Who Benefits from New Housing Supply?
In this research project, I investigated the impact of new housing supply on the distribution of rents. The project started in June 2019 and was completed in September 2021. It was funded by Deutsche Forschungsgemeinschaft.
Private markets contribute to the supply of affordable (rental) housing. The central theoretical mechanism is called „filtering“, a term referring to the fact that housing units „filter down“ the housing quality distribution as they age. The mirror-image view is that households move up the housing quality ladder when their income increases. Their previous housing unit can then be occupied by households with relatively lower incomes. Taken at face value, this suggests that it takes a long time until a newly supplied housing unit has reached a quality standard that is accessible to low-income households. However, this need not be true if households „jump“ up the housing quality ladder. This may be the case if moving costs are taken into account.
The project’s main goal is to quantify this relationship. How strongly does the rent distribution of a city react to the supply of housing by private markets? The central methodological difficulty is to find a source of variation in new housing supply that is unrelated to the state of the local housing market. This is necessary because cities with strong and growing housing demand will likely also attract more housing developers. The pure correlation between (changes in) rents and new housing supply might thus even be positive. As a solution, I propose to use weather-induced delays during the construction phase. These delays lead to a smaller number of housing completions in a given year. I find that a shock to new supply of single-family housing to a local market induces a decrease of rents at all quality levels.
In a second step, I study the decision process of households. In most simple models of filtering, households do not face any moving costs. Consequently, in such models, it is possible for household to adjust housing quality whenever household income changes (presuming that households find it optimal to spend a particular share of their income on housing services). In that case, a new housing unit of rather high quality would be occupied by a household living in a housing unit that has a slightly lower quality. It seems likely in this setting that the impact on rents of lower-quality housing units would be delayed considerably, because it takes many such „small“ moves up the housing quality ladder to get from a low- to a high-quality unit. To the contrary, if moving costs are substantial, households will only move house if the improvement in housing quality is substantial. In that case, new supply of high quality housing units would lead to a quick expansion of effective housing supply also at lower quality levels.
In order to tackle the second question, I develop a quantitative dynamic discrete choice model. The model is then estimated by combining survey data from the German Socio-Economic Panel and data on rental housing units observed in the Mikrozensus survey. I show that renter households sort into housing units by matching housing quality to household income, but moving costs are substantial also when moving only locally. This implies that the typical adjustment of housing quality is large.
Using simulations, I show that a shock to new supply to the owner-occupier market triggers moving chains in the secondary rental market. This contributes to a quick propagation of the initial supply shock. The simulated impact on rents is qualitatively similar to the reduced-form estimates.
The results are summarized in a working paper. Link
The rental housing data are available for academic research projects. Please contact me to get access.