Method and sources

Where the data comes from

Each covered state publishes a statewide parcel layer: one record per parcel, with the owner's name and mailing address, the property's own address, land and building values, and a land-use code. CountyLedger reads these records, classifies each parcel, and keeps only the county totals. Names and addresses are used for the classification and never leave it.

• North Carolina: NC OneMap's statewide parcels, which combine the records of all 100 counties and are updated weekly.
• Florida: the Florida Statewide Cadastral layer, built from the county property appraisers' tax rolls.
• Massachusetts: MassGIS property tax parcels, from each city and town's assessor.

Every county page shows the edition date of the records it was counted from.

The measures

Each share has its own denominator, because no county publishes every field for every parcel. The county page shows both numbers, for example "54,737 of 442,031 parcels with a known mailing state".

Out-of-state owners

Parcels whose owner's mailing address is in another state, out of the parcels with a known mailing state. A mailing address is where the tax bill goes, which is usually but not always where the owner lives; a trust or a management company can make an in-state owner look out-of-state, or the reverse. Government owners are left out of both the count and the total it's measured against, and show up under publicly owned instead: in Monroe County, Florida, National Park Service land billed to Washington, DC would otherwise be nearly half of the out-of-state parcels.

Vacant lots

Parcels with land value and no building, out of all parcels. "No building" depends on what the county publishes: a building value of zero, or a record saying there's no structure. A parking lot, a farm field and a lot waiting for a house all count as vacant.

Inherited (estate or heirs)

Parcels whose owner name marks an estate or heirs, such as "estate of", "heirs" or "deceased", out of the parcels with an owner name. "Et al" (and others) and life estates don't count; they usually mean co-owners or estate planning, not an inheritance. This undercounts: many inherited parcels stay in the late owner's name with no marker at all.

Owned by individuals

Parcels whose first owner name doesn't look like a company, trust, church, association or government body. Name matching is rough; a family trust counts as an entity, and an unusual company name can count as a person.

Publicly owned

Parcels whose owner name is a federal, state or local government body, out of the parcels with an owner name. That includes public universities, housing authorities, government-owned utilities such as Florida's water management districts, and homes held by HUD or the VA. It doesn't include tribal land, or Fannie Mae and Freddie Mac, which are companies. An agency recorded under an unusual name can be missed. This counts parcels, not acres: a national park can be a few large parcels and a town's lots many small ones.

Absentee owners

Owners whose mailing address isn't the property itself: every out-of-state owner, plus in-state owners whose mailing address has a different house number or street, or is a PO box in another ZIP code. For in-state owners this needs both addresses, and some counties publish property addresses for only part of their parcels. So the in-state rate is measured on the parcels where both addresses could be compared and applied to all in-state parcels; the county page says how many parcels the estimate rests on. It's an estimate, not a count. A second home, a rental and a lot owned by the farmer down the road all look the same here, and an owner who gets mail at a PO box in the same ZIP code isn't counted as absentee. Where fewer than half of a county's in-state parcels can be compared, or where the estimate comes out above 90% (most likely because local owners use PO boxes that can't be matched), the page shows no figure. Government owners are left out here too.

Where out-of-state owners live

For each county, the out-of-state parcels are grouped by the state of the owner's mailing address. The ten largest are listed; the rest are summed. Military mail (APO/FPO) and addresses outside the US, or with a state code that isn't one, are shown as their own lines.

Housing context

Where it's available, a county page also shows HUD's fair market rent for a two-bedroom home and how it changed over five years, the median home value, price-to-rent, and the change in the county's housing units from 2020 to 2025, from the Census Bureau's housing unit estimates. HUD sets one fair market rent for a whole metro area, so neighboring metro counties often share a figure; it's a standard for housing vouchers, close to the 40th percentile of rents, not a median.

Rent change compares HUD's two-bedroom rent now with five years earlier. HUD sometimes resets a county's rent by a large step in a single year, often in small counties or when it moves a county into a different metro area, so a big change may not match what rents actually did. Where a county's rent change is more than twice that of its state's median county, the page says so.

Median home value is the Census Bureau's American Community Survey figure: what owners say their home is worth, averaged over five years of surveys (2020-2024). It lags the market when prices are rising; in the North Carolina counties we checked, it ran 3 to 24 percent below Zillow's figures. Price-to-rent divides that value by twelve months of the HUD rent. Both inputs are approximate, so read it as a way to compare counties, not as what a particular home would earn. A low figure in a rural county can simply mean HUD gave it a metro area's rent.

Neighboring counties

Each county page compares the county with the counties that border it, from the Census Bureau's county adjacency file. Neighbors in states CountyLedger doesn't cover yet are named below the table.

State totals and ranks

A state's share is its counties' counts added up over their denominators added up, not an average of county shares, so a large county weighs more than a small one. Counties with no data for a measure are left out of that measure's total and ranking, and counties with the same value share a rank.

When there's no data

"No data" means the county's records can't answer the question. It never means zero. When fewer than half of a county's parcels carry the field a measure needs (a mailing state, an owner name, something that tells a built lot from an empty one), the share is left blank rather than computed from the few that do, and the county page says why. A county whose records answer none of the questions has no page; the state page names it.

What the numbers can't tell you

• Who an owner is, or why they own the land.
• Whether a parcel is rented, used as a second home, or empty.
• Anything about value. States publish market, appraised or assessed values, which can't be compared across states, so CountyLedger doesn't use them.
• Exact comparisons across states. Each state's records have different gaps, and the vacant-lot rule differs where states publish different fields.

Refreshes

Pages are rebuilt when new records arrive. The footer shows when a page's inputs last changed.