Abstract — We study a discrete-time version of Markowitz’s mean-variance portfolio selection problem where the market parameters depend on the market mode (regime) that jumps among a finite number of states. The random regime switching is delineated by a finite-state Markov chain, based on which a discrete-time Markov modulated portfolio selection model is presented. Such models either arise from multiperiod portfolio selections or result from numerical solution of continuous-time problems. The natural connections between discrete-time models and their continuous-time counterpart are revealed. Since the Markov chain frequently has a large state space, to reduce the complexity, an aggregated process with smaller state space is introduced and the underlying portfolio selection is formulated as a two-time-scale problem. We prove that the process of interest yields a switching diffusion limit using weak convergence methods. Next, based on the optimal control of the limit process obtained from our recent work, we devise portfolio selection strategies for the original problem and demonstrate their asymptotic optimality. Index Terms — Markowitz’s mean-variance portfolio selection, discrete-time model, Markov chain, switching diffusion, linearquadratic problem, singular perturbation. I.

We develop a numerical scheme for determining the optimal asset allocation strategy for time-consistent, continuous time, mean variance optimization. Any type of constraint can be applied to the investment policy. The optimal policies for time-consistent and pre-commitment strategies are compared. When realistic constraints are applied, the efficient frontiers for the precommitment and time-consistent strategies are similar, but the optimal investment strategies are quite different.

We determine the optimal dynamic investment policy for a mean quadratic variation objective function by numerical solution of a nonlinear Hamilton-Jacobi-Bellman (HJB) partial differential equation (PDE). We compare the efficient frontiers and optimal investment policies for three mean variance like strategies: pre-commitment mean variance, time-consistent mean variance, and mean quadratic variation, assuming realistic investment constraints (e.g. no bankruptcy, finite shorting, borrowing). When the investment policy is constrained, the efficient frontiers for all three objective functions are similar, but the optimal policies are quite different.

We propose the use of a mean–quadratic-variation criteria to determine an optimal trading strategy in the presence of price impact. We derive the Hamilton Jacobi Bellman (HJB) Partial Differential Equation (PDE) for the optimal strategy, assuming the underlying asset follows Geometric Brownian Motion (GBM). We also derive the HJB PDE assuming that the trading horizon is small and that the underlying process can be approximated by Arithmetic Brownian Motion (ABM). The exact solution of the ABM formulation is in fact identical to the price-independent approximate optimal control for the mean-variance objective function in [2]. The GBM mean–quadratic-variation optimal trading strategy is in general a function of the asset price. However, for short term trading horizons, the control determined under the ABM assumption is an excellent approximation.

Abstract not found

Abstract not found

Abstract. The Modern Portfolio Theory of Markowitz maximized portfolio expected return subject to holding total portfolio variance below a selected level. Digital Portfolio Theory is an extension of Modern Portfolio Theory, with the added dimension of memory. Digital Portfolio Theory decomposes the portfolio variance into independent components using the signal processing decomposition of variance. The risk or variance of each security’s return process is represented by multiple periodic components. These periodic variance components are further decomposed into systematic and unsystematic parts relative to a reference index. The Digital Portfolio Theory model maximizes portfolio expected return subject to a set of linear constraints that control systematic, unsystematic, calendar and non-calendar variance. The paper formulates a single period, digital signal processing, portfolio selection model using cross-covariance constraints to describe covariance and autocorrelation characteristics. Expected calendar effects can be optimally arbitraged by controlling the memory or autocorrelation characteristics of the efficient portfolios. The Digital Portfolio Theory optimization model is compared to the Modern Portfolio Theory model and is used to find efficient portfolios with zero calendar risk for selected periods. Key words: portfolio optimization, portfolio theory, digital signal processing, calendar anomalies 1.

This paper studies a continuous-time market where an agent, having specified an investment horizon and a targeted terminal mean return, seeks to minimize the variance of the return. The optimal portfolio of such a problem is called mean-variance efficient à la Markowitz. It is shown that, when the market coefficients are deterministic functions of time, a mean-variance efficient portfolio realizes the (discounted) targeted return on or before the terminal date with a probability greater than 0.8072. This number is universal irrespective of the market parameters, the targeted return and the length of the investment horizon. 1. Introduction. In his seminal work, Markowitz [9] proposed the meanvariance portfolio selection model for a single investment period, where an agent seeks to minimize the risk of his investment, measured by the variance of his return, subject to a given mean return. (In Markowitz’s original setup, the model is formulated as a multi-objective optimization problem,