The Role of Risk-Neutral Probabilities in Enhancing Market Price Forecasts and Investment Strategies

Breeden-Litzenberger method is the basis for the calculation of market-based risk-neutral probabilities, for forecasting future asset prices and market trends. This article assess the predictive power of these probabilities, and establishes their efficacy by considering the confidence intervals provided by the Federal Reserve. Specifically, this article using Lasso regression, and obtain an R-squared of 87.9% in the case of return prediction and 79.4% in the case of volatility for training sets, while the KNN classifier achieves a 78.4% in-sample accuracy for reversal detection. By using these probabilities and historical pricing, the strategies produce backtest results of a 37% and 40% excess return over passive strategies, which shows that the use of risk-neutral probabilities can serve the purpose of predicting the market despite market financial and other constraints.