PANEL DATA HEDONICS: ROSEN'S FIRST STAGE AS A "SUFFICIENT STATISTIC"

Traditional cross-sectional estimates of hedonic price functions can recover marginal willingness to pay for characteristics, but face endogeneity problems for estimating nonmarginal welfare measures. This article shows that when panel data on household demands are available, one can construct a second-order approximation to nonmarginal welfare measures using only the first-stage marginal prices. With repeated cross sections of product prices, the measure can be set identified or, under a single-crossing restriction, point identified. Bounds also can be constructed when there are mobility costs. Finally, a variant remains valid when individual preferences shift over time.