An optimal trading rule under a switchable mean-reversion model (Q2247920)

The authors consider a switchable mean-reversion model for the log-price of an asset. The switching process is modelled as a two-state Markov chain with only one jump after exponentially distributed random time interval. The trading rule allows buying and selling an asset sequentially over time. Two net positions are flat (no stock holding) or long (with one share of stock holding). The objective is to determine a sequence of trading times to maximize an overall return. The corresponding value functions are characterized by a set of quasi-variational inequalities. A closed-form solution is obtained under suitable conditions. The sequence of trading times can be given in terms of a set of threshold levels. Some numerical examples illustrate the results.