pith. sign in

arxiv: 2606.31935 · v1 · pith:C4KPM27Ynew · submitted 2026-06-30 · 💰 econ.EM

Delegation Rights: Property, Agency, and Investment Incentives in the Age of AI Agents

Yukun Zhang , Kemu Xu This is my paper

Pith reviewed 2026-07-01 02:06 UTC · model grok-4.3

classification 💰 econ.EM
keywords delegation rightsAI agentsresidual controlincomplete contractsinvestment incentivesdeadweight lossplatform control
0
0 comments X

The pith

Certified delegation rights restore investment incentives by conditioning AI agent access on verifiable standards.

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

This paper models how control over AI agents operating in user accounts affects investment incentives. It shows that letting platforms hold discretionary veto power over delegation protects systems but discourages users and agents from making relationship-specific investments. Pure user control avoids this problem but may not adequately address security and privacy risks. A middle path called certified delegation ties access to verifiable conditions like revocability and auditability, which the model suggests can lower overall deadweight loss. Simulations illustrate the mechanism's potential to balance incentives and risk containment.

Core claim

The paper claims that delegation rights represent revocable authority for proxy execution, and that certified delegation allocates residual control conditionally on meeting standards for authorization, revocability, auditability, rate limits, data minimization, and risk mitigation, thereby reducing deadweight loss compared to platform or user control regimes.

What carries the argument

Residual control over the mode of account execution in a three-party model between user, AI agent provider, and platform.

If this is right

  • Platform control weakens the user-agent coalition's disagreement payoff, depressing investment.
  • User control reduces hold-up but fails to internalize risks sufficiently.
  • Certified delegation restores incentives for delegation while bounding residual risk through conditional access.
  • Mechanism simulations demonstrate reduced deadweight loss under this regime.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The approach may apply to delegation in other digital services beyond AI agents.
  • Testing could involve observing investment changes when platforms implement certified delegation options.
  • Certification functions as an economic governance tool allocating control rights.

Load-bearing premise

The model assumes that under platform control the platform's discretionary veto weakens the user-agent coalition's disagreement payoff and depresses relationship-specific investment.

What would settle it

A simulation or empirical study showing no increase in relationship-specific investments or no reduction in deadweight loss under certified delegation would falsify the central claim.

Figures

Figures reproduced from arXiv: 2606.31935 by Kemu Xu, Yukun Zhang.

Figure 1
Figure 1. Figure 1: Delegation rights at the account interface. The figure summarizes the three [PITH_FULL_IMAGE:figures/full_fig_p013_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Baseline diagnostics across control regimes. [PITH_FULL_IMAGE:figures/full_fig_p048_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Regime map in elasticity space. 50 [PITH_FULL_IMAGE:figures/full_fig_p050_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Welfare under Certified Delegation as the risk standard varies. [PITH_FULL_IMAGE:figures/full_fig_p051_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Scale-normalized counterfactual: AI-assisted commerce. [PITH_FULL_IMAGE:figures/full_fig_p053_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Scale-normalized counterfactual: closed social ecosystem. [PITH_FULL_IMAGE:figures/full_fig_p054_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Robustness diagnostic surface. 57 [PITH_FULL_IMAGE:figures/full_fig_p057_7.png] view at source ↗
read the original abstract

AI agents increasingly operate inside digital accounts by exercising privileges that users already hold, raising a new control question: whether an existing account entitlement must be exercised manually or may be exercised through a user-authorized automated proxy. We define \emph{delegation rights} as the revocable, identity-preserving, scope-limited, and mode-specific authority of an account holder to authorize such proxy execution. We develop a three-party incomplete-contracts model with a User, an AI Agent provider, and a Platform. The contested object is not platform ownership, account transferability, data portability, or unrestricted API access, but residual control over the mode of account execution. Under Platform Control, the platform can protect infrastructure, identity systems, privacy boundaries, and third parties, but its discretionary veto weakens the User--Agent coalition's disagreement payoff and depresses relationship-specific investment. Under User Control, hold-up is reduced, but security, privacy, congestion, and third-party risks may remain insufficiently internalized. We then analyze \emph{Certified Delegation}, under which access protection is conditional on verifiable authorization, revocability, auditability, rate-limit compliance, data minimization, and risk mitigation. Certification is therefore not merely a technical safety screen; it is a conditional allocation of residual control. Illustrative mechanism simulations show how this regime can reduce deadweight loss by restoring delegation incentives while bounding residual risk.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The paper defines delegation rights as revocable, identity-preserving, scope-limited authority for users to authorize AI agent proxies in digital accounts. It develops a three-party incomplete-contracts model (User, AI Agent provider, Platform) comparing Platform Control (discretionary veto weakens User-Agent coalition disagreement payoff and depresses investment), User Control (reduced hold-up but uninternalized risks), and Certified Delegation (conditional on verifiable authorization, revocability, auditability, etc.). Illustrative mechanism simulations are presented to show that Certified Delegation reduces deadweight loss by restoring delegation incentives while bounding residual risk.

Significance. If the central claim holds after robustness checks, the paper contributes a new framing of residual control rights over account execution modes in AI settings, extending incomplete-contracts theory to platform-AI interactions. The three-party structure and focus on certification as a conditional allocation of control are strengths, though the illustrative simulations limit immediate applicability without analytical backing or sensitivity results.

major comments (1)
  1. [mechanism simulations] The central claim that Certified Delegation reduces deadweight loss rests on illustrative mechanism simulations. No equilibrium derivation, closed-form comparative statics, or sensitivity table is referenced in the provided description; the ranking of regimes therefore depends on the specific payoff functions and parameter values chosen. Altering functional forms (e.g., convex veto costs or positive certification costs) could reverse the result without violating the incomplete-contracts setup.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the detailed report and constructive feedback on the mechanism simulations. We address the major comment below.

read point-by-point responses

  1. Referee: [mechanism simulations] The central claim that Certified Delegation reduces deadweight loss rests on illustrative mechanism simulations. No equilibrium derivation, closed-form comparative statics, or sensitivity table is referenced in the provided description; the ranking of regimes therefore depends on the specific payoff functions and parameter values chosen. Altering functional forms (e.g., convex veto costs or positive certification costs) could reverse the result without violating the incomplete-contracts setup.

    Authors: We agree that the simulations are illustrative and that the paper does not derive closed-form equilibria or provide exhaustive sensitivity tables. The core contribution is a three-party incomplete-contracts model that frames delegation rights as a conditional allocation of residual control over account execution mode. In this tradition, the analysis emphasizes qualitative predictions about how control rights affect relationship-specific investment and risk internalization, rather than quantitative robustness across all functional forms. The simulations demonstrate a mechanism by which Certified Delegation can reduce deadweight loss relative to the other regimes under the model's maintained assumptions. We do not claim that the ranking is invariant to arbitrary changes in functional forms. To strengthen the presentation, we will add a dedicated sensitivity section in the revision that varies key parameters (including convex costs and certification costs) and reports the conditions under which the ranking is preserved. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper develops a three-party incomplete-contracts model and presents illustrative mechanism simulations as evidence that Certified Delegation can reduce deadweight loss. The provided abstract and text contain no equations, fitted parameters, or self-citations where a claimed prediction or result reduces by construction to its own inputs. The modeling assumptions (e.g., effects of discretionary veto on disagreement payoffs) are stated explicitly as part of the setup rather than derived from the target outcome, and the simulations are described as illustrative without any indication that outcomes are forced by parameter fitting or renaming. The derivation chain remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The claim rests on the applicability of incomplete-contracts theory to AI delegation and the definition of certified delegation as a mechanism that internalizes risks while preserving incentives; no free parameters or invented physical entities are stated.

axioms (1)
  • domain assumption Incomplete contracts framework applies directly to residual control over AI agent execution mode
    The paper develops a three-party incomplete-contracts model with User, AI Agent provider, and Platform.
invented entities (1)
  • delegation rights no independent evidence
    purpose: To capture revocable, identity-preserving, scope-limited authority for proxy execution by AI agents
    Newly defined in the paper as the contested object distinct from ownership or data portability.

pith-pipeline@v0.9.1-grok · 5775 in / 1351 out tokens · 40667 ms · 2026-07-01T02:06:36.833174+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

12 extracted references · 5 canonical work pages · 1 internal anchor

  1. [1]

    Regulation (eu) 2016/679 (general data protection regulation), article 20: Right to data portability

    European Parliament and Council. Regulation (eu) 2016/679 (general data protection regulation), article 20: Right to data portability. Official Journal of the European Union, L 119,

    2016

  2. [2]

    Regulation (eu) 2022/1925 on contestable and fair markets in the digital sector (digital markets act)

    European Parliament and Council. Regulation (eu) 2022/1925 on contestable and fair markets in the digital sector (digital markets act). Official Journal of the European Union, L 265, 2022a. European Parliament and Council. Regulation (eu) 2022/2065 on a single market for digital services (digital services act). Official Journal of the European Union, L 27...

    2022

  3. [3]

    Regulation (eu) 2024/1689 laying down harmonised rules on artificial intelligence (ai act)

    European Parliament and Council. Regulation (eu) 2024/1689 laying down harmonised rules on artificial intelligence (ai act). Official Journal of the European Union,

    2024

  4. [4]

    ISO/IEC 42001:2023 — artificial intelli- gence management system

    International Organization for Standardization. ISO/IEC 42001:2023 — artificial intelli- gence management system. ISO,

    2023

  5. [5]

    Jan Krämer

    doi: 10.1086/466922. Jan Krämer. Personal data portability in the platform economy: Economic implications and policy recommendations.Journal of Competition Law & Economics, 17(2):263– 308,

    work page doi:10.1086/466922

  6. [6]

    Jan Krämer and Nadine Stüdlein

    doi: 10.1093/joclec/nhaa030. Jan Krämer and Nadine Stüdlein. Data portability, data disclosure and data-induced switching costs: Some unintended consequences of the general data protection regula- tion.Economics Letters, 181:99–103,

    work page doi:10.1093/joclec/nhaa030

  7. [7]

    Jan Krämer, Pierre Senellart, and Alexandre de Streel

    doi: 10.1016/j.econlet.2019.05.015. Jan Krämer, Pierre Senellart, and Alexandre de Streel. Making data portability more effective for the digital economy. Technical report, Centre on Regulation in Europe (CERRE),

    work page doi:10.1016/j.econlet.2019.05.015 2019

  8. [8]

    White, Margaret Mitchell, Timnit Gebru, Ben Hutchinson, Jamila Smith-Loud, Daniel Theron, and Parker Barnes

    Inioluwa Deborah Raji, Andrew Smart, Rebecca N. White, Margaret Mitchell, Timnit Gebru, Ben Hutchinson, Jamila Smith-Loud, Daniel Theron, and Parker Barnes. Clos- ing the AI accountability gap: Defining an end-to-end framework for internal algorith- mic auditing. InProceedings of the 2020 Conference on Fairness, Accountability, and Transparency, pages 33–44,

    2020

  9. [9]

    Jean-Charles Rochet and Jean Tirole

    doi: 10.1145/3351095.3372873. Jean-Charles Rochet and Jean Tirole. Platform competition in two-sided markets.Jour- nal of the European Economic Association, 1(4):990–1029,

    work page doi:10.1145/3351095.3372873

  10. [10]

    Van buren v

    71 Supreme Court of the United States. Van buren v. united states, 593 u.s. 374 (2021),

    2021

  11. [11]

    hiq labs, inc

    United States Court of Appeals for the Ninth Circuit. hiq labs, inc. v. linkedin corp., 31 f.4th 1180 (9th cir. 2022),

    2022

  12. [12]

    WebArena: A Realistic Web Environment for Building Autonomous Agents

    Shuyan Zhou, Frank F. Xu, Hao Zhu, Xuhui Zhou, Robert Lo, Abishek Sridhar, Xianyi Cheng, Tianyue Ou, Yonatan Bisk, Daniel Fried, Uri Alon, and Graham Neubig. WebArena: A realistic web environment for building autonomous agents. arXiv:2307.13854,

    work page internal anchor Pith review Pith/arXiv arXiv