This paper examines the structural determinants of output volatility in developing countries, and especially the roles of geography and institutions. We investigate the volatility effects of market access, climate variability, the geographic predisposition to trade, and various measures of institutional quality. We find an especially important role for market access: remote countries are more likely to have undiversified exports and to experience greater volatility in output growth. Our results are based on Bayesian methods that allow us to address formally the problem of model uncertainty and to examine robustness across a wide range of specifications.
Download the final version via Science Direct
This is a revised version of University of Bristol discussion paper no. 05/568. The working paper version contains more details on our use of Bayesian methods in the paper.