Septic System Statistics by State and County

They differ by unit, universe, definition, method, and year. The AHS counts occupied households that report septic. The USGS model estimates residential tanks from parcels and footprints. EPA uses a broader system definition. And the 1990 Census is a large sample that's now a generation old. None of these is directly interchangeable with another.

We built the matrix below so a writer can see exactly where two “septic statistics” diverge instead of guessing. Where a source doesn't state something, we mark it not stated rather than fill in an assumption.

Unit of measurement.A household survey and a tank model aren't measuring the same object. The AHS asks a resident how their home disposes of sewage; the USGS model estimates how many residential septic tanks sit on the landscape, using inputs a household survey never touches.

Definition.EPA's “more than one in five” deliberately includes small-community cluster systems alongside individual tanks. The AHS category is “septic tank or cesspool,” which folds in cesspools — which EPA doesn't even consider adequate treatment. Narrow the definition and the number falls; widen it and it rises.

Method and coverage. Maxcy-Brown and colleagues argue that the AHS lacks an extensive rural sample and, for returning households, can ask a respondent to confirm a previously recorded septic-or-cesspool answer rather than re-classifying from scratch. They separately note that some federal measurements count clusters of six or more homes on a shared system as public sewer. Those are the mechanics behind their view that the survey runs low.

Vintage.The 1990 Census is the only nationwide benchmark of its kind, and it's old enough to drive. Any state percentage traced to it — including the historical column in the state table below — describes the country as it was, not as it is.

One honest note on the survey itself: the 2023 AHS estimate (18.5%) is higher than the 2021 estimate (15.2%), and the published tables don't establish why it moved. What Census does say is that the 2023 survey had a 59.15% unit response rate and that nonresponse may reduce the efficacy of its weighting adjustments — which is exactly the kind of caveat a careful reader should keep attached to the number.

By modeled tank count, Texas leads with about 2.56 million residential septic tanks, followed by North Carolina (1.92 million) and Pennsylvania (1.54 million). By density, the Northeast dominates: Connecticut has the most tanks per square kilometer, then Rhode Island and Massachusetts. Count and density answer different questions — big states rack up big totals, while small, unsewered states pack tanks tightly.

The distinction matters more than it sounds. Texas has the most tanks because Texas is enormous, not because it's unusually septic-dependent. Connecticut has the highest density because a large share of its homes sit on septic in a small area. A writer covering national scale wants the count; a writer covering local concentration wants the density.

One guardrail we hold firmly: we do notpublish a “current septic share” for each state by dividing modeled tanks into modern housing counts. Those two datasets aren't built to be combined that way, and doing it would manufacture a precise-looking number with no valid basis. The 1990 shares appear for historical context only, clearly labeled. As an independent cross-check on the model, California's State Water Resources Control Board estimates about 1.2 million onsite systems statewide, and our USGS central estimate for California is about 1.11 million, which is similar in magnitude, though the two measures use different definitions and methods. The full 48-state-plus-D.C. table, with lower/central/upper scenarios and both ranks, is in the downloadable dataset.

Our county rollup covers 3,108 counties and county-equivalents in the conterminous United States, each with lower, central, and upper modeled tank counts, density, and national rank. By count, Worcester County, Massachusetts leads with about 91,458 modeled tanks; by density, Island County, Washington is highest at about 41 tanks per square kilometer.

County-level is where this dataset earns its keep. National and state numbers are easy to find; a sourced, county-by-county septic estimate is not.

The full 3,108-county file is in the download section, joinable to standard Census county boundaries by five-digit GEOID.

Our USGS rollup estimates about 904,483 residential septic tanks in South Carolina (range 793,989–1,009,987), ranking the state 18th in count and 12th in density. Separately, the South Carolina Department of Environmental Services reports that roughly 30% of the state's households rely on septic. These are different measures — a modeled tank count and an agency household share — and both belong on the record.

South Carolina is a clean case study in why you report sources side by side rather than picking one. Three credible figures describe septic here, and they don't collapse into a single number.

A few things to handle carefully. The 40.6% (1990) and roughly 30% (SCDES, current) figures are separate rows because they differ in year, source, and method — and the difference does not prove a precise decline. The 17,336 figure is applications for individual systems received in 2023–2024, not permits issued, not installations, and not the total number of systems operating in the state; SCDES reports application volume rose 132% from 7,456 a little over a decade earlier, which the agency says highlights rising demand as South Carolina experiences population growth and development. Treat it as demand, not as an installed base. And because SCDES publishes a household share rather than a statewide system total, if you need a count, the ~904,000 modeled tanks is the sourced figure, with its range attached.

There is no verified, current, nationwide septic failure rate. The most reproducible recent national measure we found is from the 2023 American Housing Survey: an estimated 396,000 occupied housing units using a septic tank or cesspool — about 1.6% — reported a completely unusable sewage-disposal breakdown in the prior three months (90% margin of error ±70,000). The widely repeated claim that “10–20% of septic systems are failing” is not a current-prevalence statistic and should not be quoted as one.

This is the statistic most often stated wrong, so it's worth slowing down. The “10–20%” figure traces to EPA language describing systems that fail at some point over their operational lifetimes— not systems failing right now. Turning a lifetime-failure figure into a present-tense “one in five septic systems is failing today” is a real and common error, and it inflates the problem by conflating two very different measurements.

What can be said with a source is narrower and more honest. The 2023 AHS estimate — 396,000 septic or cesspool homes, roughly 1.6%, reporting a completely unusable breakdown in a three-month window — is a bounded, dated measure. It is not a total failure rate: the survey question captures completely unusable sewage-disposal breakdowns over a short window, and EPA separately describes groundwater and surface-water impacts from improperly managed or malfunctioning systems, which the survey doesn't measure. But the AHS number is real, and it's quotable.

None of this means failures don't matter. EPA is clear that malfunctioning septic systems can contaminate groundwater and surface water and release pathogens and nutrients into the environment. That significance stands on its own — you can convey it accurately without attaching an unsupported nationwide prevalence number to it.

This page is a reconciliation and an aggregation, not a new survey. It shows the range of credible septic estimates, a state and county model, and the reasons the numbers differ. It does not show a single “true” count, individual system locations, current system condition, or a validated current household share for every state and county.

Stating these limits plainly isn't a hedge — it's what makes the rest of the page trustworthy. A source that tells you what it can't prove is a source you can rely on for what it can.

Terms below are drawn from the American Housing Survey (AHS), the USGS model, or created as editorial labels for this aggregation — we note which is which.

Every figure on this page traces to one of these primary or peer-reviewed sources. Retrieval date for all: July 20, 2026.