| This dissertation contains two separated papers. The topic of the first paper is Comparative Advantage Indexes, Relative Costs, and Implications on Trade Pattern: A Case of Bilateral Trade in Manufacturing Commodities between Seventeen Countries and the U.S., from 1981 to 1990. The purpose of this paper is to present an alternative method to estimate a relative unit cost of a specific aggregate commodity sector between two countries, and apply it to construct a new comparative advantage index. Theoretically, the magnitudes of the new comparative advantage indexes reflect both ranking of relative unit costs and trade pattern. Data from 18 countries are used to construct and test our relative unit costs and comparative advantage indexes on their trade pattern determinations. Relative unit labor costs are also applied with our concept of comparative advantage index to compare their performances with our relative unit costs. In term of absolute cost comparisons, comparative advantage indexes constructed from the relative unit costs are more realistic than those constructed from relative unit labor costs. The tests show strong supportive evidences on the relationship between our comparative advantage indexes and values of exports in several countries. The topic of the second paper is Ricardian Model of Comparative advantage and Its Empirical Evidences on Relationship of Technology and Trade Pattern: A case of Manufacturing Commodities in Thirty-Three Countries, from 1981 to 1990. The purpose of this work is to prove the empirical validity of the Ricardian model of chain comparative advantage. Labor productivity and total factor productivity are used as the indicators of technologies. Two approaches of regression estimation with fixed effect dummy variables and estimation of correlation coefficients with the application of nonparametric kernel density are used. The paper shows that there are significant positive nonlinear relationships between the patterns of net exports and technologies. In addition, the paper reveals the highly influences of the imports to the test results and the problems of using only exports in the empirical test. |
