External Shocks and Japanese Business Cycles: Impact of the 'Great Trade Collapse' on the Automobile Industry

摘要

Why did the Japanese economy perform worst among major industrialized nations during the Lehman crisis period? This paper looks for an answer to this question. The country's poor performance was surprising because the Japanese financial system remained mostly stable, unlike its counterparts in the US and Europe. Obviously, Japan during this period was hit by a massive contraction in external demand for its products. However, even if one takes this factor into account, it is not immediately clear why it had to experience an output decline which was disproportionately larger than the extent of the external demand contraction. We shall investigate this puzzle by focusing on the automobile industry, the country's most important exporting sector. A popular explanation for the strong negative response of Japan to the Lehman crisis is that the country is extremely "export dependent". It is often argued that, during the course of the long boom between 2002 and early 2008, Japan had become so much more dependent on exports that there was no surprise in its poor performance during the crisis period when worldwide demand collpased. In the first half of this paper, we investigate these claims by a time series estimation technique known as the time varying parameter VAR method. We find that there was no noticeable structural change in the relationship between external variables and automobile exports during the pre-crisis boom. We show that, based on the relationship that existed prior to the crisis, one could not have predicted the observed sharp declines in export and output during the crisis period, even if we knew that there was going to be a large negative external shock. There are two possible explanations to the above finding: either that Japan went through a structural change whose timing happened to coincide with the Lehman crisis, or that there is an inherent non-linearity in the relationship between external shocks and production. That is, output reacts differently to large shocks than to small shocks. The second half of the paper pursues the latter possibility. We utilize detailed data on automobile production, sales, and inventories that are available by company and by types of cars. We use a Quantile Regression approach to find that auto producers tend to undertake disproportionately more aggressive inventory adjustment against a larger negative shock to sales. At the end of the paper, we offer some insights on why such a non-linearity is observed.