IntroductionCase Study: the Greater Pittsburgh RegionEvaluation Framework for AIsReview of AIs[Table 2. Regional Distribution of AIs in the Sampling PopulConclusion and Next Steps[Table 2: Regional Distribution of AIs in the Sampling PopulEvaluating Local Studies of Barriers to Fair Housing Michael P. Johnson, Ph.D.* H. John Heinz III School of Public Policy and Management Carnegie Mellon University Pittsburgh, PA 15213-3890 [email protected] 412-268-4270 412-268-7036 (fax) Angela M. Williams Foster, Ph.D. Graduate School of Public and International Affairs University of Pittsburgh Pittsburgh, PA 15260 [email protected] C. Augustus Martin, J.D., Ph.D. School of Business and Public Administration California State University, Dominguez Hills Carson, CA 90747 [email protected] Last revised: July 27, 2003 *corresponding author. Previously presented in different form at: Association for Public Policy Analysis and Management 2002 Annual Research Conference, Dallas, Texas, November 9, 2002.Evaluating Local Studies of Barriers to Fair Housing Abstract: Analyses of Impediments to Fair Housing (AIs) are Federally-mandated studies intended to develop an agenda for remedies to structural and operational barriers to fair housing in the U.S. There are few existing evaluations of AI quality, and there is little understanding of how AI results influence fair housing policy. We have developed an evaluation instrument that determines the extent to which AIs meet statutory requirements and use appropriate data and methods. We have applied this instrument to a small national sample of AIs using multiple readers. We find that while there is broad agreement across readers as to whether specific questions are addressed or not, there is considerable variation regarding more detailed measurements of AI quality. Nevertheless, the evaluation instrument appears to provide reliable measures of AI quality across topic areas. This research will enable jurisdictions that perform AIs to develop well-defined quality metrics and to address unique local characteristics. Keywords: Fair housing, analysis of impediments, housing policy I. Introduction Analyses of Impediments to Fair Housing (AIs) are studies that must be performed by American jurisdictions that receive Community Development Block Grant funds from the U.S. Department of Housing and Urban Development (HUD). AIs are intended to develop an agenda 1for remedies to structural and operational barriers to fair housing that leverages local knowledge of housing markets and policy implementation and avoids an approach relying on centralized authority and policy mandates. The legal foundation for these studies is Title VIII of the Civil Rights Act, also known as the “Fair Housing Act” (42 U.S.C sections 3601-3619), which states, in part: “[i]t is the policy of the United States to provide, within constitutional limitations, for fair housing throughout the United States.” In addition, fair housing laws have been enacted by many states and local jurisdictions. Fair housing laws apply civil rights protections to specific members of the population: those members of defined “protected classes”. Seven Federal protected classes are: race, color, religion, national origin, sex, familial status and disability. These definitions can be extended through state and local laws. For example, the Pennsylvania Human Relations Act defines three additional protected classes to those listed above: ancestry, age, and use of guide- or support-animal because of blindness, deafness, or physical handicaps; the Pittsburgh, Pennsylvania City Code defines five additional protected classes: ancestry, place of birth, sexual orientation, age (declared policy but not actually in legal code), and use of support animals because of the handicap or disability of the user. While AIs have been performed by many jurisdictions over the past decade, few evaluations of the quality of AIs have been performed, and there is little understanding of how, or whether, AI results influence fair housing policy. In the experience of these authors, who have performed AIs for the City of Pittsburgh (Martin and Johnson, 1999) and suburban Allegheny County, PA (Martin, Johnson and Williams Foster, 2000), AIs are often viewed as a government requirement separate from policy design and law enforcement related to fair housing. 2A research and policy gap exists on the subject of evaluating AIs. Although HUD and private fair housing organizations such as the National Fair Housing Alliance have historically promoted standardization of AIs as a way to improve the quality of reporting, there is little available data to assess whether, in fact, reporting jurisdictions heed these recommendations. Proposals for AI standardization are uncomplicated. For example, in its Fair Housing Planning Guide (U.S. Department of Housing and Urban Development, 1996), HUD recommends a “Suggested AI Format” to jurisdictions receiving CDBG funds. HUD’s Suggested AI Format encourages jurisdictions to report the following data: • Demographic Data • Income Data • Employment • Housing Profile • Maps HUD also recommends a standardized “Suggested Format for the Analysis of Impediments,” composed as follows: • Introduction and executive summary • Jurisdictional background data • Evaluation of current fair housing status • Impediments to fair housing choice • Public and private fair housing programs and activities • Conclusions and recommendations Details of the AI Format are contained in the Appendix. 3HUD’s proposed model for standardization presents useful guidelines for constructing AIs. It also suggests a framework for evaluating the quality of AIs and their use in developing fair housing policy. More explicit and extensive guidance on evaluation of AIs is given in a web page provided the National Low Income Housing Coalition (2001). However, neither of these two resources provides a concise, easy-to-apply methodology by which local organizations can determine the extent to which their studies confirm to widely-accepted standards. Our experience in performing AIs has led us to three fundamental insights. First, we have realized that there is little guidance available to practitioners or researchers who seek to measure the quality of a particular AI, or a group of AIs on a regional, or national basis (an analysis of impediments “report