Essays on Dutch Disease and exchange rate pass-through: Evidence from Canadian manufacturing industries.

摘要

The dissertation consists of three essays on Dutch Disease and exchange rate pass-through. Dutch Disease refers to the adverse effects of the natural resource booms on the tradable sectors (manufacturing industries) which may occur mainly through the subsequent appreciation of the real exchange rate.The second and third essays together aim to model and estimate the degree of exchange rate pass-through into Canadian producer prices in manufacturing industries. The second essay, as a theoretical one, presents a literature review and contributes to the literature by developing a relatively more general theoretical framework. The provided model, which extends Yang's model (1997) by incorporating the role of the tradable inputs, is able to show all the major determinants of exchange rate pass-through together, while the previous studies have only analyzed the role of one or some of these factors. Specifically, the theoretical model indicates that the exchange rate pass-through should be between one and zero, while it is positively affected by the share of tradable inputs in total cost, and the domestic firms' market i share and negatively by the elasticity of marginal cost with respect to output. The sign for the degree of substitutability among the variants is not theoretically clear and remains as an empirical question.Finally, the third essay presents the empirical framework for estimation of the exchange rate pass-through and its determinants in Canadian manufacturing industries. In this essay, the short- and long-run exchange rate pass-through elasticities to the domestic producer prices are estimated for the industries at three, four and few cases of five-digit levels of NAICS (about 100 industries), using quarterly data from 1992--2007. Then, the pass-through variation across industries is explained by regressing the estimated pass-through elasticities on the variables that are hypothesized to affect the pass-through elasticities according to the developed theoretical model. The results indicate that incomplete pass-through is observed in most cases although its magnitude is different across industries. The average short- and long-run pass-through elasticities are 0.24 and 0.36 respectively. The share of intermediate materials, as the tradable inputs, in production costs (with positive effect) and the elasticity of marginal cost with respect to output (with negative effect) are the most important determinants of the exchange rate pass-through across industries.The first essay aims to investigate whether Canadian manufacturing industries have experienced Dutch Disease over the period 1992--2007 as a result of the oil boom. After a review of the literature and discussion of the theoretical considerations, the paper presents a two part empirical analysis to estimate the short- and long-run Dutch Disease effects for the Canadian manufacturing industries at three, four and few cases of five-digit levels of NAICS (about 80 industries), using quarterly data. The first part of the empirical analysis estimates the relationship between real exchange rate and energy prices as well as the other related factors and the second part estimates the effect of real exchange rate on output of the manufacturing industries. Based on these two estimated relationships, the Dutch Disease effect is derived by calculating the effect of energy prices on output of the manufacturing industries. The results indicate that the direction and magnitude of the Dutch Disease effect varies substantially across industries likely, as theory explains, because of differences in market structure in terms of the market power. Specifically, 53 out of the 80 industries suffer from the Dutch Disease with the elasticity of -0.18 in average, while Dutch Disease is beneficial for 24 industries with the elasticity of 0.21 in average. The simulation results reveal that, among the industries suffering (benefiting) from the Dutch Disease, each industry could have more annual output growth by 0.93 (-1.07) percent in average if energy prices remained at its level in 1992. This simulated value for the whole sample is 0.30 percent which is significant compared to 2.8 percent as the average of annual industrial production growth during 1992--2007.