arXiv subjects

Yongheng Hu

Publications and source records attributed to Yongheng Hu.

5 recordsLinked to original sources

arXiv2026

The Nexus between Dataization and Technological Progress in General Equilibrium of Macroeconomics

In this paper, we construct an analytical model of the data economy with empirical evidence to explain the nexus between dataization and technological progress in general equilibrium. Data originates from the dataization of firm total output and contributes to the formation and enhancement of technology. Firms use the production function with data to solve the optimal investment, while households use the endogenous interest rate from the firm problem to solve the optimal consumption. We find that dataization has a negative moderating effect on the transition of general equilibrium affected by technological progress. Policy can only facilitate a positive transition in general equilibrium by simultaneously encouraging dataization and technological progress. Furthermore, when equilibrium capital stock is in a stationary state, dataization enhances technological progress at high levels. However, when equilibrium consumption is in a stationary state, dataization enhances technological progress at low levels while weakening it at high levels. Our empirical analysis uses macroeconomic data and policy from Chinese cities between 2000 and 2021 to verify the theories proposed in this paper. We further apply the Mean Field Games in a continuous-time framework to provide an extended explanation for the nexus between dataization and technological progress in partial equilibrium.

econ.TH↗

arXiv2026

MCPO: Mastery-Consolidated Policy Optimization for Large Reasoning Models

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a promising approach to improve the reasoning abilities of Large Language Models (LLMs). Among RLVR algorithms, Group Relative Policy Optimization (GRPO) and its variants have demonstrated strong performance and high training efficiency. However, GRPO-style objectives exhibit two issues on high accuracy prompts including mastered prompts (rollout accuracy =1) and majority-correct prompts (rollout accuracy in (0.5,1)). For mastered prompts, group-relative advantages vanish, yielding no training signal and unconstrained policy drift that can cause forgetting. For majority-correct prompts, the induced query weight shrinks as accuracy increases, weakening consolidation from partial correctness to mastery. To alleviate this, we propose Mastery-Consolidated Policy Optimization (MCPO), which introduces (i) a hinge-KL regularizer applied exclusively to mastered prompts to bound harmful policy drift between successive gradient steps, and (ii) a weighting mechanism that prioritizes majority-correct prompts to better allocate optimization effort. Extensive experiments across three mathematical benchmarks demonstrate that MCPO consistently improves pass@1 performance. Counter-intuitively, rather than restricting exploration, MCPO boosts pass@k metrics, indicating that mastery consolidation further catalyzes solution diversity.

cs.AI↗

arXiv2026

Trade Dynamics with Heterogeneous Fluctuations

In this paper, we design two chapters to discuss trade dynamics with heterogeneous fluctuations, contributing new insights to macroeconomic issues related to international trade. In the first chapter, we model general exchange rate fluctuations through stochastic processes and analyze the impact of heterogeneous price shocks on export competitiveness. We find that monetary policy and innovation both show positive effects on export trade, while monetary policy stabilizes exchange rate fluctuations to comprehensively boost provincial export competitiveness, innovation reduces its reliance on exchange rate mechanisms. The optimal policy according to exchange rate fluctuations aims to solve the wealth distribution of exporters, and it suggests that optimal policy should promote dynamic transitions in trade patterns rather than maintain existing comparative advantages in heterogeneous trade structures. In the second chapter, we model labor market fluctuations and the ability to utilize production factors through stochastic processes, and we analyze the impact of heterogeneous aggregate production shocks on general international trade. We find that labor market fluctuations only benefit international trade under the cooperation policy. Moreover, for both sanction and cooperation policy scenarios, positive shocks (i.e., shocks where average wage growth in the labor market exceeds unemployment) strengthen their impact on import trade while weakening their impact on export trade, and vice versa. Regarding the theories proposed in these two chapters, we prove them through empirical analyses using the provincial data of China.

econ.EM↗

arXiv2025

Heterogeneous Agents in the Data Economy

In this short paper, we define the investment ability of data investors in the data economy and its heterogeneity. We further construct an analytical heterogeneous agent model to demonstrate that differences in data investment ability lead to divergent economic results for data investors. The analytical results prove that: Investors with higher data investment ability can obtain greater utility through data investment, and thus have stronger incentives to invest in a larger scale of data to achieve higher productivity, technological progress, and experience lower financial frictions. We aim to propose a prerequisite theory that extends the analytical framework of the data economy from the currently prevalent representative agent model to a heterogeneous agent model.

econ.TH↗

arXiv2025

How Big Data Dilutes Cognitive Resources, Interferes with Rational Decision-making and Affects Wealth Distribution ?

Big data has exponentially dilated consumption demand and speed, but can they all be converted to utility? We argue about the measures of consumption and utility acquisition in CRRA utility function under the condition of big data interaction, we indicate its weakness, i.e., irrational consumption does not lead to the acquisition of utility. We consider that big data, which is different from macro and micro economic signals, formed by general information entropy, affects agents' rational cognition, which makes a part of their consumption ineffective. We preliminarily propose the theory that how dilution mechanism driven by big data will affect agents' cognitive resources. Based on theoretical and empirical analysis, we construct the Consumption Adjustment Weight Function (CAWF) of agents interacting with big data and further apply it to a model of firm wealth distribution with financial frictions, we get analytical solutions according to the Mean Field Game (MFG) and find: Lower financial friction increases the average wealth of firms but also leads to greater wealth inequality. When agents convert effective consumption into utility, which is a weight of total consumption, the average wealth of firms increases with the weight increasing. Meanwhile, wealth inequality follows a U-shaped trend, and it will be the lowest level when the weight approaches to 0.5. In conclusion, we try to provide a new complementary hypothesis to refine the 'Lucas Critique' according to the cognitive resources as endowments involved in the decision-making of agents.

econ.GN↗