A glossary of eviction & displacement terms.

The U.S. Census Bureau's ongoing survey of households. ERN draws most of its demographic, income, and housing inputs from the ACS 5-year estimates, which pool five years of responses to produce reliable figures for small areas such as census tracts.

Standard income bands used to target housing need: Extremely Low Income (ELI, 0–30% of AMI), Very Low Income (VLI, 30–50%), and Low Income (LI, 50–80%). The HPRM models displacement risk separately for each tier.

The midpoint of household income for a county or metropolitan area, published each year by HUD. Affordability and program-eligibility thresholds are usually written as a share of AMI — for example, a unit "affordable at 50% of AMI."

A machine-learning method that sums many small regression trees to predict an outcome flexibly while quantifying uncertainty. ERN uses BART to estimate both displacement (EDR) and eviction (EER) risk from census predictors.

A small, relatively stable geographic unit defined by the Census Bureau, typically holding 1,200–8,000 people. Tracts are the primary unit of analysis for ERN's national models — roughly 83,500 of them cover the contiguous United States.

A thematic map that shades areas by the value of a variable (darker usually meaning higher). ERN's interactive risk maps are choropleths of census tracts or counties.

The combined Housing Precarity Risk Model score: the sum of the displacement (EDR) and eviction (EER) sub-scores. It runs from 0 (no measured risk) to 8 (extreme risk on both dimensions).

Spending more than 30% of household income on housing; spending more than 50% is severe cost burden. Cost-burdened renters have little cushion against a rent increase, which raises their exposure to displacement and eviction.

The involuntary movement of residents out of their home or neighborhood because of forces they cannot control — rising rents, eviction, redevelopment, or the loss of affordable units. ERN distinguishes hard and soft displacement.

An HPRM classification of how a neighborhood's two risks line up: Convergent (both displacement and eviction risk present), EDR-Dominant (market displacement without much formal eviction), EER-Dominant (formal eviction without measured out-migration), or Neither.

The HPRM sub-model for soft displacement: the predicted net out-migration of low-income renter households driven by market pressure such as rising rents and neighborhood change. It is estimated with BART across three AMI tiers and binned into At Risk, Elevated, High, and Extreme.

The HPRM sub-model for hard displacement: a neighborhood's formal eviction pressure, measured as its eviction filing rate relative to the state average and modeled nationally with BART. It is binned into At Risk, Elevated, High, and Extreme.

The legal process by which a landlord removes a tenant from a rental unit. In data terms ERN usually measures eviction through court records, which capture formal filings and judgments but not informal ("self-help" or off-the-books) removals.

A landlord's formal start of an eviction case in court. A filing is not the same as a removal — many are dismissed, settled, or used as a debt-collection tactic — but filings are the most consistently recorded signal of eviction pressure.

The number of eviction cases filed per 100 (or per 1,000) renter households in an area over a period. Dividing by renter households makes filing activity comparable across places of different size.

A tract's eviction filing rate divided by its state's average rate, so 1.0 means "typical for the state" and 2.0 means "twice the state average." The EER model predicts this ratio to keep eviction risk comparable across states.

The court record an eviction filing creates, which tenant-screening companies use to deny future housing — even when the case was dismissed or the tenant won. Sealing laws hide or expunge these records so a single filing doesn't follow a tenant for years.

The process by which an influx of higher-income residents and investment raises property values and rents in a historically lower-income neighborhood, often pricing out long-term residents. Gentrification is one driver of displacement, not a synonym for it.

The unique identifier for a Census geography. A tract GEOID is an 11-digit FIPS code — 2 digits of state, 3 of county, 6 of tract. It is the key used to join ERN model output to other datasets.

Displacement through a formal, involuntary event — most often eviction, but also condemnation or the expiration of an affordability agreement. ERN proxies hard displacement with eviction filings, via the EER model.

The largest federal rental-assistance program, which pays part of a low-income household's rent directly to a private landlord. Voucher holders can still face displacement when landlords refuse vouchers or when local rents exceed the program's payment standards.

ERN's national, census-tract index of housing instability. It combines a displacement sub-model (EDR) and an eviction sub-model (EER) into a single 0–8 score that flags where low-income renters face the greatest risk. See the live model.

A protection that lets landlords end a tenancy only for specified, legitimate reasons — such as nonpayment or a lease violation — rather than arbitrarily. Just-cause laws curb no-fault and retaliatory evictions.

A classification of neighborhoods by their stage of housing-market and socioeconomic change — for example, stable, at-risk, actively gentrifying, or advanced gentrification. Typologies help match the right intervention to the right stage.

Predicted in-migration minus out-migration of households for a tract — the raw quantity behind the EDR model. A strongly negative NMR for low-income renters signals the out-flow consistent with displacement.

The written notice a landlord must give before filing to evict, and the minimum number of days a tenant has to respond. Length is set by state law — Washington, for example, expanded its pay-or-vacate notice from 3 to 14 days, a change ERN's disparity analysis helped support.

Long-term, affordable housing paired with voluntary services — case management, health care — for people with disabilities or histories of homelessness, with no requirement to "earn" housing through treatment compliance. A core Housing First model.

The federally backed 1930s practice of grading neighborhoods for mortgage risk and denying credit to areas with Black and immigrant residents. Its maps drove decades of disinvestment; ERN research links historic redlining to present-day eviction and displacement patterns.

A household that rents rather than owns its home — the population at risk of eviction, and the denominator for filing-rate measures. The HPRM only maps tracts with at least one renter household.

A policy guaranteeing tenants legal representation in eviction proceedings, where landlords are usually represented and tenants usually are not. Jurisdictions that adopt it see markedly higher rates of tenants staying housed.

Repeated eviction filings against the same tenant at the same address, often month after month, used by some landlords as a routine rent-collection tool. Serial filings inflate filing counts and expose tenants to court costs and record damage even without a removal.

Displacement through market pressure rather than a single legal event — when rising rents and neighborhood change push low-income households to leave or keep them from moving in. ERN models soft displacement with the EDR.

The body of laws and policies that shift bargaining power toward renters — rent regulation, just-cause rules, right to counsel, record sealing, and longer notice periods. ERN evaluates where such protections are strongest and weakest.

The formal legal term, in many states, for an eviction lawsuit — a landlord's court action to recover possession of a property. Court "UD" records are a primary raw source for ERN's eviction data.