The past decade has seen a push towards rigorous policy evaluation and new, exciting
methodologies have evolved to get a better sense of how effective policy interventions are in
achieving their desired outcomes. Unfortunately, few of these efforts were aimed at governance
reform initiatives, especially not at transparency reforms which have received almost univocal
support from policy makers and practitioners as a means to cure a whole range of problems that
hamper the growth performance of resource-rich countries, namely corruption. This thesis tries
to fill this gap by triangulating a series of different methodologies to conduct both a cross and
within-country evaluation of the effectiveness of the Extractive Industries Transparency
Initiative (EITI) in reducing corruption in two EITI countries, Peru and Mali. I use firm level
data from the World Bank Enterprise Surveys (WBES) and country level economic and social
indicators from the World Development Indicators (WDI) and apply two preprocessing methods
developed by Abadie and Hainmueller – entropy balancing and synthetic control groups – as
well as within-country difference-in-difference estimations to measure the impact of EITI on
reported corruption. The results show a robust reduction of corruption in Peru, where corruption
has been reduced by as much as 14 percentage points since the introduction of EITI, while no
such effect can be found in Mali. This paradox can be explained by two standard critiques of the
iv
EITI: that it is only effective when a functioning civil society is in place prior to EITI
implementation; and that the voluntary nature of the initiative may induce large differences in
outcomes between countries.