The Latent Heterogeneity of the Wage Structure and Gender Wage Gap: Dual economy analysis revisited (Japanese)

"Dual economy theory," and more generally theory and research on the relationship between segmented labor markets and economic inequality, has historically been treated as deviant in the study of labor economics. This paper shows that part of the problem lies in the analytical method and measurement of central concepts, and provides a remedy to this problem while progressing the research pioneered by Ishikawa and Dejima (1994) on the analysis of the dual economy in Japan. Labor markets in Japan can be interpreted as a probabilistic mixture of three elementary forms, each having a distinct wage structure. The most significant factor affecting the allocation of people to distinct labor markets is the distinction between regular and irregular employees. firm size, industry, and schooling also have a strong effect on allocation. Gender has no direct effect on allocation. One labor market comprising around 35% of employees is characterized by high average wage, 99% regular employment, and a greater representation of college graduates and employees of firms with 300 or more employees. It also has a larger wage return to human capital and less gender wage gap than average. However, it is also the labor market where gender inequality in the length of tenure is greatest. The second and largest labor market comprising around 60% of employees is characterized by relatively low average wage, 61% regular employment, and a larger representation by high school and junior college graduates and employees of medium-sized firms with 30-299 employees. This is the labor market where the gender wage gap is greatest and there is a large gender inequality in the rate of wage return to tenure. These first two labor markets have a greater gender wage gap at the time of employment among people with children compared to those without. The third labor market comprising only 5.5% of employees is characterized by 27% regular employment and an overrepresentation by middle-school graduates and employees of small firms with less than 30 employees. The labor market has the lowest average wage and employees have uniformly low wages except for managers and a small proportion of regular workers. No direct gender effect on wage exists in this labor market. However, a larger representation of women, particularly those with children, exists due to the larger proportion of irregular workers among women. These results indicate that a larger allocation of women in the two peripheral labor markets is an epiphenomenon caused by the exclusion of irregular workers from the core labor market, and do not support Doeringer and Piore's (1971) argument on the relative exclusion of minority workers from the core labor market. In Japan, gender wage gap is generated mainly within the core and peripheral labor markets, and only indirectly between the labor markets due to the larger representation of irregular workers among women. These results demonstrate the theoretical merit of considering the latent heterogeneity of wage structure in labor markets.