A bivariate relative poverty line for time and income poverty : detecting intersectional differences using distributional copulas

Empirical research on poverty today often goes beyond a focus on income to consider other dimensions of well-being. However, relatively few multidimensional poverty measures explicitly consider time, despite its particular relevance to women's double burden of paid and unpaid work. We construct a bivariate relative poverty line between income and leisure, based on their joint distribution in the population. Because the strength of the dependence between income and leisure influences the vulnerability to poverty, we incorporate distributional regression into copula models. Utilizing the 2018 Mexican National Survey of Households, Income and Expenses, we investigate differences in bidimensional poverty with respect to gender and ethnicity. We find that the fraction defined as bidimensional poor is 18 percentage points higher than the poverty rate computed from separate time and income measures. Those below the relative but above the absolute poverty line are primarily non-indigenous women whose poverty is made visible by our approach.