Empirical Performance Of Stochastic Volatility Models On FX Options Regarding RMB

This thesis focuses on Foreign Exchange (FX) options regarding the Chinesecurrency: the Renminbi (RMB). More precisely, it focuses on the thriving marketof the offshore RMB, also known as CNH. While the access of the onshore RMBis limited for foreign investors and heavily regulated, the CNH market, created in2010, experienced an exponential growth over the past few years, benefiting froma high demand from both risk managers and traders.In this thesis, we perform empirical studies regarding the pricing and the hedg-ing of CNH options and we measure the impact of the pricing model on the cost ofdifferent trading and hedging strategies. In that extend, we test the empirical pric-ing and hedging effectiveness of three popular option pricing models, the Black-Scholes model, the Heston model and the Bates model. To estimate the modelparameters, we use daily quotes available on Bloomberg, reflecting the prices ofvanilla options traded over-the-counter. After ranking these models according totheir performances, we perform an out-of-sample test to see which model marksthe best the volatility structure of vanilla options. Meanwhile, we test the hedgingeffectiveness of these models on one of the most traded class of exotic options:the barrier options. Last, we take it one step further, using our empirical findingsto develop some trading perspectives and highlighting and measuring the benefitsarising from a better pricing model.Through the comparison of the pricing and hedging performances of thesemodels, it turns out that both stochastic volatility models outperform the Black-Scholes model for the pricing of vanilla options and bring more accuracy on thepricing of exotic options. More precisely, the Heston model has the best staticperformances but it has to share the credits with the Bates model when it comes todynamic performances: overall, it overperforms the latter for short-term options,but underperforms it for long-term options. From these results, we find that a buyerof vanilla options will significantly reduce the probability of overpricing an optionand that a seller of barrier options will substantially reduce the cost of his hedgingstrategy by using the Heston model instead of the Black-Scholes model.