Timing versus Buy and Hold: A Model for Determining Predictive Accuracy Required for Superior Performance

In this application, we develop a model to simulate the decisions of a trader whose subjective distribution of returns may be correlated with realized stock returns. Using empirically estimated parameters from stocks in the CRSP database, we obtain performance data on a number of measures, including mean returns, volatility, the Sharpe measure, and the probability of a correct trading decision. The model suggests that daily trading of a portfolio of 20 volatile stocks gives a Sharpe measure better than that of buying and holding the S&P 500 when timing accuracy is 53% or better.