Costly switching and investment volatility

This paper extends the literature on the optimal switching rule between two investments by considering the case where switching between investments is costly. The model builds on the classic framework of the multi-armed Bandit problem by explicitly incorporating two key assumptions. First, switching investments is costly. Second, only the investment operated by the investor evolves as a random walk. The objective of the investor is to maximize the discounted sum of expected net profits over the infinite horizon. The main result is that when the volatility of profits from investments increases, so does the minimum profit gain needed for an investor to switch investments.