| In this thesis, the trade balance-exchange rate relationship is examined, using the multivariate cointegration and error-correction methodology for the industrial countries during the floating exchange rate period. With regard to theory, an improved version of the Imperfect Substitutes model is developed that includes both the demand-side and the supply-side factors. A reduced form version of this model forms the basis of our empirical work.;The empirical models employed in this study identify the exchange rate, relative income, relative cost of production (or input prices) and the relative capital stock as the key determinants of the trade balance and of import prices. Using these empirical models, I examined four issues that are related to the trade balance-exchange rate relationship: (1) Is there a long-run equilibrium relationship between the trade balance and the exchange rate? (2) Is there a J-curve effect? (3) Is there a pass-through effect? (4) Is there a two-way causation between the trade balance and the exchange rate?;I employ aggregate data on ten industrial countries: Australia, Belgium, Canada, Finland, France, Germany, Japan, Norway, the United Kingdom and the United States. Using Johansen's cointegration method, I found that, for all ten countries in the sample, there exists a long-run equilibrium relationship between these five variables. The results thus support the traditional models that identify exchange rate as an important determinant of the trade balance.;Upon discovering that the above five variables are cointegrated, I employ an error-correction model and a VAR model to see whether the trade balance of these ten countries follows a J-curve pattern. The results of the error-correction model indicate that the trade balance of only three countries, Belgium, France and Japan, exhibits the J-curve effect. However, the VAR model demonstrated the J-curve effect for Canada, Japan and the United States. Since the VAR approach allows for the feedback effects of one variable on others, the results obtained from this approach are more reliable than those of other approaches employed in the literature that ignore these feedback effects. The results thus suggest that for these countries, if the domestic currency is allowed to depreciate, it will take some time before improvements in the trade balance materialize.;The relationship between the exchange rate and import prices (i.e., the exchange rate pass-through) plays an important role in the trade balance-exchange rate relationship. The presence of an incomplete pass-through is often considered as an important cause of the weak association between the trade balance and the exchange rate. My pass-through equation includes an income variable that was absent from the previous work on exchange rate pass-through. I first examine whether the variables of the pass-through equation are cointegrated. Using the Johansen method, I found that for all ten countries, the five variables--import price, exchange rate, relative income, relative cost of production and relative capital stock--are cointegrated, and that for five of ten countries the pass-through effect is incomplete. For two, the pass-through was found to be more than complete.;Lastly, using the error correction modeling approach, I examine whether there is a two-way causation between the trade balance and the exchange rate. I found that for only three countries (Canada, France and the United States) the causation runs from the exchange rate, as the traditional models suggest. However, for six of the ten countries, the results indicate a two-way causation between these two variables. (Abstract shortened by UMI.). |
