Hedonic Modeling of the Tucson Housing Market: The Effect of Educational Submarkets on House Prices
AuthorHolland, Sandra Carole
AdvisorPlane, David A.
Mulligan, Gordon F.
Committee ChairPlane, David A
Mulligan, Gordon F
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractThis study examines the effects of educational submarkets -- schools and districts -- on house prices in the Tucson region. The supposition that homebuyers will pay more to live in a better school district or school attendance area is examined, with the quality of education measured by per-student expenditures and academic achievement. Traditional single-market modeling of the housing market finds that education submarkets have a small but significant effect on housing price. Further modeling, taking explicit account of the spatial nature of the housing market, suggests that in the single-market approach, education submarkets act as proxies for other neighborhood effects and variables omitted from the model. Incorporating the unique location coordinates of the properties and allowing marginal attribute- and location-effects to vary across geographic space in a trend surface approach produces more robust model results and allows the educational submarket effects to be isolated. The results suggest that school districts have a small but significant price effect even after a fluid price surface has been developed, but that intra-district variation remains. These price effects have some relationship with district quality as measured by academic achievement, but the housing market does not reward per-student expenditures. At the intra-district level, middle school quality does not appear to have a significant effect on housing price, at least in the Tucson Unified School District. However, the trend surface approach still proves to be a useful methodology for modeling small, local-scale variations. The use of polynomial expansion and spatial- attribute variable interactions is successful: problems of variable omission are diminished, spatially autocorrelated error terms are reduced and removed, effects of multicollinearity are minimized, and the effects of the educational submarkets may be examined in isolation.