Flight Capital as a Portfolio Choice.

Among 51 countries studied, there are large regional differences in the proportion of private wealth held abroad, ranging from 3 percent in South Asia to 39 percent in Africa. Three variables explain capital flight in Africa: exchange rate overvaluation, adverse investor risk ratings, and high indebtedness. Collier, Hoeffler, and Pattillo examine flight capital in the context of portfolio choice. They estimate the stock of flight capital held abroad and compare it with the stock of real (nonfinancial) capital held within each country. For 51 countries, they construct estimates (as of 1990) of private domestic capital and flight capital - which combined add up to domestic wealth. There are large regional differences in the proportion of private wealth that is held abroad, ranging from 3 percent in South Asia to 39 percent in Africa. They explain differences in portfolio choice in terms of the capital to labor ratio, indebtedness, exchange rate distortions, and risk ratings-all proxies for differences in the risk-adjusted rate of return on capital. They then apply the results to four policy questions in which private portfolio choices are potentially important: the effect of the East Asian crisis on domestic capital outflows; spillovers; the effect of HIPC debt relief on capital repatriation; and why Africa has so much of its private wealth outside the continent. Their conclusions: The four most severely affected East Asian countries will eventually lose about $250 billion in domestic wealth as a result of the deterioration in risk between March 1997 and September 1998. They found some support for a spillover model. The effect of the HIPC debt relief initiative on capital repatriation will vary massively between HIPC-eligible countries. Africa has by far the lowest capital per worker, which makes massive capital flight from Africa all the more distinctive. Three variables explain capital flight in Africa: exchange rate overvaluation, adverse investor risk ratings, and high indebtedness. This paper-a joint product of the Development Research Group and the International Monetary Fund-is part of a larger effort to understand how growth in low income countries can be increased. Paul Collier may be contacted at pcollier@worldbank.org