| My dissertation provides a theoretically new and methodologically more rigorous answer to a familiar question: What determines countries' trade policy choices? Existing cross-country, cross-time explanations suffer from crucial flaws. Theoretically, they rely primarily on models that assume countries make trade decisions independently. Empirically, they focus almost exclusively on developed countries, and they employ atheoretical measures of trade protection.; In Chapter 1, I argue that a country's adoption of a trade policy is as much determined by the adoption by other countries of new policies as by domestic factors and focus on two diffusion mechanisms: coercion and demonstration effects. Chapter 2 details a variety of trade tariff measures and identifies their weaknesses. In their stead, I propose the use of a uniform tariff measure (UNITAR) based on Anderson and Neary's trade restrictiveness index and offer an expanded dataset. The UNITAR measure is a single tariff rate which is equivalent in a welfare sense to the whole set of tariff policies, allowing UNITAR to be comparable across time and across countries.; Chapter 3 recaps, extends, and tests the extensive literature on independent domestic and political determinants of trade tariff policy. While a number of the domestic political and economic explanations appear to be significant when tested on a panel of 104 countries for the period 1989 to 2000, their predictive power is limited, especially within developing countries and for changes in tariff policy. The models predict at most 26% of the change among developed countries and never more than 2% among developing countries.; In comparison, the diffusion mechanisms outlined in Chapter 4---trade partner network and emulation of regional, income, and economic peers---increase the predictive power of models to up to 40% for developing countries. Support is found for the diffusion hypotheses in both versions of a large-N analysis of developing countries: a single ten-year period (1988 to 1998) cross-section with 68 country observations and a set of pooled 4 year-period cross-sections (1981--1985, 1986--1990, 1991--1995, 1996--2000) with observations from 103 countries. Thus, countries may be more influenced by what other countries are doing than by what is happening at home. |
