(AGENPARL) – WASHINGTON mar 21 giugno 2022

June 21, 2022

How Effective Were National, State, and Local Eviction Moratoria?

Adithya Raajkumar

Introduction

Evictions can have detrimental effects on tenants and their households, such as unemployment, loss of shelter, and/or food insecurity, which can often be especially severe since households that face eviction tend to be poorer and less job-secure (Desmond and Kimbro, 2015). Even when tenants are ultimately not evicted, eviction filings themselves can represent a burden for households (Banerjee and Ghatak, 2003; Leung, Hepburn, and Desmond, 2021). Finally, the mere threat of eviction can impose substantial penalties on households (Garboden and Rosen, 2019). Beyond the household level, evictions can also have adverse effects on public health (Benfer et al., 2021; Jowers et al., 2021).

During the COVID-19 pandemic, many renters lost their primary source of income and were unable to afford their rent payments (Parrot and Zandi, 2021). To assist these tenants, Congress and the Centers for Disease Control and Prevention (CDC) implemented measures to prevent them from being evicted. In particular, the CARES Act prevented landlords of federally financed properties from filing eviction proceedings, and the CDC’s order extended and strengthened the provisions of the CARES Act after it expired (McCarty and Perl, 2021). In addition, many states and municipalities enacted their own measures halting evictions, which were of varying duration (Figure 1).

Figure 1. Temporal Coverage of Eviction Moratoria

Although the strength and enforcement of these moratoria varied across jurisdictions, their staggered implementation and expiration provides an opportunity to study their effectiveness. For example, An, Gabriel, and Tzur-Ilan (2021) find that each additional week of a statewide eviction moratorium was associated with roughly 130 fewer eviction filings, and analysis by researchers at the Princeton Eviction Lab suggests that the CDC order alone may have prevented more than 1.5 million eviction filings nationwide during the 11 months it was in force (Rangel et al., 2021). However, there has been relatively little work assessing how municipal moratoria impacted eviction filings, and minimal, if any research comparing national, state, and local moratoria. Thus, in this note I assess the effectiveness of municipal eviction moratoria and compare the effectiveness of various national, state, and municipal moratoria in preventing eviction filings.

Eviction Filings in U.S. Cities

I gather data from the Princeton Eviction Lab’s Eviction Tracking System on weekly eviction filings between December 2019 and January 2022 for 31 cities² and additional data on state and local eviction moratoria from the lab’s COVID-19 Housing Policy Scorecard. I also add information on the CARES Act and CDC moratoria from the Congressional Research Service. Figure 2 displays the rate of new filings per 100,000 renting households for two fixed groups of cities classified by the duration of their eviction moratoria: the first group (“short”) let their moratoria expire very early in 2020, whereas the second group (“long”) extended their moratoria well into 2021. The figure also displays vertical lines denoting the coverage of the CARES Act moratorium and CDC order. While filings drop for both groups at the start of the pandemic (when both groups had eviction moratoria in place), they remain at relatively low levels for the “long” group, whereas they quickly begin to return to pre-pandemic levels for the “short” group, suggesting that local moratoria played a role in preventing eviction filings for the cities in the former. Moreover, national moratoria appear to impact eviction filings as well: for example, there is a decrease in filings when the CDC order is implemented (after the first blue line), and an increase when it expires (after the second blue line).

Figure 2. Eviction Filings over Time by Moratorium Duration

However, while this figure is useful for understanding the evolution of filings over time, it does not take into account conditions in each city’s particular rental policy landscape. For example, some cities may have weaker tenant protections, which both increases the rate of eviction filings and reduces the likelihood that these cities enact their own eviction moratoria. Thus, the effects of enacting moratoria based on comparisons such as in Figure 1 may be overstated.

To account for such confounders, I construct a measure of “baseline” eviction filings by calculating the historical average filings for each week of the year for each city (across the years 2015 to 2019). I then calculate the ratio of eviction filings in 2020 and beyond relative to this baseline. Figure 3 plots this ratio (totaled over the entire sample period) for the cities in our sample against the percent of the sample period for which each city had an eviction moratorium in effect.

Figure 3. Eviction Filings vs. Moratorium Length by City

The figure shows a clear negative association between moratorium length and the ratio of total filings to their pre-pandemic baselines. The blue dotted lines denote the start (September 4, 2020) and end (August 26, 2021) of the CDC moratorium. In other words, cities that maintained their eviction moratoria for longer saw greater reductions in eviction filings relative to pre-pandemic levels. Conversely, cities with shorter or no moratoria have eviction filing levels much closer to their pre-pandemic levels. Thus, even when controlling for historical eviction filing patterns, there is strong evidence that municipal eviction moratoria helped prevent eviction filings.

The Impact of Eviction Moratoria on Eviction Filings

For a more accurate assessment of the impact of various eviction moratoria on eviction filings, I regress outcomes on a series of indicators for municipal, state, and national-level eviction moratoria. Because conditions after a moratorium expires may not be identical to conditions before it is implemented⁴, I also include indicators for the expiration of moratoria. In addition, I include a dummy for the onset of the pandemic (i.e. date beyond March 15th, 2020) to better distinguish the effects of moratoria being in place from the effects of the pandemic itself. The regressions I estimate are thus of the basic form⁵:

(1) $$y_{it} = alpha_i + beta_1{LOCAL}_{it} + theta_1{POSTLOCAL}_{it} + beta_2{STATE}_{it} + theta_2{POSTSTATE}_{it} + $$

$$beta_2{CARES}_t + theta_3{POSTCARES}_t + beta_4{CDC}_t + theta_4{POSTCDC}_t + X_{it}^{prime}gamma + {Year}_t + {Month}_t + varepsilon_{it}$$.

For example, “LOCAL” equals one when a local moratorium is in force, and zero otherwise, whereas “POSTLOCAL” equals one only when a previous local moratorium has expired, and zero otherwise (including if no such moratorium was ever enacted). Thus, since all coefficients represent changes relative to the period before a moratorium was enacted, one can estimate the change in the outcome when moratoria are enacted from $$beta_1 ldots beta_4$$ in (1)⁶.

For eviction filings, I use two related outcomes: (1) new filings per week per 100,000 renting households, which allows us to estimate the (population-adjusted) marginal effect of extending a moratorium as opposed to letting it expire; and (2) filings per week as a percent of the pre-pandemic baseline (computed for each week of the year, averaged over 2015 to 2019), which reveals how filings evolved relative to the pre-pandemic average and is more effective than simply including historical eviction filings as a control in the regression.

Finally, as there are likely city-level characteristics that are correlated both with the likelihood of enacting protections and with landlords’ eviction filing behavior, I add city-level fixed effects (represented as the $$alpha_i$$ in (1)). Moreover, to control for time trends, I include year and month effects⁷. To control for other time-varying covariates that may affect both the enaction of eviction protections and the filing rate, I control for the unemployment rate, COVID-19 deaths, and social distancing as measured by the change in time spent at home relative to January 2020.

Table 1 below presents summary statistics across the entire panel (city, year, and week), except where mentioned.

Table 1: Summary Statistics

Table 2 presents results from estimating the regressions as previously described. The first two columns report outcomes in terms of filings per 100,000 renting households, and the latter two report filings as a percent of the pre-pandemic baseline.

Table 2: Regression Results

The first row (A-I) in Table 2 suggests that local eviction moratoria were highly effective, as the coefficients are negative across the board, indicating that eviction filings in both level and percentage terms were much lower in moratorium periods relative to before the pandemic. Conversely, filings are not significantly different from pre-pandemic levels once local moratoria expire. Roughly speaking, moratoria may have prevented about 40 fewer filings per 100,000 per week, which is slightly lower than the estimates found by Rangel et. al (2021) for the CDC order (about 45). In percentage terms, they are associated with at least 20 percent fewer filings relative to the baseline (columns 3 and 4). In aggregate, this amounts more than half a million total filings prevented across our entire sample.

State-level bans appear to be slightly more effective than local moratoria (row B-I, columns 1 and 2) in level terms, but less effective in percentage terms (columns 3 and 4). This suggests that state eviction moratoria were as effective as local moratoria in preventing eviction filings, but their effects may have been concentrated in areas with a higher baseline rate of eviction filings. Interestingly, the effect of state-wide moratoria seems to be more persistent, as the coefficient associated with their expiration (row B-II) is still significantly below zero, meaning that eviction filings stayed low even after they were no longer in force. One possible interpretation is that state-wide eviction moratoria tended to expire earlier than local or national moratoria (see Figure 1), meaning that other moratoria were usually in effect when state moratoria were lifted. This may make it more difficult to estimate the effect of a state moratorium expiring relative to the pre-moratorium period, as some of the estimated effect may actually be due to the other moratoria remaining in place. However, as those moratoria are also included in the regressions in Table 2, that concern should be alleviated.

Finally, the CARES Act (row C) at first glance appears to be more effective than state or local moratoria, but the effect virtually disappears after including controls. This smaller estimated effect may be because the Act applied only to federally financed (GSE-backed) rental properties, whereas local moratoria applied to a broader class of rental properties. Conversely, the CDC order (row D-I) has a much stronger effect than state or local moratoria, even after including controls. This is likely because it had a much broader scope, even if its enforcement varied across jurisdictions. In percentage terms, it is associated with nearly a thirty percent reduction in eviction filings relative to the baseline. Although evictions filings are even lower in the period after it expires (row D-II, columns 1 and 2), this does not necessarily mean that the order was ineffective, especially as this pattern does not occur when adjusting for the baseline (D-II, columns 3 and 4). The estimated effect is about two-thirds greater than the estimates found by Rangel et al. (2021), who assume that evictions would have occurred at the pre-pandemic rate if it were not for the CDC order, and do not control for time-varying conditions such as the unemployment rate. However, it is more likely that in the absence of the CDC order, evictions would have occurred at an even higher rate relative to before the pandemic. In addition, Rangel et al. rely on extrapolation to infer eviction filings in out-of-sample areas, which may not be reliable.

Conclusion

In this note, I have shown that municipal eviction moratoria are strongly associated with a reduction in eviction filings, both in terms of filings per week and filings as a percentage of the pre-pandemic baseline. I find a similar effect for state bans in terms of filings per week, although the impact is smaller when expressed as a percentage of the baseline, potentially suggesting differences in the geographic concentration of effects. For the CARES Act, I find little effect after controlling for demographic and pandemic-related factors, whereas I consistently find a very large effect for the CDC order. This is perhaps because the CARES Act was of shorter duration and narrower scope than the CDC order. Of all eviction moratoria, the CDC order is by far the most effective, being associated with nearly twice as many fewer eviction filings when it is in force compared to state or local moratoria. Further work could more accurately quantify the strength of various moratoria, for example by grading their provisions on a qualitative scale or exploiting differences in courts’ willingness to enforce them, in order to explain why certain moratoria were more effective than others. Such work could also attempt to investigate the relationship between moratorium length and effect size or compare the impacts of national and local moratoria on secondary factors, such as rent growth.

Works Cited

An, Xudong, Stuart A. Gabriel, and Nitzan Tzur-Ilan (2021). “More Than Shelter: The Effects of Rental Eviction Moratoria on Household Well-Being,” unpublished paper, November.

Banerjee, Abhijit, and Maitreesh Ghatak (2003). “Eviction Threats and Investment Incentives (PDF),” Discussion Paper No. DEDPS/39. London: Suntory and Toyota International Centres for Economics and Related Disciplines, June.

Benfer, Emily A., David Vlahov, Marissa Y. Long, Evan Walker-Wells, J.L. Pottenger Jr., Gregg Gonsalves, and Danya E. Keene (2021). “Eviction, Health Inequity, and the Spread of COVID-19: Housing Policy as a Primary Pandemic Mitigation Strategy,” Journal of Urban Health, vol. 98 (February), pp. 1–12.

Callaway, Brantly, and Pedro H.C. Sant’Anna (2021). “Difference-in-Differences with Multiple Time Periods,” Journal of Econometrics, vol. 225 (December), pp. 200–30.

Centers for Disease Control and Prevention (2020), “Temporary Halt in Residential Evictions to Prevent the Further Spread of COVID-19,” agency order, Federal Register, vol. 85 (September 4), pp. 55292–97.

Desmond, Matthew, and Rachel Tolbert Kimbro (2015). “Eviction’s Fallout: Housing, Hardship, and Health,” Social Forces, vol. 94 (September), pp. 295–324.

Garboden, Philip M.E., and Eva Rosen (2019). “Serial Filing: How Landlords Use the Threat of Eviction,” City & Community, vol. 18 (June), pp. 638–61.

Goodman-Bacon, Andrew (2021). “Difference-in-Differences with Variation in Treatment Timing,” Journal of Econometrics, vol. 225 (December), pp. 254–77.

Jowers, Kay, Christopher Timmins, Nrupen Bhavsar, Qihui Hu, and Julia Marshall (2021). “Housing Precarity & the COVID-19 Pandemic: Impacts of Utility Disconnection and Eviction Moratoria on Infections and Deaths Across US Counties (PDF),” NBER Working Paper Series 28394. Cambridge, Mass.: National Bureau of Economic Research, January.

Leung, Lillian, Peter Hepburn, Matthew Desmond (2021). “Serial Eviction Filing: Civil Courts, Property Management, and the Threat of Displacement,” Social Forces, vol. 100 (September), pp. 316–44.

McCarty, Maggie, and Libby Perl (2021). “Federal Eviction Moratoriums in Response to the COVID-19 Pandemic,” Congressional Research Service, March.

Parrot, Jim and Mark Zandi (2021). “Averting an Eviction Crisis”, The Urban Institute, January.

Hepburn, Peter, Renee Louis, and Matthew Desmond. Eviction Tracking System: Version 1.0. Princeton: Princeton University, 2020. http://www.evictionlab.org.

Rangel, Jasmine and Jacob Haas, Emily Lemmerman, Joe Fish, and Peter Hepburn (2021). “Preliminary Analysis: Eleven Months of the CDC Moratorium”. Princeton Eviction Lab (August).

0http://www.nber.org/papers/w28394.pdf’>http://www.nber.org/papers/w28394.pdf

Fonte/Source: https://www.federalreserve.gov/econres/notes/feds-notes/how-effective-were-national-state-and-local-eviction-moratoria-20220621.htm