Hull-White Model

The Hull-White model is a mean-reverting Gaussian short-rate model fitted to the current yield curve for valuing interest-rate derivatives.

The Hull-White model is a short-rate model that represents the instantaneous interest rate as a mean-reverting Gaussian process with a time-dependent drift chosen to fit the current yield curve. The one-factor version is widely used as a tractable framework for valuing interest-rate options, callable bonds, swaptions, caps, floors, and other rate-sensitive contracts.

The model describes risk-neutral rate dynamics for valuation. Fitting today’s term structure does not mean it predicts the future path of policy rates or bond yields.

Key Takeaways

  • The one-factor Hull-White model extends the Vasicek framework with a time-dependent drift that can fit the initial yield curve.
  • One random factor drives the short rate, while mean reversion limits how persistent a rate shock is under the model.
  • The Gaussian specification is analytically convenient but permits negative rates.
  • The current discount curve and the option-volatility calibration are separate inputs and checks.
  • European bond options and some rate-option components have analytical formulas; callable and Bermudan products often require a tree, lattice, finite-difference method, or simulation.
  • Model value depends on curve construction, volatility inputs, mean reversion, numerical implementation, and product terms.
  • Fitting the initial curve is exact only relative to the selected curve-building inputs and interpolation framework.
  • Mean reversion has a model time scale: with constant (a), the half-life of a rate-factor shock is (\ln(2)/a).
  • Normal, lognormal, shifted-lognormal, and premium quotes must be translated consistently before calibration errors are compared.
  • A one-factor model restricts yield-curve dynamics and cannot represent independent level, slope, curvature, and basis shocks.
  • Risk-neutral paths support valuation; they should not be presented automatically as real-world forecasts.

One-Factor Hull-White Equation

A common risk-neutral specification is:

$$ dr_t = \left[\theta(t)-a r_t\right]dt +\sigma\,dW_t $$

where:

TermInterpretation
(r_t)Instantaneous short rate at time (t)
(\theta(t))Time-dependent drift function selected to fit the initial term structure
(a)Mean-reversion speed
(\sigma)Instantaneous short-rate volatility under the simplified constant-volatility form
(W_t)Brownian motion under the pricing measure

Some texts use different symbols or write the model as a mean-reverting factor plus a deterministic shift. Those forms can be mathematically equivalent under their definitions. The symbol (\theta(t)) in the equation above is not simply a constant long-run average rate.

Equivalent Shifted-Factor Form

A common implementation writes:

$$ r_t=x_t+\phi(t) $$

with:

$$ dx_t=-a x_t\,dt+\sigma\,dW_t $$

The stochastic factor (x_t) is an Ornstein-Uhlenbeck process, while the deterministic shift (\phi(t)) is selected to reproduce the initial term structure under the chosen conventions. This form separates today’s curve from future random shocks.

Parameterizations are not interchangeable by symbol alone. One implementation’s (\theta(t)), (\phi(t)), volatility function, initial factor value, and curve inputs must be mapped algebraically before prices or parameters are compared.

What Mean Reversion Does

The parameter (a) controls how quickly a short-rate shock decays in the model.

  • A larger (a) makes shocks less persistent and can reduce the effect of a current short-rate move on distant maturities.
  • A smaller (a) makes shocks more persistent across time.
  • The volatility parameter (\sigma) controls the scale of rate uncertainty.

The two parameters interact when fitting option prices. Different combinations can produce similar prices for selected instruments while generating different exposures for other maturities or exercise structures.

For constant (a>0), the remaining effect of a factor shock decays approximately as (e^{-a\Delta t}). Its model half-life is:

$$ t_{1/2}=\frac{\ln 2}{a} $$

For example, (a=0.10) per year gives a half-life of about 6.93 years, while (a=0.50) gives about 1.39 years. These values describe risk-neutral factor persistence under the model, not forecasts that observed yields will return to a policy target on those dates.

The conditional variance of the constant-volatility factor over a horizon (\Delta t) is:

$$ \operatorname{Var}[x_{t+\Delta t}\mid x_t] = \frac{\sigma^2}{2a} \left(1-e^{-2a\Delta t}\right) $$

As (a) approaches zero, the factor behaves more like a non-mean-reverting Gaussian rate process over the relevant horizon. Large (a) dampens distant effects, but (a) and (\sigma) must be interpreted together.

Fitting the Initial Yield Curve

The time-dependent drift is selected so the model reproduces the initial prices of zero-coupon bonds from the chosen discount curve. Under a common constant-(a) form, model zero-coupon bond prices can be written:

$$ P(t,T)=A(t,T)e^{-B(t,T)r_t} $$

with:

$$ B(t,T) = \frac{1-e^{-a(T-t)}}{a} $$

The function (A(t,T)) depends on the initial curve and model parameters. This affine structure contributes to the model’s analytical tractability.

Under the constant-volatility one-factor form, the magnitude of a zero-coupon bond’s instantaneous volatility is proportional to:

$$ \sigma_P(t,T)=\sigma B(t,T) $$

For a bond close to maturity, (B(t,T)) approaches zero. For a distant maturity, mean reversion limits the loading toward (1/a). This maturity pattern is one reason the mean-reversion parameter affects long-expiry and long-tenor option values.

The initial curve must still be built carefully. Instrument selection, interpolation, bootstrapping, day counts, collateral or discounting conventions, and market-data timestamps can change the fitted curve and derivative value.

Curve Fit Is Not Volatility Fit

Hull-White calibration commonly has two distinct layers:

LayerMain inputsWhat is matched
Deterministic curve fitDiscount factors or zero-coupon pricesToday’s selected term structure
Stochastic-parameter fitCap, floor, or swaption prices or volatilitiesSelected option evidence

An exact curve fit does not establish that caplet or swaption prices are matched. Conversely, a low option-calibration error can depend on a curve whose instruments, interpolation, collateral basis, or timestamp differ from the trade’s valuation framework.

Modern rate valuation can use separate discount and projection curves. The one-factor short rate does not automatically model basis between them. Implementations may treat some curves deterministically, introduce additional factors, or apply basis adjustments. The chosen architecture must be documented rather than inferred from the Hull-White label.

Hull-White vs. Vasicek and Black-Style Models

FeatureHull-White one-factorVasicek modelBlack-style market model
State variableInstantaneous short rateInstantaneous short rateForward rate, swap rate, or futures-related rate under selected formulation
Mean reversionYesYesNot represented through one universal short-rate process
Initial curve fitTime-dependent drift can fit the selected curveBasic constant-parameter form generally does not fit an arbitrary current curve exactlyCurrent forward or swap rate is an explicit market input
Rate distributionGaussian in the basic formGaussianCommon lognormal or normal quote model, depending on convention
Negative ratesPossiblePossibleDepends on model and displacement or normal convention
Typical useConsistent valuation across rate-dependent cash flows and exercise datesFoundational term-structure modelingQuoting and valuing selected caps, floors, or swaptions

No row makes one model universally superior. The choice depends on product payoff, exercise rights, market quote conventions, curve framework, and required risk measures.

What One Factor Implies

Every modeled bond and rate depends on the same Brownian shock in the basic one-factor framework. This creates a parsimonious, recombining state representation, but it also restricts how the yield curve can move.

  • A level-like rate shock can be represented efficiently.
  • Independent slope and curvature shocks are not fully represented.
  • Discount, projection, benchmark, funding, and cross-currency basis cannot all move independently through one factor.
  • Long-dated callable or Bermudan products can be sensitive to these omitted curve dynamics.

The model can still be useful when the omitted risks are measured separately, the product is not materially sensitive to them, or reserves and alternative-model comparisons address the limitation.

Practical Example: Receiver Swaption Exercise Value

Assume a European receiver swaption gives its holder the right at time (T) to enter a swap receiving a fixed rate (K=3.50%). At expiry, suppose:

  • the comparable market swap rate (S(T)) is (2.90%);
  • the swap annuity factor (A(T)) is (4.5) per unit of notional; and
  • the notional amount is $10 million.

For a simplified physically settled receiver swaption, the exercise value at (T) is:

$$ \text{Payoff} = N A(T)\max\left(K-S(T),0\right) $$

Using the illustrative values:

$$ \$10{,}000{,}000 \times 4.5 \times (0.035-0.029) = \$270{,}000 $$

The receiver right has value because receiving 3.50% is favorable relative to the 2.90% market swap rate in this simplified state. If the market swap rate were above the strike, the holder could let the option expire.

Before expiry, the Hull-White model values the discounted payoff across possible future rate states, using the fitted curve and calibrated rate dynamics. The $270,000 calculation is an expiry payoff illustration, not a model price and not a forecast.

How Calibration Works

Implementation commonly separates two tasks:

  1. Fit the initial term structure. Construct the discount and projection curves required by the product, then determine the time-dependent drift consistent with that framework.
  2. Fit rate-option prices or volatilities. Calibrate mean reversion, volatility, or time-dependent volatility parameters to selected caps, floors, or swaptions.

A price-error calibration can be represented schematically as:

$$ \min_{\mathbf{p}} \sum_{i=1}^{n} w_i \left[ P_{model,i}(\mathbf{p})-P_{market,i} \right]^2 $$

The parameter vector (\mathbf{p}) can contain (a), constant or time-dependent volatility parameters, and any approved extensions. The weight (w_i) determines which instruments and errors matter most.

Calibration choices include:

  • normal, lognormal, shifted, or premium-based market quotes;
  • selected expiries, tenors, strikes, and liquidity filters;
  • price-error, volatility-error, or vega-weighted objectives;
  • one constant volatility or a time-dependent volatility function;
  • parameter bounds, smoothing, and regularization; and
  • bid, ask, midpoint, and timestamp conventions.

A calibration can fit liquid at-the-money instruments while mispricing strikes or exercise patterns not included in the objective. Residuals should be reviewed across the relevant expiry-tenor grid.

Calibration evidenceQuestion to answer
Price and quote residualsWhere does the model miss by expiry, tenor, and strike?
Bid-ask comparisonIs the model value outside executable market evidence?
Starting-value rerunsDoes the optimizer converge to the same parameter region?
Parameter stabilityDo (a) and (\sigma) jump without a corresponding market change?
Out-of-sample instrumentsDoes the fit extend beyond the instruments used?
Hedge or P&L attributionDo model sensitivities explain subsequent value changes?

Different combinations of mean reversion and volatility can fit a limited instrument set similarly. Parameter stability and out-of-sample behavior should therefore be evaluated separately from the optimizer’s total error.

Converting Market Volatility Quotes

Interest-rate option markets can quote normal volatility, lognormal volatility, shifted-lognormal volatility, or premium. A calibration engine ultimately needs prices consistent with the curve, annuity, settlement, and quote convention.

  1. Identify the quote type, volatility unit, shift, expiry, tenor, strike, and timestamp.
  2. Reconstruct the market forward rate and annuity from the same curves and conventions.
  3. Convert the quoted volatility into a premium under the stated market formula.
  4. Compare the Hull-White model price with that premium or convert both prices back to the same quote convention.
  5. Reject or flag invalid, stale, crossed, or internally inconsistent observations.

Comparing a normal-volatility error directly with a lognormal-volatility error is not meaningful without conversion. A one-basis-point volatility difference also does not represent the same currency-value error for every instrument.

Where the Model Is Used

  • Zero-coupon bond options: exploit the model’s affine bond-pricing structure.
  • Caps and floors: value portfolios of rate-option cash flows under a consistent short-rate process.
  • European swaptions: use analytical or numerical approaches under the chosen calibration.
  • Callable and putable bonds: apply backward induction because exercise depends on future rate states.
  • Bermudan swaptions: compare continuation and exercise value across multiple dates.
  • Exposure simulation: generate rate scenarios for valuation adjustment or risk calculations, subject to model-purpose controls.

The product’s reference rate, fallback language, collateral currency, multi-curve framework, exercise settlement, and day-count conventions must be represented separately from the short-rate equation.

Trees, Lattices, and Numerical Methods

A recombining short-rate tree is a common implementation for products with early or repeated exercise. At each node, the engine:

  1. values future contractual cash flows;
  2. discounts them through the calibrated rate process;
  3. compares continuation with call, put, or exercise value; and
  4. works backward to the valuation date.

Grid spacing and time steps affect accuracy. The tree should reproduce the input curve and calibration instruments within documented tolerances. Monte Carlo methods may be used for exposure profiles or more complex multi-factor structures, although early exercise then requires an additional approximation method.

MethodTypical useMain implementation control
Analytical bond-option formulaEuropean claims reducible to bond optionsCurve, bond volatility, and formula convention
Recombining trinomial treeCallable bonds and Bermudan exerciseState range, transition probabilities, and time grid
Finite-difference PDEEarly exercise and state-dependent featuresBoundary conditions, grid, and convergence
Monte Carlo simulationExposure profiles and path-dependent cash flowsTime step, regression or exercise rule, and sampling error

Exercise and Callability

For an exercisable product, value at an exercise date compares immediate exercise with continuation:

$$ V(t)=\max\left(V_{\text{exercise}}(t),V_{\text{continue}}(t)\right) $$

The comparison is straightforward to state but model-dependent to calculate. Continuation value reflects future rates, cash flows, notice periods, settlement, accrued interest, call schedules, and any path-dependent contract terms.

A Bermudan swaption or callable bond can be especially sensitive to mean reversion because (a) affects how shocks propagate between exercise dates and across maturities. Two calibrations with similar European swaption errors can therefore produce different exercise boundaries and callable values.

Greeks and Rate Scenarios

Hull-White risk is broader than one parallel-rate delta. A useful report can include:

  • sensitivities to discount-curve and projection-curve nodes;
  • mean-reversion and volatility parameter sensitivity;
  • caplet or swaption volatility-bucket sensitivity;
  • parallel, steepening, flattening, and curvature scenarios;
  • exercise-boundary and call-date exposure; and
  • model-choice comparisons against normal, Black-style, or multi-factor alternatives.

Frozen-parameter Greeks hold calibrated parameters fixed while curves move. Recalibrated scenarios allow parameters to change with market quotes. Those approaches answer different questions and can produce different P&L explanations.

Model Validation and Governance

  1. Define approved products, markets, currencies, payoffs, and uses.
  2. Verify the equation, parameterization, pricing measure, and deterministic fitting function.
  3. Rebuild discount and projection curves independently and reconcile zero-coupon prices.
  4. Benchmark analytical zero-coupon bond and bond-option values.
  5. Reprice calibration instruments and inspect errors across expiry, tenor, and strike.
  6. Test tree, PDE, or simulation convergence and difficult parameter regimes.
  7. Compare Greeks with bump-and-revalue results under documented curve and volatility shocks.
  8. Challenge one-factor, Gaussian, multi-curve, and negative-rate limitations.
  9. Review out-of-sample prices, exercise boundaries, hedge attribution, and parameter stability.
  10. Document reserves, overlays, overrides, monitoring thresholds, and independent approval.

Risks and Limitations

  • One-factor limitation: a single source of rate uncertainty cannot represent every yield-curve twist, spread, and volatility movement.
  • Gaussian-rate limitation: the basic process permits negative short rates and may not match market distribution assumptions.
  • Calibration risk: parameters can depend materially on quote selection, weights, bounds, and market regime.
  • Volatility-surface risk: a simple parameterization may fit at-the-money options but miss strike-dependent skew.
  • Curve risk: discount, projection, basis, collateral, and interpolation choices can materially change value.
  • Exercise-model risk: callable and Bermudan values depend on the modeled continuation value and exercise policy.
  • Numerical risk: tree, grid, simulation, and interpolation choices can create bias or instability.
  • Parameter-instability risk: mean reversion and volatility can change when the calibration window or instrument set changes.
  • Product-basis risk: a one-curve short-rate model may not capture benchmark, funding, or cross-currency basis dynamics.
  • Measure risk: risk-neutral paths can be mistaken for real-world rate forecasts.
  • Quote-conversion risk: normal, lognormal, shifted, and premium inputs can be mixed incorrectly.
  • Exercise-boundary risk: small parameter or grid changes can alter modeled call or exercise decisions.
  • Recalibration risk: changing parameters during attribution can blur market P&L and model P&L.
  • Hedge-basis risk: liquid calibration instruments may not match the trade’s strike, tenor, settlement, or curve exposure.

Common Mistakes

  • Treating (\theta(t)) as a directly observed long-run policy-rate forecast.
  • Assuming fitting today’s yield curve proves that the model forecasts future yields accurately.
  • Calibrating to one option and using the result across all strikes, expiries, and products without testing.
  • Ignoring the model’s ability to generate negative rates.
  • Calling Hull-White a volatility model for equities rather than a term-structure model for rates.
  • Using a one-factor model for material curve or basis exposures without measuring the omitted risks.
  • Reporting a precise model value without curve, quote, parameter, and numerical-method evidence.
  • Treating the deterministic curve-fitting term as a forecast of future mean rates.
  • Assuming a low at-the-money calibration error proves a good smile or Bermudan fit.
  • Comparing normal and lognormal volatility residuals without converting them consistently.
  • Using one tree grid without convergence or exercise-boundary tests.
  • Interpreting negative simulated rates as either automatically impossible or automatically harmless.

Authoritative Sources

FAQs

What does the Hull-White model fit exactly?

The time-dependent drift can be chosen so the model matches the selected initial zero-coupon curve. Mean reversion and volatility parameters are then commonly calibrated to rate-option evidence. These are distinct calibration steps.

Can the Hull-White model produce negative interest rates?

Yes. The basic one-factor model is Gaussian, so negative short-rate states are possible. Whether that is acceptable depends on the market, product, calibration, and model-use policy.

Is Hull-White a forecasting model?

Not primarily in derivatives use. It is usually applied as a risk-neutral valuation model calibrated to current curves and option prices. Its simulated paths should not be interpreted automatically as real-world rate forecasts.

Does fitting the yield curve also fit swaption volatility?

No. The deterministic drift or shift fits the selected initial curve. Mean reversion and volatility parameters require a separate calibration to rate-option prices or volatilities.

Check Your Understanding

Loading quiz…

For an actual valuation, use current curves, option quotes, trade terms, model documentation, calibration records, validation findings, and executable market evidence. This article is for financial education only and is not personalized investment, derivatives, valuation, legal, accounting, or tax advice.

Browse Financial Instruments