How Much Should We Trust the Dictator's GDP Growth Estimates?

I study the overstatement of GDP growth in autocratic regimes by comparing the self-reported GDP figures to the night time lights (NTL) recorded by satellites from outer space. I show that the NTL elasticity of GDP is systematically larger in more authoritarian regimes. This autocracy gradient in the elasticity is robust to multiple changes in data sources, econometric specification or sample composition and is not explained by potential differences in a large set of country characteristics. The gradient is larger when the incentive to exaggerate economic growth is stronger or when the constraints on such exaggeration are weaker. The results suggest that autocracies overstate yearly GDP growth by as much as 35%. Adjusting the GDP data for the manipulation taking place in autocracies leads to a more nuanced view on the economic success of non-democracies in recent decades and affects our understanding of the effect of changes to foreign aid inflows on income per capita