Public Housing Misallocation: Evidence from the Netherlands

I empirically assess the extent of housing misallocation in the Dutch public housing sector by applying the methodology proposed by Glaeser and Luttmer (2003) and extending it in several dimensions. First, I base my analysis on market valuations of the dwellings in question, so called WOZ values, to improve the comparability across apartments and additionally I propose a monetary measure of misallocation. Moreover, I show that Glaeser and Luttmer's (2003) measure of misallocation increases with the number of categories chosen for the dependent variable in the ordered generalized probit model. Lastly, I introduce an innovative counterfactual group within the public housing sector based on waiting time data for public housing in Dutch municipalities. My results suggest that Dutch households living in public housing are on average misallocated by 6,016 euros, which amounts to 3.7% of the average WOZ value of dwellings in the public housing sector for municipalities with long waiting times of 161,438 euros. This means that households would on average choose a dwelling in the free market with a WOZ score which diers by roughly 6,000 euros compared to the apartment they were assigned in public housing. However, the placebo analysis in this paper suggests that the identifying assumption of Glaeser and Luttmer's (2003) method to calculate misallocation does not hold implying that the results should be treated carefully.