Author(s)

Yiren Liu, Yuli Bian

Corresponding Author

Yiren Liu

ABSTRACT

Using assessing pricing model and mathematical method, this paper find both return timing and volatility timing are shown to be profitable and their result has proven to be statistically sound. Given the promising performance, we are thinking about exploring ways to combine these two timing strategies. Gradient boost machine learning algorithm are also used to predict SPY return with 20 variables including some constructed variables and develop trading strategies using the forecasted return and stock volatility. After getting the forecast return, using the forecast return to determine the weight of buy and hold S&P500 to create the portfolio. The simulated strategy shows a significant increase in Sharpe ratio compared to the return of just hold S&P500. We also extend our strategy from daily rebalancing to monthly trading which also shows significant Sharpe ratio.

KEYWORDS

S&P500; Economic Indicators; Technical Indicators; Forecast; Regression