Are All World Bank Lending Instruments "Helping Countries Combat Corruption"? Preliminary Draft: June, 1st 2014, Not for quotation

This study investigates whether World Bank loans are selective in targeting the better governed and least corrupt countries. It also identifies if World Bank loans "help countries combat corruption". The originality of this paper is twofold. Firstly, we separately analyze IDA credits and IBRD loans and the various World Bank lending instrument (i.e. projects, adjustment loans and budget supports). Secondly, looking at overall assessment of projects (provided by the World Bank’s Independent Eval-uation Group, IEG), we determine whether higher quality loans are associated with lower levels of corruption in recipient countries. To our knowledge, these aspects have never been studied in the literature. Over the period 1984-1997, we observe that World Bank lending operations were incorrectly targeted. Following the implementation of the Governance and Anti Corruption (GAC) objectives in 1997, projects and Development Policy Lending (DPL) tend to be oriented toward better governed countries. However, we find no evidence that World Bank loans reduce corruption in recipient countries. Instead, unsatisfactory DPL (as measured by IEG ratings) tend to worsen corruption in recipient countries. Our results are robust to various specifications (Tobit, Heckman method, 2SLS and 3SLS) and measures of corruption and governance (CPI, WGI, and ICRG)