Environmental Regulatory Disparities and the Spatial Reallocation of Factors: Evidence from China’s National Air Quality Monitoring Network

Technical Summary: Environmental Regulatory Disparities and the Spatial Reallocation of Factors

Problem Statement
This study addresses the structural dilemma facing China's urbanization process: while industrialization has driven massive factor migration from rural to urban areas, the resulting unidirectional concentration has weakened the endogenous development capacity of counties, exacerbating urban–rural disparities. The paper investigates how environmental regulation, specifically the spatially asymmetric deployment of China's National Air Quality Monitoring Network, influences the spatial reallocation of industrial firms and labor. It seeks to explain why a seemingly uniform national policy induces directional factor flows, hypothesizing that the uneven density of monitoring stations (dense in urban cores, sparse in counties) creates regulatory differentials that alter location costs and trigger a "push-pull" dynamic between urban and county regions.

Methodology
The authors employ a quasi-natural experiment design leveraging the phased rollout of the national air quality monitoring network (2013–2023) across 2,598 Chinese counties.

• Data Sources: The study utilizes a panel dataset combining firm registration data from the National Enterprise Credit Information Publicity System (2013–2023), county-level socioeconomic data from the China County Statistical Yearbook, environmental enforcement data from the PKULaw database, and industrial land market data from the China Land Market Network.
• Identification Strategy: The primary empirical approach is a staggered difference-in-differences (DID) model. The treatment is defined by the establishment of the first monitoring station within a prefecture-level city, with the policy shock measured at the county level. The model distinguishes between urban districts and counties to capture heterogeneous effects.
• Robustness and Advanced Techniques:
  • Spatial Spillover Analysis: The authors construct distance-gradient models to examine effects within 4km, 8km, and 12km radii of monitoring stations.
  • Double Machine Learning (DML): To address potential non-linearities and high-dimensional covariates, the study employs a partially linear DML framework using Random Forest, Lasso, XGBoost, and Gradient Boosting algorithms to estimate treatment effects and identify heterogeneous factors.
  • Mechanism Testing: The paper analyzes environmental administrative penalties to test for strategic regulatory responses by local governments and examines industrial land transaction data to distinguish between autonomous firm location choices and government-led investment promotion.
  • Validation: The study conducts parallel trends tests, placebo tests (random assignment of pseudo-stations), and Bacon decomposition to assess biases from heterogeneous treatment timing.

Key Contributions

1. Joint Mobility Framework: Unlike existing literature that often treats firm and labor mobility as independent processes, this paper develops a conceptual framework integrating the joint mobility of firms and workers. It demonstrates how firm relocation expands local market size and reduces living costs in counties, creating circular cumulative effects that attract labor inflows.
2. Policy Implementation Perspective: The study shifts the focus from the existence of regulation to its spatial implementation. It elucidates how the interaction between the asymmetric spatial distribution of monitoring infrastructure and local government incentives (dual goals of growth and environmental performance) shapes regulatory intensity, thereby driving factor reallocation.
3. Dynamic Spatial Restructuring: By utilizing a staggered DID approach, the paper captures the dynamic evolution of factor reallocation, moving beyond static snapshots to reveal a pattern of "urban dispersion and county agglomeration."

Empirical Results

• Factor Reallocation: The establishment of monitoring stations significantly reduces industrial firm density and labor density in urban areas while increasing both in counties. Specifically, urban industrial firm density decreases by approximately 0.42 firms per square kilometer, while county density increases by 0.09. Similarly, urban labor density declines, while county labor density rises.
• Spatial Heterogeneity: The effects exhibit a clear distance-decay pattern. The negative impact on urban firms is strongest within 4km of monitoring stations. Conversely, counties show significant agglomeration effects, particularly for manufacturing firms.
• Mechanism of Strategic Regulation: Evidence indicates that local governments respond strategically to the monitoring network by tightening enforcement in urban cores (where accountability pressure is high) while maintaining relatively softer constraints in counties. This regulatory disparity lowers the relative compliance costs for firms in counties, driving relocation.
• Market-Driven Relocation: Analysis of industrial land markets reveals that the agglomeration in counties is accompanied by rising land prices and increased market-based allocation (public listing). This suggests the relocation is driven by firms' autonomous cost-minimization choices rather than government-subsidized land promotion.
• Heterogeneity:
  • Industrial Structure: The agglomeration effect is strongest in counties with lower tertiary sector shares (dominated by manufacturing). In highly urbanized areas with large tertiary sectors, the policy induces a strong dispersion effect.
  • Urbanization Level: Higher urbanization levels in cities correlate with stronger dispersion effects. In counties, agglomeration effects emerge once urbanization surpasses a certain threshold (approx. 30%), reflecting improved infrastructure.
  • Ecological Constraints: The agglomeration effect is significantly weaker in counties with national-level nature reserves due to rigid ecological constraints, whereas it is strongest in areas with lower-level or no reserves.

Significance and Claims
The paper claims to provide critical insights for designing spatially differentiated environmental policies in emerging economies. It argues that the "urban-dense, county-sparse" monitoring configuration inadvertently creates a regulatory arbitrage that facilitates the spatial rebalancing of production factors. The findings suggest that environmental regulation is not merely a constraint but a structural force reshaping regional economic landscapes.

The authors conclude that policy should move beyond uniform regulatory standards. Instead, they advocate for differentiated strategies: maintaining strict standards in core urban areas to drive green upgrading, while dynamically adjusting standards in counties to guide selective, green-oriented industrial upgrading. Furthermore, they emphasize the need to integrate ecological constraints (e.g., nature reserve boundaries) into spatial planning to prevent the transfer of pollution to ecologically sensitive areas, and to establish market-based green incentive systems to shift firm behavior from regulatory avoidance to green innovation.