Assessing the Economic Value of Implied Volatility Estimates

Sammanfattning: This thesis studies the value of implied volatility estimates for portfolio allocation under the modern portfolio theory (MPT) framework introduced by Markowitz and compares the pricing performances of several common option pricing models. The thesis consists of two parts. The first part compares the NAGARCH framework with the bidirectional Markov switching models (B-MSM) of (Duan et al., 2002) and develops an efficient pricing algorithm for European options under a finite-state model. The variance-implied portfolios are then evaluated on the S&P 500, giving further evidence for using implied volatility estimates for asset allocation. The second part compliments the first by analyzing the value of implied volatility estimates using practical alternatives, namely the Black-Scholes (BS) model, the ad-hoc Black-Scholes (ABS) model and the VIX. The ABS model shows very promising results, outperforming standard BS in form of option pricing and under MPT can create value via volatility timing, beating a buy-and-hold strategy on the S&P 500 in some cases and showing improvements over a VIX induced strategy. However, the simple, practical approaches fall short when compared to the models in part 1, both in form of option pricing as well as MPT asset allocation, suggesting that there is a considerable payoff in estimating implied volatility under these more complex frameworks.