Vasicek Model: Mean-Reverting Interest Rates & Bond Pricing

Published: April 13, 2026

Article by Ryan O'Connell, CFA, FRM

The Vasicek model is the foundational framework for understanding how interest rates evolve stochastically over time. Whether you’re a fixed income portfolio manager assessing interest rate risk, a derivatives trader pricing bond options, or a risk officer running Monte Carlo simulations, the Vasicek model provides the analytical foundation for modeling uncertain rates. This guide covers the model’s stochastic differential equation, mean reversion mechanics, zero-coupon bond pricing formulas, yield curve implications, and practical limitations.

What Is the Vasicek Model?

The Vasicek model, introduced by Oldrich Vasicek in 1977, is a one-factor short-rate model that describes how the instantaneous interest rate (the “short rate”) evolves over time. It belongs to the class of Ornstein-Uhlenbeck processes — stochastic processes that exhibit mean reversion.

Unlike stock prices, which can grow exponentially over time, interest rates tend to fluctuate within a range dictated by monetary policy and macroeconomic conditions. The Vasicek model captures this behavior mathematically, making it useful for:

• Bond pricing — deriving closed-form prices for zero-coupon bonds
• Interest rate derivatives — pricing caps, floors, and bond options
• Risk management — simulating interest rate scenarios for VaR calculations
• Asset-liability management — modeling duration gaps for banks and insurers

The Vasicek model is the simplest member of the affine term structure models family, where bond prices have a specific exponential-affine form that allows analytical solutions. More complex models like the Cox-Ingersoll-Ross (CIR) model and Hull-White model build directly on Vasicek’s framework.

The Vasicek Stochastic Differential Equation

The Vasicek model describes the evolution of the short rate r(t) through a stochastic differential equation (SDE):

Where:

• r — the instantaneous short rate (not a Treasury yield or YTM)
• κ (kappa) — the mean reversion speed; how quickly rates return to the long-term mean
• θ (theta) — the long-term mean level; the equilibrium rate toward which r is pulled
• σ (sigma) — the volatility of rate changes (annualized)
• dW — the increment of a Wiener process (Brownian motion); the random shock

This SDE is an Ornstein-Uhlenbeck process — one of the few stochastic processes with a known analytical solution. The short rate at any future time t, given an initial rate r₀, follows a normal distribution with known mean and variance.

The Short Rate Distribution

One of the Vasicek model’s key advantages is that the future short rate r(t) has a closed-form distribution. Given an initial rate r₀ at time 0, the rate at time t is normally distributed:

As t → ∞, the expected rate converges to θ and the variance converges to σ²/2κ. The stationary distribution is N(θ, σ²/2κ).

Mean Reversion in Interest Rates

Mean reversion is the empirical observation that interest rates tend to fluctuate around a central level rather than trending indefinitely in one direction. This behavior contrasts sharply with stock prices, which can compound over decades.

The economic intuition for mean reversion comes from central bank policy and economic cycles:

• High rates slow economic activity, eventually prompting central banks to cut
• Low rates stimulate borrowing and inflation, eventually prompting central banks to raise
• Arbitrage forces prevent rates from diverging too far from fair value

The half-life of mean reversion measures how long it takes for rates to move halfway back to the long-term mean:

For example, if κ = 0.1 (per year), the half-life is about 6.9 years. If κ = 0.3, the half-life drops to 2.3 years. Empirical estimates of κ for short rates typically range from 0.05 to 0.30, implying half-lives of 2 to 14 years.

Zero-Coupon Bond Pricing Under Vasicek

The Vasicek model yields a closed-form solution for zero-coupon bond prices. This is one of the model’s most valuable properties — it allows analytical pricing without Monte Carlo simulation.

The price at time t of a zero-coupon bond maturing at time T, under the risk-neutral measure, is:

The functions B(τ) and A(τ) are:

Note that these formulas use risk-neutral parameters. The parameters estimated from historical rate data (physical measure) differ from those implied by current bond prices (risk-neutral measure). For pricing, always use risk-neutral calibration.

Coupon bonds are priced as the sum of their cash flows, with each cash flow treated as a separate zero-coupon bond. For more on basic bond pricing and yield to maturity, see our dedicated article.

How to Calculate Vasicek Bond Prices

Let’s work through a complete numerical example to see the Vasicek bond pricing formula in action.

The Vasicek Yield Curve

The Vasicek model generates an entire term structure of interest rates — the relationship between yields and maturities. The continuously compounded yield for a τ-year zero-coupon bond is:

As maturity τ → ∞, the yield converges to a long-run yield:

The long-run yield is always below the long-term mean θ due to the “convexity adjustment” term σ²/2κ². Higher volatility or slower mean reversion (lower κ) increases this adjustment.

The Vasicek model can produce three yield curve shapes, depending on the relationship between the current short rate and model parameters:

The exact boundaries between these shapes depend on all model parameters (κ, θ, σ), not just the relationship between r and θ. Higher volatility σ or slower mean reversion (lower κ) shifts the boundaries.

For more on the relationship between short rates and the term structure, see our article on spot rates and forward rates.

Vasicek Model vs Other Short-Rate Models

The Vasicek model is the simplest analytical short-rate model, but it has limitations that led to the development of alternative models. Here’s how the main one-factor models compare:

When to use each model:

• Vasicek — best for simple analytics, teaching, or when negative rates are acceptable. Ideal starting point for understanding rate modeling.
• CIR — when negative rates must be prevented. The square-root volatility structure keeps rates non-negative when started positive. The Feller condition (2κθ > σ²) ensures rates stay strictly positive. Used in credit modeling and inflation.
• Hull-White — when you need to match the current observed term structure exactly. Essential for derivative pricing where arbitrage-free calibration is required.

All three are one-factor models with a single source of randomness. They cannot separately capture parallel shifts, slope changes, and curvature changes in the yield curve. For richer dynamics, see the Heath-Jarrow-Morton framework or multi-factor extensions.

Limitations of the Vasicek Model

Despite its elegance, the Vasicek model has several important limitations that practitioners must understand:

1. Negative Rates Are Possible — The normally distributed short rate can go negative with non-zero probability. While this matches post-2008 reality in some markets, it can cause numerical issues in applications requiring positive rates.

2. Constant Parameters — The model assumes κ, θ, and σ are constant over time. In reality, volatility regimes shift and long-term rate expectations change. The Hull-White model addresses this with time-varying parameters.

3. Single Factor — With only one source of randomness, the model can only capture one type of yield curve movement at a time. It cannot simultaneously model changes in level, slope, and curvature — the three principal components that drive most yield curve variation.

4. Risk-Neutral vs Physical Measure — Parameters estimated from historical data (physical measure) differ from those needed for pricing (risk-neutral measure). The market price of risk must be specified to bridge this gap.

5. Mean Reversion May Not Hold — While interest rates have historically mean-reverted, structural changes in monetary policy or the economy could alter this behavior. The model assumes the mean-reverting structure persists indefinitely.

For managing interest rate risk in practice, the Vasicek model provides useful intuition, but more sophisticated models are typically required for production trading systems.

Common Mistakes

When working with the Vasicek model, watch out for these common errors:

1. Confusing the Short Rate with Treasury Yields — The instantaneous short rate r(t) in the Vasicek model is a theoretical construct, not a directly observable rate. It is not the same as the 3-month T-bill rate, fed funds rate, or any Treasury yield. These market rates are related but not identical.

2. Mixing Physical and Risk-Neutral Parameters — Parameters estimated from historical rate movements (physical measure) cannot be directly used for pricing. Bond pricing requires risk-neutral parameters, which incorporate the market price of interest rate risk.

3. Inconsistent Units — All parameters must use consistent time units. If κ is annual, then σ must be annual volatility and τ must be in years. Mixing daily volatility with annual mean reversion speed produces nonsensical results.

4. Confusing Mean Reversion Speed with Half-Life — Higher κ means faster reversion (shorter half-life), not slower. The half-life is ln(2)/κ, so doubling κ cuts the half-life in half.

5. Ignoring Negative Rate Probability — In Monte Carlo simulations, negative rate paths will occur. If your application cannot handle negative rates (e.g., certain option pricing formulas), you should use CIR or another non-negative model. Ad hoc fixes like flooring rates at zero change the model’s statistical properties and can introduce bias.