An economic analysis of the differential effects of neighborhood amenities on housing prices in Minneapolis and suburbs, 1970-1980

This dissertation investigates differences in intra-urban housing prices in Minneapolis and suburbs between 1970 and 1980. The relevant literature on filtering and arbitrage models of neighborhood change is reviewed along with housing research based on a hedonic price index. Housing prices and related attributes are gathered from Multiple Listing Service data on 684 randomly selected houses and U.S. Census data on 231 neighborhoods in Hennepin County as defined by 1970 census tract boundaries. A Neighborhood Succession Index is constructed to compare geographic differences in housing appreciation. A five-part method is developed to distinguish between significant housing structure and neighborhood attributes. Three regression models are developed to test five hypotheses. Initially, the Neighborhood Median Price Model and the Neighborhood Median Price Change Model use census data to examine the influence of neighborhood attributes on median housing prices. Finally, a Housing Attributes Price Model is tested to distinguish between the influence of ten housing structure and twelve neighborhood attributes using individual housing sales as the dependent variable. A Box-Cox transformation is used to correct for problems of heteroskedasticity. Six restricted models are tested to determine the significance of various neighborhood attributes. These models seek to identify the extent to which intra-urban differences in housing prices can be explained by a neighborhood premium and penalty price index. The analysis reveals seven housing-structure attributes explain eighty percent of the variation. Nine neighborhood variables explain one-third of the remaining variation. All neighborhood attributes when dropped from various restricted models were significant at the.05 level. Neighborhood price premia display a strong geographic pattern and challenge prior research on the determinants of housing value based either on one-dimensional distance gradients or the use of housing sub-markets. A discussion of research methodology reveals the limitations of aggregated data suggesting future research efforts incorporate neighborhood and proximity characteristics unique to each house in the sampled area. Policy implications for local municipalities and development organizations are discussed in the context of using imputed housing prices to construct indicators to estimate differences in neighborhood quality.