Tariffs and income: a time series analysis for 24 countries

We argue for a new approach to examining the relationship between tariffs and growth. We demonstrate that more can be learned from time series analyses of the experience of individual countries rather than the usual panel data approach, which imposes a causal relation and presents an average coefficient for all countries. Tentative initial results using simple two-variable cointegrated VAR models suggest considerable heterogeneity in the experiences of the countries we look at. For most, however, there was a negative relationship between tariffs and levels of income for both the pre- and post-Second World War periods. However, in the second half of the twentieth century, the causality ran from income to tariffs: that is, countries simply liberalized as they got richer. Policy decisions based on the usual panel approach might thus be very inappropriate for individual countries.