Cumulative Returns Research Articles

Leveraging LSTM Predictions for Enhanced Portfolio Allocation with Markowitz Mean-Variance Optimization

This research investigates the application of Long Short-Term Memory (LSTM) networks for predicting expected returns and integrating these predictions into the Markowitz Mean-Variance Optimization (MVO) framework. The study utilized historical data from eight Indonesian stocks: BBCA, BBRI, TLKM, EXCL, UNVR, ICBP, ASII, and SMGR. The dataset covered the period from 2018 to 2024. The LSTM model was employed to predict cumulative returns over a 90-day horizon, and its performance was compared to the Exponentially Weighted Moving Average (EWMA) method. The findings indicate that LSTM achieved lower Root Mean Squared Error (RMSE) than EWMA for four stocks (BBCA, BBRI, UNVR, ICBP), while EWMA demonstrated better performance for the remaining four stocks. MVO results revealed that LSTM-based predictions achieved an average return of 4.285%, surpassing EWMA's 1.856% but falling short of the 12.298% obtained using actual returns. These results highlight the potential of LSTM models to enhance portfolio allocation strategies.

A hybrid stock prediction method based on periodic/non-periodic features analyses

Stock investment is an economic activity characterized by high risks and high returns. Therefore, the prediction of stock prices or fluctuations is of great importance to investors. Stock price prediction is a challenging task due to the nonlinearity and high volatility of stock time series. Existing deep learning models may not capture the periodic and non-periodic features of stock data effectively. In this paper, we propose a novel model that leverages Complete Ensemble Empirical Mode Decomposition (CEEMD), Time2Vec, and Transformer to better capture and utilize various patterns in stock data for enhanced prediction performance, and we call it ETT. CEEMD decomposes the stock data into different frequency components based on their intrinsic scales. Time2Vec provides a time vector representation that captures both periodic and non-periodic patterns while being invariant to time scaling. Transformer learns the long-term dependencies and global information from the data. We apply ETT to predict stock prices in the Chinese A-share market and compare it with several baseline models. The results show that ETT reduces the mean squared error (MSE) by an average of 4% and increases the average cumulative return by 58% on the CSI 100 and Hushen 300 datasets.

A Quantitative Strategy for value investing in Small Cap Stocks Based on Positive Feedback Asymmetry

Significant positive feedback trading exists in the Chinese market, and there is a significant difference between the extent of chasing up and killing down in the context of positive feedback trading. The purpose of this paper is to investigate whether a quantitative investment strategy with this asymmetric trait as the core factor can achieve significant excess returns and reduce the risk impact of irrational investors in the market. This paper is based on the CSI 500 index represented by small-cap stocks as the strategic stock pool, and further constructs a quantitative stock selection strategy, which firstly estimates the positive feedback trading asymmetry index of individual stocks in the market through the framework of SW, and at the same time, combines the Chinese three-factor model with Amihud's illiquidity index to construct the factor portfolios, and conducts a validity test to ultimately select the four portfolios with the highest cumulative returns as factor portfolios. This paper concludes that the strategy has a return of 316.73% during the back testing period, a weekly Sharpe ratio of 0.33, and a maximum backtest of 36.41%, realizing a significant excess return with controlled risk.

Optimal portfolio selection of China's green bond and stock markets: Evidence from the multi-frequency extreme risk connectedness

This paper investigates the optimal portfolio of industry stocks with China's green bond (CBGB) at different time scales, through comparing minimum connectedness portfolio (MCoP1) and modified minimum connectedness portfolio (MCoP2) under the condition of minimizing extreme risk connectedness, as well as minimum VaR portfolio. First, we examine the ability of CBGB to provide diversification benefits for industry stocks at different scales, which employs extreme risk connectedness. Then, the portfolio performing optimally at each scale is explored by comparing specific performances of various portfolio methods. The empirical results show that: (1) CBGB is weakly related to industry stocks, whether in the time or frequency domain, and is the net receiver. (2) Most industry stocks play different roles (transmitter or receiver) in different scales, with time-varying characteristics. Therefore, industry stocks benefit highly from the diversification effects of CBGB, and their portfolio should possess frequency and time-varying characteristics. (3) Based on the comparison results of weight allocations, cumulative returns, and Sharpe ratios, the short-term optimal performance is achieved by MCoP2, while in medium- to long-term, MCoP1 outperforms MCoP2 and minimum VaR portfolio. This implies that considering extreme risk connectedness in the construction of portfolios can effectively enhance returns, but in the long term, market integration should also be taken into account.

Adaptive peak price with lazy updates for short-term portfolio optimization

This paper introduces the novel Adaptive Peak Price with Lazy Updates (APPLU) approach for short-term portfolio optimization (SPO), a method that innovatively combines the Radial Basis Function (RBF) and a new lazy update approach to address the unique challenges of SPO. Our approach is tailored to balance the dual objectives of maximizing returns and minimizing transaction costs, which are critical concerns in dynamically allocating wealth among various assets over time. Unlike conventional methods that primarily rely directly on peak price information, APPLU introduces an adaptive peak price using RBFs to capture continuous depreciation information of assets, thereby alleviating the aggressive nature of traditional peak price strategies and avoiding frequent trades. Furthermore, we propose a new lazy update approach that employs unsquared l2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$l^{2}$$\\end{document}-norm regularization to represent portfolio changes. This approach contrasts with squared l2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$l^{2}$$\\end{document}-norm regularization, which disproportionately penalizes larger portfolio changes while being more lenient towards smaller changes. Thereby, our methodology offers a more balanced and effective approach to portfolio adjustment. Extensive experiments conducted on seven real datasets demonstrate that APPLU outperforms existing strategies in terms of cumulative return and risk-adjusted return, while effectively controlling transaction costs and maintaining a moderate wealth accumulation strategy.

Enhancing Portfolio Optimization: A Two-Stage Approach with Deep Learning and Portfolio Optimization

The portfolio selection problem has been a central focus in financial research. A complete portfolio selection process includes two stages: stock pre-selection and portfolio optimization. However, most existing studies focus on portfolio optimization, often overlooking stock pre-selection. To address this problem, this paper presents a novel two-stage approach that integrates deep learning with portfolio optimization. In the first stage, we develop a stock trend prediction model for stock pre-selection called the AGC-CNN model, which leverages a convolutional neural network (CNN), self-attention mechanism, Graph Convolutional Network (GCN), and k-reciprocal nearest neighbors (k-reciprocal NN). Specifically, we utilize a CNN to capture individual stock information and a GCN to capture relationships among stocks. Moreover, we incorporate the self-attention mechanism into the GCN to extract deeper data features and employ k-reciprocal NN to enhance the accuracy and robustness of the graph structure in the GCN. In the second stage, we employ the Global Minimum Variance (GMV) model for portfolio optimization, culminating in the AGC-CNN+GMV two-stage approach. We empirically validate the proposed two-stage approach using real-world data through numerical studies, achieving a roughly 35% increase in Cumulative Returns compared to portfolio optimization models without stock pre-selection, demonstrating its robust performance in the Average Return, Sharp Ratio, Turnover-adjusted Sharp Ratio, and Sortino Ratio.

Economic policy uncertainty in OFDI host countries and the cross-section of stock returns

This study examines the association between foreign economic policy uncertainty (FEPU) originating from the host countries (regions) of outward foreign direct investment (OFDI) and expected stock returns. We construct a novel variable, FEPU, based on the year-end OFDI amounts of Chinese listed multinational enterprises (MNEs) and the EPU indices of 23 host countries (regions). Our findings reveal that stocks with higher FEPU outperform those with lower FEPU by 4.96 % annually. Beyond predicting short-term expected returns, FEPU also exhibits strong positive predictive power for firms’ long-term cumulative returns. Additionally, through mechanism tests, we demonstrate that this excess return is attributable to compensation for risk premium. Unlike the majority of studies that focus solely on the impact of EPU within a single country or region, we examine the cross-border implications of EPU. This approach offers an innovative perspective on the pricing of MNEs’ stocks by characterizing the exposure to external EPU.

Portfolio optimization by enhanced LinUCB

We study portfolio optimization through reinforcement learning, employing the contextual LinUCB approach to build an investment portfolio with changing market conditions. Contextual information enhances the LinUCB technique recommendation process, maximizing the efficacy of the model and the cumulative return. The recommendation criteria are to achieve the maximum level of confidence regarding the uncertainty surrounding the upward trend of asset returns within the pool of assets. Employing stocks from the Chinese A-share market, Hong Kong market, and US market as empirical tests, our findings demonstrate that the proposed model consistently outperforms the benchmark indices in our backtesting process.

Portfolio Optimization Using Machine Learning Method and Monte Carlo Simulation

Investment portfolio optimization is a crucial aspect of quantitative finance, aiming to maximize returns and minimize risks. This study focuses on optimizing investment allocation among five selected stocks (BAIC Blue Valley, BYD, Chang’ an Automobile, Kweichow Moutai, Sunwoda) over 100 days using Monte Carlo simulation and machine learning methods. This study uses the sliding window approach with the linear regression model to predict future returns and calculate the variance-covariance matrix to determine the optimal portfolio weights, leading to two key strategies: maximizing the Sharpe ratio and minimizing the risk. The results reveal that the Max Sharpe Ratio portfolio achieved a cumulative return of 0.0909, significantly outperforming the CSI 300 index’s return of -0.0348. Additionally, the Min Risk portfolio exceeded the market index with a cumulative return of 0.0062. These findings demonstrate that both the Max Sharpe Ratio and Min Risk strategies are effective in achieving a balance between risk and return and surpassing the market, offering valuable insights for investors seeking optimized portfolio allocations.

Is Bitcoin a hedge or safe-haven asset during the period of turmoil? Evidence from the currency, bond and stock markets

Market turmoil, such as that induced by Covid-19, tends to create huge pressure on financial markets, forcing decision-makers and investors to analyze risks and manage their investment portfolios. Taking the market turmoil induced by Covid-19 as an example, this study addresses the question of whether Bitcoin exhibits a hedging and safe haven property during the period of turmoil. We employ the TVP-VAR approach to investigate return spillovers and interconnectedness among Bitcoin and stock, money, and bond markets. In addition, we employ three portfolio techniques, including minimum variance portfolios, minimum correlation portfolios, and minimum connectedness portfolios, to evaluate the Sharpe ratios of the portfolios as well as the cumulative return before and during the epidemic. The results demonstrate that incorporating Bitcoin decreases total connectedness across markets, and Bitcoin provides stronger hedging and safe haven capabilities. Notably, the portfolio with minimum connectedness approach reaches the highest Sharpe ratio during the pandemic.

Reinforcement learning-based portfolio optimization with deterministic state transition

Portfolio optimization has attracted substantial interest within the artificial intelligence community due to its significant impact on financial decision-making, risk management, and market analysis. Reinforcement learning fits well with portfolio optimization because their goal is to maximize cumulative returns. In reinforcement learning, state transition probabilities are often unknown and must be estimated. However, in portfolio backtesting experiments, these probabilities are deterministic, making the conventional reinforcement learning approach to estimating state transitions suboptimal for portfolio optimization. Addressing this issue, this study decomposes the portfolio optimization into two core tasks: prediction and profit policy optimization and proposes a novel reinforcement learning framework that assumes deterministic state transition probabilities, comprised of three main modules: feature extraction, prediction, and profit strategy optimization. To model assets more effectively and comprehensively, we capture their temporal features, relational features, and market state. We introduce a patch-wise correlation method and attribute based gate to enhance feature extraction. In the profit policy module, we utilize a deterministic strategy, employing a recursive reinforcement learning method based on Monte Carlo sampling to train the policy network. This enables dynamic adjustments of asset investment weights, ensuring the maximization of cumulative returns. Extensive experiments conducted on cryptocurrency datasets demonstrate the superior performance of our approach, and achieving 36.6%-75.6% improvements in main measurements on cryptocurrency datasets.

Effect of charcoal balls on distillation productivity in hemispherical solar stills: Experimental investigation and performance analysis

This study investigates the enhanced performance of hemispherical solar stills by integrating charcoal balls. The method showcases improved efficiency by augmenting heat transfer mechanisms and enhancing solar radiation absorption. Reducing the surface tension of basin water is a solution to increase its heating, enhancing the softwater yield of solar stills. The novelty of the research lies in the innovative integration of charcoal balls to improve the performance of hemispherical solar stills. This method presents a unique approach to enhancing heat transfer and evaporation dynamics, potentially revolutionizing the efficiency of solar desalination systems. Therefore, the impact of charcoal balls at a diameter of 1 cm on the performance of distillation productivity was studied. A hemispherical solar still containing charcoal balls (HSS-CB) was compared with a simple hemispherical solar still (SHSS) in the same conditions. Experiments were conducted at El-Oued-Algeria in June 2024. The method employed in this research offers distinct advantages, primarily in its ability to enhance solar still efficiency by incorporating charcoal balls. These advantages stem from the increased surface area for heat transfer, improved solar radiation absorption, and potential enhancement of evaporation and condensation processes. Cumulative return results for SHSS and HSS-CB are 4.80 and 6.20 L/m2/day, respectively. The rate of improvement in the cumulative return was 29.16 % for SHSS using the quantity of coal powder. It is recommended that charcoal balls be used to boost solar stills productivity.

Quantile connectedness among fintech, carbon future, and energy markets: Implications for hedging and investment strategies

This study proposes a novel framework to explore the quantile connectedness among fintech, carbon future, and energy markets with implications for hedging and investment strategies. We aim to construct portfolio strategies for extreme market conditions based on the results of quantile connectedness to address the complex price fluctuations in the energy markets. Using the daily data from January 4, 2017, to March 6, 2023, we find that (i) carbon future and crude oil are the net receivers, while fintech and clean energy are the net contributors. However, the net spillover direction of fintech, carbon futures, and energy markets change frequently under extreme market conditions; (ii) the COVID-19 pandemic and Russia-Ukraine conflict significantly enhance connectedness across all quantile levels; (iii) the sample period is dominated by the lower quantile connectedness, indicating that investors react more strongly to bad news. Furthermore, the result of portfolios shows that, compared to normal market conditions, the dynamic investment weights of minimum connectedness (MCOP) exhibit greater volatility under extreme market conditions. This highlights the differences in portfolio adjustments across different market conditions. In addition, the result shows that the minimum connectedness and minimum correlation portfolio have a greater cumulative return than the minimum variance portfolio. The analysis based on the Sharpe ratio shows that in extreme market conditions, carbon futures and fintech can serve as effective hedging tools for the energy markets. Robustness tests further confirmed our conclusions. Our results have positive implications for decision-makers and investors about risk management and diversified portfolio selection.

Integrating LSTM Networks with Mean-Variance Optimization for Enhanced Portfolio Construction: An Empirical Study on UK Stock Market

Portfolio optimization has long been a central theme in finance. With ongoing advancements in machine learning, there is a significant opportunity to integrate predictive methods into portfolio optimization. This paper proposes leveraging Long Short-Term Memory (LSTM) networks alongside the established Mean-Variance (MV) optimization framework to construct optimal portfolios. These portfolios aim to help financial investors effectively manage and mitigate risk while maximizing returns. The study meticulously screened the leading stocks of 12 prominent UK companies listed on the London Stock Exchange (LSE), known for their influence and visibility. Initially, the study applies the LSTM networks to predict stock price volatility and integrates these predictions into the MV model to allocate portfolio weights effectively. To underscore the superiority of the proposed approach, the study compares cumulative returns from portfolios optimized for maximum and minimum variance Sharpe ratios with real data against the FTSE100 index over the same period. The results demonstrate that the portfolio optimized with the maximum Sharpe ratio significantly outperforms both the conservative minimum variance portfolio and the FTSE100 index. This empirical evidence underscores the practical value of integrating advanced machine learning techniques with established financial theory, enabling more innovative and efficient investment strategies to meet the evolving needs of financial investors.

Portfolio Optimization Based on 9 Chinese Stocks

Geopolitical conflicts and the impact of the pandemic have triggered volatility and downward trends in the Chinese stock market, prompting investors to pay more attention to stabilizing returns and controlling risks through optimizing portfolios. This study selected nine stocks listed on the Shenzhen and Shanghai Stock Exchanges in China and first identified the efficient frontier composed of these nine stocks, then constructed portfolios with maximum Sharpe ratio and minimum volatility. The results indicate that China Yangtze Power Co., Ltd. holds the highest investment weight in all portfolios, followed by Kweichow Moutai Co., Ltd. which also holds significant weights in both portfolio types. By comparing the cumulative returns of the two portfolios with the HS300 index, it was found that the performance of all portfolios outperformed the benchmark index. This result can provide guidance for some investors in managing investments during this particular period of market downturn.

Strategic portfolio rebalancing: Integrating predictive models and adaptive optimization objectives in a dynamic market

Adjusting investment strategy is one of the ways to handle dynamic market conditions. This study proposes a novel portfolio management strategy using appropriate optimization objectives for different stock market trends while also incorporating market trends and stock return predictions The optimization objectives that will be evaluated for different market trends are maximizing the Sharpe ratio, minimizing risk, and minimizing expected shortfall. This study utilizes simulation modelling with various predictive models on building the portfolios. The results show that, in an upward market trend, the strategy is to choose stocks with positive returns, and the objective is to maximize the Sharpe ratio. The portfolio that follows this strategy during upward market trends has greater returns than both the Indonesian Composite Index and LQ45, which serve as stock market benchmarks, with 90% certainty. Meanwhile, during the downward market trend, the strategy is to choose stocks with a negative correlation with the Indonesian Composite Index, and the proper optimization objective is to minimize risk. A portfolio that follows this strategy during downward market trends has greater returns than stock market benchmarks with 95% certainty. Across the evaluation period from 2018 to 2023, the portfolio using the proposed strategy outperforms both stock market benchmarks, with a higher quarterly Sharpe ratio of 0.3047 and cumulative return of 107.90%. The proposed portfolio has a higher quarterly return than the stock market benchmark with 99% certainty. Therefore, the proposed strategy shows a promising result in a dynamic market.

Contagion effect between fuel fossil energies and agricultural commodity markets and portfolio management implications

This study investigates the dynamic spillover effects between fossil fuels (Brent oil and natural gas) and major agricultural markets (maize, soybeans, and wheat) using a time-varying parameter vector autoregression (TVP-VAR) model. Additionally, we quantify the cumulative return of the investment portfolio, optimal weights, time-varying risk coverage, and hedging effectiveness of bivariate commodity portfolios using the Minimum Variance Portfolio (MVP), Minimum Correlation Portfolio (MCP), Minimum Connectedness Portfolio (MCoP) under bearish, tranquil, and bullish market conditions. The results show that maize and natural gas are the main net receivers of spillovers, whereas wheat and Brent oil are the main net transmitters of shocks to the network, respectively. Our findings of optimal weights in multivariate portfolio suggest more weights for agri-commodities against fossil fuels in all three modes of market. In addition, the optimal weights of bivariate portfolio analysis show that the most optimal portfolio which has the highest HE value is related to natural gas pairs with agri-commodities. Moreover, depending on the status of the market; in an average and bearish modes, the highest optimal weight is related to maize/natural gas. However, in the bullish market, the highest optimal weights are related to natural gas/Brent oil pair. Furthermore, the results of optimal bivariate hedge ratios show that natural gas/wheat (wheat/maize and vice versa) portfolio is the cheapest (the most expensive) investment strategy in all modes of market and based on MVP and MCoP approaches.

Empirical Analysis of Portfolio Allocation During Market Turbulence Based on Efficient Frontier

The global financial markets have experienced considerable turmoil due to the COVID-19 pandemic, in this context, portfolio management becomes critically important. This paper employs Monte Carlo Simulation to determine Efficient Frontier of a portfolio comprising fourteen US technology stocks in the period of COVID-19 pandemic by python, which also allocates portfolios for Minimum Volatility and Maximum Sharpe Ratio, along with their key parameters. According to this study, Maximum Sharpe Ratio portfolio outperforms Minimum Volatility portfolio in terms of cumulative return. However, its important to note that Maximum Sharpe Ratio portfolio also experiences a greater maximum drawdown. Additionally, this paper visually depicts the optimal allocation of asset weights across different levels of risk aversion using a chart. Furthermore, it identifies NVDA as the optimal choice for investors who have low degree of risk-aversion due to its high volatility and expected return. Conversely, IBM and CSCO emerge as the preferred choices for risk-averse investors, as they effectively mitigate the portfolios entire risk.

Dynamic Returns Connectedness: Portfolio Hedging Implications During the COVID‐19 Pandemic and the Russia–Ukraine War

ABSTRACTWe apply a Time‐Varying Parameter Vector Auto Regressive (TVP‐VAR) connectedness approach on global assets to investigate time‐varying dynamic connectedness, portfolio performance, and hedge effectiveness during COVID‐19 and the Russia–Ukraine war. With increased connectedness and the changing role of energy and soft commodities during these two events, we find the minimum correlation (connectedness) portfolio performing better during COVID‐19 and the Russia–Ukraine war and that cumulative returns of portfolios are higher during COVID‐19. Additionally, we find varying (stable) hedge effectiveness of equity market indices and soft commodities (cryptocurrencies). This paper provides specific insights to investors about using optimal portfolios and hedging during pandemics and military conflicts.

Curriculum learning empowered reinforcement learning for graph-based portfolio management: Performance optimization and comprehensive analysis

Portfolio management (PM) is a popular financial process that concerns the occasional reallocation of a particular quantity of capital into a portfolio of assets, with the main aim of maximizing profitability conditioned to a certain level of risk. Given the inherent dynamicity of stock exchanges and development for long-term performance, reinforcement learning (RL) has become a dominating solution for solving the problem of portfolio management in an automated and efficient manner. Nevertheless, the present RL-based PM methods just take into account the variations in prices of portfolio assets and the implications of price variations, while overlooking the significant relationships among different assets in the market, which are extremely valuable for managerial decisions. To close this gap, this paper introduces a novel deep model that combines two subnetworks; one to learn a temporal representation of historical prices using a refined temporal learner, while the other learns the relationships between different stocks in the market using a relation graph learner (RGL). Then, the above learners are integrated into the curriculum RL scheme for formulating the PM as a curriculum Markov Decision Process, in which an adaptive curriculum policy is presented to enable the agent to adaptively minimize risk value and maximize cumulative return. Proof-of-concept experiments are performed on data from three public stock indices (namely S&P500, NYSE, and NASDAQ), and the results demonstrate the efficiency of the proposed framework in improving the portfolio management performance over the competing RL solutions.