Abstract: This paper investigated the performance of non-parametric option hedging methods in the Chinese market, in which investors minimized their single-period mean-squared hedging errors. Experiments were conducted using SSE (Shanghai stock exchange) 50 ETF (exchange traded fund) options. It was proposed the use of feed-forward neural networks and linear regression for model mapping from option-observable variables to hedging strategies. Results showed that non-parametric methods significantly outperformed the benchmark parametric models with hedging errors reduced by over 10% due to the fact that non-parametric models could capture the leverage effect in the SSE 50 ETF option market.

Key words: option hedging, machine learning, linear regression, leverage effect