On October 29, 2026, a one-year Treasury yielded 4.42% and a two-year Treasury 4.61%. An investor with 10,000 dollars and two years to wait could lock in the two-year yield and end with 10,943.25 dollars. Or the investor could buy the one-year bill, collect 10,442 dollars in October 2027 and reinvest at whatever the one-year rate is then. The two plans finish level if that future rate is 4.80%.

Expectations theory says 4.80% is the market's forecast of the one-year rate a year ahead. Every upward or downward slope in the yield curve becomes a statement about where short-term rates are going, and nothing else.

Expectations theory
The theory that a long-term interest rate equals the average of the short-term rates investors expect over the life of the bond. It is also called the expectations hypothesis, the pure expectations theory or the unbiased expectations theory.
Forward rate
The future short-term rate implied by today's yields, set so that holding one long bond and rolling shorter bonds give the same result.
Term premium
The extra yield a long bond pays over the average of expected short rates. Pure expectations theory sets it to zero, and the competing theories explain where it comes from.

What expectations theory says about the yield curve

The theory rests on one assumption: investors care only about expected return, so bonds of different maturities are perfect substitutes. If a two-year bond were expected to pay more than two one-year bonds in a row, investors would buy it until its yield fell back into line. The same argument drives the efficient market hypothesis: any predictable extra return is competed away.

For a bond of n years, the theory gives:

(1 + yn)n = (1 + r1) × (1 + E[r2]) × … × (1 + E[rn])

Here yn is today's n-year yield, r1 is today's one-year rate and E[r2] to E[rn] are the one-year rates expected in each later year. The long yield is a geometric average of the expected short rates. Friedrich Lutz set out this version in the Quarterly Journal of Economics in 1940.

Three readings of the curve follow directly:

  • An upward-sloping curve means short rates are expected to rise.
  • A flat curve means they are expected to stay where they are.
  • An inverted curve, with long yields below short ones, means they are expected to fall.

The word "unbiased" refers to the forward rate: under the theory it is an unbiased forecast of the future short rate, right on average with errors nobody can predict in advance. Once future rates are uncertain, the different ways of stating the theory stop being exactly equivalent, because compounding turns an average of rates into a slightly different average of returns. John Cox, Jonathan Ingersoll and Stephen Ross showed in 1981 that the traditional versions cannot all hold at once in an equilibrium model.

How to calculate a forward rate

The one-year rate expected one year from now comes from the one-year and two-year yields:

1 + f1,2 = (1 + y2)2 ÷ (1 + y1)

With the October 29, 2026 yields, 1.0461 squared is 1.094325, and dividing by 1.0442 gives 1.048003. The forward rate is 4.80%. The general formula divides one plus the longer yield, raised to its number of years, by one plus the shorter yield, raised to its number of years, then takes the root for the number of years in between.

Forward rateCalculationResult
One-year rate, one year ahead1.0461² ÷ 1.04424.80%
One-year rate, two years ahead1.0473³ ÷ 1.0461²4.97%
Two-year rate, one year ahead(1.0473³ ÷ 1.0442)1/24.89%
Two-year rate, three years ahead(1.04785 ÷ 1.0473³)1/24.86%

The inputs are the Treasury's one, two, three and five-year yields of 4.42%, 4.61%, 4.73% and 4.78% on October 29, 2026. A quick approximation works when rates are moderate: twice the two-year yield minus the one-year yield, 2 × 4.61 − 4.42, also gives 4.80.

The Treasury's daily curve uses par yields on coupon bonds, and precise work uses a zero-coupon curve built from the same bonds.

Read through pure expectations theory, that curve forecast a one-year rate of 4.80% in October 2027 and 4.97% in October 2028. The Federal Reserve had raised its target range to 3.75% to 4.00% in September 2026, its first increase since 2023, so the curve priced further increases.

Liquidity premium theory and the term premium

A long bond loses more value than a short one when rates rise. The liquidity premium theory says investors demand extra yield for carrying that risk, so a forward rate equals the expected future rate plus a premium that grows with maturity. Under this theory the curve slopes upward even when no change in rates is expected, which is its usual shape.

Take one-year rates that everyone expects to stay at 4.42% for two years, and a premium of 0.20 points on the two-year bond. The two-year yield is about 4.62%, and the curve slopes upward although nobody expects rates to move. Read through pure expectations theory, the same curve forecasts a one-year rate of about 4.82% a year later, and the 0.40-point rise it shows is all premium.

The premium cannot be observed directly. The Federal Reserve Bank of New York publishes an estimate built on the model of Tobias Adrian, Richard Crump and Emanuel Moench, known as the ACM term premium, which splits each Treasury yield into an expected path of short rates and a premium. For the 10-year Treasury, the monthly series shows:

  • an average premium of 1.43 percentage points from June 1961 to September 2026
  • a peak of 5.20 points in May 1984
  • a low of minus 1.38 points in July 2020, when investors accepted less than the expected path of short rates to hold long bonds
  • 0.89 points at the end of September 2026

The same data separates forecast from premium in the forward rate. At the end of September 2026, the model's one-year and two-year yields were 4.56% and 4.84%, a forward of 5.13% for the one-year rate a year later. With the premium removed, the expected yields were 4.27% and 4.42%, a forward of 4.57%. About 0.55 points of the forward rate was premium. The model uses continuous compounding, so twice the two-year yield minus the one-year yield gives the forward exactly.

Market segmentation and preferred habitat theory

John Culbertson argued in 1957 that short and long bonds trade in largely separate markets. In the segmentation view, banks hold short paper, pension funds and insurers hold long bonds, and each yield is set by supply and demand within its own segment. In its strict form the theory has trouble with a plain fact: short and long yields usually move in the same direction.

Franco Modigliani and Richard Sutch proposed a middle position in the American Economic Review in 1966. Investors have a preferred maturity, or habitat, and leave it only for extra yield. Expectations still link the segments, but a shortage or glut of bonds at one maturity can move its premium up or down. The premium can therefore be negative, which liquidity premium theory does not allow.

The Federal Reserve's bond purchases after the 2008 crisis put the idea to a large test. From late 2008 to March 2010 it bought more than 1.7 trillion dollars of longer-term securities, 22% of the stock of longer-term agency debt, mortgage-backed securities and Treasuries. The announcements cut the 10-year term premium by 50 to 100 basis points, depending on the window measured, and the term premium accounted for most of the fall in the 10-year yield around those announcements. Under pure expectations theory, buying bonds would change long yields only by changing expected short rates.

Pure expectations, liquidity premium, preferred habitat and segmentation compared

TheoryWhat sets long-term yieldsWhat a forward rate measuresSign of the term premium
Pure (unbiased) expectationsExpected short-term rates onlyThe expected future short rateZero by assumption
Liquidity premiumExpected short rates plus a premium for interest rate riskThe expected rate plus a premiumPositive, rising with maturity
Preferred habitatExpected short rates plus premiums set by supply and demand at each maturityThe expected rate plus a premiumPositive or negative
Market segmentationSupply and demand inside each maturity segmentNo fixed link to expected ratesNot defined

The liquidity premium and preferred habitat theories are sometimes grouped as "biased expectations theories", because both keep expectations at the center and add a premium that makes the forward rate a biased forecast.

Empirical tests of expectations theory

The standard test regresses the change in rates that actually happened on the change the forward rate predicted. If expectations theory holds, the slope coefficient is 1.

At a one-year horizon the forward rate barely helps. Regressing the change in the one-year rate over the next year on the change implied by the forward rate gave a slope of 0.09 over 1964 to 1984 in Eugene Fama and Robert Bliss's tests, far from 1. If short rates followed a random walk, today's rate would be the best forecast of next year's, and the slope would be zero. The picture improves with distance: the forecasting power of the curve rises as the horizon lengthens from two to five years. Fama found the same longer-horizon result out of sample from 1986 to 2004, with spot rates reverting toward a level that itself shifts over time.

John Campbell and Robert Shiller ran the test across maturities from one month to 10 years on monthly data from January 1952 to February 1987. They checked two predictions of the theory:

  1. The long yield. A steep curve should be followed by a rise in long yields. The slope should be 1. It came out at minus 3.10 for the five-year yield against the one-month rate and minus 5.02 for the 10-year yield against the one-month rate. Long yields tended to fall after the curve steepened, with mostly wrong signs in every subperiod except 1952 to 1959.
  2. Short rates over the long bond's life. A steep curve should be followed by rising short rates. Here the data went the right way, though the spread moved roughly twice as much as the theory implies.

Their reading was that the spread behaves like a multiple of its expectations-theory value, which time-varying risk premia could produce.

John Cochrane and Monika Piazzesi measured that premium directly. One combination of forward rates predicted one-year excess returns on Treasury bonds of one to five years with an R-squared of up to 0.44. A premium that today's curve can forecast changes over time, which pure expectations theory rules out.

Forward rates against the rates that followed, 2019 to 2026

The test can be run by hand on recent Treasury data. The table takes the one-year and two-year yields at each year end, computes the one-year rate they implied for a year later and compares it with the one-year yield on the last trading day of the following year. The par yields are treated as annual rates.

Curve date1-year yield2-year yieldImplied 1-year rate a year laterActual 1-year yield a year laterMiss, points
Dec 31, 20191.59%1.58%1.57%0.10%−1.47
Dec 31, 20200.10%0.13%0.16%0.39%+0.23
Dec 31, 20210.39%0.73%1.07%4.73%+3.66
Dec 30, 20224.73%4.41%4.09%4.79%+0.70
Dec 29, 20234.79%4.23%3.67%4.16%+0.49
Dec 31, 20244.16%4.25%4.34%3.48%−0.86
Dec 31, 20253.48%3.47%3.46%4.42% on Oct 29, 2026+0.96 so far

Over the six complete years the average absolute miss was 1.24 points, in both directions. The two largest misses came from policy surprises. In March 2020 the Fed cut its target range by 1.5 points in two weeks to 0% to 0.25%. In 2022 it raised the range from 0% to 0.25% to 4.25% to 4.50%, while the curve at the end of 2021 had priced a one-year rate of about 1.07%.

What the 2022 to 2024 inversion implied

The 10-year Treasury yield closed below the two-year yield on every trading day from July 6, 2022 to August 26, 2024, 782 calendar days, with two single-day returns to inversion in early September 2024. The gap was widest on July 3, 2023, at 1.08 points. Against the three-month bill, the 10-year stayed lower from October 25, 2022 to December 12, 2024, with the widest gap at 1.89 points on June 1, 2023. The dates are counted from the Treasury's daily par yield curve.

Under pure expectations theory, an inversion of that size meant one thing: investors expected short rates to fall a long way. The Fed held its range at 5.25% to 5.50% from July 2023, then cut by 1 point between September and December 2024 and by another 0.75 points in late 2025, to 3.50% to 3.75%. The cuts came, later than the curve had priced at the end of 2023.

The ACM estimates show the inversion was only partly a rate forecast. At the end of June 2023, the model's one-year yield was 5.33% and its 10-year yield 3.85%, a gap of 1.48 points. The expected path of short rates averaged 4.62% over 10 years, and the term premium was minus 0.77 points. About half of the inversion came from expected cuts and about half from investors accepting a negative premium to hold long bonds.

Pure expectations theory makes no recession forecast. An inverted curve says only that short rates are expected to fall, and a falling inflation rate can produce that as well as a downturn. The NBER's business cycle chronology still lists April 2020 as the most recent trough, so no recession has been dated since the inversion began.

By October 2026 the curve had turned upward again. The 10-year yielded 4.98%, the two-year 4.61% and the three-month bill 4.18% on October 29, and the ACM 10-year term premium had gone from minus 0.58 points at the end of 2022 to 0.89 points at the end of September 2026.

What expectations theory means for a bond investor

Under pure expectations theory, locking in a long yield, rolling short bills and buying a longer bond to sell early all have the same expected return over a given horizon. The choice changes the risk. A bond that matures on the date the money is needed fixes the result in advance. Rolling bills leaves the result to future short rates. A longer bond sold before maturity leaves it to future long yields, and the 10-year Treasury lost 17.83% in 2022, as our article on the 60/40 portfolio shows.

A positive term premium tilts that comparison: it means long bonds are priced to earn more than rolling bills on average, in exchange for losses in years when rates rise faster than forward rates implied. A negative premium, as in most months from 2017 to mid-2024, means the reverse. This article is general information, not investment advice.

Frequently asked questions

What is expectations theory in simple terms?

Expectations theory says that a long-term interest rate is the average of the short-term rates investors expect over the life of the bond. If the two-year Treasury yields more than the one-year Treasury, the theory reads the gap as a forecast that one-year rates will rise. On October 29, 2026, yields of 4.42% for one year and 4.61% for two years implied a one-year rate of 4.80% a year later.

What is the difference between unbiased and biased expectations theory?

Unbiased, or pure, expectations theory says forward rates are unbiased forecasts of future short rates, with no extra yield for holding long bonds. Biased expectations theories, mainly the liquidity premium and preferred habitat theories, add a term premium, so a forward rate equals the expected rate plus a premium. The New York Fed's estimate of the 10-year term premium averaged 1.43 percentage points from 1961 to 2026 and was 0.89 points in September 2026.

How do you calculate the expected future interest rate with expectations theory?

Divide one plus the longer yield, raised to its number of years, by one plus the shorter yield, raised to its number of years, then take the root for the years in between and subtract one. With a one-year yield of 4.42% and a two-year yield of 4.61%, the one-year rate expected a year from now is 1.0461 squared divided by 1.0442, minus one, or 4.80%.

What does an inverted yield curve mean under expectations theory?

Under pure expectations theory, an inverted curve means investors expect short-term rates to fall. The theory says nothing about why they would fall, so a recession forecast is a separate reading. When the US curve inverted from July 2022 to August 2024, rate cuts came from September 2024, while the NBER still lists April 2020 as the most recent recession trough. About half of the gap between the one-year and 10-year yields in June 2023 came from a negative term premium.

Does expectations theory hold in practice?

Only partly. On US data from 1952 to 1987, the slope of the yield curve predicted the direction of short-rate changes over several years, as the theory says, but long yields tended to fall after the curve steepened, the opposite of what it predicts. Forward rates also forecast bond excess returns, which means the term premium varies over time. The theory now serves mainly as a benchmark, with the gap between forward rates and expected rates treated as a time-varying premium, as in the New York Fed's term premium estimates.