Water loss, as filed
Every public water system in Texas and Massachusetts files what it produced and what it billed for. Almost nobody reads those filings. We read all of them, check whether the numbers add up, and put a dollar figure on the gap.
- Systems covered
- 2,987 Across 11,657 filed utility-years. Every one gets a page, a percentile, and a source.
- Figures too unstable to act on
- 402 Systems whose reported loss moves more than three quarters of their state's do. Real losses do not change that fast — these are describing measurement, not water.
- Filings that do not balance
- 1 Of 9,744 filings where a component balance is required, the parts do not sum to the reported whole beyond rounding.
Why this exists
A percentage on its own tells a utility nothing. Fifteen percent sounds bad and is better than half the state; nine percent sounds fine and is worse than most of its neighbours. Every page here says where a system sits among the ones that filed the same year, whether it has been over its state's standard for one year or for all of them, and whether its own reported figure is stable enough to act on. None of those questions can be answered from a single filing, which is why nobody publishes them.
A water loss audit is an identity, not an opinion: everything that entered the distribution system left it as billed consumption, authorized unbilled use, apparent loss, or real loss. When a filing's own components do not sum to its own reported input, that is checkable, and it is checkable by anyone. Regulators collect these filings; nobody publishes the arithmetic.
Three verdicts, not two. A filing either closes, fails to close by a stated amount, or does not publish enough to be checked at all. The third is the most common and the most useful, because a check that cannot fail is not a check — and a dashboard that reports it as a pass is worse than no dashboard.
Coverage
- Texas — 2,706 systems. Median loss 17.9%, 90th percentile 44.0%. 0 above the standard every year, 335 reporting figures too unstable to act on.
- Massachusetts — 281 systems. Median loss 12.0%, 90th percentile 27.0%. 68 above the standard every year, 67 reporting figures too unstable to act on.
Methodology and definitions
- Apparent loss — Water that reached a customer and was never billed — the component of water loss that is recoverable revenue inside a budget year rather than a capital project.
- Non-revenue water — Water a utility produced and paid to treat but never billed anyone for, split into the part that is a revenue problem and the part that is a cost problem.
- Real loss — Water that escaped the distribution system before reaching a customer, valued at what it cost to produce rather than at what it would have sold for.
- Unaccounted-for water (UAW) — The Massachusetts Water Management Act performance metric for distribution loss, and how it differs from the AWWA non-revenue water framework used elsewhere.
- How the water loss index is built — Every source, every computation, and every place the index refuses to produce a number — including why most filings are marked unverifiable rather than passing.
Download the whole dataset as CSV or JSON. Attribution appreciated, not required.
Is your system on here?
If the numbers look wrong, they are wrong in your filing or wrong in our reading of it. Either one is worth a conversation, and the first one is worth a conversation with your finance director.
Email us The Recovery Snapshot