بررسی دوباره دستمزد، اشتغال و ساعات تعدیلات: چه تفاوت در تعدیلات حیاتی است؟

Manufacturing wage, employment, and hours adjustments are significantly different among developed countries, typically between Japan and the US. The problem of explaining the differences is closely related to that of what effect the extent of wage flexibility has on labor market adjustments. However, prevailing theories are invalid or incomplete for explaining the differences. The differences in the adjustments cannot be unrelated with institutional features of the labor market. Indeed, the extent of wage flexibility depends on the differences in the institutional features. Using a new model with such an idea, this paper theoretically re-examines the cause of the differences in the adjustments. The crucial causes are the characteristics of wage flexibility as well as the differences of the institutional features. The characteristics are that wage adjustment is discontinuous, infrequent, and has a lower limit in comparison with changes in demand.

Manufacturing wage, employment, and working hours (per person) adjustments are significantly different among developed countries, typically between Japan and the US. According to Gordon (1982), in cyclical variations in demand, wage and working hours (per worker) are both more flexible, and employment is much more invariable in Japan than in the US (see Table 1). Furthermore, in Japan hours are first adjusted, and then employment adjusted if the hour adjustment is not enough. Empirical evidence after Gordon (1982) demonstrates that those differences remain unchanged.1 The question of why such differences arise is still one of the important issues not only in labor economics but also in macroeconomics. It is closely related to the question of what effect the extent of wage flexibility has on labor market adjustments. However, prevailing theories to explain the above differences, e.g. the human capital theory and flexible wage theory, are invalid or incomplete, as stated later. In addition, the conclusions derived from both theories are different. This paper theoretically re-examines the cause of such differences in the adjustments as observed between a rigid wage country like the US and a flexible wage country like Japan, focusing on the process of a deep recession. In the paper, wage flexibility indicates the flexibility of nominal wage, and the flexibility of real wage is not explained. It is because empirical evidence for the difference in real wage flexibility is not so enough as in the case of nominal wage.In the human capital theory the cost of on-the-job training plays a crucial role in explaining the inter-country differences about the adjustments. Hashimoto (1993) attributed the stability of employment in Japan to greater investments in firm-specific human capital in Japan than in the US, and flexible hours in Japan to its stable employment. Although these relationships appear to be valid, they are based on a supposition. Incidentally, the latter relationship is not always obvious in itself, since wage adjustment has a negative effect on hours, which may, in turn, offset the relationship. Fukao and Otaki (1993) analyzed the influence of training cost functioning as sunk cost on real wage, employment, and hours adjustments in the real business cycle model. They explained stable employment, flexible hours and real wage in Japan under the supposition that training cost is born more in Japan than in the US. Their analysis, however, includes a very serious problem. This is because it is questionable whether training cost in their model is truly investment cost on human capital. What they call training cost is cost required for acquiring a certain level of skill for production. If it is so, larger training cost implies more inefficient labor, not higher skill. They, in fact, consider the effect of lower quality of labor on real wage, employment, and hours adjustments (regarding this point, see Fukao and Otaki, 1993, pp. 75–77 or Otaki, 1994, pp. 200–203, 209–211.) It is, thus, doubtful whether their analysis really explains the cross-country differences. Topel (1982), on which the Hashimoto (1993) study relies, also has the same problem about the effect of the cost. Even if Fukao and Otaki take account of the positive effect of training cost on skill, they would not be able to accomplish their purpose. The reason is that it is ambiguous whether the marginal contribution of training cost is larger or smaller than its marginal cost, and consequently, the effect of the cost on the adjustments becomes ambiguous. This suggests that the cost is not a crucial factor for our purposes, and that the level of the cost should not be assumed but explicated theoretically.2 As far as employment adjustment is concerned, however, it has been prevalently accepted that much smaller employment adjustment in Japan may be owing to its larger investments in specific human capital. This is well known as the labor hoarding hypothesis since Becker (1962) and Oi (1962). They explained the hypothesis using a two-period model in which specific investments in the first period improve workers’ skills and productivity in the second period. But it is only the difference in employment adjustment that the human capital approach can account for. The approach cannot account for the mechanism of labor market adjustment as a whole containing wage and hours as well as employment. It is our model that can explicate the mechanism.3 We analyze the adjustment mechanism through a “flexible wage approach” instead of the “human capital approach”. However, the existing flexible wage approach has not been able to actually explain employment and hours adjustments. According to the approach, flexible real wage, which arises owing to institutional factors, brings about more stable employment. This is usually illustrated as the effect of a leftward shift of a labor demand curve due to deep recession in the neoclassical model of a competitive labor market. However, it is not employment in itself but labor input that is then more stable. The model, thus, does not succeed in actually illustrating more stable employment and more flexible hours in Japan. Hence, wage flexibility in itself is never crucial for our purposes. Incidentally, there are very few theories accounting for hour adjustment. As we stated before, flexible hours are not a necessary result of stable employment. It is also invalid to say that hours become flexible, since the adjustment cost of hours is less than that of employment.4 By elaborating a “new flexible wage model”, we provide a satisfactory explanation of the adjustment mechanism. The model is more or less similar to the insider–outsider model.5 Dynamic models such as those of business cycle and labor contracts are not used, since there is no necessity that we should employ the former model as in Fukao and Otaki (1993), and also since the latter model, which explains rigid real wage, cannot explain the difference in the adjustments. Furthermore, a negotiation model of a firm and its union is not explicitly employed due to the difficulty of its explaining the difference. As will be demonstrated in Section 2, the reason is that the difference of the adjustments depends greatly on institutional features in the labor market, and that it is impossible in terms of time and cost that negotiations are conducted with changes in demand. And, hence, there is no necessity that the difference in the adjustments should be pursued in the difference in each objective function of the firm and union. Even if we do not rely on their models, we will show that a meaningful analysis is sufficiently possible, and a new hypothesis about our purpose can be derived. We focus on the process of deep recession, because the cross-country differences in the adjustments become conspicuous in that phase. We take account of institutional features in the labor market, since the extent of wage flexibility depends on the institutional features. Indeed, it is well known that wages in Japan are more flexible than those in the US because of the following two characteristics. One is the shorter duration of contracts and the closer point of time of consecutive wage revisions than in the US; and the other is cooperative labor-management relations owing to longtime employment, enterprise unions, and information-sharing in joint consultation. The latter characteristic is very few in the US. Based on the difference in such institutional settings, it is assumed that wages in Japan are downwardly flexible corresponding to deep recession, while those in the US are rigid. The assumption about Japan means “risk sharing” or “profit sharing”, which is usually done by bonus adjustment. Even if institutional differences determine the degree of wage flexibility, the degree does not immediately lead to the differences in the adjustments. We elucidate that it is the characteristics of wage flexibility, not the degree in itself, which cause the differences in the adjustments. The characteristics are that wage adjustment is discontinuous, infrequent, and has a lower limit compared with changes in demand. So far, only the difference in the institutional settings has been intuitively stressed as a key factor causing the differences in the adjustments. We will clearly demonstrate that the crucial factors accounting for such differences in the adjustments as observed between rigid wage and flexible wage economies are the characteristics of wage flexibility as well as the difference in the institutional settings. In the following section we present a model. In Section 3 the adjustment mechanism in a labor market is considered according as wage is flexible or rigid. Section 4 summarizes the conclusions.

We indicated that the prevailing theories have never given a satisfactory explication for the differences in wage, employment and hours adjustments among developed countries. We provided an appropriate explication of such differences in the adjustments as observed between rigid wage and flexible wage countries, focusing on the process of deep recession. It was supposed that wages of Japan are downwardly flexible with deep recession owing to the institutional distinction, while those of the US with little such distinction are sticky. The supposition about Japan is “risk sharing” or “profit sharing”. The following is the summary of our results under such a supposition. In such a rigid wage economy as the US, it was shown that employment decreases, but hours are inflexible through the process of deep recession. As for such a flexible wage economy as Japan, it was, however, elucidated that wage flexibility in itself does not cause the differences in the adjustments. We showed that in the stage of wage adjustment, discontinuous and infrequent wage adjustment keeps employment unchanged even if demand deteriorates, and the constant employment and the adjusted wage decrease hours with a declining demand. These results imply that although it is often mentioned that flexible hours cause stable employment, the opposite result holds. Furthermore, we showed that if economic conditions deteriorate further beyond the stage of wage adjustment, employment decreases, but hours are inflexible because of the existence of a lower limit in wage adjustment. From the results above, it follows that in the stage of wage adjustment hours are adjusted, and in the following severer stage employment is adjusted. Even if employment adjustment is done in Japan, it does not necessarily mean the collapse of the longtime employment practice. Whether or not employment is adjusted depends on the extent of deterioration of economic conditions even under Japanese employment practices stated earlier. Taking account of the comparison above between a rigid wage and a flexible wage economies, it turned out that in addition to the differences in the institutional characteristics, the key factors causing the differences in the adjustments are the features of wage flexibility, not the flexibility in itself. The features are that in comparison with changes in demand, wage adjustment is discontinuous, infrequent, and has a lower limit. The crucial points of the analysis are the following two: one is that the institutional characteristics are taken into consideration, and that two exogenous variables are assumed to change in a different way; that is, wage adjustment is definitely different from changes in demand in terms of a frequency, continuity, and lower limit, respectively; and the other is that a corner solution is used in order to explicate the stable employment of Japan in the stage of wage adjustment.