Information, prices and valuation in real estate markets

This dissertation introduces two models of price formation in a real estate market characterized by incomplete information, costly search and bargaining. The first is a static model where agents optimally search for the bargaining game which provides them with the highest expected payoff. Agents do not acquire any additional information about the item of exchange as they search. In the dynamic model, we relax this last condition and allow buyers to acquire new information as they progress in their search. Thus at any given time, the buyer's optimal stopping rule encompasses the possibility of acquiring more information in the future.; In both models, it is shown that for reasonable parameter values, any observable market transaction price is a useful signal of the true price for the purpose of inference. Under such an interpretation, the task of an appraiser is formalized: an appraiser filters information from market prices to make inference on properties which are not sold. This interpretation is used to rationalize the appraisal "smoothing" issue. An empirical procedure consistent with this notion of information filtering is estimated using transaction prices of income producing properties.; A procedure to calibrate appraisal errors for the purpose of determining the reliability of appraisal based return indices is also proposed. Transaction prices as well as appraised values of income producing properties were used and it is shown that the bias is modest. This section also discusses the possibility of sample selectivity bias when transaction prices are used to make inference on properties which are not sold. A maximum likelihood estimation procedure is proposed which would explicitly account for such biases.