"Do We Know How Much Our Homes Are Worth? (with Talya Miron-Shatz)"

Grace Wong Bucchianeri, Wharton

Abstract

This paper investigates the self-reported home value, an indicator of housing wealth used in the economic and other literature extensively. Given no observable “true” value of a given home at any given time, we come up with two benchmarks, both based on a market transaction of a specific home. First, we study the gap between a reported home value and the subsequent sales price for around 2,000 retirement-age households across in the US, all of whom have sold a house within one year of self-reporting a home vlaue. We find systematic and significant differences in reported home values from the sales prices by income, housing finance status and health. This analysis focuses exclusively on households who have sold a home. We also document the evolution of reported home values on the same home by the same households across time, whether the home is subsequently sold or not. Second, we perform similar analysis on about 600 women in Franklin County, OH, by tracing back the historical sales price of their residence in the tax records. We use a hedonic-adjusted historical sales price as a benchmark, after adjusting for housing characteristics, geographical location and time trends. This enables an analysis on those who have not sold a home. The main findings remain robust. Additionally, we use a hypothetical experimental setting to demonstrate that even given identical information about a house, there exists initial variations in self-reported price evaluations , and those variation further influence home sale outcomes, including asking prices, probability of a sale and sales price. This suggests that variations in price evaluations affect economic actual behavior. We conclude that researchers need to be mindful of systematic variations in the reported home value indicator, especially when income, health or financial correlates cannot be sufficiently controlled for or accurately measured. We discuss how the interpretation of relevant literature can be affected by a potential bias in the reported home value indicator.