Guaranteeing low-deposit mortgages puts the Commonwealth behind the riskiest segment of the housing market. The obvious question is how that risk has performed, and the obvious temptation for the administering body is to report volumes and stay quiet about losses.

Housing Australia reports the risk in detail. As at 30 June 2025 loan arrears were 0.3 per cent and 0.9 per cent of borrowers had accessed hardship support. Two claims were paid in 2024-25, taking total claims paid since launch to three. Figure 15 plots hardship and arrears against every Reserve Bank cash rate decision from September 2022 to June 2025, and the text notes that arrears and hardship run proportionally higher under the Family Home Guarantee, where serviceability rests on a single borrower with dependents.

Two details lift this above routine disclosure. The report identifies which of its own products carry the most stress rather than reporting a portfolio average. And it forecasts against itself: "there may be a pickup in claims in 2025-26, consistent with the natural lifecycle of the loan book". Publishing a warning that your own numbers are likely to worsen is rare.

Read the annual Trends and Insights report. Six consecutive editions exist, so a series can be built, at the cost of transcribing six PDFs.

Opened and read on 2026-09-03, and the entry's headline figure was wrong.

It previously claimed "12 defaults on more than 240,000 guarantees since 2020", taken from reporting about Housing Australia rather than from the report. Neither "12" nor "240,000" appears in the 2024-25 document. What it publishes is three claims paid since launch, two of them in 2024-25, alongside arrears of 0.3 per cent and hardship at 0.9 per cent.

The distinction matters and the original phrasing blurred it. A claim is paid when a borrower defaults and the sale proceeds fall short of the loan. Claims are therefore a subset of defaults, and the scheme reports claims, not defaults. The question has been reworded to ask what the report can actually answer.

Recording this correction is worth more than the entry. The register was praising a body for a disclosure it had not read, using a number that body had not published, and the true disclosure turned out to be better than the invented one. A false win is as damaging to this project as a false gap, and harder to notice, because nobody complains about being praised.

Included partly to keep the register honest about the same publisher. Housing Australia does this well and reports Future Fund completions poorly, and both entries carry its name -- see How many homes has the Housing Australia Future Fund actually finished, and where are they?. Scoring bodies rather than holdings would obscure that, which is why publisher pages report a median as well as a mean.

The one criticism is format. A PDF a year means a reader who wants the series transcribes six documents, and there is no reason an arrears rate could not be released as data.