the various players may nonetheless mean that one oligopolist carries more weight than
the other. Again, regions with tight links with the more powerful player would be
favored. The third case is that of a (partly) segmented duopoly (or oligopoly), where
‘markets’ for IFI lending are not fully integrated. In this context, even relatively weak
shareholders may nonetheless carry considerable weight in a specific region (‘market’).
These models offer different predictions as to the pattern of lending. The monopoly
model would predict that only the links with the dominant shareholder affect the pattern
of lending. By contrast, the collusive oligopoly model would hold that a region’s access
to lending will be determined by the aggregate - trade or financial - exposure of the main
shareholders. Even in this model, though, if one shareholder carries more political weight
than the others, its exposure will disproportionately affect the distribution of lending.
Finally, in the segmented market case, different shareholders may carry different
influence depending on the lending region being considered. These are all testable
implications.
With these considerations in mind, we first look at the main links between major
shareholders of the World Bank and IMF and regions to which lending goes. In Figures
6-8 we report the weights of different regions in US and EU exports. We see how exports
to Latin America are much more significant for the US than for the EU, with the gap
increasing in the aftermath of the NAFTA agreement. The opposite holds for Africa.
Finally, the shares of US and EU exports to Asia are pretty similar, with the EU catching
up with the US in the more recent years. As shown by Figure 8, Asia carries a
disproportionate weight in Japan’s exports. If Japan is to play at least a regional role, it is
therefore in Asia where this would be most visible.
Financial links are reported in Figures 9-11. We take as a measure of such linkages the
ratio to the lending country’s GDP of total consolidated foreign claims of the respective
banking sector toward a given region. An alternative, and possibly superior measure,
would have normalized regional foreign claims with their aggregate value. Unfortunately,
our source - the Bank of International Settlements (BIS) - provides comprehensive
information on the aggregate value of consolidated foreign claims only from 1999
onwards. Still, our measure offers some interesting insights. The patterns largely reflect
those observed for trade. For Africa, in particular, the EU plays a predominant role in
finance as well as in trade. However, banking exposure toward Latin America is not
substantially different between the US and the EU during the 1980s, with the EU
however taking the lead in the 1990s15. Differences in the level of exposure between the
US and the EU are also much less marked in Asia. Interestingly enough, though, US
banks have been substantially more successful in reducing their exposure to this region in
the aftermath of the recent crisis. Moreover, as highlighted by Figure 11, Asia also plays
a key role in the case of Japan and this may have therefore affected the pattern of IFI
disbursements toward the region.
15 European banks played a leading role in the banking consolidation process in Latin America during the
nineties and in the financing of privatization.
12