Apply
Itô’s lemma to the derivative price, hold the amount of stock needed to match its instantaneous Brownian exposure, and put the remaining value in a
cash account. A self-financing strategy that replicates the option must then match its drift as well. With no dividends and a constant
risk-free rate r, the result is
Vt+21σ2S2VSS+rSVS−rV=0.
At one instant, a call and its stock are exposed to the same unpredictable shock. A call worth
10 can move by roughly
0.60 for each
1 stock move if its
delta is
0.60. Holding
0.60 shares reproduces that first-order movement. The stock position usually costs more than the call, so the
replicating portfolio also holds a negative cash balance. The interest on that balance is part of the portfolio's return.
The drift of the real-world stock process appears on both sides of the
replication equation. It cancels because the amount of stock exposed to that drift is the same on both sides.
Volatility survives through curvature. A convex
payoff reacts differently to two small moves in opposite directions, and the squared Brownian increment contributes at the order of elapsed time.
Match the exposures before matching the drift
Treat V(t,S) as a price per option and VS as its value change per dollar of stock move. Holding VS shares matches the coefficient of dW because the stock itself contributes σSdW per share. The cash position is whatever remains of the option value after buying those shares. The hedge is self-financing if changes in the number of shares are paid for from cash. That is why the portfolio gains contain ΔdS and the interest on cash, with no extra injection of money when delta changes.
Suppose dS=μSdt+σSdW. Before expiration, Itô’s lemma gives
dV=(Vt+μSVS+21σ2S2VSS)dt+σSVSdW.
A portfolio worth V holding Δ shares has cash V−ΔS. When it is self-financing, its gains are
dV=ΔdS+r(V−ΔS)dt.
Self-financing means trades between stock and cash are paid from within the portfolio. It is why this gains equation contains
ΔdS even when
Δ changes with time. Comparing diffusion terms gives
Δ=VS. Comparing drifts and cancelling
μSVS yields the PDE. A
European call supplies the terminal condition
V(T,S)=(S−K)+. Other payoff conditions can lead to other solutions of the same pricing equation.
Write the two drift coefficients side by side. Itô gives Vt+μSVS+σ2S2VSS/2. The hedge earns μSVS+r(V−SVS) from the stock and cash. The μSVS terms are identical because the two strategies hold the same instantaneous stock exposure. Subtracting them leaves Vt+σ2S2VSS/2=r(V−SVS). Rearranging moves the stock financing term to the left, producing rSVS−rV. This explains why the physical drift drops out even though the stock itself still has that drift.
What the equation says about theta
Take a moment when S=100, V=10, VS=0.60, and VSS=0.02. Use r=5% and σ=20% per year. The cash position is 10−60=−50, so it contributes −2.50 per year at that instant. The curvature term contributes 21(0.20)2(100)2(0.02)=4.00 per year. The PDE therefore gives Vt=−2.50−4.00=−6.50 per year.
Now let the assumed physical stock drift be
8%. The option’s local expected change is
−6.50+4.80+4.00=2.30 per year. The portfolio earns
4.80−2.50=2.30 per year in
expectation, and both sides carry the same random term. Changing
μ changes both expected gains by the same amount. These figures describe an instantaneous calculation, so a one-year forecast cannot be read off this local rate.
The
options pricer lets you vary spot while holding volatility fixed and see the curvature that enters the PDE.
References and next steps