From democratization to political liberalization: formalizing, operationalizing and testing political regime choice and change

Over a decade has gone by since the renaissance of autocratic studies, which greatly contributed to refining the knowledge accumulated from the democratization studies literature. However, almost all contributions within the field have focused on classifying autocracies along "autocratic types" (e.g. placing countries into categories that comprise civil, military, royal or personalistic dictatorships), while close to none asked whether we should be interested in explaining variation along political liberalization levels. I propose a measure of political regime openness which is correlated but significantly different from indexes of democracy. I then produce a tripartite classification of the political space linked to a theoretical model (dividing countries between closed autocracies, open autocracies, and democracies), and empirically test the correlates of these liberalization types and pro- and anti-liberalization transitions in 154 countries of the world between 1970 and 2007. I produce multiple tripartite classifications according to different political regime openness thresholds, and show that empirical findings are highly robust to different conceptualizations of this latent variable. To assuage the problems of high missingness levels in a host of theoretically-relevant correlates, I employ known but seldom-used empirical methods (multiple imputations). This allows me to include a much higher number of regressors without resorting to listwise deletion or to bias-inducing single/"expert" imputation techniques. At the same time, by incorporating the uncertainty around missing data instead of doing away with it, this allows me to retain only very significant findings and to do away with the rest. Empirical findings show that, after controlling for relevant correlates, no political liberalization or autocratic retrenchment (time-clustering) "wave" occurs within the sample period. On the contrary, spatial autocorrelation and regime duration are key to understanding autocratic and democratic resilience. In terms of substantive socio-economic correlates, I find that inequality levels have a crucial role to play in determining political liberalization levels at the early stages of a polity, but tend to become irrelevant later on. I am also able to adjudicate between long-standing theories of inequality and democracy, finding in favor of those arguing that higher liberalization levels should be found at average, not low, levels of inequality. My findings also show that the modernization theory should be problematized: while countries with higher levels of economic well-being tend to be more politically liberal, the same higher levels of economic well-being tend to tip the balance in favor of regime resilience in general. In autocracies, this process is liable to open up a rift between the preferences of the population and those of the governing elite(s). This is most evident in non-rentier countries, where governing actors have less leeway to buy out consensus, and is also currently applicable to hydrocarbon-dependent countries who have to cope with low international prices. These findings are relevant to scholars and policymakers alike. First, they suggest that the international community should think twice about supporting democratization attempts, which may have a higher likelihood of failing whenever structural conditions are not favorable in the first place. Sometimes, political liberalizations within autocracies could be preferred. Second, while GDP growth is correlated with changes in political liberalization levels, in fact economic recessions are only found to increase the likelihood of autocratic retrenchment. Policymakers arguing in favor of economic sanctions are advised to take into account that the most likely outcome is for the existing government – or for a different government succeeding it – to veer towards a less liberal political regime. Current sanctions against Russia and Syria (not to mention North Korea) may be a case in point. Third, findings emphasize the role of agency. Structural relations within a polity do not perfectly determine political regime choice and change, and the same trend can have controversial effects over the likelihood of choosing a more liberal political regime and the likelihood of transitioning in the first place. Therefore, whenever policymakers and the international community are determined into “nudging” a country into a more liberal political regime, they are advised to concentrate over specific actors – e.g. governments in exile, oppositions within and outside the country, the military, etc. – and work with them towards a common objective. The bottom line is: while economic sanctions appear to be the easiest way to coordinate international action against an illiberal government, broader political action, while more costly, has a higher chance of succeeding in the longer run. Fourth, political regime transitions are associated with higher degrees of political violence. Whenever politicians, or the public opinion, side in favour of regime change in illiberal countries, they should be aware that this may not come at a small cost in terms of human lives lost or infrastructural damage. As conflicts in Syria and Libya that are dragging on since the 2011 Arab Spring show, attempted transitions may degenerate into longstanding violent confrontations. Finally, as Egypt’s autocratic retrenchment shows, any regime transition “resets the clock” of a regime’s durability, making it more likely to change again within the next few years. At the same time, both Egypt on the one hand and Tunisia on the other are evidence that, aside from differing structural conditions, the main actors of a polity are the ultimate source of political regime choice and change.