This paper uses two different historical accounts of the occurrence of earthquakes to identify the effects of these shocks on aggregate economic outcomes. We find that the use of a widely popular dataset Emergency Events Database (EM-DAT) that records natural disasters restricted to occurrence with high level of damage points to statistical negative consequences of earthquakes on economic growth. Yet, these results do not hold when using a more comprehensive dataset from the United States Geological Survey (USGS) systematically recording earthquakes irrespective of the associated damage. The two results can be reconciled when isolating case of high-damage earthquakes in the context of poor countries often associated with weaker state capacity. These findings confirm the negative consequences of natural disasters' role on economic development in poor countries and highlight the importance of systematic data collection of natural disasters.