An inverted yield curve does not forecast a downturn from the sidelines. It charges a running fee to every institution that funds itself short and invests long, and the rate cuts it appears to predict often arrive because one of those institutions can no longer pay.
The summer of 2007 makes the point better than any model. The Federal Reserve had held its policy rate at 5.25% since June 2006, and the 10-year Treasury yield spent most of the following year below that level. On paper, the bond market was betting that the Fed would eventually cut.
Something more concrete was happening in the plumbing. Structured investment vehicles and bank-sponsored conduits had financed long-dated mortgage securities with asset-backed commercial paper, short-term IOUs that had to be rolled every few weeks. With short rates at or above long rates, the spread those vehicles lived on had thinned to almost nothing. Any wobble in the value of their collateral left them with no cushion at all.
On August 9, 2007, BNP Paribas froze three funds whose assets it said it could no longer value. Commercial paper buyers stepped back, conduits could not roll their funding, and sponsoring banks were forced to absorb the assets. The Fed cut by 50 basis points in September, and by December 2008 the policy rate sat in a range of zero to 0.25%.
The inversion had been right, but not because bond traders foresaw the housing bust. It was right partly because the shape of the curve had made the most leveraged carry structures fragile, and when they broke, the Fed had to cut.
What the curve is actually made of
A long-term yield is two things stacked together. The first is the expected policy path, the average short-term rate the market expects over the life of the bond. The second is the term premium, the extra yield investors demand for locking money up for ten years instead of rolling short bills: compensation for inflation surprises, supply shocks, and the plain discomfort of holding duration.
10-year yield ≈ average expected short rate over 10 years + term premium
The slope of the curve is the difference between that sum at the long end and the same sum at the short end. The term premium on a 2-year note is small, so the 2-year yield is close to a pure read on where policy is headed over the next two years. The 10-year carries the full weight of the premium.
Why does this matter? Because inversion can come from either ingredient. The textbook route runs through expectations: the market judges current policy too tight to last, so the expected average short rate over ten years falls below today's short rate. The second route is a collapse in the term premium, which flattens the curve even when nobody expects cuts at all.
The 2006 to 2007 inversion had both. In February 2005, Alan Greenspan told Congress it was a "conundrum" that long yields had barely risen while the Fed kept hiking; heavy foreign reserve buying of Treasuries was squeezing the term premium. By the time expectations of easing took hold, the premium was already thin, so only a modest bet on future cuts was needed to push the 10-year below the funds rate.
The practical consequence for you: the slope is never a single fact. Before reading anything into an inversion or a steepening, establish which of the two components moved.
Why inversion taxes maturity transformation
Maturity transformation means borrowing at short maturities to fund long-maturity assets. Commercial banks do it with deposits and loans. Mortgage REITs do it with repo funding and mortgage-backed securities. The 2007 conduits did it with commercial paper. Each of them earns, roughly, the long yield minus the short funding cost, which makes the curve's slope their gross margin.
When the curve inverts, that margin compresses toward zero or turns negative, and these actors face three bad choices. They can shrink, which means less lending into the real economy. They can tighten credit standards, which has the same effect more slowly. Or they can reach for longer duration or riskier assets to rebuild the lost spread, which leaves them exposed to exactly the price shock that eventually arrives. The first two choices slow growth. The third sets the trap.
Silicon Valley Bank walked into the third option. It had invested a surge of deposits in long-dated Treasuries and agency mortgage securities while yields were low. As the Fed hiked through 2022, those holdings lost market value, and depositors could earn far more in money market funds than the bank paid them. The short funding repriced; the long assets could not. When the bank sold securities at a loss to meet outflows, confidence collapsed, and regulators closed it on March 10, 2023. The 2-year/10-year spread was inverted by roughly a full percentage point through much of that year, reaching about minus 108 basis points in July.
The actor matters here, not the abstraction. The curve does not tighten credit by itself; specific balance sheets with specific funding mismatches do, under pressure from the curve. What this means for you: when the curve inverts, find out who is carrying long assets on short money, because that is where the break is most likely to come from.
The dollar channel that pushes back on the curve
The cross-asset transmission that gets the least attention runs through currency hedging. A Japanese insurer or a European pension fund buying a 10-year Treasury usually hedges the dollar exposure with FX forwards. Under covered interest parity, the no-arbitrage rule linking forward exchange rates to interest rates, the cost of that hedge roughly equals the gap between US and home-country short rates.
Hedged yield ≈ US 10-year yield − (US short rate − home short rate)
Take illustrative numbers in the spirit of 2023: a US 10-year at 4.0%, US short rates at 5.3%, and a home short rate near zero, as Japan's was then. The hedged yield works out to 4.0 − (5.3 − 0.0) = −1.3%. A deeply inverted US curve makes Treasuries a money-losing purchase for a hedged foreign buyer.
That buyer then either stops buying or buys unhedged, which ties Treasury demand to currency moves. Fewer hedged foreign buyers means a higher term premium is needed to clear supply, so inversion can contain some of the pressure that later steepens the curve. We would treat this as one force among several rather than the dominant one, since the size of foreign demand shifts is hard to pin down in real time.
Carry flows can also snap. In early August 2024, as the Bank of Japan tightened and markets priced US cuts, positions funded in cheap yen unwound violently, and the Nikkei fell 12.4% on August 5. A rate-differential trade that looked like a currency story became an equity crash in a single session. For you, the lesson is to watch hedged yields for the big foreign holders, not just the nominal yield on the screen.
Reading a 41-basis-point slope with the 10-year at 5.29%
As of September 30, 2026, the 10-year Treasury yielded 5.29% and the 2-year yielded 4.88%, a positive 2-year/10-year spread of 0.41 percentage points. The curve is upright, and a headline reader would call that normalization. The decomposition says it can mean two opposite things.
Start with the forward rate, the yield the curve implies for a future period. A simple linear approximation of the 8-year rate starting two years from now:
2y8y forward ≈ (10 × 10-year yield − 2 × 2-year yield) / 8
(10 × 5.29 − 2 × 4.88) / 8 = (52.90 − 9.76) / 8 ≈ 5.39%
Now strip out the term premium. We do not have a model estimate for it here, so test two assumptions. If the 10-year carries a 0.5 point premium, its expectations component is 4.79%, and the implied average short rate for years three through ten is (47.90 − 9.76) / 8 ≈ 4.77%. That is a nearly flat policy path sitting just below the 2-year. If the premium is 0.8 points instead, the expectations component is 4.49%, and the years-three-to-ten average falls to (44.90 − 9.76) / 8 ≈ 4.39%. That is a meaningful easing cycle priced in.
Same slope, opposite stories. Under the first assumption, the entire 41 basis points is term premium wrapped around a flat policy outlook. Under the second, the market expects cuts and is also demanding more pay for duration.
There is a second distinction. A bull steepening happens when short yields fall faster than long yields, typically because the central bank is cutting into weakness. A bear steepening happens when long yields rise faster than short ones. A positive slope with the 10-year above 5% looks more like the second kind, which shifts the pressure on maturity transformers from funding margins to mark-to-market losses on long holdings. The fee is being collected through a different window. What this means for you: a positive slope does not retire the risks the inversion created if the long end got there by selling off.
When the inversion signal misleads
The mechanism has clear failure modes, and each one has a recent example.
The first is term premium distortion. When a central bank buys long bonds through quantitative easing, or foreign reserve managers buy them for policy reasons, the premium is suppressed and inversion becomes cheaper to achieve. A curve that inverts mainly because the premium collapsed is saying less about future cuts than one that inverts because expected short rates fell. Greenspan's conundrum is the template.
The second is the lag. The 2-year/10-year spread inverted in July 2022 and stayed inverted for more than two years. Anyone who sold risk assets on the first day of inversion watched the S&P 500 rise more than 24% in 2023. The curve can identify stress correctly and still be useless for timing.
The third is that breaks do not always produce cuts. When Silicon Valley Bank failed, the Fed did not ease; inflation was still too high. It created the Bank Term Funding Program, which lent against Treasuries and agency securities at par value, and it kept hiking. A liquidity facility absorbed the break that, in another inflation regime, might have forced cuts. The curve's message about fragility was accurate. Its message about the policy path was overridden.
The fourth trap is reading un-inversion as an all-clear. Steepening after a long inversion has often coincided with the start of trouble rather than the end of it, because a bull steepening frequently reflects a central bank cutting in response to something that has already broken. Treat that pattern as a signal to examine, not a rule: the 2024 re-steepening did not fit it neatly, which is exactly why the component decomposition matters more than the sign of the spread.
What to do with the curve
The yield curve is an expected policy path plus a term premium, and it inverts when expected cuts exceed the pay for waiting; the cuts themselves often come because the inversion has squeezed the carry structures that fund long assets with short money until one of them fails.
Three habits follow. Decompose every move in the slope into its policy and premium components, even with rough assumptions, because the same spread can hide opposite outlooks. Map the balance sheets that live on the slope, from regional banks to mortgage REITs to hedged foreign buyers, and track their funding: deposit outflows, commercial paper rates, and repo spreads tell you where the strain is building. And when the curve steepens, check whether the short end fell or the long end rose, because only one of those relieves the pressure.
