WHEN IS TRADE PROTECTION GOOD FOR GROWTH?

Abstract

I. INTRODUCTION There is a shared belief among economists, policy makers, and the general public that more open economies grow faster than closed economies. Many international organizations emphasize this correlation: for example, the World Trade Organization proclaims on its website that it has helped to create a strong and prosperous trading system contributing to unprecedented However, the empirical evidence is mixed. While many studies demonstrate a positive relationship between trade openness and economic growth (Edwards 1998; Frankel and Romer 1999; Sachs and Warner 1995; and many others), others find no evidence of such a relationship. Harrison and Hanson (1999) demonstrate the lack of robustness of several proxies for openness, for example, while Vamvakidis (2002) finds that the positive correlation between openness and growth exists only after 1970, and Clemens and Williamson (2004) show that the correlation between tariffs and growth was positive prior to World War II and negative after it, which they attribute to changes in the world economy. Perhaps it is not surprising that empirical work has failed to reach a consensus about the correlation between trade protection and growth, because theory does not generally provide an unambiguously negative relationship. In endogenous growth models in open-economy frameworks such as Grossman and Helpman (1990) and Matsuyama (1992), the relationship between trade policy and growth is frequently a contingent one, in which the effect of trade barriers on growth depends on the pattern of comparative advantage across countries. Rodr{guez and Rodrik (2000) expand upon this idea, and also provide a comprehensive and critical review of the empirical literature. As discussed in more detail below, DeJong and Ripoll (2006) allow for contingencies in the correlation between tariffs and growth, based primarily on countries' levels of economic development. Motivated by open-economy endogenous growth models, we examine the relationship between trade protection and growth in a nonlinear framework, and investigate whether the relationship between trade barriers and growth depends on the pattern of comparative advantage across countries. Using a cross-country regression model, we find that such contingencies do in fact exist, and that the correlation between tariffs and growth is strong and positive for skill-abundant countries. Our results are robust to a number of different specifications of conditioning variables and dependent variables. This paper is related to a large literature on openness and growth. As indicated above, many previous empirical studies have assumed a linear relationship between openness and growth. Notable exceptions in which nonlinearities in this relationship are the main focus of the paper are DeJong and Ripoll (2006) and Papageorgiou (2002). (1) In Papageorgiou (2002), trade openness is a threshold variable separating countries into distinct growth regimes; however, openness is not directly included in the growth specification, but operates indirectly, as a separating variable. In contrast, we focus on the potential nonlinearity of the direct relationship between trade barriers and growth, and we find evidence that such nonlinearities do, in fact, exist, in that the relationship between tariffs and growth is contingent on the pattern of comparative advantage across countries. Using a panel of 60 countries, DeJong and Ripoll (2006) investigate the relationship between tariffs and growth and how this relationship depends on income levels. Among higher-income countries, they find a negative correlation between tariffs and growth. Unlike DeJong and Ripoll (2006), we allow for the possibility that the relationship between trade barriers and growth is contingent on the pattern of comparative advantage. Relying on the pattern of comparative advantage, rather than income, as the source of differences in the correlation between growth and tariffs is more consistent with the endogenous growth models, although we do include an interaction between income and tariffs in our sensitivity analysis as a comparison to their paper. …