LIHTC Revitalization

LIHTC Revitalization

Racial Segregation of Dallas Low Income Housing Tax Credit units and lack of Community Revitalization

Large infusions of Tax Credit Housing alone does not correlate with revitalizing neighborhoods. The following is an analysis of Hatcher Square - a proposed development in the South Dallas Frazier Neighborhood that explores the effects of high concentrations of Tax Credit housing and the conditions it produces. 

Hatcher Square Analysis

DBPC submitted a comment to the IRS in response to the IRS request for comments on the content of concerted community revitalization plans. See IRS Notice 2016-77.

 


Re: The defense of racial segregation in Who Wants Affordable Housing in Their Backyard should be subjected to additional scrutiny given the questions about the underlying data and several conclusions in the study.

July 20, 2017

Re: The defense of racial segregation in Who Wants Affordable Housing in Their Backyard should be subjected to additional scrutiny given the questions about the underlying data and several conclusions in the study.

The Rebecca Diamond and Tim McQuade study, Who Wants Affordable Housing in Their Backyard: An Equilibrium Analysis of Low Income Property Development,[1] (Who) is interpreted by some to confirm the basic working principle of racial segregation that housing for low income minority renters should be kept in low income minority areas in order to maximize social utility. The study is reported as a credible claim that LIHTC projects increases property values and social utility when placed in minority areas which also avoids decreasing property values in White non-Hispanic (White) areas by keeping it out of those areas and thus accommodating White aversion to minority residents.

The newspaper references include:

Washington Post in “A surprising way to increase property values: build affordable housing” by Tracy Jan

New York Times in “Program to Spur Low-Income Housing Is Keeping Cities Segregated” By JOHN ELIGON, YAMICHE ALCINDOR and AGUSTIN ARMENDARIZJULY 2, 2017

“But a study published last year by two Stanford professors made a case for building tax-credit housing in high-poverty areas, finding that home values around the developments rose by about 6.5 percent and that segregation decreased modestly.” 

The Diamond and McQuade study gives mention to the proven injuries to children from living in high poverty, racially segregated areas. The study claims these injuries are offset in a cost benefit analysis by the increase in area property values. Who pages 1, 2, 32.

Based on the study’s claims, the principal winner in the difference between placing LIHTC properties in poor white areas and placing LIHTC properties in poor minority areas is, of course, the absentee landlords. According to the study, the benefit to absentee landlords from a LIHTC project in a poor white area is $29,048,000. The benefit to absentee landlords from an LIHTC project in a poor minority area $74,236,000. Who, page 61, Table 5. The increase in value to absentee landlords from using the LIHTC project to perpetuate racial segregation is 236%.

The reports on the study do not include any common sense observations on the study’s conclusions and the underlying data. There are several such observations that counsel caution in applying the conclusions to justify racial segregation in the LIHTC program or in any other affordable, low income housing program.

The first common sense observation is that the study’s conclusion that building LIHTC projects in high minority areas “will then spark in-migration of high income and a more racially diverse set of residents” is directly contradicted by the study’s conclusion on the same page that White’s will pay more to live further away from tax credit housing in minority areas and have a lower willingness than minorities to pay for proximity in lower income, high minority areas.

Such results suggest that white households may have a preference for neighborhood homogeneity which interacts with how they view the amenities/disamenities provided by LIHTC construction. Who, page 33.

The comment and the reporting on the study focuses on the benefits to putting LIHTC projects in low income minority areas. The study actually reports a substantially higher average benefit per home owner in low income, non-minority areas. The Average Benefit per Home Owner in low income, Q1, Low Minority areas is $23,403. The Average Benefit per Home Owner in low income, Q1, High Minority areas is substantially less, $16,857. Who, page 60,Table 4. This would seem to support placing LIHTC units in poor White areas rather than in poor minority areas as the study suggests.

The data supporting the economic utility of racial segregation espoused in the study is from only a few states and at least one half of the data is from one state, California. The data does not include any LIHTC units in 35 of the 50 states or the District of Columbia. The data was not chosen by random sampling but by whether home sales data is made available in public records. This selection criteria omits data from most of the Southern states except for Tennessee and Florida. Who, pages 5, 6, Figure 1. No data from Texas is included even though approximately 10% of the LIHTC units are in the state. More than 50% of the LIHTC sites in the study are located in California. Who, page 38. Only 8.2% of the LIHTC projects developed from a substantial part of this period, 1995 through 2006, were in California. HUD, Updating the Low Income Housing Tax Credit (LIHTC) Database: Projects Placed in Service Through 2006, 2009, pages 83 - 84.

The changes in property values that are said to support racial segregation are extremely limited in geographic scope. The study reports increases in minority areas and decreases in White non-Hispanic areas only for home sales within 528 feet or 0.1 of a mile from the LIHTC project. For example:

We find that the average household that desires living near LIHTC sites in low income areas is willing to pay approximately 6% of their house price to live 0.1 miles from a LIHTC site. In higher income areas with low minority populations, on the other hand, the average household who chooses to live near LIHTC is willing to pay approximately 1.6% of their total house price to avoid living within 0.1 miles of a LIHTC site.5 Correlating these preference estimates with information on home buyer demographics, we further find that higher income households are more willing to pay for proximity to or distance from the LIHTC site, consistent with results of Diamond (2016). Who, page 3. (Emphasis added).

These narrow results are based on sales of only small total numbers of owner occupied units in the 0.1 mile, 528 feet range. The data shows an average of 3.76 homeowners impacted by an LIHTC Development in Low Minority Areas with Income Q1 in each neighborhood observed. The data shows an average of 1.67 homeowners in High Minority Areas with Income Q1 in each neighborhood observed. This is not counter intuitive. There would not usually be many homeowners in such a small area. According to the data, homeowners are a minority of the households impacted by a LIHTC project. Who, Table A3: Median # of Households Impacted by LIHTC Development.

Perhaps it is this very limited data set of homes with 528 feet or 0.1 mile of a LIHTC project that results in the very high income of the home buyers and very low percentage of Black buyers in even in the very low income areas. Table A1 reports:

• for block groups with median income < $26,017, 14.59% of the home buyers were Black and had income of $92,547 and

• for block groups with median income < $38,177, 9.07% of the home buyers were Black and had income of $82,374. Who, page 74.

If the White aversion to LIHTC projects is strong as reported in the study, and the home buyers are not likely to be Black, then some other race or Hispanics are buying the units. This may be a form of integration but not a reduction in Black/White or Hispanic/White racial segregation.

The study’s definition of high minority area is ambiguous and may explain some of the results. A census block group is classified as high minority if it “has a population at least 50% Black or Hispanic based on the 1990 census.” Who, page 15. If this definition requires an area to be at least 50% Black or at least 50% Hispanic, then it may exclude from the definition of high minority areas those block groups that are substantially more than 50% minority but are not either 50% or more Black or 50% or more Hispanic. A block group that was 40% Black and 40% Hispanic and 20% White non-Hispanic would not be a high minority area if the “or” is disjunctive. Common use would describe a 20% White non-Hispanic block group as a high minority population block group. The study does not give examples to clarify the definition and the authors have not made the underlying data available for review.

The data uses only 1990 U.S. Census data to analyze the universe of LIHTC units placed in service from 1987 through 2012. Who, page 6.This introduces a bias to the analysis. For example, the data concludes that LIHTC projects cause a decline in predominantly White non-Hispanic areas. But areas that were majority or significantly White non-Hispanic in 1990 have frequently become substantially less White non-Hispanic in the passing years. In Dallas, the use of 1990 census data would place numerous LIHTC projects and units in White non-Hispanic areas even though those locations are and have been predominantly minority since 2000 or 2010 or 2015. The use of 1990 census data for the existing LIHTC projects and units in the City of Dallas would show 19 LIHTC projects with 3,719 LIHTC units in 50% or greater White census tracts. This is 13.6% of all LIHTC projects and 15.2% of all LIHTC units in the City of Dallas. The reality is significantly different using 2010 U.S. Census data. Only 6 LIHTC projects, 4.3% of the total projects, with 999 LIHTC units, 4.1% of the total units, are in majority White non-Hispanic census tracts using 2010 U.S. Census data. If there are property value decreases in the tracts that are no longer White, the study would still attribute those decreases to LIHTC projects in White areas. This would overstate the decrease in White areas and understate the decreases in minority areas.

The use of 1990 census is not consistent in the study. The number of renter and owner households in the study neighborhoods are not calculated with 1990 census data even though the income and racial characteristics are from 1990 census data. The renter households and owner households are calculated using 2007 to 2012 American Community Survey data. Who, page 76, Table A3. This combination of data sources is accurate only if there are only small differences between 1990 census data and 2012 census data.

The study’s conclusion on the reduction of crime in LIHTC neighborhoods is more limited than the data used for the property value estimates. The crime data conclusions are based on several years of data in three cities: Chicago, San Francisco, and San Diego. There are only 127 LIHTC projects in this data, 0.3% of the nation’s LIHTC projects. Who, pages 18 - 19. There are 39,094 LIHTC projects in the overall LIHTC inventory. The locations were not chosen at random but by whether there was data available. Who, page 18.

Other than the cursory look at crime in three cities, the only neighborhood effect examined for the entire data base is the Chetty future income measure. By this measure, the benefits to the tenants in a high income, Q4, low minority area would be $26.7 million. The study admits that this underestimates the total benefits to tenants which have been documented in other studies on neighborhood effects. Who, page 32. But, according to the authors, the continuation of racial segregation is more beneficial because the $26.7 million integration benefit to the families is outweighed by the $116 million benefit to everyone from building in a low minority, low income area. Who, page 32. Of course the majority of the benefits in those areas accrue to the absentee landlords and the homeowners according to the study. Who, pages 61, Table 5. The study’s defense of racial segregation should be put in the context of all the costs of that segregation including all of the injuries to children such as high crime, challenged schools, adverse industrial and other environmental uses, and the absence of adequate neighborhood services. The study’s claims that the injuries to children from segregation are justified by the increases in property values to absentee landlords are meant to counter the neighborhood effects from racial segregation research by Chetty and others.

The Who study skips over most of the country in the data used but then makes claims that racial segregation is appropriate everywhere. The basis for the claim should receive more scholarly attention.  One way to examine the claims is to investigate whether the claimed increases in values and benefits have been achieved in LIHTC neighborhoods not included in the study.

[1] Rebecca Diamond and Tim McQuade, Who Wants Affordable Housing in Their Backyard: An Equilibrium Analysis of Low Income Property Development, November 2016, available at https://web.stanford.edu/~diamondr/LIHTC_spillovers.pdf;.