| Summary: | 碩士 === 國立高雄第一科技大學 === 風險管理與保險所 === 95 === ABSTRACT
This thesis studies the stochastic dynamic programming based on Duffie, Pan, and Singleton model (2000) to construct optimal problem for investors and include event risk in the price and volatility jump-diffusion model. However, we can’t obtain the analytic solution for the complicated optimal problem and thus we provide modified Euler to acquire the numerical solutions of the weights of risky assets. Then, we use the solutions to explore the return and risk of the dynamic investment programming under different scenarios. Empirically, we found that investors tend to decrease their risky asset allocation for either negative or positive jump size enlargement and the drop of risky asset amount for negative jumps is more significant, i.e. the asymmetric effect of jumps. Both price and volatility jumps will affect the risky asset allocation of the less risk averse investors and volatility jumps otherwise would make the more risk-averse investors add somewhat their risky asset amount. Finally, withγ=2 and 5 and planning horizon = 3.5 and 6.25 years, the performances of dynamic investment programming measured by Sharpe ratio are superior to buy-and- hold substantially.
|
|---|
