|
on Computational Economics |
| By: | Yicai Xing |
| Abstract: | Contemporary societies face a severe crisis of demographic reproduction. Global fertility rates continue to decline precipitously, with East Asian nations exhibiting the most dramatic trends -- China's total fertility rate (TFR) fell to approximately 1.0 in 2023, while South Korea's dropped below 0.72. Simultaneously, the institution of marriage is undergoing structural disintegration: educated women rationally reject unions lacking both emotional fulfillment and economic security, while a growing proportion of men at the lower end of the socioeconomic spectrum experience chronic sexual deprivation, anxiety, and learned helplessness. This paper proposes a computational framework for modeling and evaluating a Stratified Polyamory System (SPS) using techniques from agent-based modeling (ABM), multi-agent reinforcement learning (MARL), and large language model (LLM)-empowered social simulation. The SPS permits individuals to maintain a limited number of legally recognized secondary partners in addition to one primary spouse, combined with socialized child-rearing and inheritance reform. We formalize the A/B/C stratification as heterogeneous agent types in a multi-agent system and model the matching process as a MARL problem amenable to Proximal Policy Optimization (PPO). The mating network is analyzed using graph neural network (GNN) representations. Drawing on evolutionary psychology, behavioral ecology, social stratification theory, computational social science, algorithmic fairness, and institutional economics, we argue that SPS can improve aggregate social welfare in the Pareto sense. Preliminary computational results demonstrate the framework's viability in addressing the dual crisis of female motherhood penalties and male sexlessness, while offering a non-violent mechanism for wealth dispersion analogous to the historical Chinese Grace Decree (Tui'en Ling). |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.20678 |
| By: | Oleksandr Castello (Ca’ Foscari University of Venice); Marco Corazza (Ca’ Foscari University of Venice) |
| Abstract: | The availability of sufficiently large, reliable, and high-quality datasets represents a fundamental prerequisite for quantitative analysis and data-driven decision-making in economics and finance. In practice, however, financial data are often limited, noisy, or subject to restricted access, creating significant empirical constraints for both researchers and practitioners. Recent advances in Generative Machine Learning (GenML) provide promising tools to overcome these limitations by enabling the generation of synthetic data capable of preserving the main statistical features of original data. Despite the rapid diffusion of these techniques, most existing studies focus on replicating stylized facts of financial time series or producing forward-looking simulations, while less attention has been devoted to a systematic assessment of the generative fidelity and generalization capacity of alternative models across different distributional environments. Motivated by this gap, this study provides a comparative evaluation of several Deep Generative Machine Learning (Deep-GenML) families by assessing their ability to reproduce both theoretical statistical distributions and empirical financial and commodity market data. The analysis spans multiple Deep-GenML architectures, distributional settings and market regimes, while also examining model performance under alternative training configurations that reflect varying degrees of data availability. The empirical evidence indicates that deep generative models are capable of accurately reproducing complex distributional features—including heavy tails, asymmetry, and multimodality—across a wide range of scenarios. Overall, the results highlight the potential of deep generative approaches as flexible tools for synthetic data generation and distributional modeling in financial and energy market applications. |
| Keywords: | Deep Generative Machine Learning, Synthetic data generation, GAN, VAE, EBM, Financial and Energy market data |
| JEL: | C45 C46 C58 C63 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ven:wpaper:2026:11 |
| By: | Hui Gong |
| Abstract: | Recent advances in large language models, tool-using agents, and financial machine learning are shifting financial automation from isolated prediction tasks to integrated decision systems that can perceive information, reason over objectives, and generate or execute actions. This paper develops an integrative framework for analysing agentic finance: financial market environments in which autonomous or semi-autonomous AI systems participate in information processing, decision support, monitoring, and execution workflows. The analysis proceeds in three steps. First, the paper proposes a four-layer architecture of financial AI agents covering data perception, reasoning engines, strategy generation, and execution with control. Second, it introduces the Agentic Financial Market Model (AFMM), a stylised agent-based representation linking agent design parameters such as autonomy depth, heterogeneity, execution coupling, infrastructure concentration, and supervisory observability to market-level outcomes including efficiency, liquidity resilience, volatility, and systemic risk. Third, it develops an illustrative empirical application based on event studies of AI-agent capability disclosures and heterogeneous market repricing. The central argument is that the systemic implications of AI in finance depend less on model intelligence alone than on how agent architectures are distributed, coupled, and governed across institutions. In the near term, the most plausible equilibrium is bounded autonomy, in which AI agents operate as supervised co-pilots, monitoring systems, and constrained execution modules embedded within human decision processes. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.13942 |
| By: | Hanyong Cho; Jang Ho Kim |
| Abstract: | This study introduces a benchmark framework for evaluating the financial decision-making capabilities of large language models (LLMs) through portfolio optimization problems with mathematically explicit solutions. Unlike existing financial benchmarks that emphasize language-processing tasks, the proposed framework directly tests optimization-based reasoning in investment contexts. A large set of multiple-choice questions is generated by varying objectives, candidate assets, and investment constraints, with each problem designed to include a unique correct solution and systematically constructed alternatives. Experimental results comparing GPT-4, Gemini 1.5 Pro, and Llama 3.1-70B reveal distinct performance patterns: GPT achieves the highest accuracy in risk-based objectives and remains stable under constraints, Gemini performs well in return-based tasks but struggles under other conditions, and Llama records the lowest overall performance. These findings highlight both the potential and current limitations of LLMs in applying quantitative reasoning to finance, while providing a scalable foundation for developing LLM-based services in portfolio management. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.09301 |
| By: | Takayuki Sakuma |
| Abstract: | We present a differential machine learning method for zero-days-to-expiry (0DTE) options under a stochastic-volatility jump-diffusion model that computes prices and Greeks in a single network evaluation. To handle the ultra-short-maturity regime, we represent the price in Black--Scholes form with a maturity-gated variance correction, and combine supervision on prices and Greeks with a PIDE-residual penalty. To make the jump contribution identifiable, we introduce a separate jump-operator network and train it with a three-stage procedure. In Bates-model simulations, the method improves jump-term approximation relative to one-stage baselines, keeps price errors close to one-stage alternatives while improving Greeks accuracy, produces stable one-day delta hedges, and is substantially faster than a Fourier-based pricing benchmark. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.07600 |
| By: | Pei-Jun Liao; Hung-Shin Lee; Yao-Fei Cheng; Li-Wei Chen; Hung-yi Lee; Hsin-Min Wang |
| Abstract: | Predicting stock prices presents challenges in financial forecasting. While traditional approaches such as ARIMA and RNNs are prevalent, recent developments in Large Language Models (LLMs) offer alternative methodologies. This paper introduces an approach that integrates LLMs with daily financial news for stock price prediction. To address the challenge of processing news data and identifying relevant content, we utilize stock name embeddings within attention mechanisms. Specifically, we encode news articles using a pre-trained LLM and implement three attention-based pooling techniques -- self-attentive, cross-attentive, and position-aware self-attentive pooling -- to filter news based on stock relevance. The filtered news embeddings, combined with historical stock prices, serve as inputs to the prediction model. Unlike prior studies that focus on individual stocks, our method trains a single generalized model applicable across multiple stocks. Experimental results demonstrate a 7.11% reduction in Mean Absolute Error (MAE) compared to the baseline, indicating the utility of stock name embeddings for news filtering and price forecasting within a generalized framework. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.19286 |
| By: | Hanyong Cho; Geumil Bae; Jang Ho Kim |
| Abstract: | This paper investigates how large language models (LLMs) form and express investor risk profiles, a critical component of retail investment advising. We examine three LLMs (GPT, Gemini, and Llama) and assess their responses to a standardized risk questionnaire under varying prompts. In particular, we establish each model's default investment profile by analyzing repeated responses per model. We observe that LLMs are generally longterm investors but exhibit different tendencies in risk tolerance: Gemini has a moderate risk level with highly consistent responses, Llama skews more conservative, and GPT appears moderately aggressive with the greatest variation in answers. Moreover, we find that assigning specific personas such as age, wealth, and investment experience leads each LLM to adjust its risk profile, although the extent of these adjustments differs across the models. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.09303 |
| By: | Haochen Luo; Zhengzhao Lai; Junjie Xu; Yifan Li; Tang Pok Hin; Yuan Zhang; Chen Liu |
| Abstract: | Large Language Models (LLMs) excel at general code generation, yet translating natural-language trading intents into correct option strategies remains challenging. Real-world option design requires reasoning over massive, multi-dimensional option chain data with strict constraints, which often overwhelms direct generation methods. We introduce the Option Query Language (OQL), a domain-specific intermediate representation that abstracts option markets into high-level primitives under grammatical rules, enabling LLMs to function as reliable semantic parsers rather than free-form programmers. OQL queries are then validated and executed deterministically by an engine to instantiate executable strategies. We also present a new dataset for this task and demonstrate that our neuro-symbolic pipeline significantly improves execution accuracy and logical consistency over direct baselines. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.16434 |
| By: | Keonvin Park |
| Abstract: | Portfolio construction traditionally relies on separately estimating expected returns and covariance matrices using historical statistics, often leading to suboptimal allocation under time-varying market conditions. This paper proposes a joint return and risk modeling framework based on deep neural networks that enables end-to-end learning of dynamic expected returns and risk structures from sequential financial data. Using daily data from ten large-cap US equities spanning 2010 to 2024, the proposed model is evaluated across return prediction, risk estimation, and portfolio-level performance. Out-of-sample results during 2020 to 2024 show that the deep forecasting model achieves competitive predictive accuracy (RMSE = 0.0264) with economically meaningful directional accuracy (51.9%). More importantly, the learned representation effectively captures volatility clustering and regime shifts. When integrated into portfolio optimization, the proposed Neural Portfolio strategy achieves an annual return of 36.4% and a Sharpe ratio of 0.91, outperforming equal weight and historical mean-variance benchmarks in terms of risk-adjusted performance. These findings demonstrate that jointly modeling return and covariance dynamics can provide consistent improvements over traditional allocation approaches. The framework offers a scalable and practical alternative for data-driven portfolio construction under nonstationary market conditions. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.19288 |
| By: | Chung-Hoo Poon; James Kwok; Calvin Chow; Jang-Hyeon Choi |
| Abstract: | Anti-money laundering (AML) systems are important for protecting the global economy. However, conventional rule-based methods rely on domain knowledge, leading to suboptimal accuracy and a lack of scalability. Graph neural networks (GNNs) for digraphs (directed graphs) can be applied to transaction graphs and capture suspicious transactions or accounts. However, most spectral GNNs do not naturally support multi-dimensional edge features, lack interpretability due to edge modifications, and have limited scalability owing to their spectral nature. Conversely, most spatial methods may not capture the money flow well. Therefore, in this work, we propose LineMVGNN (Line-Graph-Assisted Multi-View Graph Neural Network), a novel spatial method that considers payment and receipt transactions. Specifically, the LineMVGNN model extends a lightweight MVGNN module, which performs two-way message passing between nodes in a transaction graph. Additionally, LineMVGNN incorporates a line graph view of the original transaction graph to enhance the propagation of transaction information. We conduct experiments on two real-world account-based transaction datasets: the Ethereum phishing transaction network dataset and a financial payment transaction dataset from one of our industry partners. The results show that our proposed method outperforms state-of-the-art methods, reflecting the effectiveness of money laundering detection with line-graph-assisted multi-view graph learning. We also discuss scalability, adversarial robustness, and regulatory considerations of our proposed method. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.23584 |
| By: | Takashi Kameyama; Masahiro Kato; Yasuko Hio; Yasushi Takano; Naoto Minakawa |
| Abstract: | Large language models (LLMs) are trained on enormous amounts of data and encode knowledge in their parameters. We propose a pipeline to elicit causal relationships from LLMs. Specifically, (i) we sample many documents from LLMs on a given topic, (ii) we extract an event list from from each document, (iii) we group events that appear across documents into canonical events, (iv) we construct a binary indicator vector for each document over canonical events, and (v) we estimate candidate causal graphs using causal discovery methods. Our approach does not guarantee real-world causality. Rather, it provides a framework for presenting the set of causal hypotheses that LLMs can plausibly assume, as an inspectable set of variables and candidate graphs. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.04276 |
| By: | Minxuan Hu; Ziheng Chen; Jiayu Yi; Wenxi Sun |
| Abstract: | The deployment of autonomous AI agents in derivatives markets has widened a practical gap between static model calibration and realized hedging outcomes. We introduce two reinforcement learning frameworks, a novel Replication Learning of Option Pricing (RLOP) approach and an adaptive extension of Q-learner in Black-Scholes (QLBS), that prioritize shortfall probability and align learning objectives with downside sensitive hedging. Using listed SPY and XOP options, we evaluate models using realized path delta hedging outcome distributions, shortfall probability, and tail risk measures such as Expected Shortfall. Empirically, RLOP reduces shortfall frequency in most slices and shows the clearest tail-risk improvements in stress, while implied volatility fit often favors parametric models yet poorly predicts after-cost hedging performance. This friction-aware RL framework supports a practical approach to autonomous derivatives risk management as AI-augmented trading systems scale. |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.06587 |
| By: | Mohammad Al Ridhawi; Mahtab Haj Ali; Hussein Al Osman |
| Abstract: | Stock markets exhibit regime-dependent behavior where prediction models optimized for stable conditions often fail during volatile periods. Existing approaches typically treat all market states uniformly or require manual regime labeling, which is expensive and quickly becomes stale as market dynamics evolve. This paper introduces an adaptive prediction framework that adaptively identifies deviations from normal market conditions and routes data through specialized prediction pathways. The architecture consists of three components: (1) an autoencoder trained on normal market conditions that identifies anomalous regimes through reconstruction error, (2) dual node transformer networks specialized for stable and event-driven market conditions respectively, and (3) a Soft Actor-Critic reinforcement learning controller that adaptively tunes the regime detection threshold and pathway blending weights based on prediction performance feedback. The reinforcement learning component enables the system to learn adaptive regime boundaries, defining anomalies as market states where standard prediction approaches fail. Experiments on 20 S&P 500 stocks spanning 1982 to 2025 demonstrate that the proposed framework achieves 0.68% MAPE for one-day predictions without the reinforcement controller and 0.59% MAPE with the full adaptive system, compared to 0.80% for the baseline integrated node transformer. Directional accuracy reaches 72% with the complete framework. The system maintains robust performance during high-volatility periods, with MAPE below 0.85% when baseline models exceed 1.5%. Ablation studies confirm that each component contributes meaningfully: autoencoder routing accounts for 36% relative MAPE degradation upon removal, followed by the SAC controller at 15% and the dual-path architecture at 7%. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.19136 |
| By: | Stephan Ludwig; Peter J. Danaher; Xiaohao Yang |
| Abstract: | The growing use of unstructured text in business research makes topic modeling a central tool for constructing explanatory variables from reviews, social media, and open-ended survey responses, yet existing approaches function poorly as measurement instruments. Prior work shows that textual content predicts outcomes such as sales, satisfaction, and firm performance, but probabilistic models often generate conceptually diffuse topics, neural topic models are difficult to interpret in theory-driven settings, and large language model approaches lack standardization, stability, and alignment with document-level representations. We introduce LX Topic, a neural topic method that conceptualizes topics as latent linguistic constructs and produces calibrated document-level topic proportions for empirical analysis. LX Topic builds on FASTopic to ensure strong document representativeness and integrates large language model refinement at the topic-word level using alignment and confidence-weighting mechanisms that enhance semantic coherence without distorting document-topic distributions. Evaluations on large-scale Amazon and Yelp review datasets demonstrate that LX Topic achieves the highest overall topic quality relative to leading models while preserving clustering and classification performance. By unifying topic discovery, refinement, and standardized output in a web-based system, LX Topic establishes topic modeling as a reproducible, interpretable, and measurement-oriented instrument for marketing research and practice. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.03623 |
| By: | Pranjal Rawat |
| Abstract: | This survey (re)introduces reinforcement learning methods to economists. The curse of dimensionality limits how far exact dynamic programming can be effectively applied, forcing us to rely on suitably "small" problems or our ability to convert "big" problems into smaller ones. While this reduction has been sufficient for many classical applications, a growing class of economic models resists such reduction. Reinforcement learning algorithms offer a natural, sample-based extension of dynamic programming, extending tractability to problems with high-dimensional states, continuous actions, and strategic interactions. I review the theory connecting classical planning to modern learning algorithms and demonstrate their mechanics through simulated examples in pricing, inventory control, strategic games, and preference elicitation. I also examine the practical vulnerabilities of these algorithms, noting their brittleness, sample inefficiency, sensitivity to hyperparameters, and the absence of global convergence guarantees outside of tabular settings. The successes of reinforcement learning remain strictly bounded by these constraints, as well as a reliance on accurate simulators. When guided by economic structure, reinforcement learning provides a remarkably flexible framework. It stands as an imperfect, but promising, addition to the computational economist's toolkit. A companion survey (Rust and Rawat, 2026b) covers the inverse problem of inferring preferences from observed behavior. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.08956 |
| By: | Yogesh Agrawal (University of Central Florida); Aniruddha Dutta (University of Central Florida); Md Mahadi Hasan (University of Central Florida); Santu Karmaker (University of Central Florida); Aritra Dutta (University of Central Florida) |
| Abstract: | Real-world financial decision-making is a challenging problem that requires reasoning over heterogeneous signals, including company fundamentals derived from regulatory filings and trading signals computed from price dynamics. Recently, with the advancement of Large Language Models (LLMs), financial analysts have begun to use them for financial decision-making tasks. However, existing financial question answering benchmarks for testing these models primarily focus on company balance sheet data and rarely evaluate reasoning over how company stocks trade in the market or their interactions with fundamentals. To take advantage of the strengths of both approaches, we introduce FinTradeBench, a benchmark for evaluating financial reasoning that integrates company fundamentals and trading signals. FinTradeBench contains 1, 400 questions grounded in NASDAQ-100 companies over a ten-year historical window. The benchmark is organized into three reasoning categories: fundamentals-focused, trading-signal-focused, and hybrid questions requiring cross-signal reasoning. To ensure reliability at scale, we adopt a calibration-then-scaling framework that combines expert seed questions, multi-model response generation, intra-model self-filtering, numerical auditing, and human-LLM judge alignment. We evaluate 14 LLMs under zero-shot prompting and retrieval-augmented settings and witness a clear performance gap. Retrieval substantially improves reasoning over textual fundamentals, but provides limited benefit for trading-signal reasoning. These findings highlight fundamental challenges in the numerical and time-series reasoning for current LLMs and motivate future research in financial intelligence. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.19225 |
| By: | Rahul D Ray |
| Abstract: | Deep learning models in quantitative finance often operate as black boxes, lacking interpretability and failing to incorporate fundamental economic principles such as no-arbitrage constraints. This paper introduces ARTEMIS (Arbitrage-free Representation Through Economic Models and Interpretable Symbolics), a novel neuro-symbolic framework combining a continuous-time Laplace Neural Operator encoder, a neural stochastic differential equation regularised by physics-informed losses, and a differentiable symbolic bottleneck that distils interpretable trading rules. The model enforces economic plausibility via two novel regularisation terms: a Feynman-Kac PDE residual penalising local no-arbitrage violations, and a market price of risk penalty bounding the instantaneous Sharpe ratio. We evaluate ARTEMIS against six strong baselines on four datasets: Jane Street, Optiver, Time-IMM, and DSLOB (a synthetic crash regime). Results demonstrate ARTEMIS achieves state-of-the-art directional accuracy, outperforming all baselines on DSLOB (64.96%) and Time-IMM (96.0%). A comprehensive ablation study confirms each component's contribution: removing the PDE loss reduces directional accuracy from 64.89% to 50.32%. Underperformance on Optiver is attributed to its long sequence length and volatility-focused target. By providing interpretable, economically grounded predictions, ARTEMIS bridges the gap between deep learning's power and the transparency demanded in quantitative finance. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.18107 |
| By: | Wei Wei; Jin Zheng; Zining Wang |
| Abstract: | Recent advances in artificial intelligence (AI) and natural language processing (NLP) have enabled tools to support systematic literature reviews (SLRs), yet existing frameworks often produce outputs that are efficient but contextually limited, requiring substantial expert oversight.The framework employs a human-in-the-loop process to define sub-SLR tasks, evaluate models, and ensure methodological rigor, while leveraging structured knowledge sources and retrieval-augmented generation (RAG) to enhance factual grounding and transparency. LR-Robot enables multidimensional categorization of research, maps relationships among papers, identifies high-impact works, and supports historical, fine-grained analyses of topic evolution. We demonstrate the framework using an option pricing case study, enabling comprehensive literature analysis. Empirical results reveal the current capabilities of AI in understanding and synthesizing literature, uncover emerging trends, reveal topic connections, and highlight core research directions. By accelerating labor-intensive review stages while preserving interpretive accuracy, LR-Robot provides a practical, customizable, and high-quality approach for AI-assisted SLRs. Key contributions: (1) a novel framework combining AI and expert supervision for contextually informed SLRs, (2) support for multidimensional categorization, relationship mapping, and fine-grained topic evolution analysis, and (3) empirical demonstration of AI-driven literature synthesis in the field of option pricing. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.17723 |
| By: | Takanobu Mizuta; Isao Yagi |
| Abstract: | Some investors say increasing investors with the same strategy decreasing their profits per an investor. On the other hand, some investors using technical analysis used to use same strategy and parameters with other investors, and say that it is better. Those argues are conflicted each other because one argues using with same strategy decreases profits but another argues it increase profits. However, those arguments have not been investigated yet. In this study, the agent-based artificial financial market model(ABAFMM) was built by adding "additional agents"(AAs) that includes additional fundamental agents (AFAs) and additional technical agents (ATAs) to the prior model. The AFAs(ATAs) trade obeying simple fundamental(technical) strategy having only the one parameter. We investigated earnings of AAs when AAs increased. We found that in the case with increasing AFAs, market prices are made stable that leads to decrease their profits. In the case with increasing ATAs, market prices are made unstable that leads to gain their profits more. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.03671 |
| By: | Yicai Xing |
| Abstract: | As large language models (LLMs) and vision-language-action models (VLAs) become widely deployed, the tokens consumed by AI inference are evolving into a new type of commodity. This paper systematically analyzes the commodity attributes of tokens, arguing for their transition from intelligent service outputs to compute infrastructure raw materials, and draws comparisons with established commodities such as electricity, carbon emission allowances, and bandwidth. Building on the historical experience of electricity futures markets and the theory of commodity financialization, we propose a complete design for standardized token futures contracts, including the definition of a Standard Inference Token (SIT), contract specifications, settlement mechanisms, margin systems, and market-maker regimes. By constructing a mean-reverting jump-diffusion stochastic process model and conducting Monte Carlo simulations, we evaluate the hedging efficiency of the proposed futures contracts for application-layer enterprises. Simulation results show that, under an application-layer demand explosion scenario, token futures can reduce enterprise compute cost volatility by 62%-78%. We also explore the feasibility of GPU compute futures and discuss the regulatory framework for token futures markets, providing a theoretical foundation and practical roadmap for the financialization of compute resources. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.21690 |
| By: | Federico Vittorio Cortesi; Giuseppe Iannone; Giulia Crippa; Tomaso Poggio; Pierfrancesco Beneventano |
| Abstract: | Neural networks applied to financial time series operate in a regime of underspecification, where model predictors achieve indistinguishable out-of-sample error. Using large-scale volatility forecasting for S$\&$P 500 stocks, we show that different model-training-pipeline pairs with identical test loss learn qualitatively different functions. Across architectures, predictive accuracy remains unchanged, yet optimizer choice reshapes non-linear response profiles and temporal dependence differently. These divergences have material consequences for decisions: volatility-ranked portfolios trace a near-vertical Sharpe-turnover frontier, with nearly $3\times$ turnover dispersion at comparable Sharpe ratios. We conclude that in underspecified settings, optimization acts as a consequential source of inductive bias, thus model evaluation should extend beyond scalar loss to encompass functional and decision-level implications. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.02620 |
| By: | Kevin Mott |
| Abstract: | This paper introduces a no-arbitrage, Monte Carlo-free approach to pricing path-dependent interest rate derivatives. The Heath-Jarrow-Morton model gives arbitrage-free contingent claims prices but is infinite-dimensional, making traditional numerical methods computationally prohibitive. To make the problem computationally tractable, I cast the stochastic pricing problem as a deterministic partial differential equation (PDE). Finance-Informed Neural Networks (FINNs) solve this PDE directly by minimizing violations of the differential equation and boundary condition, with automatic differentiation efficiently computing the exact derivatives needed to evaluate PDE terms. FINNs achieve pricing accuracy within 0.04 to 0.07 cents per dollar of contract value compared to Monte Carlo benchmarks. Once trained, FINNs price caplets in a few microseconds regardless of dimension, delivering speedups ranging from 300, 000 to 4.5 million times faster than Monte Carlo simulation as the state space discretization of the forward curve grows from 10 to 150 nodes. The major Greeks-theta and curve deltas-come for free, computed automatically during PDE evaluation at zero marginal cost, whereas Monte Carlo requires complete re-simulation for each sensitivity. The framework generalizes naturally beyond caplets to other path-dependent derivatives-caps, swaptions, callable bonds-requiring only boundary condition modifications while retaining the same core PDE structure. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.12375 |
| By: | Chang Chen; Duy-Minh Dang |
| Abstract: | We develop a neural-network framework for multi-period risk--reward stochastic control problems with constrained two-step feedback policies that may be discontinuous in the state. We allow a broad class of objectives built on a finite-dimensional performance vector, including terminal and path-dependent statistics, with risk functionals admitting auxiliary-variable optimization representations (e.g.\ Conditional Value-at-Risk and buffered probability of exceedance) and optional moment dependence. Our approach parametrizes the two-step policy using two coupled feedforward networks with constraint-enforcing output layers, reducing the constrained control problem to unconstrained training over network parameters. Under mild regularity conditions, we prove that the empirical optimum of the NN-parametrized objective converges in probability to the true optimal value as network capacity and training sample size increase. The proof is modular, separating policy approximation, propagation through the controlled recursion, and preservation under the scalarized risk--reward objective. Numerical experiments confirm the predicted convergence-in-probability behavior, show close agreement between learned and reference control heat maps, and demonstrate out-of-sample robustness on a large independent scenario set. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.06563 |
| By: | Khem Raj Bhatt; Krishna Sharma |
| Abstract: | We study the deployment performance of machine learning based enforcement systems used in cryptocurrency anti money laundering (AML). Using forward looking and rolling evaluations on Bitcoin transaction data, we show that strong static classification metrics substantially overstate real world regulatory effectiveness. Temporal nonstationarity induces pronounced instability in cost sensitive enforcement thresholds, generating large and persistent excess regulatory losses relative to dynamically optimal benchmarks. The core failure arises from miscalibration of decision rules rather than from declining predictive accuracy per se. These findings underscore the fragility of fixed AML enforcement policies in evolving digital asset markets and motivate loss-based evaluation frameworks for regulatory oversight. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.04328 |
| By: | Shogo Fukui |
| Abstract: | Non-survey methods have been developed and applied for estimating regional input-output tables. However, there is an ongoing debate about the assumptions necessary for these methods and their accuracy. To address these issues, this study presents a deep learning method for estimating regional input-output tables. First, the quantitative economic data for regions is augmented by linear combinations. Then, deep learning is performed on each item in the input-output table, treating these items as target variables. Finally, regional input-output tables are estimated through matrix balancing to the predicted values from the trained model. The estimation accuracy of this method is verified using the 2015 input-output table for Japan as a benchmark. Compared to matrix balancing under the ideal assumption of known row and column sums, our method generally demonstrates higher estimation accuracy. Thus, this method is anticipated to provide a foundation for deriving more precise estimates of regional input-output tables. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.13823 |
| By: | Reiner Martin (National Bank of Slovakia); Piroska Nagy Mohacsi (London School of Economics and Political Science); Tatiana Evdokimova (Joint Vienna Institute); Jan Klacso (National Bank of Slovakia); Olga Ponomarenko (Caplight) |
| Abstract: | Central bank communication on financial stability has been less studied than on monetary policy. Our paper aims to contribute to the growing literature in this area. Our focus is the region of Central Europe, where financial sectors are intertwined through close cross-border ownership, and about half of the countries are members of the euro area. Using large language models (LLMs) combined with country-specific contextual analysis, we study executive summaries of Financial Stability Reports (FSRs) published since the early 2000s by seven Central, Eastern, and Southeastern European (CESEE) central banks, as well as by Austria and the European Central Bank (ECB). We construct a novel financial stability sentiment index and document that central bank communication is strongly risk-focused, most notably in the case of the ECB. In addition, prior to the Global Financial Crisis, the Austrian central bank was much less concerned than other central banks in the region although Austria plays a pivotal role in the financial system in the region. Our analysis of the link between financial stability sentiment communication and macroprudential policy action highlights that many central banks actively use and communicate about borrower-based measures, while most countries activated non-zero counter-cyclical capital buffers belatedly or not at all. Finally, comparing central banks’ communication on financial stability and monetary policy, we find that euro area national central banks and the ECB’s FSR communicated about the rising risks of post-Covid inflation in a timely manner, ahead of the ECB’s monetary policy communication. |
| JEL: | C55 E58 E61 H12 D83 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:svk:wpaper:1139 |
| By: | Anna Baiardi; Paul S. Clarke; Andrea A. Naghi; Annalivia Polselli |
| Abstract: | Panel data methods are widely used in empirical analysis to address unobserved heterogeneity, but causal inference remains challenging when treatments are endogenous and confounding variables high-dimensional and potentially nonlinear. Standard instrumental variables (IV) estimators, such as two-stage least squares (2SLS), become unreliable when instrument validity requires flexibly conditioning on many covariates with potentially non-linear effects. This paper develops a Double Machine Learning estimator for static panel models with endogenous treatments (panel IV DML), and introduces weak-identification diagnostics for it. We revisit three influential migration studies that use shift-share instruments. In these settings, instrument validity depends on a rich covariate adjustment. In one application, panel IV DML strengthens the predictive power of the instrument and broadly confirms 2SLS results. In the other cases, flexible adjustment makes the instruments weak, leading to substantially more cautious causal inference than conventional 2SLS. Monte Carlo evidence supports these findings, showing that panel IV DML improves estimation accuracy under strong instruments and delivers more reliable inference under weak identification. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.20464 |
| By: | Anders G Fr{\o}seth |
| Abstract: | We develop spectral portfolio theory by establishing a direct identification: neural network weight matrices trained on stochastic processes are portfolio allocation matrices, and their spectral structure encodes factor decompositions and wealth concentration patterns. The three forces governing stochastic gradient descent (SGD) -- gradient signal, dimensional regularisation, and eigenvalue repulsion -- translate directly into portfolio dynamics: smart money, survival constraint, and endogenous diversification. The spectral properties of SGD weight matrices transition from Marchenko-Pastur statistics (additive regime, short horizon) to inverse-Wishart via the free log-normal (multiplicative regime, long horizon), mirroring the transition from daily returns to long-run wealth compounding. We unify the cross-sectional wealth dynamics of Bouchaud and Mezard (2000), the within-portfolio dynamics of Olsen et al. (2025), and the scalar Fokker-Planck framework via a common spectral foundation. A central result is the Spectral Invariance Theorem: any isotropic perturbation to the portfolio objective preserves the singular-value distribution up to scale and shift, while anisotropic perturbations produce spectral distortion proportional to their cross-asset variance. We develop applications to portfolio design, wealth inequality measurement, tax policy, and neural network diagnostics. In the tax context, the invariance result recovers and generalises the neutrality conditions of Fr{\o}seth (2026). |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.09006 |
| By: | Enoch Hyunwook Kang |
| Abstract: | AI agents are increasingly deployed in interactive economic environments characterized by repeated AI-AI interactions. Despite AI agents' advanced capabilities, empirical studies reveal that such interactions often fail to stably induce a strategic equilibrium, such as a Nash equilibrium. Post-training methods have been proposed to induce a strategic equilibrium; however, it remains impractical to uniformly apply an alignment method across diverse, independently developed AI models in strategic settings. In this paper, we provide theoretical and empirical evidence that off-the-shelf reasoning AI agents can achieve Nash-like play zero-shot, without explicit post-training. Specifically, we prove that `reasonably reasoning' agents, i.e., agents capable of forming beliefs about others' strategies from previous observation and learning to best respond to these beliefs, eventually behave along almost every realized play path in a way that is weakly close to a Nash equilibrium of the continuation game. In addition, we relax the common-knowledge payoff assumption by allowing stage payoffs to be unknown and by having each agent observe only its own privately realized stochastic payoffs, and we show that we can still achieve the same on-path Nash convergence guarantee. We then empirically validate the proposed theories by simulating five game scenarios, ranging from a repeated prisoner's dilemma game to stylized repeated marketing promotion games. Our findings suggest that AI agents naturally exhibit such reasoning patterns and therefore attain stable equilibrium behaviors intrinsically, obviating the need for universal alignment procedures in many real-world strategic interactions. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.18563 |
| By: | Grzeskiewicz, M. |
| Abstract: | State dependence is empirically important in repeat-purchase demand and can materially change welfare conclusions from price variation. To this end, we introduce a flexible neural demand system for continuous budget allocation that allows current choices to depend on a low-dimensional summary of purchase history. In Dominick’s scanner data on analgesics, augmenting demand with a habit state reduces out-of-sample prediction error by about 33% relative to standard share systems, and a shuffled-history placebo eliminates the gain, indicating that the improvement reflects meaningful dynamics rather than additional covariates. State dependence also changes economic conclusions: conditioning on the habit state col-lapses the apparent aspirin–ibuprofen cross-price effect toward zero while preserving robust acetaminophen–ibuprofen substitution. These differences translate into welfare: for a 10% ibuprofen price increase, the habit specification implies compensating-variation losses about 15–16% larger than a static model. We also provide simulation evidence with known ground truth and report diagnostics of near-integrability to support welfare calculations. The code is available at https://github.com/martagrz/neural_deman d_habit. |
| Keywords: | Neural Demand Systems, Habit Formation, Machine Learning in Econometrics, Scanner Data, Welfare Measurement |
| JEL: | C45 C14 C51 D12 D61 |
| Date: | 2026–03–02 |
| URL: | https://d.repec.org/n?u=RePEc:cam:camdae:2613 |
| By: | Mohammad Mosaffa; Omid Rafieian |
| Abstract: | Firms collect vast amounts of behavioral and geographical data on individuals. While behavioral data captures an individual's digital footprint, geographical data reflects their physical footprint. Given the significant privacy risks associated with combining these data sources, it is crucial to understand their respective value and whether they act as complements or substitutes in achieving firms' business objectives. In this paper, we combine economic theory, machine learning, and causal inference to quantify the value of geographical data, the extent to which behavioral data can substitute for it, and the mechanisms through which it benefits firms. Using data from a leading in-app advertising platform in a large Asian country, we document that geographical data is most valuable in the early cold-start stage, when behavioral histories are limited. In this stage, geographical data complements behavioral data, improving targeting performance by almost 20%. As users accumulate richer behavioral histories, however, the role of geographical data shifts: it becomes largely substitutable, as behavioral data alone captures the relevant heterogeneity. These results highlight a central privacy-utility trade-off in ad personalization and inform managerial decisions about when location tracking creates value. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.12374 |
| By: | Ziqin Gong; Ning Li; Huaikang Zhou |
| Abstract: | Artificial intelligence matches or exceeds human performance on tasks with verifiable answers, from protein folding to Olympiad mathematics. Yet the capacity that most governs scientific advance is not reasoning but taste: the ability to judge which untested ideas deserve pursuit, exercised daily by editors and funders but never successfully articulated, taught, or automated. Here we show that fine-tuning language models on journal publication decisions recovers evaluative judgment inaccessible to both frontier models and human expertise. Using a held-out benchmark of research pitches in management spanning four quality tiers, we find that eleven frontier models, spanning major proprietary and open architectures, barely exceed chance, averaging 31% accuracy. Panels of journal editors and editorial board members reach 42% by majority vote. Fine-tuned models trained on years of publication records each surpass every frontier model and expert panel, with the best single model achieving 59%. These models exhibit calibrated confidence, reaching 100% accuracy on their highest-confidence predictions, and transfer this evaluative signal to untrained pairwise comparisons and one-sentence summaries. The mechanism generalizes: models trained on economics publication records achieve 70% accuracy. Scientific taste was not missing from AI's reach; it was deposited in the institutional record, waiting to be extracted. These results provide a scalable mechanism to triage the expanding volume of scientific production across disciplines where quality resists formal verification. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.16659 |
| By: | Daria Dzyabura (New Economic School); Renana Peres (Hebrew University of Jerusalem); Irina Linevich (MIT Sloan School of Management) |
| Abstract: | Color is an important component in brand visual communication. Firms select brand colors to align with the brand's strategic positioning goals. Despite their importance, brand color decisions are often driven by intuition and trial and error. We introduce BRACE (BRand Attribute and Color Engine), a predictive model and genetic-algorithm based optimization framework, that generates color palettes that reflect combinations of brand characteristics. Using theory on color combinations and color harmonies, the model avoids contradictions across characteristics while maintaining visual harmony. For example, if a brand seeks to be perceived as Friendly and Glamorous, or highly Outdoorsy but not Young, we recommend aesthetically appealing color palettes that best capture these attribute combinations. We validate the algorithm through a series of experiments. We also find that real ads recolored with recommended palettes are rated significantly higher on the intended brand characteristics. We further use topic modeling to provide interpretable insights into the relationships between characteristics and colors, and how these relationships vary across product categories. This paper is a major step towards data-driven brand visual communication that can better align creative choices with communication goals. |
| Keywords: | Image analytics, branding, color, machine learning, genetic algorithm, topic modeling, brand personality, BRACE. JEL Classifications: M31, M37, C45, C55, D12 |
| Date: | 2025–12 |
| URL: | https://d.repec.org/n?u=RePEc:abo:neswpt:w0292 |
| By: | Mahesh Kumar; Bhaskarjit Sarmah; Stefano Pasquali |
| Abstract: | As organizations increasingly integrate AI-powered question-answering systems into financial information systems for compliance, risk assessment, and decision support, ensuring the factual accuracy of AI-generated outputs becomes a critical engineering challenge. Current Knowledge Graph (KG)-augmented QA systems lack systematic mechanisms to detect hallucinations - factually incorrect outputs that undermine reliability and user trust. We introduce FinBench-QA-Hallucination, a benchmark for evaluating hallucination detection methods in KG-augmented financial QA over SEC 10-K filings. The dataset contains 755 annotated examples from 300 pages, each labeled for groundedness using a conservative evidence-linkage protocol requiring support from both textual chunks and extracted relational triplets. We evaluate six detection approaches - LLM judges, fine-tuned classifiers, Natural Language Inference (NLI) models, span detectors, and embedding-based methods under two conditions: with and without KG triplets. Results show that LLM-based judges and embedding approaches achieve the highest performance (F1: 0.82-0.86) under clean conditions. However, most methods degrade significantly when noisy triplets are introduced, with Matthews Correlation Coefficient (MCC) dropping 44-84 percent, while embedding methods remain relatively robust with only 9 percent degradation. Statistical tests (Cochran's Q and McNemar) confirm significant performance differences (p |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.20252 |
| By: | Adjemian, Stéphane; Juillard, Michel; Karamé, Fréderic; Mutschler, Willi; Pfeifer, Johannes; Ratto, Marco; Rion, Normann; Villemot, Sébastien |
| Abstract: | Dynare is a software platform for handling a wide class of economic models, in particular dynamic stochastic general equilibrium (DSGE), overlapping generations (OLG) models, heterogeneous agent models (HA) and semi-structural macroeconomic models. The models solved by Dynare include those relying on the rational expectations hypothesis, wherein agents form their expectations about the future in a way consistent with the model. But Dynare is also able to handle models where expectations are formed differently: on one extreme, models where agents perfectly anticipate the future; on the other extreme, models where agents have limited rationality or imperfect knowledge of the state of the economy and, hence, form their expectations through a learning process. Dynare offers a user-friendly and intuitive way of describing these models. It is able to perform simulations of the model given a calibration of the model parameters and is also able to estimate these parameters given a dataset. Dynare is a free software, which means that it can be downloaded free of charge, that its source code is freely available, and that it can be used for both non-profit and for-profit purposes. |
| Keywords: | Dynare; Numerical methods; Perturbation; Rational expectations |
| JEL: | C5 C6 C8 |
| Date: | 2026–03–19 |
| URL: | https://d.repec.org/n?u=RePEc:cpm:dynare:087 |
| By: | Martin Jaraiz |
| Abstract: | How much macroeconomic information is contained in a single input-output table? We feed FIGARO 64-sector symmetric tables into DEPLOYERS, a Darwinian agent-based simulator, producing genuine out-of-sample GDP forecasts. For each year, the model reads one FIGARO table for year N, self-organizes an artificial economy through evolutionary selection, then runs 12 months of autonomous free-market dynamics whose emergent growth rate predicts year N+1. The I-O table is the only input: no time series, no estimated parameters, no expectations formation, no external forecasts. We present five results. First, a 9-year Austrian panel (2010-2018) using 12-seed ensembles produces MAE of 1.22 pp overall; for five non-crisis years, MAE falls to 0.42 pp -- comparable to the best professional forecaster (WIFO: 0.48 pp). Second, cross-country portability is demonstrated across multiple FIGARO countries with zero parameter changes. Third, a German 9-year panel reveals systematic +3.7 pp positive bias from export dependency -- an informative negative result. Fourth, a COVID-19 simulation demonstrates the I-O structure as a shock propagation mechanism: a 19-month timeline produces Year 1 GDP -4.62% vs empirical -6.6%. Fifth, emergent firm size distributions match European Commission data without micro-target calibration. These results establish the I-O table as serving a dual purpose: structural baseline engine and dynamic shock propagation mechanism. Since FIGARO covers 46 countries, the approach is immediately portable without retuning parameters. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.12412 |
| By: | Abraham Itzhak Weinberg |
| Abstract: | We present a hybrid classical-quantum framework for portfolio construction and rebalancing. Asset selection is performed using Ledoit-Wolf shrinkage covariance estimation combined with hierarchical correlation clustering to extract n = 10 decorrelated stocks from the S&P 500 universe without survivorship bias. Portfolio weights are optimised via an entropy-regularised Genetic Algorithm (GA) accelerated on GPU, alongside closed-form minimum-variance and equal-weight benchmarks. Our primary contribution is the formulation of the portfolio rebalancing schedule as a Quadratic Unconstrained Binary Optimisation (QUBO) problem. The resulting combinatorial optimisation task is solved using the Quantum Approximate Optimisation Algorithm (QAOA) within a walk-forward framework designed to eliminate lookahead bias. This approach recasts dynamic rebalancing as a structured binary scheduling problem amenable to variational quantum methods. Backtests on S&P 500 data (training: 2010-2024; out-of-sample test: 2025, n = 249 trading days) show that the GA + QAOA strategy attains a Sharpe ratio of 0.588 and total return of 10.1%, modestly outperforming the strongest classical baseline (GA with 10-day periodic rebalancing, Sharpe 0.575) while executing 8 rebalances versus 24, corresponding to a 44.5% reduction in transaction costs. Multi-restart QAOA (4096 measurement shots per run) exhibits concentrated probability mass on high-quality schedules, indicating stable convergence of the variational procedure. These findings suggest that hybrid classical-quantum architectures can reduce turnover in portfolio rebalancing while preserving competitive risk-adjusted performance, providing a structured testbed for near-term quantum optimisation in financial applications. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.16904 |