Effect of Location, Neighborhood Quality, and House Quality on Property Values in Memphis

Effect of Location, Neighborhood Quality, and House Quality on Property Values in Memphis

Chris Kamphaus

April 26, 2007

________________________________________________________________________

Abstract

As the cost of living in America climbs, so does the importance of measuring the value of homes and the inputs that affect them. If one can accurately estimate the value of a home, it makes for a better informed consumer, who can make smarter decisions when it comes to purchasing a residence. As the average American spends 62% of annual income on housing costs, it is evident that if misinformed, money can be unnecessarily foregone. This study was designed to examine the differences between housing prices in different geographical locations of Memphis, Tennessee, Hedonic pricing models allow us to get accurate estimates of these factors that affect value. The main objective was to test the effect of dwelling specific inputs on price in three different regions of Memphis, and conduct variable sensitivity analysis to changes in geographic location. This will enable us to better understand which inputs are significant in which areas, and formulate hypothesis as to why these disparities exist. My results suggest that median income, poverty rate, geographic region, and number of stories, have the greatest effect on the price of a home. Further, all dwelling (structure) specific inputs are sensitive to geographic location, but not necessarily to income in all cases.

________________________________________________________________________

1. Introduction

Measuring the value of homes is an important aspect of our society because a house is the most valuable asset for the average American. Deciding where to establish residence is a big decision. Buying a home has implications not only for the owner, but for their family as well. There are many factors that people evaluate when purchasing a home, and if they don’t know the true value of these factors, their decision is no longer rational. This is a topic of considerable scrutiny, due to the fact that each model predicts something a little different. Geographic location causes many inconsistencies in findings because each test uses modified inputs. Therefore, each test has proven to be model specific, and thus, few reasonable, overlying assumptions of house characteristics can be made across all locations. When coupling house characteristics with demographic data however, we get a better explanation for the changes of home values from region to region. Trends in the housing market, however segregating they may be, fall perfectly in line with certain demographics. Income level has become the best proxy for neighborhood quality and thus mean home value. Poverty rate, although not perfectly collinear with median income, is another good proxy for neighborhood quality, which has emerged as a key determinant of home value after years of tests.

In this paper I will examine the effect of certain dwelling specific variables in different regions and use demographic data to elucidate the differences between neighborhoods. Analyzing the demographics and regression results will allow assumptions to be made as to the reason for changes in price. Why certain variables have less of an effect in certain areas? Why certain variables wash out the effect of others? How we can justify the effect of alterations to a house? Variable sensitivity will be key to this study, because it will allow questions to be raised as to why the effects of certain variables change from neighborhood to neighborhood. From the variable sensitivity analysis we may better understand why some zip codes will benefit from, say more rooms, and why others will be adversely affected. The results I have attained will show the significance of house specific characteristics, how demographics can be used to answer segregation questions, and illustrate better decision making opportunities for consumers and for builders to maximize welfare.

This paper is catalogued into four components. First, I will discuss some of the empirical literature on this topic; address some previous findings that parallel my results and introduce those that have come to different conclusions. Following the literature review, I will introduce my data, where it comes from, and how it was used to predict housing prices. This section will also reveal the models I have formed from the data and the tests I have run to ensure the accuracy of the results. Finally,