Business, Economics and Management

Abstract: Research on environmental, social and governance performance often asks whether ESG improves firm value and whether profitability or reporting quality transmits that effect. This study shows that the usual linear formulation can conceal the economically relevant pattern. We analyze 1,759 firm year observations for 358 companies listed on the Ho Chi Minh Stock Exchange from 2020 to 2024. The conventional linear model produces only a weak positive ESG coefficient for Tobin's Q and no significant ESG effect on return on assets or absolute discretionary accruals. Standard linear mediation therefore finds no profitability or earnings quality channel. Quadratic models tell a different story. Firm value follows a statistically validated U shape: the ESG coefficient is negative, the squared term is positive, the turning point is 0.335, and the Lind and Mehlum test rejects monotonicity. Profitability displays a closely aligned U shape with a turning point of 0.382. Earnings management exhibits a weaker inverse U, peaking near 0.385, which implies that reporting quality is lowest at intermediate ESG levels. The valuation threshold remains visible in both the 2020 to 2021 and 2022 to 2024 subsamples. Random effects estimates preserve the curvature, whereas firm fixed effects lose precision because most identifying variation is cross sectional. Dynamic system GMM retains the expected signs but does not identify the nonlinear terms precisely. The evidence supports a transition from symbolic to substantive ESG: early engagement may impose costs and invite skepticism, while stronger commitment is associated with higher profitability, better earnings quality and higher market value.

Abstract: The rapid diffusion of decentralized finance (DeFi) protocols across emerging econo-mies raises two related policy questions: whether these platforms broaden access to fi-nancial services and whether their expansion affects incumbent commercial banks. Us-ing an unbalanced panel of ten BRICS Plus economies (2015–2024), we estimate four complementary specifications—two-way fixed effects, system GMM, structural equa-tion modelling with mediation, and panel smooth transition regression—to investigate the DeFi–inclusion–stability nexus. Estimators yield mixed evidence on the DeFi–inclusion link: fixed effects and SEM specifications produce positive but non-significant coefficients, while system GMM yields a small negative estimate once path dependence is absorbed. The direct association between DeFi adoption and non-performing loans is negative and robust across all specifications once multicollinearity between the two Chainalysis indicators is remedied. The PSTR suggests a possible threshold on the Fi-nancial Inclusion Index at ĉ ≈ 0.641, with the DeFi–Z-score association switching from negative below the threshold to positive above it. However, we cannot reject the null hypothesis of linearity at conventional 5% significance (p = 0.057), bootstrap confidence intervals for the regime slopes are wide and straddle zero, and Monte Carlo simulations confirm limited statistical power. Alternative transition variables yield mixed results, suggesting the threshold reflects broader financial development rather than inclusion specifically. The findings are best interpreted as pattern evidence from methodological triangulation rather than definitive causal identification. Given the borderline statisti-cal evidence, we offer policy observations as directional hypotheses for future research rather than definitive prescriptions.

Abstract: Purpose. This study examines whether voluntary environmental, social, and governance (ESG) disclosure improves the financial performance of commercial banks in an emerging market, and whether the effect differs between accounting-based and market-based performance. Design/methodology. Using an unbalanced panel of Vietnamese commercial banks over 2010–2025, we construct a voluntary ESG disclosure index from annual and sustainability reports through content analysis across the environmental, social, and governance pillars. Financial performance is captured by two accounting measures (ROA, ROE) and two market measures (Tobin’s Q, price-to-book). We estimate fixed-effects models for the accounting measures and random-effects models for the market measures, based on F, Breusch–Pagan and Hausman tests, with bank-clustered robust standard errors; we address multicollinearity through a reduced-form specification and endogeneity through a two-step system GMM estimator. Findings. Voluntary ESG disclosure is positively and significantly associated with accounting performance—most strongly with ROE (β = 0.0021, p < 0.01)—and the effect on ROA becomes significant once multicollinearity between disclosure, size and age is mitigated. In contrast, ESG disclosure shows no significant association with either market measure. The positive effect on accounting performance is confirmed by a dynamic system-GMM estimator that controls for endogeneity. Leverage, bank age, liquidity, media pressure and CEO stability are additional determinants of accounting performance. Originality/value. The study provides among the first long-window (16-year) evidence for a frontier banking market and documents a disclosure–performance asymmetry: banks are rewarded in their books but not (yet) in their market valuation, indicating that the capital market has not yet fully priced sustainability transparency.

Abstract: Geopolitical Risk (GPR) has become a central concern for financial stability. This study examines the intellectual and conceptual structure of GPR research in financial markets, banking, and insurance, using a Bibliometric-Systematic Literature Review (B-SLR) framework that combines PRISMA-guided screening with bibliometric performance analysis and science mapping (bibliometrix, VOSviewer). A total of 395 journal articles were retrieved from Scopus, along with a dedicated subsample of 45 documents on banking and insurance applications. Results show that scientific production accelerated sharply after 2018, peaking at 125 articles in 2025. The field’s intellectual core remains organized around the Caldara and Iacoviello GPR Index. Within the banking and in-surance subsample, growth is even steeper, reflecting the very gap that motivated this study, since research on GPR’s transmission to the banking and insurance sectors had remained fragmented and no comprehensive bibliometric mapping of the field had previously been available. Dedicated citation clusters have emerged around bank-ing-focused outlets, while a comparable structure for insurance has yet to develop. These findings carry both theoretical and practical value, offering scholars a structured map for future research while cautioning risk managers and regulators about the still-limited evidence base on insurance.

Abstract: This paper provides a comprehensive analysis of leveraged Exchange-Traded Funds (ETFs) and Exchange-Traded Notes (ETNs), financial instruments that have grown significantly in popularity and market presence since their introduction in the early 2000s. Using data from 2020-2025, we examine the performance characteristics, risk profiles, and potential market impacts of these complex investment vehicles across different asset categories and market conditions. Our analysis reveals significant volatility drag and tracking errors that increase with holding period length and underlying asset volatility. We find that leveraged ETFs tracking technology and semiconductor indices experience the most extreme performance patterns, while fixed income leveraged ETFs show more moderate but still significant decay effects. The paper demonstrates that these products generally fail to deliver their stated multiple of underlying index returns over periods longer than their daily rebalancing horizon, with the divergence increasing during periods of high market volatility. These findings have important implications for individual investors, financial advisors, and regulators concerned with market stability and investor protection.

Abstract: This study constructs a two-period model where a regulator determines the level of precautionary capital, information-generating reporting, and the design of supervisory information systems, while providers respond by participating. When reporting successfully produces a verified diagnostic, continuation capital is set afterward, but reporting incurs various costs, including variable, participation, and fixed setup costs, before its information is utilized. In a Bayesian context, the diagnostic's gross decision value is weakly nonnegative because it can be disregarded. Reporting is only activated if its discounted decision value and any screening benefits outweigh its net costs and setup expenses. Under recursive maxmin assumptions, an admissible prior that makes the adverse state certain introduces a certainty boundary: diagnosis cannot alter the continuation capital once the conditions are met, and precaution substitutes for learning. Priors that are uniformly interior allow for ongoing learning. Recursive smooth ambiguity models positive reporting at finite levels of ambiguity aversion and converges to the maxmin boundary under specified conditions, without implying overall monotonicity. A joint-selection theorem compares scenarios with no reporting, common, and specialized reporting architectures after optimizing intensity. An architecture that keeps the common experiment and adds an ignorable signal slightly improves gross information but may reduce net surplus. Kenya’s virtual-asset framework provides a dated institutional example but neither calibrates the model nor reveals its core mechanism. The results are supported by analytical proofs, independent recalculations, and reproducible code.

Abstract: Carbon pricing raises the cost of fossil-fired electricity inside the regulated jurisdiction but cannot price emissions beyond the regulator's border. Proof-of-work cryptocurrency mining is an unusually clean setting in which to test whether this asymmetry produces carbon leakage: mining is electricity-intensive, geographically footloose, and reallocates across grids far faster than conventional heavy industry. Using daily data from January 2019 to January 2026 (N = 2,550), we estimate quantile regressions of power-sector CO2 emission growth in the EU27, the Russian Federation and the rest of the world on the interaction between Bitcoin returns and European carbon-allowance returns. In Russia the interaction is positive and significant in the lower tail (beta = 0.066, p = 0.001) and survives ordinary least squares with robust standard errors (p = 0.012) and a dynamic specification controlling for emission persistence (p = 0.010); Russia is the only region whose model is jointly significant. In the EU27 and the rest of the world the interaction is not robustly distinguishable from zero at any quantile. The Russian effect is entirely a post-2020 phenomenon, coinciding with the Chinese mining prohibition and the tenfold rise in European allowance prices, while neither other region shifts across the same break. We find no support for a green-paradox reading of green finance: the Green Bond index enters negatively and significantly across specifications. Instruments confined to the emissions-trading perimeter are structurally unable to govern a mobile, electricity-intensive load.

Abstract: Research on airport proximity has mostly concentrated on permanent residents, ignoring vacation resort markets despite their increasing economic significance. Our study examines 9,841 resort property sales from 1983 to 2023 in Mobile and Baldwin Counties, Alabama. We use a hedonic pricing model and geographically weighted regression (GWR) to analyze how airport proximity impacts resort values. A global regression indicates no overall relationship, but GWR shows that this masks strong, spatially varying effects that a single coefficient cannot capture. Properties near airports tend to sell for less, with larger discounts near general aviation airports than commercial airports. This reverses the accessibility premium seen in residential markets for nonurban general aviation. We introduce the Amenity Separation Effect to explain that resort buyers prioritize tranquility over mobility. As a result, airport accessibility does not compensate for its disadvantages, unlike in residential markets. These findings can help improve property assessment, appraisal practices, noise policies, and disclosure requirements in tourism-dependent coastal economies.

Abstract: Green bonds and green bank loans coexist as instruments for financing the low-carbon transition, yet the strategic mechanism through which issuers select between them remains weakly formalized. This paper models green debt instrument choice as a three-stage extensive-form game with perfect information in which a firm first selects a financing route, investors then decide whether to subscribe to a green bond issuance, and a bank sets lending conditions under regulatory incentives. The subgame-perfect equilibrium is characterized analytically by backward induction and is then implemented numerically: the equilibrium correspondence is mapped over the parameter space and evaluated through a Monte Carlo experiment of 200,000 parameter drawing across the eight admissible states of demand, issuer credibility and regulatory regime. Equilibrium green bond issuance requires two conditions to hold jointly, namely investor participation and a demand-driven advantage exceeding the fixed issuance and certification cost. The simulation shows that green bond adoption falls from 79.2% under high demand and strong credibility to 1.5% under moderate demand and weak credibility, with a sharp discontinuity at the participation threshold, and that bank concessionally shifts financing levels without altering the bond-versus-loan margin. The framework provides a tractable basis for assessing sustainability-related financing risk and for the empirical separation of feasibility, optimality and institutional transmission channels.

Abstract: This paper presents a comprehensive and critical examination of the theoretical and empirical underpinnings of liquidity management within the commercial banking sector, with a specific focus on the Republic of Armenia. We argue that liquidity, far from being a monolithic metric, is a multifaceted construct that encapsulates an institution's capacity to meet its financial obligations in a timely and cost-effective manner. This study advances the discourse by integrating seminal theoretical frameworks—from the classical "golden banking rule" to the modern dynamics of asset-liability management (ALM) and funding liquidity risk—with a granular empirical analysis of regional financial heterogeneity. The analysis is structured around two central pillars. First, we decompose the evolution of key regional economic indicators, including firm registrations, wage dynamics, and unemployment, to map the landscape of real economic activity against the backdrop of national liquidity conditions. Second, we employ a correlation matrix to dissect the intricate, often non-linear, relationships between bank-specific variables (e.g., liquid assets to total assets, non-performing loans (NPLs), profitability) and macroeconomic aggregates (e.g., GDP growth, wage inflation). Our findings challenge the conventional wisdom of a homogeneous national liquidity framework, revealing that regional disparities are profound and that the predictive power of traditional liquidity measures is contingent upon the specific economic context. The paper concludes by positing that effective liquidity management is predicated on a "state-contingent" mechanism, one that is responsive to both micro-prudential signals and macroeconomic shocks, and that this is particularly crucial in small, open economies where external vulnerabilities are pronounced.

Abstract: This study aims to identify the financial and non-financial factors affecting audit opinions by using explainable artificial intelligence (XAI) models, which provide an important advantage in terms of transparency in the auditing sector. In this context, 2,415 firm-year observations from 238 manufacturing firms listed on Borsa Istanbul for the period 2010–2024 were used. In the study, the independent audit opinion was used as the dependent variable, while 18 financial and non-financial attributes were employed as independent variables. Descriptive statistics and a correlation heatmap analysis were conducted. To identify the most influential factors affecting audit opinions and to comparatively reveal which factors are more important across different models, explainable predictive models were developed using logistic regression, decision tree, random forest, XGBoost, LightGBM, and CatBoost models. Three different analyses were performed: permutation feature importance analysis, SHapley Additive exPlanations (SHAP) analysis, and SHAP beeswarm plot analysis. The built-in feature importance heatmap and the SHAP feature importance heatmap for all models were presented comparatively. A final heatmap was then constructed based on the averages of the built-in feature importance scores and SHAP values. According to the results of the final heatmap, the type of audit opinion in the previous year (X17) was identified as the most influential factor with a score of 0.78, followed by asset turnover ratio (X7) with 0.57, operating sector (X18) with 0.54, current ratio (X1) with 0.54, acid-test ratio (X2) with 0.51, and the natural logarithm of total assets (X10) with 0.50, all of which were found to have the highest explanatory power. The final comparison table, constructed by averaging both the models’ built-in importance scores and SHAP values, indicates that the results are not random but are based on a robust foundation. The findings of this study make significant contributions to both the academic literature and practitioners by guiding decision-makers in identifying the factors that should be prioritized.

Abstract: This study introduces a reliable portfolio optimization framework that utilizes Mean-Variance (MV) Optimization together with Kullback-Leibler (KL) divergence under the AR-GJR-GARCH filtering process for capturing non-normality, volatility clustering, and model uncertainty in the returns of financial assets. The introduced Mean-Variance-KL (MV-KL) method incorporates the maximization of expected return, risk, and proximity to investor-specific target return distributions, while increasing the efficiency and robustness of the portfolio through portfolio diversification. The MV-KL portfolio is developed based on the returns from the African and worldwide financial assets using the AR-GJR-GARCH model to extract the conditional mean and volatility processes before constructing the optimal portfolios. The portfolio evaluation metrics used include residual return, residual volatility, Sharpe and Sortino ratios, as well as VaR and CVaR. The results indicate that although the classical MV approach generates the best stable risk-adjusted returns among the other methods. The proposed hybrid MV-KL method is more diversified and provides better downside risk management than the conventional portfolio techniques. The integration of KL divergence allows superior flexibility in adapting investor preferences and improves portfolio resilience under varying market conditions. These findings add to the increasing literature on robust portfolio optimization by providing an information-theoretic approach to dynamic asset allocation.

Abstract: Predicting whether asset prices will rise or fall is essential for investment decision-making, since even modest improvements in directional accuracy can produce substantial economic benefits. This study proposes a dynamic sliding-window (DSW) framework for the daily directional classification of Exchange-Traded Funds (ETFs). Unlike conventional forecasting approaches that rely on either a fixed training sample or an expanding window, the DSW method allows the length of the estimation window to change at every prediction step. The underlying premise is that observations associated with market conditions that are most relevant to the current regime may provide more useful predictive information than a larger volume of older data. For each forecast, the optimal dynamic sliding-window size is selected through Bayesian optimisation applied to an internal validation segment of the available training sample, with a Gaussian process used as the surrogate model. A linear Support Vector Machine is then estimated using the observations contained in the selected window and a set of 39 technical indicators representing different dimensions of market behaviour. The empirical evaluation is conducted on 156 US-listed ETFs obtained from Yahoo Finance over the period from September 2019 to September 2024, yielding more than 170,000 daily observations. The dynamic sliding-window model achieves statistically significant improvements across all evaluation measures when compared with both an expanding, or stretching, window and a conventional train–test split. These results show that dynamically adapting the amount of historical data used for model estimation can improve financial time-series classification. The proposed DSW framework therefore provides a scalable and transparent approach to forecasting in markets characterised by changing regimes and persistent volatility.

Abstract: International climate finance is widely presumed to accelerate green technological innovation in recipient economies, and recent studies report positive effects. This study examines whether that proposition survives systematic scrutiny. Using OECD ENV-TECH patent indicators and OECD DAC Rio Marker climate-mitigation commitments for 118 developing economies between 2004 and 2022, the analysis estimates 180 specifications based on two estimators, three treatment scalings, five denominator thresholds, alternative trend specifications, and sample restrictions. Although 90% of the estimates are positive and 34% reach conventional significance, significance is confined to two identifiable dependencies. Count-based (quasi-Poisson) models are dominated by China, which accounts for 78.7% of the sample’s inventions despite representing only 10.7% of the world total: all thirty count specifications are significant with China included, and none is significant once China and India are excluded. Fractional-response models on the green invention share are invariant to dropping large recipients but derive their significance entirely from country-years with fewer than four total inventions, where a single patent moves the outcome by tens of percentage points. Of the 48 specifications that both exclude China and impose a minimally informative denominator, none is significant; the median standardised effect is 0.03. The findings indicate no reliable evidence that bilateral climate-mitigation finance raises green innovation, and identify a structural weighting problem that plausibly explains positive results in smaller samples.

Abstract: In a financial environment increasingly shaped by persistent volatility and recurrent phases of instability, understanding the conditions under which trend-following strategies remain effective has become a central concern in empirical economic research. This article reassesses the performance of widely used technical indicators, including moving averages, the Relative Strength Index, the Moving Average Convergence Divergence (MACD), the stochastic oscillator, and Bollinger Bands, across distinct market regimes. The analysis is based on a dataset of 500 daily financial observations obtained from a financial trading platform and covering distinct market conditions. The dataset exhibits key characteristics commonly observed in financial time series, including volatility clustering and non-linear price dynamics, allowing a focused examination of the structural behaviour of technical indicators across different market regimes. A GARCH model is applied to capture the persistent and clustered nature of volatility, a structural element that significantly influences the reliability of technical signals. The findings indicate that these indicators tend to retain satisfactory performance during periods of market stability, but their effectiveness deteriorates markedly during episodes of crisis, when volatility shocks become more intense and persistent. These dynamics substantially increase the frequency of incorrect signals and weaken the capacity of trend-following approaches to provide reliable guidance. The study also shows that strategies combining trend-based indicators with volatility-sensitive measures offer greater robustness in turbulent environments. From a practical standpoint, these findings suggest that systematic traders and risk managers should embed volatility regime detection into their signal generation processes rather than applying technical indicators indiscriminately across all market conditions. Taken together, the results contribute to a renewed economic reading of trend-following by highlighting the structural conditions that drive its effectiveness and by identifying the limitations imposed by unstable market environments.

Abstract: The focus and objective of this study are to conduct a comprehensive evaluation of the impact of the trade war between the United States and China, examining the degree of integration and causal relationships, as well as its spillover effects on stock and international financial markets in Asia and the United States. The data used were obtained by selecting a sample of 11 stock market indices from both regions, with monthly data covering the period from 2017 to 2024. The methodological approach to data analysis involved applying the Vector Error Correction Model (VECM) test, combined with the Granger Causality test, Johansen Cointegration, and the Impulse Response Function (IRF), as well as forecasting using the Forecast Error Variance Decomposition (FEVD) approach. The results of this study provide evidence that the greatest spillover effects and impacts are found in the stock and financial markets of Singapore, Japan, the Philippines, China and the United States. This implies a strong relationship between the stock markets in these two regions. Using the IRF and FEVD methods, it has been demonstrated that external shocks result in significant changes, whilst also showing that market volatility is temporary and will return to equilibrium once economic and financial stability is restored in both Asia and the United States. Fundamentally, the novelty of this research lies in simultaneously testing 11 stock markets across Asia and the United States, demonstrating that the trade war between the United States and China has triggered a paradigm shift: it is not only bilateral relations that are affected by this conflict, but the trade war also has implications for other countries or regions, where market integration, spillover effects and dynamic inter-market interdependence are evident. Thus, the findings of this study demonstrate that the interconnectedness of capital markets, both regionally and globally, is very strong and simultaneously reduces the effectiveness of international diversification strategies in decision-making due to the trade war between these two countries.

Abstract: This paper examines analyst earnings forecast accuracy for S&P 500 and Russell 2000 firms from 2015 to 2025, focusing on the COVID-19 pandemic and the subsequent AI era, which began in late 2022. Results show S&P 500 firms consistently have lower forecast errors than Russell 2000 firms, although the difference narrowed after the widespread adoption of AI-assisted tools. During the pandemic, forecast accuracy declined slightly for S&P 500 firms but much more for Russell 2000 firms, while it improved substantially for both groups during 2023–2025. These results remain similar after reducing the influence of extreme observations. The analysis of signed forecast errors indicates a persistent pessimistic bias for S&P 500 firms, while forecast bias for Russell 2000 firms was weaker and less consistent. An interesting finding is that the historical difference in forecast bias between S&P 500 and Russell 2000 firms disappeared in the later period and was no longer statistically significant.

Abstract: We examine how restricting high-frequency trading (HFT) affects stock liquidity in China’s A-share market. Using China’s 2024 Provisions on Program Trading in the Securities Market as a quasi-natural experiment, we construct a stock-level high-frequency trading intensity index from tick-level order data and apply a difference-in-differences design. The regulation significantly improves liquidity by reducing bid-ask spreads and price impact. The effect is concentrated during stock price declines, suggesting that algorithmic traders withdraw liquidity when liquidity provision is most valuable. The improvement is stronger for small- and mid-cap stocks and margin-tradable stocks, which are more vulnerable to liquidity fragility. These findings show that targeted regulation of algorithmic trading can strengthen market resilience and investor protection in retail-dominated emerging markets.

Abstract: AI-driven credit scoring is supervised as a high-risk application in banking and insurance, yet unfairness is rarely operationalized as a measurable category of model, conduct, legal, and reputational risk. Using 20,000 anonymized applications from a Southern European digital lender (15.2% twelve-month default rate), we compare a regularized logistic regression with gradient-boosting and neural-network classifiers, estimated under baseline and fairness-aware configurations (reweighing, adversarial debiasing, and reject-option adjustment), evaluating performance (AUC-ROC, Brier score, F1) jointly with group fairness (demographic-parity and equal-opportunity differences, disparate-impact ratio, Theil index). The interpretable benchmark attains an AUC of 0.734 and a Brier score of 0.162 yet already disadvantages the marginalized ethno-socioeconomic subgroup (disparate-impact ratio 0.67; equal-opportunity difference −0.165), despite excluding sensitive attributes. Higher-capacity models raise accuracy while widening disparities: the neural network reaches AUC 0.792 but a ratio of only 0.58. Mitigation reverses this cheaply, lifting the ratio to 0.85 for 1.1 AUC points under reweighing and to 0.84 while retaining an AUC of 0.765 under adversarial debiasing, so the presumed fairness–accuracy trade-off is weaker than assumed. We map these results onto the EU AI Act, GDPR, EBA, Solvency II, EIOPA, and IAIS frameworks and propose fairness-risk controls based on impact assessment, validation, and three-lines-of-defense governance.

Abstract: This study examines whether renewable energy production is capable of covering energy costs in European Union family farms and identifies the determinants of the Energy Cost Coverage Ratio across different economic size classes. The analysis was based on Farm Sustainability Data Network (FSDN) data for 2014-2023 and combined descriptive statistics with panel data models. The results indicate that the Energy Cost Coverage Ratio remained relatively stable, fluctuating between 30.9% and 39.2%, despite a substantial increase in renewable energy production from €1594 to €2999 per farm. This limited improvement resulted from a simultaneous rise in energy costs, which increased from €5162 to €8465 per farm over the analysed period. Considerable differences in the Energy Cost Coverage Ratio were observed between economic size classes, while panel data models showed that its determinants vary across farm classes, with no single factor being significant for all classes. The findings suggest that renewable energy has strengthened the economic resilience of European farms, but its contribution remains insufficient to fully offset rising energy expenditures, highlighting the need for farm-size-specific policy support.