Search

Under the pressures of environmental pollution and overcapacity, environmental regulation is adopted by the Chinese government to reduce excess capacity of China’s manufacturing industry. Sector-level empirical research on the impact of environmental regulation on overcapacity is necessary since both of them are highly variable among different manufacturing sectors. Based on the panel data from 2001 to 2010, capacity utilization and its two factors, namely technical efficiency and equipment utilization, of 28 manufacturing sectors in China have been calculated using data envelopment analysis (DEA). The results of empirical tests indicate that environmental regulation has a significant impact on equipment utilization rather than on technical efficiency, leading to an overall negative effect on capacity utilization. In addition, the impact of environmental regulation on capacity utilization shows sector heterogeneity when the manufacturing sectors are classified according to the degree of overcapacity or pollution. To be specific, environmental regulation is not an effective means of reducing excess capacity of mild-pollution sectors and severe-overcapacity sectors. Capacity utilization of moderate- and severe-pollution sectors can be increased by combining environmental regulation and other policies to stimulate technical innovation. Excess capacity of mild- and moderate-overcapacity sectors can be reduced by environmental regulation; meanwhile, policy guidance on promoting technical innovation should also be provided to mild-overcapacity sectors. Therefore, it is crucial to formulate differentiated policies and maintain a reasonable range of intensity in the process of reducing overcapacity by means of environmental regulation.

In this paper, we proposed an assessment system of forest environmental carrying capacity from many aspects and comprehensively evaluated and predicted the forest environmental carrying capacity of 40 cities in the Yangtze River Delta of China by using the proposed deep learning-based model. In addition, the proposed model is used to dynamically evaluate the comprehensive scores of forest environmental carrying capacity of 34 provinces and cities in China between 2015 and 2020. This paper also has the following contributions: (1) a deeply integrated unidirectional bidirectional LSTM considering the correlation of time series is proposed. The bidirectional LSTM network is used to deal with the backward dependence of input data to prevent the omission of some useful information, which is beneficial to the prediction results. (2) In terms of missing data processing, the Mask layer is added to subtly skip the processing of missing data, which will help to enhance the evaluation results. [ABSTRACT FROM AUTHOR]