“Measure not by the scale of perfection”: fixed-effects versus peers in hedonic valuation of spatial amenities

A fine spatial scale is essential to the widely applied omitted variable bias treatment of small-area-fixed-effects in Hedonic Pricing. However, amenity valuation is subsumed into the fixed-effects when amenities vary at a coarser spatial scale. To recover amenity valuation while retaining the treatment of spatial omitted variable bias, we derive a new “Differenced-Price-Peers” specification by integrating Hedonic Pricing and the prices and attributes of spatiotemporal peers. We show that small-area-fixed-effects and Differenced-Price-Peers treat omitted variable bias and capitalize the distance to the city center equally well in a spatiotemporally dense US data context. Conversely, in a sparse Greek data context with anisotropic noise pollution where small-area fixed-effects fail, we show Differenced-Price-Peers to successfully recover aviation noise capitalization. The noise discount of house prices is -0.71% per decibel and about 70% higher than the magnitude of the non-fixed-effects Hedonic Pricing, which suggest potential undervaluation due to spatial omitted variable bias.