Abstract
Drawing on over 8 million eviction court records from twenty-eight states, this study shows the role that eviction filings play in extracting monetary sanctions from tenants. In so doing, it documents an unanticipated feature of housing insecurity: serial eviction filings. Serial eviction filings occur when a property manager files to evict the same household repeatedly from the same address. Almost half of all eviction filings in our sample are associated with serial filings. Combining multivariate analysis with in-depth interviews conducted with thirty-three property managers and ten attorneys and court officials, we document the dynamics and consequences of serial eviction filings. When legal environments expedite the eviction process, property managers use the housing court to collect rent and late fees, passing costs on to tenants. Serial eviction filings exacerbate tenants’ housing cost burden and compromise their ability to find future housing. Using tract-level rent and filing fees, we estimate that each eviction filing translates into approximately $180 in fines and fees for the typical renter household, raising their monthly housing cost by 20%. The study challenges existing views of eviction as a discrete event concentrated among poor renters. Rather, it may be better conceived of as a routinized, drawn-out process affecting a broader segment of the rental market and entailing consequences beyond displacement.
Introduction
In recent years, rents have increased much faster than wages. Nationwide, median rent more than doubled over the last two decades, from $483 in 2000 to $1,002 in 2019 (U.S. Census Bureau 2019). Yet between 1979 and 2011, the bottom 90 percent of workers saw only 15 percent cumulative growth of real annual wages (Mishel and Finio 2013). Increasing housing cost burden among renting families places a growing number at risk of eviction (Desmond 2016; Myers and Park 2019). Over a million households are evicted each year (Desmond et al. 2018a). By and large, researchers have treated eviction as a discrete event involving the displacement of poor, severely rent-burdened tenants (Desmond 2016; Lundberg and Donnelly 2019). In describing serial eviction filings, this study challenges that understanding and shows the threat of eviction to be a routine, drawn-out process affecting those beyond the bottom of the rental market and not always resulting in displacement, yet nonetheless exacerbating financial precarity.1
Drawing on over 8 million court records from 28 states, we describe and analyze the prevalence of serial eviction filings. In 2014, nearly one-third of households associated with eviction filings were filed against more than once from the same address. Serial filings were most common in neighborhoods with mid-range rental prices and in areas where more evictions were filed by businesses rather than individual landlords. Serial filing rates were lowest in areas where legal and regulatory barriers increase the cost of eviction.
To remain in their homes, tenants threatened with eviction must pay late fees and court costs in addition to past-due rent. Using tract-level median rent and county-level filing fee data, we estimated that each eviction filing translated into $180 in fines and fees for renter households in the average tract, effectively raising monthly housing costs by 20%. By documenting these hidden costs, we show that a comprehensive analysis of housing cost burden must go beyond conventional measures to capture monetary sanctions imposed through the civil courts.
To understand landlord-tenant and legal dynamics driving serial filings, we interviewed thirty property managers and ten attorneys and court officials in Charleston, South Carolina, and Mobile, Alabama. Charleston and Mobile shared a range of sociodemographic and housing market characteristics, including similar overall eviction rates, but had strikingly different serial filing rates. Our interviews highlight the work of professional property managers, who play an increasingly influential role in U.S. housing markets (Thacher 2008; U.S. Census and Statista 2018). Interviews demonstrate how property managers respond to different local laws and regulations. In Mobile, where the law has made formal evictions costly, property managers used eviction court primarily to reclaim property. In Charleston, where formal evictions were cheap and quick, eviction court functioned like the court of first, not last, resort. These findings reveal that eviction dynamics are steered not only by economic trends, but also by legal regimes, ownership structures, and property management strategies.
Eviction as Tenant Discipline
A formal eviction occurs when a civil or housing court denies a household tenancy rights to a rental unit in which they have been living. Once an eviction judgement has been granted, a property owner may summon local law enforcement agencies to remove tenants and their belongings. Forced displacement from housing has been linked to job loss, maternal depression, homelessness, downward neighborhood moves, and other negative outcomes (Collinson and Reed 2018; Desmond and Kimbro 2015).
Researchers have devoted less attention, however, to how property owners may use the eviction process to accomplish goals other than reclaiming property. Property owners may rely on the threat of eviction to collect rent and discipline tenants (Garboden and Rosen 2019; Immergluck et al. 2019). This strategy may also allow property owners to profit from late payments. Tenants threatened with eviction not only must pay their rent in full; they are also charged late fines and legal fees, including costs associated with attorneys and court filings. These hidden costs of the housing crisis may be common if landlords routinely rely on eviction court for rent collection.
Changes to the ownership of rental housing may play an important role in explaining serial eviction filings. The rise of “automated” and private equity landlords (Fields 2017, 2019)—professional property-owners who are less inclined to “give weight to social and personal considerations” (Gilderbloom and Appelbaum 1988: 93) when dealing with their tenants—means that renters may find it difficult to establish informal arrangements with property management when late on rent. Rental property ownership is consolidating in fewer hands,2 and an increasing number of owners partner with property management companies (Mallach 2007; Thacher 2008). Professionalized landlords and managers rely on a suite of software tools to track housing stock, expenditures, and income, as well as automate eviction filings (Hartman and Robinson 2003; Immergluck 2013). The consolidation of tenant information and availability of these tools, along with expanding property management portfolios, encourage an explicit formalization and monetization of the landlord-tenant relationship (Kleysteuber 2007).
Serial eviction filings may also be a function of misalignment between rent deadlines and payment schedules. Traditionally, rent is owed on the first of the month “in part because many tenants get their paychecks on the last workday of the month” (Stewart et al. 2016: 71). However, only 5.4 percent of private businesses in the United States now pay their employees on a monthly basis. The remainder pay workers on a semimonthly (18.6%), biweekly (42.2%), or weekly (33.8%) basis (U.S. Bureau of Labor Statistics 2019). Consequently, wages sufficient to cover monthly rental expenses often are not fully available when rent comes due. Moreover, income instability has increased since the 1970s (Gottschalk and Moffitt 2009), as rising schedule uncertainty affects workers’ take-home pay (Lambert et al. 2014). Together, these trends have compromised families’ ability to consistently make rent by the first of the month (Morduch and Schneider 2017).
What at first presents itself as a liability to property owners can be transformed into opportunity through the eviction process. Because the loss of housing is significantly costlier than paying fines and fees associated with the eviction process (Desmond 2012; Desmond and Shollenberger 2015), tenants are incentivized to pay these costs and remain in place. That incentive grows each time a property owner files for eviction, as each filing generates a court record that can limit tenants’ future housing options. In screening prospective tenants, property owners tend to view recent eviction records as disqualifying, even in cases where the court process did not render an eviction judgement (Desmond 2016; Hubert 1995). Ironically, then, an eviction filing may be a strategy used to retain tenants, not displace them, by rendering them ineligible for competing housing options.
Hypotheses
These factors motivate hypotheses on how different neighborhood, property-owner, and court characteristics are associated with serial eviction filings. Many of the mechanisms are assumed to operate at the level of the landlord-tenant relationship. As a function of available data—which provide limited information about either party in eviction cases—we conduct analyses at the tract level and higher, which yields an imperfect representation of the underlying mechanisms.
First, regarding neighborhood characteristics, up-market neighborhoods typically cater to financially advantaged tenants able to consistently make rent; while in down-market neighborhoods, where rent burdens are most acute, tenants cannot afford additional fines and fees to avoid eviction (Immergluck et al. 2019). Tenants in mid-range rental housing markets, however, may routinely miss rent deadlines but have the means to eventually pay rent and accompanying late and court fees. Accordingly, we hypothesize that neighborhoods in middle-range rental markets will have the highest serial eviction filing rates.
Gentrifying neighborhoods may prove the exception to this pattern. Property owners in these neighborhoods may be motivated to replace tenants with new customers who can afford higher rents (Mele 2000; Stabrowski 2014). When this opportunity presents itself, landlords might be more likely use eviction to displace, rather than to discipline, tenants. Accordingly, we hypothesize that gentrification will be associated with lower serial eviction filing rates.
Otherwise similar housing markets may be controlled by different types of landlords. When property owners have a personal relationship with tenants, issues may be solved without involving the legal system. This arrangement can make rent more flexible: it may be paid late without sanction or satisfied with in-kind efforts (Gilderbloom 1989; Sternlieb 1969). The professionalization of rental management compromises tenants’ ability to form personal relationships with landlords and encourages property owners to rely on uniform rent-collection strategies (Kleysteuber 2007). We hypothesize, then, that neighborhoods with a higher concentration of corporate property owners will have higher serial eviction filing rates.
Last, serial eviction filings may vary in different legal environments. Civil courts often accentuate and manifest landlord power (Bezdek 1992). The vast majority of tenants facing eviction lack legal counsel (Engler 1997; Seron et al. 2001), while property owners have been instrumental in shaping housing law to protect their interests (Glendon 1982). However, landlord-tenant laws vary widely and may affect filing behavior (Hatch 2017; Sullivan 2017). In jurisdictions with high filing fees and requirements to hire attorneys, for instance, eviction court may prove a poor rent-collection strategy because tenants will not be able to routinely bear the associated costs. Accordingly, we hypothesize that counties with more legal and regulatory barriers that increase the cost of eviction will have lower serial eviction filing rates.
Data and Methods
This study combines quantitative and qualitative data. Quantitative analysis of millions of eviction records enables us to examine the prevalence of and variations in serial filings from national perspective. Qualitative analysis of forty-three in-depth interviews describes property managers’ rent-collection strategies and eviction behaviors.
Eviction Records
We analyzed the prevalence of serial eviction filings across 958 counties in 28 states, covering roughly one-third of U.S. renter households.3 We included all counties for which court records were available and could be validated against publicly available data sources between 2010 and 2016.4 We drew on 8,108,674 eviction cases filed over this period, with a primary focus on 1,082,845 cases filed in 2014. We focused on 2014, the most recent available year with sufficient subsequent years of court records. While most serial eviction cases are filed within a year—the median time between serial filings is 2.4 months—some take place across longer periods of time. As such, we included records from years prior and subsequent to 2014. Our sample included Alabama, Delaware, Florida, Georgia, Iowa, Illinois, Indiana, Kansas, Kentucky, Massachusetts, Michigan, Minnesota, Missouri, Mississippi, North Carolina, Nebraska, New Mexico, Nevada, Ohio, Oklahoma, Oregon, South Carolina, Texas, Utah, Virginia, Washington, Wisconsin, and West Virginia.
Eviction records include case numbers, names of plaintiffs (landlords, property managers, or their agents) and defendants (tenants), defendant addresses, and filing dates. These records contain no information about the lease contract or socio-demographic characteristics of either party, limiting individual-level analysis of serial filing risk. We cleaned these records, removed duplicate and commercial eviction cases, geocoded, and validated records against publicly available data sources published by county- and state-court systems (Desmond et al. 2018b). We identified serial filings in the data using case numbers, tenant names, and tenant addresses: distinct cases that shared the same (standardized) defendant name and address were linked.5 We calculated the serial filing rate: the number of households filed against two or more times divided by the total number of unique households in the records. The serial filing rate measures how many households facing an eviction filing within a given geographic unit have been repeatedly filed against at the same address.
Our analysis described the prevalence of serial filings and explored differences across subgroups. We began by describing variations in serial eviction filing rates between and within the states in our sample. We analyzed variations in the frequency with which households were filed against and the timing of these filings. We simulated the cost of “paying to stay” for serially evicted households. In each tract, we added the county filing fee—the court fee required to initiate an eviction, typically shifted onto tenants—to 10 percent of tract median rent, a standard used to set late fees commonly heard in our interviews. We collected data on filing fees for 3,118 counties and county-equivalents across the United States; median rent figures are from the American Community Survey (ACS) 5-Year estimates for 2012–2016.6 We divided this measure by tract median rent to estimate the additional share of rent collected through filings. Resulting figures should be treated as approximate; we derived our assumptions from a limited set of interviews, and institutional patterns in other states may result in higher or lower costs per eviction. Factors that may result in higher costs, such as attorney fees for either party, were not included because such data are not systematically reported.
The dependent variable (|${Y}_{ij}$|) was the count of serial filings in tract i of county j in 2014; it followed a negative binomial distribution with mean parameter |${\mu}_{ij}$| and dispersion parameter |$\theta$|. At level 1, we modeled the |${\mu}_{ij}$| parameter with a log link function. We included in the model an offset, |$\mathit{\log}\big({n}_{ij}\big)$|: the natural logarithm of the total number of unique renter households receiving an eviction filing in the tract in 2014. The resulting coefficients (both the |$\beta$| and |$\gamma$| terms) should be interpreted in terms of the log rate of serial eviction filings.
We included variables that addressed our hypotheses and controlled for characteristics commonly associated with eviction. The independent variables at level 1 were tract-level characteristics described below; all of these variables were grand-mean centered. The level 2 model predicted the county-level intercept |$\big({\beta}_{0j}\big)$|as a function of six county-level legal barriers to eviction, also described below. Whenever possible, covariates focused on those at risk of eviction: renter households. Unless specified otherwise, variables were drawn from ACS 5-Year estimates for 2012–2016 (U.S. Census Bureau 2016).
Neighborhood Characteristics
To test whether serial eviction filings were most common in mid-range markets, we included tract median household rent, recoded as a categorical variable to allow for nonlinearities across the distribution. Less-educated workers, particularly those in the service sector, are disproportionately at risk of income volatility (Bania and Leete 2009; Lambert et al. 2014). To assess the degree to which serial filings were associated with such volatility, we observed the percentage of renting households whose householder had a bachelor’s degree or more and the percentage of working adults in the tract employed in the service sector. Because the educational status or occupation of individual tenants facing eviction is unavailable, tract-level measures served as imperfect proxies.
We examined whether serial filings were less prevalent in gentrifying neighborhoods, where landlords may be incentivized to displace tenants (Mele 2000; Stabrowski 2014). We measured gentrification using data from the 2000 Census and the 2012–2016 ACS. We used principal components analysis (PCA), separately by period, to combine three tract-level variables: (1) the percentage of residents aged 25 and over with a bachelor’s degree or higher; (2) the percentage of residents employed in managerial, technical, or professional occupations; and (3) median household income. In both 2000 and 2012–2016, PCA showed that these variables loaded onto a single factor (eigenvalues of 2.59 in 2000 and 2.60 in 2012–2016). We assigned each tract within a Core Based Statistical Area (CBSA) its percentile in the period- and CBSA-specific distribution of this factor. Following Timberlake and Johns-Wolfe (2017), any tract in the bottom 60% of the distribution in 2000 was “gentrifiable.” Among that set, tracts were labeled “gentrified” if they did not experience population decline over 50% but did experience at least a 10% increase in the CBSA-specific distribution between periods. Across the 22,068 tracts in our sample, 11.20% gentrified since 2000.
We explored how serial eviction filing rates were associated with four additional neighborhood characteristics. First, larger rental markets, in which landlords can leverage economies of scale through bulk filings, may have more serial filings (Bogdon and Ling 1998). We examined the association between the natural logarithm of the total number of renting households and the serial eviction filing rate. Second, because landlords might be more reluctant to file for eviction in slack markets in which replacing displaced tenants may prove costly, we analyzed variations with the neighborhood’s rental vacancy rate (Garboden and Rosen 2019). Third, because eviction disproportionately affects nonwhite communities (Desmond and Shollenberger 2015), we assessed how serial eviction filing rates vary by tract racial majority: white, Black, Latinx, or other/none. Fourth, because women and households with children face a higher risk of eviction (Desmond 2012; Desmond et al. 2013), we analyzed whether neighborhoods with larger shares of female-headed renter households and renter households with children had higher serial filing rates.
Landlord Characteristics
We included two tract-level measures to evaluate the relationship between landlord characteristics and serial eviction filings. First, we observed the extent to which the concentration of corporate landlords was associated with serial eviction filing rates. Using names in eviction court records, we identified plaintiffs as either individuals or businesses.7 We analyzed how the share of evictions initiated by a corporate entity in each neighborhood varied with the serial eviction filing rate. Second, we examined the association between a neighborhood’s serial filing rate and its eviction filing rate—the number of eviction filings divided by the total number of renting households—to account for variations in landlord practices and the normalization of eviction filing.
Legal Characteristics
Little empirical evidence exists on the relationship between landlord-tenant law and eviction rates. Interviews with landlords suggested the role of six legal characteristics associated with landlords’ eviction filing behavior: (1) the cost of filing an eviction; (2) the number of courts hearing cases; (3) the process time required; (4) whether either party is required to hire attorneys, which increases the cost of eviction; (5) whether the court automatically schedules a hearing, which incurs time costs for landlords; and (6) whether landlords are required to notify tenants of late rent for a set number of days before they can file an eviction (see also Hatch 2017). We analyzed how each characteristic was associated with variations in serial eviction filing rates. To do so, we used the filing fee data described above. We also collected information on which courts hear eviction cases in each county.8 Using court records, we determined for each county the median number of days it took to resolve an eviction case. We used information published on government websites to code indicators of attorney requirements, automatic hearing scheduling policies, and notice requirements. Summary statistics of variables used in our analysis are presented in Table 1.
Summary Statistics
| Mean | SD | Minimum | Maximum | |
|---|---|---|---|---|
| Tract-level variables (n = 22,063) | ||||
| Serial filing rate | 19.2 | 18.6 | 0.0 | 100.0 |
| Number of households with serial filings | 11.9 | 31.2 | 0.0 | 2,005.0 |
| Neighborhood characteristics | ||||
| Median household rent ($) | 927.5 | 347.2 | 114 | 3,501 |
| % Renter householder with bachelor’s degree | 22.3 | 18.5 | 0.0 | 100.0 |
| % in service occupations | 18.8 | 7.8 | 0.0 | 100.0 |
| Gentrification indicator | 0.0 | 0.0 | 0.0 | 1.0 |
| Number of renting households | 621.6 | 473.3 | 0.0 | 4,729.0 |
| Rental vacancy rate | 6.3 | 7.3 | 0.0 | 92.9 |
| % white | 65.9 | 27.5 | 0.0 | 100.0 |
| % black | 16.0 | 22.3 | 0.0 | 100.0 |
| % hispanic | 11.7 | 15.8 | 0.0 | 98.9 |
| % other | 6.4 | 7.1 | 0.0 | 90.2 |
| % renter households with female head | 36.1 | 19.2 | 0.0 | 100.0 |
| % renter household with kids | 36.0 | 16.7 | 0.0 | 100.0 |
| Landlord characteristics | ||||
| % of plaintiffs that are businesses | 52.5 | 23.0 | 0.0 | 100.0 |
| Eviction filing rate | 7.2 | 13.6 | 0.1 | 1,415.5 |
| County-level variables (n = 958) | ||||
| Legal characteristics | ||||
| Filing fee ($) | 109.1 | 67.4 | 30 | 350 |
| Number of courts handling evictions | 1.5 | 1.4 | 0 | 16 |
| Median eviction processing time (days) | 25.4 | 19.7 | 3 | 288 |
| Attorney requirements | 0.4 | 0.5 | 0 | 1 |
| Automatic hearings | 0.8 | 0.4 | 0 | 1 |
| Notice requirements | 0.9 | 0.3 | 0 | 1 |
| Mean | SD | Minimum | Maximum | |
|---|---|---|---|---|
| Tract-level variables (n = 22,063) | ||||
| Serial filing rate | 19.2 | 18.6 | 0.0 | 100.0 |
| Number of households with serial filings | 11.9 | 31.2 | 0.0 | 2,005.0 |
| Neighborhood characteristics | ||||
| Median household rent ($) | 927.5 | 347.2 | 114 | 3,501 |
| % Renter householder with bachelor’s degree | 22.3 | 18.5 | 0.0 | 100.0 |
| % in service occupations | 18.8 | 7.8 | 0.0 | 100.0 |
| Gentrification indicator | 0.0 | 0.0 | 0.0 | 1.0 |
| Number of renting households | 621.6 | 473.3 | 0.0 | 4,729.0 |
| Rental vacancy rate | 6.3 | 7.3 | 0.0 | 92.9 |
| % white | 65.9 | 27.5 | 0.0 | 100.0 |
| % black | 16.0 | 22.3 | 0.0 | 100.0 |
| % hispanic | 11.7 | 15.8 | 0.0 | 98.9 |
| % other | 6.4 | 7.1 | 0.0 | 90.2 |
| % renter households with female head | 36.1 | 19.2 | 0.0 | 100.0 |
| % renter household with kids | 36.0 | 16.7 | 0.0 | 100.0 |
| Landlord characteristics | ||||
| % of plaintiffs that are businesses | 52.5 | 23.0 | 0.0 | 100.0 |
| Eviction filing rate | 7.2 | 13.6 | 0.1 | 1,415.5 |
| County-level variables (n = 958) | ||||
| Legal characteristics | ||||
| Filing fee ($) | 109.1 | 67.4 | 30 | 350 |
| Number of courts handling evictions | 1.5 | 1.4 | 0 | 16 |
| Median eviction processing time (days) | 25.4 | 19.7 | 3 | 288 |
| Attorney requirements | 0.4 | 0.5 | 0 | 1 |
| Automatic hearings | 0.8 | 0.4 | 0 | 1 |
| Notice requirements | 0.9 | 0.3 | 0 | 1 |
Summary Statistics
| Mean | SD | Minimum | Maximum | |
|---|---|---|---|---|
| Tract-level variables (n = 22,063) | ||||
| Serial filing rate | 19.2 | 18.6 | 0.0 | 100.0 |
| Number of households with serial filings | 11.9 | 31.2 | 0.0 | 2,005.0 |
| Neighborhood characteristics | ||||
| Median household rent ($) | 927.5 | 347.2 | 114 | 3,501 |
| % Renter householder with bachelor’s degree | 22.3 | 18.5 | 0.0 | 100.0 |
| % in service occupations | 18.8 | 7.8 | 0.0 | 100.0 |
| Gentrification indicator | 0.0 | 0.0 | 0.0 | 1.0 |
| Number of renting households | 621.6 | 473.3 | 0.0 | 4,729.0 |
| Rental vacancy rate | 6.3 | 7.3 | 0.0 | 92.9 |
| % white | 65.9 | 27.5 | 0.0 | 100.0 |
| % black | 16.0 | 22.3 | 0.0 | 100.0 |
| % hispanic | 11.7 | 15.8 | 0.0 | 98.9 |
| % other | 6.4 | 7.1 | 0.0 | 90.2 |
| % renter households with female head | 36.1 | 19.2 | 0.0 | 100.0 |
| % renter household with kids | 36.0 | 16.7 | 0.0 | 100.0 |
| Landlord characteristics | ||||
| % of plaintiffs that are businesses | 52.5 | 23.0 | 0.0 | 100.0 |
| Eviction filing rate | 7.2 | 13.6 | 0.1 | 1,415.5 |
| County-level variables (n = 958) | ||||
| Legal characteristics | ||||
| Filing fee ($) | 109.1 | 67.4 | 30 | 350 |
| Number of courts handling evictions | 1.5 | 1.4 | 0 | 16 |
| Median eviction processing time (days) | 25.4 | 19.7 | 3 | 288 |
| Attorney requirements | 0.4 | 0.5 | 0 | 1 |
| Automatic hearings | 0.8 | 0.4 | 0 | 1 |
| Notice requirements | 0.9 | 0.3 | 0 | 1 |
| Mean | SD | Minimum | Maximum | |
|---|---|---|---|---|
| Tract-level variables (n = 22,063) | ||||
| Serial filing rate | 19.2 | 18.6 | 0.0 | 100.0 |
| Number of households with serial filings | 11.9 | 31.2 | 0.0 | 2,005.0 |
| Neighborhood characteristics | ||||
| Median household rent ($) | 927.5 | 347.2 | 114 | 3,501 |
| % Renter householder with bachelor’s degree | 22.3 | 18.5 | 0.0 | 100.0 |
| % in service occupations | 18.8 | 7.8 | 0.0 | 100.0 |
| Gentrification indicator | 0.0 | 0.0 | 0.0 | 1.0 |
| Number of renting households | 621.6 | 473.3 | 0.0 | 4,729.0 |
| Rental vacancy rate | 6.3 | 7.3 | 0.0 | 92.9 |
| % white | 65.9 | 27.5 | 0.0 | 100.0 |
| % black | 16.0 | 22.3 | 0.0 | 100.0 |
| % hispanic | 11.7 | 15.8 | 0.0 | 98.9 |
| % other | 6.4 | 7.1 | 0.0 | 90.2 |
| % renter households with female head | 36.1 | 19.2 | 0.0 | 100.0 |
| % renter household with kids | 36.0 | 16.7 | 0.0 | 100.0 |
| Landlord characteristics | ||||
| % of plaintiffs that are businesses | 52.5 | 23.0 | 0.0 | 100.0 |
| Eviction filing rate | 7.2 | 13.6 | 0.1 | 1,415.5 |
| County-level variables (n = 958) | ||||
| Legal characteristics | ||||
| Filing fee ($) | 109.1 | 67.4 | 30 | 350 |
| Number of courts handling evictions | 1.5 | 1.4 | 0 | 16 |
| Median eviction processing time (days) | 25.4 | 19.7 | 3 | 288 |
| Attorney requirements | 0.4 | 0.5 | 0 | 1 |
| Automatic hearings | 0.8 | 0.4 | 0 | 1 |
| Notice requirements | 0.9 | 0.3 | 0 | 1 |
In-depth Interviews
To obtain a more detailed understanding of the dynamics of serial filings, the first author interviewed property managers and court officials in Charleston, South Carolina, and Mobile, Alabama in August and September of 2018. These sites were selected because—as displayed in Table 2—they had vastly different serial filing rates despite broad demographic similarities. In 2014 Charleston had a serial filing rate of 52.7%, compared to 5.9% in Mobile.
Demographic and Eviction Statistics for Charleston County, SC, and Mobile County, AL, 2014
| Charleston, SC | Mobile, AL | |
|---|---|---|
| Total population | 380,673 | 414,291 |
| Number of renting households | 59,679 | 52,495 |
| % renter households | 40% | 34% |
| % white renters | 56% | 43% |
| % black renters | 36% | 49% |
| Median renter household income (thousands) | 36,308 | 25,261 |
| Poverty rate | 11% | 16% |
| Eviction rate | 3.39% | 2.42% |
| Serial filing rate | 52.46% | 5.63% |
| Charleston, SC | Mobile, AL | |
|---|---|---|
| Total population | 380,673 | 414,291 |
| Number of renting households | 59,679 | 52,495 |
| % renter households | 40% | 34% |
| % white renters | 56% | 43% |
| % black renters | 36% | 49% |
| Median renter household income (thousands) | 36,308 | 25,261 |
| Poverty rate | 11% | 16% |
| Eviction rate | 3.39% | 2.42% |
| Serial filing rate | 52.46% | 5.63% |
Demographic and Eviction Statistics for Charleston County, SC, and Mobile County, AL, 2014
| Charleston, SC | Mobile, AL | |
|---|---|---|
| Total population | 380,673 | 414,291 |
| Number of renting households | 59,679 | 52,495 |
| % renter households | 40% | 34% |
| % white renters | 56% | 43% |
| % black renters | 36% | 49% |
| Median renter household income (thousands) | 36,308 | 25,261 |
| Poverty rate | 11% | 16% |
| Eviction rate | 3.39% | 2.42% |
| Serial filing rate | 52.46% | 5.63% |
| Charleston, SC | Mobile, AL | |
|---|---|---|
| Total population | 380,673 | 414,291 |
| Number of renting households | 59,679 | 52,495 |
| % renter households | 40% | 34% |
| % white renters | 56% | 43% |
| % black renters | 36% | 49% |
| Median renter household income (thousands) | 36,308 | 25,261 |
| Poverty rate | 11% | 16% |
| Eviction rate | 3.39% | 2.42% |
| Serial filing rate | 52.46% | 5.63% |
To select an initial sample for interviewing, we identified plaintiffs who appeared most frequently in court records. Because managers with a large number of units had higher odds of appearing repeatedly in the court records, almost all interviews were conducted with professional property managers. The average property manager with whom we spoke oversaw 229 units.9 We supplemented with a snowball sample, asking interviewees for referrals at the end of each interview. Lastly, we recruited attendees from a risk management course for property managers in Mobile. We interviewed 17 property owners or managers in Charleston and 16 in Mobile. Following an interview guide—available in the Online Appendix—the semi-structured interviews focused on their work experiences and responsibilities, with an emphasis on rent collection and evictions. Each interview lasted between 30 and 90 min; participants were compensated $50. All but two interviews with property managers were conducted at their offices.10 All interviews were recorded and later transcribed. With their consent, we use participants’ real names in this paper (Jerolmack and Murphy 2017).
In both states, we also spoke with legal experts and observed eviction proceedings at courts. In South Carolina, we visited three magistrate courts in North Charleston, where the majority of serial filings were processed. We spoke to two legal aid attorneys, a judge, and a court clerk at those courts. In Alabama, we visited the Mobile County District Court, where unlawful detainer hearings took place, and interviewed a judge, two landlord attorneys, and three legal aid attorneys. These interviews and observations improved our understanding of legal procedures and allowed us to fact-check statements made by landlords.
Quantitative Results
In 2014, 862,151 unique households across the 958 counties in our sample received an eviction filing. This represents an overall eviction filing rate (the ratio of eviction filings to renter households) of 7.9%.11 Of the 1,082,845 evictions filed in 2014, 44.6% were associated with serial filings. Among households at risk of eviction—those that received at least one eviction filing—30.4% were filed against multiple times at the same address. Using filing fee and median rent data, we estimated that the average cost of “paying to stay,” across our sample of tracts, was approximately $180 in late fees and court costs. This represented, on average, a 20.3% increase in monthly rent per eviction filing.12
Figure 1 displays the distributions of tract-level serial filing rates for the states in our sample, ordered from top to bottom by descending median tract-level serial filing rate. Delaware had the highest median rate (51.8%). The next four highest states—the Southeastern states of South Carolina, Virginia, North Carolina, and Georgia—had median tract-level serial filing rates above 30%. By contrast, the median rate fell below 10% in Florida, Washington, Utah, Alabama, West Virginia, Oklahoma, Mississippi, Kansas, Indiana, and Illinois.
Serial filings can be evaluated in terms of the number and timing of filings received by households. The median in our sample was two filings, but this masks variation across states. Figure 2 presents the statewide proportion of serial cases with three or more filings per household. The states are ordered from left to right by ascending serial filing rate. In Delaware, Georgia, Kentucky, Michigan, Mississippi, North Carolina, South Carolina, and Virginia, over 50% of households filed against serially were threatened with eviction three or more times.
Distribution of tract-level serial filing rates, by state.
Distribution of tract-level serial filing rates, by state.
Percentage of serially-evicted households with more than two filings, by state.
Percentage of serially-evicted households with more than two filings, by state.
In states with the highest serial filing rates, at-risk households were not only more likely to receive multiple eviction filings, they also received filings more frequently. In our sample, the median household facing serial eviction received a follow-up filing within 2.4 months, and 88.0% received subsequent eviction filings within a year of their first eviction filing. However, the median time between serial cases varied considerably across states (see Appendix Table A2). States with higher serial rates had lower median time between filings. In South Carolina and Delaware, approximately 1 in every 10 serially evicted households received ten or more eviction filings between 2010 and 2016. A nontrivial share of households that had been filed against—roughly one in twenty in South Carolina and one in thirty in Georgia—were filed against six or more times in a year.
In Table 3, we present results from regression analysis. Relative to neighborhoods with median rents between $1,201 and $1,400 per month—controlling for all other covariates—neighborhoods with lower rents had significantly lower serial eviction filing rates. These rates were also significantly lower in areas with median rents above $2,000. We illustrate this pattern in Figure 3, which plots the marginal effects from the model across the distribution of rents, holding all other covariates at their means.
Tracts with proportionately more college-educated renters had higher serial eviction filing rates, while those with more service sector workers had lower rates. Together, results support the hypothesis that serial eviction filings affect mid-range rental markets most. Renters in these areas who fall behind may have the means to prevent eviction by paying arrears, late fees, and court costs, whereas poorer renters may not be able to afford to stay.
All else equal, neighborhoods that gentrified since 2000 had significantly lower serial eviction filing rates. This conforms with our hypothesis that landlords in such neighborhoods may favor removing tenants rather than serially filing evictions for rent collection. Serial eviction filing rates were higher in neighborhoods with larger renter populations, which aligns with the claim that economies of scale may favor serial filings. The serial eviction filing rate was estimated to decline as a neighborhood’s rental vacancy rate increased. This suggests that the repeated threat of eviction is deployed to a greater extent in tighter markets, where vacated homes can be replaced more quickly. Compared to tracts without white, black, or Latinx racial majorities, majority-Latinx neighborhoods had lower serial eviction filing rates. The serial eviction filing rate was positively correlated with both the percentage of female-headed renter households and the percentage of renter households with children.
Two-Level, Varying-Intercept Negative Binomial Regression Estimates of Tract-Level Serial Filings
| Coefficient | Robust S.E. | |
|---|---|---|
| Neighborhood characteristics | ||
| Median household rent | ||
| $600 or less | −0.131* | (0.051) |
| $601–$800 | −0.193*** | (0.051) |
| $801–$1,000 | −0.113** | (0.035) |
| $1,001–$1,200 | −0.046* | (0.019) |
| $1,201–$1,400 | ref | |
| $1,401–$1,600 | −0.024 | (0.026) |
| $1,601–$1,800 | −0.051 | (0.040) |
| $1,801–$2,000 | −0.061 | (0.084) |
| $2,001 or more | −0.304*** | (0.085) |
| % renter householder with bachelor’s degree | 0.004*** | (0.001) |
| % in service occupations | −0.002*** | (0.001) |
| Gentrification indicator | −0.037** | (0.013) |
| Number of renting households (logged) | 0.098*** | (0.013) |
| Rental vacancy rate | −0.007*** | (0.001) |
| Neighborhood racial majority | ||
| Majority white | −0.017 | (0.012) |
| Majority black | 0.011 | (0.021) |
| Majority hispanic | −0.063** | (0.024) |
| No/other racial majority | ref | |
| % renter households with female head | 0.001** | (0.000) |
| % renter household with kids | 0.002*** | (0.000) |
| Landlord characteristics | ||
| % of plaintiffs that are businesses | 0.005*** | (0.001) |
| Eviction filing rate (logged) | 0.240*** | (0.029) |
| Legal barriers | ||
| Filing fee (log) | −0.127 | (0.132) |
| Number of courts handling evictions | 0.025 | (0.025) |
| Median eviction processing time | −0.004 | (0.004) |
| Attorney requirements | −0.322 | (0.204) |
| Automatic hearings | 0.019 | (0.177) |
| Notice requirements | −0.315* | (0.140) |
| Constant | −0.924 | (0.601) |
| Observations | 21,831 | |
| Number of groups | 957 | |
| Coefficient | Robust S.E. | |
|---|---|---|
| Neighborhood characteristics | ||
| Median household rent | ||
| $600 or less | −0.131* | (0.051) |
| $601–$800 | −0.193*** | (0.051) |
| $801–$1,000 | −0.113** | (0.035) |
| $1,001–$1,200 | −0.046* | (0.019) |
| $1,201–$1,400 | ref | |
| $1,401–$1,600 | −0.024 | (0.026) |
| $1,601–$1,800 | −0.051 | (0.040) |
| $1,801–$2,000 | −0.061 | (0.084) |
| $2,001 or more | −0.304*** | (0.085) |
| % renter householder with bachelor’s degree | 0.004*** | (0.001) |
| % in service occupations | −0.002*** | (0.001) |
| Gentrification indicator | −0.037** | (0.013) |
| Number of renting households (logged) | 0.098*** | (0.013) |
| Rental vacancy rate | −0.007*** | (0.001) |
| Neighborhood racial majority | ||
| Majority white | −0.017 | (0.012) |
| Majority black | 0.011 | (0.021) |
| Majority hispanic | −0.063** | (0.024) |
| No/other racial majority | ref | |
| % renter households with female head | 0.001** | (0.000) |
| % renter household with kids | 0.002*** | (0.000) |
| Landlord characteristics | ||
| % of plaintiffs that are businesses | 0.005*** | (0.001) |
| Eviction filing rate (logged) | 0.240*** | (0.029) |
| Legal barriers | ||
| Filing fee (log) | −0.127 | (0.132) |
| Number of courts handling evictions | 0.025 | (0.025) |
| Median eviction processing time | −0.004 | (0.004) |
| Attorney requirements | −0.322 | (0.204) |
| Automatic hearings | 0.019 | (0.177) |
| Notice requirements | −0.315* | (0.140) |
| Constant | −0.924 | (0.601) |
| Observations | 21,831 | |
| Number of groups | 957 | |
*** p< 0.001; **p < 0.01; *p < 0.05.
Two-Level, Varying-Intercept Negative Binomial Regression Estimates of Tract-Level Serial Filings
| Coefficient | Robust S.E. | |
|---|---|---|
| Neighborhood characteristics | ||
| Median household rent | ||
| $600 or less | −0.131* | (0.051) |
| $601–$800 | −0.193*** | (0.051) |
| $801–$1,000 | −0.113** | (0.035) |
| $1,001–$1,200 | −0.046* | (0.019) |
| $1,201–$1,400 | ref | |
| $1,401–$1,600 | −0.024 | (0.026) |
| $1,601–$1,800 | −0.051 | (0.040) |
| $1,801–$2,000 | −0.061 | (0.084) |
| $2,001 or more | −0.304*** | (0.085) |
| % renter householder with bachelor’s degree | 0.004*** | (0.001) |
| % in service occupations | −0.002*** | (0.001) |
| Gentrification indicator | −0.037** | (0.013) |
| Number of renting households (logged) | 0.098*** | (0.013) |
| Rental vacancy rate | −0.007*** | (0.001) |
| Neighborhood racial majority | ||
| Majority white | −0.017 | (0.012) |
| Majority black | 0.011 | (0.021) |
| Majority hispanic | −0.063** | (0.024) |
| No/other racial majority | ref | |
| % renter households with female head | 0.001** | (0.000) |
| % renter household with kids | 0.002*** | (0.000) |
| Landlord characteristics | ||
| % of plaintiffs that are businesses | 0.005*** | (0.001) |
| Eviction filing rate (logged) | 0.240*** | (0.029) |
| Legal barriers | ||
| Filing fee (log) | −0.127 | (0.132) |
| Number of courts handling evictions | 0.025 | (0.025) |
| Median eviction processing time | −0.004 | (0.004) |
| Attorney requirements | −0.322 | (0.204) |
| Automatic hearings | 0.019 | (0.177) |
| Notice requirements | −0.315* | (0.140) |
| Constant | −0.924 | (0.601) |
| Observations | 21,831 | |
| Number of groups | 957 | |
| Coefficient | Robust S.E. | |
|---|---|---|
| Neighborhood characteristics | ||
| Median household rent | ||
| $600 or less | −0.131* | (0.051) |
| $601–$800 | −0.193*** | (0.051) |
| $801–$1,000 | −0.113** | (0.035) |
| $1,001–$1,200 | −0.046* | (0.019) |
| $1,201–$1,400 | ref | |
| $1,401–$1,600 | −0.024 | (0.026) |
| $1,601–$1,800 | −0.051 | (0.040) |
| $1,801–$2,000 | −0.061 | (0.084) |
| $2,001 or more | −0.304*** | (0.085) |
| % renter householder with bachelor’s degree | 0.004*** | (0.001) |
| % in service occupations | −0.002*** | (0.001) |
| Gentrification indicator | −0.037** | (0.013) |
| Number of renting households (logged) | 0.098*** | (0.013) |
| Rental vacancy rate | −0.007*** | (0.001) |
| Neighborhood racial majority | ||
| Majority white | −0.017 | (0.012) |
| Majority black | 0.011 | (0.021) |
| Majority hispanic | −0.063** | (0.024) |
| No/other racial majority | ref | |
| % renter households with female head | 0.001** | (0.000) |
| % renter household with kids | 0.002*** | (0.000) |
| Landlord characteristics | ||
| % of plaintiffs that are businesses | 0.005*** | (0.001) |
| Eviction filing rate (logged) | 0.240*** | (0.029) |
| Legal barriers | ||
| Filing fee (log) | −0.127 | (0.132) |
| Number of courts handling evictions | 0.025 | (0.025) |
| Median eviction processing time | −0.004 | (0.004) |
| Attorney requirements | −0.322 | (0.204) |
| Automatic hearings | 0.019 | (0.177) |
| Notice requirements | −0.315* | (0.140) |
| Constant | −0.924 | (0.601) |
| Observations | 21,831 | |
| Number of groups | 957 | |
*** p< 0.001; **p < 0.01; *p < 0.05.
Serial filings may reflect a property-management approach more common among corporate landlords. Our model supports this hypothesis, finding that, all else held equal, neighborhoods with a larger percentage of businesses listed as plaintiffs had significantly more serial filings. We also found that areas where eviction has been more routinized—measured by higher eviction filing rates—had significantly higher serial filing rates.
Predicted serial filing rate by tract median rent.
Predicted serial filing rate by tract median rent.
Results indicated limited associations between the six legal characteristics and serial eviction filing rates. Only the presence of notice requirements was significantly (negatively) associated with serial eviction filings. In practice, however, these six legal characteristics do not exist independent of one another. Simplifying several variables, we created a measure tallying the number of legal barriers to eviction in each county: the higher the measure, the more difficult or costly it is for a landlord to formally evict a tenant.13 In Figure 4, we display county-level distributions of serial eviction filing rates by the number of legal barriers in place. We see a steep drop from zero to one barrier and consistent declines as the number of legal barriers increased. Counties with more barriers to eviction had lower serial eviction filing rates.
Distribution of county-level serial eviction filing rate by number of legal barriers to eviction.
Distribution of county-level serial eviction filing rate by number of legal barriers to eviction.
Our quantitative results analyzed the prevalence of and patterns in serial eviction filings across neighborhoods. To better understand how these factors played out at the level of landlords and tenants—and to gain greater insight into how property managers responded to legal environments—we turn to interview results.
Qualitative Results
The Property Manager’s Perspective
Rent collection was a central responsibility for the property managers with whom we spoke. Several used software to track each property’s monthly income and delinquency rates, metrics closely monitored by upper management. Shona Littlejohn, a property manager at Pinecrest Apartments in Charleston, described the pressure she faced.
As a manager, my reports are due every Monday. So, I actually work week by week to get those done because I’m only as good as my week is. I have to make sure my numbers are always on the incline and never on the decline.
Many property managers felt that they were at the mercy of their spreadsheets. Adherence to corporate policies, including on rent collection and eviction, left little room to negotiate with tenants. Shay Lawrence, who managed 120 units at Ashford Place Apartments in Mobile, had to file on the sixteenth of each month: “If it’s not paid on that date, we have to process for eviction. That’s our rules from compliance. … That’s from head office.” All but one property manager in South Carolina provided a cutoff date in each month when they would begin eviction proceedings. Meeting corporate targets was essential to keeping their jobs. As Littlejohn put it, “I love staying in the black, never in the red, in my delinquency and so forth, and that’s my job security.” Property managers who displayed good performance also received perks. Three in Charleston disclosed that they received bonuses based on their properties’ financial performances.
Property managers were familiar with legal procedures for filing evictions. In Alabama, we observed a course offered by the Mobile Realtors Association, with 17 property managers attending. A landlord attorney gave advice on risk management, including on filing evictions and avoiding Fair Housing Act (FHA) violations. She gave out templates of notices to vacate and promoted additional templates sold by her law firm. These forms ensured that managers adhered to legal requirements so that the case would not be dismissed on procedural grounds.
This is only one example of the many supplementary business offerings that facilitated evictions, including rent-management software and automated eviction services. Several of the property managers we interviewed used Nationwide Evictions, a service claiming to operate in over thirty states and process roughly 180,000 eviction cases a year (Nationwide Evictions n.d.). To begin the eviction process, Crystal Brown, a property manager in Charleston, needed only to input tenant information, lease, and rent ledger. “And then we just send it off to Nationwide Evictions,” Brown said. “What they do on their end, I’m not sure. I know it gets to the courthouse. I know it gets filed. I know the residents get something on their door.”
Property managers with whom we spoke in Charleston and Mobile shared many similarities. They managed properties of similar scale, dealt with oversight, and pressure from owners and corporate headquarters, and benefitted from industry support. Nonetheless, serial filings were much more common in Charleston than Mobile. Comparing legal systems is key to understanding and explaining differences in eviction filing patterns between the two sites. We therefore turn to understanding how property managers in each city interpreted and responded to their legal environments.
South Carolina: Filing for Rent
In South Carolina, property managers were primed to file for eviction upon nonpayment of rent and regularly did so. Property managers reported that they did not always file for eviction to reclaim property but to extract rent and additional fees. For these property managers, filing an eviction was easy, cheap, and part of the rent collection process. The state’s landlord-tenant law allowed the lease to include language that made the lease itself the notice of late rent and eviction, instead of the five- or seven-day eviction notices required in other states (South Carolina Code of Laws Section 2020). This saved landlords the time spent serving notices, allowing them to file as early as the fifth of the month. It cost $40 to file an eviction, less than half the national average of $112.14 As there was no attorney requirement in the state, property managers could represent themselves in court. This would be a time-consuming process if tenants requested hearings, but the vast majority did not. Of the forty evictions filed in July 2018 by Jennifer Williams at Bridgeview Apartments, only two households requested court hearings. In combination, these factors dramatically lowered the barriers to filing an eviction.
Managers in South Carolina expected most tenants would “pay to stay” once an eviction was filed. Courtney Elliot, property manager at Palmetto Creek, observed that tenants “generally pay even after [an eviction has been filed]. We just don’t move forward with the eviction process.” As long as tenants paid in full, including late fees and filing costs, they could remain in their homes. The manager thereby maintained the tenants, transferred court costs, and gained additional income from late fees. These fees typically were applied after the fifth of the month (regardless of eviction filing) and ranged from $50 to $100, plus up to $10 each additional day the rent went unpaid.
Misalignment between strict rent due dates and pay schedules also contributed to patterns of serial eviction filings. Following company policy, Jennifer Williams at Bridgeview Apartments filed evictions on the eleventh of each month for all tenants behind on rent. In July 2018, Williams filed against 40 of the 300 units in her complex. Thirty-eight of those households paid in full by the end of the month. “Because they’re paid usually biweekly,” she explained, “they’ll pay half their rent with one check and then pay the other half with the other check, and [so] by the time they’re getting that second check, it’s after the eleventh of the month.”
Unexpected expenses compromised tenants’ ability to make rent. Millie Moore, property manager at Oak Ridge Townhouses, explained that missed rent payments were often due to “death in the family, unexpected car repair, school expense[s]… When you live paycheck to paycheck, it’s not as easy to absorb an unexpected expense.” Almost all property managers we spoke with only accepted payments in full, filing for eviction when their tenants—who could often make rent by the end of the month—initially fell short. By paying to stay, tenants reinforced the effectiveness of filing evictions as a strategy for rent collection. If tenants were unable to pay in full, property managers had already initiated the eviction process, thereby reducing the time needed to remove the tenant.
Filing evictions early and frequently also sent a strong signal to tenants. Elliot, property manager at Palmetto Creek, saw an eviction filing as a disciplinary tool. She instructed her team to “file evictions on the eleventh of every month for any and all accounts that have not been satisfied, regardless if we have heard from them or not.” Moore, property manager at Oak Ridge Townhouses, echoed Elliot’s sentiment: “Basically, my experience as I’ve been here, [an eviction filing] gets everybody’s attention.”
Property managers also claimed that they were anxious to avoid being accused of FHA violations, involving discriminating against protected classes of people. If they were to extend leniency in one case, managers told us, they could be accused of discriminating. As such, property managers filed against all late-paying tenants on the same day, regardless of their circumstances. This caution extended to blanket policies on tenant screening and rent collection procedures, including fees charged. Heidi Townsend, manager at Wedgewood Apartments and Chester Apartments, explained:
You pretty much have to enforce the same late fee, or the same eviction fee, the same procedure with every resident across the board so that they cannot come back and say, “Well you didn’t charge my neighbor the late fee, but you charged me the late fee, and I feel like I’m discriminated against.”
Victoria Cowart, Vice President of Property Management at Darby Development Co. Inc., concurred.
We are concerned that if we do not file that will set legal precedent that says we have permitted the tardy payment. We’re very concerned by fair housing laws and civil rights laws that we not treat customer different from another. We will file the ejectment, but it is not tantamount to us saying we want you to leave. We hope that they do not leave.
The equal-treatment mandates of the FHA were repeatedly invoked to explain limited property manager discretion, including toward leniency. The same law that upheld equal access to housing also upheld equal access to eviction. In so doing, the FHA may have contributed to increasing tenants’ exposure to costs associated with the eviction process.
Alabama: Filing for Possession
Legal barriers to eviction were much higher in Alabama. Before initiating eviction action in court, landlords were required to serve the tenant with a seven-day written notice (Code of Alabama 1975 § 35-9A). Filing an eviction cost $256 in Mobile, the nation’s third-highest filing fee. In Alabama, only individual owners listed on a property’s deed could initiate eviction actions without an attorney. When a corporation was listed as owner, property managers were required to hire attorneys, incurring additional costs (Code of Alabama 1975 § 35-9A). According to Amber Stroud, assistant property manager at Montlimar Apartments, an eviction could drag on for “five or six months without us collecting rent.” She continued: “Then we’re having to pay the attorney on top of it, so that’s another month of rent, basically.”
Given these constraints, filing an eviction was a serious commitment made only when property managers sought to take possession of a unit. Just as in South Carolina, should tenants facing eviction wish to remain in place, they were required to pay back rent, late fees, court costs, and (often) attorney fees. In Alabama, higher court costs and attorney fees made it much harder for tenants to pay and stay. Amber Harrington at Bel Air Apartments observed that only “two people reinstate[d] in the year that I’ve been here. So, no, it’s not very common. … ‘Cause if you can’t pay your rent then, and then you have to pay probably two months’ rent plus all of the lawyer’s fees [to stop the eviction]: it’s a lot to come up with.” Most property managers that we interviewed in Alabama did not expect tenants to reinstate once an eviction process had been initiated. “You kind of don’t want to take a little bit of money and start the process back over because it’s just going to go on and on,” explained Caroline Ivory at Azalea Pointe Apartments. “If they have a good bit of [money], I’m going to try and work with them on the rest.”
In Mobile, property managers relied on other strategies for rent collection before taking eviction actions. They emphasized the importance of communication. At Turtle Creek Apartments, the managers, Amanda Odle, and Sandy Foley, spent most of their time on rent collection. On the fifteenth of the month, they sent out letters urging tenants to pay rent or make special arrangements. They made phone calls, sent notes, and knocked on doors. Candis Blackmon, assistant property manager at Southern Oaks Apartments, described a similar process: “I call your phone. I call your emergency contact. I call your supervisor. I email. I send out letters. I basically hound until I hear from that person.” Her strategy appeared to work: of the 224 units she managed, only one tenant was unable to pay rent in the month we spoke with her. That tenant moved out voluntarily to avoid eviction.
In some communities, payment arrangements were part of the property management policy. Bel Air Apartments accepted rent until the twentieth of the month and allowed the tenant to extend the deadline once per lease term. Harrington explained, “I will pretty much do a payment arrangement for anybody that asks for one if they haven’t had one already because … it’s hard to determine somebody else’s financial struggles.” Similarly, Blackmon tried to work with her tenants. “I keep up with whenever someone says they’re going to promise to pay, and I calculate how much their late fees are going to be and give them that amount. And as long as they pay that amount on the date that they say they’re going to pay, then they can avoid going to eviction.” Property managers still collected late fees from tenants behind on rent but offered solutions that helped those tenants avoid incurring court costs and attorney fees, as well as an eviction record.
Payment arrangements were not unique to Alabama. We observed similar strategies in South Carolina. However, in South Carolina property managers reported that they would continue the eviction filing process even when they had arranged payment plans with tenants. As Shona Littlejohn described:
We do do payment arrangements, but we still have to go through the process, filing the eviction and all that. If somebody comes in and says, “Hey, I’m gonna be late with my rent for September. I will be able to pay it on the sixteenth.” I will tell that resident that’s absolutely fine … but you still will have to incur those charges such as the late fee and eviction costs.
Several property managers in Alabama told us that when they recognized a tenant could not pay the rent, they sometimes advised the tenant to leave before the eviction process began. That way, tenants could avoid an eviction record. As Paige Collins, a property manager in Mobile, said:
I try to let everyone know in advance. If you get evicted, it’s terrible for your credit. It’s going to be really hard to get an apartment anywhere else because we kind of all abide by the same criteria. So, the best thing to do … if you have an issue, [is] to just move out.
While this spared property managers from the cost and time associated with the formal eviction process, it did not necessarily spare tenants from arrears and late fees. To collect from former tenants, landlords often hired collection services, which could affect tenants’ credit history or even lead to wage garnishment (Lieberman 2017; Spector 2007).
Discussion
Drawing on millions of eviction records filed in 28 states between 2010 and 2016, this study documented serial eviction filing, a process involving tenants being threatened with displacement multiple times from the same address. Nearly one-third of households facing eviction in 2014 were filed against repeatedly. In these cases, eviction was not a single event but a repetitive and routinized process. Understanding serial filings allow us to appreciate eviction as a property management strategy facilitated by the legal system.
We documented significant within- and between-state variation in serial filing rates. Serial filing was characteristic of mid-range rental markets: areas with tenants well-off enough to catch up on rent and avoid displacement but not financially secure enough to consistently pay on the first of the month. Previous research has characterized eviction as a condition affecting poor, severely rent-burdened tenants (Desmond 2016; Lundberg and Donnelly 2019). Our results show that the threat of eviction extends beyond the bottom of the rental market. Researchers must consider how a broader population of tenants may be affected by previously unobserved forms of housing insecurity.
Eviction is not simply a product of the housing crisis or persistent poverty but also the result of different ownership structures and management strategies. Urban sociologists have only recently returned their focus to property owners and landlords (Desmond and Wilmers 2019; Rosen 2014), a critical topic in the scholarship of a previous generation (Cherry and Ford 1975; Logan and Molotch 1987). As segments of the housing market consolidate under corporate ownership, new tools, and methods have been devised to extract profit from rental investments (Immergluck 2013; Mallach 2007). We found that serial filings were concentrated in neighborhoods where corporate landlords initiate a larger share of eviction filings. Eviction should not exclusively be viewed as a market failure, the unfortunate result of a contract breach between landlords and tenants. The threat of eviction is also commonly employed as a strategy to both collect rent and increase revenue.
Interviews with property managers illuminated the dynamics driving serial filings. In both states, strict rent collection policies created and reified the problem of tenant lateness, which ultimately enhanced the profitability of rental housing. In South Carolina, when tenants were late, property managers used repeated eviction filings more for rent and fee collection than for displacement. In Alabama, where the cost of executing an eviction was considerably higher, eviction served instead as a means of recovering possession of a unit. Whereas eviction was among the first moves in South Carolina, it was among the last in Alabama.
In these interactions between owners and tenants, civil courts are not neutral arbiters. We found market actors to be responsive to lax regulation and more than willing to use the courts to collect rent and fees. Courts with low barriers to eviction are frequently contracted by property owners to manage and discipline tenants. In this way, those courts act more like an extension of the residential rental business than an impartial arbitrator between landlords and tenants. While this study focuses on the property management industry, future research should extend to segments of the rental market occupied by other types of landlords.
We found suggestive evidence that the FHA may create unexpected motivation for property managers to file in bulk and repeatedly against the same tenants. Property managers cited equal treatment and non-discrimination as grounds for policies to file for eviction against all past-due tenants. Previous studies have found stricter housing regulations to result in landlords’ heightening screening, more aggressive monitoring of tenants, and increasing rents (Greif 2018; Ambrose and Diop 2018). Policies intended to integrate public housing have been found to have adverse, exclusionary effects (Ganapati and Frank 2008; Simon 1991). Our study documents another possible unintended consequence of housing policies and calls for further research on the relationship between the FHA and eviction.
Housing loss is not the sole consequence of eviction filing. Even when not removed from their homes, tenants still suffered consequences from serial filings. Eviction records have durable consequences on the lives of tenants (Desmond et al. 2015), and repeated filings create barriers to future mobility. Nearly every landlord we interviewed claimed to categorically reject applicants with any negative rental history, be it money owed or an eviction filing, even those that did not culminate in an eviction judgement. According to Courtney Elliot, the manager at Palmetto Creek, “a bunch of eviction filings [is] a big red flag” when reviewing applicants. Tenants also faced financial consequences from eviction filings which, by definition, accrued to a population already struggling to pay rent. For a renting household in a typical tract, fines and fees associated with serial filings cost approximately $180 each time they were threatened with eviction, effectively raising their housing costs by 20% per month—housing costs that scholars and policymakers have previously overlooked. No national database with information on late fees and legal fees exists. Our estimates and interviews may not have fully captured these costs and are thus likely conservative. Future studies should investigate the full financial costs associated with eviction filings. Serial filings likely affected tenants’ credit ratings, ability to pay other bills, and deferred expenditures, consequences future research should more fully enumerate.
The findings of this study hold several legislative and policy implications. First, they provide evidence that municipalities could potentially lower serial filing rates by raising the barriers to eviction filing, either through higher filing fees or requirements that lengthen the process. Our interviews with property managers in Alabama suggest that raising legal barriers to eviction could incentivize more flexibility and problem-solving between landlords and tenants. Future research should attempt to exploit geographic variation to assess the effects of county-level policies on serial eviction filing rates. Second, this study found that property managers leverage court systems to collect additional fees, which can deepen renters’ financial precarity. Policymakers can regulate these fees. For example, Washington, DC, limits late fees to 5 percent of monthly rents (D.C. Law 2016). In Massachusetts, landlords cannot impose a late fee until rent is 30 days late (Massachusetts General Laws 2017). Third, more thought should be given to revising rent payment practices to align with income volatility and pay schedules. Interviews revealed that mismatches between rent deadlines and pay schedules often led to unnecessary eviction filings. First-of-the-month rent payment requirements are burdensome to workers paid on a bimonthly or weekly basis. Rent collection practices should be more sensitive to the reality that monthly pay schedules are, for the vast majority of American workers, a relic of the past.
About the Authors
Lillian Leung is a PhD student in Sociology and Social Policy at Princeton University. She is interested in applying mixed methods to understand urban poverty and inequality in America. This is her first publication.
Peter Hepburn is an Assistant Professor of Sociology at Rutgers University—Newark. His research examines how changes to three core social institutions—work, criminal justice, and housing—serve to produce and perpetuate inequalities. His recent work has appeared in Demography, Social Problems, and the Journal of Marriage and Family.
Matthew Desmond is the Maurice P. During Professor of Sociology at Princeton University and Principal Investigator of the Eviction Lab.
Endnotes
An eviction filing is when a landlord initiates court action against their tenant. Not all filings lead to an eviction, which occurs when households are removed from their homes. Households threatened with eviction are those that have received at least one eviction filing. Households are marked as receiving serial eviction filings if they have received more than one eviction filing at the same address.
In 2002, the top 50 residential rental firms made 22.3% of all rental revenue (U.S. Census Bureau 2002); their share rose to 31.5% 10 years later (U.S. Census Bureau 2012)
Differences between tracts in our sample and all tracts nationwide were small but statistically significant. Appendix Table A1 provides a complete summary. Appendix Table A2 includes the percentage of tracts in each state in our sample. Appendix Figure A1 maps counties included in the sample.
Validation was necessary because of potential low coverage, which may result from several causes. Individual-level eviction records are sealed or barred from public release in a number of counties and states. In areas in which records go un-sealed, bulk and online data collection was not always available. In such cases, data collection companies sent individuals to collect records in-person at court. This process resulted in less-than-comprehensive collection in some areas and periods. Publicly available data sources used in validation include individual-level court records and aggregate court statistics
The Eviction Lab established a protocol using the Levenshtein distance, a measure of edits required for two strings to match, to compare defendant names sharing the same street address within and between case numbers. A threshold of two or fewer edits, depending on the fields matched, were used to determine whether two versions of defendant names match (Desmond et al. 2018).
For this analysis, we are missing filing fee information for 7 out of 956 counties. In these cases, we use the state’s average filing fee.
We calculated the percentage of unique eviction cases in the tract in 2014 in which the plaintiff name included one of the following terms: Acres, Annex, Apartment (or Apt), Associated, Associates, Association, Bank, Capital, Community, Company, Condo Trust, Cooperative, Corporation, Crossing, Development, Enterprise, Estate, Foundation, Holdings, Housing, Housing Authority, Housing Board, Incorporated, Investment, Leasing, Limited (or Ltd), Living Trust, LLC, LP, Management (or Mgmt), MHP, Mobile Home, National, Nominee, Partner, Partnership, Property, Real Estate, Realtor, Realty, Redevelopment, Rental, Residential, Revocable, The, Townhome, Townhouses, Trust, Trustee, and Village.
This information comes from two sources: BRB Publications, LLC’s Public Record Research System (2018), and documentation accompanying court records received from LexisNexis Risk Solutions.
Each property manager reported the number of units they managed. We summed these counts and divided the total by the number of property manager interviewed.
Interviews took place at settings of the participants’ choosing. Among the property managers’ offices where interviews took place, 23 were located on the properties and 8 were at the main offices of the property management companies. Two interviewees requested to meet in other public locations.
Not all eviction filings result in eviction judgments. The eviction rate is typically about one-third of the eviction filing rate. For example, Desmond et al. (2018) estimates a nationwide eviction filing rate of 6.6% and an eviction rate of 2.5% in 2014.
Averages are weighted by the number of households serially filed against in 2014.
We produced dummy variables indicating a high filing fee ($200 or above; 90th percentile within our sample of counties); whether a single court heard all eviction cases within the county; and whether evictions in the county take 21 days or more to process (the median across the counties in our sample). We added these to the three dummy variables for attorney requirements, automatic hearings, and notice requirements.
We collected filing fees across the United States, as described above. The average is unweighted across all counties in the dataset.
