|
on Computational Economics |
| By: | Tyler H. McCormick |
| Abstract: | Modern Machine Learning (ML) and Artificial Intelligence (AI) models, especially large language models (LLMs), are increasingly used to generate scientific hypotheses and mechanistic explanations from observational data. This position paper argues that in the high-dimensional proxy regimes where modern ML excels, mechanistic learning is generically underdetermined: many incompatible mechanisms induce essentially the same observational relationships on the support of the data, so predictive success and coherent explanations are insufficient evidence of mechanism discovery. This underdetermination becomes uniquely hazardous with large language models (LLMs), which tend to collapse large equivalence classes of explanations into a single fluent narrative. This paper proposes concrete standards for ``mechanistic ML, '' and argues these norms are necessary if LLM-centered workflows are to support science rather than merely simulate it. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.02632 |
| By: | Yuqi Li; Siyuan Liu; Bingjun Liu |
| Abstract: | While deep learning has excelled in various domains, its application to sequential decision-making in finance remains challenging due to the low Signal-to-Noise Ratio (SNR) and non-stationarity of financial data. Leveraging the reasoning capabilities of Large Language Models (LLMs), we propose \textbf{PandaAI}, a closed-loop neuro-symbolic LLM agent with market regime modeling and constrained alpha generation, which bridges general LLM reasoning with financial rigor and suppresses the financial toxicity of LLM-generated outputs. To bridge the gap between general linguistic capability and financial rigor, we fine-tune a domain-specific LLM. Furthermore, we integrate this LLM into a modular architecture and form a closed-loop system. Unlike traditional models that optimize isolated prediction metrics, \textbf{PandaAI} is designed as a neuro-symbolic agent that navigates the complex, real-world financial environment with explicit risk awareness. Extensive experiments on CSI 300 stock data show that \textbf{PandaAI} achieves a $18.2\%$ higher Rank IC and $25.7\%$ lower maximum drawdown than state-of-the-art time-series models. Our constrained LLM generation and dual-channel adaptation method provide a general paradigm for LLM deployment in high-stakes sequential decision-making scenarios. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.06823 |
| By: | Damian Lebied\'z; Robert \'Slepaczuk |
| Abstract: | This study aims to determine whether the application of Deep Reinforcement Learning (DRL) as a specialized execution overlay can enhance pair trading in highly volatile cryptocurrency markets. Although classical implementations of the strategy have proven successful in traditional equities, they frequently exhibit rigidity and suffer from severe divergence risks when applied to high-variance environments. To address this need, this research introduces novel concepts. To construct a robust system, we developed a hierarchical "Filter-then-Rank" pair selection methodology and a proprietary "Fixed Risk, Adaptive Mean" execution model. The system employs a Proximal Policy Optimization (PPO) agent with a Long Short-Term Memory (LSTM) layer to govern execution decisions within strict deterministic risk management boundaries. Evaluated on 1-hour interval data from the Binance USD-M Futures market, the optimized RL policy achieved an out-of-sample performance that substantially outperformed the heuristic baseline. A stationary circular block bootstrap robustness check confirms that the agent's risk-adjusted outperformance is statistically significant at the 10 percent level. Although falling marginally short of the stricter 5 percent threshold, this result highlights the extreme idiosyncratic variance characteristic of digital assets. Ultimately, this thesis contributes to the quantitative finance literature by introducing a hybrid architecture that combines statistical arbitrage with DRL execution policies. Furthermore, it delivers a novel framework for safe reinforcement learning via deterministic shielding, proving that anchoring a neural policy to statistically robust boundaries successfully mitigates severe divergence risks. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.04574 |
| By: | Ferrara, Andreas (University of Pittsburgh, Department of Economics, and NBER) |
| Abstract: | Large language models (LLMs) are lowering the entry barriers to working with exciting data sources that used to require strong data science skills, such as handwritten ledgers, text, images, or sound recordings. This guide provides an introduction for researchers who are new to LLMs. It sets out a step-by-step workflow for turning a research idea into working code and data, and describes the four main ways of interacting with an LLM: the chat window, editor-integrated assis tants, agentic coding tools, and the API. It then works through the decisions a practitioner meets in sequence, beginning with whether an LLM is the right tool and whether the data are allowed to be sent to one, then how to select models, write prompts, manage context limits, and control costs, and finally how to validate, reproduce, document, and correct LLM-generated measures in regression settings. A review of recent research shows how these tools already extract, link, har monize, and classify historical data at scale. Four worked examples with replication files illustrate the use of LLMs. They classify emotions in paintings, link census records without names, measure newspaper salience and sentiment around the 1882 Chinese Exclusion Act, and score the emotional delivery of Franklin D. Roosevelt's wartime speeches. The guide also condenses the workflow, the best-practice recommendations, and the preparation of replication packages into summary tables and checklists to aid applied economists. |
| Keywords: | Large Language Models, Artificial Intelligence, Economic History, Practitioner's Guide JEL Classification: C8, N0, C55 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:cge:wacage:810 |
| By: | Helyette Geman |
| Abstract: | In 2017, eight scientists from the Google research team published in the journal Advances in Neural Info Processing Systems the remarkable article “Attention is all you need, ” which introduced a Transformer neural network architecture. The paper has been cited over 173, 000 times and ranks among the top 100 most cited papers of the 21st century. It builds on the attention principle introduced in 2014 by Bahdanau, Cho and Turing Award winner Bengio, who proposed neural machine translation by jointly learning to align and translate. This transformer approach has become the main architecture for a wide variety of AI tasks, including large language models. In machine learning, “attention” refers to a mechanism that allows models to focus on specific parts of the input data during the learning process and to determine the relative importance of each component within a sequence. Turning to financial economics, financial news—whether in terms of volume, unusual frequency, or sentiment (positive versus negative tone)—has long attracted the attention of researchers seeking to forecast market dynamics—“buy on rumors, sell on news.” Financial bubbles, however, remain among the most challenging phenomena to model and trade. Traditional models relying solely on price dynamics often fail to detect bubbles in real time, a key objective for stock picking and portfolio selection. Advances in natural language processing (NLP) now enable researchers to quantify market attention and sentiment using financial news and social media activity. This paper builds on recent research on sentiment in financial markets and integrates these insights into quantitative bubble detection models derived from the Log-Periodic Power Law (LPPL) literature, while incorporating a Hype Index that measures disproportionate news attention at a given moment, in order to obtain a hype-adjusted view of speculative dynamics. Within this framework, sentiment and news intensity modify bubble scores derived from price dynamics. The resulting Hyped Log-Periodic Power Law (HLPPL) model improves the identification of emerging bubbles and enables the detection of negative bubbles, corresponding to temporarily overvalued assets. The approach further highlights the importance of the choice of numéraire with respect to which prices are expressed (e.g., gold versus the dollar), emphasizing that bubbles must be assessed relative to a chosen reference asset. Empirical illustrations across equities and cryptocurrencies show how media attention and narrative amplification interact with price dynamics during speculative episodes. Taken together, these results suggest that incorporating information flows and market narratives can significantly improve the early detection and interpretation of financial bubbles. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:ocp:pbecon:pb27_26 |
| By: | Helyette Geman |
| Abstract: | In 2017, eight scientists from the Google research team published in the journal Advances in Neural Info Processing Systems the remarkable article “Attention is all you need, ” which introduced a Transformer neural network architecture. The paper has been cited over 173, 000 times and ranks among the top 100 most cited papers of the 21st century. It builds on the attention principle introduced in 2014 by Bahdanau, Cho and Turing Award winner Bengio, who proposed neural machine translation by jointly learning to align and translate. This transformer approach has become the main architecture for a wide variety of AI tasks, including large language models. In machine learning, “attention” refers to a mechanism that allows models to focus on specific parts of the input data during the learning process and to determine the relative importance of each component within a sequence. Turning to financial economics, financial news—whether in terms of volume, unusual frequency, or sentiment (positive versus negative tone)—has long attracted the attention of researchers seeking to forecast market dynamics—“buy on rumors, sell on news.” Financial bubbles, however, remain among the most challenging phenomena to model and trade. Traditional models relying solely on price dynamics often fail to detect bubbles in real time, a key objective for stock picking and portfolio selection. Advances in natural language processing (NLP) now enable researchers to quantify market attention and sentiment using financial news and social media activity. This paper builds on recent research on sentiment in financial markets and integrates these insights into quantitative bubble detection models derived from the Log-Periodic Power Law (LPPL) literature, while incorporating a Hype Index that measures disproportionate news attention at a given moment, in order to obtain a hype-adjusted view of speculative dynamics. Within this framework, sentiment and news intensity modify bubble scores derived from price dynamics. The resulting Hyped Log-Periodic Power Law (HLPPL) model improves the identification of emerging bubbles and enables the detection of negative bubbles, corresponding to temporarily overvalued assets. The approach further highlights the importance of the choice of numéraire with respect to which prices are expressed (e.g., gold versus the dollar), emphasizing that bubbles must be assessed relative to a chosen reference asset. Empirical illustrations across equities and cryptocurrencies show how media attention and narrative amplification interact with price dynamics during speculative episodes. Taken together, these results suggest that incorporating information flows and market narratives can significantly improve the early detection and interpretation of financial bubbles. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:ocp:pbcoen:pb27_26 |
| By: | Zhu, C.; Wang, X.; Zhang, W. |
| Abstract: | Large language models (LLMs) are increasingly delegated tasks once reserved for trained researchers: generating hypotheses, choosing specifications, drafting conclusions. Whether this delegation produces trustworthy science is not solely a technical question about model capability; it also depends on how cognitive labour is structured between humans and machines. We study this problem by organising an AI-assisted research workflow around behavioural-science principles, pre-commitment, decision sequencing, accountability, and attention allocation, rather than by directly measuring human psychology at the gates. We propose that reliability in AI-assisted research is a property of decision architecture: the placement, sequencing, and binding force of the choices that humans and AI components are each permitted to make. In a pre-specified 2 × 4 factorial experiment (N = 280 complete research runs across four datasets), an unconstrained multi-agent baseline produced critical failures in 72% of runs; the same underlying model and the same agent decomposition, with identical prompts on the reasoning agents shared by both arms, failed in 16% once organised by three architectural commitments (LLMs restricted to reasoning, data and estimation executed deterministically, three human decision gates; Fisher’s exact p |
| Date: | 2026–06–12 |
| URL: | https://d.repec.org/n?u=RePEc:cam:camdae:2643 |
| By: | Eugene Park |
| Abstract: | We study how frontier large language models (LLMs) behave as financial forecasters during boom-bust market cycles when made progressively aware of Soros's theory of reflexivity. Standard AI-assisted forecasting treats the market as an exogenous system. Reflexivity theory holds otherwise: prices shape fundamentals, and every forecaster is a participative agent in the loop it analyzes. We evaluate three frontier models - GPT5, Claude Sonnet 4.6, and Gemini 3 Pro - under four accumulating zero-shot conditions across two historically distinct episodes: the dot-com bubble (1996-2001) and the global financial crisis (2004-2009). The primary metric is directional forecasting accuracy; we also report the Sharpe ratio of an implied long/cash strategy to capture the risk-adjusted economic value of the forecasts. All inputs are anonymized and normalized to guard against memorization. We find that conditions incorporating reflexivity awareness improve forecasting accuracy differently across models and context windows, revealing that the same theoretical awareness can produce qualitatively different forecasting behavior across frontier LLMs. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.00061 |
| By: | Yoshiyuki ARATA |
| Abstract: | Recent research has made substantial progress in studying shock propagation through inter-firm transaction networks, and empirical studies have directly documented firm-level shock propagation. Despite these advances in both theory and empirics, no method has yet been established to accurately predict the effects of large-scale future shocks, such as natural disasters, financial crises, or pandemics. A central challenge is the heterogeneity inherent in firms and transaction relationships, which makes it difficult to identify which firms are important for shock propagation and which links amplify it. To address this issue, this study uses firm-level data and a graph neural network (GNN) to predict firm growth rates with a model that explicitly incorporates network structure. In particular, by analyzing the trained GNN model, we quantitatively identify the firms and transaction links that are important for shock propagation. Using the global financial crisis, specifically the sharp decline in exports, as a case study, we show that incorporating network structure significantly improves predictive performance and enables us to identify specific firms and links that are important for propagation. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:eti:dpaper:26045 |
| By: | Prashik N. Somkuwar; K. Srinivasan; G. Raghavan |
| Abstract: | Quantum combinatorial optimization offers theoretical advantages for complex financial modeling, but physical implementation on Noisy Intermediate Scale Quantum (NISQ) devices is severely constrained by hardware topology. This study presents a hardware benchmarking analysis between a Hardware Efficient Variational Quantum Neural Network (HE-VQNN) and the Warm Start Quantum Approximate Optimization Algorithm (WS-QAOA) for a hybrid Mean Variance and Conditional Value at Risk (CVaR) portfolio objective. By implementing a novel classical quantum hybrid proxy matrix to bypass the CVaR auxiliary qubit bottleneck, we map up to 16 assets from the NIFTY 50 index onto an IBM heavy hex processor. We systematically quantify algorithmic resilience to the "SWAP tax" incurred during routing. Empirical results reveal a critical operational trade-off: WS-QAOA provides exact theoretical mapping but suffers catastrophic hardware decoherence due to exponential nonlocal gate overhead. Conversely, HE-VQNN preserves hardware coherence but lacks the mathematical expressibility to capture dense tail risk asset correlations. This study exposes the limitations of dense financial optimization on current architectures forces an nonviable choice between algorithmic inexpressibility and hardware decoherence. This is indicative of a deeper limitation as to what can and cannot be done with NISQ computers lacking in all-to-all connectivity. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.07727 |
| By: | Andrei Bysik; Robert \'Slepaczuk |
| Abstract: | This paper investigates whether machine learning forecasts of hourly BTC-USDT returns can be converted into economically meaningful trading performance after transaction costs. Using approximately 70, 000 hourly observations from 2018-2026, XGBoost, LSTM, and iTransformer are evaluated in a 27-fold walk-forward protocol. All three models produce positive gross trading performance in selected configurations, but naive sign-based strategies fail once transaction costs of ten basis points are imposed. A cost-aware execution filter, which prevents trades only when the forecast magnitude exceeds a transaction-cost-based threshold, sharply reduces turnover and restores profitability in selected configurations. The strongest long-only XGBoost strategy produces annualised returns above 65% with a Sharpe ratio above one. Additional tests show that technical indicators improve performance in selected cases, EGARCH-derived features do not provide uniformly robust gains, and XGBoost is descriptively stronger than the neural alternatives, although bootstrap evidence does not support formal statistical dominance. Loss-function and model-selection effects are secondary and statistically fragile. The results show that the main obstacle in hourly cryptocurrency trading is not only weak predictability, but also the way forecasts are converted into trades. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.00060 |
| By: | Bryan T. Kelly; Semyon Malamud; Johannes Schwab; Teng Andrea Xu |
| Abstract: | Large language models trained on unrestricted internet corpora inevitably embed information from the future, introducing lookahead bias that compromises the validity of backtests and causal inference in finance and the social sciences. Point-in-time language models—trained exclusively on text available up to each calendar date—eliminate this leakage by construction, but existing efforts typically produce models that lag substantially behind their unconstrained counterparts. We show that this performance gap can be narrowed through scale. Training decoder-only transformers with up to 4 billion parameters on 1 trillion chronologically filtered tokens from FineWeb, we construct a sequence of monthly model checkpoints spanning 2013–2024. Across a range of common-sense reasoning and language understanding benchmarks, our models approach the performance of leading open-weight models of comparable size (such as Gemma-3-4B and LLaMA-7B) trained on temporally unrestricted data, although a performance gap remains on several tasks. Finally, in a strict out-of-sample economic evaluation task, portfolios built from point-in-time embeddings achieve robust positive Sharpe ratios and perform close to full-sample counterparts that violate temporal validity, indicating that chronologically consistent language models can extract economically meaningful signals without relying on look-ahead bias. We release the complete pipeline—including dataset construction, training infrastructure, and evaluation code—to enable reproducible point-in-time language modeling and to support research applications that require strict temporal validity. |
| JEL: | C14 C45 G11 G14 G17 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35247 |
| By: | Albanese, Andrea (Luxembourg Institute of Socio-Economic Research (LISER)); Marguerit, David (Luxembourg Institute of Socio-Economic Research (LISER)) |
| Abstract: | Cross-border work is expanding in the EU, yet its labor-market effects on the cross-border workers themselves remain largely undocumented. Using linked Belgian administrative registers that identify cross-border spells in Luxembourg, we estimate the effects of cross-border employment on post-return labor-market outcomes through dynamic double machine learning. Returnees face a short-run employment penalty that fades with cross-border tenure and time since return. They are also more likely to receive Belgian unemployment benefits than comparable stayers, with higher daily benefit levels among recipients. |
| Keywords: | cross-border commuting, return migration, unemployment insurance, EU labor mobility, administrative data |
| JEL: | J61 J65 J64 I38 C21 C14 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18674 |
| By: | David Arbour; Eli Ben-Michael; Avi Feller; Apoorva Lal; Lo-Hua Yuan |
| Abstract: | Generative AI and large language models can produce realistic predictions of human behavior from rich, unstructured inputs with little to no task-specific training data. Recent work uses these ``digital twin'' predictions to supplement human responses in surveys and experiments. We study the special case of using AI-generated predictions to reduce variance in randomized experiments. We argue that doing so requires no new estimators and that researchers can simply include AI predictions as covariates in standard regression adjustment, analogous to adjusting for a prognostic score. A benefit of this approach is a ``do no harm'' property whereby the adjusted estimator reverts to the unadjusted difference in means when predictions are uninformative. Other methods, such as variants of prediction-powered inference, do not have this guarantee. We provide implementation guidance, including how to obtain continuous scores from discrete LLM outputs and how to use LLMs to featurize unstructured inputs as auxiliary covariates. We demonstrate these ideas in simulations and three empirical applications: a survey mega-study, an email marketing A/B test, and a large-scale technology platform experiment. Overall, efficiency gains are real if modest, with greater benefits in studies that contain substantial text and other unstructured data. We also confirm the do no harm property empirically. Given these gains and limited costs, we recommend adjusting for AI-generated predictions as a regular empirical practice. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.08853 |
| By: | Yanchen Jiang; David C. Parkes; Tonghan Wang |
| Abstract: | Characterizing revenue-optimal auctions for multi-item, multi-bidder settings remains a fundamental open problem, with no known closed-form solution existing beyond restrictive binary-type instances. This has motivated interest in computational approaches to optimal auction design. In this paper, we introduce the first computational framework that directly tackles the dual problem for multi-item, multi-bidder auctions and dominant-strategy incentive compatibility (DSIC), generating certified revenue upper bounds. Our approach parametrizes Lagrange multipliers with a structurally guaranteed strict flow-conservation property using neural networks, enabling efficient optimization over feasible dual solutions via gradient descent. To bridge the gap between discrete computational methods and theoretical guarantees for continuous types, we develop a novel lifting technique that maps dual certificates from coarse discretizations to fine refinements. We prove that lifting gives valid revenue upper bounds for multi-item, multi-bidder auctions with continuous uniform valuations. Furthermore, we give a generalized lifting construction for arbitrary continuous distributions and demonstrate that these lifted duals converge to the revenue of the original continuous problem in the discrete limit. We validate this computational framework for the dual auction design problem by recovering known analytical mechanisms for canonical instances. For multi-item multi-bidder problems, our framework establishes a small gap between the optimal revenue and best-known DSIC mechanisms, providing computational certificates of near-optimality. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.10112 |
| By: | Arvind Ashta |
| Abstract: | Artificial intelligence (AI) is emerging as a transformative force in microfinance and financial inclusion, addressing long-standing barriers such as credit invisibility, high operational costs, and limited access to formal financial services. This paper systematically examines AI applications across key financial domains (payments, savings, lending, insurance, investments) highlighting how machine learning, natural language processing, and generative AI are enabling innovative solutions tailored to the needs of marginalized populations. Drawing on contemporary research and case studies from the Global South, the analysis demonstrates AI’s potential to democratize financial services through alternative credit scoring, automated underwriting, and adaptive tools. However, the deployment of AI also presents significant challenges, including algorithmic bias, proxy discrimination, privacy violations, and the risk of exacerbating digital divides. The paper underscores the need for robust governance frameworks, ethical oversight, and inclusive policies to mitigate these risks and ensure that AI-driven financial inclusion serves the most vulnerable without creating new forms of exclusion. Future directions include advancing fairness-aware AI, improving transparency, and fostering cross-sector collaboration to align technological innovation with social justice and human dignity. |
| Keywords: | Artificial Intelligence; Microfinance; Financial Inclusion; Machine Learning; Alternative Credit Scoring; Algorithmic Bias; Digital Divide; Ethical AI; Global South |
| JEL: | G21 G23 O16 O33 D81 I25 C45 C55 |
| Date: | 2026–06–05 |
| URL: | https://d.repec.org/n?u=RePEc:sol:wpaper:2013/408010 |
| By: | Daniel Cunha Oliveira; Kieran Wood; Stefan Zohren; Mihai Cucuringu; Andr\'e Fujita |
| Abstract: | Deep learning models show promise in financial forecasting, yet their generalization is often undermined by small datasets, noisy signals, and non-stationarity. While meta-learning and related techniques mitigate some of these issues, they typically do not account for a core limitation in macro-financial prediction: the scarcity of distinct macroeconomic regimes that drive asset returns. We introduce HANET (Hierarchical Attention Network), a hybrid LSTM-based architecture that integrates macroeconomic domain knowledge through attention over long-run macro contexts while preserving high-frequency market dynamics. HANET organizes information in a hierarchical mixed-frequency structure, with daily asset-return signals nested within monthly macroeconomic windows, and introduces a Hierarchical Cross-Attention mechanism that reconciles low-frequency macro signals with high-frequency returns without discarding granular daily information. By framing regime selection as attention over macroeconomic contexts, the model adapts to scarce and shifting regimes. Empirically, across 55 liquid futures spanning multiple asset classes, HANET consistently outperforms neural forecasters that ignore macroeconomic information, particularly during turbulent periods, improving risk-adjusted returns and mitigating losses. Ablation studies show that these gains rely on structured macro conditioning rather than naive feature augmentation: an LSTM with the same macro representation performs poorly, and shuffling macro contexts substantially degrades performance. Finally, HANET provides interpretability through attention weights, highlighting which historical regimes are most influential for each forecast and linking macro conditions to portfolio outcomes. These results establish HANET as a systematic approach to integrating macroeconomic information into attention-based deep learning for financial forecasting. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.00624 |
| By: | Enoch Hyunwook Kang |
| Abstract: | In the forward reinforcement-learning problem, the reward is fixed and known; the learner is asked to find a good policy or value function. Here we turn the question around. Given offline data generated by an expert, can we recover the reward the expert was optimizing? This is the inverse reinforcement learning problem, and remarkably, two communities, structural econometricians studying dynamic discrete choice (DDC) and machine learners studying entropy-regularized IRL, have been working on exactly the same probabilistic model under different names. We begin by proving their equivalence. We then develop the classical identification result of Magnac and Thesmar and the classical computational paradigms that grew out of it: Rust's nested fixed-point algorithm, the conditional-choice-probability approach of Hotz and Miller, and the two temporal-difference approaches of Adusumilli and Eckardt: linear semi-gradient TD and approximate value iteration. Each route has its limits: dimensionality, transition-kernel estimation, the deadly triad, or projected fixed-point bias. We then walk through the modern ML/IRL strand: adversarial IRL, occupancy matching, IQ-Learn, and offline ML-IRL, deriving each method's actual objective and stating precisely what it does and does not identify. We close with the empirical-risk-minimization framework of Kang et al., which yields a gradient-based estimator for offline IRL/DDC. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.30843 |
| By: | Daniil Mikriukov (University of Liverpool; Xi'an Jiaotong-Liverpool University); Ruoyu Sun (Xi'an Jiaotong-Liverpool University); Angelos Stefanidis (Xi'an Jiaotong-Liverpool University); Jionglong Su (Xi'an Jiaotong-Liverpool University); Zhengyong Jiang (Xi'an Jiaotong-Liverpool University) |
| Abstract: | Deep reinforcement learning (DRL) frameworks for portfolio optimization have shown promise for their ability to learn allocation rules dynamically from market data. However, these models fail to account for fat-tailed returns, which characterize actual market behavior with more frequent extreme events. Furthermore, historical data is treated homogeneously, without accounting for temporal importance, leading models to fail during regime changes. We propose a new BAVAR-BLED algorithm that combines methods derived from Bayesian-Averaging Vector Autoregressive (BAVAR) and the Black-Litterman model using Elliptical Distributions (BLED) within a TD3 architecture. BAVAR captures a set of vector autoregressive representations that consider multi-scale temporal features, enabling adaptive allocation decisions based on regime-aware estimates of return expectations and dispersion matrices. These estimates serve as prior inputs to BLED, a model that uses Student's t-distributions, allowing for more realistic fat tail return estimates. The BAVAR-BLED algorithm uses transformer networks for view construction and CNNs for risk-aversion estimates, which modify dynamic allocation decisions based on market conditions. An evaluation of 29 Dow Jones Industrial Average constituents over a decade-long market period shows that BAVAR-BLED significantly outperforms state-of-the-art methods, achieving Sharpe and Sortino ratios of 1.72 and 2.70, respectively, and total returns of 57.26%. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.09104 |
| By: | Jan Rovirosa; Jesse Schmolze |
| Abstract: | Modeling non-stationary stochastic systems requires balancing the representational capacity of deep learning with the structural transparency of classical probabilistic models. Markov transition matrices provide such a framework, but traditional frequency-based estimation collapses at high resolutions due to data sparsity. We propose a hybrid approach that parameterizes the manifold of stochastic matrices through a neural network, enabling estimation of time-inhomogeneous Markov chains in sparse-data regimes, and use financial markets as a testbed to investigate the Markov state variable as a critical inductive bias. We show that conditioning on realized volatility produces a more internally consistent Markovian structure than return-based states, achieving a $5.6\%$ reduction in Chapman-Kolmogorov discrepancy and superior held-out likelihood in 9 of 10 assets. Unlike black-box sequence models, our approach generates explicit matrices amenable to direct geometric analysis, surfacing structural findings such as the universal homogenization of transition probabilities under high-volatility regimes. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.30943 |
| By: | Mingxuan Yi; Vidal Mehra; Jing Chen; John Cartlidge |
| Abstract: | Regime shifts in financial markets reorganise the joint dynamics of asset prices and macro variables, breaking any single-regime calibration. They are nonetheless difficult to detect reliably because the data signal is noisy and heavily multicollinear, while the contemporaneous text that announces them is unstructured. Standard regime shift detection methods rely solely on structured time-series data and ignore policy communications, even though these texts often signal shifts before they materialise in observed prices. We propose a text-enhanced regime shift detection pipeline that combines large language model (LLM) reasoning over central-bank communications with statistical validation on multivariate financial time series. The framework is detector-agnostic: text-proposed candidates are validated using a bootstrap likelihood-ratio test on a vector autoregression (VAR), while data-driven candidates from arbitrary regime detectors are ratified through a lenient LLM text check. We evaluate the framework on 2010-2024 FOMC minutes paired with a 14-variable U.S. Treasury and macroeconomic panel, using four interchangeable data-driven detectors. The proposed pipeline achieves F1 = 0.82 against a verified anchor list of monetary-policy regime shifts, with same-day modal detection latency and consistently stronger performance than pure data-driven baselines. The results demonstrate that combining unstructured policy text with statistical structural-break detection improves the robustness and interpretability of regime shift identification in financial markets. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.30363 |
| By: | Andreas Aigner |
| Abstract: | We present a hybrid news sentiment engine that continuously learns market sentiment from paired news headlines and concurrent asset-price snapshots without requiring any neural network training or GPU compute. The system uses a three-way ensemble combining (1) a financial-domain lexicon (FinBERT-style keyword scoring), (2) an adaptive statistical TF-IDF cluster learner that organizes headlines into semantic neighborhoods and tracks their average realized price reactions, and (3) an auto-calibrating weighting mechanism that adjusts ensemble contributions based on each signal's historical correlation with actual price movements. The engine runs on a 3-hour polling cycle from the Tradeflags NewsFeed API, which provides 22 price-snapshot fields per news item spanning equity indices (ES, NQ, SPY, DJIA, NDX, IWM), commodities (CL), and cryptocurrencies (BTC, ETH). All processing occurs at sub-second latency on a CPU-only server at effectively zero marginal cost per analytic cycle. We compare our approach against established methods -- FinBERT, GPT-based scoring, VADER, and commercial sentiment APIs -- across dimensions of cost, latency, accuracy, and adaptability. Our statistical cluster learner, which adapts to changing market regimes without retraining, represents a novel contribution not found in existing sentiment systems. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.03457 |
| By: | Yichi Zhang; Ke Zhu; Zhoufan Zhu |
| Abstract: | Learning Value-at-Risk (VaR) and Expected Shortfall (ES) is important for managing financial risks effectively. Existing approaches with limited parameters are vulnerable to model misspecification in the era of big data. To address this limitation, we propose a large tail risk model, the retrieval-enhanced self-grouping autoencoder (ReSGA), which is designed with millions of parameters to exploit the rich cross-sectional dependence and long-term temporal dynamics of assets using their characteristics. Applied to monthly US equity returns from 1926 to 2023 with 153 firm characteristics, ReSGA outperforms twelve econometric and machine learning competitors in terms of out-of-sample loss and statistical backtesting. In addition, its forecast advantages can translate into significant economic gains from long-short decile portfolios that are constructed by a new size-enhanced left-side momentum strategy. To clarify the role of complexity, we further conduct a systematic scaling analysis and demonstrate that improvements in joint VaR-ES forecasting are primarily driven by data complexity rather than model complexity. Finally, our analyses of group-importance and transfer-learning exhibit the interpretability and cross-market generalizability of ReSGA. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.04576 |