In theory, the IMF could influence economic growth via several channels, among them advice to policy makers, money disbursed under its programs, and its conditionality. This paper tries to disentangle those effects empirically. Using panel data for 98 countries over the period 1970-2000 it analyzes whether IMF involvement influences economic growth in program countries. Consistent with the results of previous studies, it is shown that IMF programs reduce growth rates when their endogeneity is accounted for. There is only weak evidence that compliance with conditionality mitigates this negative effect. IMF loans have no statistically significant impact.