The yield curve

The yield curve is a single picture of what the entire bond market expects the future to look like. It plots the interest rate the US Treasury must pay across every maturity it issues — from three months out to thirty years — and the shape of that line is the most concentrated piece of macroeconomic information on the planet. Every fixed-income trader, every central banker, every pension fund manager looks at this chart first thing in the morning. There is a reason.

In normal times the curve slopes upward. Lending the government money for thirty years carries more risk than lending it overnight — more time for inflation to surprise, more time for policy to drift, more time for the world to break — so long-dated bonds offer a higher yield. This is called the term premium, and it is the default shape of a healthy economy with a Federal Reserve that is not aggressively fighting anything.

The inverted curve is what makes this instrument extraordinary. When short-term rates exceed long-term rates — when a two-year Treasury yields more than a ten-year — the bond market is making a specific and consequential claim: that the average overnight rate over the next decade will be lower than today's overnight rate. The only way that arithmetic works is if the Federal Reserve is about to cut. And the Federal Reserve, in modern monetary history, only cuts aggressively when the economy is breaking.

Every US recession in the last fifty years was preceded by an inversion of the 10Y-2Y spread. The 1980 recession, the 1981-82 double-dip, the 1990 recession, the 2001 dot-com bust, the 2008 financial crisis, the 2020 pandemic shock — each one had a yield curve inversion in front of it, typically six to twenty-four months before the recession officially began. The only false positive in the entire post-war record was a 1966 inversion that was followed by a near-recession but not an NBER-dated one. No other indicator in macroeconomics has a track record this clean.

The 2022-23 inversion is what makes this lesson urgent. At its deepest point in July 2023 the 10Y-2Y spread touched approximately -108 basis points — deeper than 2006, deeper than 2000, deeper than 1989. The only modern parallel is the Volcker inversion of 1981, when the Fed Funds rate was held above nineteen percent to break the back of the 1970s inflation. That earlier inversion was followed by the worst recession since the Great Depression. The current one has, so far, been followed by the most resilient labor market in living memory. Decades of evidence and trillions of dollars of positioning hang on which interpretation is correct.

What the curve does not tell you is the timing or the severity. The lag from inversion to recession has ranged from six months (2020) to twenty-four months (2007). The depth of the recession that follows is not correlated with the depth of the inversion. The curve is a binary signal that something is wrong with the calibration between the Fed's policy stance and the economy's actual capacity for growth. It does not tell you when the breakage will become visible, and it does not tell you how bad the breakage will be.

The transmission mechanism from inversion to recession runs through the banking system. Banks borrow short and lend long — they take deposits at near the overnight rate and issue mortgages, business loans, and commercial credit at rates closer to the ten-year. When that spread inverts, every new loan they write is unprofitable at the margin. Credit creation stalls. The pipeline of new investment that the economy depends on slows to a trickle. By the time you see it in unemployment data, it has been happening in bank loan officer surveys for a year.

For asset allocation, the curve is one of the cleanest signals available. When the 10Y-2Y is deeply inverted, the historical playbook has been brutally consistent: cash, gold, long Treasuries, and the dollar have outperformed risk assets in six of the last six cycles. Stocks, high-yield credit, and small caps have underperformed. Long-duration bonds in particular become a leveraged bet on the Fed eventually capitulating to disinflation — a 2% drop in long yields produces a roughly 36% gain on a 30-year position with no options and no margin. The math is mechanical and the historical edge is real.

The most important thing to understand about the yield curve is what it actually represents. It is not a prediction from any single analyst or model. It is the consensus bet of every bond investor in the world, weighted by the size of their positions, updated continuously. When that consensus says short rates will be lower than long rates a decade from now, it is saying the economy cannot sustain current monetary policy. That collective signal has been right every single time it has been deeply persistent. The 2022-23 inversion may yet break that record, or it may not. Either way, ignoring what the curve is saying has not historically been a winning strategy.