Abstract:

In economics researchers have access to a large variety of different methods. The question is how well these methods work if the aim is to advice politicians. In this paper theoretical and empirical methods are evaluated with respect to two dimensions: The ability to detect causal relationships and the external validity of their results. It turns out that there is a tradeoff between these two criteria. This leads to the conclusion that all methods available within the “causality – external validity space” should be used for the sake of policy advice and this includes the experimental and behavioral methods.