WEAI/AERE 2009 - Individual Paper Abstract
Title: Dynamic Estimation of Open Space and Housing Values: A Novel Repeated Sales/Matching Approach
Author(s): Bowman CUTTER, Pomona College, Department of Economics, Bowman.Cutter@Pomona.edu, 425 N. College Avenue. 909-607-8182; Linda Fernandez, Ritu Sharma (photo credit: Rick Burnham)
Abstract:
This research employs a novel quasi-experimental method to value proximity to open space for residential home values in Riverside county. Research on the value of proximity to open space typically uses cross-section or repeated cross section data where open-space areas are fixed and constant over the time period of the sample. This type of analysis could result in biased estimates of open space values because open space proximity could be correlated with unobservables that influence house values.
We use house sales data from Western Riverside County (in Southern California). This county maintained an active program of open space acquisition from 1988 through the end of our sample period in 2004. The addition of new open space reserves allows us to test whether creating nearby open space adds to the value of a residence. The addition of the open space reserves can be viewed as an experiment with a treatment group where proximity to open space changes and a control group. We use a repeat sales approach that measures whether the rate of house price appreciation is greater in a time period where the proximity to open space declines for that house. The repeat sales methodology allows us to control for all time-invariant house characteristics, whether observed or not.
It is possible that open space preserves were placed in areas where home values were increasing more rapidly than other areas in any case. For instance, perhaps people are more motivated to improve their houses if they are close to open space. One method we use to diagnose this effect is a false treatment dummy. Our data contains many houses that sold several times over the period and where the proximity to open space changed in one sale pair, but not in others. In these cases, we create a false treatment dummy that is positive for an observation if : 1) the house had a change in proximity to open space at some transaction pair; and, 2) this particular observation (transaction pair) did not have an open space proximity change. If the coefficient on the false treatment dummy is positive and significant, it suggests that open space preserves are placed in areas where house values are appreciating in any case. In our case, the false treatment dummies are insignificant and near zero.
We also use a matching methodology to test whether the control and treatment groups are similar. We employ a variety of propensity-score based matching techniques to obtain weights for control and treatment groups. Using these weights we re-examine our base specifications and find that the match-based-weighting has little effect on our overall results.
Our results appear robust to a number of different specification tests and we believe they are the first results to rigorously identify open-spaced proximity values based on quasiexperimental methods. In addition, the results are particularly policy applicable because they apply to open space preservation on the wildland-urban frontier. This frontier is often where the ecological value of open space is high because there is room to maintain contiguous habitats. Our research suggests that there are significant benefits to residential use values in addition to these ecological benefits and measuring these values over time is important.