Association for Public Policy Analysis & Management

The use of community revitalization initiatives has grown considerably in recent years, from federal programs like Choice Neighborhoods to initiatives backed by philanthropic and anchor institutions. While the prevalence of these efforts has grown considerably, our knowledge of their effects has not grown commensurately. Like the communities they seek to change, revitalization initiatives are complex and have proven extremely difficult to evaluate well.

Given the challenges of randomly assigning interventions at a neighborhood-level, quasi-experimental evaluation approaches are needed to estimate the impact of most revitalization efforts. This study uses three different modeling approaches to assess neighborhood-level impacts of three interventions. The modeling approaches are: matching, fixed effects, and synthetic control (Abadie, Diamond, and Hainmueller 2010). The study tracks neighborhood trajectories over an extended time horizon, using data from the Decennial Census (1970 through 2000), and ACS (2006-10 continuing annually through 2012-16). The research questions are: 1) What effects did large place-based initiatives have on their target communities along key impact measures?; 2) How similar or different are the effects of the initiatives?; 3) To what extent do benefits “spill over” into communities surrounding the areas targeted by these initiatives?; and 4) How comparable are estimated results using different modeling approaches?

The three initiatives studied here are Atlanta’s East Lake Initiative, Baltimore’s East Baltimore Revitalization Initiative, and San Diego’s City Heights Initiative. All three share some common elements. They are large, having invested in the hundreds of millions of dollars. They are spatially focused on small, spatially congruent areas. They are multi-dimensional strategies that include a mix of physical redevelopment and human services. For each, these supports have been deployed over several years. Each initiative has a “backbone” organization, but the efforts combine public, philanthropic, and private capital. Finally, all three began in the late 1990s or early 2000s—meaning they have been active long enough to observe any neighborhood-level effects. The three initiatives have differences as well that make comparison interesting. Most notably, the market strength of the local areas varies, as does the mix of services and redevelopment.

Key outcome measures of interest for the study include whether the initiatives have generated local changes in economic (income, poverty), housing (home values, rent levels), and demographic (population levels, race/ethnicity) conditions.

Through quantitative case study impact investigation, this study explores multiple initiatives in terms of their development, context, and localized effects. The study holds several implications for practice, policy, and research. Relatively little is known, about whether revitalization initiatives can change community attributes over the long-term, and if so, along which dimensions. Little is known about how much capital needs to be invested to achieve gains; what types of capital are needed; what mix of services is beneficial; and whether a series of concentrated, sustained local investments can have meaningful “spill over” benefits for broader areas. Additionally, this study will inform research method development for place-based evaluation by comparing traditional impact estimation approaches with a relatively new approach (synthetic control) that has not yet been used in this research context.