Variance Reduction Techniques in Variance Gamma Option Pricing

Research on options models is still relevant to help buyers determine the fairness of option prices. The Black-Scholes model assumes that the stock price is lognormally distributed, whereas, in real applications, the stock price data does not match this assumption because it has different skewness and kurtosis values from the normal distribution. This condition is more suitable to be solved by non-normal models such as the Gram-Charlier and Variance Gamma. To reduce the variances, there are Antithetic Variate and Importance Sampling techniques. In this paper, we discuss an empirical study of option prices under skewness and kurtosis conditions using reduced variance techniques from two options of automotive stock (NIO) and technology stock (INTC), where we want to investigate the performance of those methods in the estimation of the stock call option price model in those two stocks. From the analysis, we found that those techniques can reduce the option price variance and give a more accurate price, where the Variance Gamma models produce the smallest MAPE compared to the other models used.