Brownian Motion

Quantitative Finance · advanced · CC-BY-4.0

Brownian motion (also called a Wiener process) is a continuous-time stochastic process in which the change in value over any time interval is normally distributed with mean zero and variance equal to the length of the interval, with increments being independent. It is the mathematical foundation of modern options pricing, stochastic calculus, and the modeling of asset price dynamics in continuous time.

Key takeaways

• A standard Brownian motion W(t) has W(0) = 0, continuous paths, independent increments, and W(t) − W(s) ~ N(0, t−s) for any t > s.
• Geometric Brownian motion (GBM) — the log-price version — is the stochastic process underlying the Black-Scholes model and models asset prices as positive, continuously compounding with a drift (μ) and diffusion (σ) component.
• Brownian motion has quadratic variation equal to t (not zero, as for differentiable functions), which requires the use of Itô's lemma rather than ordinary calculus when applying functions to Brownian-motion-driven processes.
• The 'no-drift' martingale property — when μ=0, W(t) is a martingale, meaning E[W(t)|F_s] = W(s) — is foundational to risk-neutral pricing and the construction of hedging strategies.
• Real asset prices depart from pure GBM in important ways: fat tails (excess kurtosis), volatility clustering (GARCH effects), and mean reversion in some assets all indicate that Brownian motion is a first-order approximation, not a complete description of price dynamics.

Explanation

Standard Brownian motion, denoted W(t) or B(t), satisfies four conditions: (1) W(0) = 0; (2) the process has independent increments — changes over non-overlapping intervals are statistically independent; (3) increments are stationary — W(t) − W(s) depends only on t−s, not on s itself; and (4) W(t) − W(s) ~ N(0, t−s), meaning the increment is normally distributed with variance equal to the elapsed time. The resulting paths are continuous but nowhere differentiable — they exhibit infinitely jagged, fractal-like behavior at any time scale.

In finance, raw Brownian motion is rarely used directly. Instead, practitioners work with Geometric Brownian Motion (GBM), which models the log-price increment: dS = μS dt + σS dW, where μ is the drift (expected return), σ is the volatility, and dW is the Brownian increment. The solution to this SDE is: S(t) = S(0) × exp[(μ − σ²/2)t + σW(t)], ensuring prices remain positive. This is the direct basis for the Black-Scholes model (with μ replaced by the risk-free rate r under the risk-neutral measure Q) and for most closed-form derivatives pricing formulas.

Itô's lemma is the key tool for applying calculus to Brownian-motion-driven processes. Unlike ordinary calculus, where a Taylor expansion to first order is exact, Brownian paths have non-zero quadratic variation (∫₀ᵀ (dW)² = T), which means second-order terms in the Taylor expansion do not vanish. For a function f(t, S) of time and a GBM process S, Itô's lemma gives: df = (∂f/∂t + μS ∂f/∂S + ½σ²S² ∂²f/∂S²) dt + σS ∂f/∂S dW. Setting up a delta-hedged portfolio and applying Itô's lemma to the option price is precisely how Black and Scholes derived their famous PDE.

Extensions to the basic Brownian framework include: mean-reverting Ornstein-Uhlenbeck processes (used for interest rate models like Vasicek); jump-diffusion models (Merton, 1976) that add a Poisson jump component to capture fat-tail events; stochastic volatility models (Heston, SABR) where σ itself follows a separate diffusion; and fractional Brownian motion (with Hurst exponent H ≠ 0.5) to model long-range dependence in asset returns. Each extension addresses a specific empirical deficiency of pure GBM while preserving the mathematical tractability of the Brownian framework.

Formula

Standard Brownian Motion: dW ~ N(0, dt)
Geometric Brownian Motion: dS = μS dt + σS dW
Solution: S(t) = S(0) × exp[(μ − σ²/2)t + σW(t)]
Itô's Lemma: df = (∂f/∂t + μS ∂f/∂S + ½σ²S² ∂²f/∂S²)dt + σS ∂f/∂S dW

Example

Consider pricing a European call option using Black-Scholes. The underlying asset price S = $100, strike K = $100, time to expiration T = 1 year, risk-free rate r = 5%, volatility σ = 20%. Under GBM, S(T) = 100 × exp[(0.05 − 0.02)×1 + 0.20×W(1)] where W(1) ~ N(0,1). Applying the Black-Scholes formula (which analytically integrates the lognormal distribution implied by GBM), d1 = [ln(100/100) + (0.05 + 0.02)×1] / (0.20×1) = 0.35, d2 = 0.15, call price = 100×N(0.35) − 100×e^(−0.05)×N(0.15) ≈ $10.45. The entire derivation relies on Brownian motion's properties — specifically, the normality of log-returns and the Markov property that makes delta hedging feasible.

Related terms

Autoregressive Model Basis Black Scholes Model Call Option Cross Sectional Momentum Delta Geometric Brownian Motion Hedging Hurst Exponent Interest Rate Option Random Walk