Understanding Blocking Behavior on Airbnb: An Empirical Study of Airbnb in New York City

This study investigates what makes hosts in New York Citys blocked listings on Airbnb; that is, when hosts mark properties as unavailable when there are no corresponding bookings. While hosts blocking behavior is frequently overlooked in evaluations of short-term home rental markets, it is crucial for understanding platform supply and availability. In order to investigate the factors influencing blocking, the research conducted both descriptive analysis and logistic regression analysis on a combined dataset with over 47 million amount of data that includes both daily listing statues (A for available, B for blocked, R for rented) and static proper-level variables like Average Daily rate (price), Number of Bedrooms and Bathrooms, Rating, etc. Based on the descriptive analysis, blocking occurs frequently but disperses unevenly. A small fraction of listings is blocked regularly or even unavailable most of the times, while most housings are merely occasionally blocked. The study then applies a logistic regression model to identify key predictors of whether a listing is ever blocked and is then able to discover that higher average daily rates and larger number of reviews are associated with lower odds of a listing being ever blocked, while higher ratings are entire-home listings are linked with increased blocking likelihood. Finally, the paper concludes by pointing out limitations related to variable creation as well as data scope and suggests further research possibilities.