November 2013 Volume 26, Number 8 The Abell Report What we think about, and what we’d like you to think about Published as a community service by The Abell Foundation Sandtown-Winchester—’s Daring Experiment In : 20 Years Later, What Are the Lessons Learned?

By: Stefanie DeLuca, Johns Hopkins non-profit Enterprise Foundation that In this report, we examine the long- University and Peter Rosenblatt, would re-invest in Baltimore’s strug- term legacy of these reforms by looking Loyola University Chicago gling inner- neighborhoods. The over time at changes in the fortunes Enterprise Foundation was established of the neighborhood and its residents. n the early 1990s, Sandtown-Win- a decade earlier by James Rouse, the We find increases in homeownership chester was a struggling neighbor- widely-celebrated developer behind the that were likely associated with the Ihood. Entering that decade, more city’s “renaissance,” and Neighborhood Transformation Initia- than 40% of the families living in the founder of Columbia Township. Rouse tive. The neighborhood also benefited neighborhood were below the poverty was looking for a project that could from declines in poverty and crime that line, and more than one in six were showcase a comprehensive approach to represent, in part, trends throughout unemployed. A neighborhood of 72 neighborhood renewal, one that would the whole city of Baltimore during the square blocks just northwest of down- address all of a community’s needs at same time period, making it hard to town Baltimore, Sandtown-Winchester once (Brown, Butler, and Hamilton conclude that the revitalization effort in 1990 was home to roughly 11,500 2001). In Sandtown-Winchester, he led to these changes. We also uncover people in just under 5,000 homes.1 found a neighborhood that could serve challenges that remain in the neigh- Journalistic accounts paint a picture of as an example of this kind of renewal, borhood, including high rates of fore- a neighborhood rife with crime, where one where, in partnership with resi- closure, unemployment, and weakly prostitutes walked the main streets and dents, the Enterprise Foundation could performing schools. open air drug markets were common “demonstrate the successes of raising (Baltimore Sun July 21, 1991). Once up understandable models… to trans- Background a thriving part of Baltimore’s middle form the neighborhoods in which the class African American community, the very poor people live in this country in The Sandtown-Winchester Neigh- neighborhood saw Billie Holiday sing a serious and constructive way” (quoted borhood Transformation Initiative at the Royal Theatre on Pennsylvania in The Baltimore Sun, July 23, 1995). Since the late 1980s, residents, local Avenue, Thurgood Marshall gradu- Over the next decade, the neighbor- government, and private foundations ate from the local Frederick Doug- hood transformation would grow into a have worked together to bring new lass High School, and families live on comprehensive effort to not just rebuild opportunities to Sandtown. Physical steady paychecks from local employers physical structures, but also empower redevelopment of the area has been the like Schmidt’s Bakery (Baltimore Sun residents, through employment, edu- most high-profile aspect of the neigh- November 23 1993). But by the 1990s, cation, and health services outreach. borhood transformation. The transfor- the bakery had closed, many middle Drawing on public and private funding, mation began in the later 1980s with class families had left, and the glory the neighborhood transformation efforts a plan to construct 223 new houses days of the neighborhood seemed to be raised more than $130 million to invest for low-income homeowners. Former in the past. in the neighborhood through the year President was on hand In the late 1980s, newly elected Bal- 20002. These efforts have attempted in 1992 as Habitat for Humanity timore Mayor Kurt Schmoke helped to transform the quality of life in Sand- announced that it would renovate 100 to orchestrate a partnership between town-Winchester and make the neigh- vacant homes in the area, a pledge that the city, community leaders including borhood a place where families and was fulfilled in 1998. Almost 600 units Baltimoreans United in Leadership children can realize opportunities long of public housing in Gilmor Homes and Development (BUILD), and the denied to many of the nation’s poor. were modernized in the early 1990s. continued from page 1 jobs,” such as health outreach work- Overlapping efforts, which incor- ers, family and youth counselors, or porated resident input throughout, Mayor Schmoke also pledged to reno- writers in the community newspaper made Sandtown-Winchester’s Neigh- vate 600 vacant houses in the neigh- (Costigan, 1996). In 1996, Sandtown borhood Transformation Initiative one borhood within one year, although this Works began job readiness training and of the most well-known urban revital- timeline proved overly ambitious—in placement services to teens and adults ization projects in the country (Goetz 1996, the Sun reported that the city in the neighborhood, and the following 1996; Schorr, 1993). The neighbor- was only beginning to make good on year Jobs Plus was established to pro- hood transformation spurred residents, its pledge. Funding for further devel- vide employment services to residents local government, and private develop- opment came in 1997, when Sandtown of Gilmor Homes, a 571-unit low-rise ers to act together to overcome what was selected as one of six neighbor- housing project in the neighborhood. James Rouse referred to as the “great hoods in the country to receive a $5.2 In 1993, Sandtown was included in failure in the United States” to recog- million federal “homeownership zone” Baltimore’s “empowerment zone,” a nize the deplorable conditions of inner grant from the Department of Housing Clinton-administration program that city neighborhoods and come together and Urban Development. This grant provided tax credits to employers who to do something about them (quoted attracted another $30 million in public hired residents from the area. in the Baltimore Sun September and private funding, and in 2000, 30,1995; see also Brown, Butler, and the Enterprise Foundation began sell- s Hamilton 2001). ing renovated and newly built homes to first-time homebuyers. By 2005, Data and Methods Enterprise’s Sandtown Square develop- “Overlapping efforts, ment was complete, and included 236 which incorporated resident We marshal a variety of data sources homes for low and moderate-income on the Sandtown-Winchester neigh- homebuyers in the neighborhood. By input throughout, made borhood, spanning the period from the early 2000s, 700 houses and apart- Sandtown-Winchester’s 1990 (before any of the reforms began) ments had been built or renovated in Neighborhood Transformation to 2011 (twenty years later). We use the neighborhood (Proscio 2004). Initiative one of the most well- multiple sources to capture the pos- Housing was not the only aspect sible effects of the reforms on hous- of the project, however. In 1994 the known urban revitalization ing, education, employment, and to community established a “Compact projects in the country.” explore changes in quality of neighbor- Schools” agreement to increase student hood life, including crime and over- achievement through developing cur- t all life expectancy (for more detailed riculum, instruction, and increasing analysis, see the full report, available at parental involvement in the three ele- www.abell.org/publications). mentary schools in the neighborhood— On the health front, the Vision for Our analysis involves two com- Gilmor Elementary, George C. Kelson Health Consortium (VFH), made up ponents that help us understand the Elementary, and William Pinderhughes of staff from established Baltimore hos- scope of the changes in Sandtown- Elementary. Curriculum development pitals including Bon Secours, the Uni- Winchester. First, we compare census took place over the next three years, versity of Maryland School of Nursing, and school data from before the major with Direct Instruction, a teacher- and the Baltimore City Health Depart- reinvestment provisions were started, to directed curriculum, introduced in the ment, undertook a number of initiatives the most recent census and school data 1997 and 1998 school years. to improve resident well-being. Door to available. These comparisons give us a The reforms also extended to door visits and case management were basic picture of how each component employment and health. From its undertaken to reduce infant mortality of the intervention (i.e., housing revi- inception, the Neighborhood Trans- in the neighborhood, and in 1998 the talization, school reforms) changed the formation Initiative employed resi- VFH began substance abuse treatment neighborhood. Second, we compare dents in “community improvement and prevention. the changes in Sandtown-Winchester

The Abell Report is published quarterly by The Abell Foundation 111 S. Calvert Street, 23rd Floor, Baltimore, Maryland 21202-6174 • (410) 547-1300 • Fax (410) 539-6579 The Abell Reports on the Web: www.abell.org

2 continued from page 2 neighborhood had a poverty rate almost the demographic profile of Sandtown- twice as high as the average Baltimore Winchester over the past 20 years. The to those seen in similar neighbor- neighborhood, and almost four times poverty rate dropped significantly, by hoods in Baltimore City and their zone as high as the average neighborhood in more than eight percent, to 33.1% by schools. We have chosen three “con- the Baltimore . Fewer 2009. This is especially significant trol” neighborhoods that were similar than half of the residents of the neigh- given the overall lack of change in the to Sandtown-Winchester in 1990: The borhood had completed high school, poverty rate of the average Baltimore Upton/Druid Heights neighborhood while only 4.4% had college degrees. city neighborhood. The neighborhood just east of Sandtown-Winchester, the More than one in six adults in the neigh- also saw an increase in the number of Penn North/Reservoir Hill neighbor- borhood – or 17.6% -- were unable to high school graduates and people with hood just to the north of Sandtown- find work in 1990, an unemployment college degrees. The census data do not Winchester, and the Greenmount East rate which was higher than that in the follow the same families over time, so neighborhood in east Baltimore. average Baltimore neighborhood, much we do not know if these changes are While not a perfect research design higher than the metropolitan average, due to more educated families moving for testing whether or not the Neigh- and even higher than in the city’s other into the neighborhood, or greater num- borhood Transformation Initiative segregated neighborhoods. To put this bers of long-term residents finishing caused the neighborhood changes seen in context, the 1990 unemployment school. There was also a significant in Sandtown-Winchester from 1990- rate in neighborhoods that were more increase in the percentage of home- 2010, this analysis allows us to imagine than 70% black was 14%, compared owners in Sandtown-Winchester, from what the impacts might have been had to 5.9% in the city’s majority white fewer than one-quarter of the neighbor- the program never occurred. We do this neighborhoods (Source: 1990 Census). hood’s population in 1990 to more than by comparing the change in Sandtown- The neighborhood was 98% African- a third (35.6%) in 2009. This increase Winchester’s neighborhood and educa- American, part of the city’s segregated is likely a result of the Neighborhood tional opportunity to that in the three west side that stretched northwest along Transformation Initiative’s focus on matched “control” neighborhoods that Liberty Heights Ave almost to the Bal- homeownership, especially given the did not take part in policy interven- timore County line. large increase in homeownership from tions like those that occurred in Sand- 24.2% to 32.6% of the population town-Winchester. These “control” s between 1990 and 2000. neighborhoods did not take part in the Despite these positive changes, the Sandtown-Winchester Neighborhood neighborhood still lagged behind the Transformation Initiative, but they “There was also a rest of the city and metropolitan area in were other very low income, racially significant increase in the important ways. The median household segregated neighborhoods that were income increased slightly between 1990 facing similar social and economic chal- percentage of homeowners in and 2000, but remained well below city lenges in the 1980s and 90s. However, Sandtown-Winchester, from and metropolitan averages. On average, given the time period covered in this fewer than one-quarter of the families in Sandtown-Winchester were analysis, it is not surprising that they of lower socio-economic status (both were sites of other reform efforts, nota- neighborhood’s population income and education level) than most bly the demolition of Murphy Homes, in 1990 to more than a third families in Baltimore and in the metro- a large housing project in Upton, which (35.6%) in 2009.” politan area. It isn’t clear whether this was demolished and rebuilt as a mixed is a function of more people moving to income community as a result of the the area who are above the poverty line HOPE VI program in the early 2000s. t but still low income, or changes in the We further discuss these neighborhood existing residents’ incomes. There were changes and their implications below. Table 1 (page 4) also shows how fluctuations in the unemployment rate the neighborhood has changed in the across the two decades following the Findings two decades following the Neighbor- start of the neighborhood transforma- hood Transformation Initiative. The tion initiative, but by 2009 more than Neighborhood lower sections of the table profile the one in five adults—or 21%—in the Demographic Changes change in the average Baltimore city Sandtown CSA was unemployed. In 1990, Sandtown-Winchester neighborhood and the change in the Table 1 provides some context for was among the poorest neighborhoods average metropolitan neighborhood. the changes in Sandtown-Winchester in Baltimore. Table 1 shows that the There have been some notable shifts in by showing city and metropolitan-level

3 Table 1

Sandtown-Winchester Baltimore City Baltimore Metro Area

Change Change Change Value Value Value since 1990 since 1990 since 1990 1990 Percent Below Poverty 41.4 % 23.0 % 11.0 % Median Household income $23,776 $42,683 $65,918 (2009 dollars) Percent African-American 98.2 % 55.6 % 25.5 % Percent with High School 45.6 % 58.9 % 73.0 % Diploma Percent with BA 4.4 % 15.0 % 22.4 % Unemployment Rate 17.6 % 10.0 % 5.6 % Percent Owner- 24.3% 48.6% 63.4% Occupied Homes 2000 Percent Below Poverty 37.0 % - 4.4% 24.6 % + 1.6% 12.1 % + 1.1% Median Household income $25,079 + $1,303 $39,063 - $3,620 $64,107 - $1,811 (2009 dollars) Percent African-American 98.6 % + 0.4% 62.8 % + 7.2 % 30.5 % + 5.0% Percent with High School 54.1 % + 8.5% 66.2 % + 7.3 % 78.8 % + 5.8% Diploma Percent with BA 4.8 % + 0.4% 9.9 % - 5.1% 15.5 % - 6.9% Unemployment Rate 18.2 % + 0.6% 12.4 % + 2.4% 6.6 % + 1.0% Percent Owner- 32.6% + 8.3% 50.3 + 1.7% 66.7 + 3.3% Occupied Homes 2009 Percent Below Poverty 33.1 % - 8.3%v 22.4 % -0.6% 12.1 % + 1.1% Median Household income $22,237 - $1,539 $40,036 -$2,647 $65,005 + $913 (2009 dollars) Percent African-American 99.0 % + 0.8% 64.4 % +8.8% 33.2 % + 7.7% Percent with High School 64.5 % + 18.9% 75.0 % +16.1% 84.3 % + 11.3% Diploma Percent with BA 6.2 % + 1.8% 23.4 % +8.4% 31.1 % + 8.7% Unemployment Rate 21.0 % + 3.4% 12.3 % +2.3% 7.5 % + 1.9% Percent Owner- 35.6% + 11.3% 51.1% +2.5% 68.4% + 5.0% Occupied Homes

4 Table 2

Penn North/ Sandtown-Winchester Greenmount East Upton/Druid Heights Reservoir Hill Change Change Change Change Value Value Value Value since 1990 since 1990 since 1990 since 1990 1990 Percent Below Poverty 41.4 % 42.0 % 49.0 % 38.8 % Median Household $23,776 $ 27,702 $22,735 $28,110 income (2009 dollars) Percent 98.2 % 98.6 % 98.1 % 94.0% African-American Percent with High 45.6 % 40.7 % 47.1 % 55.0 % School Diploma Percent with BA 4.4 % 2.7 % 6.1 % 12.6 % Unemployment Rate 17.6 % 21.2 % 16.3 % 11.4 % Percent Owner- 24.3% 30.2% 22.7% 23.1% Occupied Homes 2000 Percent Below Poverty 37.0 % - 4.4% 37.1 % - 4.9% 47.4 % - 1.6% 35.1 % - 3.7% Median Household $25,079 +$1,303 $24,099 -$3,693 $19,543 -$3,192 $29,175 +$1,065 income (2009 dollars) Percent 98.6 % + 0.4% 97.0 % - 1.6% 95.7 % - 2.4% 93.1 % - 0.9% African-American Percent with High 54.1 % + 8.5% 53.5 % +12.8% 52.6 % + 5.5% 64.2 % + 9.2% School Diploma Percent with BA 4.8 % + 0.4% 4.3 % + 1.6% 9.3 % + 3.2% 13.9 % + 1.3% Unemployment Rate 18.2 % + 0.6% 20.8 % - 0.4% 20.7 % + 4.4% 19.0 % + 7.6% Percent Owner- 32.6% + 8.3% 40.0% + 9.8% 16.1% - 6.6% 28.4% + 5.3% Occupied Homes 2009 Percent Below Poverty 33.1 % - 8.3% 34.1 % - 7.9 % 47.4 % - 1.6% 23.0 % - 15.8% Median Household $22,237 -$1,539 $19,691 -$8,011 $16,995 - $5,740 $31,116 +$3,006 income (2009 dollars) Percent 99.0 % + 0.8% 95.5 % - 3.1% 94.0% - 4.1% 93.5 % - 0.5% African-American Percent with High 64.5 % +18.9% 66.8 % +26.1% 61.9 % +14.8% 79.0 % +24.0% School Diploma Percent with BA 6.2 % + 1.8% 8.1 % + 5.4% 10.6 % + 4.5% 16.0 % + 3.4% Unemployment Rate 21.0 % + 3.4% 19.7 % - 1.5% 17.5 % + 1.2% 19.0 % + 7.6% Percent Owner- 35.6% +11.3% 38.0% + 7.8 % 16.2% - 6.5% 32.3% + 9.2% Occupied Homes

5 continued from page 3 graduates than Sandtown-Winchester the Sandtown-Winchester neighbor- in both the 1990s and 2000s. Upton hood still faces a number of challenges. changes over the same time period. But and Greenmount East had comparable The decline in neighborhood poverty how did the neighborhood change rela- proportions of college and high school and increase in the proportion of high tive to other places that were also dis- graduates to Sandtown-Winchester in school and college educated residents is advantaged in 1990? Table 2 (page 5) both 2000 and 2009, although the promising for the overall prosperity of compares demographic change in Sand- median income of individuals in these the neighborhood, but our comparison town-Winchester to our three “com- neighborhoods lagged behind that analyses show that similar improve- parison” neighborhoods: Greenmount of Sandtown-Winchester residents. ments were taking place in other poor East, Upton/Druid Heights, and Penn Greenmount East and Penn North neighborhoods that did not have the North/Reservoir Hill. also saw an increase in homeowner- same revitalization strategy as Sand- The top rows of Table 2 show that ship, although the increase of 11% in town. Most striking in this respect is the three comparison neighborhoods Sandtown was the most dramatic of the change in Greenmount East, which were similar to Sandtown-Winchester the four neighborhoods. We look fur- was similar to Sandtown-Winchester in many ways in 1990. All were much ther at changes in the neighborhood in 1990 in terms of concentrated poorer and more segregated than the housing market in the next section. poverty and neighborhood socio-eco- average city neighborhood, had median Table 2 further highlights the chal- nomic profile. Yet despite not having incomes well below the citywide average, lenge of reducing unemployment in the same kind of intensive community and double-digit unemployment rates. Sandtown-Winchester. While the transformation efforts as Sandtown- They were also located in the inner-city, unemployment rate also increased in Winchester, the Greenmount East and were of similar geographic size to Upton and Penn North/Reservoir Hill neighborhood saw similar declines in Sandtown-Winchester.3 Greenmount since 1990, Sandtown-Winchester was poverty, and by 2009 was still compa- East was the most similar to Sandtown- the only neighborhood of the four rable on a number of selected census Winchester in 1990, but we also chose that saw an increase in unemploy- measures to Sandtown-Winchester, one neighborhood that was slightly ment during the 2000s (i.e. that had including home ownership. worse off (Upton) and one that was a higher unemployment rate in 2009 slightly better off (Penn North/Reser- than in 2000). By 2009, Sandtown- Housing voir Hill) to give upper and lower bound Winchester had a slightly higher Rebuilding the neighborhood’s comparisons. These three comparison unemployment rate than any of the housing stock was one of the central neighborhoods allow us to explore the three comparison neighborhoods4. components of the Neighborhood changes that were happening in similar On the whole, Table 2 suggests that Transformation Initiative. While places during the past 20 years, and so Figure 1 – Median Housing Sale Price give some idea of what might have hap- pened in Sandtown-Winchester had the Neighborhood Transformation Initia- 250000 tive not taken place. The comparisons show that the 200000 drop in poverty rate we saw in Table 1 was not unique to Sandtown-Win- chester. Greenmount East, which 150000 did not have a comparable neighbor- hood transformation strategy, saw a

Sale Price Sale 100000 similar drop in poverty during both the 1990s and 2000s, although the neighboring Upton community did 50000 not. Penn North/Reservoir Hill had an even greater drop in poverty rate 0 during the 2000s, down to 23% poor 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 by 2009. Other measures of neigh- Year borhood socio-economic status also improved in Penn North, which saw Baltimore City Sandtown-Winchester Upton greater gains in the median income and Greenmount East Penn North/Reservoir Hill the number of high school and college

6 continued from page 6 Figure 2 – Foreclosures by Year most of the Sandtown housing develop- ments were initiated during the 1990s, 180 these data allow us to look at how the neighborhood housing market fared in 160 the decade following the major reforms, 140 and how it compared to similar Balti- more neighborhoods. 120 Figure 1 looks at changes in neigh- 100 borhood housing markets by showing the median sale price of housing in 80 each of the four neighborhoods, as well 60 as in Baltimore city as a whole. The solid red line in the figure shows that 40 the median housing sale price more 20 than tripled during the housing market boom between 2004 and 2006 but that 0 prices plummeted when the national City Avg Sandtown Greenmount Penn North Upton housing bubble burst in 2007. In 2006, East the median house in the neighborhood 2008 2009 2010 was selling for slightly more than the median house in the average Baltimore city neighborhood, and the median sale price of a home in Sandtown- annual number of foreclosures in the Education Winchester increased by more than average Baltimore city neighborhood. Education reforms in Sandtown- $100,000 since 2001. This increase We can see that, of the four neighbor- Winchester focused on the neigh- far surpassed that in Greenmount hoods in our comparison, Sandtown borhood’s three elementary schools, East, and the gentrifying community was the hardest hit by the foreclosure Gilmor Elementary, George C. Kelson of Penn North/Reservoir Hill was the crisis. In each year between 2008 and Elementary, and William Pinderhughes only neighborhood to have a higher 2010, the community area saw more Elementary. In this section we explore median sale price during this time. than 100 homes face foreclosure filings, student achievement in the years fol- Yet Figure 1 also indicates that with a peak of 167 filings in 2009.5 lowing the “Compact Schools” agree- the fluctuation in housing prices in Sandtown-Winchester benefitted ment to develop curriculum, improve Sandtown-Winchester was more likely from the booming housing market of instruction, and increase parental due to the local economic climate than the mid-2000s. Not only did home involvement in these three schools. As to any neighborhood-specific reforms. prices rise, but the neighborhood saw we did with neighborhoods, we also All three comparison neighborhoods an increase to more than $10 million compare changes in elementary school saw a similar cycle of housing price in conventional mortgage loans in the performance in the three Sandtown- increases followed by declines. The mid-2000s (see longer report for more Winchester elementary schools with figure also emphasizes the fleeting detailed analysis). Unfortunately, three elementary schools that serve each nature of housing speculation: by 2009, however, this level of investment was of our comparison neighborhoods6. All housing in all four communities was not sustained as Sandtown was hard of these schools were more segregated again selling for much less than the Bal- hit by the foreclosure crisis, and also (99% African-American on average) timore city average. saw an increase in vacant and aban- and served a predominantly lower- The bursting of the housing bubble doned buildings toward the end of the income population (84% eligible for led to a wave of foreclosures in many 2000s. This increase in foreclosures reduced price lunches) than the average communities, and Sandtown-Win- and vacant buildings suggests that the elementary school in Baltimore City chester was not spared. Figure 2 shows housing reforms were not enough to or Central Maryland during the two how the number of foreclosures in stave off the kind of decline that was decade period of this study. Sandtown and the comparison neigh- happening in the larger community, Our research looked at school perfor- borhoods changed between 2008 and and also in similar neighborhoods mance on statewide tests between 1993 2010. The leftmost columns show the during the past decade. and 2011. Most of the neighborhood

7 continued from page 7 across the decade. As a whole, the have been affected by the other reforms schools in the Sandtown-Winchester as residents became more empowered schools we profiled struggled to reach neighborhood did not experience the by the Neighborhood Transformation state and federal benchmarks for stu- same increase in reading and math Initiative. Figure 4 shows changes in dent achievement during this period. proficiency that other neighborhoods’ reported crime per 1,000 neighborhood Schools in Sandtown-Winchester schools did. residents for Sandtown-Winchester, the showed some improvement in math Overall, the education data suggest three comparison neighborhoods, and following the implementation of new that Sandtown-Winchester’s three ele- Baltimore city as a whole. curriculum in 1998, but as a whole the mentary schools continued to struggle We can see that the overall crime schools continued to perform below in the two decades following the Com- rate declined across the city during the the average for Baltimore City across pact Schools agreement. While our 2000s. The crime rate in Sandtown- the 1990s in both math and reading. data are limited in their ability to tell us Winchester was very close to the Bal- Figure 3 compares the performance of what triumphs or barriers teachers and timore average, as was the crime rate elementary schools in the four neigh- students have met in the classroom, the in Greenmount East (dashed green borhoods to city and statewide averages persistent difficulties to reach state and line) and Penn North/Reservoir Hill in reading under the No Child Left federal benchmarks for achievement (dash-and-dotted light blue line). Only Behind assessments, which began in tests show that the Sandtown schools Upton had a markedly higher rate of 2003. The dark dashed and dotted line have not reached the performance level crime during this period, although it shows the proficiency rate in all Cen- of other city and metropolitan schools. declined by more than half between tral Maryland schools, while the lighter 2000 and 2009. dashed line shows the proficiency rate Health and Public Safety Finally, we look at the overall life in all Baltimore City schools. There is While crime or policing programs expectancy in each neighborhood, a general increase in the number of stu- were not as central to the Neighbor- as calculated by the Baltimore City dents scoring at proficiency levels across hood Transformation Initiative as Health Department. This figure pro- the decade of the 2000s. Yet we also housing and education were, changes vides a dramatic way to understand see that Sandtown-Winchester schools in crime and violence in the neighbor- the quality of life in a neighborhood, (solid red line) performed worse than hood give us a picture of the overall by looking at how long residents live the schools in other neighborhoods quality of community life, and could on average, compared to the city as a Figure 3 – Percent Proficient or Higher, Reading

100

90

80 Central Maryland 70 Baltimore City 60

50 Sandtown-Winchester

40 Upton

30 Penn North/Reservoir Hill

20 Greenmount East 10

0 2003 2004 2005 2006 2007 2008 2009 2010 2011

8 continued from page 8

Figure 4 – Crime Rate The number of reported criminal offenses per 1,000 residents (as reported by 2000 Census); Part I offenses include murder, aggravated assault, rape, attempted rape, burglary, larceny, and auto theft.

180

160

140

120

100

80

60

# Reported Crimes per 1,000 Crimes # Reported 40

20

0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year

Baltimore City Penn North/Reservoir Hill Upton

Sandtown-Winchester Greenmount East whole. Data from 2008 and 2011 are Figure 5 – Life Expectancy in Years presented as a comparison. All four comparison neighborhoods had shorter 74 life expectancies than the city average. Upton, with a life expectancy eight to 71.8 72 70.9 nine years lower than the city average, 69.4 had the lowest of any neighborhood in 70 the city. Yet Sandtown was not much 68.1 68 better off, with an overall life expec- 65.3 65.9 tancy of around 65 in both 2008 and 66

Years 64.9 65 2011, six to seven years shorter than the 64 city average. 62.862.9 Conclusions 62 60 Sandtown Winchester Over 20 Years 58 In this report, we profile the neigh- Baltimore Sandtown Upton Penn North Greenmount borhood-level changes that residents City East of Sandtown-Winchester have experi- 2008 2011 enced over the past two decades. There

9 continued from page 9 findings also highlight the challenges environment that high crime and open of community development strategies air drug dealing create. We show that are some positive changes to report— that focus heavily on housing. Home- crime decreased in Sandtown across the the poverty rate of the neighborhood ownership brings the potential for gen- 2000s, mirroring a citywide trend. Yet dropped in this period, and more people erating wealth, and is especially signifi- there were still more than 8 homicides in the community became homeown- cant given the history of redlining and per 10,000 people in the neighborhood ers. The neighborhood also became lending discrimination that has limited in 2011, a rate that is double the Balti- home to a greater proportion of people homeownership in African-American more city average and higher than com- with high school and college degrees. communities, especially in Baltimore. parison communities. Sandtown resi- Crime data indicates that Sandtown- However, homeownership can become dents also have shorter life expectancies Winchester became a less violent place a burden on poor families if the rest of than the Baltimore city average, speak- during the 2000s, as both the violent the neighborhood does not improve, by ing to the level of material and physical crime rate and the overall crime rate anchoring them to low-resource com- hardship still present in the community. fell. Baltimore City as a whole also saw munities and also putting them at risk We have attempted to put the a decrease in crime during this period. of predatory lending practices which changes seen in Sandtown-Winchester Challenges still remain in the have been shown to target families in in context by comparing them to three neighborhood, however. Unemploy- less advantaged and minority neighbor- neighborhoods that were similar in ment remained high throughout the hoods (Shlay 2006; Stein 2001; HUD 1990, to see which changes might have 1990s and 2000s. This could be due 2000). Further research is needed to been brought on by the redevelopment, to the recession during the later years uncover the factors driving the foreclo- and which were likely rooted in broader of the 2000s, which affected employ- sures in Sandtown. economic or political processes. For ment opportunities for many families, example, the drop in the neighborhood and has also hit African-American s crime rate, while important for under- families particularly hard (Weller and standing the quality of life in the neigh- Fields 2011). The unemployment rate borhood, was shared across all four in 2009 was higher in Sandtown-Win- “Overall, the picture that neighborhoods and the city as a whole, chester than in the city’s other segre- emerges from the data under- suggesting that it was probably a result gated neighborhoods. scores the durability of social of citywide trends, rather than reforms Our housing analysis shows some specific to Sandtown. By contrast, the potential benefits from the reforms in inequality and the persistence increase in Sandtown’s homeownership Sandtown, as the neighborhood saw a of overlapping social problems rate between 1990 and 2009 was more dramatic rise in homeownership, from in high poverty and racially dramatic than in other neighborhoods, less than a quarter of the population in a result that is likely due to the housing 1990 to more than 35% in 2009, and segregated neighborhoods.” focus of the Neighborhood Transfor- millions in mortgage dollars flowed mation Initiative. into the neighborhood in 2006 and t Overall, the picture that emerges 2007, while housing prices reached from the data underscores the durability levels higher than the city average. The data also show that despite of social inequality and the persistence While this rate was not as dramatic as the educational investments made in of overlapping social problems in high that which took place in neighboring Sandtown-Winchester, the neighbor- poverty and racially segregated neigh- Penn North, it was much greater than hood elementary schools have struggled borhoods (Sampson 2012). All four the level of investment in Greenmount to improve their academic performance neighborhoods profiled lag behind Bal- East. However, Sandtown was also hit in statewide reading and math assess- timore City and the greater metropoli- hard by the foreclosure crisis —more ments. Across the 1990s and 2000s, tan area on most measures. Low family than 100 homes in the community Sandtown schools lagged behind Cen- incomes, high unemployment, high area were foreclosed on in recent years. tral Maryland and Baltimore city on crime, high poverty rates, and racial seg- The reinvestment in Sandtown-Win- reading and math assessments, and regation are intimately linked in Ameri- chester also seems to have done little did not stand out when compared to can , and have proven tenacious to arrest the growth of vacant housing schools in similar poor neighborhoods. in the face of changing economic and in the wider community area, which Journalistic accounts of Sandtown policy contexts (Sampson 2012, Massey in 2009, comprised nearly one third of in the early 1990s portrayed a com- and Denton 1993; Wilson, 2012). Yet properties in the neighborhood. These munity struggling with the oppressive rather than give up on the families and

10 continued from page 10 programs through rigorous evaluation partnership with community members. provides a picture of what works or does While mobility programs and com- children who continue to face danger not work, but understanding how to munity development are sometimes in their neighborhoods and diminished repeat successes or adjust to challenges seen as at odds with each other, these prospects for educational and employ- requires understanding the details “on results suggest that we need to under- ment opportunities, we believe that this the ground.” Prior work on the imple- stand how both types of programs can descriptive look at Sandtown points to mentation of the Neighborhood Trans- be used to help families escape per- the need for more evidence-based social formation Initiative by Brown and col- sistent neighborhood disadvantage. policy intervention alternatives. leagues (Brown, Butler, and Hamilton Mobility programs allow poor families 2001) has highlighted the importance Implications for of nurturing partnerships with the s Research and Policy neighborhood’s residents and negotiat- In the recent past, we have seen a ing the race and class divides that make “We need to develop rigorous great deal of research attention and comprehensive community initiatives scrutiny given to programs that pro- challenging. Future evaluations could standards for research that eval- vide poor families with resources to live include examinations of: how much local uates community level interven- in less poor neighborhoods (Ludwig, community involvement and “buy-in” tions with methods that allow us 2012; Edin, DeLuca, and Owens exists; how partnerships are formed (and to distinguish between program- 2012). However, despite significant severed) between community organiza- specific outcomes and outcomes federal and local investments, we rarely tions, state, local, and federal agencies; see community level or place-based how the populations of these commu- which are due to broader policies held to the same empirical nities change over time; cost-effective economic or social trends.” evaluation standards—which prevents ways to handle these dynamics and their us from learning as much as we could related administrative challenges; and t from these efforts. We need to develop best practices culled over more than 20 rigorous standards for research that years of experience. This research should to leave violent neighborhoods in the evaluates community level interven- also seek to understand the interface short run, instead of being trapped in tions with methods that allow us to between policy change and the dynam- the low-performing schools and poor distinguish between program-specific ics of daily life for low-income families. quality housing that exist while their outcomes and outcomes which are due These families face myriad challenges, communities await larger redevelop- to broader economic or social trends. from mental and physical health con- ment investments. Rigorous research Our design has the advantage of using cerns, to family and kin networks that on mobility programs that considers comparison neighborhoods that could provide social support but also present causality and compliance has revealed have been candidates for a revitalization challenges, to exposure to trauma and some of the pitfalls as well as the suc- effort, but stops short of being able to parse violence. All of these factors influence cesses of that strategy (DeLuca, 2012). out the degree to which the programs in the way low-income families respond The same evaluation standards applied Sandtown’s Neighborhood Transforma- to policy opportunities (Sharkey, to community development approaches tion were a direct cause of the changes 2010; Theodos et al, 2012; DeLuca can help us understand what works and observed. Future research on place-based and Rosenblatt 2010). We strongly what does not, and allow us to develop programs could work to improve basic encourage the Obama administra- policies to preserve inner city commu- pre-test / post-test evaluations through tion to support such evaluation and nities and increase the scope of ben- the use of more sophisticated comparison implementation research alongside the efits to extend to families already living groups and statistical models that account administration of its more compre- there. While there are no single policy for differences between neighborhoods (see hensive urban development initiatives, solutions to reduce inequality in our also Galster et al. 2004). such as Promise Neighborhoods and urban communities, we need to ensure In addition to understanding how Choice Neighborhoods. that high quality research is conducted to estimate the effects of community The findings here point to the dif- on all of these innovative programs to redevelopment efforts on neighbor- ficulties of overcoming the overlap- show us what kinds of resources matter, hood, family and child outcomes, it is ping disadvantages that accrue in poor and how they should be targeted. The also important to evaluate the imple- and segregated neighborhoods, even depth of this challenge means that we mentation of place-based programs. with ample resources and, as in the cannot afford to leave any potential Understanding the consequences of case of Sandtown-Winchester, in full policy solution out of the equation.

11 About the Authors

Stefanie DeLuca is an Associate Professor of Sociology Peter Rosenblatt grew up in Baltimore and holds a at the Johns Hopkins University. She conducts research Ph.D in Sociology from Johns Hopkins University. He involving sociological considerations of education and is currently an Assistant Professor of Sociology at Loyola housing policy. She has carried out mixed-methods stud- University Chicago. His research focuses on housing ies of programs to help public housing residents relocate policy, urban inequality, and education. He uses mixed to safer neighborhoods and better schools, understand methods to understand the way individuals respond to the tradeoffs low and moderate income parents make in constraints and opportunities in the housing market, deciding where to live, and the behavior of urban land- and the way housing markets themselves shape and lords. Her other research examines the timing of educa- are shaped by housing policies. His work has resulted tional transitions, and the transition to college and work in multiple publications for both academic and policy for inner city young adults. She contributes regularly to audiences, and includes studies of policy interventions national and local media, and her work has been pub- such as Moving to Opportunity and HOPE VI. lished in numerous academic journals.

Endnotes

1 These data are aggregated from 1990 census There are 55 CSAs in Baltimore City. In analyses block groups. not shown here, but available in our longer report, we 2 In August 1996, The Baltimore Sun reported “$100 also profiled the changes in the Sandtown-Winchester million spent or earmarked for housing alone.” The neighborhood using census block-group data. largest of these expenditures were on housing, includ- 4 It is possible that more residents of the neighborhood ing $17.4 million for the Nehemiah homes, a $5.2 were looking for work in 2009. In analyses not shown million “homeownership zone” grant from HUD, here, we looked at the percentage of people out of which attracted an additional $30 million that the the labor force and found that it declined since 1990, Enterprise Foundation used for its Sandtown Square down to 45%, although this percentage matches development in the neighborhood. These figures do that in Greenmount East, which had a lower not include non-housing expenses, such as the $4 unemployment rate. million renovation of Lafayette Market, but housing 5 A foreclosure filing is a legal action in the Baltimore was certainly the largest expenditure over the course City Circuit Court, representing the initiation of a of the neighborhood transformation initiative. In legal procedure by a lending agency to reclaim the title other words, the financial investment in Sandtown- to a property due to non-payment on a mortgage loan. Winchester was significant. By way of comparison, Not all filings result in actual foreclosures. Under the city was granted $128 million through the HOPE Maryland law, foreclosure filings take place no sooner VI program in the 1990s to demolish more than 3200 than 90 days after the loan is in default. units of public housing and rebuild mixed income 6 The comparison schools are Eutaw Marshburn, communities in five neighborhoods in the city, includ- Furman L. Templeton, and Samuel Coleridge Taylor ing $31.3 million to demolish and rebuild the George in Upton; Westside Elementary, John Eager Howard, Murphy Homes public housing high rise in neighbor- and Mt. Royal Elementary/Middle in Penn North; ing Upton in 1997. Dr. Bernard Harris, Johnston Square Elementary, 3 For this analysis, we used Community Statistical and Harford Heights Primary/Intermediate in Green- Areas (CSAs), which are locally-defined geographic mount East. We use test score data from 3rd and 5th areas comprised of 3-6 census tracts. Sandtown-Win- grade reading and math in order to maintain our focus chester and the three “counterfactual” neighborhoods on comparing elementary school performance. each consist of one CSA, and so are of similar size.

12