GIS for Housing and Urban Development (2003)

National Academies Press: OpenBook

Chapter: 4. Research and Policy Development

Visit NAP.edu/10766 to get more information about this book, to buy it in print, or to download it as a free PDF.

Suggested Citation:"4. Research and Policy Development." National Research Council. 2003. GIS for Housing and Urban Development. Washington, DC: The National Academies Press. doi: 10.17226/10674.

4
Research and Policy Development

PD&R is directly involved in formulating potential urban policy solutions and monitoring and evaluating policy once it is in place. These policy issues include homelessness, fair housing, housing assistance programs, mortgages and lending, and assisting in community and economic development. PD&R is beginning to realize the potential of GIS in these areas. GIS offers multiple benefits to PD&R in terms of thinking about housing and urban issues and developing coherent public policy responses. Two of the most significant benefits are the ability to layer data from multiple sources and look at data at different scales or geographies.

HUD can use GIS to produce maps that show the distribution of public and federally assisted housing. Information about the spatial distribution of housing could show public housing authorities and other HUD field offices where their clients are and where public housing should be. Public Housing Authorities normally operate their facilities as specialized enterprises, concentrating primarily on the housing units themselves and rarely considering the surrounding framework of neighborhoods, cities, or metropolitan areas. Using GIS locally and building relationships to gather and make available data on housing and other urban conditions could inform policies that affect public housing. Data showing areas of growth in employment opportunities, public transit stops, school district data, prevalence of crime, and other themes relevant to the targeting of HUD resources could improve the agency’s efficiency and effectiveness in meeting mission goals. Currently, HUD’s field offices often lack both adequate data and staff who are proficient in GIS. GIS is a tool for

data management and spatial analysis but the information derived from GIS is only as accurate as the data that went into the system in the first place, and as relevant as the questions posed. Understanding housing markets and the demand and supply of different types of housing is important. These gaps in data and staffing leave local HUD agencies with inadequate information for making decisions about how and where they should allocate their resources for maximum effectiveness.

Using GIS to collect, store and deliver data, and ensuring the quality of the data are important, but the application of these data to policy analysis and planning depends on the relevance of the research questions posed. In addition, the relevant data (e.g., Census of Population and Housing; American Housing Survey, Department of Labor employment data, satellite imagery, EPA air quality data, DOT traffic and accident data, airport noise exposure data) have been collected for many different applications and must be adapted if HUD’s clients and partners are to use them. Data have meaning only within the context of an argument or hypothesis about how something works.

This report adopts a regional/metropolitan-level focus for addressing urban and housing issues, as described in Chapter 1. HUD can expand its research at the regional and metropolitan level to include geographic analysis of the spatial dimensions of urban poverty, the dynamics of neighborhood change, and market trends that affect the U.S. housing markets. This chapter discusses the potential of an expanded urban research agenda that is appropriate for HUD as a federal agency and identifies priorities for geographic analysis of urban and housing issues.

THE SPATIAL DIMENSIONS OF URBAN POVERTY

Understanding urban poverty requires attention to processes at the regional and metropolitan levels that result in inner-city poverty. GIS can help integrate data from multiple levels to facilitate regional analyses. The dynamics of neighborhood change and the factors that concentrate poverty in urban areas can also be analyzed using geographic data and tools. The poor are often spatially segregated from the middle class and physically removed from basic services, such as health care, childcare, and retail, and from cultural amenities, such as libraries and museums. Both the percentage of inner city neighborhoods that are poor and the percentage of poor people living in those neighborhoods have risen in recent decades (Jargowsky, 1997). Similarly, although poverty rates have declined for many groups, the income gap between the rich and the poor is widening (Lichter and Crowley, 2002). Understanding the spatial dimensions of urban poverty and neighborhood change is essential to carrying out HUD’s mission.

Regional Analyses

There is growing scholarly and political recognition of the importance of regional analysis in dealing with the problems of low-income localities. For example, Orfield (1997) and Jargowsky (1997) identify processes that shape the conditions within which housing programs are situated. These processes operate at multiple geographic scales. In the past, poverty and segregation in urban housing markets were often explained in terms of locally specific analyses of individual behavior or neighborhood characteristics, without examining the processes operating at broader scales. Examples of broader or multi-scale processes that impact urban housing markets include middle-class flight to suburbs, patterns of service and high-tech industry siting, and the administrative structures of local, urban, and state government.

Regional spatial analysis provides a more comprehensive account of the problems of poor localities. Efforts toward community empowerment should address the regional processes that create the problems confronting communities and localities. Khadduri and Martin (1997) suggest that data on the positive factors affecting families should be included in analysis in addition to the negative neighborhood factors such as crime and homelessness that are often the focus of analysis. Positive factors may include accessibility, services, formal and informal support networks, and income diversity. GIS can address these multi-scalar questions relevant to urban poverty in terms of these broader forces and processes that shape neighborhoods. Regional spatial analyses of this kind are not simple or user-friendly (Luc Anselin, University of Illinois, Urbana-Champaign, personal communication, 2002); rather they require a trained workforce. Box 4.1 presents an example of sophisticated statistical analysis that supports anti-discrimination efforts.

The committee suggests the following research questions that lend themselves to spatial analysis and may illuminate the relationships between urban and suburban processes, and housing market conditions and trends within low-income communities. Addressing these issues using spatial data requires experience and expertise in geographic analysis and research.

To what extent do policies that promote or allow rapid urban decentralization contribute to decline in inner-city neighborhoods?
How effective are place-based investments in inner-city neighborhoods if rapid decentralization undermines inner-city housing markets? What are the alternative policies?
How effective are growth management policies such as urban growth boundaries and smart growth initiatives in curbing the rate of urban expansion?

Do growth management policies contribute significantly to housing affordability problems?
Is there a threshold on the size of the region for metropolitan governance, beyond which regionalism may not work? How can we integrate “things-regionalism” with “people-regionalism” 1 in urban and housing development?
What is known about urban processes such as sprawl, economic polarization, and population growth; and the interactions among these and other processes at global, national, state, regional, metropolitan, and local scales?

GIS and Section 8 Housing Policy Analysis

This section describes the uses of GIS in analyzing Section 8 housing policy (Box 4.2). The examination of the application of GIS to Section 8 housing issues illustrates the importance of research on the spatial dimensions of urban poverty, the role of data intermediaries, and data issues including privacy concerns and the determination of causality. Using GIS for Section 8 Housing Policy Analysis encompasses understanding the needs and concerns of the residents that can be addressed using GIS, responding to those needs, keeping data on what worked and what did not, and exercising judgment about what should be done next.

Policy research that can be addressed using GIS include questions about the concentration of people and assisted housing; and the concentration of HUD Section 8 tenant-based assistance program households and employment opportunities (Thompson and Sherwood, 1999). Many internal HUD policy analysts, as well as external HUD research partners, are interested in population and income distribution and in the need for housing and services at neighborhood to regional scales. Thompson and Sherwood (1999) developed a guidebook for GIS use in which many of the examples focus on data for Section 8 housing. The guidebook presents a method for using micro-data (individual households) and a corresponding method for addressing confidentiality concerns.

Confidentiality may be a concern when collecting and disseminating data on the spatial distribution of low-income housing. Methods should be devised

“Things-regionalism” refers to the most common local government attempts at regional initiatives that focus on a single function such as transportation, watersheds, sewers, and emergency management. The most extreme poverty in America is typically geographically concentrated, suggesting a need for “people-regionalism” to promote diversity, balance, and stability in every area of a region (Cisneros, 1996).

to avoid identification of individual households but still provide high-resolution data that addresses policy concerns. Methods involve deletions and geo-coding procedures; restricted access to data; rounding numbers into larger groups, aggregation units, and scale; and random displacement of certain point data. Determining causality is another difficulty. GIS data analysis, like other data analyses, cannot compute causality. It may take many maps to explore a topic before an analyst may gain insight into relationships among housing variables.

A robust data management approach is fundamental to having access to data for mapping purposes because multiple maps are often necessary to develop an “analysis scenario.” Figure 4.3 depicts HUD-assisted housing relative to employment concentrations mapped by traffic analysis zones (TAZs)—the units most often used to compile employment data for transportation purposes. Figure 4.4 shows the TAZ employment data aggregated to concentrations.

Although maps can be used to answer many questions, they can also prompt questions such as: Do people in Section 8 housing get higher paying jobs first, then move? Or, do people move and then get higher paying jobs? Is this pattern a result of household mobility? GIS can inform questions and guide next steps for research on the distribution of Section 8 housing allocation in relation to employment.

Dynamics of Neighborhood Change

The process of neighborhood change has been a subject of academic research in several disciplines for many decades, and yet significant gaps persist in our understanding of the dynamic processes that produce decline, revitalization, gentrification, and other urban processes. Since the early twentieth century, researchers in various disciplines have studied neighborhoods. Early examples include application of the ecological lens of invasion and succession to neighborhood studies (Park, 1926), analysis of economic and social factors that contribute to neighborhood decline and revitalization (Downs, 1981), and theories of tipping behavior 2 (Schelling, 1978). Sociologists have explored neighborhood patterns of racial segregation (Farley and Frey, 1994; Massey, 1990) and the emergence of concentrated poverty

A model explaining that the collective action of individuals may produce segregation even when the individuals prefer integration. Tipping behavior is the racial make-up of a neighborhood that prompts flight from the neighborhood (Schelling, 1971, 1972, 1978).

(Jargowsky, 1997; Wilson, 1987); geographers have studied spatial patterns of gentrification (Smith, 1996).

In spite of this sustained and broad spectrum of social science research on neighborhoods, a coherent synthesis and policy responsive to the dynamics of neighborhoods—especially poor neighborhoods—remains elusive. Among the pressing questions that remain are:

What internal and external factors, in varying combinations, influence the decline and revitalization of neighborhoods?
How do the mechanisms of decline and revitalization work, and how susceptible are they to intervention?
How are expectations of neighborhood residents and outsiders about the future of a neighborhood formed and changed?
How effective are different forms of intervention, under varying market conditions, at stemming or reversing neighborhood decline?
Is gentrification becoming more widespread; if so, under what conditions, and how can displacement of low-income residents be mitigated in gentrifying neighborhoods?

BOX 4.1 Race and Mortgage Lending

Race reporting is required from mortgage lenders as a result of the Home Mortgage Disclosure Act, 1 which monitors lending practices in minority communities, however, racial disclosure to lenders when borrowing by phone or Internet is not required. Subsequently, the second largest racial/ethnic group of those seeking mortgage credit in the United States is listed as “Not Reported.” A recent study set out to analyze the geographic expression and causes of the “Not Reported” racial/ethnic designation. To this end, a GIS-enabled spatial analysis of nondisclosure reporting in Atlanta, Georgia, was conducted using the following three econometric models demonstrating:

The differences in individual characteristics between disclosing and non-disclosing loan applicants;
The degree to which non-reporting of racial/ethnic identity results from institutional factors rather than deficiencies of loan applications; and

The Home Mortgage Disclosure Act (HMDA), enacted by Congress in 1975 and implemented by the Federal Reserve Board’s Regulation C, requires lending institutions to report public loan data.

The influence of neighborhood-level characteristics on nondisclosure rates after controlling for both individual and institutional measures.

The study found that the geography of nondisclosure coincided with the areas that lacked accurate data for lending research and had a high proportion of African-American neighborhoods. Based on these findings, the study concludes that there is a need for coordinating outreach efforts to publicize the importance of reporting race/ethnicity data to enforce civil rights.

SOURCE: Wyly and Holloway, 2002.

FIGURE 4.1 Share of home mortgage applications in Atlanta, Georgia without race-ethnicity information, 1999. Pattern confirms that nondisclosure rates are highest in predominantly African-American neighborhoods. SOURCE: Wyly and Holloway, 2002.