Agentic AI systems combine probabilistic reasoning with delegated action through tools, context, memory, orchestration, and external workflow integration. This note develops a formal and managerially usable model that distinguishes Agentic Technical Debt from Stochastic Tax. Agentic Technical Debt is a stock of accumulated design and governance liability. Stochastic Tax is a recurring flow of operating burden that arises when stochastic agents are used in business workflows. The two constructs are related, but they are not the same: debt can amplify the tax, while the tax can remain positive even when debt is minimized. The note starts from a compact dashboard expression, expands it into a fuller structural model, defines all variables and parameters, shows how each cost category can be estimated from operational data, and illustrates the framework with an accounts-payable simulation and companion spreadsheet.

We study the problem of allocating a finite estate among agents whose total claims exceed the available resources, a standard framework in the theory of claims problems. Two canonical rules embody competing fairness ideals: the Proportional rule allocates in proportion to claims, while the Constrained Equal Awards (CEA) rule equalizes awards as much as possible subject to claim-boundedness. We introduce the P-CEA family of compromise rules, which assigns each agent a fixed baseline award, capped by her claim, and distributes the remaining estate proportionally to residual claims. By varying the baseline parameter, this family generates a continuum of allocation rules that interpolates between the Proportional and CEA benchmarks. We provide an axiomatic characterization based on two threshold-dependent principles: No Advantageous Reallocation, which prevents agents with claims above the threshold from benefiting through coordinated claim redistribution that preserves the threshold condition, and Sustainable Lower Bound, which guarantees each agent at least the minimum of her claim and the threshold. We further develop a dual analysis that reallocates losses instead of awards and characterize the corresponding dual family using the dual versions of our axioms.

Urban decarbonization requires scaling rooftop solar across millions of fragmented producers, yet cities face a fundamental tension: energy data is easily manipulated, and economic incentives often reward speculation rather than actual infrastructure deployment. We present SolarChain, a platform that resolves both problems by anchoring digital accountability to the thermodynamic limits of solar energy conversion. Using real-time meteorological data, geospatial coordinates, and first-principles calculations of solar yield, the system establishes a hard physical boundary for every panel's maximum possible output; any reported generation exceeding this limit is automatically rejected before entering the shared ledger. This trustless verification enables a peer-to-peer marketplace with programmatic reward structures that continuously reinvest value into equipment maintenance and market liquidity, preventing the speculative hoarding that typically destabilizes blockchain-based marketplaces. When electricity is consumed, the corresponding digital credits are permanently retired in direct proportion to physical energy dissipation, creating an auditable one-to-one mapping between urban consumption and carbon accounting. Deployed across heterogeneous city nodes, the prototype demonstrates resilience against data injection attacks while lowering capital barriers for community-level solar expansion. Beyond energy, the framework offers a general model for coordinating economic activity with physical law in any domain where distributed infrastructure demands both data integrity and sustainable investment. We release the data and code as open-access on GitHub.

This paper examines how estimates of AI use in scientific writing can be biased when evaluation methods ignore contextual differences across countries and fields. Using large-scale data on journal publications from Dimensions, we construct AI-likeness benchmarks based on differences between human-written and LLM-rephrased abstracts. We show that a pooled benchmark may confound pre-existing stylistic variation with AI-generated text, producing substantial distortions across country-field groups even in pre-LLM publications. In contrast, country-field-specific benchmarks attenuate such distortions and provide a more credible baseline for comparison. Applying these methods to publications in 2025 reveals that the pooled benchmark systematically overestimates AI use in certain countries and fields while underestimating it in others. These findings highlight the importance of context-aware measurement for accurate and equitable evaluation of AI use in science.

We study a tractable two-player contest built on a truncated cubic contest success function. Its defining feature is a strategic-feedback parameter whose sign determines whether a leading player's effort lowers (suppression) or raises (empowerment) the marginal effectiveness of the trailing player's effort; standard lottery contests impose suppression by construction. The benchmark yields closed-form mixed equilibria under complete information and a unique affine Bayesian Nash equilibrium under IID private information. Expected effort is typically single-peaked in the feedback parameter. Uncertainty lowers effort under suppression but raises it under empowerment, and the same asymmetry governs information disclosure: an effort-maximizing designer withholds information under suppression and discloses fully under empowerment. Several familiar conclusions of contest theory turn out to reflect suppressive benchmarks rather than contests as such.

Mechanism-mediated service markets with polymatroidal feasibility admit efficient, dominant-strategy incentive-compatible (DSIC) allocation, but these guarantees implicitly assume truthful execution by the marketplace operator. Modelling the operator as a strategic player, we establish a credibility trilemma: for single-parameter agents on a non-modular polymatroid, no static sealed-bid mechanism is simultaneously revenue-optimal, DSIC for agents, and credible for the operator. We introduce the Cost of Non-Credibility (CoNC) as a price-of-anarchy-style welfare-loss measure and obtain tight $Θ$-bounds across five topology classes (single-edge, series, parallel, tree, series-parallel), plus a matching upper bound $O(|\mathcal{S}|)$ on general DAGs realised by an $Ω(|\mathcal{S}|)$ witness on the SP-augmented sub-family, turning the trilemma into a structural quantity. Three structurally distinct resolutions follow: public broadcast or deferred-revelation commitment, administrative domain separation under settlement separation and four side conditions, and integrator competition orthogonal to mechanism execution under disjoint actors. An instance-level grounding over the edge-pricing market of Amin et al. confirms the trilemma's robustness on a refereed external setting. The result establishes marketplace neutrality as a first-order design constraint on polymatroidal service markets rather than an implementation detail: where the operator is a strategic player, credibility trades off against revenue optimality and agent incentive compatibility along structurally characterised lines.

Tabular foundation models achieve strong accuracy on choice prediction tasks, but their predictions often violate the economic logic those tasks require: raising a price sometimes increases predicted demand, and implied willingness-to-pay estimates are frequently negative or implausible. We propose a two-stage adapter that embeds foundation model predictions within a utility-maximization framework. In the first stage, we estimate a standard choice model whose parameters are constrained to obey economic theory. In the second stage, we freeze those parameters and train a correction term that incorporates the foundation model's predictions as additional information. The result is a model that inherits the foundation model's accuracy gains while guaranteeing monotonic price-demand relationships under policy perturbation and producing analytically computable trade-off measures. On two transportation datasets, the adapter recovers up to 13 percentage points of accuracy over a standard logit model while maintaining perfect economic consistency, something neither the raw foundation models nor conventional distillation achieve.

This paper introduces state-robust equilibrium (SRE), a local validity test for Nash predictions in finite-strategy population games when the payoff-relevant aggregate state may be misspecified. The reported prescription and payoff map are held fixed; only the state used to evaluate payoff comparisons varies. SRE is equivalent to local best-response invariance, absence of structural exposure, and validity along every vanishing interior aggregate-state error. In affine games, the tangent-cone, normal-cone, and linear-program tests characterize exposure and identify the exposing population, the pure strategy, and the aggregate-state direction. The main implication is a sharp negative result: robust mixing requires local payoff identity on the support; in generic affine games, SRE reduce to strict pure Nash equilibria, although weak boundary equilibria can survive through feasible-set protection. In affine games with polyhedral local uncertainty regions, the same inequalities yield a deterministic finite diagnostic for reported-state validity.

Researchers have started using LLM agents in place of human subjects in behavioural and political-science experiments, often as a cheaper substitute for laboratory pools. The substitution does not hold up in strategic settings: humans and LLMs reliably make different choices, and neither fine-tuning on human response data nor persona conditioning has closed the gap. The behavioural-economics literature has, since Simon's introduction of bounded rationality, modelled human strategic behaviour as a classical baseline plus an additive correction term $δ$. The framework proposed here reads $δ$ as the mathematical signature of bounded computation: the gap between what an unboundedly-rational agent would compute and what a computationally bounded agent actually produces. For canonical games whose solutions are present in standard training corpora, LLMs retrieve and recombine corpus material, bypassing the bound that produces $δ$ in humans. The framing extends to reasoning-distilled models through cognitive-hierarchy theory: their accessible level-$k$ strategic reasoning is bounded by compute budget and context length rather than by the cognitive constraints that bound humans, and the $δ$ they produce, if any, carries different structural signatures. Four operational tests (conditional dependence, distributional asymmetry, path-dependence under repetition, and paraphrase-robustness) are proposed to discriminate human-shaped $δ$ from LLM-shaped $δ$. A moderator prediction is that $|δ|$ scales with peer-signal individuation in the decision environment, with a quantitative bound of Cohen's $d \geq 0.5$ between named-opponent and aggregate-opponent settings.

We study stochastic object assignment problems in which objects may have minimum and maximum requirements, such as with classes with upper and lower enrollment bounds. We construct a new random assignment mechanism, the minimums probabilistic serial (MPS) mechanism, which generalizes the Probabilistic Serial mechanism of Bogomolnaia and Moulin (2001). The random allocation produced by MPS is guaranteed to be Pareto efficient; that is, there is no other implementable allocation that all agents prefer via first order stochastic dominance. We also show that MPS is i) envy-free, in that no agent will strictly prefer another agent's assignment, and ii) weak strategyproof, in that agents cannot achieve a better assignment by misreporting their preferences.

We study a nonlinear factor model in which observed responses depend on low-rank latent factors through an unknown monotone link function. This setting is challenging and largely underexplored due to severe nonconvexity and identifiability issues. The link function is assumed to lie in a reproducing kernel Hilbert space (RKHS), enabling flexible nonparametric modeling while preserving identifiability. We formulate the problem as the joint recovery of the low-rank factors, loadings, and the nonlinear link function from possibly incomplete and noisy observations and propose a projected block coordinate descent (BCD) algorithm with explicit regularization to address scale and rotational ambiguities. Under mild incoherence of factors and standard sampling conditions, we establish convergence guarantees in both noiseless and noisy regimes, along with sublinear regret bounds for the link-function updates. Our results extend classical linear factor models to a broad nonlinear regime and provide a principled framework for learning nonlinear latent structures. We evaluate the proposed approach using controlled synthetic experiments, indicating promising performance.

Contract theory typically assumes full commitment by the principal, but many contracts fix some payoff-relevant decisions while leaving others discretionary. We ask when imperfect commitment is equivalent to full commitment. For contracts in which a committed baseline is followed by a bounded discretionary adjustment, as in commercial-insurance schedule rating or civil penalties, bounded discretion is allocation-neutral. When contractible and non-contractible decisions are distinct instruments, the equivalence fails. We characterize optimal single-principal contracts and show that simple-offer equilibria are robust under competing principals. The methodological contribution is an extended taxation principle that makes these analyses more tractable.

We consider debiased inference on finite-dimensional functionals of infinite-dimensional least-squares solutions to inverse problems as a way to avoid having to assume exact solutions exist. Such assumptions are substantive and not innocuous, and their failure may imperil inference when we impose them on the statistical model. Our approach instead allows us to conduct inference on a quantity that is defined regardless of solutions existing and coincides with the usual estimands when they do. For the case of instrumental variables, this means we can motivate the analysis with structural models but these do not need to hold exactly for the semiparametric inferential procedure to remain valid.

Score-driven models update time-varying parameters using conditional likelihood scores. This paper develops a Bayesian interpretation of such updates through Tweedie's formula, which connects posterior mean corrections with marginal scores. In Gaussian signal extraction, this gives an exact posterior-correction identity. For natural exponential families, related identities characterize posterior means in natural- and expectation-parameter spaces. Building on these identities, we show that conjugate Bayesian filtering in expectation space coincides exactly with an inverse-Fisher-scaled conditional score update under local precision discounting. For general conditional densities, the exact Bayesian correction involves a generally unavailable predictive-marginal score. A local Gaussian approximation shows that the conditional likelihood score provides the leading approximation to this posterior correction; under local precision discounting, the predictive covariance becomes proportional to inverse Fisher information, yielding the familiar inverse-Fisher-scaled score recursion. The results clarify when score-driven updates are exact Bayesian filters and when they should instead be viewed as tractable local approximations.

Mandatory waiting periods have been instituted for medical procedures, gun purchases, and other high-stakes decisions. Are these softly paternalistic policies substitutes for harder restrictions, and are delayed decisions more respected? In a general population survey experiment, Choice Architects make rules for decision-makers facing a high-stakes Bomb Risk Elicitation Task. Treatments vary when the decision is made: on the spot or after one day, and whether the initial decision can be revised. Choice Architects set a cap on the decision-maker's risk taking; in one treatment, they can additionally implement a mandatory waiting period. Exogenous deliberation has no effect on the cap; equivalence testing (TOST) and Bayesian analysis ($\text{BF}_{01} \approx 12$) provide strong positive evidence for the absence of an effect. Endogenously prescribed waiting periods are add-on restrictions that do not substitute for the cap. Choice Architects believe that, with time, the average decision-maker will take less risk and -- because of the distribution of Choice Architects' bliss points -- come closer to Choice Architects' subjective ideal choice; the resulting reduction in forecasted errors is small. Soft and hard paternalistic instruments are not substitutes: waiting periods are deployed as add-on restrictions.

How informative are job ads about the actual pay and non-pay attributes offered by employers? Using a comprehensive database of job ads posted by Norwegian employers, we develop a methodology to systematically classify the pay and non-pay job attributes advertised in vacancy texts. About 60% of job ads provide pay-related information and nearly all ads feature information on non-pay attributes. We link these advertised attributes to the employers posting the ads and validate this information against revealed-preference measures of employer quality, realized attributes, and choices from a survey experiment. All three strategies confirm that job ads provide reliable signals of employer quality. We then incorporate the detailed job attributes in a monopsony framework and quantify their contribution to labor market inequality.

This paper studies the robustness of estimated policy effects to changes in the distribution of covariates, a key determinant of the external validity of (quasi)-experimental results. I propose a novel robustness metric $δ^*$ which measures the smallest covariate shift needed to invalidate an empirical claim about the policy effect (e.g., $ATE > 0$). I estimate $δ^*$ via de-biased GMM, achieving a parametric rate of convergence while accommodating machine-learning estimators of treatment-effect heterogeneity (e.g., LASSO, random forests, neural networks). I develop benchmarking and calibration exercises to interpret the magnitude of $δ^*$. I illustrate these tools in an application to the Oregon Health Insurance Experiment. Researchers can report $δ^*$ alongside the point estimate and standard error as a third number gauging external validity under covariate shifts.