The effects of IMF programs on income inequality: a semi-parametric treatment effects approach

Abstract

Purpose This paper aims to provide new insights regarding the impact of International Monetary Fund (IMF) programs on income inequality. Design/methodology/approach The paper uses a novel methodological approach proposed by Acemoglu et al. (2019), using (1) the regression adjustment, (2) the inverse probability weighting and (3) the doubly robust estimator, which combines (1) and (2), and a sample of annual data for 135 developing countries over the time period 1970 to 2015. Findings The findings show that IMF programs are associated with greater income inequality for up to five years. By differentiating the effect of IMF programs, the authors find that only IMF non-concessional programs have a significant detrimental effect on income inequality, while IMF concessional programs do not have a consistent effect on income inequality. In addition, the authors find that only IMF programs with a higher number of conditions have a detrimental and statistically significant effect on income inequality, compared to IMF programs with a smaller number of conditions, where their effect on income inequality is found to be insignificant. Originality/value To the best of the authors’ knowledge, the analysis developed in this paper contributes to the existing literature by applying the most methodologically sound identification strategy, which does not rely on the linearity assumption, the selection of instruments or matching variables and additionally takes into account the selection bias related to IMF program participation.