🏛️GovAuctions
← Surplus Price Index

How the Surplus Price Index is built

This page does not change when the numbers do. It describes what the index measures, what it does not, and the specific ways a reader can get it wrong. Every monthly release links here.

What it measures

The median final sale price of completed government surplus auction lots, by category, for one market. Governments dispose of surplus in public and nobody aggregates the result, so there is no official series to compare this against; it is built from the sales themselves.

Source and window

Every lot we record as completed on the official platforms agencies actually sell on, over a trailing 365-day window. The archive began in May 2026, so until May 2027 the window holds everything recorded since then. A long window keeps category samples healthy at the cost of responsiveness: a figure here describes months of the market, not this month's.

Prices are hammer prices. They exclude buyer's premium, sales tax, transport and any resale fee, which on a real purchase are substantial. A lot whose outcome the platform does not publish is excluded entirely rather than assumed either way.

Flat-rate lots, and why they are disclosed rather than removed

Some agencies clear routine lots at a fixed administrative price. Where one exact price dominates a category, that category's median is partly a disposal policy rather than a market outcome, and no amount of like-for-like matching catches it: the flat-rate lots span every item type.

So each release reports, per category, the single most common exact price and its share of lots (modePrice and modeShare), and flags any category above 10%. They are not stripped out, because they are real sales and removing them would publish a median of a market that does not exist.

In the current US release (US-2026-09), 2 categories are flagged: office furniture lots (14% at $1), military lots (21% at $5).

Why there are no movement figures

A category median is a statistic about what sold, not a price. Comparing two medians compares two baskets, and with a corpus this young the basket moves more than the market does. Our own first attempt at comparative headlines produced three confident claims and all three were artifacts: a state-versus-state gap that halved once bulk pallets were matched against single machines, a “101% jump” that was a two-point comparison across a dip, and a 163% rise over a basket that changed from racking to cabinets.

So releases publish levels only. Movement will be published when the corpus has six or more complete months and a change can clear like-for-like matching, direction agreement across sub-buckets, a volatility gate against that series' own history, and the flat-rate check above. Until then the honest answer is that we cannot support the claim, and a release that says so is worth more than one that guesses.

What a release does not tell you

  • What a particular item is worth. A pooled median is not a valuation.
  • Whether prices rose or fell. See above; that figure is withheld, not zero.
  • How states compare. Their baskets differ sharply, and some are dominated by flat-rate disposal.
  • What the government sold in total. We cover the platforms we aggregate, not private or purely local disposal.

Versions

Each monthly release is frozen at publication and never edited. If a figure is later found to be wrong, the correction appears as a new release with a changelog line, so a citation made in good faith stays checkable.

Reuse

Released figures are free to use with attribution to GovAuctions.app and a link to the release. The machine-readable form of every published number, each with its sample size and method, is at /research/facts.json. Questions about the data are welcome via contact.