The prevalence of poverty in a given population is determined by both the level of average income and the shape of income distribution. Accordingly, the difference in poverty between two populations can be attributed to disparities in their average incomes and in the levels of income inequality. In this paper, we decompose the differences in relative poverty between each of the twenty-one Croatian counties and Croatia as a whole into the contributions of the mean income and income inequality, using the Household Budget Survey data for 2010. The decomposition framework that we utilize here is one usually applied for decompositions of intertemporal poverty changes, and is based on the concept of Shapley value from cooperative game theory. Poverty is measured by three conventional measures – the headcount ratio, the poverty gap, and the squared poverty gap – and robustness of the results to switching from one measure to another is discussed. The results of decompositions show that in most cases both the mean income and inequality differences contribute to poverty variation across the counties, relative to poverty in Croatia as a whole. When poverty is measured by the headcount ratio, the income contribution dominates the inequality contribution, while when we switch to the other two measures, which give more weight to the poorer among the poor, the inequality contribution starts to dominate.