Skip to main content
archive Search
Submit Donate Log in

Press Enter to search · Advanced search

Economics

  • New submissions
  • Cross-lists
  • Replacements

See recent articles

Showing new listings for Tuesday, 17 November 2026

Total of 62 entries
Showing up to 2000 entries per page: fewer | more | all

New submissions (showing 20 of 20 entries)

[1] arXiv:2611.18767 [pdf, html, other]
Title: China's Local Government Debt, Strategic Competition, and Bitcoin Demand: A Conditional Transmission Framework
Bit Sun
Comments: 13 pages, 2 figures, 2 tables
Subjects: General Economics (econ.GN)

How might the resolution of China's local-government debt affect strategic competition and Bitcoin demand? This paper develops a conditional framework connecting debt accounting, bank balance-sheet adjustment, public spending priorities, and constrained portfolio substitution. A debt swap can increase recognized government debt without increasing consolidated principal, while its effect on credit capacity depends on capital relief, loss recognition, and recapitalization. A protected strategic-spending floor can preserve current effort while shifting adjustment toward civilian productive services. In the asset block, stronger relative preference for Bitcoin does not ensure greater effective demand: access constraints and global funding conditions can reverse the sign. We derive a finite-shock threshold and deterministic sign bounds, and distinguish Bitcoin demand from demand for dollar stablecoins. Reproducible numerical illustrations and channel-removal exercises show that opposite outcomes are compatible with the same assumed fiscal stress. The analysis uses dated primary sources to establish institutional context, but estimates neither a debt-to-Bitcoin causal coefficient nor a price-impact scale. Its contribution is a compact set of conditions and measurement distinctions for evaluating transmission claims, rather than an unconditional price forecast or an inference about conflict probabilities.

[2] arXiv:2611.18872 [pdf, html, other]
Title: The opposing informational effects of AI trading
Hajime Shimao
Subjects: General Economics (econ.GN); Artificial Intelligence (cs.AI)

Better artificial intelligence (AI) can improve individual forecasts without making prices better at revealing new information. In a competitive market with resource-constrained forecasting, more capable algorithms can substitute analysis of public records for private observations, reducing the private information aggregated into prices. The same trades can convey sophisticated public-data analysis to investors unable to perform it themselves. We establish conditions under which these effects coexist and the private benefit of further computation falls. Numerical equilibrium calculations and finite-data neural-market experiments exhibit corresponding patterns. Out-of-sample earnings forecasts and controlled language-model tasks show related patterns of information substitution. In the neural experiment, wider adoption increases the predictive value of price for limited observers, reduces it for advanced observers, and lowers the benefit of upgrading by 44\%. Retraining by investors who did not upgrade can amplify the loss of private-information precision. The informational consequences of AI trading depend on what prices reveal and who can use them.

[3] arXiv:2611.19021 [pdf, html, other]
Title: Aggregating Rankings of Coalitions to Rank Individuals
Takahiro Suzuki
Subjects: Theoretical Economics (econ.TH)

We develop an ordinal framework for aggregating a profile of rankings of coalitional performances to determine the ranking of individuals. From university ratings to the assessment of team-sports players, we often evaluate individuals based on their coalitional performances. Although such evaluation has been addressed by cooperative game theory, it is only recently that its ordinal counterpart, social ranking theory (Moretti, 2015), has emerged. As most of the literature considers a single coalitional ranking (which corresponds to a real-valued characteristic function), we extend the model to allow a profile of coalitional rankings (which corresponds to a multidimensional characteristic function). We tackle the aggregation problem by introducing three families of inter-profile axioms, each representing how the rankings should interact or contribute to the output. After revealing fundamental impossibilities concerning the reconciliation of different families, we characterize a new solution, Pareto-extension Lex-cel (PEL), via some weaker versions of the inter-profile axioms. PEL extends Lex-cel, which is a central solution in social ranking theory, by analogy with Sen's Pareto extension rule. Our results develop the connection between classical Arrovian aggregation, cooperative game theory, and social ranking theory.

[4] arXiv:2611.19068 [pdf, html, other]
Title: Corporate Stances on Societal Issues in the US
Jacob Conway, Andrew Hall, Magdalena Larreboure, Caroline Le Pennec, Vincent Pons, Anna Sun
Subjects: General Economics (econ.GN); General Finance (q-fin.GN)

When do firms take public positions on divisive societal issues and why? We assemble a comprehensive dataset on corporate stance-taking in the United States, using large language models to read 52,000 news articles and press releases about 836 companies that belonged to the S&P 500 at any point between 2008 and 2025. We identify 9,917 stances on nine partisan issues such as climate change, abortion, and racial discrimination. We show that stance-taking increased fourfold after 2008 and peaked in 2021. It is overwhelmingly liberal and concentrated among the largest firms. A quarter of stances give no explicit business justification. Stance-taking is only weakly related to lobbying and campaign contributions and, unlike them, does not follow the electoral cycle. Instead, firms tend to speak when media coverage and public interest in an issue peak. Together, these findings suggest that stance-taking can be understood as a contribution to broad policy debates rather than a strategy to shape policies that affect firms' core business. Finally, we investigate whose views corporate stances represent. Consumer-facing firms take more stances, and firms whose consumers lean Democratic take more liberal stances on issues other than climate. Companies with larger shares of Democratic employees, executives, and board members also take more liberal stances. Exploiting within-firm variation, we find that having more Democrats among a company's leadership increases liberal stance-taking, while having more women increases stance-taking on gender-related issues and abortion.

[5] arXiv:2611.19177 [pdf, html, other]
Title: Alternative Computational Formulations for Testing Discrete-Choice Random Utility Models: Half-Space and Network-Flow Formulations
Haoge Chang, Kota Saito, Mingxi Zhu
Subjects: Econometrics (econ.EM); Theoretical Economics (econ.TH)

This paper develops computational algorithms for testing discrete-choice random utility models (RUM). The set of RUM-consistent choice probabilities forms a polytope, and existing computational methods rely on its vertex representation. We exploit alternative half-space (Block--Marschak inequalities) and network-flow representations to develop two methods based on the alternating direction method of multipliers and a flow-based column-generation algorithm. Compared with current vertex-based methods, our algorithms are better suited to GPU implementation. We establish convergence guarantees and provide optimality certificates for all three algorithms. The resulting algorithms achieve $100$--$600\times$ speedups over existing methods. The solver based on our methods is publicly available on Github at \href{this https URL}{\texttt{d-rum-testing-gpu}}.

[6] arXiv:2611.19209 [pdf, html, other]
Title: Unit-Level Conformal Inference for Panel Counterfactuals
Ziyi Liu, Yiqing Xu
Subjects: Econometrics (econ.EM); Methodology (stat.ME)

We develop conformal inference for panel counterfactual estimators using the exchangeability of prediction errors across units rather than over time. Existing conformal methods for panels with few treated units rely on temporal properties of treated-unit residuals, which are vulnerable to strong temporal heterogeneity such as nonstationarity. We instead hold out each control unit, hide its relevant outcomes, and predict them using the same rule applied to the treated unit. Target-date estimators, including difference-in-differences and synthetic control, use controls' pre-treatment and target-date outcomes to construct the counterfactual, allowing both full conformal confidence sets and Jackknife+ intervals. Full-path estimators, including factor-model imputation and matrix completion, also use controls' outcomes at other post-treatment dates; for these, we use the Jackknife+. Both procedures yield finite-sample, period-wise confidence sets under this across-unit exchangeability, without requiring stationarity or random assignment. We characterize when the coverage holds and show why a common prediction rule alone does not ensure it. For multiple treated units, a block Jackknife+ covers their average effect, while a relabeling test is exact for sharp nulls over all treated cells. We show that the synthetic control placebo test is a full conformal test of no effect that extends to full-path estimators. We also relate our procedures to a parametric bootstrap and classify common estimators by the inference they allow. We illustrate our procedures in three empirical applications.

[7] arXiv:2611.19417 [pdf, html, other]
Title: Are Revealed Preferences All You Need for AI Alignment?
Joshua S. Gans, Richard T. Holden
Comments: 86 pages, 6 figures
Subjects: Theoretical Economics (econ.TH)

AI agents learn what principals want from what principals choose. We offer a criterion for when a principal's decisions, at every belief an AI might hold, are all the AI needs. For acting, they are. Cautious actions, such as deferring to a person, typically divide the principal's actions into blocks, and for choosing costless evidence, decisions are enough exactly when the same evidence is best in every block. Otherwise, the AI acts as its principal would yet can gather the wrong evidence, and greater capability can turn that error into incorrect actions. Choices between two mistakes reveal what decisions miss.

[8] arXiv:2611.19432 [pdf, html, other]
Title: Cluster-Robust Inference in Small Samples Using Bias Corrections
M. Hashem Pesaran, Yimeng Xie
Subjects: Econometrics (econ.EM)

This paper derives an exact expression for the bias of the linearized estimator of the variance for cluster-robust inference, which generalizes an earlier result derived by Bell and McCaffrey (2002). Both cases of continuous and binary treatments are discussed. Under the former it is shown that the bias of the linearized variance estimator and its jackknife variant are of order $G^{-1}$ relative to the true variance, where $G$ denotes the total number of clusters. A bias-corrected variance estimator (BCCV) is then proposed and shown to have a relative bias of order $G^{-2}$. In the case of models with binary treatments the orders of the bias of standard estimators and the proposed bias-corrected one are shown to be $\max (G_{1}^{-1},G_{0}^{-1})$ and $\max (G_{1}^{-2},G_{0}^{-2})$, where $G_{1}$ and $G_{0}=G-G_{1}$ are the numbers of treated and untreated clusters. Scale and degrees of freedom adjustments to the $t$-statistic computed using bias-corrected standard errors are also proposed. Small sample properties of the proposed BCCV test are investigated using Monte Carlo experiments. It is shown that the proposed test generally outperforms tests based on conventional and jackknife cluster standard errors, and performs comparably to wild bootstrap inference and the recent jackknife distribution-adjusted testing procedure by \cite{hansen2026jackknife}. The utility of the proposed test is illustrated\ using three empirical applications from the literature.

[9] arXiv:2611.19476 [pdf, html, other]
Title: Team Stability: Theory and Evidence
Anton Badev
Subjects: Econometrics (econ.EM); Social and Information Networks (cs.SI)

Teams emerge organically when individuals decide on their own their activities and collaborators. I develop and test a theory in which teams form with consent, interactions reflect team composition, and payoffs feature externalities and several forms of congestion. Stability constraints concern group interactions and require each team member to be willing to remain rather than withdraw or pursue a project alone. Empirical analysis of research collaborations among economists at the Federal Reserve Board suggests that: (i.) these constraints help explain observed teams beyond the payoffs to collaboration; (ii.) estimated violations point to large teams of entrants that appear to be assigned; (iii.) a compensatory mechanism mitigates the effects of variation in policy workload, which appears to affect team composition more than total research output; and (iv.) shifting workload toward experienced economists increases entrants' research output and collaboration, at the cost of lower output among experienced economists who compensate for this shift in policy workload with increased interactions with entrants.

[10] arXiv:2611.19662 [pdf, html, other]
Title: Identifying Time Preference Shocks: Stated and Revealed Preferences after an Italian Earthquake
Krisztina Molnar, Concetta Rondinelli, Po Yin Wong
Comments: 31 pages, 9 figures, 5 tables. Online Appendix available from the authors upon request
Subjects: General Economics (econ.GN)

Identifying time preference shocks is challenging: shocks large enough to shift survey-elicited stated impatience also tighten financial resources, confounding the interpretation of stated time preference. Jointly examining stated and revealed preferences resolves this identification problem: impatience shocks and reductions in financial resources generate opposite consumption responses, even though both raise stated impatience. Applying our identification framework to the 2009 L'Aquila earthquake, we combine panel data with spatially granular seismic and damage data in a difference-in-differences design that compares undamaged households that experienced the earthquake with a spatial control group at similar distances from the epicenter. Treated households report greater impatience and increase nondurable consumption by 11 percent relative to controls. Effects are consistent with a psychological stress channel, concentrated among the most vulnerable to stress, and persist as households deplete their liquid wealth buffer over time.

[11] arXiv:2611.19690 [pdf, html, other]
Title: Comparing Two-Sample IV Estimators under Weak Instruments
Jeffrey Tse
Comments: 41 pages, 2 tables
Subjects: Econometrics (econ.EM); Methodology (stat.ME)

We compare three estimators in a two-sample summary-data instrumental variables (IV) model: inverse-variance weighting (IVW), debiased inverse-variance weighting (dIVW), and the profile score (PS) estimator. The framework assumes independent samples and instrument-specific summary estimates, with two-sample Mendelian randomisation as a leading application. We compare existing fixed- and many-weak-instrument results and the sufficient conditions for the three estimators. Under the variance scaling used in the profile-score literature, the principal collective-strength conditions for PS and dIVW are equivalent, although their additional restrictions on instrument concentration and growth differ. We then compare their asymptotic variance formulae for fixed and diverging numbers of instruments. For fixed $L$, IVW and dIVW have the same first-order asymptotic variance. Under a common variance-ratio condition, the PS variance formula exceeds the IVW variance by a positive term of order $n^{-2}$, compared with the order-$n^{-1}$ leading variance. For diverging $L$, we establish variance orders across instrument-strength regimes and the exact ordering $\mathsf{V}_{\mathtt{IVW}} < V_{\mathtt{PS}} \leq \mathsf{V}_{\mathtt{dIVW}}$ under the same variance-ratio condition. Absolute variance gaps vanish when $\sqrt{L}\kappa\to\infty$, whereas relative equivalence holds if and only if $\kappa\to\infty$. These comparisons concern asymptotic variance formulae rather than finite-sample efficiency or mean squared error.

[12] arXiv:2611.20030 [pdf, html, other]
Title: High-Dimensional Approximate Factor Models with Score-Driven Dynamics: Estimation and Inference
Matteo Barigozzi, Enzo D'Innocenzo
Subjects: Econometrics (econ.EM); Methodology (stat.ME)

We propose a factor model for high-dimensional time series where the dynamics of the latent factors is generated by a multivariate score-driven model, thus allowing to model nonlinearities and heavy tails in both the conditional mean and the conditional variance. Estimation is in two steps: first the factors are extracted via Principal Components and then the parameters of the score-driven model are estimated via Quasi Maximum Likelihood based on the estimated factors. Since factors are not identified, we first establish set-consistency and asymptotic normality for the estimated parameters governing their dynamics, as both the number of time series and the sample size diverge to infinity. Then, we construct valid asymptotic prediction intervals, which do not depend on the chosen identifying constraints for the factors. Numerical results confirm the goodness of our estimator.

[13] arXiv:2611.20209 [pdf, html, other]
Title: Knowing Less After a Diagnostic Test: Dilation in Clinical Decision-Making
Filip Obradović, Gabriel Ziegler
Subjects: Theoretical Economics (econ.TH); Econometrics (econ.EM); Applications (stat.AP)

Dilation (Seidenfeld and Wasserman 1993) occurs when conditioning on new information \textit{only} increases uncertainty regardless of what the information reveals. Often viewed as an undesirable feature of models of ambiguity, dilation has sparked sustained debate about belief updating. This paper shows that following established clinical guidance can induce dilation, even for diagnostic tests that meet conventional performance benchmarks. We characterize when a test is dilation-inducing and show that a ubiquitous feature of performance studies -- evaluating tests against an imperfect reference -- creates scope for dilation. To facilitate detection before clinical use and inform regulatory standards, we develop a finite-sample valid hypothesis test of the null that a diagnostic test is dilation-inducing in this context. Applying the method to an influential study that preceded and contributed to the clinical adoption of chest CT screening for COVID-19, we find that the data are consistent with dilation.

[14] arXiv:2611.20561 [pdf, html, other]
Title: When Better Artificial Intelligence Widens the Adoption Gap in Health Care
Jonas Fooken, Ali Furkan Kalay, Metin Uyanik
Subjects: Theoretical Economics (econ.TH)

Artificial intelligence (AI) can assist healthcare providers in clinical practice or replace services for which they are paid. We model how provider influence shapes two policy choices. The first is adoption, the share of cases in which coverage is switched to care using AI. The second is the review requirement, the number of paid reviews that must accompany AI when it is used in clinical care. Assistance can lower providers' costs and increase their profits while leaving paid reviews unaffected. However, increasingly capable AI may support autonomous use that can replace reviews and reduce providers' profits. Provider gains from adoption first rise and subsequently fall as AI capability improves. Hence, providers oppose sufficiently autonomous AI if their lost income from reviews exceeds savings from better assistance. Opposition to AI adoption arises earlier for AI that displaces more reviews. When providers can influence adoption but not the review requirement, their opposition causes under-adoption. The patient value forgone at the margin grows as AI improves. At interior policies, provider influence over both decisions widens adoption and raises review requirements above the patient optimum.

[15] arXiv:2611.21070 [pdf, html, other]
Title: Trade Tensions Meet the Gig Economy: Evidence from the 2025 Halving of Chinese Travel to Japan
Suguru Otani, Chang Yang
Comments: 18 pages, 20 pages appendix
Subjects: General Economics (econ.GN)

How do geopolitical shocks to services exports reach local labor markets? Using China's November 2025 Japan travel advisory and postings on Japan's largest spot-work platform, we provide the first such evidence for the gig economy. Posted hourly wages fall immediately and persistently in prefectures more dependent on Chinese visitors, relative to others, whereas postings, hours, and employment show no robust change. Wages fall more for early morning and daytime shifts, when hotels and shops hire, than for evening and night shifts. Where each shift is priced anew, a services-trade shock lowers the price of labor in the hours it hits.

[16] arXiv:2611.21230 [pdf, html, other]
Title: Shrink Responsibly: Empirical Bayes in Economics
Peter Hull, Michal Kolesár
Comments: 60 pages including all appendices
Subjects: Econometrics (econ.EM)

Empirical researchers are often interested in the relative causal impact of individuals or institutions such as teachers, schools, hospitals, or neighborhoods. Regression-based estimates of such unit-specific parameters are often noisy; empirical Bayes (eB) methods can help isolate the signal from the estimation noise. This article provides an overview of eB methodology, showing how eB shrinkage can improve compound decisions---such as selecting individuals with the highest causal impacts---even when the working model motivating the shrinkage is misspecified. Not all analyses benefit from eB shrinkage, however, and we discuss which types of causal summaries are best estimated directly and how to do so robustly. We illustrate practical takeaways with a running analysis of public elementary and middle school value-added.

[17] arXiv:2611.21364 [pdf, html, other]
Title: A Game of Throngs: How Spotlight Tourism Reshapes Local Economies
Mafalda Batalha, Javier Viviens
Subjects: General Economics (econ.GN)

In July 2017, San Juan de Gaztelugatxe, a modest chapel in the Basque Country, appeared as "Dragonstone" in Game of Thrones, transforming a locally known site with limited tourism activity into a global tourist destination overnight. We exploit this exogenous spotlight shock to estimate its causal effects on the local economy using a novel administrative dataset and a Synthetic Control approach. We find an abrupt increase in tourism activity, far exceeding the effects documented in the literature. Despite this surge, employment gains in tourism-related sectors were largely offset by losses elsewhere, resulting in no net effect on total employment. We also document signs of housing-market overheating, with increases in both property transactions and prices. We find no evidence that the shock displaced the local population. Instead, it led to a roughly twofold increase in arrivals from abroad. Taken together, these findings suggest that a substantial increase in tourism activity does not necessarily translate into broad-based local economic gains. Instead, tourism-led growth may entail significant sectoral reallocation and housing-market pressures, underscoring the importance of looking beyond tourism activity alone when assessing its wider economic effects.

[18] arXiv:2611.21474 [pdf, html, other]
Title: Sharp High-Dimensional Gaussian Approximations over Rectangles without Covariance Restrictions
Victor Chernozhukov, Denis Chetverikov, Kengo Kato, Yuta Koike
Comments: 107 pages. Supplemental Material (28 pages) and archived Lean verification: this https URL
Subjects: Econometrics (econ.EM); Statistics Theory (math.ST)

We study how well a sum of $n$ independent random vectors with $p$ coordinates can be approximated by a Gaussian vector with the same mean and covariance. We measure the error over rectangles, so the bounds apply to maxima and simultaneous confidence regions. The coordinates may be arbitrarily dependent, even exactly linearly dependent, and the bounds do not deteriorate as the covariance matrix becomes singular. After standardizing the coordinate variances, bounded or sub-exponential coordinates give the $n^{-1/2}$ sample-size rate up to logarithmic factors, with the size or tail bounds held fixed. The main upper bounds retain the Gaussian expected maximum, improving the rates when the coordinates are strongly dependent. Fourth-moment refinements give further improvements when large observations contribute little to individual coordinate variances. If we instead impose a fixed moment bound of order $q>4$ on the largest absolute coordinate, together with a coordinate fourth-moment bound, the rate becomes $n^{-1/2+1/q}$ up to logarithmic factors. At $q=4$, the slower $n^{-1/4}$ rate is unavoidable, confirming our earlier conjecture. We also allow the bounds on observation size to grow with $n$. Lower-bound examples show that the dependence on sample size, observation size, and Gaussian width is sharp up to logarithmic factors in the regimes studied, and in several regimes the upper and lower bounds match including the logarithmic terms.

[19] arXiv:2611.21496 [pdf, html, other]
Title: Never too LATE: Exploiting restrictions for inference on the Local Average Treatment Effect
Xavier D'Haultfœuille, Yagan Hazard, Philipp Ketz
Subjects: Econometrics (econ.EM)

Under the exclusion restriction, independence, and monotonicity assumptions, Imbens and Angrist (1994) show that the Wald estimand is equal to the local average treatment effect (LATE). However, these assumptions entail additional restrictions, which we exploit for inference on the LATE. Specifically, we propose a conditional likelihood ratio test that we show controls asymptotic size uniformly over a large class of data-generating processes. The test has very low computational cost: it requires no simulation and involves solving at most four quadratic programming problems. Monte Carlo simulations show that our test can offer substantial power gains over the Wald and Anderson-Rubin tests, including when the instrument is weak.

[20] arXiv:2611.21568 [pdf, html, other]
Title: The Geoengineering Put: Ecological Debt, Termination Risk, and Asset Prices
Samuel Hartzmark, Kelly Shue
Subjects: General Economics (econ.GN)

Solar geoengineering offers an inexpensive and fast-acting way to lower global temperatures without removing carbon from the atmosphere. As warming intensifies, climate-vulnerable countries face growing incentives to deploy geoengineering unilaterally. The economic and financial consequences of this prospect remain largely overlooked. We develop a general equilibrium macro-finance framework in which geoengineering acts as a put option on climate damages, funded by the "ecological debt" of masked warming. If deployment stops, temperatures snap back in a catastrophic termination shock. Anticipation of geoengineering raises current brown firm valuations and lowers the optimal carbon tax, helping explain two puzzles in the existing literature: high brown-firm valuations and low carbon taxes. As ecological debt grows, however, termination risk transforms brown firms into toxic assets. Geoengineering also reshapes green innovation: it buys time for a breakthrough, but can weaken innovation incentives by lowering energy and carbon prices. Termination risk partly restores research incentives and directs them toward carbon removal, which can retire ecological debt. Our calibrations show that the geoengineering put can raise welfare across a range of parameter assumptions. Our framework clarifies how this conclusion depends on uncertain parameters and where further research is most needed.

Cross submissions (showing 8 of 8 entries)

[21] arXiv:2611.18785 (cross-list from cs.LG) [pdf, html, other]
Title: Setting Optimal Reserve Prices using Fine-Tuned Language Models
Joshua Foster, Fredrik Odegaard
Subjects: Machine Learning (cs.LG); Econometrics (econ.EM); Machine Learning (stat.ML)

Online auction houses must set reserve prices for highly differentiated products whose unstructured descriptions contain valuable demand information. Existing approaches generally require managers to specify product attributes in advance. We ask whether a language model can instead learn from listing descriptions and historical bids to set item-specific reserve prices before bidding begins. We fine-tune a pretrained text embedding language model using the structural likelihood of auction bids as its training objective. This directs representation learning toward the valuation distribution that determines the reserve decision. We show that consecutive upper bids identify the tail information needed to set an optimal reserve without observing or modeling the number of potential bidders. A split-sample estimator learns the representation on one half of the data and conducts structural estimation and inference conditional on that representation on the other. We apply the method to approximately 65,800 online used-vehicle auctions. With seller values calibrated to match the sale rate among listings, the model implies a counterfactual net increase of $165.8 million, representing a 14.6% increase above the incumbent reserve policy. The framework translates a listing description and a seller's value into an item-specific reserve recommendation without constructing a product taxonomy or requiring a bid history for the item being priced. It provides an approach to adapting language models for operational decisions in marketplaces where products are heterogeneous and demand-relevant information is recorded in text.

[22] arXiv:2611.20110 (cross-list from cs.GT) [pdf, html, other]
Title: Who Observes Whom: Network Rationalizability and Its Complexity
Mojtaba Fayaz-Bakhsh, MohammadAmin Fazli
Subjects: Computer Science and Game Theory (cs.GT); Computational Complexity (cs.CC); Multiagent Systems (cs.MA); Theoretical Economics (econ.TH)

Nash equilibrium and rationalizability sit at opposite informational extremes: the former presumes that each player knows all opponents' strategies, the latter that players reason from common knowledge of rationality alone. When visibility is partial and asymmetric, we study the solution concepts induced by a directed \emph{observation network} $G$ in which an edge $i\to j$ means that $i$ observes $j$'s strategy. A set of profiles is \emph{self-sustaining} if every profile in it can be justified, for every player, by conjectures from the same set that agree with everything the player observes. We consider \emph{point} conjectures (single profiles) and \emph{belief} conjectures (distributions over profiles). Both range from Nash equilibrium on the complete network to point- and correlated rationalizability on the empty network, and shrink monotonically as links are added.
Network topology governs both behavior and complexity. A predicted profile can be found in polynomial time on the empty network, whereas a single link makes search $\mathsf{PPAD}$-hard: under one-sided two-player monitoring, the observed player's strategies are exactly her Nash equilibrium strategies. On hierarchical networks (transitive tournaments), membership reduces recursively to smaller games and admits polynomial-size certificates, placing verification in $\exists\mathbb{R}$, and in $\mathsf{P}$ for two players. On the directed three-cycle, however, iterated elimination may never terminate under either variant, and under point conjectures interior predictions may admit only infinite witness sets, a pathology that belief conjectures avoid in the same game. Finally, when only realized pure actions are observed, the solution set is guaranteed to be non-empty exactly when the network has no edges, and is computable in time $O(n|A|^2)$.

[23] arXiv:2611.20204 (cross-list from cs.GT) [pdf, html, other]
Title: Paying Creators for Training Data Can Increase AI Displacement
Lijuan Luo, Emaad Manzoor, Nathan Yang, Fan Yao
Comments: 53 pages, 6 figures. Previously circulated under the title "Platform Design When Creators Train Their AI Substitutes"
Subjects: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Theoretical Economics (econ.TH)

Generative AI increasingly competes with creators while using their content as training data. In response, policymakers have called for platforms to compensate creators for training data use and allow opting out. We study how such policies affect creator displacement in a model of a platform where creators choose whether to contribute training data and consumers choose between purchasing from creators and prompting generative AI trained on their content. When compensation is required, the low-quality creator chooses to remain solely as a training data supplier while AI fully displaces her demand, despite AI quality and the price the platform charges for AI-generated content being identical with and without compensation. Without compensation, the risk of losing creators and their training data constrains the platform to set the price of creators' content to preserve their demand. This constraint also prevents further AI displacement as AI improves, with the platform raising AI prices while keeping creators' prices and demand unchanged to prevent exit. Compensation increases creators' payoffs and total surplus but reduces consumer surplus. Our findings highlight how compensation policies for training data can shape creators' economic roles in a future with more capable generative AI.

[24] arXiv:2611.20322 (cross-list from stat.ME) [pdf, html, other]
Title: Maximum Scores and Treatment Feedback in Regression Discontinuity Designs
Subir Hait
Comments: 20 pages, 3 figures, 4 tables. Code and replication-level results in ancillary files
Subjects: Methodology (stat.ME); Econometrics (econ.EM)

Hospitals use repeated risk scores to decide when to send an alert. Some evaluations summarize these scores by their maximum and use the maximum as the running variable in a regression discontinuity (RD) design. The maximum shows correctly whether a patient was ever alerted. However, it can include scores recorded after an earlier alert has already changed care. In this paper, we study what the maximum-score RD estimates when an earlier intervention changes a later score. We use a two-assessment model and derive a formula that splits the RD contrast into the causal effect and a composition term. When feedback moves scores in one direction, the composition term equals the relative jump in the density of the maximum at the cutoff times a prognostic contrast. Therefore, an exactly continuous density means no bias, and the population density jump bounds the bias. Valid inference on the jump turns this into a confidence bound under a stated sensitivity assumption. Within this model, bias behind a continuous density requires feedback in both directions. For this case, we bound bias by total boundary movement and construct conservative confidence sets for binary outcomes under explicit sensitivity and smoothness assumptions. We give closed forms for independent uniform scores and for correlated Gaussian scores. In one example, a true risk reduction of four percentage points appears as a one-point increase. Simulations confirm the formulas. Narrower bandwidths, robust bias correction, and machine-learning covariate adjustment do not remove the bias. A fuzzy RD on the first score recovers the causal effect when early and late treatment have the same effect. Our results concern how the running variable is built. We do not assess any published clinical study.

[25] arXiv:2611.20806 (cross-list from eess.SP) [pdf, html, other]
Title: Do Hierarchical Networks Actually Learn? Insights from Word-of-Mouth Estimation
Andrea Da Col, Cristian R. Rojas
Comments: Submitted for publication to the 21st IFAC Symposium on System Identification (SYSID2027)
Subjects: Signal Processing (eess.SP); Econometrics (econ.EM)

We study recursive estimation of a constant state vector by a hierarchical network of asymmetrically informed linear estimators, modeled as a directed tree. Only the leaves observe the regression signal, whereas every other estimator learns by word of mouth from noisy signal estimates relayed by its children. Authority within the hierarchy is expressed through common prior information: after each round, every estimator adopts the estimate and error covariance broadcast by the root. The more accurate this common information, the larger the effective noise variance on local observations, which makes estimation intrinsically self-defeating. The resulting performance is the subject of this paper. We show that recursive least-squares (RLS) estimators do learn the state in mean square, albeit slowly. Under strong persistency of excitation, the broadcast error covariance decays polynomially in time, at a rate that we characterize by upper and lower bounds of matching order. The exponent is determined by the smallest number of relays between a leaf and the root, and it decreases exponentially with this number.

[26] arXiv:2611.21113 (cross-list from physics.soc-ph) [pdf, html, other]
Title: Intransigence or compromise: concatenated opinion dynamics with voting feedback
Lingfei Wang, Johan Swartling Sennhed, David Ohlin, Emma Tegling
Subjects: Physics and Society (physics.soc-ph); Theoretical Economics (econ.TH); Systems and Control (eess.SY); Adaptation and Self-Organizing Systems (nlin.AO)

In this paper, we study opinion evolution in a sequence of Friedkin-Johnsen (FJ) models, concatenated so that the final opinions of each model serve as the initial opinions of the next. Each FJ model is followed by a vote, and each agent becomes partially or fully stubborn in the next model, depending on how close its opinion is to the voting result. With the state-dependent switching of stubbornness, the dynamics are not guaranteed to be contractive. We show that, over any fixed strongly connected network, the opinions always converge. A sufficient condition on the initial opinions is proposed, under which consensus is reached exponentially fast. For the complete graph, we further show that, if the agents' stubbornness levels do not differ too much, an agent that is partially stubborn in one FJ model can never become fully stubborn in a subsequent one: once it has started compromising, it keeps doing so.

[27] arXiv:2611.21207 (cross-list from stat.AP) [pdf, html, other]
Title: Who Is the Regression About? Auditing AI-Labeled Treatments When Effects Differ
Ayman Farahat
Subjects: Applications (stat.AP); Econometrics (econ.EM)

Researchers increasingly regress outcomes on treatments labeled by large language models. When effects differ, OLS averages the effect on the treated (ATT) and on the untreated (ATU), weighting the smaller group more. A label's bias then has two parts: classical attenuation from misclassification, and a selective term when its errors are correlated with the outcome. In LinkedIn postings, the model misses better-paid remote jobs: a precision check passes, yet the remote pay gap is understated by nearly 40 percent, a bias that more data cannot remove. A stratified, blind audit of both labeled groups corrects it and recovers both effects.

[28] arXiv:2611.21262 (cross-list from q-fin.RM) [pdf, html, other]
Title: Strategic Repayment and Latent Fragility in Financial Networks: Why Banks Withhold Payable Liquidity
Hicham El Kaissi
Subjects: Risk Management (q-fin.RM); Computer Science and Game Theory (cs.GT); Theoretical Economics (econ.TH); Mathematical Finance (q-fin.MF)

A network in which every bank pays in full can nevertheless admit a self-fulfilling settlement freeze under the same balance-sheet fundamentals. Thus, observing full repayment does not establish the uniqueness or robustness of the payment equilibrium. We derive this result from a precautionary-liquidity choice: banks may delay otherwise feasible payments to insure an interim funding need, trading a positive settlement-discipline cost against the marginal value of retained cash. The optimal reserve is a critical quantile of future liquidity requirements, and the repayment rule nests Eisenberg-Noe maximal feasible clearing exactly. Counterparty underpayment both reduces current cash and raises desired reserves. We characterise network amplification and the conditions for coexistence, including a homogeneous two-bank GPD frontier in which interbank exposure and maximum funding pass-through enter through one dimensionless index. A two-date settlement accounting rule distinguishes payment timing from ultimate default. The main result concerns equilibrium coexistence; the conditional best-response adjustment diagnostic does not assign selection probabilities.

Replacement submissions (showing 34 of 34 entries)

[29] arXiv:2203.11820 (replaced) [pdf, html, other]
Title: Dealing with Logs and Zeros in Regression Models
David Benatia, Christophe Bellego, Louis Pape
Comments: Substantially revised version. 55 pages, including main Appendices A-D; separate 126-page Supplemental Appendices S1-S7 available as an ancillary PDF. Software and replication links are provided on the first page: this https URL and this https URL
Subjects: Econometrics (econ.EM); Methodology (stat.ME)

Log regression is widely used to express effects in percentage terms, but zeros, heterogeneity and retransformation pose distinct difficulties. We unify the interpretation, identification and estimation of proportional effects through weighted individual percentage changes. We connect these parameters to pseudo-maximum likelihood estimation, characterize finite-sample differences between consistent estimators, and establish relationships with log regression. We develop a new shifted-log transformation that yields unit-invariant proportional-effect parameters and use it to construct iterated ordinary least squares (IOLS), an algorithm that estimates power-mean models through repeated OLS regressions. We prove global numerical convergence, consistency and asymptotic normality, and extend the approach to endogenous regressors. Furthermore, we develop fixed-effect cancellation and bias-correction methods that address incidental-parameter bias in common panel settings under general working variances, extending beyond the familiar Poisson-PML case. Finally, simulations document numerical reliability, competitive speed and finite-sample performance, while an application to economic development illustrates their empirical relevance.

[30] arXiv:2304.07855 (replaced) [pdf, html, other]
Title: Selective Inference for Survey-Weighted Generalized Linear Models: Growing Dimension and Average Marginal Effects
Joann Jasiak, Purevdorj Tuvaandorj
Comments: 54 pages, 4 tables
Subjects: Econometrics (econ.EM); Statistics Theory (math.ST); Machine Learning (stat.ML)

This paper extends the $\ell_1$-penalized likelihood Selective Inference method to survey settings. We establish its asymptotic validity for generalized linear models with an increasing number of covariates and survey weights, allowing inference on pseudo-true parameters of possibly misspecified selected models. Moreover, we generalize the method to facilitate inference on nonlinear functionals, such as the average marginal effect in survey-weighted logit models. The effectiveness of the proposed approach is illustrated using simulated data and the 2020 Canadian Internet Use Survey.

[31] arXiv:2307.06190 (replaced) [pdf, html, other]
Title: Stationarity with Occasionally Binding Constraints
James A. Duffy, Sophocles Mavroeidis, Sam Wycherley
Comments: ii + 36 pp., 4 figures. Split from arXiv:2211.09604v1. Author accepted manuscript
Subjects: Econometrics (econ.EM); Statistics Theory (math.ST)

This paper studies a class of multivariate threshold autoregressive models, known as censored and kinked structural vector autoregressions (CKSVAR), which are notably able to accommodate series that are subject to occasionally binding constraints. We develop a set of sufficient conditions for the processes generated by a CKSVAR to be stationary, ergodic, and weakly dependent. Our conditions relate directly to the stability of the deterministic part of the model, and are therefore less conservative than those typically available for general vector threshold autoregressive (VTAR) models. Though our criteria refer to quantities, such as refinements of the joint spectral radius, that cannot feasibly be computed exactly, they can be approximated numerically to a high degree of precision.

[32] arXiv:2405.11142 (replaced) [pdf, html, other]
Title: Pricing Expertise with Information Spillovers
Klajdi Hoxha
Subjects: Theoretical Economics (econ.TH)

Why do experts charge value-based fees for some services and flat fees for others? For example, an analyst charges the same fee for a stock report whether the client invests a thousand dollars or a million, while an investment bank charges a higher fee for a fairness opinion on a large merger than on a small one. We consider the problem of an expert who sells expertise to a client whose private type determines his project and how much he has at stake. Because project states are correlated, cheaper expertise about a related project can substitute for expertise about his own. We develop a Markov model of information spillovers and provide an explicit characterization of the optimal menu. With perfect spillovers, as with a stock report, a common flat fee for full information is optimal. With sufficiently weak spillovers, as with a fairness opinion, each client receives full information at a value-based fee equal to his willingness to pay. Between these extremes, low types can receive distorted expertise at their full willingness to pay, while middle and high types receive full information. Middle types pay a common base fee plus a project-specific fee that rises with what is at stake, and high types pay a value-based fee equal to their willingness to pay. Because higher types lose more by buying expertise about another project, the project-specific fee can rise faster than willingness to pay, so client surplus can peak among middle types and fall to zero at the top.

[33] arXiv:2501.15422 (replaced) [pdf, html, other]
Title: TTC Domains
Sumit Goel, Yuki Tamura
Comments: 25 pages, 3 tables. v5: proves that the top-two condition is also necessary, yielding a full characterization of TTC domains
Subjects: Theoretical Economics (econ.TH); Computer Science and Game Theory (cs.GT)

The Top Trading Cycles (TTC) rule is fundamental to object reallocation: it is the only rule that is individually rational, pair efficient, and strategyproof on the unrestricted domain. Despite its desirable properties, at some profiles, TTC precludes longer trading cycles that could lead to more appealing allocations. We explore restricted preference domains as an avenue for finding alternative rules satisfying the three axioms. Our main contribution is to introduce a richness condition on domains, the top-two condition, and show that it characterizes TTC domains: domains to which the above characterization of TTC extends. The condition requires that, within every subset of objects, any two objects that can each be ranked first can also be ranked as the top two, in both possible orders. For every domain that fails the top-two condition, we construct a rule distinct from TTC that is individually rational, Pareto efficient, and strategyproof, and that executes longer trading cycles that TTC precludes.

[34] arXiv:2506.13936 (replaced) [pdf, other]
Title: The Anatomy of Value Creation: Input-Output Linkages, Policy Shifts, and Economic Impacts
Sourish Dutta
Comments: Presented at the 13th Statistical Forum of the International Monetary Fund (IMF) in Washington, D.C. on November 19, 2025 | this https URL. Paper link... this https URL | Presentation Link... this https URL
Subjects: General Economics (econ.GN)

This paper describes the construction of high-resolution 7-digit product-level Supply-Use Tables (SUTs) and symmetric Input-Output Tables (IOTs) for the Indian economy from 2016-17 to 2022-23, primarily utilising microdata from the Annual Survey of Industries (ASI). It outlines the intricate methodology for generating comprehensive input and output flows, reconciling data from registered and unregistered manufacturing sectors using NPCMS-NIC concordance, and transforming SUTs into analytically robust IOTs via the Industry Technology Assumption. The paper demonstrates the application of this framework through an in-depth case study of India's mobile phone manufacturing sector (NPCMS 4722200), analysing its economic impact, particularly on Domestic Value Added (DVA) and employment. The findings reveal significant structural shifts, including output growth, import substitution, export surges, and evolving labour dynamics. These detailed tables represent a valuable advancement for analysing sectoral interdependencies, the efficacy of industrial policy, and enhancing India's National Accounts Statistics, resonating with the NSS's legacy of data-driven policymaking.

[35] arXiv:2507.14621 (replaced) [pdf, html, other]
Title: Testing Clustered Equal Predictive Ability with Unknown Clusters
Oguzhan Akgun, Alain Pirotte, Giovanni Urga, Zhenlin Yang
Subjects: Econometrics (econ.EM)

We develop tests of clustered equal predictive ability (C-EPA) in panels where the clusters are unknown and estimated by Panel Kmeans. Since the hypotheses depend on data-driven clusters, we adopt a selective conditional inference framework. We show that the square root of a Wald statistic for the equality of two estimated cluster centers is asymptotically a truncated $\chi$ variable, conditionally on the selected clusters and their number, with multiple random initializations, and that the truncation set is a finite union of intervals that can be computed exactly. Unlike existing methods of selective inference for Kmeans, this holds under serial and cross-sectional dependence and heteroskedasticity. For the C-EPA hypothesis, we combine the pairwise $p$-values with a test of overall EPA. In Monte Carlo experiments, the proposed tests control size where naive tests severely over-reject. An application to exchange rate forecasting with time series and machine learning methods illustrates the procedure.

[36] arXiv:2512.03930 (replaced) [pdf, html, other]
Title: A choice-based axiomatization of Nash equilibrium
Michele Crescenzi
Subjects: Theoretical Economics (econ.TH); Computer Science and Game Theory (cs.GT)

An axiomatization of Nash equilibrium is provided for games of complete information in normal form. The Nash equilibrium correspondence is shown to be the unique solution concept that satisfies four simple and intuitive axioms. Two axioms reflect a mild form of individual rationality: one states that any profile of dominant strategies in a game must be a solution, whereas the other says that any solution to a game should not depend on strictly dominated strategies. The remaining two axioms are adaptations to non-cooperative games of standard choice-theoretic axioms about contraction and expansion consistency.

[37] arXiv:2602.09189 (replaced) [pdf, html, other]
Title: Affirmative Action in India with Hierarchical Reservations
Orhan Aygün, Bertan Turhan
Comments: arXiv admin note: text overlap with arXiv:2310.02660
Subjects: Theoretical Economics (econ.TH)

We study the design of affirmative action mechanisms, focusing on India's system of horizontal reservations within vertical reservations. We consider a one-to-all policy, under which an applicant eligible for multiple horizontal categories simultaneously contributes toward each category's minimum guarantee. When horizontal reservations have a hierarchical structure, we introduce a choice rule that selects the unique merit-optimal set of applicants subject to these guaranties, satisfies the substitutes condition and size monotonicity. In a matching-with-contracts framework, we show that the generalized deferred acceptance mechanism is stable, strategy-proof, and eliminates justified envy under aggregate choice rules using the hierarchical choice rule in each vertical category. We establish a fundamental trade-off between merit and affirmative action between the one-to-all and one-to-one counting policies.

[38] arXiv:2602.09490 (replaced) [pdf, html, other]
Title: Robust Trust
Piotr Dworczak, Alex Smolin
Subjects: Theoretical Economics (econ.TH); Artificial Intelligence (cs.AI); Computer Science and Game Theory (cs.GT)

An agent chooses an action based on her private information and a recommendation from an informed but potentially misaligned adviser. With a known probability, the adviser truthfully reports his signal; with the remaining probability, he can send any message. We characterize robustly optimal decision rules that maximize the agent's worst-case expected payoff. Every optimal rule is equivalent to a trust-region policy in belief space: the adviser's reported beliefs are taken at face value if they fall within the trust region but are otherwise projected onto the trust region's boundary. We derive alignment thresholds above which advice is strictly valuable, explicitly characterize the solution in binary-state environments, and develop a procedure for computing the trust region numerically.

[39] arXiv:2604.00468 (replaced) [pdf, html, other]
Title: When AI Improves Answers but Slows Knowledge Creation: Matching and Dynamic Knowledge Creation in Digital Public Goods
Keh-Kuan Sun
Subjects: General Economics (econ.GN)

Generative AI helps users solve problems more efficiently, but can also replace exchanges between people with private conversations with chatbots. When more problems are solved privately, fewer problem-solving episodes are converted into reusable public knowledge for others. We build a dynamic model of public good provision where agents post and answer questions on a public platform. AI can shrink the platform's archive through two margins. Private diversion can reduce the posted volume of knowledge-enhancing questions (the flow margin), while contributor exit can lower conditional resolution (the resolution margin). Low-archive traps can arise with or without AI, because a thin archive makes public answering too costly to sustain; AI can narrow the viable region. The model gives conditional empirical predictions that distinguish demand-side diversion from supply-side pool thinning under explicit identifying restrictions. Logging private solutions adds archive inflow, while rewards for answering strengthen contribution incentives.

[40] arXiv:2606.04356 (replaced) [pdf, html, other]
Title: Sequential algorithm for structural estimations with equilibrium constraints
Takeshi Fukasawa
Subjects: Econometrics (econ.EM)

This study examines sequential algorithms with the Zero Jacobian Property (ZJP) for estimating structural models subject to equilibrium constraints. For Maximum Likelihood Estimation (MLE) and the Generalized Method of Moments (GMM), the current study shows that these algorithms attain fast (near-quadratic) local convergence in large samples to the solution of the constrained optimization problem. If consistent initial estimates of the parameters satisfying a suitable rate condition are available, the algorithms yield an asymptotically efficient estimator even after one iteration.
It then proposes a novel algorithm called the Sequential Linearly Constrained (SLC) algorithm, which is applicable to a broader class of structural models than existing methods. A key advantage of the SLC algorithm is that it can be implemented without explicitly computing the Jacobian of the equilibrium constraints and can be several times faster than the Nested Fixed Point (NFXP) approach. The current study illustrates its performance through two numerical experiments: a dynamic discrete game with time-varying unobserved heterogeneity and a dynamic demand model.

[41] arXiv:2606.12585 (replaced) [pdf, html, other]
Title: Revisiting the ABCs of Working with AI: Evidence from Radiologists
Daniel Martin
Subjects: General Economics (econ.GN); Human-Computer Interaction (cs.HC)

I examine the relationship between the benefits of working with AI and two baseline traits - ability and belief calibration - among all 227 radiologists available in Collab-CXR, a dataset from an experiment where AI assistance was provided during X-ray diagnosis. When analyzed jointly, both lower unassisted accuracy and smaller miscalibration, measured by the absolute gap between average confidence and accuracy, are associated with greater improvements in accuracy with AI. In addition, the association between unassisted accuracy and improvement with AI becomes more negative and statistically significant when controlling for calibration.

[42] arXiv:2607.00504 (replaced) [pdf, html, other]
Title: How optimistic inflow forecasts distort dispatch, prices, and contracts in hydro-dominated power systems: evidence from Brazil
Arthur Brigatto, Alexandre Street, Joaquim Dias Garcia
Subjects: General Economics (econ.GN); Systems and Control (eess.SY)

Centralized hydrothermal planning models determine generation schedules and electricity spot prices based on inflow forecasts in audited-cost power systems, such as those prevalent in Latin America, and provide operational benchmarks and decision support in hydro-dominated competitive electricity markets. Consequently, biased forecasts can propagate directly into both operational decisions and market outcomes. This paper studies how persistent optimistic inflow forecast bias can propagate through a hydrothermal power system, from system operation to market outcomes. For a stylized hydrothermal model, we show that optimistic bias weakly reduces water values and weakly increases first-stage hydro discharge relative to the unbiased optimum, thereby lowering reservoir storage and postponing thermal commitment. Using official Brazilian planning and operational data, we provide empirical evidence consistent with this mechanism. We then conduct controlled SDDP experiments to compare policies trained under biased and bias-corrected inflow-forecast processes, evaluating both under the same bias-corrected inflow scenarios. The policy based on biased forecasts produces lower reservoir levels, delayed dry-season thermal dispatch, sharper spot-price peaks, higher reliability risk, and higher expected operating costs. Finally, a stylized contracting exercise indicates that these distortions increase the price-quantity risk of a representative hydropower producer and reduce its willingness-to-contract. The results indicate that inflow-forecast bias can be a source of operational inefficiency, reliability risk, and distorted market incentives in hydro-dominated power systems. The insights and policy implications drawn in this paper may be relevant to hydro-dominated systems and electricity markets that are increasingly reliant on energy storage.

[43] arXiv:2607.05882 (replaced) [pdf, html, other]
Title: Predictive Uncertainty with Revised Economic Data
Yizhou (Kyle)Kuang
Comments: 49 pages
Subjects: Econometrics (econ.EM)

Data revisions create uncertainty that outlives the first publication of an economic indicator. We develop a framework that turns an existing point forecast of the first release into a predictive distribution for a later release. The framework combines forecast-error records with official-release histories, including periods without recorded forecasts. We establish identification conditions and, from the component distributions, construct sharp bounds for directional probabilities and dependence-robust prediction intervals. We also show when additional revision records strictly improve estimation precision for a fixed rare-tail bound. Monte Carlo simulations quantify the predictive gains from longer release histories. Across four US inflation series, full histories improve interval scores relative to forecast-matched histories, with the largest gains in PCE. The conditional procedures lower pooled interval scores by 27--28 percent relative to a joint stochastic-volatility benchmark. With recency weighting, they match its observed coverage using intervals 34--36 percent narrower.

[44] arXiv:2607.09990 (replaced) [pdf, html, other]
Title: Economic Power in International Trade
Ashwin Bhattathiripad, Vipin P Veetil
Subjects: General Economics (econ.GN); Social and Information Networks (cs.SI)

Economic power is a country's capacity to harm another more than itself by withdrawing from a trading relationship. This paper develops a general equilibrium model of the world trade network to measure that power through counterfactual sanctions. Using a world input-output table of eighty economies and fifty industries, we solve for $3{,}160$ such sanction equilibria, one for each bilateral severance of a trading relationship. We read the asymmetry between the two sides' losses as power. Our results show that mutual trade dependence is anything but mutual. The distribution of power in world trade is highly skewed. In the median trade relationship, one country inflicts on its trading partner more than four times the loss it bears. Only one pair in seven is close to balanced. The United States is more powerful than all seventy-nine of its partners and China against all but one. Economic power is not determined by the bilateral trade balance.

[45] arXiv:2608.02311 (replaced) [pdf, html, other]
Title: AI Governance for Institutional Readiness in Finance
Irene Aldridge, Steve Krawciw
Comments: 50 pages
Subjects: Econometrics (econ.EM); Risk Management (q-fin.RM); Statistical Finance (q-fin.ST)

Agentic AI is gaining acceptance in asset management, but governance has not kept pace: 88\% of surveyed finance professionals report no operational governance framework for agentic AI, and only 24 of 75 large U.S. money managers disclosing AI use in Form ADV filings report a formal governance policy. We argue this gap is architectural: governance built for static validation does not survive continuously retrained agentic policies. We propose a four-layer framework (Policy, Engineering, Composition, Systemic) grounded in two distinct kinds of evidence, kept explicitly separate: two calibrated synthetic illustrations (a regret-covariance drift monitor; a crowding simulation showing joint drawdown risk rising from 39.2\% to 79.3\%), and three real, documented cases (a deployed LLM-embedding trading strategy, a \$45 billion discretionary fund's forced-deleveraging blowup, and a tribunal ruling holding an airline liable for its chatbot). The synthetic examples demonstrate computability from observable data; the cases demonstrate that the failure modes are not hypothetical. We provide a 90-day implementation sequence spanning trading and payments/customer-facing systems.

[46] arXiv:2608.10294 (replaced) [pdf, html, other]
Title: Testing Sign Agreement
Deborah Kim
Subjects: Econometrics (econ.EM)

This article studies the problem of testing sign agreement among a finite number of parameters, with applications to dynamic treatment effects, instrument validity, and labor-market discrimination. We propose least favorable and two-step tests that accommodate correlated estimators. Both have parametric and nonparametric bootstrap implementations and uniform asymptotic validity. Although neither test dominates the other in general, local power analysis shows that the two-step test can gain power when several parameters provide strong directional information. Simulations show substantial power gains for the two-step test over the least favorable and competing procedures. Empirical applications examine unilateral divorce laws and the college-proximity instrument.

[47] arXiv:2609.09859 (replaced) [pdf, html, other]
Title: How an Economy Shrinks in Space: Concavity-on-Jobs and Upward Consolidation under Demographic Decline
Miki Ogawa, Tomoya Mori
Comments: 55 pages, 15 figures, 8 tables. The 95-page Online Appendix is included as an ancillary file (anc/online_appendix.pdf)
Subjects: General Economics (econ.GN)

When a country's population declines, the aggregate economy appears to contract on the intensive margin: industrial diversity intact, every industry a little smaller. At the regional level, contraction is uneven and takes the extensive form: entire industries disappear, one after another. The relevant unit is the city: industries are nested by size - the hierarchy property of industrial location - each viable only above a minimum population. Necessity industries' thresholds bunch at the low end, so a city's industry count - and its jobs - is sharply concave in size (concavity on jobs). A modest loss pushes a small city below many thresholds at once; a large core sheds a few specialized industries, one at a time. Lost industries consolidate upward to the next city large enough to host them; for the worker it means a step down to a lower-paid local job if one is left, and a fall to the bottom of the hierarchy if none is. To recover the lost wage, workers move up to the apex - the only city hosting the full industry range. In Japan - two decades ahead of the OECD, Tokyo at its apex - a wage regression in real, housing-inclusive wages, on worker-level panel data for the young workers who carry the migration, finds a Tokyo-bound migration incentive that varies across origins exactly as concavity on jobs predicts.

[48] arXiv:2609.12976 (replaced) [pdf, html, other]
Title: Complements or Substitutes? Technology Adoption and Clinical Care Utilization: Evidence from Automated Insulin Delivery
Moslem Rashidi, Cristina Ugolini, Gianluca Fiorentini
Subjects: General Economics (econ.GN)

Whether medical technology reduces or increases demand for professional care is central to understanding its effects on healthcare utilization and costs. We study automated insulin delivery (AID) adoption among 1,608 adults with type 1 diabetes treated in four specialist clinics of the Italian National Health Service, 283 of whom adopt AID. Because adoption is clinically targeted and staggered, we combine risk-set coarsened exact matching with a staggered event-study design. Matching retains 181 adopters with comparable untreated controls. We further adjust for differential pre-adoption trends by extrapolating the untreated trajectory in event time. After adjustment, AID adoption is associated with greater routine outpatient engagement. The estimated effect on the probability of a diabetologist visit is 16.3 percentage points one semester after adoption and 39.3 percentage points four semesters after adoption, relative to a 55.2% visit probability in the semester before adoption. HbA1c testing and a broader process-of-care index show similar positive patterns, although the evidence is less precise. The visit estimates remain sensitive to the trend-extrapolation assumption and should therefore be interpreted conditionally on that restriction. The findings are consistent with a task-based view in which automation substitutes for routine dosing decisions while complementing clinical labor through interpretive, adjustment, and supervisory tasks. Patient-facing automation may therefore reorganize rather than simply reduce healthcare use.

[49] arXiv:2609.31744 (replaced) [pdf, html, other]
Title: Hidden Eligibility and the Credibility of Adoption
Georgy Lukyanov, Mikhail Zolotarev
Subjects: Theoretical Economics (econ.TH)

Subsidies encourage adoption, but they also change what others can learn from it. We study a government that chooses the amount and coverage of an early-adoption rebate and the information released about eligibility. Both early and late adoption create positive externalities. When eligibility is hidden, expanding the programme makes adoption less informative. As the benefit of early adoption rises, the optimal programme jumps from a small one that preserves imitation to a larger one that gives it up. We show that partial disclosure can support intermediate coverage and strictly improve on both optimally chosen opacity and full disclosure. An optimal report certifies the eligibility of some users, leaving the remaining adoptions sufficiently persuasive to induce imitation. We characterise the conditions for this policy and compare coverage under the alternative reporting rules. Analytical extensions allow delivery failures, private information, and heterogeneous adoption costs. The discontinuous change in optimal coverage can persist even when cost heterogeneity makes welfare continuous in coverage.

[50] arXiv:2610.15786 (replaced) [pdf, html, other]
Title: Resource Allocation with Midterm Review
Yijun Liu
Subjects: Theoretical Economics (econ.TH)

A principal funds an agent's project over two periods, with a midterm review providing a public signal about project quality. The agent privately knows the quality and always prefers more funding, while the principal wants to match funding to quality. We identify two reversal phenomena in the optimal mechanism. First, the principal's expected payoff is strictly higher in the first period than in the second, even though she is better informed after the review. Second, higher-quality projects receive more funding initially but less funding following a bad review, and may receive less funding following a good review as well.

[51] arXiv:2610.34629 (replaced) [pdf, html, other]
Title: Bayesian Model Comparison for Nonlinear State Space Models without Posterior Sampling
Joshua C.C. Chan, Wei Zhang
Subjects: Econometrics (econ.EM)

Marginal likelihoods are central to Bayesian comparison of nonlinear state space models, but commonly used estimators require posterior simulation for each model, with computational costs that grow with the sample size $T$. We propose a deterministic estimator that requires no posterior simulation. A grid filter integrates out the latent states, and a Laplace approximation with a second-order correction integrates over the static parameters. Under derivative-growth and posterior-concentration conditions, the corrected Laplace approximation has error $O(T^{-2})$, and the size of the correction indicates the error that remains. In stochastic volatility experiments, the estimator is more accurate than Chib's method, the Gelfand--Dey estimator and importance sampling with a fitted proposal at equal computing time for $T\geq 2{,}000$. Applications to two exchange rates and to $T=19{,}266$ five-minute stock returns provide the strongest support for heavy-tailed observations and a transitory volatility factor.

[52] arXiv:2611.05608 (replaced) [pdf, html, other]
Title: Debiased Conditional Independence Tests for Causal Effects
Susan Athey, Maria Nareklishvili, Jann Spiess
Subjects: Econometrics (econ.EM)

We propose robust tests for causal effects on potentially high-dimensional outcomes where there may be selection into treatment. These tests measure conditional independence of outcomes from treatment assignment by asking whether treatment can be predicted from outcomes more accurately than from covariates alone. We design three such tests of conditional independence. First, we propose a conditional randomization test based on an inverse propensity weighted mean squared error test statistic. If the propensity score is known, the test controls size in finite samples. Second, we propose a double-residualization approach, where we partial out covariate effects from both treatment and outcome-based treatment predictions, and test whether the correlation between the residuals is zero. Third, we propose a test based on an augmented inverse propensity weighted estimator built around doubly robust, Neyman-orthogonal population moments that are necessary and sufficient for conditional independence. Simulations show that all three tests improve power over a joint Wald test based on a seemingly unrelated regression across many non-linear settings, particularly under heavy tails, outliers, mixtures, heterogeneous effects, and effects with opposite signs.

[53] arXiv:2611.05692 (replaced) [pdf, html, other]
Title: A General Theory of Dynamic Mechanism Design: Joint Optimization of Contracting, Information-Production Technology, and Disclosure
Guoqiang Tian
Comments: 73 pages
Subjects: Theoretical Economics (econ.TH)

We develop a general theory of dynamic mechanism design that jointly optimizes contracting, information-production technology, and disclosure across a broad class of economic environments with general allocation constraints. Hidden actions, endogenous information acquisition, learning, and disclosure interact through dynamic incentives. We accommodate correlated private information, interdependent preferences, and potentially infinite-dimensional allocations. We introduce joint perfect Bayesian equilibrium, establish the Dynamic Revelation Principle, and recursively characterize incentive compatibility. Under private values with informational interdependence, we characterize joint incentives through choice obedience, reporting-envelope first-order conditions, and global integral monotonicity. We derive an exact dynamic information-rent identity, dynamic revenue equivalence, and dynamic virtual surplus. Comparing first-best, two second-best, and third-best contracts distinguishes information rents, hidden-action implementation costs, and direct information-production expenditures, revealing nonadditive interactions of either sign and branched outcomes. Applications show Milgrom-Weber revenue-ranking reversal with endogenous acquisition, no efficient first-price equilibrium under Gaussian signals, and absent, interior, or full financial-contract disclosure.

[54] arXiv:2611.12565 (replaced) [pdf, other]
Title: That Time of the Year Again: Government Pay Revisions in the Aggregate Economy
Faisal Quaiyyum
Subjects: General Economics (econ.GN)

Public pay is set by law in nominal terms, so inflation erodes its real value between legal acts until the government revises it. I model each revision as the choice of a forward-looking government, shaped by two costs that push in opposite directions. A fixed cost of acting makes revisions large and rare, whereas a convex cost of financing the increase at once makes them small and frequent. Through the budget, each revision also moves debt and the risk of fiscal stress under which the next one is made. In the documented pay decisions of eleven countries, pay set at discretion erodes against prices and private earnings and is restored in discrete steps, every cut was decided during market stress or an adjustment program, and court rulings drove a wedge between the cuts legislated, the cuts paid and the cuts that lasted. Estimated on the legal acts, the model reproduces the documented paths of pay in most countries when data and simulations are measured the same way. In Portugal's estimated model, the savings that the Treasury requires once spreads rise produce the cuts, the debt overhang holds pay down after the crisis, and a court that strikes cuts after they take effect invites cuts it will strike. Review of cuts before they took effect would have lowered the social cost of Portugal's crisis relative to review after the fact, and partial indexation would have paid in Bangladesh unless its automatic increases were costly to finance.

[55] arXiv:2208.13323 (replaced) [pdf, html, other]
Title: Safe Policy Learning under Regression Discontinuity Designs with Multiple Cutoffs
Yi Zhang, Eli Ben-Michael, Kosuke Imai
Subjects: Methodology (stat.ME); Econometrics (econ.EM)

We study policy learning under the regression discontinuity (RD) design, where deterministic treatment assignment requires extrapolation to identify improved treatment cutoffs. We develop a robust optimization approach that leverages the multiple cutoffs common in RD settings to minimize worst-case welfare loss relative to the status quo policy. Under the maintained assumptions, the resulting cutoffs guarantee overall welfare at least as high as that under the existing policy. We also establish oracle-gap bounds for the learned policy. Finally, we illustrate the proposed methodology through simulation studies and apply it to data from an educational assistance program in Colombia.

[56] arXiv:2406.09732 (replaced) [pdf, html, other]
Title: Finding pure Nash equilibria in large random games
Andrea Collevecchio, Tuan-Minh Nguyen, Ziwen Zhong
Comments: 28 pages, 4 figures, 1 table. To appear in Mathematics of Operations Research
Subjects: Probability (math.PR); Computer Science and Game Theory (cs.GT); Theoretical Economics (econ.TH)

Best Response Dynamics (BRD) constitute a family of strategy update rules designed to identify Pure Nash Equilibria (PNEs) in strategic games. In each iteration, a player is selected uniformly at random and adopts a best response strategy, maximising their payoff given the current strategies of the other players. If no improvement is possible for the selected player, another is chosen. The process terminates upon reaching a state where no player wishes to deviate, corresponding to a PNE. However, convergence is not guaranteed as either PNEs may not exist, or the dynamics may become confined within substructures of the game, which we call traps, that admit no PNEs. We investigate this behaviour in the context of random games with $N$ players, each choosing between two actions and receiving independently and identically distributed payoffs. The payoff distribution is permitted to have atoms, allowing for the possibility of ties. The induced dynamics define a class of random walks on the $N$-dimensional hypercube, where the landscape includes two types of absorbing sets: PNEs and traps. Our analysis encompasses BRD, simple random walks, and a broad class of nearest-neighbour processes. We prove that, with high probability, these dynamics reach a PNE before encountering any trap. Our results complete and generalise the ones by Amiet, Collevecchio, Scarsini and Zhong (2021), using a different approach.

[57] arXiv:2410.05634 (replaced) [pdf, html, other]
Title: Identification and estimation for matrix time series CP-factor models
Jinyuan Chang, Yue Du, Guanglin Huang, Qiwei Yao
Journal-ref: Annals of Statistics 2026, Vol. 54, pp. 1372-1397
Subjects: Methodology (stat.ME); Machine Learning (cs.LG); Econometrics (econ.EM)

We propose a new method for identifying and estimating the CP-factor models for matrix time series. Unlike the generalized eigenanalysis-based method of Chang et al. (2023) for which the convergence rates of the associated estimators may suffer from small eigengaps as the asymptotic theory is based on some matrix perturbation analysis, the proposed new method enjoys faster convergence rates which are free from any eigengaps. It achieves this by turning the problem into a joint diagonalization of several matrices whose elements are determined by a basis of a linear system, and by choosing the basis carefully to avoid near co-linearity (see Proposition 5 and Section 4.3). Furthermore, unlike the generalized eigenanalysis-based method which requires the two factor loading matrices to be full-ranked, the proposed new method can handle rank-deficient factor loading matrices. Illustration with both simulated and real matrix time series data shows the advantages of the proposed new method.

[58] arXiv:2512.09337 (replaced) [pdf, html, other]
Title: Balancing Weights for Causal Mediation Analysis
Kentaro Kawato
Comments: 93 pages, 2 figures, 5 tables
Subjects: Methodology (stat.ME); Econometrics (econ.EM)

Causal mediation analysis decomposes the average treatment effect into components that operate through intermediate variables and a component that does not. This paper studies weighting estimation of generalized mediation functionals, whose contrasts include natural direct and indirect effects, for any fixed finite number of causally ordered mediators. Inverse propensity score weights are vulnerable to (a) weight instability, because they are products of estimated probability ratios, and (b) finite-sample covariate and mediator imbalance, which can bias the estimator. We instead propose sequential balancing weights, which address (a) by minimizing weight dispersion and (b) by imposing exact balance. Stage one balances covariates between a treatment group and the full sample. Each later stage adds the next mediator and balances a treatment group against the preceding weighted group. A finite-sample bias decomposition of the augmented weighting estimator determines these variables and targets. When exact balance cannot be attained in a small sample, the procedure relaxes the balancing conditions by tolerances selected without using the outcome. Under the stated assumptions, the estimated weights converge at every stage to their population inverse-probability counterparts, with errors from preceding stages entering additively. We also establish asymptotic normality and semiparametric efficiency of the augmented weighting estimator, together with conditions for consistent variance estimation. In the main simulation design with two ordered mediators, the proposed weights outperform the compared methods in bias, root mean squared error, and coverage. In a small media-framing experiment with two ordered mediators, the proposed weights reduce the imbalance left by propensity-score weighting.

[59] arXiv:2512.18515 (replaced) [pdf, html, other]
Title: The Narrow Corridor of Stable Solutions in an Extended Osipov--Lanchester Model with Constant Total Population
Sergey Salishev
Comments: 20 pages, 1 figure. Submitted version. This preprint has not undergone peer review or any post-submission improvements or corrections. A revised version has been accepted for publication in Operations Research Forum
Subjects: Dynamical Systems (math.DS); General Economics (econ.GN); Optimization and Control (math.OC)

This paper studies a modification of the classical Osipov-Lanchester model in which the total population of the two sides, $N=R+B$, stays constant. In this model the ratio $y=R/B$ obeys the Riccati equation $\dot y=\alpha y^2-\beta$, which is solvable in elementary functions for all values of the parameters. The main result is a complete classification of the stable regimes. An asymptotically stable invariant set in the quadrant $R,B\ge 0$ exists only in three cases: $\alpha<0,\ \beta<0$ gives a stable interior equilibrium; $\alpha=0,\ \beta<0$ makes the boundary $B=0$ stable; $\alpha<0,\ \beta=0$ makes the boundary $R=0$ stable. If $\alpha>0$ or $\beta>0$, every solution except one unstable equilibrium reaches a boundary of the quadrant in finite time, where the model stops being applicable. If the parameters vary in time, a fixed distance from the boundaries is kept when $\alpha(t)$ and $\beta(t)$ obey buffer dynamics that push them away from zero. Finally, in the case $\alpha<0,\ \beta<0$ the original system grows exponentially, while in the two degenerate cases it grows at most linearly. Growth demanded of a given cost, the sum of the two channels, bounds their imbalance and, with the perturbation bound, confines the parameters to an explicit wedge about $|\alpha|=|\beta|$ truncated near the origin: the narrow corridor of the model.

[60] arXiv:2610.23917 (replaced) [pdf, html, other]
Title: Announced Breaks Separate Conformal Reliability from Frequency Calibration
Karl Li
Comments: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026). 41 pages, 7 figures
Subjects: Machine Learning (cs.LG); Econometrics (econ.EM); Machine Learning (stat.ML)

When the process behind a time series changes on an announced date, like Australia's switch to five minute electricity settlement on October 1, 2021, intervals calibrated before it can stop covering at their stated level. Whether a forecaster needs the next interval to cover or a finished period's share of misses on target decides the restart. We recompute the next conformal interval from post-break forecast errors, and for n independent ones from the same continuous law, its expected coverage error is $O(n^{-1/2})$, the optimal order in a fixed class among methods told the date. For a finished period's share of misses to match its target, we restart adaptive conformal inference (ACI), which on bounded errors brings that share within $O(n^{-1})$ of target if its corrections at a lower limit stay bounded in total, for that period only. After a simulated mean jump of 5.0 standard deviations, recomputed 90% intervals lie an expected 3.5 percentage points from target at the 50th forecast, a third closer than the 5.2 of the best of five adaptive methods. Of 9 announced changes in Texas, Australian and PJM power markets and a Federal Reserve decision, only Australia's switch moves the errors enough for a 5% test at 80% power to resolve a gain from restarting ACI, and dates help restarted ACI only after large changes.

[61] arXiv:2610.28907 (replaced) [pdf, html, other]
Title: JaxMARL-BC: Multi-Agent Reinforcement Learning for Macroeconomic Modelling, in Seconds
Daniele Montagnani, Andrea Coletta, Antonietta Mira, Aldo Glielmo
Comments: 9 pages, 7 figures
Journal-ref: Proceedings of the 7th ACM International Conference on AI in Finance (ICAIF '26), Milan, Italy, 2026
Subjects: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); General Economics (econ.GN)

Multi-agent reinforcement learning (MARL) offers a principled way of populating simulated economies with many heterogeneous decision makers. Its adoption, pioneered in finance, has nonetheless been slow in macroeconomics because of two main limitations: the lack of a shared frame of reference connecting MARL systems to dynamic stochastic general equilibrium (DSGE) models the profession already trusts, and the very long training times. The first limitation was recently addressed by the Multi-Agent Reinforcement Learning Business Cycle (MARL-BC), which places deep RL households in a real business cycle economy and reduces to classic DSGE models in specific limits. Here we address the second limitation. We present JaxMARL-BC, a JAX-accelerated simulator that trains MARL-BC economies in seconds (between 75x and 724x faster than the reference implementation), and scales to 200,000 independent learning households. We verify its correctness against the known DSGE limits, characterise its scaling through a simple cost law, and demonstrate its value by a proof-of-concept calibration of an extended model with heterogeneous returns on capital, in a two-hour experiment that would previously have taken over a month. We release JaxMARL-BC as open source.

[62] arXiv:2610.36176 (replaced) [pdf, html, other]
Title: Probabilistic Causal Panel Analysis
Carl Bonander, Qinyun Lin, Ulf Strömberg, Marta Blangiardo
Subjects: Methodology (stat.ME); Econometrics (econ.EM)

Many causal panel estimators construct counterfactuals using weights or latent structure estimated directly from observed outcomes, making their estimates sensitive to observation noise. Bayesian hierarchical models can account for this noise and pool information across related observations. Existing Bayesian approaches to causal panel analysis, however, typically specify counterfactual imputation within the probabilistic model, so changing estimators may require a new formulation. We develop a framework that learns the latent untreated panel with a probabilistic model and applies established causal estimators to its posterior draws. The model accommodates non-Gaussian observations and structured pooling, while re-estimation across draws propagates uncertainty through counterfactual imputation. We establish conditions for posterior consistency and characterize how pooling affects imputation risk. Under correct posterior specification, we show that the excess Bayes risk of our approach is bounded by the causal estimator's expected squared imputation error when given the true latent outcomes in observed untreated cells. We also develop a fast cross-validation approach for model selection and calibration of predictive uncertainty for treatment effects. We use simulations to examine when pooling helps or harms imputation and cross-validation to assess whether probabilistic modeling improves accuracy in multiple empirical applications.

Total of 62 entries
Showing up to 2000 entries per page: fewer | more | all
We gratefully acknowledge support from our major funders, member institutions, , and all contributors.
About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab)