Abstract
This article examines within-group inequality, which has been typically neglected in regression-based analyses, as well as between-group inequality. We argue within-group differences of dependent variables are substantively interesting objectives for the social sciences. By focusing on the heterogeneity of within-group variances, we analyze what kind of people are more vulnerable to social and economic risks. We adopt the model which simultaneously estimates the parameters of covariate vectors for the mean and the variance. Using 1995 and 2005 Social Stratification and Social Mobility Survey in Japan, we test whether the variances of individual and household incomes are larger for particular groups. The results indicate that (1) non-regular workers are not only likely to have lower than average incomes but also greater variability in incomes, and that (2) Married people tend to have higher than household income and less household income variability. Consistent with earlier literature on social stratification, we confirm that economic insecurity is not evenly distributed across social groups. Especially, those who have been included the male breadwinner model enjoy more stability.
