COVID-19 will have cascading but unknown impacts on housing, as an unprecedented wave of layoffs decimate household balance sheets, particularly among low-wage workers in the service sector who tend to be the most cost-burdened of rental households.1 The virus's effects exacerbate an existing affordable housing shortage across the Federal Reserve Bank of Atlanta's District, which lacked 1,001,605 affordable rental units before the virus outbreak began.2 Some state and local policymakers have responded to the inevitable impact on renter incomes—and therefore, ability to pay rent—by halting evictions, resulting in a patchwork of policies across and within states.

On the federal level, Congress enacted the Coronavirus Aid, Relief, and Economic Security (CARES) Act, which included a 120-day eviction moratorium that will be in effect until July 25, 2020.3 In doing so, Congress leveraged its federal reach to reduce evictions, even in states where eviction moratoriums are not in place. This move brought relief and stability to many households that have lost income and cannot pay rent. However, the CARES Act moratorium only applies to properties that have an existing relationship with the federal government, and therefore does not cover all rental properties nationwide. It is difficult to know exactly how many or which properties are covered in any given location.

Exploring the CARES Act moratorium
We estimate that the CARES Act moratorium covers between 28.1 percent and 45.6 percent of occupied rental units nationally. Our estimate's wide range and lack of a state- or regional-level breakdown results from the unavailability and often fragmented nature of data on the nation's rental stock and housing finance. Without these data, federal, state, and local actors are left unable to understand the impact the federal eviction moratorium may have within their jurisdictions, stunting their ability to predict and respond to housing needs. Given better data, policymakers, advocates, and service providers could craft evidence-based solutions to stabilize rental housing both in the midst of the crisis and as the economy strengthens.

In this article, we explore the categories of covered properties under the federal CARES Act's eviction moratorium and estimate the extent of its coverage. In doing so, we reveal two chief data challenges that, if addressed, would lead to a better understanding of not only the CARES Act moratorium, but also of the broader availability and stability of rental housing. First, much of the data needed is already collected by public and quasi-public entities, but the data are not currently available for public use. With access to these data in a versatile and useful form, policymakers, researchers, and advocates will be able to make better evidence-based recommendations and decisions regarding the housing impacts of COVID-19. Second, some information related to rental housing, its ownership, and its underlying financing is not being collected in a centralized and standardized manner or, in some cases, at all.4 We encourage state and federal actors to develop methods and mechanisms that will generate more robust data on the presence and financing of rental housing stock, such that threats to housing infrastructure can be better assessed and responses more finely crafted.

Which rental units are covered under the CARES Act?
Since the CARES Act does not uniformly prevent evictions in all rental properties, in order to know whether a tenant is protected by the act, one must determine whether the unit has characteristics that fall within the parameters of the act's coverage. The CARES Act's definition of “covered dwellings” includes occupied dwellings that have been subsidized, financed, insured, guaranteed, or otherwise supported by the federal government.5 Generally, “covered properties” fall into one of two categories: federally subsidized affordable housing, or a property with a federally backed mortgage. Chart 1 depicts these two categories and the overlapping relationship between them.

The CARES Act's federally subsidized affordable housing category includes properties that are federally owned, receive federal funding, or were constructed or purchased using federal financing or active incentive programs designed to create affordable housing. The act draws its list of covered federal programs from the Violence against Women Act (34 U.S.C. 12491(a)(3)) and adds the rural housing voucher program (42 U.S.C. 1490r) to that list. Chart 1 represents federally subsidized rental housing in blue and breaks it down into boxes representing programs included in this category.

The second category of “covered properties” includes any single-family or multifamily property encumbered by a federally backed mortgage loan. This means that during a purchase, refinance, or renovation of a property, the owner secured a mortgage that has been insured, guaranteed, owned, supplemented, or otherwise assisted by a federal officer or agency. Properties that fall into this category may have loans insured by the Veterans Administration (VA), the Federal Housing Administration (FHA), or the U.S. Department of Agriculture (USDA).6 This category of properties may also have loans originated as conventional mortgages that have since been purchased or securitized by the Federal Home Loan Mortgage Corporation (Freddie Mac) or the Federal National Mortgage Association (Fannie Mae). An unknown number of properties with a federally backed mortgage may also fall into the federally subsidized affordable housing category, shown as an overlap in green and blue boxes in chart 1. Chart 1 demonstrates that overlap exists between project-based subsidized units and tenant-based subsidized housing choice vouchers as well as properties with a federally backed mortgage.

CARES Act moratorium coverage: numbers and data access
Public officials at all levels of government, service providers, charitable organizations, researchers, and advocates want to make fast, data-informed decisions as they tackle the fallout from COVID-19. Service providers may need to know whether a given household is protected by the eviction moratorium, and those studying or planning the larger impact on renters may need to know what share of the overall rental market is covered. Knowing the extent of the CARES Act moratorium's impact on their area would help these stakeholders plan for overflow shelter space, understand the likelihood of increased eviction numbers, and develop programs to support affected parties such as tenants and landlords. Tenants may not know or be able to access information about whether their unit is protected by the moratorium, and thus would be unlikely to benefit from it should they face the threat of eviction.7 Unfortunately, much of the data needed to address these concerns thoroughly is incomplete or unavailable. However, we are able to use what data are available to estimate a range of CARES Act eviction moratorium coverage.

The National Housing Preservation Database (NHPD) has combined much of the property-level data sets on federally subsidized affordable housing (the category represented in blue in chart 1).8 Combining this with other subsidy program data, we estimate that nationally 11.6 percent of rental units are federally subsidized affordable housing, covered by the CARES Act's eviction moratorium. These estimates range from 11.1 percent to 19.0 percent in individual southeastern states, as seen in chart 2.9

However, this estimate likely undercounts the number of units in the federally subsidized category because it does not include housing choice vouchers (HCV)—a tenant-based housing subsidy that provides 2.6 million households with affordable housing nationwide.10 The U.S. Department of Housing and Urban Development provides HCV allocations by subarea, but does not release cross tabulations by program that can account for overlap between programs. This estimate also does not include the following rural housing programs: Section 516 Farm Labor Housing Grants, Section 542 Rural Development Vouchers, or Section 533 Rural Housing Preservation grants.11

Nationally, HCVs comprise approximately 5.8 percent of total rental units.12 However, some units contained in our 11.6 percent national estimate of federally subsidized rental units likely also benefit from an HCV.13 Therefore, although these programs both account for portions of the overall American Community Survey count of rental units nationwide, we cannot simply add 2,556,270 HCV units (5.8 percent) to our 5,103,430 federally subsidized affordable units (11.6 percent) estimate of nationally subsidized units. Doing so would very likely result in double-counting. Given these ambiguities, we can only calculate that the range of CARES Act covered, federally subsidized affordable housing units is likely more than 5,103,430 units (11.6 percent) and less than 7,659,700 (17.5 percent) of the national rental market.14

With regard to understanding the number of rental properties with federally backed mortgages, the necessary data are even more difficult to access than that of federally subsidized properties. The respective agencies administering these programs collect data that would provide insight into the scope of this category of covered properties.15 However, as of June 2020 when this article posted, these entities do not share this information such that policymakers and researchers can use it to understand CARES Act moratorium coverage. Recently, Fannie Mae and Freddie Mac released lookup tools that allow tenants who live in properties with five or more units to see if their building has a mortgage that has been securitized by either entity.16 The tool does not permit a user to access the overall database, preventing broader analysis of the location and prevalence of these properties across rental markets. Providing broader access to the data would help to better understand and respond to the impact that COVID-19 will have on rental properties in communities across the country. That said, even if these entities provide expanded access to information they hold, it would still be very difficult to know how many rental units are located on properties with federally backed mortgages. Most notably, no comprehensive database exists of single-family dwellings (one- to four-unit properties) that are rental properties, making it very difficult to account for this portion of rental properties. The Urban Institute estimates that 2,767,433 of the 22,326,507 (12.4 percent) total single-family units are federally backed.17

The U.S. Census's Rental Housing Finance Survey (RHFS) provides a measure of financial, mortgage, and property characteristics of rental housing properties in the United States. This survey includes a question on whether rental properties and units have federally backed mortgages. The most recent available results date back to 2015.18 These results indicate that approximately 11.2 percent of rental units are federally backed by the FHA, VA, Fannie, or Freddie.19 However, the RHFS estimate likely undercounts federally backed units because an estimated 31 percent of rental units that self-reported as having a mortgage also reported not knowing whether their mortgage was federally backed.20 Given this information, it is likely that many landlords—of both multifamily and single-family properties—do not know that their property is subject to an eviction moratorium, and may move to evict tenants from units that should be protected. Given the data limitations noted in this article, an updated RHFS is needed to gain insights into the national rental housing stock. Due to the RHFS data's age and likely underreporting, an alternative estimate was sought. The Urban Institute recently triangulated RHFS data with data on mortgage-backed securities and other census and HUD data to estimate that 28.1 percent of all rental units nationwide have a federally backed mortgage and are covered by the federal eviction moratorium.21

To get an understanding of national CARES Act coverage for rental units, we use the Urban Institute's estimate of federally backed mortgages, but we cannot simply add that number to our estimate of federally subsidized affordable housing units (the first CARES Act category), because substantial overlap likely exists between the two categories. Property-level data are not available across data sets to identify the extent of overlap. However, combining our estimated range of federally subsidized affordable housing units (5,103,430 to 7,659,700 units, or 11.6 percent to 17.5 percent of rental units nationwide) with the Urban Institute's estimate of 12,327,936 (28.1 percent nationwide) rental units with federally backed mortgages, we can approximate that the national coverage of the federal eviction moratorium is likely more than 12,327,936 rental units (28.1 percent) and less than 19,987,636 (45.6 percent) rental units.22

In closing: data access makes a difference, now and in the future
The lack of data necessary to make more precise calculations described in this article—whether because it has not been released or because it has not been recorded to begin with—carries implications beyond the federal CARES Act eviction moratorium. The list of covered properties in the CARES Act represents a body of rental properties that, due to their participation in a federal program, are more easily targeted for federal policy interventions. A federal effort that applies universally to all rental property may be more susceptible to legal challenges, and consequently less likely to be politically viable. Therefore, having access to robust data reflecting the prevalence and location of rental properties affected by these federal programs would contribute to vital solutions-oriented research at national and local levels. This research is urgent in the face of the present public health and economic crisis, and will continue to be vital as stakeholders tackle future housing affordability challenges.

The CARES Act eviction moratorium and other federal interventions in rental housing can have significant impacts in the states, cities, and even the neighborhoods where those properties are located. The ability to understand the location and concentration of properties with current protections against eviction would unlock the potential of these federal interventions to stabilize housing concerns and would allow policymakers at all levels to target the gaps that remain. In the case of the COVID-19 pandemic, the gap amounts to 7,659,700 (17.5 percent) renter households with uncertain coverage under the CARES Act eviction moratorium, in addition to 23,824,150 (54.4 percent) renter households that are likely not covered by these important CARES Act protections at all, depending on factors that cannot be accounted for given existing data limitations.23

By Sarah Stein, CED adviser, and Nisha Sutaria, CED research analyst II


1 The Shimberg Center for Housing Studies determined that statewide workers in the leisure and hospitality (14 percent) and retail trade (13 percent) categories are heavily affected by the economic slowdown and layoffs. See also an issue brief by the Federal Reserve Bank of Boston outlining the impact of COVID-19 on service workers in New England.

2 The 1,001,605 units are calculated by adding together the deficit of affordable and available units for extremely low-income renter households among Alabama, Florida, Georgia, Louisiana, Mississippi, and Tennessee. Extremely low income is defined as 30 percent of the area median income or less. Data are available at the Federal Reserve Bank of Atlanta's Southeastern Rental Affordability Tracker data tool, based on U.S. Census Bureau American Community Survey (ACS), 2015 one-year estimates from IPUMS. Data compiled by CED and the Shimberg Center for Housing Studies at the University of Florida.

3 U.S. Congress CARES Act, 2020. “Temporary Moratorium on Eviction Filings,” Section 4024.

4 For example, there is no database of one- to four-unit rental properties with federally backed mortgages.

5 For an in-depth legal analysis of federally subsidized properties that fall under the CARES Act, see the National Housing Law Project's Summary and Analysis of Federal CARES Act Eviction Moratorium.

6 Some FHA and USDA multifamily properties are both federally subsidized and encumbered by a federally backed mortgage.

7 Fannie Mae and Freddie Mac have created lookup tools for renters to check if their unit is in a property with a federally backed mortgage. This tool only lists multifamily properties (five-plus units). Renter in one- to four-unit properties do not have a way of knowing if their landlord has a Fannie- or Freddie-backed mortgage. The National Low Income Housing Coalition (NLIHC) has created a tool that combines available data from Fannie Mae, Freddie Mac, and other project-based affordable housing subsidies. NLIHC warns that the tool does not include all covered properties.

8 Due to the reporting requirements placed on these programs, publicly available data can be used to gain a general understanding of the scope of the eviction moratorium as it applies to federally subsidized properties. In a Q&A it published addressing CARES Act programs, the House Financial Services Committee suggested that NHPD could be used as a lookup tool for properties covered by the act.

9 Authors' calculation of rental units with federal subsidies. The number of units with federal subsidies are based on data aggregated from 2020 NHPD data and 2019 HOPWA data (state-based data, and national totals). The number of total rental units is based on 2018 Census American Community Survey one-year estimates. The following properties are included in the NHPD database: Low Income Housing Tax Credit (LIHTC), Section 8 Project Based Rental Assistance, Public Housing, USDA (Sections 414, 515, 521, and 538), Section 202 Direct Loans, Section 236, HUD Insured Mortgages, HOME Investments Partnership Program, and State Subsidies.

10 Data are from HUD Assisted Housing Data Portal, 2019 Summary of All HUD Programs Reports.

11 CARES Act coverage in unit totals for these programs could not be determined. As an indicator of the comparative size of these programs with HCVs, total 2019 congressional appropriations for these three programs were $51.8 million, or 0.2 percent of the 2019 congressional appropriation for HCVs ($22.6 billion).

12 Authors' calculation; 2.56 million HCV from HUD 2019 as a percent of 43.8 million estimate of total occupied rental units from census ACS 2018.

13 For example, as a requirement of participating in the Low Income Housing Tax Credit program, property owners must accept HCVs, indicating a likely overlap between these two programs. Although efforts have been made to account for the coincidence of HCVs and LIHTC, accurate reporting is lacking, and no consensus has been reached on the amount of overlap between LIHTC units and HCVs. A 2009 HUD study found that 47 percent of all LIHTC properties placed in service between 1995 and 2006 had one or more tenants with an HCV. This overlap could account for a significant portion of federally subsidized properties because LIHTC provides 2.4 million of the 5.1 million federally subsidized units in our national count. Based on data from Florida Housing Finance Corporation, the University of Florida's Shimberg Center estimates that about 15 percent of Florida's HCVs are used in LIHTC units, and conversely about 9 percent of LIHTC units are occupied by voucher holders.

14 Authors estimated a range of federally subsidized units based on the most conservative to the least conservative possible number of units. The lower-bound conservative figure is calculated assuming that all HCVs coincide with federally subsidized units, therefore, the share remains 11.6 percent. The upper-bound, least conservative figure assumes that no HCVs coincide with federally subsidized units, therefore, an additional 5.8 percent of all rental units yields an estimated share of 17.5 percent.

15 FHA publishes a database of its insured multifamily mortgages. The Shimberg Center has extracted a list of multifamily FHA properties in Florida. In an examination of that list, Shimberg found that 38 percent of FHA multifamily units appeared to overlap with federally subsidized properties, most of which would already be covered by CARES Act protections.

16 Tenants in one- to four-unit properties and researchers or policymakers with questions about market coverage cannot use these tools to understand CARES Act coverage.

17 Goodman, Laurie, Karan Kaul, and Michael Neal. (2020). “The CARES Act Eviction Moratorium Covers All Federally Financed Rentals—That's One in Four US Rental Units.” Urban Institute.

18 The 2018 RHFS data were released after this article posted.

19 The 2015 version of RHFS did not separately account for USDA mortgages. Data from U.S. Census Bureau's Rental Housing Finance Survey (RHFS) tool, 2015.

20 Authors' calculation of 31 percent of property owners with a mortgage do not know if their mortgage is federally backed calculated using 2015 RHFS Public Use File (PUF)

21 The Urban Institute methodology uses data from the 2018 American Community Survey, eMBS Inc. data, the 2015 Rental Housing Finance Survey, the U.S. Department of Housing and Urban Development, and the Federal Housing Finance Agency to estimate the portion of rental units with federally backed mortgages. See also Laurie Goodman, Karan Kaul, and Michael Neal. (2020.) “The CARES Act Eviction Moratorium Covers All Federally Financed Rentals—That's One in Four US Rental Units.” Urban Institute.

22 Authors' estimate of a range of total units covered by CARES is based on the most conservative to the least conservative possible number of units. Authors and Urban Institute use the same data source, the 2018 American Community Survey, for total number of rental units, making shares additive. The lower-bound conservative figure is calculated assuming that Urban Institute's estimate of federally backed units and the authors' upper range estimate of federally subsidized rental units overlap with each other entirely, therefore, the share remains 28.1 percent. The upper-bound figure is calculated assuming that Urban Institute's estimate of units with a federally backed mortgage and the authors' estimated upper range of federally subsidized rental units do not overlap at all, therefore, an additional 17.5 percent share of units yields 45.6 percent.

23 7,659,700 renter households (17.5 percent) is the difference between the upper bound of coverage (45.6 percent) and lower bound of coverage (28.1 percent). 23,824,150 million renter households (54.4 percent) is the difference between full coverage of 43,811,786 renter households (100 percent) and the upper bound of coverage (45.6 percent).