The documents behind the rows
Documents held here in full rather than summarised. Two are studies on how long housing takes to deliver — the material the register rows were drawn from, published so a reader can check one against the other rather than take the summary on trust. One is this project’s own working thesis on the commercial question, published for the reasoning rather than the numbers. None of the three is a graded claim.
Every figure this project stands behind is a row in the evidence register with a source, a grade and a review date, and the ones taken from these studies are argued on what the world publishes. The documents themselves assert a great deal more than that, on another project’s tiering and confidence scale. They are published for the same reason the closed papers are published uncorrected: a reader who wants to check a row needs the thing it came from, not a description of it.
Six defects were corrected in both before they were published here, found by a review that checked them against primary source files rather than against the sessions that wrote them. Each correction is recorded against the unedited original in the knowledge base, so the repair is checkable rather than asserted.
Four articlesRead them, or take the data
Global · August 2026
Separates the time spent getting permission, the time spent waiting after permission, and the time actually spent building — then finds that no country in it publishes all three as one regular series. Country sections, comparability warnings, a methodology note on what would close the biggest gaps, and the full dataset as a sortable table at the foot.
Published in its own presentation rather than restyled, because it is a document rather than a page of this site.
United States · August 2026
Anchors every state to its Census region and applies no state-level adjustment to construction time, because no published evidence supports one. Sections on why the ranking everyone asks for is a trick, where permitting variation is real, the owner-built gap, and a methodology appendix listing what would replace the models with measurements.
Arrived as plain text and is set here in the site’s own typography.
This project · August 2026 · pre-validation
The constraint moved. It is no longer what can be built — it is what actually gets delivered. Zoning interpretation is being commoditised at national scale, and Arizona has already made ADU placement by-right, so the layer worth holding is narrower and further down than the previous revision argued. Sheets on what the ROAD Act changed, the competitive landscape, the HUD-code lane and its ceiling, a bottom-up market model, unit economics, and a capital plan.
This one is ours, and it is the least settled thing on the site: no customer interviewed, no parcel screened, no revenue earned. Every figure in the model, the economics and the capital plan is a stated assumption rather than a measurement, and none of it is graded. It ends on what is missing and on what would kill the thesis, which is the part worth reading first.
This project · August 2026
Neither study above could be assembled from a single source, because no authority publishes the whole clock. This is that gap measured rather than asserted: what three phases a delivery time needs, what the two American states that compel the reporting actually produce when their own filings are joined, and the four different ways a published permitting figure stops short of the wait it appears to describe.
Unlike the two studies above, every figure on it is this project’s own and traces to a graded register row. It closes on what would shut the gap, which needs no new survey.
The two commissioned studies sit outside the figure gate that governs the
rest of this site, because that gate requires every number on a page to trace to a register
row and these pages are full of numbers this project has not verified. That exclusion is
written down rather than incidental: check_site.py refuses to build if a document
is hosted here without a stated reason, and the withdrawn-figure sweep still reads both of
them like any other page.
Three doorsWhat was made of this
01 · the argument
The graded claims drawn from both documents, with the counter-evidence carried in the body rather than a footnote.
→ /international
02 · browse
Every figure this project asserts, with a source, a grade and a review date.
→ /evidence
03 · downloads
Both datasets, the register as a CSV, and the closed papers with their errata.
→ /documents