Corruption and trade regulations: an instrumental variable approach

Abstract

We find that tariff rates decreased less rapidly in more corrupt countries in instrumental variables (IV) estimation, using mortality rates of European settlers at the time of colonization as IV. Our finding implies that tariff rates are endogenous to corruption.

Notes

¹ In a political economy model with endogenous regulations as discussed here, one can build a model where corruption can contribute the persistence of inefficient regulations.

² Unfortunately, the average tariff rates are available only for about 20 to 40 countries each year. This is the main reason for the relatively small sample size.

³ We experimented with dropping China, dropping transitional economies, using only data from the WDI source, adding EU countries, and adding observations with imputed tariff rate from tariff revenue. The main results remain. The results are available upon request.

⁴ Ades and Di Tella (1999) presented evidence that corruption is negatively associated with competition from foreign firms proxied by share of imports in GDP. Treisman (2000) also reports that corruption is associated with exposure to imports and exports. He, however, acknowledges that he could not find convincing instruments for trade openness to control for the possibility of reverse causality.

⁵ We also used changes in tariff rates as dependent variable instead of final year tariff rates. In IV regressions with these changes in tariff rates, the coefficient of ICRG is estimated to be −7.62 with SEs 2.36.

⁶ In the first-stage regressions, the corruption level is quite strongly associated with settler mortality with t-value larger than 3.9.