Online Appendix for Controlling Unobserved Heterogeneity in Repeat Sales Models : Application to Anchoring to Purchase Price

This is the Online Appendix for: Controlling Unobserved Heterogeneity in Repeat Sales Models: Application to Anchoring to Purchase PriceStocks, art, automobiles, housing, corporate bonds and other assets are often studied in a repeat sales framework in order to control unobserved heterogeneity. But unobserved variables that influence both first and second sales remain a troubling source of omitted heterogeneity for a large group of two-stage repeat sales models. We develop a simulation algorithm to correct omitted variable bias. Our algorithm is based on a function of the variance of the unobserved as a proportion of a measurable value: the total variance of unobserved plus random noise associated with the first sale. We illustrate our model with an application to seller anchoring to the price paid for a house. Our simulation results suggest that premiums on expected losses are reduced by about 50% compared to conventional models, as expected if our algorithm reduces unobserved heterogeneity. We use property tax assessed value data to control endogeneity and to validate our simulation algorithm. Our algorithm can be estimated with standard hedonic data and it is relevant to many two-stage studies, including boundary effects models, studies of credit spreads in bond markets and firm acquisitions