Is paying bribes worthwhile? Corruption and innovation in middle-income countries

Corruption research in economics has a long history. Seminal early articles, and older findings contrast with newer developments which have as yet not been measured empirically; in particular the link between corruption and innovating activities suffers from multiple results, on both a national country and company or firm level. This paper examines the corruption-innovation link in transition and emerging countries as the decision to corrupt, and the ability to innovate may not be independent. An endogenous switching regression model is advocated as a suitably methodological way of modeling the joint determination of a firm's innovation and possible bribes as it implies not only a selection between corrupted and non-corrupted firms, but also heterogeneous effects on innovative activities. The paper shows that, when the selection effect is adequately considered, different firms' strategies arise. In particular, the treatment effect of corruption on innovation is positive for corrupting firms and negative for non-corrupting firms. Corrupting firms appear rational because paying bribes increases their innovative activities. However, non-corrupting firms also appear rational because in the presence of bribes, their innovating activities would be fewer. Thus, when the selection effect is adequately considered, the effects of so-called "greasing-and-sanding-the-wheels" can co-exist. Finally, the role of competition is also considered. Building on these results, future research can move forward to re-examine economic outcomes such as the productivity or the economic impact of corruption, in the presence (or absence) of selection processes.