Imperfect Information, Learning and Housing Market Dynamics

Résumé

This paper examines the decision problem of a homeowner who maximizes her expected profitfrom the sale of her property when market conditions are uncertain. Using a large dataset of realestate transactions in Pennsylvania between 2011 and 2014, I verify several stylized facts aboutthe housing market. I develop a dynamic search model of the home-selling problem in which thehomeowner learns about demand in a Bayesian way. I estimate the model and find that learning,especially the downward adjustment of the beliefs of sellers facing low demand, explains some of thekey features of the housing data, such as the decreasing list price overtime and time on the market.By comparing with a perfect information benchmark, I derive an unexpected result: the value ofinformation is not always positive. Indeed, an imperfectly informed seller facing low demand canobtain a better outcome than her perfectly informed counterpart thanks to a delusively strongerbargaining position.