Effects of Tax on Investment Portfolios and Financial Markets Under Mixed Integer Stochastic Programming

This paper investigates the micro and macro effects of income tax on large scale portfolio optimization. Stochastic integer programming is used to optimize post-tax large-scale portfolios when the global market is segmented by regional tax rules. A broad range of realistic trading rules and inequality constraints as well as a large number of assets are considered. The increased complexity and scale of the problem renders theoretical methods infeasible. A new numerical approach based on basic Greedy heuristics, in which integer and non-linear restrictions are considered simultaneously, is proposed. The superiority of this approach is demonstrated through a comparison with extant BONMIN’s Branch-and-Bound (B&B) methods, for problems with up to 288 assets whereas previous efforts under mixed-integer non-linear programming (MINLP) were limited to 200 assets. The approach is used to test and extend extant theoretical work on post-tax portfolio management. The same generic conclusion that taxation affects the portfolio composition dramatically, is reached, but the new approach reveals greater detail of the implications of the combination of various factors. A study of price effects finds that market equilibrium prices are affected by relative as well as absolute tax rates across assets and global regions. The implication is that the market's response to tax rate changes across assets and countries can now, at least partially, be predictable prior to implementation by government. Investors are also better able to forecast how the markets will react in the short term.