|
on Computational Economics |
| By: | Aquilina, Matteo; Araujo, Douglas; Gelos, Gaston; Park, Taejin; Perez-Cruz, Fernando |
| Abstract: | Predicting financial market stress has long proven to be a largely elusive goal. Advances in artificial intelligence and machine learning offer new possibilities to tackle this problem, given their ability to handle large datasets and unearth hidden nonlinear patterns. In this paper, we develop a new approach based on a combination of a recurrent neural network (RNN) and a large language model. Focusing on deviations from triangular arbitrage parity (TAP) in the Euro-Yen currency pair, our RNN produces interpretable daily forecasts of market dysfunction 60 business days ahead. To address the “black box†limitations of RNNs, our model assigns data-driven, time-varying weights to the input variables, making its decision process transparent. These weights serve a dual purpose. First, their evolution in and of itself provides early signals of latent changes in market dynamics. Second, when the network forecasts a higher probability of market dysfunction, these variable-specific weights help identify relevant market variables that we use to prompt an LLM to search for relevant information about potential market stress drivers. |
| JEL: | G14 G15 G17 |
| Date: | 2025–10 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:20768 |
| By: | Veni Arakelia; Guglielmo Maria Caporale; Mirto M. Gasparinatou; Menelaos Karanasos |
| Abstract: | This paper examines the forecasting of liquidity dynamics in European stock markets by means of traditional econometric models and machine learning techniques. It uses daily data for the DAX, CAC 40, FTSE 100, FTSE MIB, and IBEX 35 over 2010–2026, liquidity being measured by the logarithmic Amihud illiquidity indicator. The empirical framework compares ARIMA models and a dynamic panel specification with Random Forest, Extreme Gradient Boosting (XGBoost), and Support Vector Regression (SVR) within a common rolling one-step-ahead forecasting framework. The results show that liquidity is highly persistent and that the dynamic panel model achieves the lowest forecast errors, although Diebold–Mariano tests indicate no significant predictive advantage over the leading machine learning models. SHAP analysis reveals that trading activity, lagged liquidity, and market uncertainty are the main determinants of liquidity forecasts. The findings highlight the complementary role of explainable machine learning in empirical finance. |
| Keywords: | liquidity dynamics, european stock markets, forecasting, econometric models, machine learning (ML), artificial intelligence (AI) |
| JEL: | C22 C33 C53 G17 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_12829 |
| By: | Yang Liu; Yuhao Liu; Yunran Wei |
| Abstract: | We propose a noise-robust elicit-to-optimize framework that integrates inverse reinforcement learning (IRL) and reinforcement learning (RL) for eliciting agents' risk preferences and optimizing policies under a broad class of risk objectives characterized by distortion riskmetrics. On the elicitation side, we propose an adaptive Bayesian IRL method that infers agents' latent risk objectives from their noisy observed decisions, explicitly allowing agents to take stochastic and suboptimal actions. We establish the existence of a finite set of distinguishing questions that identifies the preferred distortion riskmetric within the candidate class and prove that the convergence rate of the algorithm is of order $O(\exp(-cm+O(\sqrt{m\log m})))$ under general settings, where $c>0$ is a constant and $m$ denotes the number of algorithm iterations. On the optimization side, we develop a model-free RL algorithm for optimizing policies under conditional distortion riskmetrics. By representing the objective as an integral of the conditional cost quantile function with respect to the distortion function, the method unifies distortion-riskmetric objectives. We optimize diverse risk objectives by extending the Proximal Policy Optimization (PPO) algorithm with policy, value, and quantile neural networks, where the quantile network estimates the full conditional cost quantile function and enables numerical evaluation of general risk objectives. A comprehensive empirical study demonstrates the framework's elicitation accuracy and effectiveness in complex financial environments. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.14373 |
| By: | Anand, Kartik; Kazinnik, Sophia; Leonello, Agnese; Panetti, Ettore |
| Abstract: | Does artificial intelligence (AI) pose a threat to financial stability? We study AI investor behavior, specifically Q-learning and large language model (LLM) investors, in a mutual fund redemption problem with economic and strategic uncertainty. Different AI architectures generate systematically different outcomes. Q-learning investors coordinate well but under default risk exhibit excessive redemption that amplifies fragility. LLM investors internalize equilibrium structure but display belief heterogeneity, weakening coordination and predictability. Our findings show that AI architecture is a first-order determinant of financial stability. |
| Keywords: | Coordination games; Financial stability; Q-learning; Large Language Models; Artificial intelligence; Strategic uncertainty |
| JEL: | G01 G23 C63 |
| Date: | 2025–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:20681 |
| By: | Xianhua Peng; Wu Guo; Songyan Wang; Jianfei Zhu |
| Abstract: | This paper proposes the certainty-equivalent first-order learning (CEFOL) algorithm, a deep learning algorithm for solving discrete-time dynamic programming problems with recursive utility. Dynamic programming with recursive utility is challenging because nonlinear certainty equivalent appears in the Bellman equation and the first-order optimality conditions but is difficult to evaluate. By introducing a separate neural network to represent the certainty equivalent, CEFOL enables the exploitation of the Bellman and model-specific first-order optimality conditions. In addition to certainty equivalent, CEFOL also uses neural networks to learn the value functions, policy functions, and Lagrange multipliers by using model-specific first-order conditions to construct residuals for minimization. By using first-order and KKT residuals to learn the policy, CEFOL directly accommodates general equality and inequality constraints on the controls, including occasionally binding constraints, without requiring penalty functions or problem-specific reformulations. We apply the algorithm to risk-sensitive and Epstein--Zin consumption-saving problems, a small-noise robust-control problem, and a DSGE model with recursive preferences and stochastic volatility. Across these applications, out-of-sample Bellman diagnostics and model-specific optimality residuals, including Euler or first-order residuals where applicable, are generally of order 1.0e-4 to 1.0e-3 over the relevant state regions, with larger values mainly near binding constraints, and the learned value and policy functions closely match VFI benchmarks when available. The CEFOL algorithm also works for dynamic programming problems with expected utility, as expected utility is a special case of recursive utility. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.09461 |
| By: | Bartosz Zi\'o{\l}ko; Kacper Dobrzeniewski |
| Abstract: | In this study, we examine the opportunities brought by Large Language Models (LLMs) to various aspects of fundamental analysis of companies based on their reports as well as data and documents describing macroeconomic situation like GDP and inflation changes as well as documents filled to the U.S. Securities and Exchange Commission (SEC) which can be found in EDGAR. We were preprocessing those data and than sending via API to gpt-4o model in a Retrieval-Augmented Generation (RAG) like regime. We prepared as well a document describing an exemplar investor knowledge based on Kitchin cycles. We were scanning data important for analysis of 9 companies for 4 weeks. Using LLM we were producing automatic briefs about them. They were sent to nine participants who are individual investors to evaluate usefulness of such approach to data analysis. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.09121 |
| By: | Fengzhuo Zhang; Zhuoran Yang; Dirk Bergemann |
| Abstract: | Large Language Models (LLMs) have revolutionized AI services, but a critical tension emerges: while personalization improves model performance, it consumes scarce computational resources that users must share. When should a user invest in expensive Supervised Fine-Tuning (SFT) versus lightweight In-Context Learning (ICL)? How does congestion from other users' personalization choices reshape these incentives? And what strategies should platforms adopt when offering multiple personalization algorithms? We develop a tractable framework for LLM serving that captures the statistical-economic trade-offs users face. Our analysis yields several surprising insights. First, we show that ICL and SFT dominate in different regimes, determined by an interplay between pretraining coverage and data signal-to-noise ratios, but congestion can flip these rankings. Second, equilibrium resource consumption exhibits pronounced non-monotonicity: improving pretraining precision reduces the congestion, while broader pretraining coverage and harder tasks sometimes increase it. Third, we prove that offering both personalization methods never hurts the platform's maximal profits, despite potentially increasing computational load. Experiments with GPT-2 on linear regression tasks validate our theoretical predictions about algorithm performance. Complementing these results, our review of documentation from 21 major AI platforms shows that the share offering both SFT and ICL increased from 9.5% in 2021 to 71.4% in 2025, consistent with our platform-design implications. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.14371 |
| By: | Griesshaber, Niclas; Ogilvie, Sheilagh |
| Abstract: | What can we learn about institutional transplantation by analyzing craft guilds in colonial Latin America? We use large language models (LLMs) to investigate colonial guild ordinances, addressing two major bottlenecks in assessing institutions: digitizing qualitative sources efficiently and analyzing them quantitatively. Our newly designed methodology reveals both long-term continuities and striking differences between craft guilds in colonial Mexico and Peru, particularly with regard to human capital and product quality. The LLM-based approach identifies patterns that were previously not discernible using standard methods in economic history, its results are reproducible, and it can easily be extended to other historical settings. |
| Keywords: | Institutions; Large Language Models |
| JEL: | N86 N46 C55 C63 |
| Date: | 2025–08 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:20556 |
| By: | Yufei Wu; Daniel Schmierer; Dan Zylberglejd |
| Abstract: | In two-sided marketplaces with heterogeneous products, it is important to understand the causal relationship between additional supply and marketplace outcomes, such as the total quantity transacted or transaction value in the marketplace. This paper studies a causal machine learning approach to estimating this relationship across product segments. We use the Airbnb marketplace as an example, focusing on the impact of additional listing supply on total bookings, but the methodology applies to other two-sided marketplaces. Our approach combines double/debiased machine learning with a hierarchical Bayesian framework that leverages pre-existing knowledge as priors. We construct tractable and informative features for the model by leveraging measures of product segment similarity from the geospatial literature. We find that such a model provides plausible estimates of the marketplace returns to additional supply and strong out of sample performance. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.30999 |
| By: | Xianhua Peng; Wu Guo |
| Abstract: | We propose the first deep learning algorithm, the Certainty Equivalent Learning (CEL) algorithm, for solving high-dimensional discrete-time dynamic programming problems with recursive utility. Dynamic programming with recursive utility is numerically challenging because the recursive utility does not have an explicit representation and the Bellman equation contains a certainty equivalent that is difficult to evaluate. The CEL algorithm learns this certainty-equivalent value directly with neural networks and jointly approximates value functions, policy functions, and certainty-equivalent functions. The CEL algorithm is mesh-free and simulation-based, allowing high-dimensional state and control spaces, and does not rely on Euler equations, first-order conditions, or differentiability of the state transition function. The CEL algorithm also works for dynamic programming problems with expected utility as expected utility is a special case of recursive utility. We apply the CEL to discounted linear exponential quadratic Gaussian control, small-noise robust control, Epstein-Zin DSGE, and multivariate strategic asset allocation problems. Compared with closed-form and VFI-based benchmarks, the CEL delivers accurate value and policy approximations, remains effective in high-dimensional problems, achieves accuracy comparable to VFI in the small-noise robust-control case, and produces out-of-sample Bellman errors and Euler or first-order residuals that are in the range from 1.0e-4 to 1.0e-3 for most problems. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.04278 |
| By: | Torben S. D. Johansen; Julius Koschnick; Christian Vedel |
| Abstract: | Machine learning (ML) has rapidly transformed economic history, lowering costs of digitization, data linkage, and imputation, and making information in historical text usable at scale. This paper offers a practical guide to using these tools well. However, ML tools have also created new problems. Prediction errors are often systematically correlated with covariates of interest, so even highly accurate models can distort and sometimes reverse coefficients, and standard validation cannot detect this. Given that ML tools often perform worse for historical data, this problem is especially severe for the field of economic history. We also identify a solution to this problem. We show that recent debiasing methods can correct such bias for a wide class of applications, using a small, randomly sampled set of expert-coded labels while retaining the efficiency of large-scale prediction. We organize the field with a taxonomy of three ML tasks, survey the literature along it, and indicate where debiasing applies and where validation against proxies remains the only recourse. We close with best-practice guidance on digitization, model choice, and reproducibility. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.28063 |
| By: | Fernández-Villaverde, Jesús |
| Abstract: | The ongoing revolution in artificial intelligence, especially deep learning, is transforming research across many fields, including economics. Its impact is particularly strong in solving equilibrium economic models. These models often lack closed-form solutions, so economists have relied on numerical methods such as value function iteration, perturbation, and projection techniques. While powerful, these approaches face the curse of dimensionality, making global solutions computationally infeasible as the number of state variables increases. Recent advances in deep learning offer a new paradigm: flexible tools that efficiently approximate complex functions, manage high-dimensional problems, and expand the reach of quantitative economics. After introducing the basic concepts of deep learning, I illustrate the approach with the neoclassical growth model and discuss related ideas, including the double descent phenomenon and implicit regularization. |
| Date: | 2025–09 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:20669 |
| By: | Anastasis Kratsios; Giulia Livieri; Philipp Schmocker |
| Abstract: | We address fundamental challenges in representing and computing $\mathbb{R}^{d}$-valued predictable square-integrable processes over $[0, T]$, collected in the space $\mathcal{H}^2_T(\mathbb{R}^{d})$. These processes are central to continuous-time stochastic control, reinforcement learning, and mathematical finance. Although Wiener-chaos expansions offer strong theoretical tools, traditional computational methods are hindered by the need for large chaos dictionaries and high-order iterated integrals. To overcome these obstacles, we introduce NeuralChaos -- a neural operator architecture that produces elements of $\mathcal{H}^2_T(\mathbb{R}^{d})$ using only finitely many evaluations of the driving Brownian motion, while preserving predictability and square-integrability. We prove that NeuralChaos is dense in $\mathcal{H}^2_T(\mathbb{R}^{d})$ and achieves the best $N$-term chaoslet approximation rates for compressible and Malliavin--Sobolev regular processes. Moreover, compressibility is shown to be typical for processes from $\mathcal{H}^2_T(\mathbb{R}^{d})$ under non-degenerate sub-Gaussian sampling. In contrast, we show that finite-dimensional Markovian neural SDE models constitute a meagre and Gaussian-null subset in $\mathcal{H}^2_T(\mathbb{R}^{d})$, regardless of discretization, whereas compressible processes are generic. Numerical experiments on a stochastic optimal control problem and dynamic hedging highlight the practical effectiveness of our approach. Our results enable more efficient and expressive modelling in stochastic analysis and mathematical finance. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.14361 |
| By: | Aldasoro, Inaki; Hördahl, Peter; Schrimpf, Andreas; Zhu, Sonya |
| Abstract: | Using newly constructed market conditions indicators (MCIs) for three pivotal markets centered around the US dollar (Treasury, foreign exchange, and money markets), we demonstrate that tree-based machine learning (ML) models significantly outperform traditional time-series approaches in predicting the full distribution of future market stress. Through quantile regressions, we show that the random forest method achieves up to 27\% lower quantile loss than autoregressive benchmarks, particularly at longer horizons (up to 12 months). Shapley value analysis reveals that variables related to macro expectations and uncertainty — especially about the monetary policy stance — are important predictors of future tail realizations of market conditions. For individual market segments, the state of the global financial cycle, as well as liquidity conditions, also play important roles. These results highlight the value of ML in forecasting tail risks and identifying systemic vulnerabilities in real time, bridging the gap between high-frequency data and macroeconomic stability frameworks. |
| Keywords: | Shapley value |
| JEL: | G01 C53 G17 G12 G28 |
| Date: | 2025–07 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:20439 |
| By: | Garau, Alessio |
| Abstract: | Can a language model improve how economists classify their own papers? Only 15% of four million IDEAS/RePEc records carry usable JEL codes, and similar papers often receive different ones. I use a large language model (LLM) to solve this problem and assign three-digit codes from titles and abstracts, evaluating it on 69, 503 coded articles published from 1991 to 2023 in the top 100 economics journals. Two tests do not assume that author codes provide the correct classification. Across semantic neighbors identified by a separate embedding model, model codes are 1.8 times as consistent as author codes. Holding codes per paper fixed, a blind check finds that 83% of model codes fit official American Economic Association (AEA) guidelines, compared with 67% of author codes. The classifier expands coverage, and the evaluation framework applies whenever human labels are incomplete or noisy. |
| Keywords: | JEL codes; field classification; generative artificial intelligence; large language models; text as data; semantic similarity |
| JEL: | A14 C45 C81 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:pra:mprapa:130163 |
| By: | Corrado Di Guilmi; Takashi Kamihigashi |
| Abstract: | The paper proposes a computational approach for including forward-looking intertemporally optimizing agents in agent-based models. Optimization is implemented considering, on the one hand that revision of economic behavior does not occur continuously over time but only when individual circumstances suggest or impose it, and, on the other hand, that, given the inherent uncertainty and complexity of the economic system, the planning horizon is finite. We propose a macroeconomic model with a large population of household agents. Each period a random sample of them resets their propensities to consume and invest by maximizing their intertemporal utility. They then stick to these optimally set quantities until they are again selected for optimization. The study is a primer in considering the joint effect of heterogeneous agents' interaction and forward-looking behavior, and provides novel insights into the mechanism of transmission of individual choices to the macroeconomy. The heavy computational tasks are managed through the development of new programming tools. The coexistence of interaction and forward-looking behavior generates interesting coordination dynamics. The results suggest that even a tiny fraction of optimizing agents over the whole population has a significant effect of aggregate output, but this effect is nonlinear and conditional on the length of the planning horizon. |
| Keywords: | intertemporal optimization, computational agent-based model, forward-looking behavior |
| JEL: | C63 E21 E70 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:een:camaaa:2026-53 |
| By: | Hyung Joo Kim; Dong Hwan Oh |
| Abstract: | Despite documented heterogeneity in volatility dynamics across the option surface, standard implied volatility forecasting models apply homogeneous parameters throughout. We introduce a machine-learning framework that uses regression trees to partition the surface along both moneyness and maturity dimensions, identifying data-driven regions where distinct forecasting models perform best. Extending the Surface Heterogeneous Autoregressive (SHAR) framework of Dufays, Jacobs, and Rombouts (2025), we develop tree-based SHAR specifications that preserve interpretable structure while allowing model parameters to vary across the surface. Empirical analysis using S&P 500 options demonstrates that the boosted tree-based specification achieves the lowest out-of-sample forecast errors across all horizons, reducing one-month-ahead RMSE by 13 percent versus the benchmark SHAR model. The improvements are statistically significant and particularly pronounced during stress periods. The estimated tree presents economically interpretable segmentation: short-dated options exhibit higher daily persistence but lower monthly persistence than long-dated options, while deep out-of-the-money calls or puts display distinct dynamics from near-the-money contracts. |
| Keywords: | implied volatility forecasting; option surface; machine learning; regression trees; ensemble methods; heterogeneous autoregressive models |
| JEL: | C14 C22 C32 C51 C53 C58 G12 |
| Date: | 2026–07–06 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedgfe:103519 |
| By: | Ralph S. J. Koijen; Bradford Levy |
| Abstract: | Evaluating optimized AI systems for asset pricing is fundamentally difficult for two reasons. First, models are trained on all data, implying that any backtest or analysis using historical data suffers from look-ahead bias. In addition, markets are reflexive — as investors adopt AI, prices adjust — which may erode the very patterns the AI system was trained to exploit. We introduce a real-time, out-of-sample benchmark designed to sidestep both problems. The benchmark measures how well AI systems can explain contemporaneous stock returns around earnings announcements using only information available at announcement time, including the text of the announcement itself. Applying this benchmark to a range of agentic AI systems — which extract structured signals from earnings call transcripts and optimize over those signals — we find that the best-optimized systems more than double the explained variation in returns relative to standard benchmarks (R2 increasing from 8% to close to 20%). We show that AI-based optimization can deliver efficiency gains relative to traditional machine learning methods while also improving interpretability as our approach produces human-readable economic mechanisms that explain price movements. These learned rules can be compared to the drivers of realized returns in existing asset pricing models to identify missing sources of variation in a data-driven, self-evolving way that integrates empirical learning with economic structure. We release an SDK for researchers to improve on our results. Saturating this benchmark would represent fundamental progress in understanding how capital markets process firm-level information. |
| JEL: | C10 G1 |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35431 |
| By: | Diego Franco (Central Reserve Bank of Peru); Delia Ruiz (Central Reserve Bank of Peru); Walter Cuba (Central Reserve Bank of Peru) |
| Abstract: | This paper develops an intraday early-warning framework to predict deviations of the volumeweighted overnight interbank rate from the BCRP policy rate after the close of the Central Bank’s second intervention window. We study both upward deviations, associated with liquidity-scarcity episodes, and downward deviations, associated with liquidity-abundance episodes. Using a unique high-frequency dataset spanning 2015-2025, we evaluate whether morning liquidity indicators can anticipate rate deviations exceeding 5 basis points. We compare a regularized logistic regression with a nonlinear artificial neural network (ANN), estimating separate models for each direction of deviation. Both models are calibrated on a chronological development sample and evaluated on a held-out test period. The logit model outperforms the ANN in both cases, with a statistically significant ranking advantage (ROC-AUC of 0.95 vs. 0.88 for upward deviations; 0.77 vs. 0.75 for downward deviations). Average marginal effects reveal an economically coherent asymmetry. Market concentration, measured by the HHI, and cross-bank dispersion in reserve requirement compliance reduce the probability of upward deviations and increase the probability of downward deviations. We interpret this as reflecting the presence of a small number of large, readily identifiable liquidity providers: their visibility reduces search frictions and prevents rate spikes when the market is short, while the same concentration shifts bargaining power toward borrowers, who become the scarce side of the negotiation, when the market is long. Overall, the findings support the feasibility of a simple, interpretable early-warning tool for BCRP money market operators. |
| Keywords: | Interbank money market ; Monetary policy implementation ; Earlywarning models ; Machine learning ; Market concentration ; Peru |
| JEL: | E58 E43 G21 C53 |
| Date: | 2026–07–16 |
| URL: | https://d.repec.org/n?u=RePEc:gii:giihei:heidwp17-2026 |
| By: | Chi, Ta-Chung; Fan, Ting-Han; Ghigliazza, Raffaele; Giannone, Domenico; Wang, Zixuan (Kevin) |
| Abstract: | We forecast the full conditional distribution of macroeconomic outcomes by systematically integrating three key principles: using high-dimensional data with appropriate regularization, adopting rigorous out-of-sample validation procedures, and incorporating nonlinearities. By exploiting the rich information embedded in a large set of macroeconomic and financial predictors, we produce accurate predictions of the entire profile of macroeconomic risk in real time. Our findings show that regularization via shrinkage is essential to control model complexity, while introducing nonlinearities yields limited improvements in predictive accuracy. Out-of-sample validation plays a critical role in selecting model architecture and preventing overfitting. |
| Keywords: | Regularization |
| JEL: | C22 C52 C53 C55 |
| Date: | 2025–10 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:20727 |
| By: | Po Han Teo |
| Abstract: | Large Language Models (LLMs) are increasingly used as stand-ins in behavioural games. These stand-ins rely on the assumption that the LLM's distribution of choices meaningfully matches how humans play the same game. This study tests that assumption through two games. The first is a p-beauty contest, and the second one is a public goods game. The study first investigates five local-model settings within the same model family. These settings are varied together in a 360-cell factorial, which balances temperature, scale (0.5-32B), quantisation, instruct vs base, and framing. Each cell's distribution is then compared against whole choice distributions in published human data. Each deployment setting, except for quantisation, governs a different aspect of fidelity. Mechanically, while the dispersion of human players can be somewhat recovered through deployment settings, the strategic process behind it cannot. Through the lens of the level-k cognitive theory, we find that LLMs act as static, category-retrieved level-k players, where k is set by the model scale. The models also do not run within-game belief-updating or backward induction throughout multiple-round horizon settings. While human contributions decayed in the public goods game, LLMs stayed flat or rose at every scale. When the horizon test was administered, LLMs were more cooperative under an indefinite horizon compared to a finite one. However, LLMs ignore their relative round position, so no last-round defection was displayed. This implies that LLMs retrieved levels relative to the horizon category rather than working out iteratively from the specific game setting. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.27845 |
| By: | Ioanna-Yvonni Tsaknaki; Andrea Macr\`i; Fabrizio Lillo |
| Abstract: | In this paper, we investigate whether a model-free RL agent can identify and exploit price manipulation opportunities more effectively than a traditional model-based approach that assumes correct specification of the data-generating process but relies on noisy parameter estimates. We consider a single-asset market in which prices evolve according to an Almgren-Chriss framework with non-linear permanent impact and linear temporary impact. We first establish the existence of price-manipulative strategies in discrete time and compute the optimal benchmark strategy using Sequential Least Squares Quadratic Programming under full information. We then compare two finite-sample learning approaches: a model-based procedure that estimates impact parameters from simulated execution data and an agnostic RL approach based on Deep Deterministic Policy Gradient, trained directly on the same amount of data. For intermediate volatility, the RL agent successfully discovers profitable manipulative strategies without explicit knowledge of the underlying model, even when training data are quite limited. More importantly, RL consistently outperforms the model-based approach when parameter estimates are affected by sampling error, despite the latter benefiting from the correct model specification. For large volatility, all methods are unable to identify manipulation opportunities, while for small volatility, the model based approach outperforms RL. These findings highlight both the effectiveness of RL in complex control problems and the risks associated with deploying learning algorithms in financial markets without appropriate safeguards. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.06121 |
| By: | Danial Ramezani; Mostafa Abouei Ardakan; Mohamadreza Dehghani Ahmadabad |
| Abstract: | Passive management has increasingly won popularity over the past few years because of its advantages, such as lower management fees and transaction costs. Index tracking endeavors to reproduce the performance of an index with smaller sets of assets. In this paper, a novel formulation is proposed that is not only more robust than the existing ones but also performs better on out-of-sample data and tracks indices over long periods without any considerable deviation or the need for rebalancing. Solving index tracking problems in a polynomial time is a challenging task due to their NP-hard nature. To address this issue, a novel heuristic based on metaheuristic algorithms and local branching is also developed to solve the proposed model. The heuristic enjoys not only the exploration capabilities of a genetic algorithm but the characteristics of local search algorithms as well. The data from the OR library is used to verify the capabilities of the proposed heuristic in comparison with commercial solvers. Results indicate that not only is the heuristic able to converge to optimal solutions for not-so-large problem sizes, but the portfolios it generates also outperform those yielded by commercial solvers in terms of both in-sample and out-of-sample data. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.09556 |
| By: | Ryan Parker; Mark Stedman; Luca Capriotti |
| Abstract: | Using the path-integral formalism, we develop an accurate and easy-to-compute semi-analytical approximation for a general class of {default intensity} models. We illustrate the accuracy of the method by presenting results for the Black-Karasinski model for which the proposed approximation provides remarkably accurate results, even in regimes of high volatility and multi-year time horizons. The accuracy and the computational efficiency of the proposed approximation makes it a viable alternative to fully numerical schemes for a variety of applications in econometrics and derivatives pricing, including the computation of XVA for credit products. As a practical example, we consider the pricing of a quanto Credit Default Swap (CDS) under stochastic intensity of default and an FX devaluation model. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.21800 |
| By: | Rui Yao; Kenan Zhang |
| Abstract: | Large-scale traffic assignment requires equilibrium models that are both behaviorally plausible and computationally tractable. This paper develops a perturbed utility Markovian equilibrium (PUME) framework that preserves the scalability of link-based Markovian traffic equilibrium models and extends their applicability to settings with boundary choice probabilities, undiscounted network loading, and general link interactions. As the behavioral basis of PUME, we first develop the perturbed utility Markovian choice model (PUMCM) in which the Bellman optimality operator is defined through a convex surplus function whose gradient directly yields the optimal policy. The model generalizes existing additive random utility (ARUM) Markovian choice models and admits both interior and boundary choice probabilities. Accordingly, unattractive links can receive zero flow without imposing ex ante choice-set restrictions as in existing ARUM models. We establish conditions under which the corresponding Markov decision problem is well posed and yields a proper demand mapping. We then formulate the equilibrium as a variational inequality (VI) problem on the dual cost space and establish its existence and uniqueness. Particularly, the VI formulation of PUME accommodates non-separable and asymmetric cost structures and thus offers a more flexible modeling framework than existing Markovian traffic equilibrium (MTE) models. For computation, we develop a modified policy iteration method for network loading and a safeguarded accelerated meta-algorithm for computing equilibrium. Both algorithms are proven to be globally convergent and have demonstrated satisfactory numerical performances. Experiments on benchmark and synthetic networks further show that the proposed framework is highly scalable and robust towards a wide variety of demand-supply settings. |
| Date: | 2026–07 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2607.09568 |
| By: | Piguillem, Facundo; Shi, Liyan |
| Abstract: | We analyze the dynamic problem of decision makers with quasi-hyperbolic discounting and random shocks to temptation. We show that this problem is equivalent to that of a standard consumer-saver who assigns biased weights to future shocks. This equivalence provides a straightforward methodology for finding, theoretically and numerically, the Markov equilibrium with hyperbolic agents. Through this equivalent problem, we provide conditions for the existence and uniqueness of the equilibrium. If the weights constitute a probability measure, the decision maker can be interpreted as optimistically biased, ensuring a unique equilibrium with continuous decision rules and implying no value for commitment devices. Otherwise, there is intertemporal "conflict" between present and future selves: if the conflict is limited, uniqueness is guaranteed. |
| Keywords: | Quasi-hyperbolic discounting; Optimism; Political economy; Time inconsistency |
| JEL: | E2 E7 H1 |
| Date: | 2025–08 |
| URL: | https://d.repec.org/n?u=RePEc:cpr:ceprdp:20579 |
| By: | Boero, Riccardo (NILU - the Climate and Environmental Research Institute) |
| Abstract: | This work presents a stylized computable general equilibrium (CGE) model for analysing circular economy policy strategies in an intentionally abstract setting. The model is not calibrated to a specific economy; instead, it uses round-number calibration data and systematic numerical experiments to identify the behavioural and physical conditions under which circular strategies reduce primary material use, represented by virgin metal. The analysis compares virgin-metal taxation, recycling support, and life-extension support through refurbishment, repair, and reuse in single-country and two-country model variants. The results show that circular strategies cannot be ranked by label alone. Virgin-metal taxation operates through an upstream material-price channel and is comparatively insensitive to how end-of-life products are allocated among circular routes. Recycling and life-extension support depend more strongly on collection and allocation of end-of-life products, route yields, recycled-metal quality, route substitutability, and final-product service-demand response. For some combinations of demand elasticity, substitution elasticities, route yields, and recycled-metal quality, a policy can reduce virgin-metal use even while demand for the final product service increases; for others, the same demand response can weaken or reverse the expected primary-material reduction. Support-efficiency comparisons show that refurbishment support performs better than recycling support in the closed-economy case, while this gap narrows when virgin-metal production is separated from consumption and circular processing in the two-country extension. The nested life-extension analysis further shows that refurbishment, repair, and reuse are not interchangeable policy mechanisms. Overall, the model provides a reproducible equilibrium framework for identifying where circular economy strategies work, where they fail, and which assumptions drive the difference. |
| Date: | 2026–06–20 |
| URL: | https://d.repec.org/n?u=RePEc:osf:socarx:69gcx_v1 |
| By: | Alexandros Gelastopoulos |
| Abstract: | Many bipartite social networks exhibit pronounced asymmetries in selectivity and matching opportunities: members of one side can afford to be highly selective, while members of the opposite side are forced to accept less desirable matches. While it is natural to try to explain this asymmetry in terms of the intrinsic characteristics of the two sides or other exogenous factors, here we show that such asymmetries can also emerge endogenously through a feedback process generated by the matching process itself: as one side becomes more selective, the other side is pushed to be less selective due to reduced matching opportunities, and vice versa. We develop a model in which individuals repeatedly form one-to-one matches across two groups and adapt their selectivity to achieve a target matching rate. Using both analytic and numerical methods, we show that when encounters are sufficiently frequent, the unique equilibrium is for one group to be highly selective and the other non-selective. This qualitative outcome holds even for heterogeneous groups with overlapping, almost indistinguishable distributions of target matching rates. The model makes several testable predictions, and it provides a mechanism for behavioral differentiation in repeated matching environments, with applications ranging from online dating to hiring and housing markets. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.31802 |