Abstract:
This study investigates the share of open space that maximizes total private property values in urban areas. Open space poses a number of trade-offs to city managers. On the one hand, previous studies have shown that certain kinds of open space can increase property values, which tends to increase tax revenues. On the other hand, open space typically requires substantial capital to establish and perpetual maintenance costs to maintain. This means that in order to keep the city budget balanced, financing open space requires either taking money away from other municipal services, which may be of greater value to residents than open space, or increasing the property tax rate. Both these courses of actions tend to reduce property values, and therefore, lower tax revenues. Open space also incurs an opportunity cost, in that land used for open space could be developed and taxed.
While previous research has modeled these trade-offs, there is still more to be learned by empirically estimating the share of open space that maximizes property values in urban areas. According to the theoretical underpinnings of this study, one of the primary determinants of a city's value-maximizing, or "optimal", share of open space is the price elasticity of housing supply. Therefore, in order to estimate the optimal share of open space, this study estimates the price elasticity of housing supply for 349 U.S. Metropolitan Statistical Areas (MSAs). According to theory, the other factors that determine the optimal share of open space are the price elasticity of housing demand, the economies of scale in the provision of municipal services, the elasticity of property values with respect to municipal services, and the elasticity of housing demand with respect to open space. For these factors, an example value is establish based on prior research and used commonly among all MSAs to estimate the optimal share.
Once the estimated and example values are determined, they are inserted into the equation that determines the optimal share of open space. The result provides an estimate of this optimal share of open space for 349 MSAs. On average, the model, combined with the estimated and assumed values, produces very low estimates for the optimal share of open space. The mean optimal share was 1.5%, and 95% of the estimates were 5% or less. For shares based on statistically significant supply elasticity estimates, optimal shares ranged from 0.2% to 27%.
In order to gauge how far cities were from their estimated optimal share of open space, this study compared the estimated optimal share to observed shares of open space in 72 MSAs. When compared to observed shares of open space, the model (along with the estimated and assumed values) showed that 89% of the observed MSAs displayed "excesses" of open space, or, an observed share of open space that exceeded their optimal share. The other 11% demonstrated a "shortage" of open space. The average deviation between optimal and actual share was an excess of 6.3 percentage points.
Further analysis was conducted in order to account for the error inherent in the supply elasticity estimates (and subsequently inherent to the estimates of optimal share). Once this error was accounted for, only two cities still showed evidence of having open space shortages: Stockton, CA and Miami, FL. However, both cities were within a percentage point of their optimal share's confidence interval, making it possible these cities are not experiencing meaningful shortages of open space. In contrast, 92% of the cities in the sample set showed statistically significant excesses of open space. Of these, five MSAs exceeded their confidence intervals by 15 or more percentage points: Austin, TX; Albuquerque, NM; Akron, OH; New Orleans, LA; and Anchorage, AK. Because these cities' actual share of open space lies so far above their optimal share, it is very likely that decreasing open space area would increase
property values. Two cities in the sample fell within their optimal share's confidence interval: Washington, DC and Virginia Beach, VA. Of all the cities in the sample set, these two are the most likely to be at their optimal share of open space, and therefore, are the most likely to decrease property values by making any changes to their share of open space.
After this primary analysis, a sensitivity analysis was conducted in order to determine how assumptions regarding the variables impacted the estimated optimal share of open space. A reasonable range for each variable was established based on the literature, and this range was used to test each variable’s effect on the optimal share of open space. These tests revealed that the optimal share is not especially sensitive to the assumed values for the price elasticity of housing demand, nor to the economy of scale in the provision of municipal services. However, the elasticity of property values with respect to municipal services and the elasticity of housing demand with respect to open space both have large influences on the optimal share. The impact of all the other variables increased as supply elasticity decreased, and as the elasticity of housing demand with respect to open space increased.
Because the elasticity of housing demand with respect to open space has such a disproportionate influence on the optimal share of open space, and because there is very little empirical evidence surrounding its value, further analysis was done to investigate this variable. By assuming that the 72 MSAs for which there is an observed share of open space are at their optimal share, in conjunction with the estimated and assumed values for the other variables, one can estimate the implied value for the elasticity of housing demand with respect to open space.
Using this method, this study found that the average implied elasticity was 0.57. While this result could indicate that open space has a higher-than-assumed effect on housing demand, evidence from the literature suggests this value is too high. It is more likely that this result provides further evidence that the model is indicating actual shares of open space are higher than optimal.
In the final portion of the sensitivity analysis, cities' observed shares of open space were again assumed to be at the optimal level, however, the other variables assume the limits of their reasonable range so as to make the implied elasticity as
high or as low as possible. This analysis provided further evidence of discrepancies between actual and optimal shares of open space. While the evidence for open space shortages was fairly slim, the analysis reinforced evidence that some cities have excess open space. Those that presented the highest implied elasticities (and therefore show the strongest evidence of open space excess) are Austin, TX; Akron, OH; and Greensboro, NC.
By comparing optimal shares of open space to observed shares of open space, the results show that the majority of cities could likely increase property values by decreasing their share of open space. This study also sheds new light on the relationship between housing demand and open space. By defining a reasonable value and range for the elasticity of housing demand with respect to open space, this study adds to the scarce information on this variable.
While the results of this study indicate that many urban areas in the U.S. have larger shares of open space than would maximize property values, it is important to emphasize that the value-maximizing share of open space is not the socially optimal share. Open space provides a number of other social benefits that are not capitalized into property values, and are therefore not considered in this study. Environmental benefits are one important example. Further research is needed to determine the socially optimal amount of open space that maximizes social welfare.
The appendices of this study list both the estimates of housing supply elasticity and the estimates of value-maximizing share for the 349 MSAs in the sample set. For the managers of cities that were included in the study, these figures can provide valuable information to help them better understand the implications of open space provision. The housing supply elasticities serve to better illuminate housing markets in their area. The optimal share estimates allow managers to better understand the relationship between property values and open space.
For the managers of cities not included in this study, the methods presented here offer a relatively simple way to calculate their own housing supply elasticity and optimal share. With data on housing prices, housing construction, and population, interested parties can estimate the supply elasticity in their cities. Using this estimate in combination with the values this study assumed for the other variables, they can estimate their own value-maximizing share of open space. By comparing this figure to their present share of open space, city managers can gain a better understanding of the effect open space has on the property values within their city.