Three Theories That Explain the Yield Curve's Shape

No single theory fully accounts for why long-term rates sit where they do relative to short-term rates, and treating any one of them as the whole story leads to confident but wrong conclusions about what a curve is signaling. Three complementary theories, taken together, do a far better job.

By Karanvir Singh, founder of Quantix Invest

The three theories

The pure expectations hypothesis holds that a long-term yield is simply the compounded average of expected future short-term rates, full stop, with no additional compensation demanded purely for holding a longer maturity. Under this theory alone, an upward-sloping curve means the market expects short rates to rise, a flat curve means expected rates are stable, and an inverted curve means the market expects short rates to fall.

The liquidity preference theory (sometimes framed together with the related term premium theory) starts from the same expected-rate foundation but adds that investors generally prefer shorter maturities, since they carry less interest rate risk, and so must be paid a maturity-specific premium to accept longer commitments. Under this theory, long yield = average expected short rate + term premium, with the premium growing with maturity. This alone would tilt the curve upward on average even if the market genuinely expected flat future rates, which is exactly the mechanism demonstrated numerically below and in the companion article on interest rate uncertainty.

The market segmentation theory (with its close variant, the preferred habitat theory) takes a different approach entirely: it argues that different classes of investors operate largely within separate maturity "habitats," driven by their own liabilities and regulatory constraints, rather than freely arbitraging across the whole curve. Banks and money market funds gravitate toward short maturities to match short-term liabilities and liquidity needs; pension funds and insurers gravitate toward long maturities to match decades-long payout obligations. Because these investor bases do not fully substitute across maturities, concentrated supply or demand within one segment, a wave of pension-fund buying at the 20-to-30-year point, for instance, can move yields at that specific point on the curve independent of what rate expectations or a uniform term premium alone would predict.

The distinction between strict segmentation and preferred habitat matters for how rigid the theory's predictions are. Strict segmentation assumes an investor never crosses their maturity boundary under any circumstances, which is a fairly extreme assumption. Preferred habitat softens this: it holds that investors have a natural home maturity range but can be induced to venture outside it if the compensation on offer is attractive enough. This softer version matches observed institutional behavior more closely, since large asset owners do occasionally shift allocation toward maturities outside their typical range when relative pricing becomes sufficiently compelling, but the practical implication for curve-watchers is the same either way: yields at specific points on the curve are not purely a function of rate expectations and a smooth, uniform term premium, but also of how strongly capital resists or accepts moving across habitat boundaries at any given time.

A worked example: the same curve under two theories

Suppose the observed 1-year spot rate is 3.00% and the observed 2-year spot rate is 3.60%. Under the pure expectations hypothesis, the implied forward rate for year 2 is calculated exactly as it would be for a rate forecast, since the theory assumes the forward rate equals the expected future short rate with nothing added:

f_1,2 = (1.036)² / 1.030 − 1 = 1.073296 / 1.030 − 1 = 1.042035 − 1 = 4.20%

Under pure expectations, this curve is telling you the market expects the 1-year rate to rise from 3.00% today to about 4.20% one year from now, a full 1.2 percentage point increase priced in as a pure forecast.

Now reinterpret the identical curve under liquidity preference theory, assuming a term premium of 0.30% is embedded in the 2-year rate specifically, compensation for the extra interest rate risk of the longer maturity. Using the standard additive approximation, the 2-year rate equals the average of the two years' expected short rates plus the premium:

r₂ = [r₁ + E(r_1,2)] / 2 + L₂

3.60% = [3.00% + E(r_1,2)] / 2 + 0.30%

Solving: [3.00% + E(r_1,2)] / 2 = 3.30%, so 3.00% + E(r_1,2) = 6.60%, giving E(r_1,2) = 3.60%.

The exercise generalizes cleanly to any assumed premium size. If the true term premium embedded in the 2-year rate were instead 0.10%, a smaller figure appropriate to a calmer market environment, the same algebra would give [3.00% + E(r_1,2)] / 2 = 3.50%, so E(r_1,2) = 4.00%, an implied expected rate increase of a full percentage point rather than 0.60 points. This sensitivity is exactly why the size of the assumed term premium matters so much to the conclusion: a modest change in that single assumption swings the implied rate forecast meaningfully, which is a caution worth carrying into any analysis that leans heavily on liquidity preference theory to interpret a curve.

Which theory the evidence favors

Empirical work over many decades has not crowned a single winner; instead, it has generally concluded that all three mechanisms operate simultaneously, with their relative importance shifting across time and across market conditions. Pure expectations alone fails a basic empirical test: if it held exactly, forward rates should be unbiased predictors of future short rates, and extensive testing has repeatedly found a persistent, systematic bias consistent with a real, positive term premium layered on top, supporting liquidity preference theory's core addition. At the same time, market segmentation effects show up clearly in specific historical episodes, regulatory changes that push insurers or pension funds to concentrate buying at particular long maturities have been observed to compress yields at exactly those maturities relative to a smooth curve fitted through neighboring points, a pattern pure expectations or a uniform term premium cannot explain on their own.

The most defensible modern synthesis treats the observed curve as the sum of three layers: a base level reflecting genuinely expected future short rates, a term premium that generally grows with maturity but varies over time with policy and inflation uncertainty, and localized segmentation effects that can push specific maturities away from what the first two layers alone would predict.

How this changes what you infer from a real curve

For a bond investor, the practical value of knowing all three theories is resisting overconfident single-cause narratives that show up constantly in financial commentary. A headline claiming the curve's steepening "shows the market now expects sharply higher rates" is implicitly assuming pure expectations theory and ignoring that steepening is exactly what a rising term premium would also produce, particularly during periods of elevated policy uncertainty. A separate headline claiming a specific maturity looks "cheap" relative to its neighbors might be picking up a genuine segmentation effect worth understanding, rather than a rate-expectation anomaly.

Practically, this argues for looking at the whole curve's shape and its recent history rather than fixating on any single spread, and for being specifically alert to segmentation-driven distortions around maturities where regulatory or institutional demand is known to concentrate, commonly the very long end of the curve where pension and insurance liability matching is heaviest.

This has a direct implication for anyone comparing government bond yields across countries as well. Two countries with structurally similar inflation and growth outlooks can nonetheless show meaningfully different curve slopes if their domestic pension and insurance sectors differ in size, regulatory design, or typical liability duration, since those differences change how strong the segmentation and habitat effects are at each country's long end. A steep curve in one market and a flatter curve in another are not automatically telling you the two central banks or economies are on different paths; part of the gap can simply be a structural difference in how much of each country's long-duration debt is absorbed by habitat-driven, rather than expectations-driven, demand.

Actionable breakdown

• Diagnosing what is driving a curve's shape
  • Check if the whole curve moved or just one maturity.
  • Compare the move against typical term-premium behavior.
  • Consider institutional demand shifts for isolated distortions.
• Applying the theories together
  • Expect a positive term premium to bias the curve upward.
  • Use pure expectations only as a rough starting baseline.
  • Watch for segmentation effects near pension-heavy maturities.
• Reading financial commentary skeptically
  • Question single-theory explanations for curve moves.
  • Ask which theory a claimed interpretation assumes.
  • Look at the whole curve, not one isolated spread.

Common pitfalls

Relying on pure expectations exclusively: as the worked example shows, a curve's slope can substantially overstate the market's true rate forecast once a realistic term premium is accounted for.

Ignoring institutional demand shifts: a regulatory or structural change that pushes insurers or pension funds toward specific maturities can move yields at that point independent of any macro rate story.

Treating the theories as mutually exclusive: the strongest empirical case is that all three operate together at different strengths depending on conditions, not that one theory is simply correct and the others wrong.

Assuming a fixed term premium across all conditions: as the worked sensitivity example shows, even a small change in the assumed premium meaningfully shifts the implied expected-rate forecast, so a single static assumption applied across very different market environments can badly mislead.

The bottom line

The yield curve's shape reflects a blend of genuine rate expectations, a maturity-driven term premium, and localized institutional demand effects, and any explanation crediting just one of the three should be treated with real skepticism.

The practical upshot for anyone reading financial commentary is to ask, whenever a curve move is explained with confidence, which of the three theories the explanation is implicitly leaning on, and whether the other two might just as plausibly account for part of what happened. A curve move attributed entirely to a shifting rate outlook may, on closer inspection, owe as much to a widening term premium during an uncertain policy period, or to a large institutional buyer or seller concentrated at one specific maturity, as to any genuine change in the market's collective forecast for where rates are headed.

Holding all three theories in mind at once is admittedly less satisfying than a single clean story, but it is the version that actually survives contact with decades of curve data. A curve is the joint output of what the market collectively expects, how much it charges to bear the uncertainty of being wrong, and how freely capital moves across maturity boundaries when the price is right, and no single one of those three forces, on its own, has proven sufficient to explain the shapes curves have actually taken across different countries and different decades.

Treat the three-theory framework as a checklist rather than a formula to solve precisely. Faced with an unusual curve shape, working through each theory in turn, what would pure rate expectations alone imply, how much of the gap could plausibly be term premium, and is there a specific institutional or regulatory story that could be distorting one particular maturity, produces a far more grounded read than reaching immediately for whichever single explanation is dominating that week's commentary.

Related reading: bonds fundamentals, reading the yield curve, the term premium explained, interpreting the term structure in practice, yield curve, defined.

All articles · The deep guides