Measuring Idiosyncratic Risk: Implications for Capital Flows

This paper offers two refinements of the traditional risk measure based on the volatility of growth. First, we condition GDP growth on structural characteristics of the host country that move only slowly and therefore can be partly predicted by an investor. Second, we adjust conditional risk for the systematic components due to the global and regional interdependence between alternative investment locations. The decomposition of conditional risk into its systematic and idiosyncratic components reveals that not only are African countries on average characterised by a larger conditional risk than Asian and Latin American countries, but the idiosyncratic risk factor also represents a larger share than in other developing countries. As a final contribution, we search the empirical literature on foreign direct investment and risk in order to determine which of the suggested risk measures provide the best description of idiosyncratic risk. Using a general-to-specific methodology, we find that both economic and political risk factors are important elements in the investment decision. We also find that commercial risk factors applied in the literature so far are poor determinants of idiosyncratic risk.