Navigating the digital divide: unraveling the impact of ICT usage and supply on SO2 emissions in China’s Yangtze River Delta

Descriptive statistics and correlation matrix

Descriptive statistics are crucial for understanding data, identifying outliers, and making initial observations. Table 2 exhibits descriptive statistic values of the variables. The mean SO₂ value during the sample period is 5.426, with a standard deviation (SD) of 6.409. Notably, Shanghai reported the lowest SO₂ values from 2013 to 2018, with a minimum of 0.001. These lower values place it among the least SO₂-producing cities in the YRD. Although most of the cities in YRD have reduced their overall emissions since the 2000s due to technological advancements, some others remain high compared to other cities. Despite the lowest value being reported for Shanghai, the maximum SO₂ value, 37.871, is also attributed to Shanghai in 2012, with other top-emitting cities including Suzhou, Zhenjiang, Ma’anshan, Tongling, Jiaxing, and Huainan.

Furthermore, the ICTU exhibits mean 0.139 and SD 0.082 values in the dataset. However, cities like Nantong, Yancheng, Ma’anshan, Chizhou, Anqing, Suqian, and Bengbu show relatively lower ICTU levels, with Luan having the lowest value of 0.028 in 2012. It could be due to a lack of necessary arrangements for widespread ICTU adoption or limited resources dedicated to these cities. Conversely, Nanjing reported the highest ICTU value of 0.424 in 2018, while Hangzhou and Shanghai also demonstrated higher ICTU levels. It is attributed to advanced infrastructure and robust economic development. These cities typically invest more in technology and have better access to high-speed Internet. Supportive government policies further encourage the adoption of ICTU across various sectors, fostering an environment conducive to technological innovation, productivity, and competitiveness.

The mean of ICTS is 2.569, with an SD of 7.358, indicating diverse levels of ICT supplies across different cities in the panel. Notably, cities such as Chuzhou and Xuancheng exhibit lower ICTS values, implying potential challenges in employing the workforce in the industry. The city of Chuzhou reported minimum values of 0.051 in 2018 in the YRD dataset. On the other hand, Nanjing and Shanghai consistently demonstrate high ICTS values, with the latter achieving a maximum value of 65.870 in 2019. This output suggests that urbanization in Shanghai is higher compared to other cities in China because it is among the most developed cities and serves as a major financial and economic hub. It attracts a large number of people from rural areas seeking better job opportunities, well-developed infrastructure, and a skilled workforce available to be employed in the ICT sector.

The SDI average is 0.525, with an SD of 0.198, while Luan reported the lowest value of 0.183 in 2016. Apart from Luan, Chizhou, Zhoushan, Lishui, and Bozhou cities also show lower values for SDI, which implies the shift towards automation and technological advancements, leading to a decrease in the need for manual labor. Also, rising labor costs and competition from other cities with lower labor costs may contribute to lower employment in these cities. However, SDI values for all cities exhibit a positive trend, reaching a highest of 2.870 in 2017. The mean GDP is 0.091 with an SD of 0.026, reflecting the economic activity level of YRD cities, which constitute 20% of the Chinese economy in 2022. The cities of Huai’an, Tongling, Huainan, and Tongling reported negative values once, with the lowest of −0.063 in 2017 for Huai’an. Other than that, GDP revealed positive outcomes, which demonstrate strong infrastructure, technological advancements, a skilled workforce, stable political environments, and access to capital contribution to YRD cities. The mean INS is 0.228 with an SD of 0.181; Hefei reports a maximum value of 1.051, highlighting its significant investment portfolio. Tongling, Chizhou, and Huangshan have lower values for INS, but Quzhou exhibited a minimum value of 0.001 in 2018. The POP average value is 6.390, with an SD of 0.569. The minimum reported value of POP is 4.977 for Lishui, and the highest is 8.272 for Shanghai. YRD is one of the most densely populated regions in China and includes one of the world’s largest cities.

The correlation outcomes are presented in Table 3, indicating that multicollinearity issues are less likely to arise since none of the correlation coefficients are over 0.70.

Furthermore, Fig. 2 shows that around 2011, the haze problem in Zhejiang, Nanjing, Hefei, Maanshan, and Wuxi threatened the population’s health. It was formerly one of the most pressing problems in China; however, SO₂ emissions have declined, suggesting that the country’s air quality has improved from 2015 to 2019.

Spatial distribution of SO₂.

In addition, the initial phase of the sample period indicates that YRD cities are located in a highly developed region, including Shanghai, Hangzhou, Suzhou, Nanjing, Wenzhou, Wuhan, Chongqing, and Chengdu (Fig. 3). These cities represent the central hubs within major urban agglomerations, encompassing municipalities under the central government, provincial capitals, and selected highly developed urban centers. In 2015, the overall performance of ICTU had improved. Zhejiang Province has the highest percentage of rapid and advanced development. In 2019, the value of ICTU was highest on the whole, and the majority of cities in YRD appear to have a strong cluster pattern and are in the fast promotion and highly advanced areas.

Spatial distribution of ICT usage.

Regression results

The current research explores the extent to which ICTU and ICTS impact SO₂ emissions in YRD. Table 4 provides the results of panel RE-GLS regression (columns 1–3) and SDM (columns 4–5). In column 1, SO₂ emission regresses with control variables, i.e., GDP, SDI, POP, and INS. Column 2 added ICTU and ICTS, and the outcomes of the inverted U-shape relationship were reported in column 3. The SDM results are mentioned in columns 4 and 5, in which the spatial spillover effect of ICTU and ICTS were examined.

The coefficient value of ICTU and its square term under both models reveal an inverted U-shape relationship with SO₂ emission. Based on this result, Hypothesis 1 is supported, which suggests that the impact of ICTU on SO₂ emission has an inverted U-shaped relationship. The outcome is supported by the earlier study of Cheng et al. (2023), who found a significant inverted U-shaped relationship between ICTU and SO₂ emissions in China.

Furthermore, the outcome shows that ICTS is negatively and significantly related to SO₂ emissions. The findings indicate that ICTS helps to reduce SO₂ emission at 0.01 significance levels, which confirms hypothesis 2. The results are endorsed by the prior study of Khan et al. (2022), who claimed that ICTS should be promoted to diminish emissions and enhance environmental conditions. At the same time, the outcomes contradicted the previous study by Weili et al. (2022), which stated that ICTS raised emission levels.

The spatial spillover effect of ICTU on SO₂ pollution was found to be insignificant, which does not provide sufficient proof to support H3. Meanwhile, the spillover effect of ICTS on SO₂ pollution has a significant negative impact at a 5% significance level. The finding endorsed Hypothesis 4.

Robustness results

To further support the study’s findings, a dynamic panel model was employed to address the issue of endogeneity, with the outcomes reported in Table 5. Based on the findings, the signs and significance of almost all variables are coherent with the regression results presented in Table 4, showing the robustness of our research.