What you should know about eviction.

An eviction is, at bottom, the legal removal of a renter from their home. But the word covers a much wider event than the court case that carries its name. The case begins when a renter falls behind: the typical case is filed for nonpayment of rent, and 90% of evicted tenants owe at least a month of rent when they are removed. What it does not record is the question that matters most for the family: did they manage to stay, or were they displaced?

Two households with identical court records can live completely different lives afterward. One negotiates, pays what they can, and keeps the apartment. The other packs up — sometimes before a judge ever rules — and begins the search for somewhere they can afford, near work, near a child's school. The record looks the same. The outcome does not. Eviction data tells you that a household reached the edge of losing a fundamental need. It does not tell you whether they fell.

That gap is the first thing to understand about eviction numbers, and it shapes everything that follows. The figures on this page describe a system under enormous strain. They cannot, on their own, describe the life of any one family inside it.

The reason a case is recorded — nonpayment — is almost never the whole reason. It is the charge that stuck.

A nonpayment filing records the most easily proven fact — money owed — but rarely the full context around it. Behind a single filing can sit a rent increase, a change in management, an unresolved repair request, a dispute, a new child, a disability, a building-code complaint. The court writes down only the dollars.

It is tempting to imagine the evicted tenant as careless. The data describes someone else: a working household one shock away from the edge. About 90% of evicted tenants owe at least a month of rent when they are removed, and the typical amount owed at move-out is around $1,250 — roughly one and a quarter months' rent. People are not evicted because they stopped trying. They are evicted because a thin margin gave way.

That last number is the engine. When more than a third of adults cannot absorb a $400 surprise, the distance between "current on rent" and "in eviction court" is a single car repair, a cut shift, an emergency-room bill. In one ERN study of tenants in eviction, a large share carried recent medical debt. The causes stack: income that swings week to week, a layoff six months earlier, a health crisis. The rent is simply the bill that comes due first and most visibly.

And the sums are often smaller than people assume. Many eviction cases are filed over modest amounts — sometimes a fraction of a month's rent — even though missed months can compound the total by the time a household is removed. Sometimes the unpaid rent is itself tied to the condition of the home: mold, a broken furnace, an air-conditioner that can't keep up with a Central Valley summer, a repair left undone. The ledger records none of that. It records the shortfall.

"Eviction" is not one process — it is four phases that every state assembles in its own way: a grace period (in a few states), a notice period, a court phase from filing to judgment, and a post-judgment window from the judgment to the sheriff's lockout. ERN compiled the typical clock for all fifty states and the District of Columbia. End to end it runs from under a week to about two months — but two states with the same total can get there by very different routes.

How a state spends those days varies as much as the total. Some front-load the notice period; Massachusetts, Oregon, and Vermont have the longest court phases; Pennsylvania and D.C. pile most of the days after the judgment, in the wait for the sheriff. The full state-by-state chart — every phase, every citation, every effective date, sortable — lives on its own page:

→ State Eviction Timeframes
The verified statutory clock for all 50 states and D.C.: floors of 5–56 days, ceilings only where the law sets one, per-stage citations and effective dates, and a downloadable dataset. Louisiana and West Virginia can lawfully reach a lockout in about five days; Massachusetts takes about two months.

A renter's odds depend enormously on which of these fifty-one clocks they happen to live under — and on how the days are arranged within it. A long notice period is time to find rent or a lawyer; a long post-judgment window is time to plan a move. The same total can be far more, or far less, survivable depending on where the days fall.

This is the part the data is most hopeful about: evictions are not weather. They respond to policy, and several interventions have a measurable record. Start with representation. In eviction court, the sides are profoundly unequal — nationally only about 4% of tenants have a lawyer, against 84% of landlords — the National Coalition for a Civil Right to Counsel's average across roughly forty local studies, absent a right to counsel. The gap shapes outcomes before a word is spoken.

A lawyer is not magic — it is navigation. Eviction is a civil case, so there is no public defender; a renter facing a seasoned landlord attorney is usually negotiating alone, often without understanding the process. Many never appear at all: across New York State courts, 54% of nonpayment filings went unanswered by the tenant (2016–mid-2022) — and while a default judgment requires the landlord's request rather than following automatically, defaults account for a full 40% of the eviction warrants the state issues. Representation closes that gap, and where tenants get it, large majorities reach a favorable outcome and most keep their homes.

The largest test of the money side is now published. ERN and the Housing Initiative at Penn evaluated the pandemic's $46.55 billion Emergency Rental Assistance program for HUD — Keeping People in Their Homes (2026) — using filing records, Treasury payment data, and household surveys across 16 states. Counties that received more ERA saw fewer filings, especially where unemployment and rent burden ran high, and fact-based proxy rules that cut application paperwork amplified the effect. In the case-study cities, Los Angeles delivered nearly $2.4 billion to 181,117 households and Philadelphia over $235 million to 31,335; assisted tenants had 53% lower odds of moving (Los Angeles) and about 65% lower odds of street homelessness. One finding reaches beneath the court data entirely: assisted households had 45% lower odds of a forced move with no court process — rental assistance suppresses the informal evictions of Figure 1, not just the filings. Across the study counties, filings from March 2020 to early 2022 ran about 45% below pre-pandemic forecasts — roughly 673,000 fewer — the combined force of moratoriums, state protections, and ERA together.

But a lawyer alone leaves the underlying problem — the unpaid rent — in place. The most effective combination pairs counsel with money. In Phase 2 of Philadelphia's pre-filing diversion program, just 4.2% of cases that reached a mediated agreement and received rental assistance went on to face an eviction filing within six months, against 42.2% without assistance — though over 21 months the assisted rate rises to 33.3%, a sign assistance sometimes delays rather than fully prevents. Prevention beats litigation: the gold standard is stopping an eviction before it is ever filed.

ERN organizes the evidence as four P's, ordered from fastest-acting to slowest:

The funnel in Figure 2 is not hypothetical. Where researchers have linked court records to what actually happened, the attrition is dramatic: in Cook County, about 65% of eviction cases end with an eviction order, and the Sheriff's Office executes about 26% of those orders — roughly 17 filings in 100 end with a forced removal. In New York City, about 35% of nonpayment cases end in an order and 31% of orders are enforced by a marshal — roughly 11 in 100, with around 64% of non-order cases ending in settlement (Collinson et al., Quarterly Journal of Economics, 2024). Baltimore is the extreme: a rent court processing roughly 150,000 complaints a year produced about 6,500–6,900 completed evictions — under 5% of filings (Justice Diverted, 2015).

The harm does not wait for the sheriff. The same linked-records research finds an eviction order raises the probability of emergency-shelter use by 3.4 percentage points and of appearing at a new address by 8.2 points within a year, with effects concentrated among Black and female tenants — and long-run damage to credit and debt that outlasts the case.

A filing is a threat recorded; a judgment is a loss recorded; a lockout is a removal recorded. Three different measures of three different things.

Only a handful of places count eviction notices at all. Los Angeles has required landlords to file every notice with the city since February 2023: 290,560 notices in its first three years and two months — roughly double the county's unlawful-detainer filings over the same window, and a hard floor, since the notices cover only the city while the filings cover the county (LA Controller; Judicial Council of California). Washington ran the only statewide notice census to date, a 2021–23 pilot that logged 158,251 pay-or-vacate notices — more than 7,000 a month in a state that filed 15,000–20,000 eviction cases a year before the pandemic; the courts' own report called the volume "significantly higher... than anyone understood." San Francisco is the caution: its registry excludes nonpayment notices, so filings there exceed registered notices — registry design decides what you see.

Set against Figure 1's estimate of 5.5 informal forced moves per formal eviction, the arithmetic compounds: filings sit midway down a ladder most displacement never touches. They are also not harmless paperwork. A one-point rise in a county's filing rate predicts a measurable rise in sheltered homelessness the following year — an association that holds for filings even where judgments show none (Treglia, Byrne & Rai, 2023).

ERN distinguishes hard displacement — a formal, involuntary event, proxied by eviction filings — from soft displacement: the market pressure of rising rents and neighborhood change that pushes low-income households out with no legal event at all. The two substitute for each other. Where tenant protections are strong, formal filings are suppressed while displacement reroutes through non-renewals, buyouts, and owner move-ins — San Francisco's pattern. Where protections are minimal and market pressure low, eviction operates as the primary displacement mechanism — the pattern across much of the South.

This is the problem the Housing Precarity Risk Model exists to solve. Its eviction sub-model (EER) measures a neighborhood's filing rate relative to its own state's average and models it nationally — the standardization raw counts cannot give — while its displacement sub-model (EDR) estimates the net out-migration of low-income renters that filings never see. One model result belongs on this page: among the EER's 184 predictors, two time-to-evict variables — the notice period and the post-judgment window — rank in the top five. The speed of a state's legal process shapes its filing rate independent of its housing market: fast, cheap process invites filing; slow process deters it. The clock in Part I is not just a tenant's survival window — it is a cause of the counts themselves.

Most of what you have just read rests on data that does not collect itself. Eviction records arrive as a mess — sealed in one state, address-free in another, sitting as scanned images in a third. The Eviction Research Network exists to turn that mess into something a researcher, an agency, or an advocate can actually use, and to do it the same way in every state so the numbers can be compared.

Three things make ERN's reading of eviction different in practice: