Discrimination in Metropolitan Housing Markets: National Results from Phase 1 of the Housing Discrimination Study (HDS)
The Housing Discrimination Study 2000 (HDS 2000) represents the most ambitious effort to date to measure the extent of housing discrimination in the United States against persons because of their race or color. It is the third nationwide effort sponsored by HUD to measure the amount of discrimination faced by minority home seekers. The previous studies were conducted in 1977 and 1989.
This report provides national estimates of discrimination faced by African Americans and Hispanics in 2000 as they searched for housing in the sales and rental markets. It also provides an accurate measure of how housing discrimination has changed since 1989.
The results in this report are based on 4,600 paired tests in 23 metropolitan areas nationwide. The report shows large decreases between 1989 and 2000 in the level of discrimination experienced by Hispanics and African Americans seeking to a buy a home. There has also been a modest decrease in discrimination toward African Americans seeking to rent a unit. This downward trend, however, has not been seen for Hispanic renters. Hispanic renters now are more likely to experience discrimination in their housing search than do African American renters.
While generally down since 1989, housing discrimination still exists at unacceptable levels. The greatest share of discrimination for Hispanic and African American home seekers can still be attributed to being told units are unavailable when they are available to non-Hispanic whites and being shown and told about less units than a comparable non-minority. Although discrimination is down on most measures for African American and Hispanic homebuyers, there are worrisome upward trends of discrimination in the areas of geographic steering for African Americans and, relative to non-Hispanic whites, the amount of help agents provide to Hispanics with obtaining financing. On the rental side, Hispanics are more likely in 2000 than in 1989 to be quoted a higher rent than their white counterpart for the same unit.
For convenience, the report is available for download in PDF format. Your options for downloading are as follows:
- Executive Summary Only (.pdf 128KB)
- The Full Phase 1 Report without Annexes (.pdf 2.79MB)
The Executive Summary provides a general overview of the major findings and the methods used to develop the estimates.
Quick Link to MSA Summaries. This page provides quick summaries by MSA of year 2000 estimates of consistent adverse treatment and the primary types of that adverse treatment. Click on your individual MSA to see the summary results. If you want to see how patterns of adverse treatment for your MSA compares to the national level or other MSAs, click on the BW rental (adverse treatment toward African Americans seeking to rent a unit), BW sales (adverse treatment toward African American homebuyers), HW rental (adverse treatment toward Hispanics seeking to rent a unit), or HW sales (adverse treatment toward Hispanic homebuyers).
Note that for the most part, the metro-level results show much fewer items as showing statistically significant discrimination than the national estimate. It is not because discrimination is necessarily different or less in the metropolitan areas than nationally, it simply reflects that the number of tests conducted in each metro area was small (relative to the total national sample), the lower-bound (net) estimates of discrimination are often not statistically significant. Generally, we conducted about 70 tests per tenure and per ethnic group in each metro area, a very challenging volume of testing for local organizations conducting the tests. However, at this sample size we would need to see net measures of about 10 percent or higher to be sure they were statistically significant. In general, because of the wide confidence intervals, we report the overall incidence of consistent white-favored treatment was comparable across most metro areas. The national estimates have much larger sample sizes (between 700 and 1200 tests for each tenure and ethnic group), allowing us to measure discrimination with much greater precision than we do at the metropolitan level.
This is the full report, including the Executive Summary.
- Chapter 1 provides a description of the paired testing methodology generally along with an overview of the scope of the study.
- Chapter 2 presents the methodology implemented in Phase I of HDS2000, including the sample of metropolitan areas in which tests were conducted, the procedures used to draw a sample of available housing units in each of these metropolitan areas, the paired testing protocols implemented for both rental and sales housing, and the statistical procedures used to estimate the incidence of adverse treatment.
- Chapter 3 presents current national estimates of adverse treatment against African American and Hispanic renters and homebuyers, as well as estimates of change in differential treatment since 1989.
- Chapter 4 presents metropolitan-level estimates of adverse treatment against African Americans and Hispanics compared to the national level for each of the twenty large metropolitan areas in our sample, highlighting metropolitan areas with significantly higher or lower rates of adverse treatment. Chapter 4 also presents results from exploratory testing for adverse treatment against Asians and Native Americans.
- Chapter 5 uses multivariate analysis methods to test hypotheses about potential sources of random and systematic differences in treatment, and addresses some of the major methodological criticisms that have been leveled at paired testing research.
- Chapter 6 presents expanded measures of geographic steering in the sales market.
- Chapter 7 explores systematic variations in the incidence of adverse treatment, and assesses the extent to which they support hypotheses about the causes of discrimination.
- Chapter 8 reviews all the findings from Phase I of HDS2000 and discusses their implications, both for future paired testing research and for ongoing enforcement efforts.
Annex 1 of the HDS 2000 report includes the Test Assignment Guides, Forms and Instructions used by the organizations that conducted the paired tests for HDS 2000.
Annex 2 of the HDS 2000 report includes the Test Report Forms used for the Rental and Sales paired tests. After meeting with an agent, each tester independently completed these forms to record the treatment that they received. The data from these forms were then used by the analysts to construct the treatment variables.
- Annex 3: This annex describes the tests for statistical significance used in the report.
- Annex 4: This annex describes the methodology for and presents preliminary results from triad tests that were conducted in two metropolitan areas in Phase II of HDS2000. Triad tests involve visits by three testers to inquire about each randomly selected advertisement. Two of the visits in each test involve testers of the same race. A comparison of the experiences of the two same-race testers provides a direct measure of random differences in treatment during the testing process.
- Annex 5: This annex describes how and why the treatment measures reported in 2000 from the 1989 HDS paired test data are different from the treatment measures reported as part of the 1989 study.
- Annex 6: For comparability, Phase I of HDS2000 implemented the same weekly ad-sampling methodology that was used in 1989. However, the 1989 HDS found that some geographic areas within many metropolitan housing markets were under-represented in the major metropolitan newspaper. In order to learn more about this issue, two additional samples of available housing units were selected for a subset of sites in HDS2000. First, additional advertisements for units in under-represented communities were drawn from the major metro newspaper. And second, additional units available for sale or rent were identified from other sources for the most under-represented communities. In this annex, we stratify tests based on whether the advertised unit was located in a well-represented community or an under-represented community to determine whether patterns of treatment vary.
- Annex 7: This annex includes tables which show how each individual treatment variable contributes to the two overall composite indexes, the consistency composite and the hierarchical composite, for both 1989 and 2000.
Annex 8 includes estimates on every treatment variable for each Metropolitan area that was included as part of the 2000 sample. Statistical significance of the net measure is also shown.
- Annex 9 provides the methodology for the multinomial logit estimations and simulations that were reported in Chapter 5 of the main report.
- Annex 10 provides the methodology for the fixed effect logit estimation reported in Chapter 7 of the main report.
- Annex 11 provides the detailed results from the fixed effect logit estimation reported in Chapter 7 of the main report.