«A Journal of Policy Development and Research HoPe VI Volume 12, Number 1 • 2010 U.S. Department of Housing and Urban Development Office of Policy ...»
The entire redeveloped site, named the Townhomes on Capitol Hill, has housing for owner occupants. Of the 147 units on the site, 134 are affordable cooperatives and 13 are market-rate homes. About one-third of the cooperative units are reserved for public housing residents, another one-third are set aside as affordable units (for households making 80 percent or less of the metropolitan area median income [AMI]), and the remaining one-third are for moderate-income households (those earning up to 115 percent of the AMI). The goal of the redevelopment, unique among its HOPE VI contemporaries, was that resident ownership would result in better property maintenance and thus fewer problems for the neighborhood. The Townhomes complex is entirely self-sufficient and does not require any ongoing government subsidies.
Washington: Wheeler Creek Estates (Formerly Skytower and Valley Green) The Wheeler Creek Estates complex sits on the site of the former Skytower and Valley Green developments in the Washington Highlands neighborhood in southeast Washington. Both original properties consisted of three-story, walkup buildings that had been designed with the institutional starkness common in 1960s-era public facilities. The 312-unit Valley Green complex had been built in 1961 to house people displaced by slum clearance and highway construction projects elsewhere in the District of Columbia. Many of those individuals were then working at the nearby military installations but moved away when the bases closed later in the decade. With fewer working tenants, conditions at Valley Green began a steady decline. Vandalism plagued the site throughout the 1980s, and maintenance orders overwhelmed management. Eventually, DPAH decided to board up damaged units when tenants moved out. The vacancies offered havens for criminals, however; by the late 1980s, Valley Green had become a major drug market, riddled with violent crime. Conditions worsened so much that, by 1994, only eight families remained at the property (Loeb, 1997; Knox, 2007).
The adjacent Skytower property, developed at roughly the same time as Valley Green, was privately owned and managed. Containing 91 units, Skytower had received federal subsidies during its construction, housed only Section 8 voucher recipients, and continued to be insured by HUD.
Unfortunately, the property suffered from very poor management. Property upkeep was minimal, at best. By 1994, the complex had widespread internal water damage, had no emergency lighting or fire extinguishing systems, and was infested with rats. Crime was as bad at Valley Green, if not worse. In a single month in 1993, 16 people were killed and 22 were assaulted in the immediate vicinity of Skytower. HUD foreclosed on the property a few years later, and DCHA acquired it as part of a broader strategy to redevelop Valley Green.
DCHA obtained a $20.3 million HOPE VI grant in 1997, demolished both the Skytower and Valley Green complexes, and created the 314-unit Wheeler Creek Estates. The redeveloped site contains 148 public housing units (48 for families and 100 for seniors), 32 market-rate rental units, and 134 homeownership units. It also features a 13,000-square-foot community center and a daycare center. Of the 134 homeownership units, 30 were set up as lease-to-purchase homes for former public housing residents and 11 were purchased outright (for $115,000 each) by other public housing households under an arrangement in which the household was responsible for a first mortgage of $45,000 and DCHA paid a second mortgage of $70,000 if the owner resided in the home for 20 years. The remaining 93 homes were sold at market rates to the general public.
The development, which was completed in 2000 and fully occupied in 2001, earned HUD’s Best Practices Award in 2000 and an award from the National Association of Home Builders in 2002.
Cityscape 107Zielenbach and Voith
HOPE VI Effects on Residential Property Values What effects have these redevelopments had on their surrounding communities? This section examines differences in local property values before and after the redevelopments and the extent to which the changes can be attributed to the HOPE VI investments.
Residential Property Data To evaluate the effects of the HOPE VI investments on neighborhood property values, the study used First American Real Estate Solutions transaction data on single-family home sales from 1990 to 2006. In addition to sales prices and dates, the data include the addresses and sizes of the homes. We defined the surrounding neighborhood as the area within 2,000 feet of the HOPE VI property. Because city blocks typically are 500 feet long, the 2,000-foot distance captures effects within a four-block radius—the likely limit of any single development’s impact. In actuality, natural and manmade barriers, such as rivers, hills, highways, and boulevards, can limit the potential impact area for particular properties. In three of the four cases (Orchard Gardens being the exception), we revised the impact area after visiting the sites and consulting aerial and census tract maps. In the two Washington cases, we also truncated the impact area to minimize the influence of nearby public housing sites on property values. The panels in exhibit 3 show the four impact areas.
For the two Boston sites, the study used data from 28,235 residential sales. More than 400 of these sales occurred within the Mission Main ring, and nearly 200 occurred within 2,000 feet of Orchard Gardens. For the Washington sites, the study used data on 39,788 residential sales. More than 640 of these sales occurred within 2,000 feet of the Townhomes on Capitol Hill, and more than 350 sales took place within 2,000 feet of Wheeler Creek Estates. The remaining sales took place in ZIP Codes adjacent to the designated HOPE VI impact areas; we segmented these data into two categories: sales within 2,000 feet of a “traditional” (that is, not a HOPE VI) public housing property and those outside of such a radius.
Property Effects Estimation Methods To measure the effects of HOPE VI redevelopment on neighborhood home prices, we employed a hedonic interrupted time series model similar to that used by Galster, Tatian, and Smith (1999).
The empirical specification allows for identifying price levels and trends in the impact areas of both HOPE VI and traditional public housing, relative to those in areas with no public housing.
The specification further allows the price levels and trends in HOPE VI neighborhoods to diverge before and after the HOPE VI investment. In this context, we assumed that the counterfactual— what would have happened with no HOPE VI redevelopment—is the continued deviation in price trends in the HOPE VI area from those in the overall sample.
Our multivariate regression model controls for the year a home was sold, its size, its distance to the central business district, its location relative to the HOPE VI site and to traditional public housing (within or outside of a 2,000-foot radius of each), and the housing price trends in the HOPE VI and surrounding ZIP Codes both before and after the HOPE VI intervention.7 The specification included indicator variables for each year and year indicators that interacted with a variable indicating the presence of traditional public housing in the neighborhood. These variables allowed us to trace the overall house price trends and measure the difference between the overall price trend and the trend in areas with traditional public housing. We also included a variable that indicated whether the sale was in a HOPE VI neighborhood before redevelopment began, and its interaction with a linear time trend to trace pre-HOPE VI trends. Similarly, we included a post-HOPE VI variable and tracked its interaction with time to capture the changes in price levels and trend from the pre-HOPE VI period.
One challenge is determining the appropriate intervention point, the moment at which investors perceived a change in the neighborhood. In theory, that tipping point could be when the redevelopment was announced, when demolition of the former property began, when tenants began reoccupying the redeveloped site, or when the redevelopment was completed. Reasonable arguments can be made for choosing any of these times, and no moment stands out as the most likely tipping point. Thus for each development, we therefore assessed change in appreciation trends around each development, using as intervention points the years marking the start of demolition, first occupancy, and project completion, and then focused on the model that best fit the data.8 Estimated Property Value Effects Exhibit 4 shows the key estimated parameters for the four HOPE VI sites.9 The dummy coefficients represent the Y-intercept for the trend lines before and after redevelopment, and the trend coefficients represent the slope of those lines. The neighborhoods encompassing both Mission Main and Orchard Gardens experienced positive, statistically significant differences in price trends subsequent to the HOPE VI redevelopment. Before the demolition of the former public housing Appendix A includes the model and its empirical specifications.
We explored, but generally were unable to distinguish, separate effects for the redevelopments’ announcement, demolition, move-in, and completion dates.
Appendix B shows full regressions.
properties, surrounding property values were decreasing at rates of 8.2 and 9.2 percent, respectively, relative to the whole sample. After demolition, values surrounding the two sites increased at rates of 5.4 and 6.2 percent, respectively.
The story in Washington is more complicated. Residential sales prices in the area surrounding the Townhomes on Capitol Hill rose continuously during the sample period, and home prices near the HOPE VI property tended to rise more quickly than prices elsewhere in the area. Yet the price trends showed neither a sustained difference before and after the HOPE VI investment (as evidenced by the similar pre- and post-HOPE VI trend coefficients) nor any significant change in the underlying price level.10 In contrast, we found Wheeler Creek to have had a positive effect on neighborhood property prices, as evidenced by the post-HOPE VI dummy coefficient of 0.512—a statistically significant increase (at the.10 level) from the pre-HOPE VI level. No significant difference occurred in the price trends before and after redevelopment, however; both rates were slightly more negative after the HOPE VI investment. The increase in price levels in the Wheeler Creek neighborhood more than offset the slightly increased negative price trend.
The coefficients displayed in exhibit 4 can be used to simulate the effects of the HOPE VI interventions and thus estimate the effects of the redevelopments on property values. Exhibit 5 shows the model’s estimate of price trends in each of the four neighborhoods, with and without the redevelopment. The estimated trend without redevelopment assumes that the neighborhood’s prices would have maintained their same rate of appreciation relative to the parts of the selected ZIP Codes with no public housing. We estimate that the HOPE VI interventions contributed to an average property appreciation of more than $365,000 in the area surrounding Mission Main and of nearly $198,000 in the Dudley Square area. These gains translated into aggregate property value increases of $107 million and $58 million, respectively. The initial jump in prices associated with Wheeler Creek’s redevelopment translated into an average gain of slightly more than $30,000 per We used demolition as the intervention point for the Townhomes redevelopment. The findings did not change when we used other intervention points or smaller radii to define the impact area.
home, as of 2006, relative to the estimated prices in the absence of redevelopment. The aggregate impact of the Wheeler Creek redevelopment on property values in the surrounding Washington Highlands area was $14 million. Because there was no statistically significant difference in Capitol Hill price amounts or trends before and after the redevelopment of the Townhomes, we cannot attribute any property value changes to that HOPE VI intervention.
$450,000 $400,000 $350,000 $300,000 $250,000 $200,000 $150,000 $100,000 $50,000 $0
Exhibit 5 Estimated Property Value Impacts of HOPE VI Redevelopments (2 of 2) Townhomes on Capitol Hill No Hope VI Townhomes on Capitol Hill $800,000 $700,000 $600,000 $500,000 $400,000 $300,000 $200,000 $100,000 $0
$200,000 $150,000 $100,000 $50,000 $0 HOPE VI Effects on Crime The HOPE VI redevelopment process involved physically redesigning the properties to make them more conducive to public safety. The housing authorities also committed to improved policing of the sites and stricter screening of potential residents for past criminal behavior. We would expect these changes to contribute to reductions in crime rates both on site and in the surrounding neighborhoods. This section explores that hypothesis.
Crime Data We obtained data on Part I crimes (homicides, rapes, aggravated assaults, robberies, burglaries, larcenies, motor vehicle thefts, and arson) from the Boston and District of Columbia police departments. Boston’s data were available at the police reporting area level back to 1990. We compared crime trends in the areas containing and closely surrounding the redeveloped sites with trends in the areas surrounding traditional public housing as well as in the city as a whole. The District of Columbia Metropolitan Police Department (MPD), however, had data going back only to 1998,
Cityscape 113Zielenbach and Voith
and it had changed the boundaries of crime-reporting districts in the early 2000s. We therefore relied on data from DCHA11 specific to individual public housing complexes in the city, which we were able to piece together from 1995 to 2005. Although the data for Washington are not as complete as those for Boston, they nonetheless provide a reasonable sense of the extent of change that has taken place in the selected areas.