Financial inclusion is a longstanding concern across underdeveloped communities, particularly for women. However, there are limited data-driven measures to first quantitatively identify such concerns and second to inform policies. In this work, we examine the digital money service adoption and women's financial inclusion in the context of Pakistan. We use the nationally representative Global Findex data from the World Bank to analyze how mobile money usage, when moderated by phone ownership, internet access, and education, affects women's access to formal financial services. Our findings show that women who adopt mobile money services have significantly higher odds of accessing formal financial systems. Findings also reveal nuanced insights: internet access does not significantly impact inclusion, highlighting the influence of socio-cultural constraints. Despite the limitations of using cross-sectional data and the absence of qualitative dimensions, our study contributes empirical evidence on gendered digital finance adoption. The findings have important implications for policy, including the need for women-centric fintech design and digital literacy reforms to bridge the gender gap in financial inclusion.
Large Language Models (LLMs) have demonstrated significant potential across various domains, particularly in banking and finance, where they can automate complex tasks and enhance decision-making at scale. Due to privacy, security, and regulatory concerns, organizations often prefer on-premise deployment of LLMs. The ThaiLLM initiative aims to enhance Thai language capabilities in open-LLMs, enabling Thai industry to leverage advanced language models. However, organizations often face a trade-off between deploying multiple specialized models versus the prohibitive expense of training a single multi-capability model. To address this, we explore model merging as a resource-efficient alternative for developing high-performance, multi-capability LLMs. We present results from two key experiments: first, merging Qwen-8B with ThaiLLM-8B demonstrates how ThaiLLM-8B enhances Thai general capabilities, showing an uplift of M3 and M6 O-NET exams over the general instruction-following Qwen-8B. Second, we merge Qwen-8B with both ThaiLLM-8B and THaLLE-CFA-8B. This combination results in further improvements in performance across both general and financial domains, by demonstrating an uplift in both M3 and M6 O-NET, Flare-CFA, and Thai-IC benchmarks. The report showcases the viability of model merging for efficiently creating multi-capability LLMs.
This study empirically examines the impact of fiscal policy on economic growth in East Java Province during the period 2018–2024. The objective is to analyze the effects of regional taxes, subsidy expenditures, and the Value Added Tax (VAT) rate adjustment on regional economic growth. A quantitative approach with multiple linear regression analysis is employed, incorporating a dummy variable to capture the impact of the VAT increase from 10% to 11% in 2022. The findings reveal that regional taxes and the VAT dummy variable significantly influence economic growth, while subsidy expenditures are statistically insignificant. The coefficient of determination (R²) of 96.5% indicates that the model explains the majority of variations in economic growth. These results imply that optimizing tax revenues and evaluating VAT policy should be aligned with strategies for inclusive and sustainable regional economic growth. This research provides relevant insights for regional policymakers in designing effective fiscal instruments.
Eliot Brenner, Dominic Seyler, Manjunath Hegde
et al.
Despite advances in generative large language models (LLMs), practical application of specialized conversational AI agents remains constrained by computation costs, latency requirements, and the need for precise domain-specific relevance measures. While existing embedding models address the first two constraints, they underperform on information retrieval in specialized domains like finance. This paper introduces a scalable pipeline that trains specialized models from an unlabeled corpus using a general purpose retrieval embedding model as foundation. Our method yields an average of 27.7% improvement in MRR$\texttt{@}$5, 44.6% improvement in mean DCG$\texttt{@}$5 across 14 financial filing types measured over 21,800 query-document pairs, and improved NDCG on 3 of 4 document classes in FinanceBench. We adapt retrieval embeddings (bi-encoder) for RAG, not LLM generators, using LLM-judged relevance to distill domain knowledge into a compact retriever. There are prior works which pair synthetically generated queries with real passages to directly fine-tune the retrieval model. Our pipeline differs from these by introducing interaction between student and teacher models that interleaves retrieval-based mining of hard positive/negative examples from the unlabeled corpus with iterative retraining of the student model's weights using these examples. Each retrieval iteration uses the refined student model to mine the corpus for progressively harder training examples for the subsequent training iteration. The methodology provides a cost-effective solution to bridging the gap between general-purpose models and specialized domains without requiring labor-intensive human annotation.
In specialized domains, humans often compare new problems against similar examples, highlight nuances, and draw conclusions instead of analyzing information in isolation. When applying reasoning in specialized contexts with LLMs on top of a RAG, the pipeline can capture contextually relevant information, but it is not designed to retrieve comparable cases or related problems. While RAG is effective at extracting factual information, its outputs in specialized reasoning tasks often remain generic, reflecting broad facts rather than context-specific insights. In finance, it results in generic risks that are true for the majority of companies. To address this limitation, we propose a peer-aware comparative inference layer on top of RAG. Our contrastive approach outperforms baseline RAG in text generation metrics such as ROUGE and BERTScore in comparison with human-generated equity research and risk.
Xueqing Peng, Triantafillos Papadopoulos, Efstathia Soufleri
et al.
Despite Greece's pivotal role in the global economy, large language models (LLMs) remain underexplored for Greek financial context due to the linguistic complexity of Greek and the scarcity of domain-specific datasets. Previous efforts in multilingual financial natural language processing (NLP) have exposed considerable performance disparities, yet no dedicated Greek financial benchmarks or Greek-specific financial LLMs have been developed until now. To bridge this gap, we introduce Plutus-ben, the first Greek Financial Evaluation Benchmark, and Plutus-8B, the pioneering Greek Financial LLM, fine-tuned with Greek domain-specific data. Plutus-ben addresses five core financial NLP tasks in Greek: numeric and textual named entity recognition, question answering, abstractive summarization, and topic classification, thereby facilitating systematic and reproducible LLM assessments. To underpin these tasks, we present three novel, high-quality Greek financial datasets, thoroughly annotated by expert native Greek speakers, augmented by two existing resources. Our comprehensive evaluation of 22 LLMs on Plutus-ben reveals that Greek financial NLP remains challenging due to linguistic complexity, domain-specific terminology, and financial reasoning gaps. These findings underscore the limitations of cross-lingual transfer, the necessity for financial expertise in Greek-trained models, and the challenges of adapting financial LLMs to Greek text. We release Plutus-ben, Plutus-8B, and all associated datasets publicly to promote reproducible research and advance Greek financial NLP, fostering broader multilingual inclusivity in finance.
Petrukha Nina M., Petrukha Serhii V. , Tiurmenko Yaroslav M.
et al.
The article surveys the introduction of rational allocation of financial and material resources in the sphere of security and defense as a factor that significantly enhances the resilience of the State against modern challenges. The conditions of dynamic changes in the security environment that necessitate the use of the latest methods for assessing economic feasibility and optimizing strategic planning in the sphere of security and defense capability are considered. Special attention is given to the development of scientifically grounded approaches and econometric models to enhance the resilience and adaptability of the new defense economy. In the study, statistical data from 20 security and defense projects implemented in Ukraine, which encompass the development of new weaponry prototypes and the modernization of existing defense and security systems, were analyzed. The multiple regression method was used to assess the impact of key factors on current costs, as well as the least squares method to build a mathematical model for forecasting expenditure levels and optimizing the material and financial resources involved. It is determined that the largest impact on costs comes from management and operational items, main supply resources, and long-term investment needs. A level model of strategic management has been developed, taking into account innovative, risk, and inflation components, which confirms that high-tech projects with significant innovative potential ensure long-term economic efficiency. Prospects for further research include expanding the database to analyze a larger number of projects, implementing digital cost monitoring tools, and adapting the constructed econometric model to the conditions of other countries and defense systems. A priority direction is the development of integrated methods for assessing risks and effectiveness in the context of global transformational challenges.
Rural Resilience represents the ability of maintaining their core functions when facing internal changes and recovering to original conditions through transformation. Withdrawal from rural homesteads (WRH) is considered as one of critical strategy for rural revitalization of China but its systemic impacts on rural resilience remain underexplored. This study develops a multidimensional resilience evaluation framework encompassing economic, social, cultural, environmental, and governance dimensions through a Delphi-structured expert consultation process with 16 specialists. Considering the complexity of rural socio-ecological systems and the interplay among various dimensions of rural resilience, this paper uses the Fuzzy Decision-Making Trial and Evaluation Laboratory methodology to analyze the causal relationships between 22 resilience indicators. Results reveal economic resilience and social resilience as dominant causal dimensions, with economic diversification promotion and collective land marketization emerging as key drivers. Cultural and environmental dimensions exhibit effect characteristics, demonstrating dependence on economic, social, governance interventions. Notably, villagers’ income improvement and cooperative mechanisms demonstrate high centrality, while indicators related to culture and environment rank as vulnerable nodes. These findings provide policymakers with a prioritized intervention framework, emphasizing the need for economic restructuring coupled with institutional safeguards to balance developmental and conservation objectives in rural spatial reorganization processes.
This study examines the effect of income distribution on economic growth across the E6 countries, namely China, Türkiye, Mexico, Brazil, Russia, and India, from 1988 to 2021, employing the Panel Corrected Standard Errors (PCSE) method. The findings, based on the PCSE estimations and preliminary tests including cross-sectional dependence, unit root tests, and the Westerlund cointegration test, can be summarized as follows: i) Models exhibit cross-sectional dependence (CSD) according to various CD tests. ii) There is a long-term relationship among the variables according to the Westerlund cointegration test. iii) The growth-enhancing effect of absolute redistribution on the economic growth rate has been observed. iv) The growth-promoting effect of absolute redistribution is also confirmed through the use of the relative redistribution variable. v) Based on these findings, the study offers policy recommendations and outlines directions for future research on the topic.
Price volatility in grain markets, especially for maize, has substantial socio-economic impacts, particularly in low-income regions where food security remains a critical concern. Accurate forecasting of grain price volatility is therefore crucial in safeguarding the financial interests of commodity traders, as well as shielding consumers from detrimental effects of inflationary food prices. This study proposes a hybrid Bi-directional Long Short-Term Memory (BLSTM) model, integrated with generalised autoregressive conditional heteroscedasticity (GARCH)-type methods, to forecast white maize futures volatility in South Africa. By comparing the forecasting accuracy of the hybrid BLSTM model against several benchmarks, including standard LSTM and BLSTM models, our results demonstrate notable improvements in prediction accuracy, as shown through heteroscedasticity-adjusted performance metrics. The key contribution of this research is its enhancement of volatility forecasting by combining advanced machine learning with traditional econometric approaches, bridging a gap in predictive accuracy for commodity price dynamics. Additionally, this study supports the United Nations Sustainable Development Goals (SDGs), particularly Zero Hunger and Responsible Consumption and Production, by improving food price stability and risk management in agriculture. This approach exemplifies the evolving role of data science in financial analysis, offering market participants an effective tool to manage price risk and improve food security.Impact Statement This study introduces a novel hybrid forecasting model that integrates GARCH-type econometric techniques with Bi-directional Long Short-Term Memory (BLSTM) neural networks to predict the realised volatility of white maize futures. As white maize is a staple food, accurate volatility forecasting directly contributes to improved food security and price stability. The model significantly outperforms traditional approaches and standard deep learning models across multiple forecast horizons, offering a powerful risk management tool for farmers, traders, and policymakers. By enhancing the accuracy of agricultural price forecasts, this research supports the United Nations Sustainable Development Goals (SDGs), particularly Zero Hunger (SDG 2) and Responsible Consumption and Production (SDG 12), while also demonstrating the value of advanced data science methods in addressing real-world socio-economic challenges.
Dimitrios Koemtzopoulos, Georgia Zournatzidou, Konstantina Ragazou
et al.
Fintech prioritizes the progression of issues related to environmental conservation and the consequences of climate change. This study is among the first investigations exploring the relationship between fintech and sustainable energy. It presents potential financial models that might be developed to assist companies in remaining operational via the use of renewable and clean energy sources. We employ a bibliometric analysis as the statistical methodology to address the study topic. We extract bibliometric data from the Scopus database employing the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) approach, thereafter analyzing the data with the R statistical programming language and the bibliometric applications Biblioshiny and VOSviewer. The results of the research indicate that fintech companies are committed to achieving carbon neutrality and investing in strategies such as environmental, social, and corporate governance (ESG) which may help them reduce their carbon footprint and enhance their eco-efficiency. In contrast to the United Kingdom, which is frequently regarded as the world’s preeminent financial center, Chinese fintech enterprises appear to demonstrate a more fervent dedication to the improvement of their ecological transition. However, the results, ultimately, emphasize the transition of fintech to an alternative paradigm, namely greentech. Greentech is a new fintech-dependent paradigm which will help cryptocurrencies and fintech reduce their environmental impact and promote carbon-neutral financial institutions via investment. Greentech aims to decarbonize the financial industry by investing in renewable resources and clean energy, therefore enhancing the sector’s environmental sustainability.
The prevalence of maximal extractable value (MEV) in the Ethereum ecosystem has led to a characterization of the latter as a dark forest. Studies of MEV have thus far largely been restricted to purely on-chain MEV, i.e., sandwich attacks, cyclic arbitrage, and liquidations. In this work, we shed light on the prevalence of non-atomic arbitrage on decentralized exchanges (DEXes) on the Ethereum blockchain. Importantly, non-atomic arbitrage exploits price differences between DEXes on the Ethereum blockchain as well as exchanges outside the Ethereum blockchain (i.e., centralized exchanges or DEXes on other blockchains). Thus, non-atomic arbitrage is a type of MEV that involves actions on and off the Ethereum blockchain. In our study of non-atomic arbitrage, we uncover that more than a fourth of the volume on Ethereum's biggest five DEXes from the merge until 31 October 2023 can likely be attributed to this type of MEV. We further highlight that only eleven searchers are responsible for more than 80% of the identified non-atomic arbitrage volume sitting at a staggering $132 billion and draw a connection between the centralization of the block construction market and non-atomic arbitrage. Finally, we discuss the security implications of these high-value transactions that account for more than 10% of Ethereum's total block value and outline possible mitigations.
Instrumental Variable (IV) provides a source of treatment randomization that is conditionally independent of the outcomes, responding to the challenges of counterfactual and confounding biases. In finance, IV construction typically relies on pre-designed synthetic IVs, with effectiveness measured by specific algorithms. This classic paradigm cannot be generalized to address broader issues that require more and specific IVs. Therefore, we propose an expertise-driven model (ETE-FinCa) to optimize the source of expertise, instantiate IVs by the expertise concept, and interpret the cause-effect relationship by integrating concept with real economic data. The results show that the feature selection based on causal knowledge graphs improves the classification performance than others, with up to a 11.7% increase in accuracy and a 23.0% increase in F1-score. Furthermore, the high-quality IVs we defined can identify causal relationships between the treatment and outcome variables in the Two-Stage Least Squares Regression model with statistical significance.
Event extraction lies at the cores of investment analysis and asset management in the financial field, and thus has received much attention. The 2019 China conference on knowledge graph and semantic computing (CCKS) challenge sets up a evaluation competition for event entity extraction task oriented to the finance field. In this task, we mainly focus on how to extract the event entity accurately, and recall all the corresponding event entity effectively. In this paper, we propose a novel model, Sequence Enhanced BERT Networks (SEBERTNets for short), which can inherit the advantages of the BERT,and while capturing sequence semantic information. In addition, motivated by recommendation system, we propose Hybrid Sequence Enhanced BERT Networks (HSEBERTNets for short), which uses a multi-channel recall method to recall all the corresponding event entity. The experimental results show that, the F1 score of SEBERTNets is 0.905 in the first stage, and the F1 score of HSEBERTNets is 0.934 in the first stage, which demonstarate the effectiveness of our methods.
We propose a simple methodology to approximate functions with given asymptotic behavior by specifically constructed terms and an unconstrained deep neural network (DNN). The methodology we describe extends to various asymptotic behaviors and multiple dimensions and is easy to implement. In this work we demonstrate it for linear asymptotic behavior in one-dimensional examples. We apply it to function approximation and regression problems where we measure approximation of only function values (``Vanilla Machine Learning''-VML) or also approximation of function and derivative values (``Differential Machine Learning''-DML) on several examples. We see that enforcing given asymptotic behavior leads to better approximation and faster convergence.