Compared with the end-June 1997 data, the latest BIS consolidated statistics underline the movements of funds which took place at the height of the Asian crisis. In particular, the further increases in the claims on Latin American and Eastern European countries point to limited contagion outside Asia. At the same time, the reduction recorded vis-à-vis the major countries caught in the Asian currency turmoil was smaller than shown by the BIS quarterly statistics. There are three basic reasons for this discrepancy. First, banking funds to individual countries were well sustained until the very last stage of the crisis, with changes between the two half-yearly reporting dates possibly masking sharper swings within the period. Thus, while flows to Thailand (the first country to be engulfed by the crisis) were reversed in the early summer, the shift occurred later in other countries. Secondly, the retrenchment by banks often involved a trimming of local representations, with the associate d cut in inter-office accounts being netted out in the consolidated data. Thirdly, the narrower coverage in the semi-annual reporting system of the cross-border exposure of non-Japanese Asian banks means that the extent of the pull-out of banks from the region is likely to have been underestimated.
As could be expected, the shifts taking place in international banking flows in the second half of 1997 were more strongly felt in the area of short-term lending. Indeed, an analysis of the maturity composition of the claims of reporting banks shows a decline in the weight of the short-term component in the case of Asia and a renewed increase in Latin America (but not in Eastern Europe). However, the effect of non-renewal of genuine short-term credit lines in Asia on the maturity distribution may have been masked by the sizable volume of long-term debt falling due in 1998. Thus, despite the 2 percentage point decline in the relative importance of the short-term category for the region, the share of the claims of "less than one year" for Asia remained on average above 60% at the end of 1997, compared with 55% for Latin America and 43% for Eastern Europe (although with large disparities between individual countries in all three instances).
The extent of the liquidity squeeze on Asian banking systems is better captured by the sectoral composition of reported claims, which indicates that the decline vis-à-vis entities located in Asia was more than accounted for by local banking counterparties, despite the steady flows of funds to Chinese banks. There was, as a result, a further shift towards direct lending to the private non-bank sector (to 52% of outstanding claims at the end of 1997, a figure comparable to that for Latin America). While partly related to concerns about the soundness of local banks, the growing proportion of loans to the non-bank sector raises particular questions concerning the prevention and resolution of financial crises, as the recent experiences in Asia have highlighted.
Finally, the nationality composition of reporting banks' international claims reveals important differences in the behaviour of major lending groups in the second half of 1997. Thus, whereas Japanese and North American banks reduced their outstanding claims on emerging economies in Asia, their European counterparts continued to increase their presence in the region, albeit more selectively and with a possible reassessment in the course of the reporting period. There was, as a result, a further rise in the concentration of on-balance-sheet exposure towards European banks, which reached 47% at the end of the year. German, French and UK banks accounted for 13%, 11% and 8% of the total respectively. This compares with 30% for Japanese banks and 10% for North American banks. At the same time, European banks stepped up their activity in Latin America and Eastern Europe, increasing in the process their share of reporting banks' claims on the two regions (to 62% and 80% respectively at the end of 1997). However, these overall figures conceal differences in the degree of concentration within individual nationality groups, in the weight of guaranteed claims and in the volume of off-balance-sheet exposure, which suggests a need for prudence in interpreting the data.