FIBO: A New Approach to Measuring Intercity Differences in Housing Costs

Abstract

Intercity housing price indexes that rely on median house price or pooled hedonic regressions adjust imperfectly for differences in housing characteristics. Single equation hedonic methods produce estimates of intercity differences that vary substantially with the specific cities in the regression, the sample sizes from each city, and/or the housing characteristics used. To mitigate these shortcomings, we create intercity housing price indexes for both rental and asset prices, which is a Fisher Ideal aggregation of a two-way Blinder-Oaxaca decomposition, the FIBO index. Our method improves upon current house price indexes because it is far less influenced by the cities studied, sample sizes from each city, and housing characteristics included. We also find, as predicted by Carrillo and Yezer (2021), that asset prices substantially overstate variation in rental prices across cities.