Brownian Bridge

Definition

A Brownian bridge is a continuous-time stochastic process $\{W(t): 0 \leq t \leq T\}$ conditioned to start and end at specified values — most commonly $W(0) = 0$ and $W(T) = 0$. It is a Brownian motion "tied down" at both endpoints. Ball and Torous (1983) apply this process to bond price dynamics, where the terminal constraint $P(m;m)=1$ (bond returns face value at maturity) motivates the bridge specification for excess log holding returns.

Key Ideas

• SDE (Stochastic Differential Equation): $dW = -\frac{W}{1-t}\,dt + dZ$; the restoring-force term pulls the process toward zero as $t \to 1$.
• Variance: $E[W^2(t)] = t(1-t)$; rises from zero, peaks at $t=\tfrac{1}{2}$, returns to zero at maturity. Non-stationary from the terminal constraint.
• Bond price application (Ball-Torous 1983): Excess log holding return $\eta(t;m)$ is a bridge pinned at 0 at issue and 0 at maturity. The bond price satisfies $dP/P = a_1(P,t)\,dt + \sigma_1\,dZ$ with $\sigma_1$ constant for a given maturity.
• Preference-free bond option formula: Using the bond price (tradeable) as state variable and Merton's (1973) no-arbitrage framework yields a closed-form European call: $$C(K,\tau) = P(\tau;1)N(h_1) - P(0;\tau)KN(h_2)$$ where $v^2 = \sigma_1^2 + \sigma_\tau^2 - 2\rho\sigma_1\sigma_\tau$.
• Contrast with interest-rate models: Brennan-Schwartz (1977, 1982) and Courtadon (1982) use the interest rate (non-tradeable) as state variable, yielding preference-dependent solutions requiring numerical approximation; the Brownian bridge bond-price approach avoids both.

How It Works

The Brownian bridge emerges from conditioning a standard Brownian motion $B$ on the event $B(T)=0$. Explicitly, $W(t) = B(t) - (t/T)B(T)$. The SDE $dW = -W/(T-t)\,dt + dZ$ follows by Itô's lemma: the drift term removes accumulated deviation at a rate proportional to remaining time, forcing the process to zero by $T$.

For bond pricing, define $\eta(t;m) = \xi(t;m) - \mu(m)t$ where $\xi(t;m)$ is the log excess holding return. The terminal constraint $P(m;m)=1$ forces $\eta(0;m)=0$ and $\eta(m;m)=0$, so $\eta$ is a Brownian bridge. The resulting constant-$\sigma_1$ property of the bond price SDE enables closed-form option pricing via Merton's (1973) framework.

Why It Matters

• Preference-free bond option pricing: The Brownian bridge model yields the first closed-form, preference-free formula for European calls on default-free bonds. Investors can price options without agreeing on a theory of the term structure.
• Terminal constraints in finance: Many financial securities have natural terminal constraints (bond face value, barrier options, reset features) that the Brownian bridge handles automatically. Ball and Torous (1983) suggest broad applicability beyond debt options.
• Analytical tractability: The constant-$\sigma_1$ property of the bridge eliminates stochastic volatility complications, trading empirical generality for a clean closed form.

Open Questions

• The constant-$\sigma_1$ assumption is empirically restrictive; stochastic volatility extensions to bond option pricing remain active.
• Multi-factor term structure models (level/slope/curvature) are not captured by the single-bridge formulation.

Related

• Black-Scholes Option Pricing
• Rational Option Pricing Bounds
• Term Structure of Interest Rates
• Ball and Torous (1983)
• Clifford Ball
• Walter Torous