Outdated: being updated
The Fair Housing Toolkit brings together available resources to help appointed and elected leaders, municipal planning, housing and redevelopment officials, developers, citizen board members, and other volunteers understand how to affirmatively further fair housing. A core component of fair housing law as it relates to policy and practice, disparate impact analysis can help shape a municipality’s master plan and zoning, be a key to assessing fair housing, and guide affirmative marketing efforts.
This section describes how to conduct a disparate impact analysis, introduces data sets useful for this purpose, and provides examples from practice and case law of the relevance to municipal planning.
What is disparate impact?
Even when a policy or practice is not intended to discriminate or doesn’t directly limit housing opportunity based on protected class, it may still have a discriminatory effect. Disparate Impact describes policies, practices or services that appear neutral on the surface, but, in practice, disadvantage protected class members.
Disparate impact can apply to a single rule or procedure such as the administration of a housing lottery or wait list. Disparate impact can also describe widespread effects of broad reaching policies and practice.
On Feb 15, 2013, HUD issued the Final Rule on the Implementation of the Fair Housing Act’s Discriminatory Effects Standard. With this rule, HUD formally codifies longstanding administrative and legal practice on how to screen for discriminatory effects. HUD’s final language provides that ‘‘[a] practice has a discriminatory effect where it actually or predictably results in a disparate impact on a group of persons or creates, increases, reinforces, or perpetuates segregated housing patterns…” on a protected class basis.
Municipalities have the obligation to analyze and modify rules, policies, and practices that have potential discriminatory effects. Before adopting any policy or practice, it is necessary to ask:
- Is it likely that policy or practice will negatively impact on members of a protected class compared to the general population?
- Is the policy or practice necessary to achieve substantial, legitimate, non-discriminatory interests?
- Is there a less discriminatory alternative that would meet the same interest?
 Federal Register, Vol. 78, No. 32. P 11482
On July 18, 2013 HUD issued an Affirmatively Furthering Fair Housing Proposed Rule which points to a fourth step in this analysis, particularly aimed at the obligation to revers segregated housing patterns:
- Is there a policy or practice that not only does not discriminate, but increases opportunities for traditionally under-served populations?
Disparate impact in the courts
Langlois v. Abington Housing Authority, 2002: The courts found the local residency preference policy instituted by a group of Massachusetts housing authorities to have an unlawful disparate impact because it would result in white applicants being served more quickly than applicants of color.
Anti-Discrimination Center of Metro New York v. Westchester County, 2009-2011: The County’s failure to consider race-based impediments to housing choice is at the core of this suit. One outcome is that public agencies and private firms need to confirm that they are following fair housing rules before certifying that they are in compliance. Disparate impact analysis is a key component of this policy analysis.
Click here to view other fair housing case laws.
Lessons from the Westchester lawsuit
HUD’s instructions in the Westchester settlement named the following restrictive zoning practices as likely to have a discriminatory effect and/or perpetuate segregation and thus warranting assessment:
- Restrictions that limit or prohibit multifamily housing development, including limitations on the size of a development and restrictions on lot size or other density requirements that encourage single-family housing or restrict multifamily housing. Such restrictions are widespread in Greater Boston due to a longstanding focus on single-family zoning, large lot size requirements, and provisions for multi-family and single-family housing on small lots only for residents who are 55 or older. Because of the high cost of land, limits on development size, lot size, or density have the effect of increasing costs. These policies can be shown to have a discriminatory effect because demographic data shows that African Americans, Latinos, and people with disabilities have lower incomes than the general population.
- Limitations directed at Section 8 or other affordable housing, including limitations on such developments in a municipality. Not only is source of income discrimination a violation of Massachusetts’ anti-discrimination law, but housing subsidy holders are more likely to be families with children and/or people with disabilities. Therefore limits on subsidized housing can exclude members of these protected class groups.
- Restrictions that directly or indirectly limit the number of bedrooms in a unit. Bedroom limits decrease housing choices for families with children. They can also limit choices for people with disabilities in need of in home care. Bedroom limits can disproportionately affect minority families who are more likely to have more children per household than white families and/or to have extended family living together.
- Limitations on townhouse development. Inherent in their design, multi-story townhouse units with internal and external stairs limit physical access for people with disabilities. Design preferences which favor townhouse development may run counter to housing needs for individual households and families with members who have a disability, housing needs for seniors and veterans. The impact of townhouse development must take into consideration the measured need for accessible housing for individuals and families, both in the rental and homeownership markets.
Data sets useful for conducting disparate impact analyses
Determining whether a proposed policy or practice will have a disparate impact begins with demographic data. In addition to census data, resources exist for primary demographic data as well as data sets analyzing demographic impacts.
A project of MAPC and the Boston Indicators Project, Metro Boston DataCommon is “an interactive data portal and online mapping tool that provides a wealth of information about the region’s people, communities and neighborhoods through a wide variety of topics — from arts and education to the environment and transportation. A resource for all those seeking to understand how the region is changing, it helps residents, stakeholders, planners, city and town officials, educators and journalists explore options and make informed decisions.”
The user friendly site provides data and analyses of municipal and regional trends. Their custom GIS tool allows users to create detailed maps of current and future trends. In addition to online tutorials about how to make the best use of the site, they offer monthly trainings and an online learning community to connect with other users. The project also provides an extensive list of other data visualization and mapping resources.
In proposed guidance on the obligation to Affirmatively Further Fair Housing released in July 2013, HUD announced the creation of a uniform data set to assist program participants in identifying housing needs and conducting disparate impact analyses. In the proposed rule, HUD describes the data set as “…including data related to education, poverty, transit access, employment, exposure to environmental health hazards, and other critical community assets, as well as nationally uniform local and regional data on patterns of integration and segregation; racial and ethnic concentrations of poverty; disproportionate housing needs based on protected class; and outstanding discrimination findings.” 
This data set is being designed for the expressed purpose of assisting with disparate impact analyses for municipalities to assess fair housing. Currently a prototype, HUD offers instructions for exploring the data set and expects it to evolve with user input and to make it available.
 Federal Register, Vol. 78, No. 139 P 43717
Data sets analyzing demographic impacts
Kirwan Institute/Opportunity Index Data The Kirwan Institute’s 2009 report, “The Geography of Opportunity: Building Opportunity in Massachusetts” has been widely referenced in discussions of regional equity. The Census Tract data analysis for their index is available for public use on the Kirwan Institute Massachusetts Opportunity Mapping Resource Site.
DiversityData.org Created by researchers at Brandeis University and the Harvard School of Public Health, this project analyzes racial and ethnic disparities for 362 metropolitan areas across the country. The data presented shows correlations between factors such as race and income, race and family size, disability and income in a metro region.
How to analyze for disparate impact
The key to conducting disparate impact analyses is the analysis must compare two groups within a protected class category to determine whether the policy or practice disadvantages one as compared to the other. Once data have been broken out by protected class, the next step is to compare effects of proposed policies on each subgroup. For instance racial groups could be compared to each other or families with children could be compared to adult households, or people with disabilities could be compared to people without disabilities.
Based on these numbers, the next determination is whether the policy has the effect of
- Creating more housing opportunities for protected class members
- Maintaining patterns of segregation
- Excluding groups of people, whether or not that was the intention behind the policy
In order to assess the impact of a proposed policy, the regional demographics by protected class should be reviewed. Next, the impact of the policy should be examined for each group separately to see whether one group benefits or is adversely affected in comparison to the other. If the analysis shows there is a discriminatory effect, it is then necessary to search for non-discriminatory alternatives to reach the stated goal. Beyond non-discrimination, the mandate to affirmatively further fair housing seeks a policy or practice which increases opportunity to overcome patterns of segregation. Specific remedies will differ depending on the need and the arena, as detailed in the examples below.
Various statistical methods may be employed to conduct these analyses. CHAPA’s Guide to Selection Preferences describes such analyses and gives case examples to illustrate their use along with specific steps for analyzing tenant selection policies using a civil rights lens. Part IV of this Guide uses disparate impact analysis tools to examine local selection preferences, guiding the reader through the following
Checklist for Civil Rights and Fair Housing Compliance:
- An examination of the actual waiting list or anticipated waiting list, and an examination of the housing market area indicates the following racial, ethnic or other differences between the waiting list and the housing market
- An examination of the occupancy patterns or anticipated occupancy patterns, and an examination of the housing market area indicates the following racial, ethnic or other differences between the occupancy patterns and the housing market
- The actual or anticipated rate of participation on the waiting list by minorities compared to the participation rate of non-minorities is at least 80% (the “Four-Fifths Rule”):
- The actual or anticipated rate of selection for admission by minorities compared to the rate of selection for admission by non-minorities is at least 80% (the “Four-Fifths Rule”):
- Is there a delay in minority selection for occupancy compared to non-minority waiting time? Specify.
- The selection preference, or other aspects of the housing affirmatively further fair housing in the following ways:
- Other methods for meeting the identified housing need that have less of an effect on minority participation include
The CHAPA Selection Preferences Guide also presents an Appendix with a Case Study designed to illustrate the principles discussed. The Case Study uses actual demographic data and a hypothetical housing development, drawing on various statistical analysis methods to evaluate the impact of proposed plans, finally offering alternative policies which would remove the discriminatory effect.
Example: examining race and affordability using special mismatch or free market analyses
When looking at racial demographics in particular, special mismatch or free market analyses can be particularly informative. Using income data by race, it is possible to predict what the racial and ethnic make-up of a municipality should be if affordability was the only determining factor, and then compare this theoretical population to the actual demographics existing today.
The Fair Housing Center of Greater Boston (FHCGB) and the Harvard Civil Rights Project completed this analysis for homeowners in Greater Boston in 2004. They found that in 80% of Greater Boston cities and towns, African American and Latino homeownership rates are less than half of what would be predicted by affordability alone. The City of Naperville, Illinois’ Analysis of Impediments for Fair Housing Choice uses this method to help determine housing need and Affirmatively Furthering Fair Housing policies.
Example: local resident preferences
Following the Court decision in the Langlois v. Abington case described above, Massachusetts Department of Housing and Community Development (DHCD) requires municipalities to analyze the impact of a local resident preference for housing wait lists. If the analysis shows the local preference will move whites into housing more quickly than other racial groups, DHCD requires the town to balance the wait list so that all racial groups are served in an equally timely manner. This process is detailed in DHCD’s Affirmative Fair Marketing Guidance.
Example: age-restricted zoning
Housing admissions policies which do not allowing residents under the age of 55 have an implicit and inequitable impact on families with children. Additionally, MAPC has found in MetroFuture that Boston area households of color are more likely to have children when compared to white households – 48% and 32% respectively. In order to propose such a policy, it is necessary to analyze the effects on family versus adult households and the racial and ethnic demographics of seniors versus younger residents in the region. If the analysis shows families with children or people of color would be excluded by the policy, a less discriminatory alternative must be sought.
 Citizens’ Housing and Planning Association. Meeting Local Housing Needs: A Practice Guide for Implementing Selection Preferences and Civil Rights Requirements in Affordable Housing Programs. September, 2004. P 8