This study examines how peer effects, as measured by productivity changes following a shock to human capital, vary with the intellectual structure of scientific partnerships. We measure this structure using a new index of “cognitive distance” between collaborators, derived from their publication and citation patterns. To identify causal effects, we use a difference-in-differences approach that exploits the unexpected deaths of eminent life scientists as a natural experiment. Our analysis reveals

本文将政策学习（Policy Learning）前沿方法引入中国特定政策的优化设计过程，聚焦于开发区政策的福利效果，运用政策学习方法探讨如何通过政策优化与精准实施，实现社会福利改进的重要目标。本研究在评估现有开发区政策的平均消费带动作用的基础上，运用政策学习探讨了如何依据城市特征设立国家级开发区以实现城市消费福利改进，并进一步分析了当前政策区域分散倾向与区域协调发展的联系。研究发现，国家级开发区的设立对城市消费水平具有显著的正向促进作用，通过政策学习方法，重新设定的象限处理规则、线性处理规则和决策树处理规则等政策处理规则均能带来显著的福利改进和更低的政策成本，体现出“低成本、高回报”的特点。例如，在象限处理规则中以财政一般预算收入和人均 GDP 为分配变量设立国家级开发区，其城市消费水平平均提升 3255 元，较当前政策实施路径的福利改进幅度约为 55.82%。此外，本文还探讨了开发区政策与区域协调发展的关联，发现东部地区的消费带动作用的福利改进效率均明显高于中西部地区，但在特定政策变量下，政策学习可以使得中西部地区的福利改进效率更高，从而有利于区域协调发展。本研究为实现中国政策的优化

We develop a two-layer equilibrium model of foundational AI that links downstream pricing of model capability to upstream procurement of pretraining data through neural scaling laws. Downstream, where AI acts as a productivity floor for knowledge workers, competition exhibits quality-price coupling: small capability advantages allow leaders to limit-price rivals, generating endogenous winner-take-all outcomes. Upstream, we identify a novel reversed information asymmetry: model producers can veri

Governments worldwide frequently rely on official transfers from developed to underdeveloped regions to promote local economic growth, yet little is known about how institutions shape the persistence of such growth. We develop a dynamic model of local governance to highlight the role of long-term institutional reform in translating officials’ governance capacity into sustained economic performance. Using a multi-period difference-indifferences strategy and data on prefecture-level city leaders i

The rapid advancement and widespread diffusion of artificial intelligence (AI) are transforming labor markets and reshaping traditional patterns of factor income distribution. Using textual data from annual reports of Chinese listed firms and AI-related regulatory documents, this paper constructs a firmlevel measure of AI adoption and examines its impact on the labor income share. Results show that AI significantly increases labor’s share within firms, with a 10-percentage-point rise in AI inten

Our work investigates the adoption of AI technology by firms in a duopoly competitive environment with multiple products, particularly focusing on the optimal pricing strategy for AI companies when their technology could significantly impact the downstream market. AI technology enhances product quality, influencing consumer willingness to pay, but also increases unit costs for users and affects competitive dynamics across market segments. We explore a scenario where a technological leader and fo

The rapid deployment of Large Language Models and AI agents across critical societal and technical domains is hindered by persistent behavioral pathologies including sycophancy, hallucination, and strategic deception that resist mitigation via reinforcement learning. Current safety paradigms treat these failures as transient training artifacts, lacking a unified theoretical framework to explain their emergence and stability. Here we show that these misalignments are not errors, but mathematicall

GDP is a vital measure of a country’s economic health, reflecting the total value of goods and services produced. Forecasting GDP growth is essential for economic planning, as it helps governments, businesses, and investors anticipate trends, make informed decisions, and promote stability and growth. While most previous works focus on the prediction of the GDP growth rate for a single country or by machine learning methods, in this paper we give a comprehensive study on the GDP growth forecastin

We propose a deep learning framework, DL-opt, designed to efficiently solve for optimal policies in quantifiable general equilibrium trade models. DL-opt integrates (i) a nested fixed point (NFXP) formulation of the optimization problem, (ii) automatic implicit differentiation to enhance gradient descent for solving unilateral optimal policies, and (iii) a best-response dynamics approach for finding Nash equilibria. Utilizing DL-opt, we solve for non-cooperative tariffs and industrial subsidies