Determination of Working Time and Employment : Coexistence of Overwork and High Unemployment

Abstract

The main purpose of this paper is to examine the relationship between working time and employment. Long working time have been claimed as one of serious social problems in Japan, which becomes exacerbated after bubble burst. The changes of employment and working time after 1960 Japanese data are summarized according to the following three periods. The first period (high growth period from 1960 to 1970) is characterized by secular reduction in both actual working time and standard working time in concurrence with the high growth of employment. The second period (slow growth period from 1970 to 1991) still shows decreases of the actual working time as well as the standard working time, although the rates of decrease are smaller compared with the first period, but then the increase of employment comes to stop. The third period (low growth period from 1991 to 2002) shows the same decreasing tendencies of the actual working time and the standard working time as well, but in this period not only working time but also employment show decreasing trends. Therefore, a decline in scheduled or standard working time does not necessarily bring a decline in actual working time and an increase of employment. This implies the need of theoretical investigation which treats employment and working time separately. The theoretical problems on this line are hitherto rather poorly investigated with rare exceptions of Hart (1987) and Ohashi (1990). In this paper we present a theoretical model to explain these movements of working time and employment in Japan. Especially, we introduce the standard or scheduled working time into the model and discuss the effects of a decline of standard working time on actual working time and employment. We show the existence of marginal standard working time (MSWT) and if the standard working time is less than MSWT then we have overwork situation and, on the contrary, if the standard working time exceeds MSWT then we have no overwork. Moreover, MSWT is shown to depend on the standard real wage, the elasticity of wage to overwork and the job separation rate. Another characteristic of our model is to incorporate different determinants of real wage rate, namely Keynesian case where the emphasis is placed on goods markets and Classical case where labor markets matter to determine the real wage rate. We will show that the standard working time has negative effects both on employment and actual working time in Classical case, but has no effects in Keynesian case.