In sustainable portfolio management, categorizing assets as “brown“ or “green“ based solely on ESG ratings can be misleading. A positive ESG score does not inherently indicate environmental responsibility unless it is evaluated relative to a meaningful benchmark. We propose a rescaled ESG rating system that measures each asset’s environmental standing relative to a threshold set by policymakers, reflecting the urgency of the current climate crisis. In this system, assets are assigned positive scores if they exceed the threshold (green) and negative scores if they fall below it (brown), enhancing the interpretability of sustainability metrics in portfolio construction. However, a challenge arises when aggregating these scores into an overall portfolio rating. Under sustainable portfolio optimization developed in [11], short positions in brown assets, otherwise effectively betting against polluting companies, can paradoxically improve the portfolio’s sustainability score. This creates a misleading incentive structure. To address this, we introduce a constraint that prohibits short positions in brown assets, ensuring that such investments do not positively impact the portfolio’s sustainability rating. While this restriction better aligns with environmental objectives, it also introduces complexity into the optimization process. To resolve this, we present an intuitive algorithm inspired by the active set method, which we refer to as Green Portfolio Optimization, capable of handling these constraints efficiently even in high-dimensional settings.

This study examines the performance of robo-advisors within the broader digital transformation of financial services. Robo-advisors automate portfolio construction and maintenance through algorithmic frameworks that apply established investment principles and low-cost ETFs, thereby extending professional investment management to individuals lacking the time, resources, or expertise traditionally required. Focusing on Wealthfront’s Classic Portfolio from 2013 to 2023, the analysis evaluates absolute and risk-adjusted returns, volatility and drawdown dynamics, and factor exposures to distinguish systematic risks from potential investment skill. Results show that passive indexing outperformed all examined robo-advisor portfolios on both absolute and risk-adjusted bases during a decade dominated by strong U.S. equity performance. Although robo-advisors successfully delivered calibrated risk exposure, their diversified multi-asset allocations incurred notable opportunity costs in a growth-driven market. The platforms offer the greatest value to conservative investors, while more aggressive investors may pay advisory fees without receiving proportional benefits. JEL classification numbers: G11, G51. Keywords: Robo-advisor, Artificial intelligence, Performance evaluation, Risk-adjusted performance.

Asset allocation is a foundational concept in finance that determines how investment capital is distributed across different asset classes to balance risk and return. It plays a critical role in portfolio construction, long-term wealth creation, and financial stability for individual and institutional investors. This theoretical article examines asset allocation from classical and contemporary financial perspectives, drawing on portfolio theory, behavioral finance, and strategic investment frameworks. The study explores the theoretical foundations of asset allocation, including risk–return trade-offs, diversification benefits, and the role of investor objectives and constraints. It further analyzes strategic, tactical, and dynamic asset allocation approaches, highlighting their relevance under varying market conditions. By synthesizing existing literature, the paper conceptualizes asset allocation as a dynamic decision-making process influenced by economic cycles, market efficiency, and investor behavior. The article contributes to finance literature by providing an integrated theoretical understanding of asset allocation and identifying directions for future empirical research and practical application in portfolio management.

Portfolio Construction and Technical Analysis: Evidence from IT and Banking Sectors in India

This paper examines how cryptocurrencies can be incorporated to the corporate investment portfolio and how the risk management strategies can be implemented to make the cryptocurrencies usage a success. The sheer volatility and uncertainty that have been seen in the cryptocurrency markets pose great threats to management of corporate treasury and institutional investment. By thoroughly examining the processes of portfolio construction, risk measurement models, and the protective resources, the research gives recommendations to corporations thinking about using cryptocurrencies. The study compares passive and active investment style whereby the performance benchmarking is considered in a variety of market regimes such as crash periods, flat markets, bullistic market and bearish market trends. The research has revealed that the traditional diversification gains are narrow in the cryptocurrency market, and altcoins do not offer significant risk mitigation as compared to Bitcoin. Nevertheless, predictability through momentum facilitates effective downside protection tactical allocation strategies which retain the upside participation strategies. The paper puts forward an Optimal NAV Protect strategy, which is a combination of minimum-variance allocation and momentum-driven tactical exposure, and has a better performance based on risk adjustment in different market environments. The practise provides corporations with a viable model of cryptocurrency integration that is compensatory in its return targets and institutional risk limitations. The analysis adds to the knowledge of the cryptocurrency portfolio dynamics, its risk management and consideration of implementation with corporate investors that operate within this new asset class.

ABSTRACT This study examines the dynamic interconnections and portfolio implications of clean energy ETFs, artificial intelligence (AI) indices, crude oil, and Bitcoin within sustainable and technology-driven financial markets. Using a Time-Varying Parameter Vector Autoregression (TVP-VAR) framework and daily data from January 2019 to December 2024, we analyze time-varying spillovers and construct optimal portfolios based on dynamic connectedness measures. The results show that clean energy and AI-related assets display relatively stable portfolio weights, whereas Bitcoin exhibits highly volatile and generally limited allocations, particularly under risk-averse strategies. Conventional approaches such as the Minimum Variance and Risk Parity portfolios tend to favor traditional assets, while the Maximum Connectedness Portfolio enhances diversification by allocating more weight to weakly connected assets, including Bitcoin and green ETFs. The findings offer practical insights for resilience-oriented and innovation-driven portfolio construction.

The aim of this research is to explore the composition that results in an optimal investment portfolio from stocks listed in the IDXQ30 index using the Single Index Model (SIM). This research Time Frame covers January 2021 to December 2024, with 14 consistently listed stocks identified as the initial sample. According to the calculation of Excess Return to Beta (ERB) and the determination of the cut-off point (C*), six stocks were selected as portfolio candidates, namely BMRI, BBCA, UNTR, PTBA, CPIN, and ACES. The final results presented in the optimal weighting table reveal that only four stocks are belong to the optimal portfolio, comprising BMRI, BBCA, UNTR, and PTBA. The greatest portion of the fund allocation was achieved by BMRI at 44.43%, followed closely by BBCA at 44.06%. In contrast, UNTR and PTBA contributed smaller weights of 7.33% and 4.18%, respectively. This composition indicates that the financial sector, represented by BMRI and BBCA, dominates the optimal portfolio, while other sectors provide additional diversification. These findings emphasize that stocks with the highest efficiency in generating returns relative to risk are prioritized in portfolio construction, thereby producing an allocation that balances return potential and risk exposure

This paper examines whether information-theoretic complexity measures enhance industry-group return forecasting and portfolio construction within a machine-learning framework. Using daily data for 25 U.S. GICS industry groups spanning more than three decades, we augment gradient-boosted decision tree models with Shannon entropy and fuzzy entropy computed from recent return dynamics. Models are estimated at weekly, monthly, and quarterly horizons using a strictly causal rolling-window design and translated into two economically interpretable allocation rules, a maximum-profit strategy and a minimum-risk strategy. Results show that the top performing strategy, the weekly maximum-profit model augmented with Shannon entropy, achieves an accumulated return exceeding 30,000%, substantially outperforming both the baseline model and the fuzzy-entropy variant. On monthly and quarterly horizons, entropy and fuzzy entropy generate smaller but robust improvements by maintaining lower volatility and better downside protection. Industry allocations display stable and economically interpretable patterns, profit-oriented strategies concentrate primarily in cyclical and growth-sensitive industries such as semiconductors, automobiles, technology hardware, banks, and energy, while minimum-risk strategies consistently favor defensive industries including utilities, food, beverage and tobacco, real estate, and consumer staples. Overall, the results demonstrate that entropy-based complexity measures improve both economic performance and interpretability, yielding industry-rotation strategies that are simultaneously more profitable, more stable, and more transparent.

This paper introduces a new framework for high-dimensional covariance matrix estimation, the Blockwise Exponential Covariance Model (BECM), which extends the traditional block-partitioned representation to the log-covariance domain. By exploiting the block-preserving properties of the matrix logarithm and exponential transformations, the proposed model guarantees strict positive definiteness while substantially reducing the number of parameters to be estimated through a blockwise log-covariance parameterization, without imposing any rank constraint. Within each block, intra- and inter-group dependencies are parameterized through interpretable coefficients and kernel-based similarity measures of factor loadings, enabling a data-driven representation of nonlinear groupwise associations. Using monthly stock return data from the U.S. stock market, we conduct extensive rolling-window tests to evaluate the empirical performance of the BECM in minimum-variance portfolio construction. The results reveal three main findings. First, the BECM consistently outperforms the Canonical Block Representation Model (CBRM) and the native 1/N benchmark in terms of out-of-sample Sharpe ratios and risk-adjusted returns. Second, adaptive determination of the number of clusters through cross-validation effectively balances structural flexibility and estimation stability. Third, the model maintains numerical robustness under fine-grained partitions, avoiding the loss of positive definiteness common in high-dimensional covariance estimators. Overall, the BECM offers a theoretically grounded and empirically effective approach to modeling complex covariance structures in high-dimensional financial applications.

ABSTRACT This paper examines how incorporating regret aversion affects portfolio construction and performance in emerging stock markets. Using daily data from 2015–2025, the paper builds multivariate six-asset portfolios for Central and Eastern Europe (CEEC), East Asia, and Latin America, alongside a G7 benchmark. A regret-minimising portfolio is compared with traditional minimum-variance and maximum-Sharpe portfolios across pre-crisis and crisis periods. Results show that regret-averse portfolios are more diversified and place greater emphasis on low-correlation assets, reducing the likelihood of extreme underperformance relative to the best-performing asset. Emerging markets exhibit higher regret than developed markets, with the Latin American portfolios showing the largest regret levels. Notably, the CEEC portfolio consistently demonstrates the strongest performance, delivering the lowest regret and favourable risk-return outcomes in both subsamples. Robustness checks, including bootstrapping, variance-equality tests, and varying regret preferences, confirm the stability of the regret-minimising approach and highlight its practical relevance for behaviourally sensitive investors.

This study evaluates the Duration Times Spread (DTS) framework in the Korean corporate bond market. Using a comprehensive dataset covering more than 3.8 million bond–month observations from 2010 to 2020, we examine whether the key empirical properties underlying DTS—the proportional relationship between spread levels and spread volatility—hold in this market. Our results show that both systematic and idiosyncratic spread volatilities are proportional to the level of spreads, and that excess return volatility increases proportionally with DTS. These findings confirm the central assumptions of the DTS framework and demonstrate that relative spread changes provide a more stable basis for forecasting excess return volatility than absolute spread changes. When compared against the traditional spread-duration approach, DTS yields residuals that are closer to zero on average and more tightly distributed, indicating superior forecasting accuracy. Overall, the evidence extends the applicability of DTS to an understudied fixed-income market and highlights its usefulness for credit risk measurement, portfolio construction, and volatility forecasting in the Korean corporate bond market.

Traditional actuarial risk assessment relies primarily on historical return correlations to evaluate portfolio concentration and tail risk. However, corporate textual disclosures contain forward-looking information about business models, risk exposures, and strategic similarities that may not be fully captured in historical price relationships. This paper develops a systematic framework for incorporating semantic analysis of corporate disclosures into actuarial risk measurement, demonstrating applications to portfolio construction, concentration risk assessment, and stress testing. Using JSE All Share constituents, we apply natural language processing techniques to extract semantic similarity matrices from corporate narratives across business model, geopolitical, and ESG dimensions. Through shrinkage estimation, we combine these semantic correlations with traditional methods to produce enriched risk measures. Our analysis reveals that semantic correlations identify concentration risks not visible in historical data, with Effective Number of Bets reductions of 75%. The framework provides practical tools for incorporating qualitative corporate information into quantitative risk assessment processes.

We present the Resource Allocation Transformer, a deep learning framework that learns portfolio-level relationships directly through attention mechanisms, capturing how assets work together rather than only how they perform individually. Most machine learning approaches to portfolio construction reduce allocation to aggregating independent asset predictions, overlooking the complementarity between assets that drives optimal portfolios. Unlike traditional predict-then-optimise pipelines, or Economic Scenario Generators that separate the modelling of economic variables from the optimisation step, the Resource Allocation Transformer integrates correlation structure and optimisation logic within a single differentiable architecture. The framework learns constraint-satisfying allocations through self-supervised exposure to synthetic optimisation problems, providing a more stable alternative to sequential prediction-optimisation workflows. Empirical validation shows effective transfer learning from synthetic curricula to real Johannesburg Stock Exchange data (2005–2024), with the same trained model handling portfolios of varying sizes and across market regimes without retraining. By directly learning to allocate, the Resource Allocation Transformer establishes a new paradigm for asset allocation that adapts through experience rather than requiring problem-specific recalibration.

This study examines the risk profile of the Moscow Exchange’s sustainable development indices (MESG, MRRT, MRSV, MRSVR) in comparison with the benchmark Russian stock market index IMOEX over the 2024 calendar year. The objective was to quantify differences in returns, volatility, systematic risk, and tail-risk measures between ESG-oriented indices and the broad market. Monthly index values were used to compute simple returns, annualized volatility, beta coefficients, historical Value-at-Risk (VaR), Conditional Value-at-Risk (CVaR), and maximum drawdown, supplemented by correlation analysis and Levene’s test for equality of variances. The results indicate that all indices ended the period with negative annual returns, while their risk characteristics diverged considerably. MESG and MRRT demonstrated lower annual volatility and beta values below one, suggesting reduced sensitivity to market fluctuations and a relatively more stable risk profile. In contrast, MRSV and MRSVR exhibited elevated volatility, the highest VaR and CVaR levels, and deeper drawdowns. Despite visual differences, Levene’s test did not confirm statistically significant variance disparities between MESG and IMOEX, reflecting the limitations of a short time series. The scientific novelty of the study lies in the integrated comparison of Russian ESG indices using a consistent framework of traditional and tail-risk metrics. The findings expand empirical evidence on the behavior of ESG-oriented instruments in emerging markets with structural and institutional specificities. The practical value of the research is associated with the application of the obtained risk indicators in portfolio construction, assessment of market resilience, and refinement of ESG-index methodologies under conditions of heightened volatility.

The study examined the connectedness among bitcoin, green bonds (represented by the US S&P Green Bond Index), renewable energy (represented by the OMX Biofuel Index), and gold, utilizing a novel quantile connectedness approach from 14 November 2017 to 30 May 2024. This approach contributes to understanding the transmission mechanisms, influence, and connectedness among the bitcoin, green bond, renewable energy, and gold markets. The result indicates that significant values appear at specific intervals. A significant spike was observed at specific intervals around 2019, mainly due to the trade war between the U.S. and China. A subsequent shock occurred between 2020 and 2021, driven by the COVID-19 pandemic. Moreover, the US credit crisis exacerbated volatility spillovers and financial contagion across markets, worsening these effects in 2023 and intensifying volatility spillovers and financial contagion across markets, exacerbating their outcomes. Additionally, the results suggest that Bitcoin primarily serves as a receiver of shocks. At the same time, the green bond transmits the shocks, and renewable energy and gold have switched between transmission and receiving shock roles during the period. The findings offer valuable insights into sustainable portfolio construction, highlighting that green bonds serve as primary transmitters of shocks and suggest a role as diversification anchors during market stress. Additionally, recognizing Bitcoin as a shock absorber and the shifting roles of renewable energy and gold help investors optimize risk-hedging strategies and enhance portfolio resilience across varying market conditions. This indicates that understanding how these assets correlate across various market scenarios is crucial to maximizing portfolio performance while accounting for sustainability constraints.

With the digital transformation of society, the diversity of portfolios is increasingly improving. As a result, the methods of achieving optimum portfolio asset allocation attract more and more investors. Therefore, research on portfolio asset allocation is imperative. This research is aimed at a portfolio construction strategy based on the maximum Sharpe ratio. Eight A-share market stocks and three Exchange-Traded Fund (ETF) are selected as research objects. Sharpe ratio is regarded as a key indicator that balances the performance between portfolio returns and volatility. As the research showcases, a diversified portfolio among different asset categories is capable of reducing unsystematic risk and improving the stability of an asset allocation, thus increasing the Sharpe ratio. The result of the research proves the effectiveness of the asset diversification theory, whose potential for reducing unsystematic risk and improving portfolio returns is indicated. This research provides practical strategic recommendations with their practical significance and application value. However, the assumption of a normal distribution in portfolio returns, as implied by the Sharpe ratio, may not always hold, which can be further addressed by a copula-based approach.

Enhancing Portfolio Construction with Correlation Estimates from Large Language Models

ABSTRACT The capital asset pricing model (CAPM) is a widely adopted model in asset pricing theory and portfolio construction because of its intuitive nature. One of its main conclusions is that there exists a global market portfolio that each rational investor should hold in proportion to the risk‐free asset. In this paper, we demonstrate theoretically and through an example that the CAPM cannot hold in a multi‐currency environment. This is because it produces different market risk premia depending on the investor's base currency unless each exchange rate is uncorrelated with the asset prices in the portfolio.

This study aims to analyze the construction of an optimal investment portfolio using the Markowitz Model for stocks listed in the LQ-45 index. The LQ-45 stocks were chosen because they represent leading and highly liquid companies on the Indonesia Stock Exchange, consisting of diverse industrial sectors that provide broad opportunities for risk reduction through diversification. The research method includes the calculation of expected returns, standard deviations as a measure of risk, and covariance among selected stocks to examine the interrelationship of asset movements. These components are then used to determine the optimal portfolio composition that offers the maximum possible return for a given level of risk or, conversely, minimizes risk for a targeted expected return. The results of the analysis demonstrate that the Markowitz Model is effective in forming portfolios that align with different investor risk preferences. The diversification effect generated by combining stocks across various sectors significantly reduces unsystematic risk without diminishing return potential. The findings also indicate that an optimal portfolio constructed using this method can serve as a practical and strategic investment alternative for both individual and institutional investors, particularly in a volatile market environment. Furthermore, this study provides valuable insights into portfolio selection strategies that can support investment managers in developing evidence-based decision-making frameworks. By applying modern portfolio theory, investors can better understand risk–return trade-offs and improve the efficiency of their investment allocations in the Indonesian capital market.

ABSTRACT This study examines whether investor sentiment expressed on X (formerly Twitter) can be used to build outperforming stock portfolios in Brazil. Using over 1,500 daily return observations from 394 companies and a robust sentiment index built with dynamic Portuguese-language dictionaries and machine learning, we test hypotheses via VAR and multifactor OLS regressions. Results show that while sentiment does not Granger-cause returns, its momentum is positively associated with performance. The new sentiment risk factor is statistically significant in some models, especially with high-liquidity stocks. These findings offer practical implications for asset pricing and portfolio construction in Latin America’s largest emerging capital market.