Pith. sign in

REVIEW 4 major objections 4 minor 9 cited by

LLM Economist: Large Population Models and Mechanism Design in Multi-Agent Generative Simulacra

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read An LLM-based planner, optimized only through text prompts, converges near optimal tax schedules in a simulated 100-agent economy, beating the U.S. statutory baseline.

desk verdict A genuinely novel LLM simulacra framework whose headline claim is contradicted by its own experiments: the planner underperforms a Saez baseline that is itself seeded from the LLM's solution. read the letter →

arxiv 2507.15815 v1 pith:LTUFQW2K submitted 2025-07-21 cs.MA cs.LG

Seth Karten , Wenzhe Li , Zihan Ding , Samuel Kleiner , Yu Bai , Chi Jin This is my paper
classification cs.MAcs.LG
keywords in-contextreinforcementlearningmechanismdesignoptimaltaxationStackelbergequilibriumagent-basedsimulationlargelanguagemodelagentssocialwelfaregenerativesimulacra
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that a society of language-model agents can be used to design tax policy before any real policy is enacted. It builds a two-tier economy in which a planner agent proposes marginal tax brackets and one hundred worker agents, each with a distinct persona and skill level, respond by choosing how many hours to work. Both sides improve through in-context reinforcement learning: the agents read their own histories and rewards in the prompt and adjust their next text output, with no gradient updates. The paper reports that the planner converges close to a Stackelberg equilibrium whose social welfare exceeds the statutory U.S. tax schedule and reaches the neighborhood of the Saez benchmark, the standard static optimal-tax solution. It also shows that letting workers periodically vote to replace the planner can either entrench a majority or raise welfare, depending on the population.

What carries the argument

The object that carries the argument is two-level in-context reinforcement learning inside a Stackelberg game. In-context reinforcement learning means the model updates its behavior from text prompts containing scalar rewards—here, social welfare for the planner and utility for each worker—rather than from weight updates. The planner acts on a piecewise-linear marginal tax schedule over seven federal brackets, the workers act on weekly labor hours, and a lump-sum rebate keeps the budget balanced. Two design choices make the equilibrium reachable: a long enough 'tax year' for workers to finish adapting before the planner moves, and prompt cues that first encourage exploration of new schedules and then lock in the best one. A Census-calibrated skill distribution and persona text supply the heterogeneity that makes the result a population model rather than a representative-agent model.

What would settle it

Take the same 100-agent setup and keep every prompt component except the utility numbers, utility formulas, and instructions to maximize utility; if average labor hours and social welfare barely move, the utility-maximization interpretation is unsupported. Separately, replace the planner's in-context search with random bracket shifts of the same magnitudes; if welfare is comparable, the language model's role in the reported gains is not established.

Watch

Extended reading notes

Core claim

The central claim is that purely text-prompted optimization can reproduce the classic equity-efficiency trade-off of optimal tax theory. The planner's prompt contains income and utility histograms, recent social-welfare values, and the best tax schedule seen so far; the worker prompts contain pre-tax income, post-tax income, marginal rate, rebate, and a history window. From these, the planner learns to shift seven piecewise-linear marginal rates and each worker learns a labor choice, and the pair is interpreted as a stationary Stackelberg equilibrium. The paper reports that this equilibrium reaches 90% of the best observed social welfare, lands within 10-35% of the Saez benchmark's welfare in the two evaluated scenarios, and improves welfare by 93% over the U.S. federal schedule in the seven-bracket bounded case. In the democratic variant, majority voting over planner candidates produces either stable majority exploitation in a three-agent society or nearly annual leadership turnover that sometimes outperforms static optimal taxation in the 100-agent society.

Load-bearing premise

The load-bearing assumption is that a worker-agent that is told to maximize its utility actually produces the labor choice that maximizes the mathematical utility in its prompt; if the model is only following the prompt's wording or examples, the converged tax schedule is not a genuine economic equilibrium.

Editorial extensions

If this is right

  • If the central claim is right, tax-policy experiments can be run on a text-prompted population before being tried on humans or deployed in automated economies.
  • The planner's search implicitly re-estimates the behavioral elasticities that the static Saez formula takes as fixed, which removes a known practical limitation of using Saez rates directly.
  • The voting results imply that the framework can evaluate not only tax schedules but also institutional rules such as election frequency and candidate platforms.
  • Because the same 8-billion-parameter model handles 100-agent runs and scales to 1,000 agents, the testbed is cheap enough to use for repeated policy audits.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • An extension the paper does not run would be a control planner that searches the same bracket space with random or grid disturbances; absent that baseline, some of the reported welfare gain could come from the search-and-replay scaffolding rather than from the language model's economic judgment.
  • The persona population is sampled from Census marginals rather than joint demographic distributions, so correlations such as age, occupation, and income are only approximated; re-running with a joint-calibrated sampler is a direct test of how much the equilibrium schedule depends on that approximation.
  • The democratic-voting result leaves open whether voters are persuaded by the substance of platforms or by prompt style; measuring vote outcomes against platform content versus platform tone would separate persuasion from policy preference.
  • A stronger claim, that the learned schedules transfer to a freshly sampled population, is not tested; showing transfer would turn the framework from a fitted simulacrum into a reusable mechanism-design engine.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper introduces the LLM Economist, a two-tier agent-based simulation in which LLM-based worker agents choose labor hours from persona-conditioned prompts and an LLM planner proposes piecewise-linear marginal tax schedules via in-context reinforcement learning. The authors claim that the planner converges to Stackelberg equilibria that improve aggregate social welfare relative to Saez-optimal schedules, and that periodic persona-level voting further improves outcomes. Experiments use Llama-3.1-8B with populations up to 100 agents and compare against statutory U.S. tax schedules and Saez-derived baselines.

Significance. If the claims were supported, the framework would be a valuable testbed for mechanism design with heterogeneous, bounded-rational agents, and the Census-calibrated population construction is a genuinely useful ingredient. The paper's strengths include the public code release, the GB2 fit to ACS data, and explicit ablations of prompt design and tax-year length. However, the central welfare comparison is invalid and contradicted by the paper's own figures, and the Stackelberg-equilibrium interpretation is not evidenced. The main evaluation therefore does not establish the paper's headline contribution.

major comments (4)
  1. [Abstract; Sec. 4.3, Fig. 5] The abstract's claim that the planner 'improves aggregate social welfare relative to Saez solutions' is contradicted by the manuscript's own results. In the seven-bracket bounded scenario, the grid-searched Saez schedule improves SWF by 10% over the LLM Economist policy (Fig. 5a); in the three-bracket isoelastic scenario, the Saez-regression schedule outperforms the LLM Economist (Fig. 5b). The Interpretation paragraph concedes that the planner 'lands within 10–35% of the Saez optimum' and that 'Saez retains an edge.' The headline claim should be corrected or replaced by experiments that actually show an improvement.
  2. [Sec. 4.3; App. D] The Saez baselines are not independent benchmarks. In the bounded case, the Saez schedule is obtained by perturbing the LLM Economist's terminal schedule and grid-searching; in the isoelastic case, the Saez regression is solved 'given a good starting point, the LLM Economist solution' (Sec. 4.3). Appendix D states that 'Saez tax rates require a solution from the LLM Economist to be locally perturbed before finding the optimal policy.' Because the elasticity is estimated from the same simulation, the comparison cannot establish that the LLM Economist improves on Saez; it only shows that a local search around the LLM's own solution can improve welfare.
  3. [Sec. 2, Sec. 3] The interpretation of the results as convergence to a Stackelberg equilibrium is not supported. Section 2 asserts that stationary optimal responses are 'without loss of generality' and defines equilibrium via exact maximization of the utilities in Eqs. (3)-(4) and Eq. (1). Section 3, however, describes workers and the planner as producing JSON text actions from prompts, with no evidence that these actions solve the stated maximization problems. Without a validation that worker labor choices and planner schedules actually maximize the relevant objectives, the paper cannot claim that the simulated trajectories approximate Stackelberg equilibria.
  4. [Sec. 4.4, Fig. 6] The voting experiments claim to reproduce 'tyranny of the masses' and welfare-enhancing turnover, but these claims are supported only by qualitative descriptions of two runs. Figure 6 shows no quantitative welfare measures, no statistical summary across seeds, and no comparison to a non-voting control. The statement that 'electoral exploration can outperform static optimal taxation' is therefore not established by the evidence presented.
minor comments (4)
  1. [Sec. 4.2] The text says 'the dissatisfaction penalty ψ appearing in Eq. (3)' but in Eq. (3) ψ is the labor-disutility scale; the dissatisfaction penalty is φ in Eq. (4).
  2. [Fig. 5 caption] The caption mentions 'grey not shown in the three-bracket case,' but no grey curve is identified for the seven-bracket case either; please clarify which baseline is orange and what the grey reference denotes.
  3. [Table 2a] The first two rows of Table 2a display '8 310' and '16 600,' which appear to be misaligned or missing separators; the total-step entries should be formatted consistently with the later rows.
  4. [Appendix B, Algorithm 1] The pseudocode refers to a variable 'two_timescale' that is not defined in the text; the planner-update interval should be specified explicitly or aligned with the tax-year length K.

Circularity Check

2 steps flagged · score 6.0 of 10

Saez baseline is built from the LLM's own solution, so the advertised improvement over Saez is partly a self-comparison; the paper's own figures show Saez variants beating the LLM.

  1. self definitional [Section 1 (Contribution ii) and Appendix D, first paragraph]
    "We demonstrate that the planner, optimizing in-context, converges to similar social welfare to optimal Saez [60] baselines (calculated based on our solutions). ... Thus, as noted in our experiments, Saez tax rates require a solution from the LLM Economist to be locally perturbed before finding the optimal policy."

    The Saez benchmark is constructed by locally perturbing the planner's own terminal schedule; the paper states that 'Saez tax rates require a solution from the LLM Economist to be locally perturbed.' The headline comparison 'planner ... improve[s] aggregate social welfare relative to Saez solutions' therefore reduces to checking that the planner's schedule is close to a locally optimized neighbor of itself, not to an independent, theory-derived optimum. In the seven-bracket bounded case the grid-searched 'Saez' schedule actually improves SWF by 10% over the LLM planner, so the advertised improvement is also contradicted by the paper's own numbers.

  2. fitted input called prediction [Section 4.3, Three-bracket isoelastic case]
    "Since isoelastic utility is purely rational, Saez can be solved analytically (given a good starting point, the LLM Economist solution). So we follow the Saez regression recipe: estimate elasticity from the perturbation and solve the log linear system per bracket."

    The elasticity parameter that enters the Saez formula is estimated from the same LLM-generated labor-supply responses used to evaluate the planner, and the 'analytically solved' Saez schedule is initialized at the LLM Economist's own solution. The Saez benchmark is therefore a regression fitted to the model's output, not a parameter-free external optimum. Claiming that the LLM 'approaches' or 'retains an edge' against this baseline is an in-sample comparison between the planner and a locally fitted version of itself.

full rationale

The core derivation, in-context RL for workers and planner and the GB2 ACS-calibrated population, is self-contained and not circular; the statutory U.S. schedule and ACS microdata are independent anchors. The circularity is concentrated in the Saez comparison advertised in the abstract and contribution (ii). The paper itself states that Saez baselines are 'calculated based on our solutions' and that 'Saez tax rates require a solution from the LLM Economist to be locally perturbed,' so the benchmark is a locally optimized version of the LLM's own terminal schedule. In the three-bracket case the elasticity is regressed from the same simulation's perturbation response and the 'analytical' solution is initialized at the LLM schedule. Consequently, the central 'improvement relative to Saez' claim partially reduces to the planner being compared with a neighbor of itself; it is not an independent falsification. The paper's own figures also undermine the abstract: the grid-searched Saez baseline beats the LLM by 10% in the seven-bracket case and the Saez-regression schedule outperforms it in the isoelastic case. No load-bearing self-citation chain or imported uniqueness theorem was found; the problem is benchmark construction, not citation practice. Score 6 reflects that the main advertised comparison is partly by construction, while the underlying ICRL and statutory-baseline results retain independent content.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The framework rests on the assumed utility structure, the unverified assumption that LLM text search optimizes those utilities, and a benchmark that is computed from the framework's own outputs.

free parameters (6)
  • η (relative risk aversion)
    Utility Eq. (3); value not reported in the paper but directly affects labor supply and welfare.
  • ψ (labor disutility scale)
    Utility Eq. (3); set by hand, value not reported.
  • δ (labor disutility exponent)
    Utility Eq. (3); set by hand, value not reported.
  • φ (dissatisfaction penalty) = calibrated so a one-bracket misalignment halves utility
    Eq. (4) and Section 4 setup; chosen by LLM/author, not derived from data.
  • GB2 skill distribution parameters = fit to ACS 2023 microdata
    Section 4; used to sample worker skills; fitted values not listed in the paper.
  • Saez elasticity e = estimated from simulation perturbations
    Section 4.3; used to construct the Saez baseline, estimated from the same LLM simulation.
assumptions (4)
  • domain assumption LLM agents maximize the stated utility and social-welfare objectives via in-context RL
    Core premise of the framework; Section 3 states the simulation 'realizes' the Stackelberg game, but no proof or measurement confirms the LLM optimizes Eqs. (1)-(4).
  • ad hoc to paper Stationary optimal responses exist and are reached without loss of generality
    Section 2 asserts WLOG stationarity for time-homogeneous additive objectives; this is nontrivial with LLM policies and changing tax schedules.
  • domain assumption Saez formula is a valid baseline despite its assumptions being violated
    Appendix D acknowledges Saez assumes independent elasticities and purely rational utility, both violated in the experiments, yet Saez is the main benchmark.
  • domain assumption Worker skills and personas approximate the US population
    Section 4 uses a GB2 fit to ACS 2023 and persona prompts from marginals; the paper admits joint demographic correlations are only approximated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LLM Economist: Large Population Models and Mechanism Design in Multi-Agent Generative Simulacra." pith.science (2026). https://pith.science/paper/LTUFQW2K

@misc{pith2026250715815,
  author       = {Pith},
  title        = {Pith review of: LLM Economist: Large Population Models and Mechanism Design in Multi-Agent Generative Simulacra},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LTUFQW2K}},
  note         = {Machine review of arXiv:2507.15815}
}
read the original abstract

We present the LLM Economist, a novel framework that uses agent-based modeling to design and assess economic policies in strategic environments with hierarchical decision-making. At the lower level, bounded rational worker agents -- instantiated as persona-conditioned prompts sampled from U.S. Census-calibrated income and demographic statistics -- choose labor supply to maximize text-based utility functions learned in-context. At the upper level, a planner agent employs in-context reinforcement learning to propose piecewise-linear marginal tax schedules anchored to the current U.S. federal brackets. This construction endows economic simulacra with three capabilities requisite for credible fiscal experimentation: (i) optimization of heterogeneous utilities, (ii) principled generation of large, demographically realistic agent populations, and (iii) mechanism design -- the ultimate nudging problem -- expressed entirely in natural language. Experiments with populations of up to one hundred interacting agents show that the planner converges near Stackelberg equilibria that improve aggregate social welfare relative to Saez solutions, while a periodic, persona-level voting procedure furthers these gains under decentralized governance. These results demonstrate that large language model-based agents can jointly model, simulate, and govern complex economic systems, providing a tractable test bed for policy evaluation at the societal scale to help build better civilizations.

Discussion (0). Sign in to comment.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Continual Harness: Online Adaptation for Self-Improving Foundation Agents

    cs.LG 2026-05 conditional novelty 8.0 of 10

    Continual Harness automates online self-improvement for foundation-model embodied agents by refining prompts, sub-agents, skills, and memory within one run, cutting button-press costs on Pokemon Red and Emerald and cl...

  2. Learn to Match: Two-Sided Matching with Temporally Extended Feedback

    cs.LG 2026-06 unverdicted novelty 7.0 of 10

    Learn2Match is a POMG-based MARL benchmark for two-sided matching with temporally extended feedback; independent PPO yields higher social welfare and lower regret than CA-ETC but higher information-friction loss.

  3. Healthcare Mechanisms from Policy-as-Code Search under Strategic Provider Response

    cs.AI 2026-05 unverdicted novelty 7.0 of 10

    LLM-guided program synthesis in a strategic-provider simulator yields payment rules that eliminate up-coding, halve rejections, and retain most baseline funds.

  4. Agent Bazaar: Enabling Economic Alignment in Multi-Agent Marketplaces

    cs.LG 2026-05 unverdicted novelty 7.0 of 10

    Agent Bazaar is a multi-agent simulation framework that identifies economic failure modes in LLM agents, proposes stabilizing harnesses, and shows that targeted RL training can produce a 9B model with superior economi...

  5. FlowSteer: Prompt-Only Workflow Steering Exposes Planning-Time Vulnerabilities in Multi-Agent LLM Systems

    cs.CR 2026-05 unverdicted novelty 7.0 of 10

    FlowSteer is a prompt-only attack that biases multi-agent LLM workflow planning to propagate malicious signals, raising success rates by up to 55%, with FlowGuard as an input-side defense reducing it by up to 34%.

  6. CoopEval: Benchmarking Cooperation-Sustaining Mechanisms and LLM Agents in Social Dilemmas

    cs.GT 2026-04 conditional novelty 6.5 of 10

    Contracting and mediation best restore cooperative equilibria among selfish LLM agents across four social dilemmas; evolutionary pressure further increases cooperation under these mechanisms.

  7. Predicting Decisions of AI Agents from Limited Interaction through Text-Tabular Modeling

    cs.LG 2026-05 unverdicted novelty 6.0 of 10

    A tabular foundation model with LLM-as-Observer features predicts AI agent decisions in controlled games, outperforming baselines by 4 AUC points and 14% lower error at K=16 interactions.

  8. CoopEval: Benchmarking Cooperation-Sustaining Mechanisms and LLM Agents in Social Dilemmas

    cs.GT 2026-04 unverdicted novelty 6.0 of 10

    Contracting and third-party mediation enable more cooperative outcomes among LLM agents in social dilemmas than repetition or reputation, with effectiveness increasing under evolutionary pressures.

  9. Evaluation of Agents under Simulated AI Marketplace Dynamics

    cs.IR 2026-04 unverdicted novelty 6.0 of 10

    Marketplace Evaluation uses repeated-interaction simulations to assess information access systems with marketplace-level metrics such as retention and market share that complement traditional accuracy measures.

Reference graph

Works this paper leans on

79 extracted references · 45 canonical work pages · cited by 8 Pith papers

  1. [1]

    A. AL, A. Ahn, N. Becker, S. Carroll, N. Christie, M. Cortes, A. Demirci, M. Du, F. Li, S. Luo, et al. Project sid: Many-agent simulations toward ai civilization.arXiv preprint arXiv:2411.00114, 2024. 5

    arXiv 2024

  2. [2]

    K. J. Arrow et al. Essays in the theory of risk-bearing, volume 121. North-Holland Amsterdam, 1974

    work page 1974

  3. [3]

    Y. Bai, C. Jin, H. Wang, and C. Xiong. Sample-efficient learning of stackelberg equilibria in general-sum games. Advances in Neural Information Processing Systems, 34:25799–25811, 2021

    work page 2021

  4. [4]

    Brero, A

    G. Brero, A. Eden, D. Chakrabarti, M. Gerstgrasser, A. Greenwald, V. Li, and D. C. Parkes. Stackelberg pomdp: A reinforcement learning approach for economic design.arXiv preprint arXiv:2210.03852, 2022

    arXiv 2022

  5. [5]

    Brero, E

    G. Brero, E. Mibuari, N. Lepore, and D. C. Parkes. Learning to mitigate ai collusion on economic platforms. Advances in Neural Information Processing Systems, 35:37892–37904, 2022

    work page 2022

  6. [6]

    T. B. Brown. Language models are few-shot learners.arXiv preprint arXiv:2005.14165,

    arXiv 2005

  7. [7]

    R. Chetty. Sufficient statistics for welfare analysis: A bridge between structural and reduced-form methods. Annu. Rev. Econ., 1(1):451–488, 2009

    work page 2009

  8. [8]

    A. Chopra. Large population models.arXiv preprint arXiv:2507.09901, 2025. 1

    work page Pith review arXiv 2025

Show all 79 references
  1. [9]

    Chopra, S

    A. Chopra, S. Kumar, N. Giray-Kuru, R. Raskar, and A. Quera-Bofarull. On the limits of agency in agent-based models.arXiv preprint arXiv:2409.10568, 2024. 5

    2024 arXiv

  2. [10]

    J. J. Chung. Money as simulacrum: The legal nature and reality of money.Hastings Bus. LJ, 5:109, 2009

    2009

  3. [11]

    X. Deng, Y. Gu, B. Zheng, S. Chen, S. Stevens, B. Wang, H. Sun, and Y. Su. Mind2web: Towards a generalist agent for the web.Advances in Neural Information Processing Systems, 36, 2024

    2024

  4. [12]

    P. A. Diamond and J. A. Mirrlees. Optimal taxation and public production i: Production efficiency. The American economic review, 61(1):8–27, 1971. 2, 2, 5

    1971

  5. [13]

    Y. Du, L. Han, M. Fang, J. Liu, T. Dai, and D. Tao. Liir: Learning individual intrinsic reward in multi-agent reinforcement learning.Advances in Neural Information Processing Systems, 32, 2019

    2019

  6. [14]

    Dubey, A

    A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Yang, A. Fan, et al. The llama 3 herd of models.arXiv preprint arXiv:2407.21783, 2024. 13 LLM Economist

    2024 arXiv

  7. [15]

    Duetting, V

    P. Duetting, V. Mirrokni, R. Paes Leme, H. Xu, and S. Zuo. Mechanism design for large language models. InProceedings of the ACM on Web Conference 2024, pages 144–155, 2024. 5

    2024

  8. [16]

    M. F. A. R. D. T. (FAIR)†, A. Bakhtin, N. Brown, E. Dinan, G. Farina, C. Flaherty, D. Fried, A. Goff, J. Gray, H. Hu, A. P. Jacob, M. Komeili, K. Konath, M. Kwon, A. Lerer, M. Lewis, A. H. Miller, S. Mitts, A. Renduchintala, S. Roller, D. Rowe, W. Shi, J. Spisak, A. Wei, D. Wu...

  9. [17]

    M. F. A. R. D. T. (FAIR)†, A. Bakhtin, N. Brown, E. Dinan, G. Farina, C. Flaherty, D. Fried, A. Goff, J. Gray, H. Hu, et al. Human-level play in the game of diplomacy by combining language models with strategic reasoning.Science, 378(6624):1067–1074, 2022

    2022

  10. [18]

    E. Farhi. Capital taxation and ownership when markets are incomplete.Journal of Political Economy, 118(5):908–948, 2010

    2010

  11. [19]

    X. Feng, Z. Wan, M. Wen, Y. Wen, W. Zhang, and J. Wang. Alphazero-like tree-search can guide large language model decoding and training.arXiv preprint arXiv:2309.17179, 2023

    2023 arXiv

  12. [20]

    X. Feng, Z. Wan, H. Fu, B. Liu, M. Yang, G. A. Koushik, Z. Hu, Y. Wen, and J. Wang. Natural language reinforcement learning.arXiv preprint arXiv:2411.14251, 2024. 5

    2024 arXiv

  13. [21]

    Fleurbaey

    M. Fleurbaey. Normative economics and economic justice. 2004

    2004

  14. [22]

    X. Gabaix. A behavioral new keynesian model. American Economic Review, 110(8): 2271–2327, 2020

    2020

  15. [23]

    S. Garg, D. Tsipras, P. S. Liang, and G. Valiant. What can transformers learn in-context? a case study of simple function classes.Advances in Neural Information Processing Systems, 35:30583–30598, 2022

    2022

  16. [24]

    S. Hao, Y. Gu, H. Ma, J. J. Hong, Z. Wang, D. Z. Wang, and Z. Hu. Reasoning with language model is planning with world model.arXiv preprint arXiv:2305.14992, 2023

    2023 arXiv

  17. [25]

    Hoderlein

    S. Hoderlein. Nonparametric demand systems and a heterogeneous population. Technical report, Working Paper, Uni Mannheim, 2004

    2004

  18. [26]

    M. Hong, H. Wai, Z. Wang, and Z. Yang. A two-timescale framework for bilevel optimization: Complexity analysis and application to actor-critic, dec. 20.arXiv preprint arXiv:2007.05170, 2020

    2007 arXiv

  19. [27]

    J. J. Horton. Large language models as simulated economic agents: What can we learn from homo silicus? Technical report, National Bureau of Economic Research, 2023

    2023

  20. [28]

    S. Hu, T. Huang, F. Ilhan, S. Tekin, G. Liu, R. Kompella, and L. Liu. A survey on large language model-based game agents.arXiv preprint arXiv:2404.02039, 2024

    2024 arXiv

  21. [29]

    Ilut and R

    C. Ilut and R. Valchev. Economic agents as imperfect problem solvers.The Quarterly Journal of Economics, 138(1):313–362, 2023

    2023

  22. [30]

    Jeurissen, D

    D. Jeurissen, D. Perez-Liebana, J. Gow, D. Cakmak, and J. Kwan. Playing nethack with llms: Potential & limitations as zero-shot agents.arXiv preprint arXiv:2403.00690, 2024

    2024 arXiv

  23. [31]

    Karten, A

    S. Karten, A. L. Nguyen, and C. Jin. Pokéchamp: an expert-level minimax language agent. arXiv preprint arXiv:2503.04094, 2025. 5 14 LLM Economist

    2025 arXiv

  24. [32]

    Klissarov, P

    M. Klissarov, P. D’Oro, S. Sodhani, R. Raileanu, P.-L. Bacon, P. Vincent, A. Zhang, and M. Henaff. Motif: Intrinsic motivation from artificial intelligence feedback.arXiv preprint arXiv:2310.00166, 2023

    2023 arXiv

  25. [33]

    J. Y. Koh, R. Lo, L. Jang, V. Duvvur, M. C. Lim, P.-Y. Huang, G. Neubig, S. Zhou, R. Salakhutdinov, and D. Fried. Visualwebarena: Evaluating multimodal agents on realistic visual web tasks.arXiv preprint arXiv:2401.13649, 2024

    2024 arXiv

  26. [34]

    A. Korinek. Generative ai for economic research: Llms learn to collaborate and reason. Technical report, National Bureau of Economic Research, 2024. 5

    2024

  27. [35]

    Laskin, L

    M. Laskin, L. Wang, J. Oh, E. Parisotto, S. Spencer, R. Steigerwald, D. Strouse, S. Hansen, A. Filos, E. Brooks, et al. In-context reinforcement learning with algorithm distillation. arXiv preprint arXiv:2210.14215, 2022. 1

    2022 arXiv

  28. [36]

    J. Z. Leibo, V. Zambaldi, M. Lanctot, J. Marecki, and T. Graepel. Multi-agent reinforcement learning in sequential social dilemmas.arXiv preprint arXiv:1702.03037, 2017. 5

    2017 arXiv

  29. [37]

    J. Z. Leibo, A. S. Vezhnevets, W. A. Cunningham, S. Krier, M. Diaz, and S. Osindero. Societal and technological progress as sewing an ever-growing, ever-changing, patchy, and polychrome quilt. arXiv preprint arXiv:2505.05197, 2025. 5

    2025 arXiv

  30. [38]

    Leng and Y

    Y. Leng and Y. Yuan. Do llm agents exhibit social behavior? arXiv preprint arXiv:2312.15198, 2023

    2023 arXiv

  31. [39]

    N. Li, C. Gao, M. Li, Y. Li, and Q. Liao. Econagent: large language model-empowered agents for simulating macroeconomic activities.arXiv preprint arXiv:2310.10436, 2023. 5

    2023 arXiv

  32. [40]

    J. Liu, D. Shen, Y. Zhang, B. Dolan, L. Carin, and W. Chen. What makes good in-context examples for gpt-3? arXiv preprint arXiv:2101.06804, 2021. 5

    2021 arXiv

  33. [41]

    X. Liu, H. Yu, H. Zhang, Y. Xu, X. Lei, H. Lai, Y. Gu, H. Ding, K. Men, K. Yang, et al. Agentbench: Evaluating llms as agents.arXiv preprint arXiv:2308.03688, 2023

    2023 arXiv

  34. [42]

    Y. Lu, M. Bartolo, A. Moore, S. Riedel, and P. Stenetorp. Fantastically ordered prompts and where to find them: Overcoming few-shot prompt order sensitivity.arXiv preprint arXiv:2104.08786, 2021

    2021 arXiv

  35. [43]

    R. E. Lucas Jr. Econometric policy evaluation: A critique. InCarnegie-Rochester conference series on public policy, volume 1, pages 19–46. North-Holland, 1976. 1

    1976

  36. [44]

    R. D. Luce et al.Individual choice behavior, volume 4. Wiley New York, 1959. 1

    1959

  37. [45]

    K. Ma, H. Zhang, H. Wang, X. Pan, W. Yu, and D. Yu. Laser: Llm agent with state-space exploration for web navigation.arXiv preprint arXiv:2309.08172, 2023

    2023 arXiv

  38. [46]

    W. Ma, Q. Mi, X. Yan, Y. Wu, R. Lin, H. Zhang, and J. Wang. Large language models play starcraft ii: Benchmarks and a chain of summarization approach.arXiv preprint arXiv:2312.11865, 2023

    2023 arXiv

  39. [47]

    Maliar and S

    L. Maliar and S. Maliar. The representative consumer in the neoclassical growth model with idiosyncratic shocks.Review of Economic Dynamics, 6(2):362–380, 2003

    2003

  40. [48]

    N. G. Mankiw, M. Weinzierl, and D. Yagan. Optimal taxation in theory and practice. Journal of Economic Perspectives, 23(4):147–174, 2009. 2, 2

    2009

  41. [49]

    C. F. Manski. What is the general welfare? welfare economic perspectives. Technical report, National Bureau of Economic Research, 2025. 15 LLM Economist

    2025

  42. [50]

    R. D. McKelvey and T. R. Palfrey. Quantal response equilibria for normal form games. Games and economic behavior, 10(1):6–38, 1995. 1

    1995

  43. [51]

    J. A. Mirrlees. An exploration in the theory of optimum income taxation.The review of economic studies, 38(2):175–208, 1971. 2

    1971

  44. [52]

    J. A. Mirrlees. Optimal tax theory: A synthesis.Journal of public Economics, 6(4):327–358,

  45. [53]

    Moeini, J

    A. Moeini, J. Wang, J. Beck, E. Blaser, S. Whiteson, R. Chandra, and S. Zhang. A survey of in-context reinforcement learning.arXiv preprint arXiv:2502.07978, 2025. 1

    2025 arXiv

  46. [54]

    Monea, A

    G. Monea, A. Bosselut, K. Brantley, and Y. Artzi. LLMs are in-context reinforcement learners, 2024. URL https://openreview.net/forum?id=YW79lAHBUF. 1

    2024

  47. [55]

    G. H. Orcutt. Simulation of economic systems.The American Economic Review, 50(5): 893–907, 1960

    1960

  48. [56]

    Paglieri, B

    D. Paglieri, B. Cupiał, S. Coward, U. Piterbarg, M. Wolczyk, A. Khan, E. Pignatelli, Ł. Kuciński, L. Pinto, R. Fergus, et al. Balrog: Benchmarking agentic llm and vlm reasoning on games. arXiv preprint arXiv:2411.13543, 2024. 5

    2024 arXiv

  49. [57]

    J. S. Park, J. O’Brien, C. J. Cai, M. R. Morris, P. Liang, and M. S. Bernstein. Generative agents: Interactive simulacra of human behavior. InProceedings of the 36th annual acm symposium on user interface software and technology, pages 1–22, 2023. 5

    2023

  50. [58]

    Bernstein

    J.S.Park, C.Q.Zou, A.Shaw, B.M.Hill, C.Cai, M.R.Morris, R.Willer, P.Liang, andM.S. Bernstein. Generative agent simulations of 1,000 people.arXiv preprint arXiv:2411.10109,

    arXiv

  51. [59]

    Rees-Jones and D

    A. Rees-Jones and D. Taubinsky. Taxing humans: Pitfalls of the mechanism design approach and potential resolutions.Tax Policy and the Economy, 32(1):107–133, 2018

    2018

  52. [60]

    E. Saez. Using elasticities to derive optimal income tax rates.The review of economic studies, 68(1):205–229, 2001. 1, 2, 2, 5, D.1, D.2

    2001

  53. [61]

    Saez and S

    E. Saez and S. Stantcheva. Generalized social marginal welfare weights for optimal tax theory. American Economic Review, 106(01):24–45, 2016. 1, 2, 5

    2016

  54. [62]

    N. E. Sanders, A. Ulinich, and B. Schneier. Demonstrations of the potential of ai-based political issue polling.arXiv preprint arXiv:2307.04781, 2023

    2023 arXiv

  55. [63]

    Srivastava, S

    P. Srivastava, S. Golechha, A. Deshpande, and A. Sharma. Nice: To optimize in-context examples or not?arXiv preprint arXiv:2402.06733, 2024. 5

    2024 arXiv

  56. [64]

    Topsakal and J

    O. Topsakal and J. B. Harper. Benchmarking large language model (llm) performance for game playing via tic-tac-toe.Electronics, 13(8):1532, 2024

    2024

  57. [65]

    Trott, S

    A. Trott, S. Srinivasa, D. van der Wal, S. Haneuse, and S. Zheng. Building a foundation for data-driven, interpretable, and robust policy design using the ai economist.arXiv preprint arXiv:2108.02904, 2021. 5

    2021 arXiv

  58. [66]

    Census Bureau

    U.S. Census Bureau. American community survey, 2023 public-use microdata sample (pums). https://www.census.gov/programs-surveys/acs, 2023. Accessed May 14, 2025. 3, 4

    2023

  59. [67]

    Von Stackelberg.Market structure and equilibrium

    H. Von Stackelberg.Market structure and equilibrium. Springer Science & Business Media, 2010. 16 LLM Economist

    2010

  60. [68]

    Y. Wang, Q. Liu, Y. Bai, and C. Jin. Breaking the curse of multiagency: Provably efficient decentralized multi-agent rl with function approximation. InThe Thirty Sixth Annual Conference on Learning Theory, pages 2793–2848. PMLR, 2023

    2023

  61. [69]

    Willis, Y

    R. Willis, Y. Du, J. Z. Leibo, and M. Luck. Will systems of llm agents cooperate: An investigation into a social dilemma.arXiv preprint arXiv:2501.16173, 2025. 5

    2025 arXiv

  62. [70]

    C. Yang, X. Wang, Y. Lu, H. Liu, Q. V. Le, D. Zhou, and X. Chen. Large language models as optimizers, 2024. URLhttps://arxiv.org/abs/2309.03409

    2024 arXiv

  63. [71]

    J. C. Yang, M. Korecki, D. Dailisan, C. I. Hausladen, and D. Helbing. Llm voting: Human choices and ai collective decision making.arXiv preprint arXiv:2402.01766, 2024

    2024 arXiv

  64. [72]

    Z. Yang, Z. Zhang, Z. Zheng, Y. Jiang, Z. Gan, Z. Wang, Z. Ling, J. Chen, M. Ma, B. Dong, et al. Oasis: Open agents social interaction simulations on one million agents.arXiv preprint arXiv:2411.11581, 2024. 5

    2024 arXiv

  65. [73]

    W.-B. Zhang. A discrete heterogeneous-group economic growth model with endogenous leisure time. Discrete Dynamics in Nature and Society, 2009(1):670560, 2009

    2009

  66. [74]

    Zhang, S

    Y. Zhang, S. Mao, T. Ge, X. Wang, A. de Wynter, Y. Xia, W. Wu, T. Song, M. Lan, and F. Wei. Llm as a mastermind: A survey of strategic reasoning with large language models. arXiv preprint arXiv:2404.01230, 2024. 1

    2024 arXiv

  67. [75]

    Zheng, A

    S. Zheng, A. Trott, S. Srinivasa, N. Naik, M. Gruesbeck, D. C. Parkes, and R. Socher. The ai economist: Improving equality and productivity with ai-driven tax policies.arXiv preprint arXiv:2004.13332, 2020. 5

    2004 arXiv

  68. [76]

    Zheng, A

    S. Zheng, A. Trott, S. Srinivasa, D. C. Parkes, and R. Socher. The ai economist: Taxation policy design via two-level deep multiagent reinforcement learning.Science advances, 8(18): eabk2607, 2022. 5

    2022

  69. [77]

    A. Zhou, K. Yan, M. Shlapentokh-Rothman, H. Wang, and Y.-X. Wang. Language agent tree search unifies reasoning acting and planning in language models. arXiv preprint arXiv:2310.04406, 2023

    2023 arXiv

  70. [78]

    R. Zhou, S. S. Du, and B. Li. Reflect-rl: Two-player online rl fine-tuning for lms.arXiv preprint arXiv:2402.12621, 2024. 17 LLM Economist A Worker Personas In our experiments, we utilized a diverse set of worker personas to model a heterogeneous population with varying prefer...

    2024 arXiv

  71. [2022]

    URL https://www.science.org/doi/abs/10.1126/ science.ade9097

    doi: 10.1126/science.ade9097. URL https://www.science.org/doi/abs/10.1126/ science.ade9097

    doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.