Found in the (Random) Forest

Amsterdam’s beautiful Grachten


The impact of Airbnb has come under significant scrutiny. This short paper contributes to the literature by looking at Airbnb’s effect on house prices in Amsterdam. The key issue is identification due to the likely presence of unobserved confounding factors like tourism demand, which shift housing supply and demand in Amsterdam. We employ Generalised Random Forests to estimate a local average partial effect that comes closest to a causal effect of Airbnb on house prices. These results are compared to the benchmark of a panel data model with time- and area-fixed-effects. The estimated average treatment effects show a nuanced picture of the causal effect of Airbnb presence on local housing demand. Further distance to an Airbnb seems to increase house prices by 0.25% for every 100 meters on average. A 0.019% decrease in house prices per additional listing within 250 meters, on average, seems to suggest a counterintuitive negative effect on local house prices on average (possibly due to negative externalities). The spillover of Airbnb on neighbouring areas' house prices may be positive, which requires further investigation. The random forset technique also shows that the effect of Airbnb is very heterogenous.

Christian Tien
Christian Tien
PhD Student in Economics (3rd yr)

My research interests include causal inference, specifically identification, and machine learning.