Parcel Size and Land Value: A Comparison of Approaches

The analysis presented here uses simulated and real data sets to investigate the relative merits of parametric, semi- parametric and Bayesian methods in testing for the co-existence of plottage and plattage and in identifying the inflection point. Using artificial datasets generated with spatial correlation, inflection points at alternative locations over the range of sample sizes, different sample sizes and varying the relative quantities of small vs. large parcels, we find that the Bayesian approach method generally dominates the semi-parametric one. In turn, these two methods strictly dominate the parametric one.