2008 年 44 巻 4 号 p. 20-28
Arguments about disparity in the Japanese society, from the viewpoint of studies on social stratification, have pointed out two important issues: Income disparity and disparity between regular and irregular workers. Studies on social stratification have not analyzed these issues seriously. They have studied occupations rather than income; they have implicitly assumed that workers have regular jobs. To overcome these deficiencies, this paper studies the effect of employment status -regular workers versus irregular workers- on income by analyzing the 2005 Social Stratification and Social Mobility (henceforth, SSM) Survey data. The result of regression analysis with logged income as the dependent variable and occupation and employment status as independent variables shows that employment status has stronger explanatory power than occupation does. Arguments about disparity, however, have made another claim that disparity has recently been growing. To check the empirical validity of this claim, we compare regression coefficients of two regression models using the 1995 SSM and the 2005 SSM data. The result of the comparison shows that the coefficient of regular workers had become smaller from 1995 to 2005, which means that income disparity between regular and irregular workers had shrunken. To test this result rigorously, we build a sophisticated regression model and apply it to the 1995 SSM and 2005 SSM data. In the model we fix coefficients of the model using the 1995 SSM data and add interaction terms between the survey year and each independent variable. The coefficient of an interaction term shows increase or decrease in the explanatory power of the independent variable in the interaction term depending on its sign (positive or negative). The result of the regression analysis shows that the coefficient of regular workers is negative, which means that income disparity between regular and irregular workers had shrunken over the decade. It is risky to say that income disparity has shrunk based only on these results. Rather, what we would like to say is that these results would promote fruitful dialogue between arguments about disparity in the Japanese society and studies on social stratification.