The History of the Yield Curve as a Recession Indicator
The Foundations: Explaining the Shape of the Curve
Long before anyone used the yield curve to say anything about recessions, economists were trying to explain why interest rates differ by maturity at all. Irving Fisher argued in the early twentieth century that long rates embody expectations about future short rates and future inflation. Friedrich Lutz set out the pure expectations theory formally in a 1940 article, and John Hicks added the idea that lenders must be compensated for tying up funds, which introduces a term premium that normally makes long rates exceed the average of expected short rates. Later work added competing accounts: market segmentation, in which different investors are confined to different maturities, and the preferred-habitat theory of Franco Modigliani and Richard Sutch in 1966, which allowed investors to leave their preferred maturity for a sufficient price.
1965: Kessel Notices the Cyclical Pattern
Reuben Kessel's NBER study The Cyclical Behavior of the Term Structure of Interest Rates, published in 1965, was among the first systematic examinations of how the shape of the curve moves over the business cycle rather than at a point in time. Kessel documented that the spread between long and short rates behaved differently in expansions than in contractions, with short rates rising toward and sometimes above long rates late in expansions. He was working within the term-premium tradition and treating the finding as a fact about interest rate behavior rather than proposing a forecasting tool, but the observation that the curve's slope is systematically related to the cycle is the seed of everything that followed.
1986: Campbell Harvey's Dissertation
The decisive step came from a doctoral student. Campbell Harvey, working at the University of Chicago, wrote a 1986 dissertation examining what the term structure of interest rates reveals about expected future consumption and economic growth. His argument was that the yield curve's slope embeds the market's collective view of future real activity, and he showed empirically that the spread between long and short rates carried information about subsequent growth. Related work was published in the Journal of Financial Economics in 1988. Harvey later joined Duke University's Fuqua School of Business, and the finding — that a flattening or inverting curve had preceded downturns — moved from a dissertation result to a widely watched market indicator over the following decade.
The 1990s: Estrella, Hardouvelis, and Mishkin at the New York Fed
Research at the Federal Reserve Bank of New York turned the finding into a quantified tool. Arturo Estrella and Gikas Hardouvelis published The Term Structure as a Predictor of Real Economic Activity in the Journal of Finance in 1991. Estrella then worked with Frederic Mishkin, who was director of research at the New York Fed and later served as a Federal Reserve Governor, on a series of papers including a 1996 Current Issues in Economics and Finance piece on the yield curve as a predictor of U.S. recessions and a 1998 article comparing financial variables as leading indicators. Their contribution was to fit probability models — most often using the spread between the ten-year Treasury yield and the three-month Treasury bill rate — that translate a given spread into an estimated recession probability. The New York Fed has published such an estimate on an ongoing basis since.
The Modern Record and Its Caveats
By the 2000s the yield curve had entered general financial vocabulary, and the interest rate spread became a component of The Conference Board's Leading Economic Index. The historical record is genuinely striking: an inversion preceded each of the U.S. recessions beginning in 1969-70, 1973, 1980, 1981, 1990, 2001, and 2007, and the curve had inverted in 2019 before the 2020 downturn, though that downturn had an obvious external cause. The record also contains complications. The 1966 inversion was followed by a slowdown but no NBER-dated recession, a case usually cited as a false positive. Lead times have ranged widely, from months to roughly two years. And in the mid-2000s Alan Greenspan described as a conundrum the failure of long rates to rise with short rates, prompting an active literature on whether a compressed term premium — driven by global demand for safe assets, pension demand, and later central bank bond purchases — changes what an inversion means.