Dynamic Stochastic Accumulation Model with Application to Portfolio Risk Management
Darina Graczová
PhD thesis advisor: Daniel Ševčovič

PhD thesis - Full text   


Summary: Iearch we focus on the modeling of portfolio returns with fat-tailed distributions that have the property to exhibit extreme large skewness and kurtosis. We focus on the normal-inverse Gaussian distribution and analyze the impact of higher moments on the optimal choice of the portfolio composition. We consider a dynamic stochastic model of Bellman type and discuss the problem of optimal choice of portfolio composition with different level of risk dependent on proportion of risky - to - non-risky assets, especially in application to pension management. We propose a numerical scheme for calculation and perform a sensitivity analysis of the descriptive statistics of asset returns on the accumulated sum at the final time as well as in each time step during the saving. We compare the results considering the normal distribution and NIG distribution with different skewness and kurtosis and discuss the distribution of accumulated sum at the final time.
Related papers

[1] D. Graczova and P. Jacko: Generalized Restless Bandits and Knapsack Problem for Perishable Inventories, Operations Research, Volume 62 Issue 3, 2014, pp. 696-711