Skip to content
All library documents

Convertible Bond PDE and Boundary Conditions with Stochastic Rates

Article Quant Q&A · Author: user16749

Summary

The document presents a two-factor pricing setup in which the stock price follows a diffusion with constant volatility and the short rate follows a mean-reverting process. Assuming the two Brownian drivers are uncorrelated, the response gives a generalized Black–Scholes PDE with diffusion and drift terms for both state variables, along with the discounting term. It then discusses terminal payoff and boundary conditions, including zero value at a zero stock price and stock-like value at a sufficiently high stock price.

The response mixes a European call example with convertible-bond boundaries, so its conditions should not be treated as a complete, validated convertible-bond specification. In particular, it offers a proposed high-rate boundary and a homogeneous Neumann condition at very low rates without deriving their suitability. Features such as conversion terms, callability, putability, correlation, and numerical-domain choices may require different conditions.

Key ideas

  • With uncorrelated stock and rate drivers, the stated model leads to a two-state-variable pricing PDE.
  • The short rate is modeled as mean reverting, while the stock has constant volatility.
  • The response gives a maturity payoff and stock-price boundary examples for a call and convertible bond.
  • Boundary choices for extreme interest rates are proposed but not fully justified.
  • Convertible features and correlated drivers may require changes to the PDE or boundary conditions.

Tags

Full text
# Boundary Condition for Convertible Bond under Two-factor Model Interest Rate


# Boundary Condition for Convertible Bond under Two-factor Model Interest Rate












I want to find Boundary conditions for Convertible Bond under Two-factor Model Interest Rate.The portfolio contains stock where stochastic differential equation for the stock price is \begin{align} ds_t=rS_t+\sigma S_tdW_1(t) \end{align} where $\sigma$ is constant and dynamics of $r$ as follow \begin{align} dr_t=\kappa(\theta-r_t)dt+\Sigma dW_2(t) \end{align}

## Answer by user16651 (score 0, accepted)

https://quant.stackexchange.com/a/18644

For a two-factor option pricing model with underlying variables $S$ and $r$ defined as above, if we assume there is no correlation between the two Wiener processes $W_1$ and $W_2$, one finds the generalized Black-Scholes PDE \begin{align} V_t+\frac{1}{2}\sigma^2V_{SS}+r\,S\,V_S-r\,V+\frac{1}{2}\Sigma\,^2\,V_{rr}+\kappa(\theta-r)V_r=0 \end{align} This equation is subject to initial and boundary conditions. Generally speaking, derivative pricing models for different financial scenarios may share a similar pricing partial differential equation (PDE) with adjusted parameters and boundary conditions.Boundary conditions defining two portfolios will be considered. The first set of conditions will describe a European call stock option. The second set of conditions models a convertible bond. The stock price $S$ and interest rate, $r$.

- At the maturity time T, the call option price will be the payoff function \begin{align} V(S,r,T)=\max\{S-K,0\} \end{align}

- At $S = 0$, the option is worthless: \begin{align} V(0,r,T)=0 \end{align}

- For large stock price $S_{\max}$, it is almost certain that the bond will be converted to one share of the stock. Hence \begin{align} V(S_{\max},r,t)=S_{\max} \end{align}

- When $r_t$ is infinitely large, the bond component tends to zero. Since we do not enforce any time-dependent constraints of puttable and callable features, the upper bound and the lower bound to the price of the convertible bond are $\max\{S,\infty\}$ and $\max\{S,0\}$ respectively. Therefore we define the boundary condition as \begin{align} V(S,r_{\max},t)=\min\left\{\max\{S,\infty\},\max\{S,0\}\right\}=S \end{align}

- For a very small interest rate, we use homogeneous Neumann condition suggested by Bermudez and Nogueiras.

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.