The goal of this thesis is to develop an accurate model for implied volatility dynamics that can consistently beat the Black-Scholes-Merton model in delta hedging standard European options. First, we propose an improvement of the volatility-price sensitivity estimation methodology. The volatility-price sensitivity measures the relationship between the volatility surface and the underlying price process. Accounting for negative correlation between implied volatility and the underlying price we adjust the BSM delta by an amount that depends on the BSM vega and the implied volatility-price sensitivity. We extended this adjustment according to the idea of Derman who stated the regime dependent behavior of volatility. Thus, we condition our hedging parameter on the current market regime applying the Markov switching model for the estimation of the volatility-price sensitivity. We model the implied volatility sensitivity to the underlying price movements in two stages: first by applying principal component analysis to implied volatilities of the same maturity but with varying moneyness, and then by estimating a Markov-Switching model of the relationship between at-the-money volatility and the underlying price. The empirical evidence suggests, that using regime switching smile adjusted deltas consistently improves the performance of the dynamic delta-hedging, relative to all the other smile-consistent hedge ratios considered.     «

The goal of this thesis is to develop an accurate model for implied volatility dynamics that can consistently beat the Black-Scholes-Merton model in delta hedging standard European options. First, we propose an improvement of the volatility-price sensitivity estimation methodology. The volatility-price sensitivity measures the relationship between the volatility surface and the underlying price process. Accounting for negative correlation between implied volatility and the underlying price we ad...     »