Institutions and Growth: Time Series Evidence from Natural Experiments

Documenting the long term impact of institutions on economic performance has generated tremendous interest in the development literature. Contemporary or intermediate term effects of institutions over time are difficult to establish, however, since institutions seldom change significantly in the short run. In addition, accepted instruments that control for endogeneity of institutions in cross sections are inappropriate for time series analysis. In this paper we utilize an eleven year panel of 26 countries with sufficient institutional variation to identify large and significant short and intermediate effects of institutions. To control for endogeneity, we utilize the hierarchy of institutions hypothesis and find that it holds strong explanatory power. A 10 percent change in institutional quality towards OECD standards is shown to raise annual growth by 3.5 percent. In discriminating between short run and intermediate term effects, we can also document that early reformers reap the greatest benefits, but that it is never too late to begin institutional reform. *We thank Sascha Becker, Christa Heinz, Stephan Klasen, Chris Papageorgiou, Charles Nelson, Richard Startz, Farid Toubal, Steve Turnovsky, and especially John Temple for helpful comments. Any remaining errors are our own. Theo Eicher thanks the German Science Foundation for financial support. … i