Post-tax optimization with stochastic programming

In this paper, we consider a stochastic programming approach to multistage post-tax portfolio optimization. Asset performance information is specified as a scenario tree generated by two alternative methods based on simulation and optimization. We assume three tax wrappers involving the same instruments for an efficient investment strategy and determine optimal allocations to different instruments and wrappers. The tax rules are integrated with the linear and mixed integer stochastic models to yield an overall tax and return-efficient multistage portfolio. The computation performance of these models is tested using a case study with different scenario trees. Our experiments show the optimal portfolios obtained by both linear programming and mixed integer stochastic models diversify over wrappers and the original capital is distributed among assets within each wrapper.