California is in the midst of a housing affordability crisis, with over half of renters statewide considered to be “costburdened” and nearly a third considered to be “severely cost-burdened” according to federal criteria.8 Additionally, there is a 1.5 million unit shortfall of rental housing units that are affordable to very low- and extremely low-income renters. State criteria to determine which projects receive California’s limited amount of federal tax credits, as enacted in the, must be analyzed to ensure that the state is incentivizing the siting of low-income rental housing in high-opportunity areas.9,10
The QAP matters because it is a crucial determinant of where developers can receive public financing for Nine Percent Tax Credit projects, affecting whether projects are sited in neighborhoods with higher or lower opportunity and whether projects are built in racially integrated or segregated areas. It is vital to site LIHTC projects in high-opportunity areas—where quality resources and services enhance opportunity for residents. Understanding the patterns of community-level resource distribution and service provision surrounding LIHTC-financed projects provides insights as to whether government housing subsidies are fostering upward mobility and furthering fair housing for lowincome residents.
Two existing state-by-state research studies provide initial suggestions as to how California’s QAP has affected project siting within the state, though it should be noted that prior to this report’s publication no study has assessed LIHTC projects in the Bay Area relative to other metro regions. Most recently, an analysis in 2015 from HUD revealed that California is making moderate strides at reducing the poverty exposure of LIHTC units. There was a 13.1% decrease in the number of neighborhoods with LIHTC projects which had poverty rates over 30% between the 2003-2005 period and the 2011-2013 period.11 However, a study in 2006 indicated that California was not successful at promoting LIHTC projects in racially integrated neighborhoods. More than 70% of LIHTC projects in California’s metropolitan areas were sited in census tracts where the percentage of non-whites was greater than the percentage living in the respective metropolitan area, suggesting that projects were disproportionately sited in segregated neighborhoods.12,13
This report presents a series of tables, charts, and maps that display the analyses we performed to assess LIHTC projects, units, awards, construction types, and housing types in terms of neighborhood opportunity and demographic composition.
To conduct the opportunity analysis, we used the ROI methodology to recalculate the Opportunity Index for each census tract in the Bay Area and divided these tracts into 5 quintiles based on the index value—higher values as higher opportunity and vice versa. The opportunity categories, ranking from highest to lowest opportunity scores, were labeled as Very High, High, Moderate, Low, and Very Low opportunity. The opportunity categories for each census tract are presented visually in Appendix Map 1. The following tables and charts present California LIHTC developments relative to opportunity, disaggregated by Four Percent and Nine Percent Tax Credits.
Analysis of LIHTC Developments by Opportunity
Our initial interpretation of the data on LIHTC projects, units, and total awards by opportunity shows that both Four Percent and Nine Percent Tax Credits are fairly evenly distributed across five categories of opportunity: Very High, High, Moderate, Low, and Very Low (Refer to Table 1 and Chart 1). There are more Four Percent Tax Credit projects than Nine Percent Tax Credit projects sited in the Bay Area (See Table 1 and Appendix Map 2). However, 64.9% of Four Percent Tax Credit projects are sited in the Moderate, Low, and Very Low opportunity areas. In addition, the opportunity category with the highest percentage of projects (24.9%) was the Low opportunity category, a trend that is seen throughout our Opportunity Analysis of LIHTC developments that received the Four Percent Tax Credits (See Chart and Appendix Charts 1b and 1c).
The performance of the Nine Percent Tax Credits stands in stark contrast to that of the Four Percent Tax Credits, due to a relatively high percentage of projects (25.7%), units (28.7%), and awards (25.9%) sited in Very High opportunity areas (See Chart 1; see also Charts 1b and 1c in the Appendix). For the Nine Percent Credit, the Very High opportunity category has the highest number of projects (25.7%); additionally, the allocation between Moderate and Low opportunity is comparable (20.9% and 22.8%, respectively). Thus, simply disaggregating projects by Four and Nine Percent Tax Credits shows that developments which received Nine Percent Tax Credits (which are awarded through the competitive application process according to QAP criteria) are far more likely to be sited in higher opportunity neighborhoods.14
Analysis of Acquisition, Rehabilitation, and New Construction by Opportunity
To understand the distribution of Acquisition, Rehabilitation, and New Construction developments, the data—which spanned the years 1987–2014—was divided into two categories: before 2007 and after 2007.15 This data is displayed visually in Appendix Map 3. Our data analysis revealed that before 2007, there were more Acquisition, Rehabilitation, and New Construction projects in the Four Percent Credit category located in lower opportunity areas. After 2007, however, more Acquisition, Rehabilitation, and New Construction projects were sited in higher opportunity areas.
For the Nine Percent Tax Credits, there were too many fluctuations and inconsistencies in the trends before and after 2007 to draw conclusions about how projects were sited relative to opportunity. The aggregated total of Four and Nine Percent before 2007, however, showed that about two-thirds of New Construction projects were sited in Moderate and Low opportunity categories. After 2007, the distribution of New Construction projects became fairly even except for the Moderate opportunity category, which had more projects than other opportunity categories (See Table 2).
Further assessments of the construction types before and after 2007 revealed there was a clear decrease in the percentage of projects that were sited in the lowest opportunity categories across program and project types (See Chart 2b). Prior to 2007, 71.3% of Four Percent Tax Credit projects of all types were sited in Moderate, Low, or Very Low opportunity areas. After 2007, this number dropped to 59.3%. As for the Nine Percent Credit, both before 2007 and after 2007, a plurality of Acquisition and Rehabilitation projects (42.8% and 44.9%, respectively) were in Very High opportunity areas; however, the proportion of projects in the lower four opportunity categories shifted dramatically over time. Whereas before 2007, nearly 43% of Acquisition and Rehabilitation projects were in the Very Low and Low opportunity categories, after 2007 this figure diminished to 24.1%, and a far greater percentage of projects were sited in High and Moderate areas. There was also an increase of New Construction projects in Moderate and High opportunity groups after 2007. However, the percentage of Nine Percent Credits New Construction projects in the Very High opportunity category declined significantly after 2007, falling from 22.7% to 15.3%.
Analysis of Four Percent and Nine Percent Tax Credits Timeline
The timelines in Charts 2c and 2d present data from 1987 - 2014 LIHTC developments and show the trajectory of both types of tax credits and the number of projects per year. For both the Four Percent and Nine Percent program, the number of Acquisition and Rehabilitation projects financed by tax credits has historically been lower than the number of New Construction projects. Chart 2c shows that the number of Four Percent Acquisition and Rehabilitation projects increased around 1996 and remained steady from 1998 - 2009. As seen in Chart 2d, there were very few Acquisition and Rehabilitation projects financed by the Nine Percent Tax Credit, though the number of these projects temporarily increased between 1998 and 2001, and again between 2011 and 2013.
In the early years of the LIHTC program, more New Construction projects were financed using the Nine Percent credit than the Four Percent credit, though this trend reversed in the early 2000’s, and in more recent years similar numbers of projects have been financed by the two programs. The trajectory of Nine Percent New Construction projects fluctuated throughout the time frame considered, as shown in Chart 2d. The numbers rose sharply after 1989, when there were zero projects constructed, but after 1995 New Construction projects dropped and never returned to the levels seen in the early 1990’s. As seen in Chart 2c, the number of Four Percent Tax Credit New Construction projects rose from 1995 to 2005 with a peak value in 2003, but the quantity of projects dipped and increased periodically.
In this section we demonstrate how Large Family LIHTC developments were sited relative to opportunity as compared to other types of developments, and we subsequently discuss project siting relative to the racial/ethnic composition of neighborhoods. Prior to 2016, Large Family projects were defined by having at least 25% of units with three or more bedrooms, and with at least 1,000 square feet of living room space. These projects are important to consider independently from other kinds of housing because they are more likely to house families with children, and the consensus in the academic literature is that child well-being, as well as their lifetime opportunities, are correlated with the neighborhoods in which they are raised.16 Our racial and ethnic analysis demonstrated that there are a disproportionate number of LIHTC projects, units, and awards in neighborhoods where the population of non-whites is greater than 40%.
In order to better understand the spatial pattern of Large Family housing developments, we analyzed Large Family by neighborhood opportunity and demographic composition. In Appendix Map 4 we have displayed the spatial distribution of Large Family projects and “Other” types of LIHTC projects. Table 3 shows that the Four Percent Tax Credit financed a total of 223 Large Family projects (39.19%), whereas it financed 346 projects (60.80%) ‘Other’ projects. The Nine Percent category financed 195 Large Family projects, while 178 projects are classified as ‘Other’. When both federal tax credit programs are aggregated, there are fewer Large Family housing projects than all other housing types combined (See Table 3).
We recognize that aggregating all housing types that are not classified as Large Family as ‘Other’ does not show the distinctions between different LIHTC project types —which include at-risk, special needs, non-targeted, and SingleRoom Occupancy (SRO)—and we may therefore be overlooking or neglecting a variety of other dynamics. Yet, the process of combining these other categories allows us to draw clear distinctions between housing types that are likely to have smaller unit sizes and standards from Large Family housing.
Analysis of Large Family Projects by Opportunity
Analyzing tax credit projects by housing type reveals that Large Family projects in both Four and Nine Percent Tax Credit categories are disproportionately concentrated in low-opportunity areas – 46.9% of the total of Large Family projects are in Low and Very Low opportunity neighborhoods, as opposed to 36.1% of all ‘Other’ types of projects (see Table 2a in the Appendix). Furthermore, disaggregating Large Family projects and units by type shows that over 46% of New Construction projects and total units are sited in Low and Very Low opportunity neighborhoods.
It is concerning that New Construction Large Family housing projects are built in areas that are not ideal for families because resources, services, and amenities are limited or inadequate, and may be in areas with poorly-performing neighborhood schools, high crime, or environmental hazards. We urge that priorities must be given to siting Large Family developments in high-opportunity neighborhoods to benefit families with children and foster well-being and upward mobility (see Table 8, and Charts 3b and 3c in the Appendix).
Analysis of LIHTC Developments and Neighborhood Demographics
This section presents California LIHTC projects, units, and awards relative to racial or ethnic composition of census tracts disaggregated by Four Percent and Nine Percent Tax Credits. It is crucial to analyze these residential patterns relative to opportunity to determine whether LIHTC developments are reinforcing or exacerbating patterns of segregation.17 For a visual illustration of LIHTC projects and neighborhood demographic composition, refer to Map 5 in the Appendix.
To conduct this demographic analysis, we pulled data from the American Community Survey (2010 - 2014, 5-year estimates) and divided the population of non-whites into 5 categories with equal intervals of 20% each. These categories range from less than 20%, 20.01–40%, 40.01–60%, 60.01–80%, and above 80%.18
In 2014, the Bay Area has a total population of 7,360,487 residents. Of these, 1,743,954 residents identify as Hispanic or Latino; 3,050,293 as non-Hispanic white; 1,758,791 as Asian; 455,865 as Black or African American; and 351,584 as Native American, Pacific Islander, or mixed-race.19 Thus, non-whites are 58.56% of the Bay Area population and 41.44% are non-Hispanic whites (see Tables 6 and 7 in the Appendix). With this degree of diversity, it is challenging to assess whether the Bay Area is racially integrated or segregated and whether LIHTC plays a role in reinforcing or exacerbating segregation.
However, as shown in table 4, there are a disproportionate number of projects, units, and awards located in neighborhoods with non-white populations above 60%. While about 50% of the Bay Area population lives in census tracts where non-whites constitute 60% or more of the tract population, almost two-thirds (61%) of LIHTC projects are in census tracts where 60% or more of the population was non-white. The number of projects disaggregated by Four Percent and Nine Percent Credits shows similar breakdowns of developments in these neighborhoods. Furthermore, the data in table 5 (see Appendix) indicates that more than 74% (representing $5.751 billion of investments) of Bay Area LIHTC projects, units, and awards are located in areas with a concentration of non-whites greater than 50%.
As seen in charts 4a, 4b, and 4c (in the Appendix), 61% of projects, 67% of units, and 66% of awards are located in areas where 60% or more of the population is non-white. Areas with high proportions of non-whites do not necessarily suggest that they are low-opportunity neighborhoods, but do raise fair housing concerns.
- 8. California Department of Housing and Community Development, 2016. “California’s Housing Future: Challenges and Opportunities.” Statewide Housing Assessment 2025 Public Draft.
- 9. The Qualified Allocation Plan is defined in Regulation Section 10302(ee). http://www.huduser.gov/portal/publications/pdf/QAP_incentive_mdrt.pdf
- 10. The QAP applies only to the Nine Percent Tax Credit, meaning that developers must compete with one another for Nine Percent Tax Credit awards by proposing housing developments which meet the greatest number of criteria in the QAP.
- 11. U.S. HUD Office of Policy Development and Research, 2015. “Effect of QAP Incentives on the Location of LIHTC Properties.”
- 12. In this study, California LIHTC neighborhoods were roughly as segregated as those in Texas, where the Supreme Court determined in a 2013 landmark case that the segregated siting pattern of LIHTC properties created a disparate racial impact, and was therefore a violation of the Fair Housing Act. See Brief of Housing Scholars as Amici Curiae, supporting Respondent. Texas Dept. of Housing and Community Affairs v. The Inclusive Communities Project, Inc., 135 S. Ct. 2507 (2015). Texas is also a useful comparison state because it receives the second highest number of LIHTC funds after California, and because at the time data was collected, metro areas in California and Texas had roughly the same percentage of non-white residents. See Exhibit 10 in Khadduri, Buron, & Climaco, 2006. “Are States Using the Low Income Housing Tax Credit to Enable Families with Children to Live in Low Poverty and Racially Integrated Neighborhoods?” http://www.prrac.org/pdf/LIHTC_report_2006.pdf
- 13. While the findings here on racial segregation in LIHTC neighborhoods can neither confirm nor deny the 2006 findings with respect to racial segregation because of divergent methodologies, it is important to conduct additional research statewide to understand whether the Bay Area is an exception to the trends found in 2006 study, or whether LIHTC neighborhoods have become less segregated over the last ten years in California.
- 14. See page 4 in California Tax Credit Allocation Committee http://www.treasurer.ca.gov/ctcac/program.pdf.
- 15. The reason for this choice was that legislation enacted in the wake of the economic crisis of 2008 had important implications for LIHTC, namely the Housing and Economic Recovery Act (HERA) of 2008 and the American Recovery and Reinvestment Act (ARRA) of 2009. See Novogradac 2012.
- 16. See, for example, Chetty, Hendren, Kline, and Saez. 2014. “Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States.” The Quarterly Journal of Economics 129:4.
- 17. Within the academic literature on the “neighborhood effects” of poverty, and within more recent research on lifetime opportunity, there is a consensus that segregated, impoverished, non-white neighborhoods are areas of concentrated disadvantage that reduce social mobility over time. See, for example, Chetty et. al. 2014.
- 18. We gathered data for the total population and subtracted non-Hispanic whites to obtain the non-white population, comprised of Hispanics, Blacks, Asians, Native Americans, and mixed-race groups.
- 19. The demographic analysis of Bay Area’s nine counties—Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma—is done at the US Census Burea census tract level using the 2010-2014 ACS 5-year-estimates. The racial/ethnic categories are also based on the US Census Bureau demographic classification. We addressed this categorization in the limitation section