Aid and Capital Flight

Abstract

Aid is now small relative to the international movements of private capital to and from low-income countries. Although most attention has been given to foreign private capital inflows, for some regions outflows of domestically owned capital have been more substantial. In this paper we investigate whether aid affects capital flight. The issue relates to the important matter of whether aid is ‘scaled up ’ by private movements in capital. This is usually seen in the context of whether aid attracts or displaces foreign private capital. Aid will be ‘scaled up ’ if it induces domestic residents to retain a greater proportion of their wealth within the country, and it will be offset and potentially vitiated, if it induces a greater proportion of private wealth to be held outside the country. In the theoretical part we argue that even the sign of any link from aid to capital flight cannot be determined a priori. In the context of private portfolio choice we suggest that there are six distinct and sometimes opposing routes by which aid might affect portfolio choices. Using a global panel data set, spanning 48 aid recipient countries over 1970-98, we investigate the relationship empirically. We allow for the potential endogeneity of aid and find aid substantially reduces capital flight. As a consequence, aid is ‘scaled up ’ by the induced decisions of domestic wealth holders. In the short term additional aid has substantial effects on the portfolio, being matched more than dollar-for-dollar by induced reductions in capital flight. In the longer term such changes in the portfolio are only maintained if aid itself is maintained for a substantial period, perhaps a decade. We tentatively suggest that taking a long view each dollar of aid might be scaled up by around twenty to forty cents of induced domestic investment that would otherwise have left the country as capital flight. 2 1.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 24 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.