Measuring Effects of Housing Densities on Property Values Using Locally Weighted Regression

The objective of this study is to determine spatial differences in the marginal values of housing density and their implications for housing development projects. In order to achieve this objective, the values households place on neighborhood density were measured using locally weighted regression in a hedonic housing-price framework. The neighborhood housing density coefficient in the global model confirms the positive and significant value of lower neighborhood density. The spatial distribution of the housing density marginal effects from the local model shows the variation of site-specific values of neighborhood housing density. The marginal effects of lower neighborhood density gradually increase closer to the Town of Farragut, on the west end of the county. The Town of Farragut and adjacent areas show the highest marginal effects of lower neighborhood density on housing price. It was found that the marginal implicit price for one fewer house per acre in the Farragut area was $ 1 1 ,964 whereas the whole Knox County area was $3,53 1 . Based on these marginal implicit prices, open space in Farragut is valued 3 .4 times greater than open space in the whole Knox County study area. Understanding the spatial variation in values of neighborhood density across specific locations can lead to "smart growth" policies that are more appropriate for site-specific conditions. For example, a site-specific conservation subdivision ordinance could be developed based on this implicit price ratio. Each conservation subdivision is required to set aside a minimum percentage of its adjusted tract acreage as open space. Typically, the minimum amount of adjusted tract acreage is defined rather arbitrarily without systematic consideration of households' preferences about open space. For instance, based on the ratio found in this study, 3.4 times more open space within a conservation subdivision could be set aside for open space in the Farragut area compared with the overall area. Since the Farragut area has a smaller neighborhood density than the overall area, more abundant open space outside a subdivision in the Farragut area should substitute for the minimum open space requirement within the subdivision. Thus, the open space that is set aside for a subdivision in the Farragut area should be less than 3.4 times the amount set aside for a subdivision in the overall area. In conclusion, "smart growth" policies promoting only high-density development are incapable of providing a cure for urban sprawl without also considering spatial variation in the values of neighborhood density.