The worst case for real options (Q613589)

First, the authors present a standard investment problem, where the project's value follows a geometric Brownian motion and the project has finite life-time. Uncertainty is represented by a filtered probability space \((\Omega,({\mathcal F}_t)_{t\geq 0},Q)\). The firm faces an investment opportunity yielding a stochastic profit modeled as a geometric Brownian motion, its dynamics under \(Q\) given by \[ d\pi_t= \pi_t(\mu dt+\sigma dB_t). \] Here, \(\pi_0>0\), and \(\mu\) and \(\sigma\) are positive constants denoting drift and instantaneous standard deviation, respectively; \(dB_t\) is the increment of a Wiener process. The project in assumed to have a finite time \(\tau\). The discount factor which consists of two components (the rate of degreciation \(\gamma\) and the interest rate \(\rho\)) is imposed exagenously, following a deterministic process \[ dR_t=-(\rho+ \gamma)R_t dt = -rR_t dt\quad(\text{here},\;\mu< r). \] The firm wants to determine the optimal time to pay a sunk cost \(I\) to get access to the profit stream \((\pi_t)\). If the investment is undertaken at time \(t\), the value \(W_t\) of the project is given by \[ W_t= E^Q\Biggl[\int^{t+\tau}_t R_s R^{-1}_t \pi_s ds\, \Biggl|\, {\mathcal F}_t\Biggr]= \pi_t(1- \exp(-(r- \mu)\tau)) (r-\mu)^{-1}. \] The authors determine the optimal investment time and the corresponding value function. Then the authors obtain corresponding results for generalizations of the above problem where the firm is not able to assign a precise probability to future alternatives which is referred to as Knightian uncertainty or ambiguity.