Dual-track Power-sharing and Authoritarian Compensation: a Synthetic Comparative Case Study of China's 2001 Government Restructuring

Abstract: This paper investigates empirically how dictators compensate losers of reform among their constituents by conducting a synthetic comparative study on China's government restructuring in 2001. Students of dictatorships have long noticed the importance of power-sharing arrangements or patronage distribution in authoritarian politics. From this perspective, when dictators implement the reforms that alter the initial power-sharing relationship between them and their supporters, the authoritarian stability will also be disrupted. As the voluminous literature on political economy of reforms has shown, we see various kinds of dual-track solutions where, in addition to the track where reforms are implemented, a non-reform one is created at the same time for (potential) losers of them to be compensated. Drawing upon an important conceptual distinction between office- and rent-seeking politicians in the political economy literature (Persson and Tabellini, 2000), I argue that, these two sets of motivation---i.e., offices and rents---can actually be alternatively interpreted as two different "powers" in the context of full authoritarianism where both are centrally controlled by dictators. Through a comparative case study on China's 2001 government restructuring, I show empirically that dictators can not only share the power-as-offices with a certain number of elites to solicit their cooperation in policy, but also compensate the remaining losers of the competition for offices with the power-as-rents. As for the empirical design, this study leverages the fact that, during the restructuring, the bifurcated development among the pillar industries forms a clean distinction between treated and untreated sectors when the government reform is viewed as a treatment. I not only employed a synthetic control method for estimating the effect of the treatment, but also conducted a placebo test to make sure that the findings were robust.