All option guides
Stochastic calculus12 min read
Itô's Lemma in Option Pricing Explained
Learn why stochastic functions need an extra second-order term, how Itô's lemma expands option value, and how delta hedging leads to a pricing equation
Prepared by Mark · Primary sources below
Direct answer
Itô's lemma is the chain rule for a stochastic process. A Brownian increment is of order √dt, so its square contributes at order dt. The surviving second derivative connects option gamma, volatility, and time evolution
Ordinary chain rules miss stochastic variation
For a smooth deterministic path x(t), the change in f(t, x) is locally described by its time derivative and first derivative with respect to x
A diffusion path is continuous but nowhere smoothly differentiable in the ordinary sense. Its random increment over a short interval scales like the square root of time
That scaling makes a second-order Taylor term as large as an ordinary time term. Dropping it removes the effect that later appears as gamma in option pricing
Quadratic variation keeps the second-order term
Write a Brownian increment as dW. In Itô bookkeeping, dW² contributes dt, while dt² and dt·dW vanish at the relevant order
This is shorthand for a pathwise limit: the sum of squared Brownian increments converges to elapsed time as a partition becomes finer
It is not ordinary algebra saying every random shock squared equals time. The rule belongs to stochastic integration and its limiting convention
The lemma expands a function of a diffusion
Suppose dS = μSdt + σSdW and an option value is V(t, S). Itô's lemma gives its local stochastic change
The drift is Vₜ + μS Vₛ + one-half σ²S²Vₛₛ. The random term is σS VₛdW, where subscripts denote partial derivatives
Delta Vₛ multiplies the first-order shock, while gamma Vₛₛ enters the time-order drift through quadratic variation
Gamma turns volatility into expected local change
Even when the random price shock has zero conditional mean, convexity makes positive and negative moves affect option value asymmetrically
The one-half σ²S²Gamma term measures this local curvature contribution. Higher volatility increases the rate at which a curved payoff accumulates modeled value
This does not say long gamma earns a guaranteed profit. Theta, paid premium, realized paths, jumps, skew changes, and costs determine actual P&L
Delta hedging removes the modeled random term
Combine one option with minus its delta in the underlying. Under the diffusion assumptions, the dW terms cancel instantaneously
The remaining local portfolio is modeled as riskless, so no-arbitrage requires it to earn the risk-free rate after consistent dividend or carry treatment
Substituting that condition removes μ and produces the Black–Scholes partial differential equation. Replication, not a return forecast, sets the price relation
The formula also guides broader models
With several correlated diffusions, the expansion includes cross-second derivatives weighted by covariance. Those terms connect correlation with joint curvature
Changing the state process changes the generator and pricing equation. Local volatility, stochastic volatility, and rate models all use related Itô expansions
For functions that are not smooth enough, boundaries and kinks need weak, generalized, or numerical treatment rather than casual differentiation
Jumps and real hedging need extra terms
A jump process does not fit the continuous Brownian rule alone. Its change formula includes discontinuous jump contributions and often leads to an integro-differential equation
Real hedges are discrete and costly, so canceling an infinitesimal dW term does not eliminate gap, liquidity, parameter, or execution risk
Use Itô's lemma to understand model mechanics, then state the process, smoothness, measure, and trading assumptions before interpreting a price or Greek
Common questions
Why does Itô's lemma have an extra term?
Brownian increments scale with √dt, so their square is of order dt and the second-order Taylor term does not disappear
What does dW² = dt mean?
It summarizes quadratic variation in the stochastic limit; it is not an ordinary pointwise equality between a random number and time
How does Itô's lemma lead to Black–Scholes?
It expands the option's local change, after which a delta hedge cancels the diffusion shock and no-arbitrage sets the remaining return
Does Itô's lemma apply to price jumps?
The continuous version is insufficient. Jump models require an extended change formula with explicit discontinuous terms
Sources and further reading
Apply this idea to an option
Choose a contract and target to keep price, time, and volatility assumptions visible in one analysis
Analyze my option