which are not reported here, are not statistically significant for the IMF, while for the
World Bank they indicate a redirection of lending away from Asia and Latin America and
toward developing Europe. This well correlates with a new emphasis on transition
economies by the Bank and the need to diversify the portfolio of lending. The previous
findings of a larger response to crises by the IMF and a more marginal role by the World
Bank are also confirmed in this more general specification.
Finally, we turn to the role of influential shareholders. In light of the previous results, we
focus our attention on the 1990-2001 period, to reflect the fact (and finding) that both the
patterns of lending of the two IFIs and their responses to the crises that emerged were not
constant over the two decades. We focus, therefore, on a shorter, but more homogeneous
period.
We take our first shot at what we have called the monopoly model. We first estimate a
simple equation that seek to capture the notion that the US may be the only influential
shareholder in both the World Bank and the International Monetary Fund. Only US
variables - the regional share of US exports and the ratio to GDP of US banking exposure
toward regions - are thus added to the right hand side of the previously estimated
equation that already included the crisis variable and the regional fixed effects. The
results are reported in the first column of Table 4.
The key finding is that the disbursements of the two Ibis are positively and significantly
related to both the indicators of US influence that we highlighted. However, one should
not read too much in the results derived from this equation, since it was not tested against
any alternative specification. In particular, one cannot infer from the findings obtained
here that the monopoly model with the United States as the dominant shareholder
adequately describes the pattern of IFI lending. Indeed, if we simply replaced all US
variables with the corresponding indicators for the European Union we would find again
positive and significant coefficients and would conclude that the EU plays a monopolist
role in the allocation of World Bank and IMF lending!
In order to discriminate among different hypotheses, we need to specify a more general
model that includes trade and financial exposure indicators for the United States, the
European Union, and Japan. This was done and estimated. The second column of Table 4
contains the results obtained from estimating such a model. They show a mixed pattern.
Financial indicators are not significant for the European Union; conversely, for Japan
trade variables have a wrong signed impact on the pattern of IFI lending. Only for the
United States, do both trade and financial exposure indicators show a significant, and
correctly signed, influence in determining the pattern of IFI lending. These results lend
themselves to a simple, but intriguing interpretation: the pattern of European influence
seems to be motivated mainly by mercantilistic considerations, with a view to propping
up export markets, while for the United states both trade and financial considerations
matter in motivating its actions within the IFIs. For Japan, financial considerations are
paramount. However, the finding that trade links have a negative impact on IFIs lending
is quite puzzling and hard to explain.
15