Assessing the Impact of Light Trail Transit on Land Values and Tax Revenues

15. Supplementary Notes (a) No cost extension approved for this project due to time lost by PI due to medical leave (Summer 2009). (b) Project focus needed to be substantially redirected to include a thorough investigation of housing market foreclosure dynamics following the housing market crisis. 16. Abstract Light Rail Transit (LRT) has become an increasingly popular centerpiece of smart growth policies in the United States, yet considerable debate exists about its cost-effectiveness. The overarching question that we propose to address in this research is " What are the early returns to proposed and under-construction Light-Rail Transit (LRT) investment and how do these vary across neighborhoods? " It may take 10 or 20 years to generate substantial revitalization of a given neighborhood, but early increases in property value and tax collections may stimulate neighborhood change, and defray capital costs. A related question is " What is the best methodology to address the responses to LRT at the neighborhood level, while simultaneously measuring changes in property values at a regional scale rather than focusing solely on station area impacts? " We propose also to separate agglomeration effects (such as neighborhood commercial and residential development) from those associated with change in mobility (such as better access to downtown). Accordingly, a team of researchers with expertise in smart growth, spatial and temporal econometric modeling, investment decision making, and GIS/GIS-T will build a series of models to estimate the effects of LRT on the land markets. Detailed GIS-based data on land-use, new construction, and residential transactions will be analyzed for the LRT system in Phoenix, Arizona, for three distinct phases—pre-approval, planning, and construction. Innovative independent variables include dynamic measures of accessibility at the metropolitan scale derived from GIST models, and a proxy for neighborhood change derived from data on new construction. Modeling techniques capable of disentangling neighborhood and regional effects include Geographically Weighted Regression (GWR) and a local Spatial and Temporal Autoregressive (STAR) model. Comparison of out-of-sample predictive ability and spatial autocorrelation in the models' residuals will inform on performance, and provide a more nuanced understanding of the dynamics set into motion by decisions to build LRT, quantify the spatial distribution and extent of early returns on investment at the metropolitan scale, and provide more effective decision-making on this smart growth tool. The housing market and credit crisis that unfolded shortly after the study began required research efforts to be redirected towards understanding the …