Coups, revolutions and efficient policies in autocracies Mario Gilli a , Yuan Li b, ⁎ a Department of Economics, University of Milan-Bicocca, Piazza dell'Ateneo Nuovo, 1, Milan, Italy b Institute of East-Asian Studies and Mercator School of Management, University of Duisburg-Essen, LE 738, Forsthausweg 2, Duisburg, Germany article info abstract Article history: Received 15 July 2014 Received in revised form 27 April 2015 Accepted 27 April 2015 Available online 2 May 2015 The purpose of this paper is to explore the interaction of two mechanisms that might constrain the power of dictators: the threat of a coup by the selectorate and a revolution by citizens. Our results help explain a stylized fact, namely that autocracies are far more likely to be either the best or the worst performers in terms of growth and public goods policies. To this end, we focus on account- ability within dictatorships using a model where both the selectorate and the citizens are the principals and the autocrat is the agent. Our results highlight that both excessively strong and excessively weak dictators lead to poor economic performances, and that a balanced distribution of de facto political power is required to incentivize the dictator to choose efficient economic policies. © 2015 Elsevier B.V. All rights reserved. JEL classification: D02 H11 D74 Keywords: Autocracy Accountability Coup Revolution 1. Introduction A striking fact is that there are many economically efficient autocracies and many very inefficient autocracies (Besley and Kudamatsu, 2008). A few examples illustrate this point. By 1975, Spain's per capita income was eight times as large as it had been in 1945; China's per capita income has increased 16 times from 1979 to the present; Malaysia, Singapore, Taiwan, and South Korea achieved growth rates of over 10% per year when under the control of dictators. However, some of the worst economic catastrophes also occurred under dictatorial regimes. Zambia witnessed its average income fall from 1964 to 1991; the economic disaster in North Korea led to millions of people suffering starvation; in Zaire, the economy collapsed after Mobutu seized power. More generally, over 20% of all observations of autocratic countries show negative growth rates, and during the tenure of a single dictator, the annual growth rates range from -11.85 to +25.03% (Rodrik, 1997; Almeida and Ferreira, 2002; Gandhi, 2008; Jones and Olken, 2005). This stylized fact calls for an explanation that we believe can be developed by analysing how political arrangements shape policy outcomes within autocracies. In our previous work, we examined how threats of coups by the selectorate (Gilli and Li, 2013) or of revolutions by the citizens (Gilli and Li, 2014) incentivize dictators to choose policies that lead to highly heterogeneous economic performance. In this paper, we provide a more complete, detailed account of the interaction of these two threats and of the influence that they have on policy choices. Coups simply reshuffle “deck chairs” within the elite, whereas revolutions entail dramatic regime changes. Regime change is the worst possible outcome for dictators and the selectorate. In this new strategic setting, the selectorate decides to engage in a coup by considering the effects this decision will have on the probability of a revolution. Such strategic interplay between coups European Journal of Political Economy 39 (2015) 109–124 ⁎ Corresponding author. Tel.: +49 203 3792314. E-mail addresses: mario.gilli@unimib.it (M. Gilli), yuan.li@uni-due.de (Y. Li). http://dx.doi.org/10.1016/j.ejpoleco.2015.04.011 0176-2680/© 2015 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect European Journal of Political Economy journal homepage: www.elsevier.com/locate/ejpe