We present a regression-adjustment framework designed for the estimation of longitudinal treatment effects in randomized experiments under static regimes. While regression-adjustment methods are useful for variance reduction in randomized experiments by using pre-treatment covariates, they usually focus only on average effects, from which we cannot obtain valuable insights into when the effects appear and how long they continue. To address this issue, we consider intermediate outcomes and evolving post-treatment covariates over time, and we represent such dynamic trajectories using transition kernels. Furthermore, we establish the asymptotic normality and the semiparametric efficiency bound for our estimator, enabling more powerful statistical inference. Simulation studies and empirical analysis using A/B test data from a streaming platform in Japan show the practical advantages of our method.

We introduce the notion of worst-case posterior and worst-case likelihood sensitivity. These measure, respectively, the sensitivity of the expected cost to worst-case perturbations of the posterior distribution and worst-case perturbations of the likelihood of a Bayesian model. Each defines a quantitative measure of robustness. A decision maker concerned about the sensitivity of the out-of-sample expected cost to deviations from her assumptions will want a decision for which both sensitivities are small. We derive posterior and likelihood sensitivities for uncertainty sets defined in terms of deviation measures. Posterior sensitivity vanishes when the posterior variance shrinks to zero, which occurs when parameter uncertainty is eliminated from learning. Parameter learning does not eliminate likelihood sensitivity. A distributionally robust formulation of a Bayesian optimization problem makes a near-Pareto-optimal tradeoff between performance (expected cost) and robustness (posterior and likelihood sensitivity).

We develop a game-theoretic framework that compares welfare efficiency across three market mechanisms: continuous double auctions with transparent order books (lit exchanges), opaque order books (dark pools), and periodic batch auctions. Each mechanism is modeled as a queuing system where heterogeneous traders face trade-offs between the execution price, waiting costs, and transaction costs. Our main result establishes that under moderate arrival rates and bounded adverse selection, dark pools dominate both alternatives in aggregate ex-ante welfare. Observable order books create costly strategic timing games in which traders delay or rush submissions to optimize their position in the queue, generating wasteful social waiting costs. Opaque order books eliminate these timing games through information design. We formally characterize the equilibrium strategies in each mechanism and prove the welfare ranking $W^{DARK} > W^{LIT} > W^{BATCH}$. Extensions incorporate asymmetric information and endogenous venue choice. The results demonstrate how the information structure and the discipline of the service jointly determine efficiency in strategic matching environments.

AI benchmarks have well-documented limitations, with prior work examining contamination, saturation, and construct underspecification. Aggregation has received far less attention: benchmarks are typically summarized by uniformly averaging item-level scores, implicitly treating every test item as equally valuable. We model benchmarking as a multitask principal-agent game and show that the welfare loss from a benchmark is determined jointly by three item-level primitives: alignment with normative welfare priorities, marginal improvability, and performance variance. We translate the theory into an audit framework that ranks items along each of these three axes, and apply it to OLMES items using WORKBank for welfare, the EvoLM 4B suite for improvability, and the PolyPythias 410M panel for variance. The framework surfaces items that are Pareto-inferior within OLMES subject to a pro-worker welfare operationalization. All code is available at https://github.com/stair-lab/principal-agent-benchmarks.

The reduction of complex labour to simple labour is an unresolved difficulty in Marx's labour theory of value, and a key obstacle that has prevented the transformation problem from being settled definitively. This paper proposes a two-step solution framework. First, we prove that as long as the macroeconomy generates a physical surplus, the reduction coefficients that respect the floor of labour-power reproduction form a bounded ``value feasible region''; within this region the two macro aggregate equalities can hold simultaneously for a reasonable range of the profit rate. Second, we propose a linear mapping method that exploits the observable structure of nominal wages and the reproduction floor constraint to systematically construct the implicit reduction coefficients from the value feasible region. We show that this mapping is a homeomorphism between the price feasible region and the value feasible region, and that it preserves the boundary structure. An empirical calibration based on China's 2017 inter-provincial input--output table with 1272 sectors shows that the reduction coefficients obtained by the mapping method substantially outperform the homogeneous labour method and the wage-proxy method in matching the macro profit share.

We study many-to-one matching with transfers and peer effects, such as matching workers to firms, students to schools, residents to neighborhoods, or consumers to status goods. With flexible prices (as in the labor market), competitive equilibrium exists and is efficient under general conditions. We characterize when workforces are segregated by skill and matched to firms in a positively assortative manner. In general, equilibrium features alternating intervals of workforce segregation and compression (mixing). Comparative statics characterize when workforces are more segregated or more compressed, and when profits and wages are more or less unequal. With uniform prices (as in school or neighborhood choice), the value generated by peer effects accrues to schools rather than students, and equilibrium can be excessively segregated. Our model generalizes both assignment models (optimal transport) and Bayesian persuasion.

In the forward reinforcement-learning problem, the reward is fixed and known; the learner is asked to find a good policy or value function. Here we turn the question around. Given offline data generated by an expert, can we recover the reward the expert was optimizing? This is the inverse reinforcement learning problem, and remarkably, two communities, structural econometricians studying dynamic discrete choice (DDC) and machine learners studying entropy-regularized IRL, have been working on exactly the same probabilistic model under different names. We begin by proving their equivalence. We then develop the classical identification result of Magnac and Thesmar and the classical computational paradigms that grew out of it: Rust's nested fixed-point algorithm, the conditional-choice-probability approach of Hotz and Miller, and the two temporal-difference approaches of Adusumilli and Eckardt: linear semi-gradient TD and approximate value iteration. Each route has its limits: dimensionality, transition-kernel estimation, the deadly triad, or projected fixed-point bias. We then walk through the modern ML/IRL strand: adversarial IRL, occupancy matching, IQ-Learn, and offline ML-IRL, deriving each method's actual objective and stating precisely what it does and does not identify. We close with the empirical-risk-minimization framework of Kang et al., which yields a gradient-based estimator for offline IRL/DDC.

Forecasting agricultural commodity prices in emerging economies is difficult due to high volatility, frequent supply disruptions, and strong cultural influences on demand. This study introduces the Kalimati Vegetable Price Index (KVPI), a new inverse-volatility weighted composite index that aggregates 135 daily wholesale commodities from Kathmandu over ten years (2013-2023). By creating a stable macro-level signal, the KVPI reduces the noise inherent in modelling individual crops. A rich set of 64 causally valid features was developed, including festival lead-lag effects, rolling statistics, and calendar variables. Fourteen forecasting models spanning statistical, tree-based, deep learning, hybrid, and transformer architectures were rigorously evaluated across short (7-day), medium (14- and 30-day), and long-term (90-day) horizons. Tree-based ensembles proved notably robust, while classical statistical models and complex transformers struggled with the noisy dataset. The proposed Momentum-Corrected Online Stacking Ensemble achieved the strongest performance, yielding a Root Mean Square Error (RMSE) of 1.771, an exceptionally low Mean Absolute Percentage Error (MAPE) of 0.68%, and explaining 84.5% of the variance (R-squared = 0.845) at the 90-day horizon. This open-source pipeline provides policymakers and supply chain actors in Nepal and similar markets with a practical, reliable tool for anticipating price movements and strengthening food security.

Topic models are often used as dimension-reduction tools before regression, with estimated document-level topic shares treated as observed covariates. This plug-in workflow creates two inferential difficulties: valid inference requires a regular first-stage-to-second-stage expansion that propagates topic-estimation uncertainty, and, at fixed document length, a document's topic mixture cannot be consistently recovered from its own words even when the population topic matrix is known. Corrected spectral moment methods for latent Dirichlet allocation (LDA) offer a starting point: when the total Dirichlet concentration is known, low-order word moments can be corrected to yield operators diagonal in the latent topic basis. We extend this to downstream regression. Under a finite LDA model with response residuals orthogonal to the low-order token moments used for identification, response-weighted word moments admit the same correction, and the resulting supervised operator identifies the regression coefficient $β$ directly, without estimating document-level topic shares. The main obstacle is that the correction depends on the unknown total concentration $α_0$. We show that, for $k\ge3$ topics and under a generic finite-probe condition, $α_0$ is identified by commutativity: at the true value a family of corrected word-moment operators commute, whereas away from it they generically do not. This yields a feasible estimator and lets uncertainty in $\hatα_0$ propagate into inference for $β$. The estimator is asymptotically linear as the number of documents grows with fixed document length, with sandwich standard errors from document-level moment contributions. Simulations show near-nominal coverage where plug-in topic-share regressions can undercover, and an application to top economics journals illustrates contrast inference for latent topic effects.

This paper proposes ideometrics as the foundation for a generalised and potentially testable theory of human progress and civilisational progress, thus linking ideometrics to studies in economics and history. Building on prior work that conceptualises the human brain as a sensor of ideas, human progress is understood not primarily through outcomes such as wealth, health, or technological advancement, but through the dynamic process of the "idea life cycle" that shapes future states. The paper advances a formal definition of human progress as a measurable improvement in the ability of individuals and societies to generate, evaluate, prioritise, and implement ideas in a way that increasingly aligns prioritised ideas with those that truly lead to preferred future states, given available information and uncertainty, and under scarcity of human capacity, energy, time and resources. It introduces the Ideometric Index of Human Progress (IIHP) that captures the quality of idea generation (G), accuracy of their evaluation (E), efficiency of their prioritisation (P), and effectiveness of their implementation (Ie). It shows that the future progress will be realised if there is good alignment between the perceived future value of ideas and their true, realised future value, assessed as outcome monitoring (O). This formulation shifts the analytical focus from static outcomes to the quality of evaluating ideas, thereby offering a novel lens for understanding progress and regress. The concept can also be extended to long periods of history through the Ideometric Index of Civilisational Progress (IICP), where additional parameters of successful documentation of outcomes (D) and successful intergenerational transmission of gathered knowledge (T) are added. By transforming ideas into measurable units of analysis, ideometrics offers a potentially transformative approach to understanding human progress.

When redeeming open-end funds sell and natural buyers do not step in at once, some limited-capital investor must take the other side and carry the inventory until prices recover. This paper asks what return that investor requires. A continuous-time market-clearing model delivers an expected-return restriction in which the price of residual supply depends on inventory risk, trading costs, funding frictions, and the scarcity of balance sheet available to absorb it. Mapping U.S. mutual fund flows through predetermined holdings over 2003--2024, we measure one observable component of this residual supply. Forced-sale pressure predicts actual fund selling, contemporaneous price declines, and positive returns over the following one to six months. The premium roughly doubles when market-wide absorption capacity is tight, and it concentrates in stocks with thin investor bases and limited trading capacity -- precisely the cross section in which clearing the imbalance should be most costly, and a pattern that mechanical return reversal does not generate.

This paper develops a regression framework for the direct estimation of integrated functionals of conditional outcome distributions. The proposed method, termed rectified linear unit (ReLU) regression, projects the ReLU-transformed outcome onto covariates and admits a closed-form estimator. Its population regression function coincides with the integrated conditional distribution function of the outcome, and its convex conjugate, obtained via the Legendre-Fenchel transformation, recovers the integrated conditional quantile function. Both the regression and its conjugate require only mild distributional assumptions and accommodate non-continuous outcomes. We establish the uniform asymptotic distribution of the estimator and develop inference for the conjugate functional via the delta method for Hadamard directionally differentiable maps. Building on these results, we establish identification and inference for average quantile treatment effects over arbitrary subintervals of probability levels. This broadens the set of distributional parameters available to empirical work.

Collective action often requires institutions that make cooperation individually worthwhile. We ask whether democratic allocation of public-good return can transform a repeated public good into a self-sustaining cooperative institution, and how participation costs reshape that process. A simple evolutionary model shows that voted redistribution can support a prosocial allocation order, but can also sustain an antisocial allocation order or democratic free riding, in which individuals benefit from an institution maintained by others while avoiding the cost of participation. The model predicts competing effects of voting cost. Cost can suppress use of the institution to reward low contributors under strong selection, but can also thin the active electorate and erode contributor-rewarding support. We test these predictions in a preregistered online experiment with \NIncludedGroupsVone{} five-person groups. Endogenous democratic redistribution increased contributions relative to an equal-share public-goods control, with zero-cost voting producing the strongest temporal improvement. Voting costs did not mainly turn active voters toward low-contributor-rewarding allocation. Instead, they shifted behavior toward abstention and democratic free riding, made abstention locally rewarding, and widened the gap between post-task perceptions of democratic participation and the behavioral record. Democratic allocation can therefore stabilize cooperation, but participation costs can reduce the number of people actively sustaining the institution and can make that erosion less visible to participants themselves.

We develop a partial integro-differential equation (PIDE) framework for option pricing under joint stochastic volatility and jump dynamics, and evaluate its empirical content using the S&P500 index option contracts across three maturities. The framework is derived from the infinitesimal generator of an affine Lévy-type process and implemented via finite-difference discretization with FFT-based treatment of the nonlocal jump operator. Calibration via GMM reveals that stochastic volatility accounts for the dominant share of pricing improvement, where relative to Black-Scholes, the Heston specification reduces implied-volatility RMSE by 39%. Jump augmentation via either Merton or CGMY specifications yields marginal improvements concentrated at short maturities and in the deep out-of-the-money region. The calibrated CGMY activity index supports a compound-Poisson structure, consistent with high-frequency evidence on S&P500 index returns.

We study obviously strategy-proof implementation in the framework of social choice over acts introduced by Bahel and Sprumont (2020). We characterize the class of unanimous social choice functions that are implementable via obviously strategy-proof mechanisms. Our main result shows that a unanimous social choice function is obviously strategy-proof implementable if and only if it is dictatorial.

Characterising cause-effect relationships in complex systems is fundamental to understanding their underlying mechanisms. Granger causality (GC) remains a widely used computational tool for identifying causal relationships in time series data. However, like other causal discovery methods, GC has limitations and has been criticised for lacking a rigorous causal foundation. In this work, we present a fix to this criticism by reinterpreting GC through the lenses of Reichenbach's principles and causal Bayesian networks. This reinterpretation was implemented as an algorithm we call causalized Granger causality (c-GC). We demonstrate, both theoretically and graphically, that this reformulation endows GC with a robust causal interpretation under specific assumptions. c-GC yields satisfactory results on synthetic data, offering a more principled framework for causal discovery in observational datasets.

This paper generalizes several results on linear pooling from squared error loss to all kernel scores. The latter are a rich family of scoring rules that covers point and distribution forecasts for univariate and multivariate, discrete and continuous settings. Its members include the Continuous Ranked Probability Score for univariate distribution forecasting and the Energy Score for multivariate distribution forecasting. Our results indicate that forecast disagreement (measured as the average pairwise divergence of all component distributions) has important implications for the linear pool's performance. The results are useful for understanding and designing linear pools in general combination settings. In particular, they motivate using the linear pool (as opposed to other combination formulas) and yield a novel condition under which equal combination weights are optimal under a given kernel scoring rule.

Single use plastics (SUPs) impose substantial environmental costs. Following Directive (EU) 2019/904, Austria introduced producer charges and mandatory participation in collection and recycling systems. This paper exploits a monthly panel of retail offer spells drawn from a price comparison platform to estimate the extent to which compliance costs pass through to posted online prices in Austria. The treated sample comprises keyword matched SUP products including balloons, to go cups, wet wipes, plastic bags, food containers, tobacco filter items, beverage bottles, and plastic wraps observed alongside a control group of non SUP listings over 2020-2024. A two way fixed effects (TWFE) specification places the average post treatment price increase at approximately 4.1 percent. A sequential TWFE model separating the administrative reporting phase from March 2023 and the payment due phase from March 2024 reveals that the larger adjustment occurred during the earlier reporting stage, with a reporting only effect of approximately 8.1 percent and an incremental payment phase effect of 5.6 percent. For balloons, a category subject to pronounced regulatory fee exposure, event study estimates exceed 50 percent immediately following the initial payment date and remain elevated throughout most of the post treatment window. These findings indicate that Austrian online retailers adjusted prices in advance of fee payment deadlines, consistent with anticipatory pass through of expected compliance costs rather than a discrete response to realized payments. As the data contain price observations but not quantity data, the analysis speaks to price incidence and not to consumption or environmental outcomes.

Scientists often want to explain why an outcome is different in two groups. For instance, differences in patient mortality rates across two hospitals could be due to differences in the patients themselves (covariates) or differences in medical care (outcomes given covariates). The Oaxaca--Blinder decomposition (OBD) is a standard tool to tease apart these factors. It is well known that the OBD requires choosing one of the groups as a reference, and the numerical answer can vary with the reference. To the best of our knowledge, there has been no systematic investigation into whether the choice of OBD reference can yield different substantive conclusions and how common this issue is. In the present paper, we give existence proofs in real and simulated data that the OBD references can in fact yield substantively different conclusions. Our empirical exercises find that this sensitivity is more common when the OBD is extended to more complex regression models, including a pretrained transformer. Our theoretical and empirical results together establish that these conclusion reversals are not entirely driven by model misspecification, small data, or adversarial parameter choices. Our results suggest that practitioners should always report both directions of the OBD; that modern machine learning and large datasets do not automatically resolve the conclusion reversal problem; and that further work on this problem is needed.

As AI agents become more autonomous, properly aligning their objectives with human preferences becomes increasingly important. We study how effectively an AI agent learns a human principal's preference in choice under risk via stated versus revealed preferences. We conduct an online experiment in which subjects state their preferences through written instructions ("prompts") and reveal them through choices in a series of binary lottery questions ("data"). We find that on average, an AI agent given revealed-preference data predicts subjects' choices more accurately than an AI agent given stated-preference prompts. Further analysis suggests that the gap is driven by subjects' difficulty in translating their own preferences into written instructions. When given a choice between which information source to give to an AI agent, a large portion of subjects fail to select the more informative one. Moreover, when predictions from the two sources conflict, we find that the AI agent aligns more frequently with the prompt, despite its lower accuracy. Overall, these results highlight the revealed preference approach as a powerful mechanism for communicating human preferences to AI agents, but its success depends on careful implementation.

This paper provides causal evidence that temporary supply disruptions reshape firms' relationship portfolios in international trade. Using exogenous road disruptions during Colombia's 2010-11 La Niña episode, I identify exposure at the buyer-seller relationship level, exploiting variation within importers' supplier portfolios. Exposed relationships are less likely to terminate when importers have alternative non-exposed suppliers. However, firms with broader portfolio exposure gradually reduce their relationship networks and may eventually exit the market. A framework linking relationship surplus to portfolio composition explains these contrasting responses and shows how the same shock generates opposite effects at relationship and firm levels.

The goal of agents in multi-agent environments is to maximize total reward against the opposing agents that are encountered. Following a game-theoretic solution concept, such as Nash equilibrium, may obtain a strong performance in some settings; however, such approaches fail to capitalize on historical and observed data from repeated interactions against our opponents. Opponent modeling algorithms integrate machine learning techniques to exploit suboptimal opponents utilizing available data; however, the effectiveness of such approaches in imperfect-information games to date is quite limited. We show that existing opponent modeling approaches fail to satisfy a simple desirable property even against static opponents drawn from a known prior distribution; namely, they do not guarantee that the model approaches the opponent's true strategy even in the limit as the number of game iterations approaches infinity. We develop a new algorithm that is able to achieve this property and runs efficiently by solving a convex minimization problem based on the sequence-form game representation using projected gradient descent. The algorithm is guaranteed to efficiently converge to the opponent's true strategy under standard Bayesian identifiability and visitation assumptions, given observations from gameplay and possibly additional historical data if it is available.

We outline a foundation for a new field of ``AI Behavioral Science,'' covering three perspectives. First, as AI becomes ubiquitous and is increasingly proprietary and opaque, it becomes vital to develop techniques for assessing AI behavior. We outline how tools developed to assess people's behaviors by social scientists can be used to assess and infer AI's behaviors biases, tendencies, and heuristics. Second, we also discuss how AI can change the ways in which we learn about human behavior. Beyond its computational power, AI offers new techniques for simulating, inferring, and predicting human behaviors that we outline and discuss. Third, as humans and AI are interacting in increasingly complex and intertwined systems, we need to understand the implications for the resulting economic and political outcomes. We outline issues that are increasingly pressing concerning the future of human-AI interactions and potential changes and disruptions that can ensue.

We show that structural smooth transition vector autoregressive models are statistically identified if the shocks are mutually independent and at most one of them is Gaussian. This extends a known identification result for linear structural vector autoregressions to a time-varying impact matrix. We also propose an estimation method, show how a blended identification strategy can be adopted to address weak identification, and establish a sufficient condition for ergodic stationarity. The introduced methods are implemented in the accompanying R package sstvars. Our empirical application finds that a positive climate policy uncertainty shock reduces production and raises inflation under both low and high economic policy uncertainty, but its effects, particularly on inflation, are stronger during the latter.

Network effects -- the utility gains from additional consumers of a good -- are widely regarded as critical to the digital economy. Yet recent theory and evidence suggest that local network effects -- the economic value created by specific social network connections -- drive value in networked online platforms. Using incentive-compatible online choice experiments with 19,923 Facebook, Instagram, LinkedIn, and X users in the United States, we provide the first large-scale empirical measurement of local network effects in the digital economy and measure heterogeneity in connection value across platforms. Platform value ranges from \$78 to \$101 per consumer per month, with 8.1-23.7% explained by local network effects. We find that 1) stronger ties are more valuable on Facebook and Instagram, while weaker ties are more valuable on LinkedIn and X; 2) work connections are most valuable on LinkedIn and least on Facebook, and job-seekers value LinkedIn significantly more and Facebook significantly less; 3) men value connections to women significantly more than to other men, particularly on Instagram, Facebook, and X, while women value connections to men and women equally across platforms; 4) consumers value connections on any platform more if they are also connected on other platforms, suggesting that platforms are complements, not substitutes; 5) white consumers disproportionately value same-race connections on Facebook while, on Instagram, connections to alters eighteen or younger are valued significantly more than any other age group -- two patterns not seen on other platforms. Each platform generates between \$53B and \$215B in annual US consumer surplus. These results suggest that social media generates significant value, that local network effects drive a substantial fraction of it, and that the sources and contours of these effects vary across platforms, consumers, and connections.

Housing supply in dense cities depends on the ability of builders to substitute capital for scarce land. This margin is difficult to estimate because builders choose capital after observing microgeographic conditions that are only imperfectly observed by researchers. This paper develops a method for estimating revenue-based housing production functions in this setting. Because observed capital variation reflects both technological substitution and endogenous responses to latent local conditions, existing estimators can transmit unobserved heterogeneity into the estimated production function. The method treats the unobserved local conditions that affect capital choice as a scalar Markov state and combines the capital share equation with Markov moments implemented using repeated cross-sectional construction data. Monte Carlo simulations show that the estimator recovers capital and land elasticities under flexible production technologies when capital choices respond to latent local conditions. An application to newly constructed housing in Tokyo's 23 special wards illustrates how the method can be implemented in a dense single-city setting. The results show that explicitly modeling latent local heterogeneity matters for estimated capital-land elasticities and implies returns to scale close to one.

This paper studies when response time is informative about problem complexity. It revisits a canonical sequential-sampling model in which a decision-maker chooses when to stop acquiring costly information. Problem complexity is measured by the noise-to-signal ratio of the evidence process. Under exogenous stopping rules -- as when the decision-maker does not optimally adjust to problem complexity -- response time increases with complexity. By contrast, this monotonicity breaks down when the decision-maker observes problem complexity ex ante and optimally adjusts to it. Expected stopping time is then inverse-U-shaped in complexity, so choices are fast in both very simple and very complex problems. Ability and response time are similarly ambiguously related: more able decision-makers are faster on simple problems but slower on complex ones. Finally, this paper shows that complexity and ability can be inferred from the sensitivity of choices to subsidies, which is greater in more complex problems and for less able decision-makers.