International Business Cycle Transmission in a Dynamic Multi-Sectoral Heckscher-Ohlin Model

Abstract

Both, the popular opinion and the academic literature share the belief that the ongoing globalization with deepening trade channels is at least partly responsible for the appearance of a common business cycle across countries. Accordingly, one can observe a large and still growing literature that is concerned with trade as a medium of international business cycle transmission in general (cf, e.g., Kose et al. (2003), AER, Imbs (2003), IMF WP, Ambler et al. (2002), EER and Backus et al. (1992), JPE). One strand of this literature is completely empirical. Kose et al. (2003), for example, investigate annual data of a sample consisting of 75 industrial and developing countries over the last four decades. They demonstrate that the strength of the trade linkage with the G7 countries increases the correlation of domestic macroeconomic variables with the respective world variables. Typically, the empirical literature on trade as a means of international shock transmission does not reveal the particular characteristics of the trade linkage that can be seen as the causes for the certain sign of the transmission channel. In this context, the word sign refers to the assessment of the transmission channel. If, e.g., a home country H suffers an inefficiency shock and its trading partner, a foreign country F, gains - however gains are measured -, the sign of the transmission channel is termed positive. If, on the other hand, country F looses, the sign of the transmission channel is termed negative. A second strand of the literature draws conclusions from dynamic general equilibrium trade models. These models mostly intend to solve the puzzle of the so-called quantity anomaly described by Backus et al. (1992). The anomaly refers to the size of the correlation of macroeconomic variables between countries: Standard one-good aggregated dynamic general equilibrium models predict a cross-country correlation of output that is smaller than the cross-country correlation in the technology shocks, the latter in turn being smaller than the cross-country correlation in consumption. Empirical investigation, however, results in a reversed ranking of cross-country correlations. Ambler et al. (2002) succeed in removing this discrepancy between theoretical models and reality. By constructing a multi-sectoral dynamic general equilibrium model, they exploit the features that may result from a shock-induced change in the production structure. Anyhow, Ambler et al. (2002) only demonstrate that a certain correlation of macroeconomic variables between countries may exist. All the above mentioned contributions, however, are completely silent on the exact process of transmission and its welfare consequences. This is the starting point of the present study. By analyzing a dynamic multi-sectoral general equilibrium model numerically, it intends to reveal the actual character and causes of a positive or a negative transmission channel between two countries that are linked via goods trade. The base case model is characterized by Heckscher-Ohlin assumptions. This base case model is then extended by incorporating the households' labor supply decisions into the model and assuming imperfect substitutability between home and foreign produced goods. Both observed countries are characterized by a 4-sector production structure, in which the output of two sectors is used as an intermediate input in the other two sectors that each produce one final good. The model parameters are calibrated according to input-output tables from the G7 countries and China as an example of an emerging country. This results in countries as well as technologies for producing both intermediate and final goods that are quite different with respect to relative factor endowments and factor intensities. Therefore, the model may reflect business cycle transmission in the context of North-South trade. The business cycle in the model is driven by negative technology shocks. These technology shocks may be national shocks, if they occur in one country only. The shocks are termed global, if they occur in both countries. The model is solved with help of the software GAMS.