One picture for all of it
Not to scale: days, or years.
1 Why this page exists
The Fed — America’s central bank — is the lender of last resort: the one who lends when nobody else will. Only the Fed can settle a lending panic. That is what this tab explains.
2 A house with pipes
The economy is a house with pipes. They carry money from people who have it to people who need it: lending. Fine for forty years, then one winter they aren’t, and you call the plumber, who here is the Fed; everyone holding dollars pays the repair bill.
3 Borrowing is a time machine
Nobody buys a house with cash. You borrow against income you haven’t earned yet, move in now, pay for thirty years. Debt is income pulled forward — it works only while income grows.
4 The long cycle: the house itself
Every five to ten years the house overheats and cools. But each wave ends with more debt than the last began with, so the debt line climbs. Early on it climbs slower than income, because the borrowing builds something; later faster, because new borrowing only pays for the last round. A day comes when payments eat so much income that an ordinary rate cut stops working. That is the long-term debt cycle: no committee decides it, the arithmetic does.
Last time that day came in 2008: the Fed began buying bonds on 25 November and still cut its policy rate to zero. Past dates, not a schedule.
5 The four stages, one story
Four words from the site header:
- Calm. No line is past its fire level — that is today.
- The leak. One of the eleven lines goes past its fire level — the number that means it has broken, like an alarm going off. Being near one counts for nothing.
- The plumber. The Fed answers: it lends cash, or it buys bonds.
- The repair bill (monetisation). The Fed has bought the debt; everyone holding dollars pays for it a little at a time.
When debt is heavy, something specific gives way: a funding market refuses to lend, a credit spread (the extra interest a risky borrower pays) blows out, a Treasury auction gets weak bids. Those are the eleven lines, the house’s pipes. A line sat past its fire level in 163 of 287 of 287 month-ends since 2003; 1 of 252 of 252 three-year stretches passed clean. The leak is near-certain; its date is not.
It leaked in August 2007: on 9 August BNP Paribas froze three funds it could not price, and interbank lending seized. Past dates, not a schedule.
6 What counts as the Fed answering
An answer means one of five listed things: a new lending facility or repo window — the Fed lending cash; a reopened or expanded swap line; resumed bond buying; a stated long-end Treasury operation; or an unscheduled cut of half a point or more. A scheduled meeting, a speech, an ordinary quarter-point cut — not answers.
In the ten episodes where the pipes broke, the Fed answered 6 of 10 — the uncertainty under every number here. Counted a stricter way the same history gives 7 of 14, and strictest 4 of 14. That spread, not one number, is the honest uncertainty.
7 How long the wait has been
Six episodes, oldest first: first trouble, then first facility.
- US, November 1930 → 6 March 1933: 850 days — the slowest.
- Japan, November 1997 → 30 March 1998: 124 days.
- US, 9 August 2007 → 12 December 2007 (the Term Auction Facility): 125 days.
- Euro area, October 2009 → 10 May 2010 (bond buying): 203 days; and → 6 September 2012 (the fix that worked): 1,053 days.
- US, 9 March 2020 → 17 March 2020: 8 days.
- UK, 23 September 2022 → 28 September 2022: 5 days — the fastest.
Five days to 850 days. That span, not one average, is why the forecast looks as it does. Past dates, not a schedule.
8 Then what this site measures
After the break comes the repair bill: the Fed buys the debt, and everyone’s purchasing power quietly pays.
The forecast is the chance of an official answer within twelve months: 24.0%, about one in five — band 11.1% to 39.0%. Even odds arrive in March 2029 — a percentile, not a date.
What this page is NOT saying
- Not “a crash is coming.” It forecasts one listed official act — not prices, not savings.
- Not a date. Even odds in March 2029 is a percentile, not a date; so are one in ten by May 2027 and nine in ten by April 2034.
- Not certain. 24.0% is a midpoint; the band runs 11.1% to 39.0%.
- Not a rescue. The Fed answered 6 of 10; otherwise the pipes stayed broken.
- Not advice. No trade, no signal. Today reads Calm — nothing past its fire level: a measurement, not a promise.