Methodology

Rate Gazetteer summarizes published property tax information so you can compare places on a like basis. We are not a tax preparer, law firm, or government agency. Use official notices, bills, and offices for appeals, payments, and legal decisions.

How we obtain rates

Where possible, figures are built by manually reviewing primary materials from the taxing jurisdiction: county property appraiser or assessor sites, tax collector or treasurer notices, consolidated millage or rate sheets, sample TRIM notices, and other official publications. We prefer documents that show the rate stack for a representative residential parcel or column for a locality.

When a single official PDF or web table is the authority (for example, a state-prescribed consolidated millage form), we cite that source and describe which locality or column applies. Some states publish rates in mills per dollar of assessed value; others emphasize effective rates on market or assessed value. We state the assumption on each page so the headline numbers stay interpretable.

When we cannot yet tie a figure to a specific official line item, we may still publish a structured placeholder so the layout and citations are ready—those rows carry an explicit evidence label (see below) until they are replaced with verified numbers.

Benchmark value and effective rate

Each county page includes a benchmark dollar amount (for example, a typical home value or a rounded assessed value used in official examples). The modeled annual tax is the tax amount associated with that benchmark under the stated assumptions—not a median for every home in the jurisdiction.

The effective rate is that modeled annual tax divided by the benchmark, expressed as a percentage. It lets you compare jurisdictions without pretending every state values property the same way; the page text explains when the benchmark is meant as taxable assessed value rather than market price.

Millage and levy components

Many counties publish a millage (mills per thousand dollars of taxable value) or a rate table that splits the bill among schools, county general government, municipal, special districts, and debt service. When we have that detail, we show it as structured line items so you can see which portion of the stack typically goes to schools, roads, and other levies, as described in the source—not as a guarantee for your parcel.

Flat fees, non–ad valorem charges, and parcel-specific specials often sit outside a simple mill total; we call those out in the notes when the source does.

Interactive estimate on each page

County pages include a slider and value field so you can scale the published model to a home value you choose. That produces an illustrative annual tax using the same proportional relationship as the headline benchmark—not a certified estimate for your property.

State hub pages use a population-weighted average of comparable effective rates across indexed rows for that state. In most states, each row’s rate is implied from U.S. Census ACS county medians (median tax ÷ median home value). In six states we instead use primary-source county models—see Modeled primary sources below. Hubs are useful for orientation, not for choosing a final number for a specific address.

Where we publish structured exemption notes, some county pages also offer optional checkboxes or radio groups that apply rough illustrative multipliers or flat credits on top of the scaled tax. Those controls only appear when we have entered modeled parameters in our dataset; they are not a substitute for your TRIM notice or appraiser eligibility screens.

In Florida, veterans with a qualifying total and permanent service-connected disability (and certain surviving spouses) who receive the statutory exemption owe no ad valorem property tax on that homestead statewide—only non–ad valorem lines may remain. Where we model that scenario, the tool sets illustrative ad valorem tax to zero.

Modeled primary sources (six states)

Alabama, Alaska, Arizona, Arkansas, California, and Florida use other primary materials on this site (not ACS county medians) for county hub figures. Before population-weighting hub averages—or ranking those rows against ACS states—we apply the same market-comparable rules everywhere: Arkansas uses annual tax ÷ $350,000 (the same residential market illustration as county pages); Arizona uses the same $350,000 illustration in the denominator (modeled tax from ADOR’s county-average $/100 rate applied to Class 3 net assessed = 10% of that LPV); Florida and California multiply taxable-benchmark rates by 0.72 and 0.58 respectively for the same purpose. Alabama and Alaska already express comparable effective rates on each county’s disclosed benchmark in our model—no extra multiplier is applied in that layer. County pages always show the underlying benchmark and citations; these factors are for interstate comparison only.

Where the Florida and California orientation factors come from

The constants 0.72 (Florida) and 0.58 (California) are not a single federally published “true ratio” for every parcel. They are fixed, site-wide illustrative bridges we use after reviewing how each state’s published rates relate to benchmarks on our county pages. Florida and California pages often use taxable assessed values that commonly sit below current market value for residential property under assessment caps, exemptions, and acquisition-value rules. Arizona and Arkansas instead put modeled tax over the same $350,000 residential illustration (see above)—no extra multiplier in that ranking layer.

Arizona: residential Class 3 net assessed value is commonly 10% of limited property value under Title 42 (parcel variation is large). County pages multiply that net assessed benchmark by ADOR’s combined $/100 rate; for interstate tables we use modeled tax ÷ $350,000, matching the Arkansas layer.

Florida: taxable value after Save Our Homes and exemptions is frequently a large fraction of, but not equal to, just value; 0.72 is a conservative taxable-to-market orientation for interstate tables, not your appraiser’s card.

California: BOE-style burdens on net taxable assessed value often understate burden as a share of current market for long-held homes. 0.58 serves the same ranking purpose: a disclosed, revisable assumption—not precision engineering.

Values are defined in source code (comparableEffectiveRatePercent.ts) and change only when we publish a deliberate methodology update.

Census ACS counties vs. modeled primary-source counties

ACS-based rows reflect survey medians: median annual property taxes and median owner-occupied home value in the published county tabulation, so the implied rate is a reported burden snapshot for that geography and vintage—not a mill levy rebuilt from line items. Modeled rows reflect estimated tax on disclosed benchmarks under standardized rules and, where needed, the orientation factors above.

Rankings that combine both are still useful for orientation, but they do not pretend one identical instrument measured every state. Treat small gaps between neighbors (especially at the top of a ranking) as sensitive to definition, vintage, and weighting.

Evidence types we show

Each page’s source block includes a short evidence label so you know how the figure was sourced:

Confirmed: We matched the numbers to an official government publication, rate sheet, or notice from the taxing authority (or a state form that consolidates local levies) on or before the “last verified” date shown.
Estimated: The figure is compiled from indirect materials—secondary reporting, multiple partial sources, or a mix of documents where no single primary line item was used as the sole authority. Treat it as directional until replaced with a confirmed row.
Unknown: The row is illustrative, draft, or otherwise not yet tied to a clear verification path. Do not rely on it for financial or filing decisions; check the assessor, appraiser, or collector for your parcel.

Behind the scenes we track finer review states; on the public site they always resolve to one of these three labels so the evidence line stays simple.