On the geography of international banking : the role of third-country effects

International banking is a complex phenomenon. Among its determinants, distance has been found to be critical. But does distance only have a simple negative direct effect? Or is the role of geography more intricate? Applying spatial analysis techniques on BIS data of bank foreign claims in 178 countries in 2006, evidence of positive spatial autocorrelation under alternative spatial weights schemes is brought to light. The geographical aspects of international banking are further explored by a spatial autoregressive gravity model. The results obtained support that the operation of a spatial lag leads to important indirect or third-country effects. Evidence of such financial spillovers is further corroborated by results of a spatial autoregressive Tobit model. Geography is more important than the effect of distance on its own would suggest. Third-country effects operate in a manner that subsequently connects countries through links beyond those immediately involved in borrowing (destination) and lending (origin) relationships. Confirming earlier results, the economic size of sending and recipient countries, cultural similarity and inphase business cycles enhance international banking, while distance and exchange rate volatility hinder it. Also, while lower political risk has a positive role, so do higher financial and economic risks, reflecting-to some extent-some of the reasons behind the current financial crisis