Abstract
Income or expenditure data are collected at the household level and are silent about within household inequalities. As an alternative, employing a counting-based, double cut-off method allows education, health, and household living conditions to be measured separately, and a deprivation score for each person is calculated based on pre-determined weights of each dimension. Individuals whose deprivation scores are above a pre-determined threshold are deemed multidimensionally poor. In this study, we calculate multidimensional poverty for Turkey in four equally weighted dimensions using Survey of Living Conditions 2006-2015: education, health, employment and household living conditions. First, we find improvement in household living conditions across the board and including employment leads to a faster drop in the gender poverty gap. Second, overall improvement is the result of rapid improvement in younger cohorts; older individuals experienced very little gain. Moreover, there is very little convergence in regional gaps between men and women, and even a widening gap between women in Eastern provinces and women in the rest of the country. Finally, households with multidimensionally poor women and non-poor men are consistently the most common sub-group, and the overall improvement is a result of increasing share of households with no poor members pointing towards polarization between households.
