This study pioneers investigating the environmental Kuznets curve (EKC) hypothesis and its possible regional differential characteristics in China. Based on 2003e2017 years of provincial panel data, this research employs panel smooth transition auto regression (PSTR) model to analyze the impact of income levels on environmental pollution and identify the EKC threshold of energy intensity. Then, 30 Chinese provinces were categorized by energy intensity to examine the inter-provincial and interregional differences in the EKC threshold of energy intensity, rather than the traditional threshold of income. The results indicate that China’s pollution emissions and energy intensities show a stepwise decreasing pattern from the western region to the eastern region. In addition, the impact of income levels on pollution emissions is non-linear, and the critical value of energy intensity is 0.9168 between the highand low regimes. An inverted U-shaped environmental Kuznets curve is accepted for energy intensity, with 0.7670 as its threshold value. When the energy intensity is higher (lower) than the threshold value, the income elasticity of pollution emission is positive (negative). The more developed provinces and municipalities mainly in the eastern region are proved to exceed the threshold than the provinces in the central and western regions within the sample period. Given these findings, this study further divides China’s provinces into eco-friendly, low-pollution, and high-pollution provinces, and accordingly, the most important policy recommendations were discussed for policy-makers and researchers. This research can provide a new insight to investigate countries with unbalanced development levels within and a reference for environmental governance policies in sub-regions.

Introduction

Over the last few decades, since China join WTO, it has experienced substantial growth in its social and economic development and human welfare, which, in turn, has increased its demand for energy, especially the fossil fuels. Regardless of the large efforts made by China to decrease its pollution and increase the role of renewable energy, energy efficiency, and energy conservation, fossil fuels still represent the dominant source of energy use, accounting for almost 70% of the total domestic energy consumption (World Development Indicators, 2018). Therefore, the world has witnessed a large environmental degradation problem in China, which is also one of the major concerns that countries around the globe, including both developed and developing ones, are currently confronted with. This dilemma has made many researchers interested in exploring the relationship between macroeconomic variables (GDP in particular), pollution emission, and energy consumption among different countries using different econometric methodologies. Among the wide range of the literature that has investigated the relationship, one of the most discussed issues is the environmental Kuznets curve (EKC) hypothesis, analogous to the pattern Kuznets (1955) found between income inequality and economic development. According to this EKC-hypothesis, environmental pressure tends to rise faster than income growth in the early stages of economic development, then slows down, reaches a turning point and declines with further income growth (Rothman and De Bruyn, 1998).