糖心TV

"Randomized Strategies and Prospect Theory in a Dynamic Context"

We consider CPT agents who face optimal timing decisions in a dynamic setting, e.g. when to stop gambling in a casino or when to liquidate stocks. It is known that probability weighting leads to time inconsistency and a naïve agent may follow a different strategy to that which he initially planned to follow (Barberis (2012)).

However, it also leads to another feature which has not been considered to date - agents may prefer randomised strategies to pure strategies, i.e. CPT agents should spin a coin to help them reach an optimal solution. In the discrete model of Barberis (2012) we show that allowing randomized strategies leads to significant gains in CPT value. In the continuous model of Ebert and Strack (2014) we show their extreme conclusion that naïve CPT agents gamble “until the bitter end’’ is no longer valid if the agent has a coin in his pocket.