What Can We Know About Corruption? A Very Short History of Corruption Research and a List of What We Should Aim For

Corruption research has centered on three fundamental questions: What determines corruption? What effects does corruption have? And finally, what can we do to reduce corruption effectively and efficiently? Corruption has been, and continues to be researched so intensively that it is hard to keep abreast of all the latest developments in the various subfields of corruption research. Recent surveys of the vast literature include Aidt (2003), Svensson (2005), Lambsdorff (2007), Pande (2007), Olken and Pande (2011), and Kis- Katos and Schulze (2013). Instead of adding yet another survey, we focus on the main challenges of empirical corruption research and show by example how the literature has tried to address these challenges.1 The first and foremost challenge of empirical corruption research is that corruption is not directly observable, at least not in any systematic fashion as it is illegal and therefore clandestine. Thus, proxies for actual corruption levels need to be found or controlled situations need to be created in which corruption is observable. Proxies frequently used are corruption perceptions, stated corruption experiences, audit results of public expenditure tracking analyses or quantitative service delivery surveys, corruption incidents registered with law enforcement agencies, and corruption convictions. Controlled situations include laboratory and field experiments.2 Moreover, there are some natural experiments as well.