Measuring Evictions during the COVID-19 Crisis
Evictions are a serious risk for households facing job loss and economic upheaval during the COVID-19 pandemic, and temporary policies put in place to protect renters are beginning to expire. To understand how the crisis is affecting evictions, we measured eviction filing activity across 44 cities and counties. As of July 7, 2020, eviction filings have almost returned to their prepandemic levels in places where local bans have expired or where they were never enacted. We find that eviction filings tend to surge after temporary policies expire much more in places that enacted both filing bans and hearing bans than those that enacted just hearing bans while allowing filings to continue. As federal stimulus supplements for the unemployed expire, evictions are likely to increase for households that have lost work because of the crisis unless there is material improvement in the economy (Mervosh 2020).
Housing advocates and policymakers worry that the sudden upheaval of household finances resulting from the economic crisis associated with the COVID-19 pandemic will lead to a wave of evictions for tenants who are no longer able to pay their rents. In this report, we explore how evictions are evolving in real time and how state and local eviction regulations affect the trends in eviction filings as the pandemic crisis unfolds. We collected weekly1 eviction filing data for 44 jurisdictions across the United States to evaluate how eviction activity has evolved since the crisis began. Eviction filings plummeted at the beginning of the economic crisis in the United States, both in places that implemented broad eviction bans and those that did not. As of July 7,2 roughly one-third of rental units in our study are no longer covered by temporary policies, and eviction filings have now returned to their prepandemic levels in those places no longer covered; filings remain subdued in places that have continued eviction bans. In addition to expiring eviction bans, renter households face expiring supplemental unemployment benefits offered through the CARES Act in spite of a still-elevated unemployment rate, a situation which further raises the risk of eviction for households impacted by the crisis (Mervosh 2020).
Pandemic-driven Eviction Policies
The eviction risk to renters is difficult to grasp on a large scale because pandemic eviction policy is a patchwork of federal, state, and local initiatives. At the federal level, an eviction ban for nonpayment of rent covers many of the properties receiving financing through federal programs until July 25 (Goodman et al. 2020), but an enforcement mechanism for the federal ban is lacking (Ernsthausen et al. 2020). A team at Columbia Law School maintains a database of state-level policies (Benfer 2020). At the local level, many jurisdictions have enacted additional policies beyond their individual state’s policies.
Few resources are available to monitor eviction actions—particularly at a weekly or monthly frequency and in real time. Many cities lack altogether accessible data on their court proceedings. Princeton’s Eviction Lab (Desmond 2020) has recently begun tracking weekly filing data in its Eviction Tracking System for 11 jurisdictions at the census tract or ZIP code level (Hepburn et al. 2020); this resource is particularly well-suited to illustrate how eviction activity varies across neighborhoods within a city as the COVID-19 crisis unfolds. For instance, the Eviction Lab tracker details eviction rates by neighborhood demographics, demonstrating that filings both before and after COVID-19 filing bans were lifted tend to be significantly higher in majority Black or Latinx neighborhoods.
We developed a measure that captures current eviction filing activity at the county or city level in 44 jurisdictions.3 We combined these data with information on the relevant state- or local-level eviction policies to examine how the policies affect ongoing eviction filings.
The formal eviction process generally begins when a landlord finds reason to evict a tenant, because of nonpayment of rent or otherwise, and presents the tenant with a notice to vacate the premises. If the tenant does not comply, a landlord may file an eviction case. A hearing is then held, and if the landlord prevails, an order to execute the eviction is issued by the court. We focus on the filing stage of the formal eviction process, allowing for the possibility that trends in formal eviction actions before the filing stage or in evictions that occur outside the court system entirely may be different.
The jurisdictions we examined fall under a range of eviction policies, including 1) places that never enacted their own temporary bans, 2) places that restricted hearings but not filings for a period of time, and 3) places that restricted both filings and hearings. Among the places in our study, restrictions on hearings (hearing bans) generally required courts to delay the hearing phase of any eviction case filed until a future date. Filing restrictions (filing bans) generally forbade the courts from accepting new case filings from landlords until a future date. While some places banned all filings, others made exceptions for certain cases.4
Even in the places with no state or local ban on filings or hearings, some jurisdictions granted judges extra discretion for handling cases during the pandemic. Such places also fall under the federal moratorium for properties financed with federal dollars in some way. In the discussion that follows, we describe these places as having “no ban.” For more details on how we construct our measures of eviction activity and classify eviction policies, please refer to the appendix.
Different eviction policies mean different trends in eviction filings
Figure 1 shows weekly year-over-year changes in eviction filing activity in four places with a variety of eviction bans. For example, in Cleveland, Ohio, eviction filings quickly fell as citywide eviction bans on hearings and filings for nonpayment cases took effect on March 15. The bans extended until June 15. In the week after the bans expired, eviction filings briefly rose to twice the level of the same period in 2019 before falling below prepandemic levels in subsequent weeks. Milwaukee, Wisconsin, enacted shorter hearing and filing bans than did Cleveland and experienced a similar resurgence in eviction filings. In contrast with Cleveland, eviction filings in Milwaukee remain elevated above prepandemic levels in the weeks following the policy expiration. Denton County, Texas, which imposed a hearing ban only, saw a sharp fall in eviction filings during the ban period. Filings have risen since the ban expired, but they have not yet returned to prepandemic levels. Faulkner County, Arkansas, imposed no blanket eviction bans but allowed judges to handle cases at their own discretion (Baker 2020).5 It, too, saw a drop in filings from April until June, but then filings surged.
Eviction filings fell across all jurisdictions in late March and early April, regardless of the policy type
In figure 2, we calculate the average weekly year-over-year change in filing activity within groups of jurisdictions according to the eviction policy under which they operated at the time. A place, then, may move from one group to another as its policies take effect or expire.6
Eviction filings declined sharply in the latter weeks of March regardless of the type of policy a jurisdiction adopted. Declines were sharpest among jurisdictions that implemented both hearing and filing bans, though the level does not fall to zero, as not all filing bans were absolute. The declines for places with only hearing bans and for those with no temporary bans were also sharp during this period: eviction filings declined by more than 80 percent in each of these policy groups by mid-April.
Eviction filings remain low in places with hearing or filing bans and are gradually rising in places with no bans
In late April and early May, filings in places with no bans began to rise. The number of jurisdictions without bans grew from 5 to 14 as bans expired. As of July 7, just less than 40 percent of the rental units in the sample are in places with no bans. The latest data indicates eviction filings are just 3 percent below their prior-year levels in places with no bans. In contrast, filings remain far below their prepandemic levels in places that continue to impose a hearing ban and places that impose both hearing and filing bans.
A substantial number of places have lifted temporary bans since the pandemic began. As shown by Cleveland and Milwaukee in figure 1, however, the response to expiring policy can vary from place to place. To understand how filings evolve on average after temporary bans expire, we illustrate the eviction filing trends in places that transitioned either from a hearing ban only to no bans or from both hearing and filing bans to no bans.
In figure 3 we rearrange jurisdictions into two event studies based on the timing of the expiration of their temporary bans. Nine jurisdictions had both a filing ban and a hearing ban expire simultaneously, and nine different jurisdictions transitioned from hearing bans only to no bans.7 The horizontal axis shows the number of weeks since the ban expired, with zero marking the week in which it expired. As in figure 2, the jurisdictions comprising a particular week’s reading can change from week to week, and the levels in the event study are the average levels among jurisdictions in the group that are a given number of weeks from the policy expiring. The result is an estimate of the average eviction filing level corresponding to a given number of weeks before or since an eviction ban expired.8
Places with only hearing bans experienced a modest increase in filings after the ban expired
Places with only hearing bans had modestly elevated filings after the ban expired. Eviction filings were 66 percent below the level one year prior in the six weeks leading up to the expiration of the ban. During the week of expiration, filings rose to 54 percent below their level a year ago. As the bans expired, filings rose and remained elevated: in the six weeks following the expiration of the ban, eviction filings averaged 50 percent below their prior year’s levels.
Places with both filing bans and hearing bans experienced a sharp increase in filings in the weeks following the ban expiration
In places where both filings and hearings were banned, there is a sharp increase in filings following the simultaneous expiration of both bans. While the eviction filings average 99 percent below year-ago filings in the six weeks leading up to expiration, they exceed the prior year level by 19 percent in the week after the ban expired. On average, eviction filing activity returned to 2019 levels in the five weeks after the bans expired.
It is particularly important to treat these early results with caution. They may change as more jurisdictions’ bans expire, as policies expire under different economic conditions, or as households’ abilities to meet rent obligations change. These results should also not be interpreted as causal, as jurisdictions may have chosen their own eviction policies in response to local conditions that determine eviction trends.
Conclusion and Policy Implications
Eviction filings have almost returned to their prepandemic levels in jurisdictions with no bans in place. Both filing and hearing bans appear to be effective at temporarily subduing eviction filings. Furthermore, each type of ban has different implications for filing behavior both during and after the period in which the ban is in effect. Hearing bans alone depressed filings by about 66 percent, and, while there was a moderate increase in filings as bans lifted, overall filings did not reach their year-ago levels after the bans expired. As expected, combined filing and hearing bans decreased filings by about 98 percent when in effect; however, after the bans lifted, filings surged to a much higher level on average than in places with expiring hearing bans alone.
There are reasons to be concerned that the eviction filing rates in places that have lifted bans may not remain at their current levels into the future. First, there remains a large number of renter households that have experienced an economic shock as a result of the crisis—the Urban Institute estimates that in 20 percent of renter households, at least one member lost a job between February and April (Reynolds and Shroyer 2020). Second, renters report current or anticipated future difficulty in paying rent—the Census Household Pulse survey (Census 2020) from June 25 through June 30 shows that 18 percent of renters did not pay rent in May, and 32 percent had either slight or no confidence in their ability to pay their next rent bill. Third, the federal stimulus program that provides increased benefits to laid-off workers through the CARES Act expires in July. This program has increased unemployment compensation to include $600 per week of additional assistance beyond state-administered benefits. The Census Household Pulse survey reported stresses despite many workers’ receiving this increased benefit at the time.
Previous work suggests that nonpayment accounts for almost 80 percent of evictions (Lerner 2017); on-time rent payment rates may therefore provide an early warning of elevated eviction risk . Two private-market rent payment indicators produced by the National Multifamily Housing Council and Entrata suggest that, so far, on-time payments are comparable to their prepandemic levels.9 It is unclear whether the composition of renters included in these indicators is representative of the nation, however. Eviction activity, therefore, also bears close monitoring as the crisis continues to evolve and federal stimulus programs wind down.
Finally, it is important to note that we measure formal eviction filings only; by one estimate, informal evictions are twice as common as formal ones (Flowers 2016). Further work is needed to understand how informal eviction, which is largely hidden from the legal system and from systematic measurement efforts, has evolved alongside these eviction policies during this crisis.
The severe consequences of eviction will be only amplified in a pandemic. Collinson and Reed (2020) connect evictions to increased risk of homelessness, mental health hospitalizations, and emergency room visits, situations which could put evicted tenants at a greater risk of contracting, spreading, and suffering complications from COVID-19 (CDC 2020). Studies have also shown that eviction can lead to job loss (Desmond and Gershenson 2020), an especially troubling outcome in a labor market already devastated by COVID-19. These factors are worth considering for policymakers concerned with avoiding future increases in evictions as the economic crisis continues.
In this appendix, we discuss the policy landscape, sample coverage, and methods used to construct the measures we report in greater depth.
We collected counts of eviction filings from each of 44 jurisdictions10 for the consecutive periods ending on the 7th, 14th, 21st, and last day of each month from January 1 through July 7, and computed jurisdiction-level year-over-year changes based on the same period in 2019. This approach accounts for seasonality as well as weekly and monthly calendar effects while providing a timely measure of eviction activity.11 A measure of -20 percent on January 7 in Cleveland, for instance, means that eviction filings were 20 percent lower in the period of January 1–7, 2020, than they were in the period of January 1–7, 2019 (see figure 3). We first calculate year-over-year changes by jurisdiction and then computed a weighted average using the number of renter-occupied housing units in 2018 for each jurisdiction in the group. We computed measurements for the sample overall and for subsets of jurisdictions to analyze how the temporary eviction policies affect eviction filings.
Figure A1 below maps the 44 jurisdictions in our sample, which spans 11 states and include 7 percent of the rental housing stock in the United States, according to 2018 ACS estimates (Manson et al. 2019). Compared to the nation, cities and counties in our sample are more urban (94.3 versus 80.7 percent urban population), have similar income ($60,293 versus $60,715 median household income), and have slightly higher minority populations (29.5 percent versus 27.3 percent people of color).12 The smallest jurisdiction is Salem, Massachusetts, with 9,060 occupied rental units, and the largest is Sacramento County, California, with 235,947. Jurisdictions were included based on the availability of online filing data.
Details about Eviction Policies
Figure A2 shows the portion of rental units in our study that fall under a given type of pandemic eviction policy over time. Four of the 44 places we examined, containing just 5 percent of the rental units in the study, either never implemented an eviction ban or left discretion to a jurisdiction’s judges.13 Eighteen jurisdictions, representing 45 percent of rental units, implemented only a hearing ban for some period of time. In the remaining 22 jurisdictions, policymakers banned most new filings and hearings for a period of time.
The eviction policies include a wide range of policy start and end dates and vary in length: Little Rock, Arkansas, banned hearings for just two weeks in March, while Sacramento County (in California) and the entire state of Massachusetts implemented bans for four months. A number of policies have been extended as the crisis continues; as of July 7, 30 percent of rental units in the study fall under hearing bans and an additional 30 percent fall under both hearing bans and filing bans.14 Table A1 includes the length of these bans for each jurisdiction in our sample, including extensions made on or before July 7.
Table A1 displays the type and duration of eviction policies enacted in each jurisdiction in our sample.
Table A1. Eviction protections by jurisdiction
|Jurisdiction||Filing ban length (days)||Hearing ban length (days)|
|Benton County, AR||-||56|
|Brevard County, FL||-||121*|
|Brown County, WI (Green Bay)||62||62|
|Cass County, ND (Fargo)||-||-|
|Charleston County, SC||58||58|
|Comal County, TX||-||64|
|Dane County, WI (Madison)||62||62|
|Denton County, TX||-||64|
|Faulkner County, AR||-||-|
|Fort Bend County, TX||-||64|
|Greenville County, SC||58||58|
|Hamilton County, OH (Cincinnati)||-||84|
|Hennepin County, MN (Minneapolis)||111*||111*|
|Hillsborough County, FL (Tampa)||-||121*|
|Milwaukee County, WI||62||62|
|Nevada County, CA||-||135*|
|Nueces County, TX (Corpus Christi)||-||64|
|Orange County, FL (Orlando)||-||121*|
|Palm Beach County, FL||-||121*|
|Pinellas County, FL (St. Petersburg)||-||121*|
|Pulaski County, AR (Little Rock)||-||11|
|Racine County, WI||62||62|
|Ramsey County, MN (St. Paul)||111*||111*|
|Richland County, OH (Mansfield)||-||-|
|Richland County, SC (Columbia)||58||58|
|Rock County, WI (Janesville)||62||62|
|Sacramento County, CA||112*||112*|
|Sangamon County, IL (Springfield)||99*||99*|
|Sarasota County, FL||-||121*|
|Stearns County, MN||111*||111*|
|Ventura County, CA||131*||131*|
|Volusia County, FL||-||121*|
|Washington County, AR (Fayetteville)||-||45|
Note: Table A1 displays the duration of eviction policies in each jurisdiction, in days. Numbers are updated to include the extension of any ban, and are current as of July 8. All bans were implemented in either March or April; those still in effect as of July 8 are marked with an asterisk; in these cases, the length includes the future end date. Counties that include widely-recognized principle cities show those cities in parenthesis. Source: Eviction policies from Benfer (2020), public notices and authors’ calculations.
- We collect data for the consecutive periods ending on the 7th, 14th, 21st, and last day of each month, and refer to these periods, which range from 7 to 10 days in length, as weeks. For more details about our data please see the appendix. Return
- Unless otherwise stated, all dates refer to the year 2020. Return
- Our analysis overlaps with the Eviction Lab in four jurisdictions: Milwaukee, Wisconsin; Cleveland and Cincinnati, Ohio; and Boston, Massachusetts. Return
- For example, statewide directives in Massachusetts and Wisconsin, as well as a city-level order in Cleveland, Ohio, barred only nonpayment and other nonemergency filings. Among the places we study, all that enacted a filing ban also enacted a hearing ban during the same period. Return
- For Faulkner County’s judicial discretion policy, see http://www.faulknercounty.org/images/documents/COVID19/FC Press Release 3-18-20.pdf Return
- A jurisdiction appears in a particular category only if it operates under that category’s policies for the entire period. Jurisdictions are excluded from any category during the period in which they change policies mid-period. The average is weighted by the estimated number of renter-occupied housing units in each jurisdiction in 2018. Return
- While Youngstown, Ohio, had both hearing bans and filing bans, the ban expiration dates were staggered by multiple months. We excluded Youngstown from these figures to focus on places that moved from a temporary ban to no bans. Return
- Unlike in figure 2, weeks that straddle the expiration of an eviction policy are shown. The first week that includes any days following the expiration of a policy is recorded as week 1. The measures are right-censored when fewer than seven jurisdictions are available; this occurs at week 6 (panel A) and week 5 (panel B). Return
- As of June 30, the National Multifamily Housing Council reports that 80.8 percent of roughly 11 million units in its sample paid rent on time in June (within the first six days of the month), compared with 81.6 percent in June 2019. Entrata, a property management software provider, reports that more than 94 percent of households paid rent in May. Return
- We also collected data for Cook County, Illinois, but did not report it in our results. Cook County experienced a large, unexplained, and sustained increase in eviction filings beginning in the middle of 2019, making the data unsuitable for understanding changes in filings during the pandemic. Return
- An analysis of individual records on eviction notices filed with the San Francisco Rent Board reveals that there are both within-week and within-month calendar effects in eviction filings, a situation which is consistent with the most common rent payment period being monthly. The specific patterns of calendar effects may vary from jurisdiction to jurisdiction. Return
- Comparison values are the average percent urban population, median household income, and racial or ethnic composition at the jurisdiction level, weighted by total population, compared to the nation’s. Return
- For an example of an order permitting judicial discretion, see https://www.ndcourts.gov/legal-resources/rules/ndsupctadminorder/25 Return
- For a list of cities and counties and a summary of their eviction policies, see table A1. Return
- Benfer, Emily. 2020. “COVID-19 Eviction Moratoria by State, Commonwealth, and Territory.” Wake Forest University School of Law. https://docs.google.com/spreadsheets/d/e/2PACX-1vTH8dUIbfnt3X52TrY3dEHQCAm60e5nqo0Rn1rNCf15dPGeXxM9QN9UdxUfEjxwvfTKzbCbZxJMdR7X/pubhtml.
- Centers for Disease Control and Prevention. 2020. “Interim Guidance on Unsheltered Homelessness and Coronavirus Disease 2019 (COVID-19) for Homeless Service Providers and Local Officials.” https://www.cdc.gov/coronavirus/2019-ncov/community/homeless-shelters/unsheltered-homelessness.html.
- Collinson, Robert and David Reed. 2018. “The Effects of Evictions on Low-Income Households.” Working Paper, New York University. https://www.law.nyu.edu/sites/default/files/upload_documents/evictions_collinson_reed.pdf.
- Desmond, Matthew, et al. 2020. The Eviction Lab at Princeton University [Database]. Princeton University. www.evictionlab.org.
- Desmond, Matthew and Carl Gershenson. 2016. “Housing and Employment Insecurity among the Working Poor.” Social Problems. Oxford University Press, 63(1), pp. 46-67. doi:10.1093/socpro/spv025.
- Ernsthausen, Jeff, Ellis Simani and Justin Elliott. 2020. “Despite Federal Ban Landlords Are Still Moving To Evict People During the Pandemic.” ProPublica. https://www.propublica.org/article/despite-federal-ban-landlords-are-still-moving-to-evict-people-during-the-pandemic.
- Flowers, Andrew. 2016. “How We Undercounted Evictions By Asking The Wrong Questions.” FiveThirtyEight. https://fivethirtyeight.com/features/how-we-undercounted-evictions-by-asking-the-wrong-questions/.
- Goodman, Laurie, Karan Kaul, and Michael Neal. 2020. “The CARES Act Eviction Moratorium Covers All Federally Financed Rentals—That’s One in Four Rental Units.” Urban Institute, Washington, DC. https://www.urban.org/urban-wire/cares-act-eviction-moratorium-covers-all-federally-financed-rentals-thats-one-four-us-rental-units.
- Hepburn, Peter, Renee Louis, and Matthew Desmond. 2020. Eviction Tracking System: Version 1.0 [Database]. Princeton University. https://evictionlab.org/eviction-tracking/.
- Lerner, Michele. 2017. “Does Your City Rank High or Low When It Comes to Evictions?” The Washington Post. https://www.washingtonpost.com/news/where-we-live/wp/2017/12/28/does-your-city-rank-high-or-low-when-it-comes-to-evictions/.
- Manson, Steven, Jonathan Schroeder, David Van Riper, and Steven Ruggles. 2019. IPUMS National Historical Geographic Information System: Version 14.0 [Database]. University of Minnesota, Minneapolis. doi:10.18128/D050.V14.0.
- Mervosh, Sarah. 2020. “An ‘Avalanche of Evictions’ Could Be Bearing Down on America’s Renters.” The New York Times https://www.nytimes.com/2020/05/27/us/coronavirus-evictions-renters.html.
- Reynolds, Kathryn and Aaron Shroyer. 2020. “To Stay Stably Housed, Renters Need $16 Billion per Month in Housing Support during the COVID-19 Crisis.” Urban Institute, Washington, DC. https://www.urban.org/urban-wire/stay-stably-housed-renters-need-16-billion-month-housing-support-during-covid-19-crisis.
- United States Census Bureau. 2020. “Week 9 Household Pulse Survey: June 25–June 30.” https://www.census.gov/data/tables/2020/demo/hhp/hhp9.html.
The views expressed in this report are those of the authors and are not necessarily those of the Federal Reserve Bank of Cleveland or the Board of Governors of the Federal Reserve System.