Glossary term
Geometric Brownian Motion
Geometric Brownian motion is a stochastic process often used to model asset prices that evolve with drift and random volatility shocks.
Updated
Read time
What Is Geometric Brownian Motion?
Geometric Brownian motion, or GBM, is a stochastic process often used to model asset prices that evolve with drift and random volatility shocks. It is famous in finance because it sits behind the classic Black-Scholes-Merton option pricing framework.
GBM models percentage changes rather than simple dollar changes. That feature keeps modeled prices positive and produces a lognormal price distribution under the model's assumptions.
Key Takeaways
- Geometric Brownian motion models asset prices as a continuous-time stochastic process.
- It includes a drift term and a volatility term.
- It is a foundational assumption in the Black-Scholes-Merton model.
- GBM implies lognormally distributed prices under its assumptions.
- Real markets can have jumps, fat tails, changing volatility, and liquidity shocks that GBM does not capture.
Core Model Form
A common GBM expression is:
In this expression, St is the asset price at time t, μ is the drift rate, σ is volatility, and Wt represents Brownian motion.
For example, an option model may assume a stock price follows GBM so the model can simulate many possible future price paths. The model does not know which path will occur; it uses the distribution of paths to value contingent payoffs.
What GBM Assumes
Assumption | Meaning |
|---|---|
Continuous trading path | Prices move continuously in the model. |
Constant volatility | The volatility input does not change over time. |
Lognormal prices | Prices stay positive and returns follow the model's distribution. |
Random shocks | Uncertainty comes through Brownian motion increments. |
Model Risk in Practice
GBM is elegant, but it is not a complete description of markets. Asset prices can jump after earnings, policy news, defaults, geopolitical events, or liquidity shocks. Volatility can cluster and change over time. Return distributions can have fatter tails than the model expects.
That does not make GBM useless. It makes it a baseline model. Many more advanced models are best understood as attempts to relax one of GBM's clean assumptions.
GBM is often used because it is mathematically tractable. The model gives analysts a clean way to move from assumptions about drift and volatility to a distribution of possible future prices. That tractability is why it became so influential in option pricing and simulation.
The same simplicity is also the weakness. If the asset has jump risk, changing volatility, trading halts, or strong mean reversion, GBM may give a neat answer to the wrong model. The output should be treated as a baseline rather than a full market description.
GBM is useful because it turns uncertainty into a tractable model, not because markets literally follow the equation. Real prices can gap, volatility can cluster, trading can become illiquid, and returns can have fatter tails than the model assumes.
That distinction matters in derivatives and risk management. A model can be good enough to organize assumptions while still needing stress tests, scenario analysis, and judgment about events that are not well captured by a smooth continuous process.
Where It Still Helps
GBM remains useful because it provides a common language for drift, volatility, time, and randomness. Even when analysts know the assumptions are imperfect, the model can help compare option values, test sensitivities, and explain why volatility and time affect derivative prices.
The practical discipline is to avoid confusing model elegance with market truth. GBM can be a reasonable starting point for liquid assets over ordinary conditions, but it needs additional tools when jump risk, credit events, policy shocks, or liquidity breakdowns dominate the payoff.
The Bottom Line
Geometric Brownian motion is a foundational asset-price model that combines drift and random volatility shocks. It is useful for option pricing and simulation, but real markets often behave less smoothly than the model assumes.