Abstract
The empirical evidence on the effect of corruption on foreign direct investment is a mixed bag as the results show positive, negative or no effect. These results, which are typically based on cross-sectional and panel data, are highly sensitive to the selected set of explanatory variables, model specification and variable measurement. Hence, these results are subject to the Leamer critique that a regression model with a large number of potential explanatory variables can be used to prove almost anything. In a situation like this, researchers are in a position to prove prior beliefs and produce results (after extensive data mining) that support the theory they like. It is concluded that the empirical evidence on the underlying issue must be taken with a big grain of salt because it is a product of the junk-science of econometrics.
Keywords:
Disclosure statement
No potential conflict of interest was reported by the authors.