Minimum Variance Hedge Ratios using Short-Lived Arbitrage Model and Artificial Neural Network

Motivated by the short-lived arbitrage model, we jointly imply volatilities and virtual interest rates of options to develop minimum variance hedge ratios. We use artificial neural networks to capture the dynamics between the underlying asset price and correlated state variables, which improves the hedging performance. Based on several error metrics, we demonstrate that the minimum variance hedge ratio outperforms the traditional Black-Scholes delta hedge and a recent benchmark hedge based on the practitioner Black-Sholes model. We present hedging results for options on S&P 500, Russell 2000, and Nasdaq indices