Democracy, diversification and growth reversals

There exists convincing evidence that democratic countries are less volatile. This conclusion is usually reached with respect to volatility as measured by the standard deviation of annual growth rates of per capita GDP which includes both low and high frequency fluctuations. However, recently several studies have documented that significant changes in trend-growth, such as growth accelerations that last a decade or similar periods of negative growth, are quite common. In this paper we ask whether democracy also has a stabilizing effect on trend-volatility, i.e. whether more democratic countries experience fewer and milder swings of trend-growth. We find a common phenomenon medium term reversals of growth, that is periods of exceptionally high growth are, on average, followed by periods of exceptionally low growth, and vice versa. The propensity to experience large swings of trend growth is not uniform across countries – less democratic countries are more susceptible to it. When compared with factors commonly associated with volatility such as measures of quality of institutions, macroeconomic policies and financial development, we find that democracy is the most robust predictor of a country’s propensity for growth reversals. We construct a model in which non-democracies have high barriers of entry for new firms. This leads to less sectoral diversification: fewer sectors are operated but more resources are channeled to each. In a