Openness

Not long ago, Rodriguez and Rodrik (2001, p. 39) concluded that –the challenge of identifying the connections between trade policy and economic growth is one that still remains before us.” In the early 2000s, papers on growth and openness included terms like “once again” or “revisiting” in their titles. Three aspects appear to be behind this challenge:

• 1. Trade policy is highly correlated with other growth-relevant policy measures such as the rule of law. Disentangling the effect of trade policy is therefore a difficult empirical undertaking. I will avoid these difficulties by interpreting trade openness as an overall policy variable that captures a broad array of growth-relevant policies.

• 2. There is no consensus on how to best measure openness and trade policy. Measures such as tariffs, black market exchange rate premia and trade shares have been proposed. Some authors argue for adjusting trade shares for gravity variables and price differences, while others use the raw shares.

• 3. The literature is not always explicit about whether it is looking for an effect of openness on the level or on the growth rate of GDP. Theory gives conicting priors, and cross-section convergence regressions make interpretation of some of the results rather hazardous. Trying to find a link between the level of openness and the growth rate of GDP is likely to lead to inconclusive results as I will show later. I will argue that the level of trade openness is a good explanatory variable for the level of income per capita – and that the change in openness helps explain the change in income.