LONGER TRIPS TO COURT CAUSE EVICTIONS

“Controlling for census tract characteristics and residence type, we find that excess commuting time increases default rates.”

DAVID HOFFMAN, William A. Schnader Professor of Law and Deputy Dean

In the first controlled study of eviction rates across time in a large urban center, Hoffman and colleague Anton Strezhnev of the University of Chicago found that Philadelphia tenants who live further away from the city’s courthouse and rely on mass public transit are more likely to fail to show up, leading to eviction by default. In “Longer Trips to Court Cause Evictions,” Hoffman and Strezhnev report their findings that excess commuting time increases default rates.

The study reviewed nearly 235,000 evictions filed against approximately 300,000 Philadelphians from 2005 through 2021. Using datasets obtained through the non-profit Philadelphia Legal Assistance (PLA), the Pew Charitable Trusts, and elsewhere, Hoffman and Strezhnev found that 40% of tenants in eviction proceedings during that period lost due to default.

Like many cities across the nation, Philadelphia continues to suffer from an ongoing eviction crisis; Hoffman and Strezhnev write that because commuting time is plausibly unrelated to other causes of eviction, policymakers may consider it as an instrument to better identify how eviction causes downstream social problems and prevents disadvantaged people from flourishing.

Preliminary Considerations and Findings

Hoffman and Strezhnev found it “surprising that although policymakers describe defaults as a part of the eviction crisis [in Philadelphia], we lack information about their incidence across jurisdictions.” Accordingly, the authors set out to explore questions such as “How many defaults are there, really? Do they lead to evictions? Who fails to show up, and why? And given the shock of Zoom justice wrought by COVID-19 in eviction court, did making justice remotely accessible matter to outcomes?”

According to the calculations of Hoffman and Strezhnev, for every 10 minutes in additional commuting time, tenants are between .65 and 1.4 percentage points more likely to default. A one-hour increase in commuting time has a 3.9 to 8.6 percentage point average effect on the probability of tenant default. Were all tenants to be able to get to their hearing in 10 minutes or less, Philadelphia would have eliminated approximately 4,000 to 9,000 such evictions due to default during that time.

“Controlling for census tract characteristics and residence type,” write Hoffman and Strezhnev, “we find that excess commuting time increases default rates. This effect holds when comparing properties owned by the same landlord, when controlling for direct distance to the courthouse, and even when controlling for commuting time measured during the weekend.”

In contrast, when tenants were offered Zoom or virtual hearings during the COVID-19 pandemic, the commuting time effect disappeared. The result is also absent for tenants in public housing, whose eviction processes are laxer and are not built around defaulting those who show up to court late.

Moreover, Hoffman and Strezhnev also found that “petitions to reopen defaults are rarely filed and infrequently granted.”

Materials and Methods

From PLA, the authors obtained 339,172 eviction documents involving residential properties from January 2005 through July 2021. Hoffman and Strezhnev filtered the dataset to include only properties from which they could “unambiguously parse and address number and street name from the text of the listed address and obtain, from Google Maps API, a correctly matching address with a latitude and longitude.”

For demographic covariates of concern — namely, the socioeconomic characteristics of different Philadelphia neighborhoods — the authors gathered median income and median contract rent from the 2015 American Community Survey; they used 2010 census block-level data for racial demographics.

Landlord data presented a unique problem that required a creative solution. “Because the docket often only lists the filing LLC as the plaintiff, and since individual landlords may own multiple properties through different LLCs, we would be unable to identify common landlords across eviction proceedings without additional data.” Accordingly, to obtain landlord data, Hoffman and Strezhnev used a novel database of Philadelphia landlords obtained through an agreement with the Pew Charitable Trust to match roughly 55,000 landlords to 136,000 rental properties.

The authors’ primary independent variable of interest was, of course, the commuting time to the Municipal Courthouse. For this determination, they queried the Google Maps Distance Matrix API to determine the estimated distance and travel time between each building in the dataset and the Philadelphia Municipal Courthouse. They measured the time and distance to the courthouse using public transit on a weekday (when hearings are scheduled) and on the weekend.

Upon extracting the outcome of the proceedings from the docket, Hoffman and Strezhnev landed on a dataset of 223,840 eviction proceedings across 61,104 unique buildings and 283,812 unique named defendants. The plurality of judgments were in favor of the landlord and nearly all were defaults, with the second most common outcome a settlement between the parties, and the third most common withdrawal of the case by the landlord.

In non-public housing (non-PHA) cases, default rates varied over time and space with the earliest period of 2005 to 2010 having the highest rate of 40%. The data showed a steady decline in that

number over time, though, which coincided with the imposition of new landlord regulations that increased the costs of filing frivolous evictions. In public housing cases, default judgments accounted for about 20 to 25% of evictions and showed no similar pattern of decline. “Because those evictions are so distinctive,” wrote Hoffman and Strezhnev, “for our primary analysis, we focus on non-PHA cases.”

In addition to the main text, the authors provide an impressive amount of supplementary information that includes discussions and analyses of data pre-processing, sensitivity analysis, the frequency of reopening default judgments, weekday vs. weekend commuting times, effects of commuting time on judgments by agreement and complaint withdrawals, absence of seasonality in treatment effects, and regression tables for the main text results.

Finally, the authors replicated their findings using a dataset of over 800,000 evictions from Harris County, Texas. They used data accessed through the Eviction Lab and “found results extremely similar to Philadelphia, despite the radically different jurisdictions.” Since mass transit is largely unavailable in that location, the authors focused on the relationship between driving time (to the local justice of the peace office) and the likelihood of eviction. They found that a 10-minute increase in driving time resulted in an increase in the likelihood of default by 3%. Even when comparing evictions taking place in the same building, in the same month, but which are assigned to different courthouses, defaults are more likely in the courthouse that is further away.

In conclusion, Hoffman and Strezhnev find that their results “indicate that policymakers should consider the distributive effects of rules which forfeit legal rights conditional on showing up to the courthouse at a particular time.” They suggest that alternatives “from remote hearings, to easy rescheduling, to no-excuse reopening” are not only available but would also “reduce the incidence of this pathologic practice.”

Turning to the legal academy, the authors encourage scholars to consider whether other legal proceedings are similarly impacted by transit. “Essentially, we highlight the role of physical place in producing access to justice. And our results may offer a novel and better identified tool to study the downstream effects of evictions.”