Day 5: Geometric Brownian Motion

Day 5: Geometric Brownian Motion#

Today we keep with variations of the Brownian motion! Geometric Brownian Motion (GBM) is a widely used stochastic process in mathematical finance, economics, and other disciplines to model the dynamics of positive variables that evolve over time under uncertainty, such as stock prices, interest rates, and commodity prices. The advantage of modelling through this process lies in its universality, as it represents an attractor of more complex models that exhibit non-ergodic dynamics.

Definition#

The Geometric Brownian motion can be defined by the following Stochastic Differential Equation (SDE)

dX_t=μX_tdt+σX_tdW_t,t>0,

with initial condition X_0=x_0>0, and constant parameters μR, σ>0. Here, W_t denotes a standard Brownian motion.

In order to find its solution, let us set Y_t=ln(X_t). Then, Ito’s formula implies

Y_t=Y_0+(μ12σ2)t+σW_t,

or equivalently

ln(X_t)=ln(x_0)+(μ12σ2)t+σW_t.

From this expression, we obtain the solution as follows:

X_t=x_0exp{(μ12σ2)t+σW_t},t0.

Thus, for each t>0, the variable ln(X_t)|X_0 (which will be denoted simply by ln(X_t)) follows a normal distribution –since it can be expressed as a scaled marginal of the Brownian motion plus a deterministic part. Using the properties of the standard Brownian Motion we can calculate its expectation and variance, and verify that

ln(X_t)N(ln(x_0)+(μ12σ2)t,σ2t),

and consequently X_t follows a Log-normal distribution.

🔔 Random Facts 🔔#

  • In 1952, Paul Samuelson and Robert Merton adapted the Wiener process into what we now recognise as Geometric Brownian Motion. Their modifications addressed the (so-called) shortcoming of the Bachelier’s model by ensuring that the process remains positive.

  • GBM serves as the foundation for the famous Black-Scholes-Merton model for option pricing.

  • The asymptotic properties of geometric Brownian motion (GBM) are determined by the quantity μσ22:

  • μσ22>0: As t approaches infinity, X_t approaches infinity with probability 1

  • μσ22<0: As t approaches infinity, X_t approaches zero with probability 1

  • μσ22=0: As t approaches infinity, X_t has no limit with probability 1

More to Read#

P.s. If you are curious about probability distributions visit the Advent Calendar 2023