The spatial and social structure of income inequality in Germany

Abstract

Out of the multiple possible dimensions of inequality such as income, health or education, inequality of income is the highest. This study complements similar research on income inequality and socio-economic and spatial disparities with a comprehensive analysis of spatial descriptors of inequality in Germany. Using anonymized gross income tax declarations, we compute the Gini inequality index at the municipality level. We extract spatial variables from mostly open spatial datasets, and socioeconomic variables, from the openly available 2011 national census. We focus on measures of spatial variability, such as variation of spatial attributes or measures of population segregation and diversity between spatial units of one squared kilometer census cells within municipalities. Results show that a random forest model performs variable in predicting inequality across federal states, with a R2 statistic ranging between 0.3 and 0.58. Predictions are found significantly better for big municipalities with more than 10,000 inhabitants. For most states, inequality positively correlates with diversity of religion and nationality. Between the different states, inequality is associated with specific attributes of the built environment, from density of residential living space or size of residential annexes such as garages to coefficient of variation of building height or the amount of green spaces.