TL;DR
This paper introduces a novel approach to assess the environmental efficiency of Chinese industries by combining multiple Data Envelopment Analysis (DEA) models with a Gini criterion. This method provides a more robust and dynamic evaluation, revealing that industry environmental performance varies significantly over time and identifying specific sectors that are major pollution sources, offering crucial insights for targeted environmental policies.
What Problem Does This Paper Address?
China's rapid industrialization has led to significant environmental pollution, but the severity varies greatly across different industries. Existing studies often focus on regional levels or use single-method approaches, which can lead to biased or uncertain assessments of environmental performance. This lack of a nuanced, industry-specific, and dynamic understanding hinders the Chinese government's ability to formulate effective, targeted policies for green industrial development. The paper aims to fill this gap by providing a comprehensive, multi-faceted assessment of environmental efficiency at the industry level.
Methods
The core of this study lies in its innovative methodological approach:
- Multiple DEA Models with Gini Criterion: Instead of relying on a single DEA model, the researchers employed four different DEA models (super-efficiency BCC models and super-efficiency SBM models, both input/output-oriented and non-radial) to capture various facets of environmental efficiency. To address potential biases and uncertainties from individual models, they integrated these results using a Gini criterion. This criterion quantifies the "information complexity" or uncertainty in the efficiency scores from each model and assigns weights accordingly, leading to a more objective and comprehensive overall efficiency score for each industry.
- Hierarchical Cluster Analysis (Ward's Method): To understand the dynamics and patterns of environmental efficiency over time, the researchers applied hierarchical clustering. This technique groups industries with similar efficiency trends, allowing for the identification of distinct categories of environmental performance and pollution levels.
Datasets
The study utilized data for 36 non-agricultural industries in China from 2006 to 2015. The data was sourced from various official statistical yearbooks published by China Statistical Press, including:
- China Statistical Yearbook
- China Energy Statistical Yearbook
- China Labour Statistical Yearbook
- China Science and Technology Statistical Yearbook
- China Environmental Statistical Yearbook
The dataset included:
- Inputs: Labor force (quantity and quality), energy consumption, assets, and environmental governance inputs.
- Desirable Outputs: Industrial sales value and industrial profit.
- Undesirable Outputs: Waste gas emissions, wastewater discharge, solid waste production, and emissions of ammonia nitrogen, chemical oxygen demand, smoke and dust, sulfur dioxide, and carbon dioxide.
Key Insights and Ideas
- Dynamic Nature of Efficiency: The study strongly emphasizes that environmental efficiency is not static. Rankings and performance levels of industries change significantly year-to-year, underscoring the need for continuous monitoring and adaptive policy-making.
- Limitations of Single-Model Approaches: The paper highlights the inherent risks of relying on a single DEA model, demonstrating how combining multiple models with a Gini criterion can mitigate uncertainty and provide a more reliable assessment.
- Industry Heterogeneity is Crucial: The research moves beyond broad regional analyses to pinpoint specific industries that are major contributors to pollution, enabling more precise policy interventions.
- The "Pollution Haven" Effect in Reverse: The study implicitly touches upon how China, once a destination for polluting industries, is now grappling with its own domestic pollution challenges, necessitating a shift towards cleaner production.
Key Results
- Varied Efficiency Trends: The study identified six distinct patterns of environmental efficiency change over the 2006-2015 period: U-shaped, inverted U-shaped, W-shaped, growing, declining, and stable-fluctuating.
- Identification of High-Pollution Industries: Through clustering, 13 labor-intensive light industries and heavy chemical industries with high energy use and emissions were identified as medium- and high-pollution industries. These include sectors like paper and paper products, chemical materials, and special equipment manufacturing.
- Stable Green Industries: Only two industries, tobacco products (D09) and water production and supply (D36), were consistently classified as mild-pollution industries over the decade, indicating a stable green performance.
- Policy Implications: The findings directly inform policy by suggesting differentiated environmental policies for different industry clusters, emphasizing the need for targeted financial support, technology transfer, and stricter environmental supervision for high-pollution sectors.
Limitations and Future Work
The authors acknowledge several limitations:
- Data Availability: The study could only cover 36 out of 41 industries in the industrial sector due to data limitations.
- Influencing Factors: The paper did not delve into the specific influencing factors and mechanisms driving environmental efficiency changes within these sub-sectors.
- Detailed Directional Analysis: While key polluting industries were identified, a more detailed analysis of their specific industrial structures, development modes, and precise energy-saving/emission-reduction strategies was not undertaken.
Future research directions suggested include:
- Expanding the analysis to include all industrial sub-sectors and other economic departments.
- Investigating the factors that influence environmental efficiency across various industries.
- Conducting in-depth analyses of the industrial structure, development modes, and specific green strategies for the identified key polluting industries.
Summary updated