Abstract
A hedonic price model was estimated for residential developments in the province of Alberta in Canada. The model is designed to be used with the Production, Exchange and Consumption Allocation System (PECAS) land use modeling framework, which forecasts future development patterns on each parcel based on developer return-on-investment functions, allocates households and jobs, and calculates economic benefit measures. In PECAS, base year rents by land use zone (LUZ) are a calibration target, while future rents by LUZ are calculated in a bid-rent allocation, so the estimation separates out local effects from neighbourhood (LUZ) effects. Local effects include proximity to busy collectors, light rail transit (LRT) stations, schools, and amenities such as parks, as well as local road and off-road access to the nearest network link. A non-linear treatment was required to separate out the value of land from the value of space, especially on large parcels at low intensities of development. A Bayesian approach was used to 1) incorporate prior knowledge from previous PECAS work in the U.S., and 2) to provide a model that can forecast future prices on any parcel in Alberta, even in areas where certain types of development are not currently sufficient in quantity to confidentally estimate parameters. The estimation quantifies the current price landscape in Alberta by LUZ, showing higher prices (and a tendency for future development, where allowed) in major cities, major oil industry locations, and national parks. It also quantifies the positive impacts of local road access, LRT station proximity, schools, rivers, and natural areas; the negative (nuisance) impact of being too near to busy roads and schools; and the depreciation of older buildings.
Get full access to this article
View all access options for this article.