the model specification, in Model 2 to Model 4, we observe a negative association between EU membership and CO2 emissions per capita. Model 4’s results indicate that EU membership may have reduced CO2 emission per capita by 1.282 tons, which is equivalent to the effect of 106 units of the democracy stock, or the democracy stock accumulated by Serbia and Montenegro by 2014. The democracy stock coefficient is lower with the EU membership control by 0.004 (−0.016 in Table 2 vs. −0.012 in Table 3), which is a 25 percent reduction in bias. This analysis confirms that even after controlling for EU membership, we end up with evidence that democratization may have directly improved the environment, albeit to a lesser extent.

Table 3. Regressions with EU Membership and Democracy Stock (Dependent Variable: CO2 Emissions Per Capita (Post-Communist States – 1990–2015))^b

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
Democracy stock^a	−0.003	−0.025**	−0.016**	−0.012*	−0.010
	(0.013)	(0.011)	(0.008)	(0.007)	(0.007)
EU membership^a	2.129*	−0.610	−1.247*	−1.282*	−0.972**
	(1.316)	(0.993)	(0.959)	(0.912)	(0.453)
Year dummies	Yes	Yes	Yes	Yes	Yes
Country dummies	No	No	No	No	Yes
General controls	No	Yes	Yes	Yes	Yes
Institutional controls	No	No	Yes	Yes	No
Geographical controls	No	No	No	Yes	No
Observations	663	662	662	662	662
R-squared	0.054	0.540	0.629	0.672	0.939

^aIn hypothesis 3, we test whether γ₁ < 0 and γ₂ < 0. As a result, we report the p-values for the one-tailed t-tests for the coefficients associated with the democracy stock and EU membership.

^bThe same sets of controls used in regressions as reported in Table 2 and discussed in Equation (4).

Robust standard errors in parentheses; ***p < .01, **p < .05, *p < .1.

Our finding goes along with previous studies, pointing out the importance of membership in the EU and a number of factors associated with it, such as, for example, transnational environmental networks and the importance of compliance with EU environmental norms, and the EU impact on the environmental agenda of post-Communist new member-states (e.g., Andonova and Tuta 2014).

Robustness Check: Endogeneity

The main drawback of Model 4’s results reported in Tables 2 and 3 is the endogeneity concern. Despite the extensive set of controls, the estimates may still be confounded by various factors not accounted for by the model. One quick solution to check the robustness of the received results for the significance of unobserved country-level factors is to account for the country-level fixed effects instead of country-level geographic and institutional controls. The results of the extended model are reported in the columns referred to as Model 5 in Tables 2 and 3.

First, our previous analysis demonstrates that the estimates for the parameter associated with the democracy stock are more unbiased if we control for EU membership. As a result, Model 5 reported in Table 3 is our preferred specification. Based on the R² statistic, this model fits the actual data much better than the previous model, 0.939 versus 0.672. However, the most important result is that the coefficient associated with the democracy stock has changed in magnitude. It has dropped from 0.012 to 0.010 with similar standard errors. The statistical significance at the 10 percent level for the one-tailed t-test isn’t preserved at the margin. Although the coefficient associated with EU membership has also dropped in magnitude from −1.282 to −0.972, a 25 percent reduction in bias, the estimate is still statistically significant at the 5 percent level (one-tailed t-test). This analysis demonstrates that, although democracy may have directly improved air quality, the greater implication for environmental improvements in the transition period is related to EU membership.

Conclusion

How did political regime transition affect the carbon emissions of 35 post-Communist states as compared to the rest of the world over the twentieth-to-twenty-first centuries? The paper covers 46 years, almost half a century, of changes in political regime, urbanization, and CO2 across the world, focusing on the phenomenon of post-Communist states. Our findings indicate that an increase in democracy stock reduces CO2 emissions by 0.010 tons per citizen, or a one standard deviation increase in democracy stock decreases emissions by 0.86 tons per citizen, controlling for changes in income and other variables. Even if this estimate is marginally statistically insignificant at the 10 percent level, we cannot rule out the fact that democracy has a direct impact on CO2 emission. The insufficient sample size for the non-EU members wouldn’t allow us to test confidently this effect independent of the EU. Thus, based on this series of regression models, we can conclude that the democratization process has been associated with a decrease in carbon emissions in the post-communist states (despite a simultaneous increase in urbanization and EU membership). However, the EU membership played a greater role in environmental improvements. This holds even after accounting for an increase in urbanization during the transition period. Democratization has been found to be positively associated with the urbanization rate, and a one standard deviation increase in the democracy stock increases the urbanization rate by 4.4 percent.

The paper also offers an important historical perspective analysis. The analysis demonstrates that from the 1970s onwards, before the start of the transition, the former Communist countries had steadily increasing CO2 emissions in the period 1970–1989. This trend was reversed by economic and political liberalization processes. While the literature on the consequences of communism keeps on growing, this study is among the first to uncover the environmental degradation associated with communism as a political, economic, and ideological system.

The analysis also sheds more light on the relationship between economic development and environmental degradation, known as the environmental Kuznets curve. In accordance with the Kuznets curve approach, we find that after an increase of income reaches a turning point, it reduces CO2 emissions, thus decreasing environmental degradation and increasing the living standard further. Among other findings, historical legacies (religion) demonstrated significant negative impact on CO2 during the period of the study. For example, after controlling for economic and other factors, Catholic population is associated with a significantly lower level of CO2. This finding should have important implications for the literature on historical legacies and the impact of culture on environment.

The project also has a few relevant secondary findings (some of which are analyzed in detail in other contributions of this special issue). For example, in line with other essays in this issue on democracies-led versus autocracies-driven international organizations, we find that membership in the European Union is strongly associated with a decrease in CO2 emissions in post-Communist states. In this paper, our main focus remains on the nexus of democracy and carbon emissions across post-Communist states; but it is important to consider a somewhat more general trend of the effects of EU membership on the environment across post-Communist states. It is a somewhat remarkable finding that EU membership seems to have been a more powerful trigger in decreasing carbon emissions as compared to democracy alone (confirming the studies in this special issue on the role of integration into environmental transnational networks, environmental conditionality of loans, and diffusion of environmental values through membership).

Also, along with the existing literature on the importance of geographic location for sustainable development, we find that landlocked countries experience a higher rate of CO2 emissions per capita than the counterpart countries, ceteris paribus. On average, the former countries experience 2.82 tons more carbon emissions per capita, holding all other variables constant. Central Asian countries’ remote location may make the adoption of environmentally conscious production methods cost-effective. Moreover, the landlocked states are more dependent on their neighboring states across a number of issues and for connectedness with the outside world. They are thus also more susceptible to the influence of, for example, neighboring China, whose policy has been criticized for lack of sensitivity toward the goals and principles of sustainable development (e.g., Ambrosio, Hall, and Obydenkova 2021; Hall, Lenz, and Obydenkova 2021; Obydenkova et al. 2021). To sum up, the paper has placed post-Communist states within a global context and traced their development for almost half a century in terms of the nexus of democracy stock and changes in carbon emissions. Our analysis of the nexus of regime transition and CO2 across post-Communist states historically and globally contributes to a better understanding of the role and effects of international organizations and the sustainable development challenges of post-Communist states.

Acknowledgments

Anastassia Obydenkova thanks the Basic Research Program of the National Research University Higher School of Economics (HSE University) for supporting her research. She is grateful to the Center for the Institutional Studies of the National Research University Higher School of Economics and to the Institute for Economic Analysis of the Spanish Council for Scientific Research (IAE-CSIC) for hosting her as a research fellow and affiliate. The findings of this paper were presented at Columbia University in New York (2017), Universidad Autonoma de Barcelona (2017), and New York University (2018). We thank the participants of these events for their insightful comments and suggestions. We are especially grateful to Joshua Tucker, Ada Ferrer-i-Carbonell, and Elise Giuliano for their great feedback on this project. The authors thank two anonymous reviewers of this paper for their excellent comments and insights. The authors of this paper are listed alphabetically and they contributed equally to this project.

Disclosure Statement

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

Supplementary Material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/10758216.2021.1995437.

Notes

1. The authors of this paper are listed alphabetically and they contributed equally to this project.

2. While the literature on historical legacies of communism has been growing in the last decade, the impact of communism on the environmental agenda has not yet been addressed as the focus. On historical legacies, see (Beissinger and Kotkin 2014; Lankina, Libman, and Obydenkova 2016b; Pop-Eleches and Tucker 2017).

3. On democracies-driven international organizations and their implications, see Morlino (2011), Lankina, Libman, and Obydenkova (2016a), Lavelle (2011, 2007, 2003), Pillarisetti (2005), Arpino and Obydenkova (2020), Obydenkova and Arpino (2018), Obydenkova and Rodrigues Vieira (2020); on autocracies-driven international organizations, see Debre and Dijkstra (2020), Debre (2020a, 2020b), Libman and Obydenkova (2013), Libman and Obydenkova (2018a, 2018b, 2021). Other studies also detail the rhetoric of climate change policies developed by non-democracies (e.g., Poberezhskaya 2014, 2018).

4. Within the focus of this special issue, it is important to point to the potential role of membership of post-Soviet states in non-democratic regional international organizations such as the Shanghai Cooperation Organization (see Ambrosio 2008, 2012, Izotov and Obydenkova 2021). Analysis of the impact of such membership on sustainable development is beyond this paper. It should, however, remain on the agenda for future studies.

5. In the Appendix, we have more detailed discussions of empirical strategies used to test the study’s hypotheses. On the format of the presentation of equations, also see Angrist and Pischke (2009).