Measuring aggregate land values using individual city land value gradients

Aggregate land value is useful for a variety of research purposes including measuring the social surplus generated by cities and evaluating urban development policies. Nevertheless, only one previous study, Albouy et al., (2018), has attempted to measure cross-sectionally comparable aggregate land values for U.S. cities. That study relied on vacant or near-vacant land sales and used a single pooled aggregate estimate of the land value function. This research uses land values imputed by Larson et al., (2021) to estimate land value gradients for individual cities. Furthermore, the city boundary is measured using estimates of population density functions. Aggregate land value estimates from Albouy et al., (2018) and the individual city gradient approach used here are tested against the prediction of the Rosen, (1974)-Roback, (1982) model, that land value should rise with variables reflecting natural amenity. The individual city gradient approach produces estimates of intercity variation in aggregate land value that agree well with those in Albouy et al., (2018) and are consistent with theoretical expectations.