Online Reputation and Observable Quality in the Short Term Hospitality Pricing Under High-Dimensional Machine Learning: A Case of Airbnb

This paper shows that guests’ online reviews play a critical role in determining rental prices on the Airbnb platform. We identify this pricing mechanism by developing an information-cost-based hedonic machine learning approach that links guests’ online reviews, property amenities, and rental prices. We investigate the effects of guests’ online reviews and the capitalization effects of property amenities using low-dimensional versus high-dimensional regressions. We estimate those effects using U.S. Airbnb data and compute the guests’ online review sentiment scores. Our empirical findings show that, using high-dimensional amenity-vector-based hedonic regression and the double machine learning algorithms, the guests’ online review sentiment scores variable is statistically positively associated with rental prices.