Abstract
As a new pattern of urbanization, smart cities offer a set of initiatives to achieve a win-win situation for both environmental protection and economic development. This paper aims to explore the influence of smart city policy (SCP) on green technology innovation (GTI) by Chinese listed enterprises. Employing the difference-in-differences (DID) approach, the results show that SCP promotes enterprises’ GTI. This finding remains valid after various robustness tests, which include the parallel trend test, PSM-DID, placebo test, the substitution of estimation model, replacement of dependent variable, and the exclusion of the impacts of other policies. Additional tests reveal that SCP improves GTI through government environmental subsidy, corporate environmental awareness, and research and development investments. Furthermore, the heterogeneity analysis suggests that our findings are more prominent in subsamples during the growth stage, in high-tech industries, and in eastern regions. Our research is vital to the urban ecological environment and high-quality economic development.
Acknowledgements
The authors are grateful to the Editor, as well as the anonymous referees for valuable suggestions and comments that helped us improve our paper significantly.
Author contributions
Chong Guo: conceptualization, methodology, data curation, writing-review, and editing. Yuelin Wang: methodology, data curation, and formal analysis. Yiteng Hu: conceptualization, methodology, data curation, writing-original draft. Yingyu Wu: conceptualization, methodology, writing original draft. Xiaobing Lai: conceptualization, formal analysis, methodology, and validation.
Data availability statement
The datasets used during the current study are available from the corresponding author upon reasonable request.
Disclosure statement
The authors report there are no competing interests to declare.
Notes
1 ICT is the abbreviation for information and communication technology.
2 The reason we put smart economic development as the first element is that a sound economic situation is an important prerequisite for the development of smart technologies and their application in the construction of smart cities. The second element is sustainability, which involves less waste and pollution, reduced emissions, better quality of life, better social life, and greater energy efficiency (Zubizarreta, Seravalli, and Arrizabalaga 2016; Ahvenniemi et al. Citation2017). The last element is smart governance. We include this element in the definition, because it plays a crucial role in China’s smart city construction. Current overall goals for Chinese smart city development are the ubiquitous services for citizens, transparent and efficient online government (Yao, Huang, and Zhao Citation2020; Yan et al. Citation2020). Similar to China’s other national projects, the implementation of smart city projects emphasizes the crucial role of government by specifying that smart cities should be promoted mainly in a top-down procedure (Wang et al. Citation2019).
3 ST represents special treatment and means a listed company with negative net profit for two years in a row. *ST indicates a listed company with negative net profits for three continuous years. PT stands for particular transfer, which stands for special transfer service for companies with three consecutive years of losses and suspended listing.
4 Since enterprises located in higher-ranking cities usually face higher requirements for low-carbon green development and have more policy resources, financial resources and talent resources, all of which would have an impact on the enterprise's GTI. To avoid these factors from interfering with our research results, enterprises in these cities were also eliminated from the sample.
5 Specifically, we generated a ‘pseudo-policy’ variable (Random-did) by creating a random treatment group and a policy shock point in time to conduct a placebo test. We first randomly selected 90 cities in the first batch of pilot cities as the ‘pseudo-treatment group’ and the remaining cities are used as the control group. Second, we could not determine the timing of the policy shock for the randomly generated treatment group, because the smart city pilots were multiple points in time rather than occurring in a specific year; therefore, we use the year of the first batch of smart city construction in 2012 as the ‘pseudo-policy shock year. On this basis, we multiply the above two dummy variables (Random-did) to replace the did in the original regression model. To avoid the interference of other small probability events on the estimation results, we repeated the above regression process 500 times; the results are shown on the left side of . To further ensure the robustness of the results, we increased the regression process 1,000 times, and the results are shown on the right side of . It can be seen that the regression coefficients of Random-did in both graphs basically obey normal distribution around the value of 0, indicating that the effect of randomly constructed SCP on GTI is not significant, which verifies that the promotion effect of SCP on enterprise GTI is not caused by random factors.
6 To avoid the influence of heteroskedasticity, GTI in the benchmark regression, GTI-1, GTI-2 and GTI-3 are all processed with logarithm. However, logarithm treatment may cause structural changes in the data, which may lead to the “pseudo-regression” problem. Therefore, we did not take the logarithm of GTI-4 and define it as ‘the number of green patent applications’, the regression result is consistent with the baseline regression.
7 Referring to the approach employed by Chen (Citation2022), Gal et al. (Citation2019), and Tian, Li, and Cheng (Citation2022), to measure corporate environmental awareness, we first established an initial text lexicon of corporate environmental awareness based on important governmental documents such as the “State Council’s Guidance on Accelerating the Establishment of a Sound Economic System of Green, Low-Carbon and Circular Development” and the ‘Thirteenth Five-Year Development Plan of National Environmental Protection Standards’. Second, we expanded the textual lexicon of corporate environmental awareness with the help of machine learning methods. As the selection of the initial text lexicon from policy documents is subjective, we used the Python “synonyms” recognition package to expand the text lexicon. Based on this, we used Python’s Jieba word splitting package to search and match keywords in the full text of listed companies’ annual reports based on the previous text lexicon, further identifying words with a high correlation coefficient with the keywords contained in the text lexicon and incorporating them into the lexicon as new corporate environmental awareness keywords, resulting in 290 corporate environmental awareness-related keywords (see Appendix). Finally, we aggregated the frequency of keywords appearing in companies’ annual reports regarding corporate environmental awareness to derive the degree of corporate concern in the area of environmental protection and used it as a proxy variable for corporate environmental awareness.
8 Referring to the “Administrative measures for the recognition of high-tech enterprises” issued by the Chinese Ministry of Science and Technology, we classify the software and information technology service industry, computer, and research and experimental development as high-tech industries, and the rest as non-high-tech industries.