By: Phil Kernen
Financial markets are full of indicators and data relationships from which we tease conclusions. Some work better than others, but few hold our attention as do yield curve inversions and their habit of showing up before a recession. But are they still accurate?
U.S. Treasury yields are the basis from which all other bonds price. Yield curves are usually upward sloping to reflect the additional risk of lending money for longer terms. Lenders charge a higher rate for 30-year loans than 10-year loans. The difference in yields from one point to the next is called a term premium or a spread. If the 2-year Treasury yield is 1.5% and the 10-year Treasury yield is 2%, the 2-10 spread is +0.5%.
When short yields rise above longer yields, the spread turns negative, and a recession is likely to follow. In 2018, the San Francisco branch of the Federal Reserve published a study, with data since 1955, on the ability of yield curve inversions to predict recessions. It found that negative spreads, or an inversion of the yield curve, preceded every recession. Yield curve inversion historically precedes recessions because investors foresee something concerning, like a declining private investment that cheapens long-term government borrowing. Or like short-term credit that becomes more expensive, which can choke off near-term investments. In other words, they see a recession.
With inflation strengthening and the Federal Reserve planning to raise overnight interest rates in response, short rates have quickly increased, but not as much as long rates. The chart below shows the progression of the 2-10 spread over the last 35 years, with recessions overlaid in red, drawing the same conclusion as the Fed study.
However, we believe that yield curve inversion has lost its predictive power due to the monetary policy tools applied over the last decade-plus. Since 2008, to suppress intermediate and long-term bonds, the Federal Reserve has purchased trillions of U.S. Treasury Notes and Mortgage-Backed Bonds. It intended to push investors into riskier assets which would quickly help to both reopen the economy and kick-start growth. All that buying suppressed and sidelined one important market function – price discovery.
Price discovery is the organic process of determining the price of an asset at any given point in time. It is about finding where natural supply and demand meet. It relies on market structure, market participants for liquidity, and information flow. It requires a majority of participants to be price sensitive. When price discovery works appropriately, prices or interest rates settle due to thousands of independent buying and selling decisions.
Price discovery isn’t currently happening in interest rate markets. Central banks have always directly set overnight rates and indirectly influenced longer rates. But since 2008, central banks have been purchasing so many longer-term bonds for so long, without regard for price levels, that they have neutered price discovery. In 2021, the Fed purchased one-half of the total new issuance of U.S. Treasury notes to drive yields down (and prices up).
Across the U.S. Treasury yield curve, the Fed is the price-setting elephant in the room. Where would interest rates be if the Fed wasn’t the most prominent buyer by magnitudes, and what would term spreads tell us? It is no coincidence that the yield curve inverted briefly in August 2019 for the first time since 2007, but no recession followed. We also experienced a recession in 2020, which was driven not by tighter credit but by the arrival of COVID-19 instead. Price discovery won’t happen until the Fed stops buying bonds and starts selling down its portfolio.
Keep this in mind when you hear about the risks and downside of a yield curve inversion in the coming months. It may have been a reasonably reliable predictor of recession when the Fed only drove overnight rates and price discovery was allowed to work. It isn’t working today, which leaves yield curve inversion as another manipulated indicator that can no longer tell us what it once did.
