Essays on international trade

This dissertation contributes to an understanding of the effects of trade liberalization by examining new channels and looking for the evidence of the gains from trade. Chapter 1 studies the impact of cutting tariffs upon firms' productivity. Tariffs in the final market and those in the intermediate input market have different effects. In particular, I show that when the intermediate inputs are not highly differentiated, trade liberalization in the final market leads to a lower extensive margin of imports. With fewer input varieties to work with, the firms in the final market country are less productive and workers' wages are lower. Trade liberalization in the intermediate market, however, leads to more varieties of imports and therefore boosts productivity. More surprisingly, when the intermediate inputs are highly differentiated, trade liberalization in the intermediate input market, which makes imports of input cheaper, yet leads to fewer varieties of imported inputs, hence lowering firms' productivity. Cutting tariffs in the final market, as always, has the opposite effect on the number of imported input varieties and the firms' productivity. These results are valid whether labor, the only factor in my model, is allowed to move between sectors (in the long run) or not (in the short run). It can also be shown that the effects in the long run are always greater than in the short run. All of these predictions are supported by data collected by INEGI, the Mexican Institution of Geographic and Information Statistics, from 1984 to 1990. Moreover, the model developed here allows estimates of the elasticity of substitution among intermediate inputs, estimates that are consistent with other studies and can be used to quantify the gains from trade. Traditional trade theory assumes one representative firm that produces a single product. While that assumption is convenient, it does not give a full picture of the effects of trade liberalization, especially the selection effect when firms are heterogeneous in productivity, as shown by the seminal work of Melitz (2003). Chapter 2 extends Melitz's work by allowing the firms to produce multiple products. Empirical evidence shows that multiproduct firms account for the majority of exports, and therefore they cannot be neglected. Taking the assumption that the varieties must be managed by talented people, I show that more talent enhances the selection effect, and welfare is improved as a result. Indeed, when the pool of managers widens, the more efficient firms can hire more of them and expand their product portfolio, and this in turn pushes less efficient firms out of the market. As a result, the economy has only varieties produced by the more productive firms, and the average productivity increases. When two countries, differing only in their talent endowment, engage in free trade, overall welfare is enhanced, as represented by a lower price index. However, when talent is allowed to move across borders, we can replicate an integrated economy with factor price equalization. Everyone in the new integrated country is better off, with the possible exception of managers in the country with less talent. If no one is allowed to move, the country with more talent has more intense competition, so its firms need to have high enough productivity to stay and produce in the market. These firms produce fewer units of more products than similar firms in the other country. Although overall welfare is higher, the people in the country with less talent may be worse off because their firms need to cut costs in order to compete with firms in the other country. The third chapter looks at how trade liberalization affects firms in the short run and whether they improve their productivity by learning from the export market. The previous mixed evidence for this could be due to industry heterogeneity as the high technology industries are more likely to experience a spillover effect, or it could be due to biases in productivity estimates. Many measures of productivity, for example, do not take into account the possible effect of export status, so this implies a bias against the learning by exporting hypothesis. Using recent data from the Chinese automobile industry, and after correcting for the econometric biases, I find no evidence of learning by exporting. Indeed, the productivity gap between exporters and non-exporters does not widen after the time of export entry. The data also provides information about export destinations, which can shed light on why there is no evidence of learning by exporting