Portfolio Construction Approaches Research Articles

Practical Improvements to Mean-Variance Optimization for Multi-Asset Class Portfolios

In the more than 70 years since Markowitz introduced mean-variance optimization for portfolio construction, academics and practitioners have documented numerous weaknesses in the approach. In this paper, we propose two easily understandable improvements to mean-variance optimization in the context of multi-asset class portfolios, each of which provides less extreme and more stable portfolio weights. The first method sacrifices a small amount of expected optimality for reduced weight concentration, while the second method randomly resamples the available assets. Additionally, we develop a process for testing the performance of portfolio construction approaches on simulated data assuming variable degrees of forecasting skill. Finally, we show that the improved methods achieve better out-of-sample risk-adjusted returns than standard mean-variance optimization for realistic investor skill levels.

Left-tail risk and UK stock return predictability: Underreaction, overreaction, and arbitrage difficulties

Recent studies challenge the standard model risk-return trade-off by showing inverse predictive power of firm-specific left-tail risk for future returns (i.e., left-tail momentum). In this work, we investigate the pricing of left-tail risk in UK stocks. Both the portfolio construction approach and Fama-MacBeth regressions reveal the underperformance of stocks with high left-tail risk. We examine alternative channels behind this pricing anomaly, namely, investor underreaction behaviour, continuous overreaction behaviour, and limits to arbitrage. Our findings suggest that the observed underperformance associated with high left-tail risk is largely a manifestation of investor underreaction to bad performance. However, the results also show that the predictable underperformance of high left-tail risk stocks is manifest in past winners. The empirical investigation reveals that, in addition to underreaction, limits to arbitrage interacts with investor high attention levels to explain part of the anomaly. The empirical findings provided here suggest several important implications for practitioners in the equity market.

Stock price crash prediction based on multimodal data machine learning models

This study introduces multimodal data machine learning framework to predict stock crashes. It encapsulates market data, graph data cultivated from industry affiliations through node2vec, and text data derived from sentiment analysis. The LightGBM is utilized, marking an improvement by 7.13% over preceding studies, achieving 75.85% balanced accuracy. An innovative long-short portfolio construction approach is articulated, demonstrating the practical significance of the predictions, with a 4.75% portfolio return in 2022 — a 27.26% advancement over the CSI 300. This endeavour in leveraging multimodal data machine learning for stock crash prediction offers a promising performance, serving as a valuable reference for investors.

TVP-VAR based time and frequency domain food & energy commodities connectedness an analysis for financial/geopolitical turmoil episodes

Amidst the current global inflationary challenges, the concurrent rise of energy and agricultural commodity prices, which constitute the primary components of consumer prices, has emerged as a matter of significant interest among both scholars and policymakers. To this end, this study examines the dynamic interlinkages between food and energy commodity indexes from 2005:1 to 2023:3, covering turmoil episodes such as the Global Financial Crisis (GFC), the COVID-19 pandemic, and the Russia-Ukraine Conflict (RUC). Additionally, following Broadstock et al. (2022), we perform dynamic portfolio analyses to determine portfolio performances under 3 different portfolio construction approaches. The empirical results presented in this paper allow for a number of important findings. First, both the time and frequency-domain connectedness indexes associate with major financial/geopolitical stress events. Second, the fuel energy and the crude oil price indexes are the largest propagators and recipients of spillovers Third, the cumulative portfolio returns exhibit significant growth during the early phase of the COVID-19, declining during the RUC, and a notable upswing during the GFC. Finally, our findings for frequency-dependent connectedness networks indicate that the market is particularly susceptible to short-term shocks. This paper has significant ramifications for investors, market players, and policymakers.

A hybrid approach for portfolio construction: Combing two‐stage ensemble forecasting model with portfolio optimization

AbstractCombining the stock prediction with portfolio optimization can improve the performance of the portfolio construction. In this article, we propose a novel portfolio construction approach by utilizing a two‐stage ensemble model to forecast stock prices and combining the forecasting results with the portfolio optimization. To be specific, there are two phases in the approach: stock prediction and portfolio optimization. The stock prediction has two stages. In the first stage, three neural networks, that is, multilayer perceptron (MLP), gated recurrent unit (GRU), and long short‐term memory (LSTM) are used to integrate the forecasting results of four individual models, that is, LSTM, GRU, deep multilayer perceptron (DMLP), and random forest (RF). In the second stage, the time‐varying weight ordinary least square model (OLS) is utilized to combine the first‐stage forecasting results to obtain the ultimate forecasting results, and then the stocks having a better potential return on investment are chosen. In the portfolio optimization, a diversified mean‐variance with forecasting model named DMVF is proposed, in which an average predictive error term is considered to obtain excess returns, and a 2‐norm cost function is introduced to diversify the portfolio. Using the historical data from the Shanghai stock exchange as the study sample, the results of the experiments indicate the DMVF model with two‐stage ensemble prediction outperforms benchmarks in terms of return and return‐risk characteristics.

Bitcoin volatility and the introduction of bitcoin futures: A portfolio construction approach

We evaluate the introduction of Bitcoin futures on Bitcoin return and volatility using realized volatility and GARCH model pre-and post-futures. We also assess the portfolio construction implications by building two portfolios containing the top 25 S&P stocks, one without futures and one with. GARCH and realized volatility show mixed results. We provide that futures make Bitcoin riskier and more vulnerable to fluctuations over time. However, Bitcoin futures improve the portfolio's volatility and returns profile. Our findings offer implications regarding portfolio strategies implemented by risk-averse and risk-seeking investors and managers as we show how Bitcoin futures can hedge investments.

Construction of Optimal Portfolio Using Efficient Portfolio and Zero Opportunity Cost

In this manuscript, we introduce asset allocation and portfolio selection techniques based on efficiency condition, Sharpe ratio error condition and order three zero opportunity condition. Investors expect the same level of risk and return from alternative investment options unless they want the advantage of diversification for risk. There are two fundamental investment portfolios. The first one is risk free fundamental portfolio, and the second one is risky fundamental portfolio. Investors use zero opportunity cost to select portfolio objective. In this research, mathematical derivation of portfolio construction approach is described in advance. Historical data of this research show that there is a positive linear relationship between natural logarithm of standard deviation of securities’ return and square of co-variance between securities and market return. Furthermore, it shows that there is a positive linear relationship between Treynor ratio of securities’ return and Sharpe ratio of securities’ return. We use global search optimization tool in MatLab and R software to solve empirical portfolio selection and asset allocation problem. Moreover, we apply direct and indirect mathematical proof methods to prove mathematical facts of this study.

Mean–variance portfolio optimization using machine learning-based stock price prediction

The success of portfolio construction depends primarily on the future performance of stock markets. Recent developments in machine learning have brought significant opportunities to incorporate prediction theory into portfolio selection. However, many studies show that a single prediction model is insufficient to achieve very accurate predictions and affluent returns. In this paper, a novel portfolio construction approach is developed using a hybrid model based on machine learning for stock prediction and mean–variance (MV) model for portfolio selection. Specifically, two stages are involved in this model: stock prediction and portfolio selection. In the first stage, a hybrid model combining eXtreme Gradient Boosting (XGBoost) with an improved firefly algorithm (IFA) is proposed to predict stock prices for the next period. The IFA is developed to optimize the hyperparameters of the XGBoost. In the second stage, stocks with higher potential returns are selected, and the MV model is employed for portfolio selection. Using the Shanghai Stock Exchange as the study sample, the obtained results demonstrate that the proposed method is superior to traditional ways (without stock prediction) and benchmarks in terms of returns and risks.

Adaptive Optimal Risk Budgeting

In this article, the authors suggest Bayesian-style adaptive enhancement to a popular equal risk contribution (ERC) portfolio construction technique they call adaptive optimal risk budgeting (AORB). The enhancement has the potential to bring portfolios closer to mean–variance efficiency when Sharpe ratios and correlations of assets vary while retaining some of the ERC’s robustness to estimation errors. The authors test AORB’s viability by putting it in competition with ERC itself and with a version of the Bayesian shrinkage mean–variance technique in a carefully simulated setting. They find that the new method appears to deliver measurable advantages over its competition in a broad range of realistic settings. Multiple possible applications to portfolios of risk premia strategies and a multi-asset universe more generally are discussed by the authors. TOPICS:Portfolio construction, analysis of individual factors/risk premia Key Findings • We suggested a computationally simple yet powerful portfolio construction approach that is formulated in terms of risk contributions but is designed to prescribe an approximately mean-variance efficient solution even when Sharpe ratios and correlations between assets vary in magnitude and over time. We called the new method Adaptive Optimal Risk Budgeting (AORB). • When tested against popular Equal Risk Contributions and “classical” Bayesian shrinkage methods, AORB was competitive in all cases and emerged as a clear winner when Sharpe ratios and correlations between assets were substantially differentiated. • AORB can also be trivially expanded to incorporate tail risk considerations.

Portfolio Construction Techniques Applied to Traditional Multi Asset Portfolios

In this paper we describe the history of different portfolio construction approaches from a simple Risk Parity approach and its extension Equal Risk Contribution over Markovitz mean-variance to conditional Value-at-Risk and others like minimum Value-at-Risk or Average Drawdown. We implement the portfolio construction methodologies in Python and estimate a set of risk and performance metrics. We apply the approaches on a traditional multi asset portfolio and empirically show that a particular approach dominates the others in most metrics.

Practical Applications of Portfolio Optimization Strategy for Concentrated Portfolios: Models and Time Horizons

<h3>Practical Applications Summary</h3> In <b>Portfolio Optimization Strategy for Concentrated Portfolios: <i>Models and Time Horizons</i></b>, from the Fall 2018 issue of <b><i>The Journal of Wealth Management</i></b>, authors <b>Sarah Campbell</b>, <b>James Chong</b>, <b>William Jennings</b>, and <b>G. Michael Phillips</b> (all of <b>MacroRisk Analytics</b>) examine the relative merits of portfolio-construction approaches for concentrated, or high-conviction, (i.e., 15 or fewer holdings) portfolios. The authors’ study in part stems from the debate over the efficacy of an optimization approach to portfolio construction compared with an equally weighted (or “naïve”) approach. The authors’ intention is to help guide wealth managers in choosing a portfolio-optimization method best suited to their investment priorities (e.g., maximizing returns, minimizing risk, balancing risk and returns). The authors apply a Monte Carlo simulation to generate 18,000 buylists of stocks. From each buylist, the simulation derives an optimized portfolio and two equally weighted portfolios: one of stocks selected through optimization (a “hybrid” portfolio) and the other of randomly selected stocks (a “benchmark” portfolio). The study calculates each portfolio’s forward average yearly returns, annual standard deviations, and forward risk–return ratios for several different time frames, and then compares relative performance. The authors emphasize the importance of choosing the suitable portfolio-construction approach: one aligned with the number of holdings in a portfolio and investment horizon or one that best performs under the distinct circumstances faced by the wealth manager. <b>TOPICS:</b>Portfolio construction, analysis of individual factors/risk premia, tail risks, performance measurement

POST-MERGER RETURNS IN FRONTIER MARKETS, OR HOW WE LEARNED TO STOP WORRYING AND LOVE THE ACQUIRERS

This study presents the results from a comprehensive out-of-sample test of long-run returns following mergers and acquisitions (M&amp;amp;amp;amp;As). Using a unique sample from 23 frontier markets of almost 800 transactions conducted during the years 1992 to 2016, we implement both cross-sectional tests and time-series examinations based on a calendar-time portfolio approach. Contrary to evidence from developed markets, the M&amp;amp;amp;amp;As in these frontier markets do not lead to abnormal underperformance of acquirers, regardless of whether they paid for the acquisition with cash or stock. The results are robust to many considerations, including subsample and subperiod analysis, alternative formation periods, different portfolio construction approaches.

Robust equity portfolio performance

The earliest documented analytical approach to portfolio selection is Markowitz’s mean–variance analysis, which attempts to find the portfolio with optimal performance by considering the tradeoff between return and risk. The performance of mean–variance analysis has been the subject of many studies and compared to other portfolio construction approaches such as a naive equally-weighted allocation scheme. In recent years, several approaches have been proposed to improve the mean–variance model by reducing the sensitivity of the portfolio selection process in order achieve robust performance. Although robust portfolio optimization has been one of the most researched methods for improving portfolio robustness, the performance of robust portfolios has not been the major focus of studies. In this paper, a comprehensive analysis on robust portfolio performance is presented for equity portfolios constructed in the U.S. market during the period 1980 and 2014, and results confirm the advantage of robust portfolio optimization for controlling uncertainty while efficiently allocating investments.

Diversify and Purify Factor Premiums in Equity Markets

In this paper we consider the question of how to improve the efficacy of strategies designed to capture factor premiums in equity markets and, in particular, from the value, quality, low risk and momentum factors. We consider a number of portfolio construction approaches designed to capture factor premiums with the appropriate levels of risk controls aiming at increasing information ratios. We show that information ratios can be increased by targeting constant volatility over time, hedging market beta and hedging exposures to the size factor, i.e. neutralizing biases in the market capitalization of stocks used in factor strategies. With regards to the neutralization of sector exposures, we find this to be of importance in particular for the value and low risk factors. Finally, we look at the added value of shorting stocks in factor strategies. We find that with few exceptions the contributions to performance from the short leg are inferior to those from the long leg. Thus, long-only strategies can be efficient alternatives to capture these factor premiums. Finally, we find that factor premiums tend to have fatter tails than what could be expected from a Gaussian distribution of returns, but that skewness is not significantly negative in most cases.

An Econometric Approach to Incorporating Non-Normality in VaR Measurement

Following the recent financial crises, there has been a proliferation of new risk management and portfolio construction approaches. These approaches all endeavour to better quantify and manage risk by accounting for the stylised facts of financial time series mainly heavy and skewed tails, volatility clustering and converging correlations. Capturing all these stylised facts in a coherent framework has proved to be an elusive and knotty task. We here propose a pure econometric framework that captures all the stylised facts satisfactorily. We use three data sets to show how the approach is implemented in VaR forecasting and correlation analysis. We show how an investment portfolio can be constructed in order to optimise reserve capital holding. The approach employed is linear programming (LP) computable, satisfies second degree stochastic dominance and outperforms the general mean/VaR quadratic optimisation to arrive at efficient asset allocation.

Practical Applications of Risk Parity, Maximum Diversification, and Minimum Variance: An Analytic Perspective

This study provides portfolio managers with a number of insights into why particular weights of assets exist in the different risk-based portfolios. It also investigates the optimal relationships between low- and high-volatility stocks for each portfolio. Risk Parity, Maximum Diversification, and Minimum Variance: An Analytic Perspective, in the Spring 2013 issue of The Journal of Portfolio Management, presents the first analytic study of a long-only, large-set risk-parity portfolio. Co-authors Roger Clarke, Chairman of Analytic Investors, Harindra de Silva, President of Analytic Investors, and Steven Thorley, H. Taylor Peery Professor of Finance at Brigham Young University Marriott School of Management compare the findings with the previous research on long-only maximum-diversification and minimum-variance portfolios. In this Practical Applications report, co-authors Roger Clarke explains how their approach yield new insights for investors. Clarke points out that that although institutional investors and their advisors are considering alternative ways of approaching risk-based portfolios, their views may be constrained by empirical or historical limitations. Using an analytic approach rather than an empirical one, “we wanted to understand the core characteristics and tilts of three of the most popular risk-based portfolio construction approaches,” says Clarke.

A Simple Diversified Portfolio Strategy

We present a simple portfolio construction approach which is a blend of market weights and equal stock and sector weights. Our approach results in a highly diversified portfolio both on a stock level and on a sector level and generates higher portfolio returns at slightly lower risk than a market weighted index. We demonstrate that the higher returns of our diversified portfolio originate both from mitigating the link with market weights and from its higher return benefit due to diversification which we are able to capture because we rebalance our portfolio on a regular basis. Our diversified portfolio exhibits only slightly higher turnover than a market weighted index and is less concentrated in mega-cap stocks. Instead it assigns somewhat higher weight to smaller stocks and sectors than a market weighted index. At the same time the diversified portfolio retains the characteristics of a broad market index and is therefore highly implementable and has very high investment capacity.Our simple diversification strategy appears to be a superior way of capturing the equity risk premium compared to a market weighted portfolio. It therefore establishes a tougher benchmark for active fund managers than market weighted indexing as it results in well-diversified, high-capacity and low turnover portfolios that can be delivered very cheaply.

Risk-Based Asset Allocation: A New Answer to anOld Question?

In recent years, we have witnessed an alarmingly large and growing amount of literature on portfolio construction approaches focused on risks and diversification rather than on estimating expected returns. Numerous simulations applied to different universes have been documented in support of these approaches based on their apparent outperformance versus passive market capitalization–weighted or static fixed-weight portfolios. Many studies attribute the better performance of these risk-based asset allocation approaches to superior diversification. Given the absence of clearly defined investment objective functions behind these approaches as well as the metrics used by these studies to evaluate ex post performance, Lee puts these approaches into the same context of mean-variance efficiency in an attempt to understand their theoretical underpinnings. In doing so, he hopes to shed some light on what these approaches attempt to achieve and on the characteristics of the investment universe, if indeed these approaches are meant to approximate mean-variance efficiency. Rather than adding to the already large collection of simulation results, Lee uses some simple examples to compare and contrast the portfolio and risk characteristics of these approaches. He also reiterates that any portfolio which deviates from the market capitalization–weighted portfolio is an active portfolio. He concludes that there is no theory to predict, ex ante, that any of these risk-based approaches should outperform.

Drawdown-at-Risk Monte Carlo Optimization

We present a portfolio construction approach with two interesting non-standard features: First, the risk measure used is “drawdown-at-risk”, an interesting concept combining attractive features of drawdown and value-at-risk measures. Second, the efficient frontier is calculated from “random portfolios”, i.e. portfolios containing random constituent weights. We call this “Monte Carlo optimization”. Both features would deserve a detailed analysis. The goal of this note is to provide an overview and illustrate the potential of the approach with two examples: Asset allocation in a small universe consisting of three assets (Indian stocks and bonds as well as gold in INR) and drawdown optimization on single-stock level across the full S&P 500 universe.

Portfolio Construction with Qualitative Forecasts

The vast majority of active portfolio managers use a fundamental investment approach. They do not generate quantitative forecasts but instead express their market views in a qualitative manner. The portfolio construction approach in practice is usually purely ad hoc: overweighting assets with a bullish outlook by some prespecified amount, and underweighting assets for which the view is bearish. A structured approach for portfolio construction leads to a more consistent implementation of market views and to more balanced portfolios in terms of risk profile. This approach incorporates several diagnostic tools and a Bayesian model. It also allows a portfolio manager to compute the shrinkage in the information ratio when implementing a suboptimal portfolio. This mitigates the need for transactions and hence reduces transaction costs.