Portfolio optimisation: bridging the gap between theory and practice

Abstract

Portfolio optimisation is widely acknowledged for its significance in investment decision-making. Yet, existing methodologies face several limitations, among them converting optimal theoretical portfolios into real investment is not always straightforward. Several classes of exogenous (real-world) constraints have been proposed in literature with the intent of reducing the gap between theory and practice, which have worked to an extent. In this paper, we propose an optimisation-based framework which attempts to further reduce this gap. We have the explicit intention of producing portfolios that can be immediately converted into financial holdings. Our proposed framework is generic in the sense that it can be used in conjunction with any portfolio selection model, and consists of splitting the portfolio selection problem into two-stages. The main motivation behind this approach is in enabling automated investing with minimal human intervention, and thus the framework was built in such a way that real-world market features can be incorporated with relative ease. Among the novel contributions of this paper, this is the first work, to the best of our knowledge, to combine futures contracts and equities in a single framework, and also the first to consider borrowing costs in short positions. We present extensive computational results to illustrate the applicability of our approach and to evaluate its overall quality. Among these experiments, we observed that alternatives from literature are susceptible to numerical errors, whereas our approach effectively mitigates this issue.