Trade and Democracy in Developing Countries: Granger Causality Tests

Abstract

The relationship between trade and democracy according to the broad scholarship is controversial. The most significant advance in this field is empirical. However, the empirical tests that lack sound theoretical foundations imply correlation and not causation. This paper resolves the puzzle of democracy-trade literature, i.e. instability of democracy-trade relationship, by using a detailed analysis of the relationship stability and its direction across samples, datasets and different model specifications. The rolling regressions are used to identify the subsample instability of the relationship, whereas the Granger causality tests capture the asymmetries in the relationship between trade and democracy and allow testing for the direction of the causality. Among the results, first and foremost, the participation in international institutions in general mediates the relationship between trade and democracy. Second, the direction of causality depends on the number of the lags chosen. Thus, the tests signify the problems in lags selection and the importance to theoretically justify the number of lags that scholars use in the studies. Finally, I document the sample instability. Although the question, what “came first” democracy or trade, remained unanswered, this paper indicates avenues for future research and a call for solid theoretical grounds.