Policy evidence

What keeps people housed.

Eviction responds to policy. This page sorts the interventions with measured effects into four kinds, the four P's, ordered from fastest-acting to slowest. For each one it says what it does, what the evidence shows, and where it stands. A fifth section names the floor every lever stands on. Written for legislators and the staff who draft the bills.

Four P's · Thirty planks · Each marked Evidence or Reasoned · Every statistic carries its source · Updated October 2026

01 · Prevent 02 · Protect 03 · Preserve 04 · Produce 05 · The floor

3.6 million

Eviction cases landlords file in an average year. Almost 7% of renting households face one.

Gromis et al., PNAS 2022

5.5 to 1

Informal evictions for every formal one. Most forced moves never reach a court.

Gromis & Desmond, Cityscape 2021, 2017 AHS

35 per 100

Affordable and available rental homes per 100 extremely low-income renter households. The national shortage is 7.2 million homes.

NLIHC, The Gap, March 2026

22.7 million

Renter households paying over 30% of income for housing in 2024. A record, and roughly half of all renters.

Harvard JCHS, America's Rental Housing 2026

A filing count is the weakest number housing research publishes. It counts court cases, not displacement, and the 2017 American Housing Survey puts the national ratio at 5.5 informal evictions for every formal one. Where tenants have counsel and protections, more displacement passes through a court and gets counted. Where they do not, it disappears from the record. Every rate on this page carries that caveat.

The framework

No single lever stops an eviction. It takes four.

ERN organizes the evidence as four P's, sorted by where each policy meets the problem (before the courthouse, inside it, or in the housing market underneath) and by how fast it acts. The further down the list, the slower the lever, and the more it addresses the cause rather than the case.

01 · Prevent

Stop the case before it reaches court. Rental assistance, mediation, and time to pay resolve the debt instead of litigating it.

02 · Protect

Even the odds once a case exists. Counsel, sealed records, and a stated cause change what a filing costs a tenant.

03 · Preserve

Keep the low-cost homes that already exist. They leave the market quietly, one sale or one expiring subsidy at a time.

04 · Produce

Add homes, especially below market. The slowest lever, and the only one that closes the shortage underneath the docket.

Ordered by speed of effect: assistance lands in weeks, mediation in months, preservation and construction in years. The strongest single combination on the evidence pairs a Prevent tool (rental assistance) with a Protect tool (right to counsel). A fifth section names the floor under all four: the data, funding, and standing every lever depends on.

How to read each plank

Evidence the numbers in the plank were verified against the linked source Reasoned a logical case, argued from mechanism or arithmetic and labeled as such

✓ Federal law today ◑ Some states or cities ⊘ Ran nationally, then ended ○ Nowhere yet, or proposed Tier 1 = largest & fastest effect · Tier 3 = structural & slow

01 · Prevent

Nine in ten evicted tenants owe at least a month's rent. The fastest tools resolve the debt, not the tenancy.

Prevention acts before a case exists: pay the arrears, mediate the dispute, allow time to cure. The debts are small against the costs that follow. In the largest study of nonpayment cases, 90% of evicted tenants owed at least one month's rent, and across all tenants in the study the average balance at move-out was $1,252, about one and a quarter months' rent (Humphries et al., NBER 2024). The money goes to the landlord, the arrears clear, and no record is created.

T1Evidence⊘ Federal program ended

Permanent emergency rental assistance

The pandemic ran the largest test of eviction prevention to date. Emergency Rental Assistance made nearly 10.8 million payments to households from $46.5 billion made available (Treasury), and with moratoria in force, 2021 filings ran at half their historical average in tracked areas (Eviction Lab). The HUD-funded evaluation by ERN and the Housing Initiative at Penn, spanning more than 700 counties in 22 states, found that counties receiving more assistance saw fewer filings, with the largest effects where low-income renters, unemployment, and rent burden ran highest; assisted families in Los Angeles were 53% less likely to have moved and 65% less likely to end up on the street, and among unfunded applicants who moved, a third reported the landlord forced them out with no court proceeding, against about 18% of funded applicants (ERN & Housing Initiative at Penn, 2026). The design lesson carries to any successor: programs that verified eligibility with neighborhood-level data instead of pay stubs and hardship documents prevented more evictions. Minnesota has funded prevention assistance statewide since 1993 (Minn. Stat. § 462A.204); in the year to September 2024, 94% of households that received its prevention assistance kept their housing, a program exit figure rather than a counterfactual estimate (Minnesota Housing). The plank is a standing national program that expands automatically when arrears rise, instead of one built from scratch in each emergency.

T1Evidence◑ Local models

Divert the case before it reaches the docket

In Phase 2 of Philadelphia's diversion program, 7.4% of cases that received rental assistance faced an eviction filing within six months, against 42.2% without assistance, and 4.2% where the assistance came with a mediated agreement; over 21 months the assisted rate rose to 33.3%, a sign assistance sometimes delays rather than fully prevents (Dowdall & Goldstein, Reinvestment Fund 2023). Philadelphia has required every landlord to enroll in the program and participate in good faith before filing since 2022, a rule made permanent in 2024 with a 30-day minimum (Phila. Code § 9-811). Washington State required pre-filing resolution statewide from November 2021, and the courts' own evaluation found a 78% settlement rate before the program ended in 2023 (AOC legislative report). Texas ran the model statewide through its Supreme Court and housing agency: courts paused cases while both parties applied, and more than 25,000 applicants received over $243 million (TDHCA, 2021–2023). Hawaii required mediation before any nonpayment filing in 2021 and 2022, and its legal aid review found 85% of mediated tenants stayed housed (Hawaii Appleseed, as reported 2022). Across 24 court-based programs in 17 states and D.C., 89% of cases that engaged with diversion ended in a settlement or voluntary dismissal, and only 6% went to trial (National Center for State Courts, 2025). The plank makes good-faith diversion a national precondition of filing.

T2Evidence◑ State models

Slow the clock: notice, time to cure, a right to redeem

The statutory clock between falling behind and facing a filing decides whether assistance and mediation get the time they need. The cleanest natural experiment is Maryland, where for decades no notice was required, filing cost $15 (a fee the state raised in 2024), and trial came on the fifth day: Baltimore City alone recorded 144,058 landlord-tenant filings in fiscal 2017, thirty times the 4,788 in all of King County, Washington, a gap in process, not in hardship (Maryland Judiciary FY2017; Losing Home, 2018). Maryland added a 10-day pre-filing notice in October 2021 (Md. HB 18, Ch. 746). In ERN's national eviction-risk model, two of the five strongest predictors of a neighborhood's filing rate, out of 184, are process clocks: the notice period and the post-judgment window (ERN, HPRM). Washington State pairs a 14-day pay-or-vacate notice (SB 5600, 2019) with a statutory right to reinstate the tenancy by paying what is owed, before judgment or within five court days after (RCW 59.18.410); Minnesota adopted a 14-day pre-filing notice in 2024 and lets local governments set a longer one (Minn. Stat. § 504B.321). No study has yet isolated the effect of the notice length itself. Every redemption is an eviction that did not happen at a cost the tenant already owed. ERN's state timeframes table compares all fifty states and D.C.; for the full process law from pre-filing through post-judgment, the Legal Services Corporation's Eviction Laws Database, built with Temple's Center for Public Health Law Research, covers every state and territory as a snapshot of law in effect January 1, 2021.

T1Evidence◑ Programs, not a guarantee

Fund homelessness prevention like the returns are real

Families who called Chicago's homelessness-prevention line when one-time emergency funds happened to be available were 76% less likely to enter a homeless shelter within six months than families who called when funds were not, at roughly $10,300 per instance of homelessness averted (Evans, Sullivan & Wallskog, Science 2016). A Santa Clara County randomized trial of prevention assistance returned an estimated $2.47 in benefits for every dollar spent (Phillips & Sullivan, 2023). Chronic homelessness costs the public roughly $35,000 per person per year (National Alliance to End Homelessness, 2017). Prevention spends four figures per household; homelessness costs five figures per person per year. The populations differ, so this is a comparison of magnitudes, not a one-to-one ledger.

T1Evidence◑ Programs, not a right

Permanent housing for people experiencing homelessness

On a single night in January 2025, 745,652 people were homeless in the United States (HUD 2025 AHAR). The evidence for permanent supportive housing is strong exactly where a housing policy needs it. It maintains housing stability for most chronically homeless people over the studied one-to-two-year windows, while the National Academies found no substantial evidence yet that it improves health outcomes, and this page keeps that distinction (NASEM 2018). For homeless families, a long-term rent subsidy cut subsequent homelessness by more than half against usual care (HUD Family Options Study, 2016). The plank funds permanent housing at the scale of the count and evaluates it on the outcome it demonstrably moves.

02 · Protect

Landlords arrive with a lawyer 84% of the time. Tenants, 4%.

Protection changes what happens once a case exists: who has a lawyer, what the record does afterward, and whether a tenancy can end without a stated reason. None of it prevents the filing; all of it changes what the filing costs.

T1Evidence◑ States and cities, no federal right

A right to counsel in eviction court

Absent a right to counsel, about 4% of tenants have lawyers against 84% of landlords, the National Coalition for a Civil Right to Counsel's average across several dozen local studies. 28 jurisdictions (5 states, 21 cities, 2 counties) have enacted some version of the right (NCCRC, August 2026); Washington State's 2021 law was the first statewide (RCW 59.18.640), with Maryland and Connecticut following the same year. The results are measured. In New York City, 84% of households represented in court remained in their homes in FY2023 (Office of Civil Justice); in Cleveland, represented clients with a legal defense prevented an eviction or involuntary move 83% of the time in 2025 (Stout independent evaluation, 2026). Statewide programs show the same pattern: Connecticut's attorneys helped clients avoid disruptive displacement in 85% of cases and returned an estimated $3.65 per dollar spent (Stout, 2025), and Maryland's program kept 88% of clients who wanted to stay in their homes, at $3.04 per dollar (Maryland Attorney General, 2025). Counsel also answers the default problem: across New York State's covered jurisdictions, 54% of nonpayment filings went unanswered by the tenant, and defaults account for an estimated 40% of the eviction warrants issued across the state (NYU Furman Center, 2023). A 2018 Stout analysis for Philadelphia projected roughly $12.74 returned for every dollar spent on tenant counsel. The plank is a national right, funded like the public-defender obligation it parallels.

T2Evidence◑ Some states

Just cause to end a tenancy

A just-cause standard ends no-cause terminations, so a tenancy can be ended only for a listed reason. New Jersey has required good cause since 1974, a law its legislature tied to the state's housing shortage (N.J.S.A. 2A:18-61.1); Oregon and California adopted statewide standards in 2019 (SB 608; AB 1482), Washington State in 2021 (RCW 59.18.650), and Colorado and New York in 2024. The CARB anti-displacement review rates just-cause protections high-potential and fast-acting (Chapple & Loukaitou-Sideris, 2021). The direct effect is definitional, ending no-cause terminations; causal estimates of the effect on filings are thinner than for the tools above, and this page treats that as an open question rather than settled. The plank sets a national floor while leaving the list of causes to the states.

T2Evidence◑ Some states

Make voucher discrimination illegal everywhere

A voucher only prevents an eviction if a landlord will take it. In HUD's five-city paired-testing study, landlords refused voucher holders in 78% of tests in Fort Worth, 76% in Los Angeles, and 67% in Philadelphia, against 31% in Newark and 15% in Washington, D.C. The study's own summary is that landlord refusal is more common in jurisdictions without source-of-income protections, though it did not evaluate the laws' impact (Cunningham et al., Urban Institute for HUD, 2018). Adoption has spread far enough that PRRAC estimates it is likely that over 60% of voucher holders are now covered by a state or local law (Appendix B, updated March 2026). Refusing a tenant because rent arrives as a voucher wastes money the public has already committed. The plank adds source of income to federal fair-housing law, so protection stops depending on the state line.

T2Evidence◑ 19 states and D.C.

Seal the record until the case has merit

Screening reports recite any filing regardless of outcome, and the Fair Credit Reporting Act lets a civil record ride a report for up to seven years (15 U.S.C. § 1681c). Roughly 22% of eviction records carry ambiguous information on how the case was resolved or falsely represent a tenant's history (Porton, Gromis & Desmond, Housing Policy Debate 2021), landlords in a behavioral experiment treated mere filings like executed evictions (So, Housing Policy Debate 2023), and the CFPB logged more than 16,000 renter complaints about incorrect information on screening reports (2022). Black renters are 19.9% of adult renters and 32.7% of eviction-filing defendants (Hepburn, Louis & Desmond, Sociological Science 2020), which is why blanket record screening draws disparate-impact challenges under the Fair Housing Act; the first such case, by the ACLU's account, was brought for a tenant whose years-old filing never resulted in an eviction (Smith v. Wasatch, W.D. Wash. 2017, built on ERN's King County analysis). Nineteen states and D.C. have passed sealing or expungement policies (NLIHC, 2026), and NLIHC's State and Local Tenant Protections Database tracks adoption jurisdiction by jurisdiction. California hides every filing from the general public by default and unmasks it only if the landlord wins within 60 days of filing, or later by court order (CCP § 1161.2); Massachusetts lets tenants seal no-fault cases at any time, nonpayment cases after four years, and dismissals or tenant wins without a hearing (G.L. c. 239, § 16, 2024); Washington State seals case by case, at the court's discretion (RCW 59.18.367). The plank is the California default, nationally.

T2Evidence○ Largely unaddressed

Price serial filing out of the collections business

In 2014, nearly one in three households facing eviction nationally was filed against more than once (Leung, Hepburn & Desmond, Social Forces 2020), and serial filing runs highest where filing is cheap and fast. A 2023 study ties every $76 of filing fee to a 3.11-point drop in serial filing (Gomory et al., Housing Policy Debate 2023; both via the Network for Public Health Law fact sheet, 2025). The plank pairs meaningful filing fees with limits on repeat filings against the same household, so the courthouse stops functioning as a rent-collection window.

T2Evidence◑ One state, local bans, federal consent judgment

Ban algorithmic rent collusion

Rent-setting software that pools competitors' private data lets landlords coordinate increases no one would risk alone. The Department of Justice sued RealPage in 2024, and a consent judgment entered in May 2026 requires the software to stop using competitors' nonpublic data to set rents; the states' claims against RealPage and several large landlords continued past a September 2026 ruling (DOJ; Federal Register). New York enacted the first statewide ban in October 2025, making it unlawful to run or license software that coordinates rents between landlords (S. 7882, Ch. 437); Seattle banned the practice in 2025 (SMC 7.34) and King County followed for its unincorporated areas. The plank writes the same rule into national law, so it stops depending on case-by-case enforcement.

T2Evidence◑ Some states

Anti-gouging rent stabilization, designed not to chill building

Second-generation caps limit increases inside a tenancy while exempting new construction, which is where the supply risk lives. Oregon set the first statewide cap in 2019, at 7 points plus inflation with a 15-year new-construction exemption (SB 608, since tightened to a 10% ceiling); California followed the same year at 5 points plus inflation, capped at 10% (AB 1482); Washington State's 2025 law caps increases at 7 points plus inflation to a maximum of 10% (RCW 59.18.700), with a 12-year new-construction exemption (RCW 59.18.710). The honest evidence cuts both ways. San Francisco's 1994 expansion cut its rental supply about 15%, with landlords converting buildings to condos and redeveloping them rather than cutting new construction, under a law that also exempted new building (Diamond, McQuade & Qian, AER 2019). Rent regulation is a stability tool, not a supply tool, and the new-construction exemption is its single most protective design feature. The plank is a national anti-gouging ceiling of that second-generation design, not a hard freeze.

03 · Preserve

2.5 million of the cheapest rentals left the market in a decade. No court recorded it.

Preservation keeps existing low-cost homes on the market: the aging private stock that rents cheap, and the subsidized stock whose affordability runs on a contract clock. Losses here displace people without any legal event, which ERN measures as soft displacement, so the harm never appears in a court record.

T2Evidence◑ Funded, oversubscribed

A national fund for the unsubsidized low-cost stock

Between 2014 and 2024 the country lost 9.3 million units renting under $1,400 a month; 2.5 million of them rented under $600, a 30% decrease in the lowest-cost stock in one decade (JCHS, America's Rental Housing 2026). No eviction is filed when a building leaves the low-cost market; the households that would have lived there never appear in any record. About three quarters of Americans in affordable housing live in unsubsidized units, and the CARB review rates preserving them among the highest-potential anti-displacement strategies, effective in any market and fast (Chapple & Loukaitou-Sideris, 2021). Where funds exist they are swamped: Washington State's 2026 preservation round drew $97.7 million in requests against $34.1 million available (Commerce, May 2026). The plank is a standing federal acquisition fund sized to the demand states already demonstrate.

T2Evidence◑ DC and manufactured-home states

A first chance for tenants to buy the building

Under Washington, D.C.'s tenant opportunity-to-purchase law, 58% of tenants in one study sample made it through the purchase process and could afford their units in the end (Gallaher 2016, via the CARB white paper), and a 2023 review counted 16,224 affordable homes developed or preserved under the law between 1975 and 2021 (CNHED). Washington State requires notice before a manufactured-home community is marketed for sale and gives residents 70 days to form an organization and declare their intent to compete, an opportunity to compete rather than a right of first refusal (RCW 59.20.325). The plank is a national notice-and-opportunity floor for any sale of an occupied rental property, with financing to make tenant and nonprofit bids real. PolicyLink publishes a policy brief on tenant and community opportunity to purchase in its All-In Cities toolkit.

T2Reasoned◑ State patchwork

Protect the homes people own on land they rent

A manufactured home is the largest unsubsidized affordable-housing stock in the country's rural counties, and its owner is uniquely exposed. The home is theirs, the ground is not, and moving the home is often impossible in practice. Washington State caps lot-rent increases at 5% (RCW 59.20.370) with three months' notice of any increase (RCW 59.20.090); its sale-notice law names community land trusts among the eligible buyers (RCW 59.20.030). Minnesota requires 45 days' notice before a park is sold to a converter, lets residents match the price, and pays relocation costs from a state trust fund when a park closes (Minn. Stat. § 327C.095). The plank is a national floor of lot-rent protection, sale notice, and purchase financing for resident ownership. The logic is exposure. Rules written for tenants and rules written for owners both miss the household that is each on one side of the property line.

T3Evidence◑ Local, narrow reach

Community control of land

Community land trusts take land off the private market for good. The CARB review rates their displacement-prevention potential high and, unlike new construction, they do not need a hot market to work; the same review finds them rare, dependent on coordinated tenant action and economic resources, and generally not reaching the lowest incomes (Chapple & Loukaitou-Sideris, 2021). The plank funds acquisition at a scale that changes that last finding.

T2Evidence✓ Federal law, floor set high

Lower the ceiling on corporate single-family ownership

Federal law, in force from January 2027, bars an investor holding 350 or more single-family homes from buying another, with exceptions and no divestment requirement (21st Century ROAD to Housing Act, Public Law 119–101, 2026). The eviction link is measured. In Fulton County, Georgia, large corporate owners of single-family rentals were 8% more likely than small landlords to file eviction notices, after controls (Raymond et al., Federal Reserve Bank of Atlanta, 2016). The plank lowers the threshold toward the 100-home lines states are already debating, and adds beneficial-ownership disclosure so the count cannot hide behind shell LLCs.

T2Evidence○ No standing renewal fund

Renew the affordability the public already bought

Five million rental homes, one in ten rentals nationally, carry federally funded project-based subsidies, and affordability restrictions are set to expire on 374,497 of them by 2028 (NLIHC & PAHRC, 2024 Picture of Preservation). When an owner opts out, the affordability ends without a single court event, and the public usually buys the same affordability twice somewhere else. At-risk properties are tracked one by one in the National Housing Preservation Database. The plank is a standing renewal fund plus a first right to purchase for mission-driven buyers whenever a subsidized property's restrictions approach expiration. Renewing a home the public already financed is the cheapest unit the public will ever add.

04 · Produce

35 affordable homes exist for every 100 of the lowest-income renters. Nothing above closes that gap.

Production is the slowest lever and the only structural one: every tool above operates inside a shortage that none of them reduces. Its evidence runs through rents and exposure rather than courtrooms. New supply lowers nearby rents, and rent burden is where eviction risk concentrates.

T3Evidence◑ State reforms spreading

Build more homes, of every kind

Large new market-rate buildings in low-income areas lower nearby rents by 5 to 7 percent relative to comparable areas, and draw more in-movers from lower-income neighborhoods, the reverse of the displacement usually feared (Asquith, Mast & Reed, REStat 2023). The chain is measured: a new market-rate building housing 100 people leads 45 to 70 people to move out of below-median-income neighborhoods, most within three years (Mast, JUE 2023). City-scale reforms show the same direction. Auckland upzoned about three quarters of its residential land in 2016; eight years on, rents were an estimated 23% below the no-reform counterfactual (Greenaway-McGrevy & So, Economic Inquiry 2026) after a construction boom the reform demonstrably stimulated (Greenaway-McGrevy & Phillips, JUE 2023). Minneapolis grew its stock 12% from 2017 to 2022 after its land-use reforms while rents rose 1%; the rest of Minnesota added 4% and rents rose 14% (Pew, 2024). The pipeline is currently shrinking: multifamily starts fell from a three-decade peak of 547,000 units in 2022 to 416,000 in 2025 (JCHS 2026). The plank ties federal infrastructure money to a statewide zoning floor that legalizes middle housing, with Washington State's package as the model.

T3Evidence○ Far short everywhere

Build the homes the market never builds

The national shortage is 7.2 million rental homes affordable and available to the country's 11 million extremely low-income renter households; only 35 exist per 100 such households, and 74% of those renters spend more than half their income on rent and utilities (NLIHC, The Gap, 2026). Across cities, homelessness tracks housing-market conditions, absolute rent levels and vacancy rates, more than it tracks local poverty, addiction, or weather (Colburn & Aldern, Homelessness Is a Housing Problem, 2022). The market cannot price a home an extremely low-income renter can pay, because the rent that household can afford does not cover the cost of operating the building, let alone constructing it. Deep affordability exists only where the public builds it or buys it. The plank is a production program targeted at 0–30% of area median income, sized against the 7.2 million gap. No prevention or protection tool changes that arithmetic.

T3Reasoned◑ Local experiments

Social housing that pays for itself

Publicly developed mixed-income housing, where higher rents cross-subsidize lower ones, is standard practice in much of the world and newly live in the United States. Seattle voters created a social housing developer in 2023 and funded it in 2025 with a 5% payroll tax on compensation above $1 million, estimated at roughly $53 million a year (Proposition 1A). The plank is a federal revolving loan fund that lets any city capitalize a developer like it once, then recycle the money. The logic is a balance sheet. A public developer borrows once, builds, and repays from rents, so a dollar produces homes repeatedly instead of subsidizing once.

T3Reasoned◑ State models

Public land for public homes

Governments already own land in the places homes are scarcest. Washington State lets any agency transfer surplus public property for affordable housing on any mutually agreeable terms, including at no cost, with a covenant binding the use (RCW 39.33.015), and requires counties to offer tax-foreclosed parcels to the city for affordable housing before auction (RCW 36.35.150). The plank is a federal default. Surplus public land is offered for permanently affordable housing before anything else, priced at its use, not its speculation.

T3Reasoned◑ State models, oversubscribed

Turn renters into owners, and count the vacancy it creates

Every renter who buys leaves a rental behind, so ownership programs are rental-supply policy too; the chain breaks only where the purchase itself displaces a sitting tenant. That much is arithmetic, not a study. Washington State's Covenant Homeownership Program lends 0%-interest down-payment assistance to first-time buyers from communities harmed by racially restrictive covenants, funded by a $100 recording fee projected to raise $75–100 million a year; demand already outpaces its monthly funds (HB 1474, 2023). HUD opened a Fair Housing Act investigation of the program in March 2026, which is unresolved. The plank is a national down-payment program with a repair-the-harm design, sized to its own waiting lists.

T3Evidence◑ Rated, not yet funded to scale

Trade density and tax breaks for below-market homes

Inclusionary and incentive tools depend on a strong market; the CARB review rates them medium-potential and long-term (Chapple & Loukaitou-Sideris, 2021). Washington State's transit-density law attaches 50-year affordability strings to its upzones. The plank keeps the trade honest nationally. Public value created by public rezoning buys permanent affordability, not a one-cycle discount.

05 · The floor

The best national count still misses a million cases a year. Every lever above stands on this floor.

Four levers, one foundation: data that shows whether any of this works, assistance that does not run out mid-year, and the standing to invoke a right without losing your home for trying. Two of these planks are measured; four are labeled as the reasoned arguments they are.

T1Evidence◑ Six states publish; no federal count

Count every eviction

The federal government's own review concluded that “relatively little comprehensive data exist on evictions in the U.S.” (GAO-24-106637, 2024), and the best national estimates, built from 99.9 million collected court records, still undercount filings by about 1 million cases a year because some counties' records cannot be obtained (Gromis et al., PNAS 2022). In ERN's 50-state census of eviction data sources, only six jurisdictions (Alaska, Connecticut, DC, Maryland, New York, Virginia) publish statewide case-level eviction data. California will join them: SB 1160, which directs the Judicial Council to collect and publish eviction case data from every county, was signed on September 27, 2026 (Chapter 681, Statutes of 2026), with reporting phased in over the following years (leginfo). The plank is a national eviction data standard. Courts report standardized, outcome-complete case data the way public health reports disease. Nothing else on this page can be evaluated without it.

T1Evidence○ Rationed by lottery and line

Make rental assistance an entitlement, not a lottery

Federal rental assistance reaches 5.1 million households, and 3 in 4 eligible households receive none of it (CBPP, 2026). Eviction is not only a hardship; it is a cause of the harms that follow it. A study using cases randomly assigned to judges across Cook County and New York courts found an eviction order raised shelter use by 3.4 percentage points, an increase of more than 300%, cut earnings about 14% by the second year, and raised hospital visits 29%, with the housing and labor effects concentrated among female and Black tenants (Collinson et al., QJE 2024). And when voucher families in the Creating Moves to Opportunity experiment received customized search help, landlord outreach, and short-term financial assistance, the share moving to high-upward-mobility neighborhoods rose from 15% to 53% (Bergman et al., AER 2024). The plank phases federal rental assistance toward an entitlement for eligible households, paired with the search help that makes a voucher a real choice.

T2Reasoned○ Nowhere yet

Subsidy parity for renters

Federal tax law subsidizes owning a home as a standing entitlement, no waiting list attached. Assistance for renting one is rationed. The plank is a renter's tax credit that closes part of that gap for cost-burdened households. The logic is symmetry. If housing is worth subsidizing through the tax code, tenure should not decide who gets the subsidy.

T2Reasoned○ No general national right

A right to organize without retaliation

Every protection on this page is enforced, in practice, by a tenant willing to invoke it, and retaliation is how invocation gets suppressed. Labor law solved the same problem a century ago by protecting the organizing itself. The plank is a national right to form and join tenant associations, with retaliation for organizing treated the way retaliation for union activity is treated. The logic is enforcement. Rights that depend on isolated individuals asserting them against their own landlord under-enforce by design.

T2Reasoned◑ Complaint-driven almost everywhere

A habitability floor someone actually inspects

Habitability standards exist nearly everywhere and are enforced almost nowhere, because enforcement waits for a tenant complaint, and complaining invites the eviction the rest of this page is about. The plank pairs a national habitability floor with proactive rental inspection and registration, shifting enforcement from the household with the least power to the government with the most. The logic is incentives. A standard enforced only by its most vulnerable beneficiary is a standard designed to be waived. PolicyLink publishes a policy brief on rental registries, the registration half of this plank.

T2Reasoned○ Owner-centric by design

Rehouse renters after disasters like owners

Disaster recovery is built around the property owner. Rebuilding aid follows the deed, and the renter displaced from the same address holds a thinner claim on a shorter clock. Climate change is making displacement without an eviction filing a growing share of all displacement, and none of it appears in court data. The plank writes tenure-neutral rehousing into disaster law. Displacement is measured by who lost their home, not by who owned it.

The pandemic ran the experiment no legislature would have authorized. For about two years the country operated several of these planks at once, a filing moratorium and $46.5 billion in emergency rental assistance, and 2021 filings ran at half their historical average, with more than three million cases avoided between March 2020 and the end of 2021 (Eviction Lab). Filings returned as each piece lapsed. The lesson is not that emergencies work. It is that these levers move the number, and that no state holds all four at once.

The costs of doing nothing are measured. An eviction filing that never reaches a judgment is associated with 19% higher mortality, and an eviction judgment with 40% (Graetz et al., Social Science & Medicine 2024). Eviction during pregnancy is associated with lower birth weight and more premature births (Himmelstein & Desmond, JAMA Pediatrics 2021). These are associations measured on millions of records, and they fall on the same households the screening, filing, and shortage planks above describe.

The pair that works is a lawyer and a check. Cleveland's right-to-counsel evaluation shows the mechanism. About half of eviction cases carry no legal defense, and those cases, in the evaluator's words, “center on the amount of back rent owed, which often requires a monetary solution.” In 2025, represented clients with a legal defense prevented an eviction or involuntary move 83% of the time; clients without one, 70%. The program's prevention rate for 2025 cases, 76%, sat below its 83% since 2020, in the year the county's rental assistance programs closed (Stout, 2026). A lawyer wins the case with a legal defect. A check resolves the case without one. National policy funds both, permanently, and measures the result with data it finally collects.

What we know

What works is no longer the open question. A decade of evaluations points the same way. Rental assistance, diversion, representation, and record sealing all show measured results, and several return more than they cost. The open questions are implementation and speed: whether a jurisdiction is willing and able to adopt and fund what is proven, and how to keep people housed while the slow levers catch up. Where the evidence on eviction specifically is mixed or indirect, each plank says so on its face rather than rounding up. For the full literature, see the research library. For a tailored briefing on any plank, contact ERN.

How this list was made: every plank marked Evidence carries direct, quantified findings of reduced eviction filings, judgments, or displacement, or of a measurably changed case cost, from peer-reviewed studies, federal evaluations, statutes, and program records, linked in place. Each was verified against the named source in August 2026 and checked again in October 2026. Planks marked Reasoned are argued from mechanism or arithmetic and labeled as such rather than dressed up as findings. Where the evidence is thin or cuts both ways, the plank says so. The four-P framing is ERN's, and the plank structure is shared with ERN's draft national housing rights charter. For how eviction itself works (the process, the records, the data), see Understanding Eviction; for the full literature, the research library; for the adoption trackers and drafting libraries other organizations maintain, the reference shelf at the foot of this page.

For hearings and bills

Need this evidence for a hearing?

If you need a policy briefing for a hearing or a bill, we write those: your jurisdiction's data, the relevant evidence, and its limits, in plain language.

Request a briefing → Browse the research library →

The reference shelf

This page grades the evidence behind each plank. Four companion resources, maintained by other organizations, track what this page does not: which jurisdictions have adopted what, how the legal process runs in each state, and how to draft the ordinance. ERN links them here because a legislator's next question after "does it work" is "who has it, and what does the bill say."

State and Local Tenant Protections Database, National Low Income Housing Coalition. Tracks right to counsel, source-of-income protection, eviction record sealing, anti-rent-gouging measures, and just-cause laws across states, territories, tribes, and localities, with comprehensive tracking since January 2021. The live answer to where any Protect plank stands in your jurisdiction.
Eviction Laws Database, Legal Services Corporation, with the Center for Public Health Law Research at Temple University. Maps the legal process of eviction from pre-filing through post-judgment for all fifty states, D.C., and the U.S. territories, plus a 30-jurisdiction local sample. A snapshot of law in effect January 1, 2021, which makes it the baseline against which the protections passed since can be read; ERN's state timeframes table carries the current process clocks.
Housing Policy Library, Housing Solutions Lab at the NYU Furman Center, originally built as the Local Housing Solutions toolkit. Implementation briefs across the full breadth of local housing policy, organized on four pillars of its own (dedicated affordable units, housing supply, access to private-market homes, and protection against displacement and poor conditions). The drafting-level detail this page leaves to the bill writers.
All-In Cities Anti-Displacement Toolkit, PolicyLink. City- and organizer-facing policy briefs, including just cause, rent stabilization, tenant and community opportunity to purchase, rental registries, and eviction-record and tenant-screening protections.