Minimum Variance and Optimal Asymptotic Portfolios

By taking the goal of optimal portfolio management as long term maximization of wealth, whether it be via long or short term gains, portfolio selection can be based upon objective criteria instead of risk preference. In this paper we wish to relate these optimum long run or asymptotic portfolios to the set of efficient portfolios. The case of statistically identical securities is presented in detail with attention given to the effect of mutual correlation on the optimal size of the portfolio. Circumstances are discussed for which such popular policies as short term trend following and intermediate term dollar averaging are optimal in the long run. Since portfolio management can be interpreted as a problem within the framework of multistage decision process of stochastic type, further insight into optimal policies could be developed from examination of dynamic programming theory.

• 저널Management Science · 16(3) · 221–226
• 토픽Advanced Bandit Algorithms Research · Management Science and Operations Research
• DOI10.1287/mnsc.16.3.221
• 저자Samuel Bickel