sign).8 In particular, while an increase in banks’ loans to assets, in short, long
and medium term bank debt to total bank debt lending and in undisbursed credit
commitments to total bank lending to that country lead to a rise in the probability
of reschedulings, increases in unallocated credit and in foreign exchange reserves
lead to a fall in this probability.
Lanoie and Lemarbre (1996) have used the same speci...cation of Lloyd-Ellis
et al. Estimating a cross-section of data covering 93 countries in the two years
1989 and 1990, they con.rm their results and also that balance sheet variables
outperform the two other sets of variables in explaining debt reschedulings.
Backer (1992) shows how as the prediction lag is lengthened, the signi...cance
of macro-variables (as the ratio between debtservice payments and ex ports, the ra-
tio between imports and reserves, the in≠ation rate, GDP, interest rates) improves
relative to that of the balancesheet data. This might suggest that macro-variables
are proxies for more fundamental, longer-term determinants of a country’s sol-
vency, while .nancial variables provide informati on about the country’s current
liquidity. More speci.cally, he uses a logit model to estimate the debt reschedul-
ing probability for 68 debtor countries, with semi-annual data from 1981(I) to
1988(I). He integrates balance sheet variables with macro-variables and discovers
that, while the former provided a rather static description of a country’s .nan-
cial situation, the latter are more appropriate to describe the medium-long term
economic development of a country and its capacity to ful.ll its debt obligations,
that somehow have a dynamic aspect.
8These results are also con.rmed in a their subsequent paper (Lloyd-Ellis et al., 1990.
11