Hedonic Analysis, Repeat Sales, Water Quality, Non-Market Valuation, Spatial Econometrics


Hedonic property value analysis is one of the leading methods of environmental valuation. This non-market technique uses variation in home sales to infer the values of amenities or disamenities. While there have been numerous studies about air quality and hazardous waste, the number of papers focusing on water quality is much smaller. Consequently, there are still many unanswered questions about the proper handling of water quality through hedonic methods. Furthermore, estimates from hedonic property price analyses are rarely used in government cost benefit analyses. This dissertation investigates several important hedonic issues in a large analysis of water quality in central Florida. The first chapter of this paper explores the extent of water quality benefits. Almost all past studies have focused exclusively on waterfront homes. The present paper includes non-waterfront homes and investigates three hypotheses about the marginal impact of water quality. The first hypothesis is that non-waterfront homes are positively affected by water quality, but by a smaller amount than waterfront homes. The second hypothesis is about the effect of lake distance on the relationship between water quality and property prices: this relationship should be negative. The third hypothesis states that properties near larger lakes have a higher implicit price for water quality than homes around smaller lakes, all else constant. These three hypotheses are investigated in each chapter of the dissertation, and provide a unifying theme to the paper. Results from Chapter 1 support all three hypotheses. Most importantly, the empirical estimates indicate that water quality benefits extend beyond the waterfront in a declining gradient. Excluding non-lakefront homes from the analysis can therefore substantially underestimate the total benefits of a water quality improvement. Estimates of the total property price benefits from a one foot increase in water quality were found to double with the addition of non-waterfront homes. The second chapter examines the sensitivity of results to several spatial specifications. Spatial issues can be a problem in analyses of real estate data because of spatially correlated variables, unobservable neighborhood codes and covenants, identical or similar builders, and property appraisal valuation techniques. The focus of the chapter is on the spatial weights matrix (SWM). Six different SWM's are constructed, which are based on popular specifications encountered in the current spatial hedonic literature. An out-of-sample forecasting exercise is used to compare multiple spatial specifications. Results indicate that certain spatial models may be sensitive to the specification of the weights matrix. Furthermore, many popular models currently used in the literature could be improved by allowing more non-zero elements in the SWM. The third chapter investigates the definition of "water quality" and uses several additional quality indicators. Choosing the proper pollution indicator is an issue that has plagued many areas of the valuation literature. While clarity indicators have become popular in hedonic property price analysis, they are not used for the purposes of regulation by many state environmental departments. This chapter uses several indicators that are used by the state of Florida to classify lakes and implement policy. Implicit prices are computed for all of the indicators and issues of benefit extent and total benefits are explored. Instead of finding an optimal indicator for all situations, results indicate that the use of at least two types of indicators may capture a larger range of the true total benefits. The final chapter uses a repeat sales model to address potential problems with omitted variable bias. Due to the size of the data set in this paper, there are a substantial number of homes that have sold more than once. The repeat sales model analyzes differences in property sales prices for the same home over time. The three hypotheses of the first chapter are explored in this alternative model. The implicit price obtained from the repeat sales model is much larger than the regular hedonic model. However, there are some concerns with the smaller population of repeat sales.


If this is your thesis or dissertation, and want to learn how to access it or for more information about readership statistics, contact us at

Graduation Date



Milon, J. Walter


Doctor of Philosophy (Ph.D.)


College of Business Administration



Degree Program









Release Date

September 2009

Length of Campus-only Access


Access Status

Doctoral Dissertation (Open Access)

Included in

Economics Commons