The right time to sell a stock whose price is driven by Markovian noise (Q1769428)

This paper deals with the problem of finding the optimal time to sell a stock. The stock price \(Y_t\) fluctuates according to the random ordinary differential equation \(dY_t= Y_t(\mu dt+ \sigma\xi(t)dt)\), where \(\mu\) and \(\nu\) are constants and \((\xi(t))\) is an alternating Markov renewal process with value in \(\{\pm 1\}\), with an exponential renewal time with mean \(1/\alpha\). The problem is to find a stopping time \(\tau\), which maximizes the expected reward with fixed transaction cost \(a\) and discount rate \(\rho\), \(E[\exp^{-\rho\tau}(Y_\tau- a)]\), and the value function. Considering the couple \((Y(t),\xi(t))\), the authors show how the structure of solution depends on the relationship between four parameters \(\rho\), \(\mu\), \(\sigma\) and \(\lambda\). 1. The critical value of \(\rho\), under which the value function is finite, is given by \(\rho= \mu-\lambda+ \sqrt{\sigma^2+\lambda^2}\). 2. According to the possible relationship between four parameters, the authors investigate the validity of the principle of smooth fit and, using this, obtain a complete and essentially explicit solution to the problem, which exhibits a surprisingly rich structure. 3. The corresponding result for Black and Scholes model is obtained as an appropriate limit case, since \(\sqrt{\lambda}\int^t\xi(s)\,ds\) converges to a Brownian motion, as \(\lambda\to\infty\).