Deriving a Black–Scholes Price for a European Log Contract
Summary
The document derives the value of a European claim paying the logarithm of the terminal asset price relative to its strike, assuming a non-dividend-paying asset under the Black–Scholes framework. It proposes a price that is affine in the log of the current spot-to-strike ratio, substitutes that form into the pricing PDE, and matches the coefficients of the constant and log terms. The terminal payoff supplies the boundary conditions for the two time-dependent coefficients.
Solving the resulting ordinary differential equations gives a discounted log exposure and a time-dependent adjustment involving the risk-free rate and volatility. The result illustrates how a PDE ansatz can reduce a pricing problem to simpler equations. The derivation relies on constant interest and volatility rates and the stated Black–Scholes assumptions. It does not discuss dividends, transaction costs, model calibration, or empirical performance; the claim is a mathematical pricing example rather than a tested trading strategy.
Key ideas
- An affine function of the log spot-to-strike ratio is used as the price ansatz.
- The terminal payoff determines the boundary values for both time-dependent coefficients.
- Substitution into the Black–Scholes PDE yields separate ordinary differential equations by matching terms.
- The resulting price depends on the risk-free rate, volatility, and time remaining.
Tags
Full text
# Pricing log-contract with Black-Scholes PDE
# Pricing log-contract with Black-Scholes PDE
I was wondering if someone could help me with a problem, regarding the Merton Black Scholes PDE. I have an exam soon and this question on an old exam has been bothering me and a friend for quite a while. We simply don't get it.
The question goes this way:
The payoff of a so-called European log-contract is $g \left( S_T \right) = \ln \left( S_T / K \right)$ where $K$ is the strike price and S is a risky non-dividend MBS asset. Find the price $c(s,t)$ of such asset. Hint: Use the Black-Scholes PDE and give yourself the fact that c has the following form:
\begin{equation} c(s,t) = a(t) + b(t) \ln(s/K) \end{equation}
Find the functions $a(t)$ and $b(t)$.
I've tried to figure out the solution and see if there's anything online, but nothing works. Using the BS-PDE is not helping. All help and advice would be GREATLY appreciated!
## Answer by LocalVolatility (score 5)
https://quant.stackexchange.com/a/31399
The B/S PDE for a contingent claim $V(S, t)$ is
\begin{equation} \frac{\partial V}{\partial t} + r S \frac{\partial V}{\partial S} + \frac{1}{2} \sigma^2 S^2 \frac{\partial V^2}{\partial S^2} - r V = 0 \end{equation}
subject to the terminal condition $V(S, T) = \ln(S / K)$. According to the hint, the solution to $V(S, t$) takes the form
\begin{equation} V(S, t) = a(t) + b(t) \ln(S / K). \end{equation}
Since the terminal condition has to hold for all $S$, it follows that $a(T) = 0$, $b(T) = 1$.
Take the partial derivatives of the guess to get
\begin{eqnarray} \frac{\partial V}{\partial t}(S, t) & = & a'(t) + b'(t) \ln(S / K),\\ \frac{\partial V}{\partial S}(S, t) & = & b(t) \frac{1}{S},\\ \frac{\partial^2 V}{\partial S^2}(S, t) & = & -b(t) \frac{1}{S^2}.\\ \end{eqnarray}
Substituting back yields
\begin{equation} a'(t) + b'(t) \ln(S / K) + r b(t) - \frac{1}{2} \sigma^2 b(t) - r a(t) - r b(t) \ln(S / K) = 0. \end{equation}
Since for each fixed $t$, this solution has to hold for all values of $S$, we collect terms containing $\ln(S / K)$ and get the ODEs
\begin{eqnarray} 0 & = & a'(t) + \left( r - \frac{1}{2} \sigma^2 \right) b(t) - r a(t),\\ 0 & = & \left( b'(t) - r b(t) \right) \ln(S / K) \end{eqnarray}
The solution to the second ODE that satisfies $b(T) = 1$ is
\begin{equation} b(t) = e^{-r (T - t)}. \end{equation}
The solution to the first ODE that satisfies $a(T) = 0$ is
\begin{equation} a(t) = b(t) \left( r - \frac{1}{2} \sigma^2 \right) (T - t). \end{equation}
Combining these results yields
\begin{equation} V(S, t) = e^{-r (T - t)} \left[ \left( r - \frac{1}{2} \sigma^2 \right) (T - t) + \ln(S / K) \right]. \end{equation}Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.