The IMF and government health expenditure: A response to Sanjeev Gupta.

Do policy conditions attached to International Monetary Fund (IMF) lending programmes have an impact on government health expenditure in developing countries? Yes, according to a large body of literature (see Kentikelenis, 2017), and our recent article (Stubbs et al., 2017). We systematically reviewed IMF loan agreements and staff reports to generate a database of “binding” conditions that could plausibly impact health expenditure. Our database offered an alternative to the IMF's own conditionality dataset, which has been widely criticized for inaccuracies and omissions (Arpac et al., 2008; IEO, 2007). Using cross-national models covering 16 West African countries between 1995 and 2014, we found that each additional binding IMF policy reform reduces government health expenditure per capita by 0.25% (95% CI -0.44 to 0.06). The mean number of binding conditions, at 25 per year, thus corresponds to a 6.21% reduction, on average, in government health spending per capita associated with IMF conditions. To further test these findings, we performed a narrative review of these documents. They showed that IMF policy reforms reduce fiscal space for health investment, limit expansion of doctors and nurses, and undermine health system efficiency. It was clear that IMF programmes placed enormous pressure on already strained health systems, reducing health spending at times when economic