logit analysis to identify what factors characterise 45 developing countries’ adop-
tion of an IMF programme, during 1980-’85. He adds domestic credit growth
and the government’s share of domestic output to the regressors included in pre-
vious studies and discovers that countries which enter Fund programmes have
higher rates of domestic credit expansion and more expansionary policies than
“non-programme countries”. Finally, as in previous papers, the current account,
international reserves and per capita income all appear with the expected negative
coe^ients.
Conway(1994) estimates the determinants of participation in a Fund arrange-
ment, where the participation variable is either a binary variable (which is mod-
elled using a probit model) or a continuous variable censored at 0 and 1 (modelled
using a Tobit model).7 He considers a sample of 74 countries, over the period 1976-
’86, and shows that the most important variables to explain participation in IMF
programmes are past participation in Fund arrangements, lagged values of the
real GDP growth rate and of the current account (where a generally bad past eco-
nomic performance contributes signi...cantly to IMF participation), current values
of the world real rate of interest, of the terms of trade and of the outstanding
long-term external debt (which have, let’s say, the “traditional” impact on IMF
participation, that is negative coe^ientsfor the ...rsttwo variables and a positive
sign for the last one).
Finally, Knight and Santaella (1997), in a binary choice framework, reckon
that the event of a Fund approval of a .nancial arrangement is the result of two
joint events: both a country’ s need to obtain an IMF arrangement and the Fund
approval of the request (on the basis of an evaluation process of the economic
reforms a country intends to adopt). Therefore, they criticise other previ ous
papers for having considered, either explicitly or implicitly, only the so called
“demand-side” determinants of Fund arrangements. On the other hand, their
7In the latter case, the variable indicat ing “partecipat ion” is a continuous measu re of the
p ercentage of the year that a country has spent under an IMF programme.