This study proposes a framework for estimating demand in differentiated product markets with high dimensional product characteristics, building upon the seminal Berry, Levinsohn, and Pakes (1995) model, using market level data. We allow for a very large set of potential product characteristics, where the number of characteristics may exceed the number of market observations. Our contributions are twofold. First, we establish a general estimation theory for BLP models featuring high-dimensional nuisance parameters. We propose a Neyman orthogonal estimator specifically adapted to this framework, utilizing machine learning techniques, such as Lasso, to construct nuisance parameter estimators that are plugged into the Neyman orthogonal estimator. This approach offers a significant advantage: it achieves $\sqrt{T}$-asymptotic normality for parameters of interest--such as the price coefficient and price heterogeneity--even when nuisance parameters are estimated at slower rates due to their high dimensionality. Second, we apply this theory to a specialized BLP model under approximate sparsity, developing an estimation strategy for the high-dimensional nuisance parameters. The approximate sparsity condition posits that nuisance parameters can be controlled, up to a small approximation error, by a small and unknown subset of variables. In an economic context, this implies that while products have a vast array of characteristics, consumers focus on only a small subset of these due to bounded rationality. This condition makes the recovery of parameters of interest feasible by enabling nuisance parameter estimators to converge at the required rates. The practical performance of the method is evaluated through comprehensive Monte Carlo simulations, which demonstrate its efficacy in finite samples.

This paper develops a Hall-Sandpile model of economic instability that combines a Hall-like transversal stress mechanism with sandpile threshold dynamics on a real production-network substrate. In analogy with the physical Hall effect, where exposed flows under an external field generate stress in a transversal direction, we model economic shocks as fields that act on flow-intensive, low-redundancy, low-capacity nodes and produce systemic stress through a multiplicative conversion function. The accumulated stress drives a discrete toppling rule and an avalanche dynamics whose effective activation threshold declines with transversal exposure. The model is calibrated on annual World Input--Output Database (WIOD) production networks for 2000--2014 and simulated on the 2014 substrate (2{,}283 country--sector nodes) under three alternative propagation normalisations to avoid mechanical near-criticality from row-stochastic operators. Controlled Monte Carlo experiments over external field intensity and redundancy stress generate four ordered regimes: stable absorption, latent fragility, critical transition, and avalanche regime. Mean avalanche size and the probabilities of finite-size systemic events $\Pr(S\!\geq\!5)$, $\Pr(S\!\geq\!10)$ and $\Pr(S\!\geq\!20)$ rise jointly with field intensity and redundancy stress. Tail diagnostics show regime-dependent thickening of the avalanche distribution, but the estimated tail indices remain too high to interpret as evidence of universal power-law criticality. The contribution is therefore a finite-size, real-network description of how transversal stress activates structural fragility, not a claim of self-organised criticality in the global economy.

We revisit games in partition function form, i.e. cooperative games where the payoff of a coalition depends on the partition of the entire set of players. We assume that each coalition computes its worth having probabilistic beliefs over the coalitional behavior of the outsiders, i.e., it assigns various probability distributions over the set of partitions that the outsiders can form. These beliefs are not necessarily consistent with respect to the actual choices of the outsiders. We apply this framework to symmetric partition function form games characterized by either positive or negative externalities and we derive conditions on coalitional beliefs that guarantee the non-emptiness of the core of the induced games.

Offline evaluation of language models from usage logs is biased when model choice is confounded: the same user-side factors that influence which model is used can also influence how its output is judged, so raw comparisons of logged scores mix self-selected populations rather than estimating a common quantity of interest. A small randomized experiment can break this bias by overriding model choice, but in practice such experiments are scarce and costly. We study a three-source design that combines a large confounded observational log (OBS) for scale, a small randomized experiment (EXP) for unconfounded scoring, and an offline simulator (SIM) that replays candidate models on cached contexts. Our main result is an identification theorem showing that the randomized experiment and the simulator are together enough to recover causal model values; the observational log enters only afterward, to reduce estimation error rather than to make the causal comparison valid. Six estimator families are evaluated in a controlled semi-synthetic validation and in two real-task cached benchmarks for summarization and coding. No family dominates every regime; relative performance depends on the amount of unbiased EXP supervision and on how closely the target reward aligns with OBS-derived structure.

Geographic constraints have long structured access to high-growth career opportunities, concentrating upward mobility within a limited set of cities and organizations. The expansion of remote work potentially alters this opportunity structure by decoupling job matching from physical proximity, yet its implications for career mobility remain unclear. Using 48 million U.S. job transitions between 2020 and 2024 linked to employer-level measures of remote eligibility, we estimate how entering remote-eligible jobs shapes career outcomes at job transitions. Workers entering remote-eligible jobs experience significantly higher wage growth and higher rates of upward seniority mobility than comparable workers entering fully on-site roles. These transitions are also associated with greater cross-metropolitan job mobility and moves toward smaller, less prestigious employers. Importantly, effects are largest among lower-income workers and those originating from regions with limited high-skill opportunity density. Together, the findings indicate that remote work relaxes geographic constraints in job matching, reshaping the distribution of upward mobility across places and workers.

We study networked binary classification on a directed acyclic graph (DAG) where each agent observes only a subset of the feature columns of a shared dataset. Agents act sequentially along the DAG: each receives prediction columns from its parents (if any), augments its local features with these columns, fits a logistic predictor by minimizing binary cross-entropy (BCE), and forwards its prediction column to its outgoing neighbors. We ask whether this sequential distributed training procedure achieves information aggregation, meaning that some agent attains small excess loss compared to the best logistic predictor trained with access to all feature columns. This question was studied for linear regression under squared loss by Kearns, Roth, and Ryu (SODA 2026). Extending their guarantees to classification is nontrivial because their analysis relies on quadratic structure that does not directly transfer to BCE with a logistic link. We analyze the resulting sequential logit-passing protocol and prove: (i) an excess loss upper bound of $O(M/\sqrt{D})$ on depth-$D$ paths under the condition that every $M$ contiguous subsequence of $M$ agents collectively observe all features, and (ii) a close lower bound showing instances with excess loss of at least $Ω(k/D)$ where $k$ is the dimension of the feature space. Together, these results identify network depth as a fundamental bottleneck for information aggregation in networked logistic regression.

This study presents a novel, AI-driven framework for assessing Environmental, Social, and Governance (ESG) performance in European small and medium-sized enterprises (SMEs). An initial phase established expert-validated ESG baseline scores from a subset of the Flash Eurobarometer FL549 survey data. In the second phase, a scalable AI agent system, built on the n8n automation platform, applied these baselines to perform automated ESG classification and generate contextual recommendations using large language models (LLMs). The results demonstrate the AI system's high consistency with human-derived outputs, thereby supporting more effective monitoring and intervention strategies aligned with the European Green Deal.

Brazil's working-time debate is no longer only a choice between keeping the 44-hour week and moving directly to 36 hours. Alternatives around 40 hours, a five-day schedule and phased transitions are also on the table. This policy note asks a simple question for that choice: how much more productive would the economy need to become for each option not to reduce output in the short run? To answer, I combine Brazilian data on hours worked, informality, firm size and sectoral composition with a model of adjustment between formal and informal employment. The main result is that a move to 40 hours requires a productivity gain of about 2 percent. A direct move to 36 hours requires a much larger jump, between 6.6 and 8.2 percent, which is high relative to Brazil's recent productivity record. Informality also rises in the 36-hour scenario, by about 1.6 to 1.9 percentage points, but the main cost comes from fewer formal hours worked. The exercise does not say whether the reform should or should not move forward; it shows that size, timing and transition instruments change the arithmetic substantially. For policymakers, the message is direct: a phased route, with a stop near 40 hours, requires a much smaller productivity target than an immediate jump to 36 hours.

This paper develops an agentic framework that employs large language models (LLMs) for grounded persuasive language generation in automated copywriting, with real estate marketing as a focal application. Our method is designed to align the generated content with user preferences while highlighting useful factual attributes. This agent consists of three key modules: (1) Grounding Module, mimicking expert human behavior to predict marketable features; (2) Personalization Module, aligning content with user preferences; (3) Marketing Module, ensuring factual accuracy and the inclusion of localized features. We conduct systematic human-subject experiments in the domain of real estate marketing, with a focus group of potential house buyers. The results demonstrate that marketing descriptions generated by our approach are preferred over those written by human experts by a clear margin while maintaining the same level of factual accuracy. Our findings suggest a promising agentic approach to automate large-scale targeted copywriting while ensuring factuality of content generation.

Sequential search models provide a powerful framework for studying consumer search using rich data that records the sequence of consumer actions taken during the search process. In existing empirical applications, their implementation often builds on optimal policies, in which later decisions depend on outcomes from earlier actions that are often fully observed by researchers. Therefore, implementation is largely restricted by computation burden and limited model flexibility. This paper establishes a theoretical equivalence showing that, under common and mild assumptions of Independence and Invariance, a sequential search process is optimal if and only if a corresponding ranking over all feasible actions throughout the process holds, thereby introducing a ranking representation of optimal sequential search. This representation enables a novel, simple, and unified empirical strategy for implementing sequential search models. For the classic \cite{weitzman1979optimal} model, the proposed approach reduces simulation requirements while improving accuracy, computational efficiency, and ease of implementation. We further show that the same strategy extends to a broad class of sequential search settings, including partially observed action sequences and multi-stage information acquisition, such as discovery. Overall, the results enhance both the tractability and the empirical applicability of sequential search models.

We study how coordinated disinformation campaigns affect elections. We develop a constrained information design model in which a sender deploys uninformative messages that mimic voters' exogenous informative signals. Voters initially opposed to the sender's preferred outcome receive favourable messages, while those in favour are targeted with unfavourable messages to dilute adverse information. The sender's ability to manipulate political outcomes increases with greater precision of voters' independent signals, but decreases with polarisation. When messaging is costly, the sender may stop targeting marginally opposing voters while moderating message extremism among supporters.

We propose a novel method for estimating nonseparable selection models. We show that, for a given selection function, the potential outcome distributions are nonparametrically identified from the selected outcome distributions and can be recovered using a simple iterative algorithm based on a contraction mapping. This result enables a full-information approach to estimating selection models without imposing parametric or separability assumptions on the outcome equation. We propose a two-step estimation strategy for the potential outcome distributions and the parameters of the selection function and establish the consistency and asymptotic normality of the resulting estimators. Monte Carlo simulations demonstrate that our approach performs well in finite samples. The method is applicable to a wide range of empirical settings, including consumer demand models with only transaction prices, auctions with incomplete bid data, and Roy models with data on accepted wages.

We develop a unified estimation and inference framework for dyadic network formation with individual fixed effects, covering both transferable-utility (TU) and nontransferable-utility (NTU) links under general link functions. Under NTU, bilateral consent makes the fixed effects non-additive and the log-likelihood non-concave in the high-dimensional fixed effects, so differencing and profile-likelihood methods fail. We combine a joint method-of-moments initial estimator, a Le Cam one-step refinement, and a split-network jackknife bagging step that removes the incidental parameter bias without inflating variance. The resulting homophily estimator is asymptotically normal, unbiased, and attains the Cramér--Rao lower bound without requiring the log-likelihood to be concave in the fixed effects; we extend the theory to average partial effects and establish robustness to link-function misspecification. Simulations under both TU and NTU designs confirm these predictions. Applied to Thai village networks (TU), kinship and wealth differences both increase linking; in the Nyakatoke risk-sharing network (NTU), wealth differences have no significant effect, mirroring the two regimes' distinct logics.

I study a principal-agent model in which a principal hires an agent to collect information about an unknown continuous state. The agent acquires a signal whose distribution is centered around the state, controlling the signal's precision at a cost. The principal observes neither the precision nor the signal, but rather, using transfers that can depend on the state, incentivizes the agent to choose high precision and report the signal truthfully. I identify a sufficient and necessary condition on the agent's information structure which ensures that there exists an optimal transfer with a simple cutoff structure: the agent receives a fixed prize when his prediction is close enough to the state and receives nothing otherwise. This condition is mild and applies to all signal distributions commonly used in the literature.