Evaluation and Influencing Factors of Regional Green Innovation Efficiency Based on the Lasso Method
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe article is written very logically. The authors have a clear understanding of the research topic. A thorough analysis of the literature has been carried out. However, it seems that the authors have studied the publications of Chinese scholars. Perhaps this is due to the fact that the study concerns the Yangtze River.
The methodology and the analysis itself are very interesting. The authors note that: ‘provides experience and reference for further research on regional green innovation efficiency...’ However, the authors did not disclose this in the article.
‘time span’ should be replaced by ’period’
Author Response
Thanks for the comments. All the responses are shown in the next file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors1. In terms of Introduction, the research issue should be more clearly clarified.
2. In terms of Literature Review, please review the connotation of green innovation, the relationship between green development and green innovation, etc, not just list the references. Besides, you should summarize your main research innovation after reviewing related studies.
3. In terms of research methods, i doubt whether this Lasso Regression Model can be applied into panel data, since your green innovation efficiency and its influencing factors is balanced panel data.
4. In terms of Index Selection, why just select the undesirable output includes the total amount of emissions and wastewater discharge in each region, PM2.5 carbon emissions also could be included in the undesirable output. in terms of Other explanatory variables, please add related references to support variables selection.
5. In terms of results, why highlight the ESG measurement and evaluation, which is not consistent with the title. Moreover, the analysis is only superficial and does not reveal the internal mode of action.
6. No further discusion about your main implications for other studies. no deficiencies is noted for further reserch. In addition, there are too few references.
Comments on the Quality of English Language
Therefore, better implement the innovation-driven development strategy is of great significance for the stability and sustainable development of China's economy.
Author Response
Thanks for the comments. All the responses are shown in the next file.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for Authors This article addresses the Assessment and Influencing Factors of Regional Green Innovation Efficiency Based on the Lasso Method.
Comments #1: 1. Introduction
The introduction adequately addresses the definition of regional green innovation efficiency and its importance in China's future economic development, but does not make clear at the end the real purpose of identifying a set of variables. Having identified the dependent variable which is regional green innovation efficiency, what is the role to be played by the variables to be selected, explaining the main determinants of such efficiency? If so, this should be explained more clearly.
Comments #2: 2. Literature Review
Many references are made to the literature, but I believe that others are still missing that may be important and are related to the work done:
Super-SBM methodology:
Nan Zhao, Xiaojie Liu, Changfeng Pan, Chenyang Wang, (2021). The performance of green innovation: From an efficiency perspective, Socio-Economic Planning Sciences, Volume 78, 2021, 101062, ISSN 0038-0121, https://doi.org/10.1016/j.seps.2021.101062.
Spatial correlation and convergence in the efficiency of green innovation in China:
Peiyang Zhao, Zhiguo Lu, Jiali Kou, Jun Du (2023), Regional differences and convergence of green innovation efficiency in China, Journal of Environmental Management, Volume 325, Part A, 2023, 116618, ISSN 0301-4797, https://doi.org/10.1016/j.jenvman.2022.116618
In the following article there is an inverted U relationship between environmental regulation and green innovation efficiency:
Jingxiao Zhang, Le Kang, Hui Li, Pablo Ballesteros-Perez, Martin Skitmore, Jian Zuo, (2020), The impact of environmental regulations on urban Green innovation efficiency: The case of Xi'an, Sustainable Cities and Society, Volume 57, 2020, 102123, ISSN 2210-6707, https://doi.org/10.1016/j.scs.2020.102123.
Different techniques based on the estimator with penalty: ridge and lasso estimator.
Li, M.; Sun, H.; Agyeman, F.O.; Heydari, M.; Jameel, A.; Salah ud din Khan, H. Analysis of Potential Factors Influencing China’s Regional Sustainable Economic Growth. Appl. Sci. 2021, 11, 10832. https://doi.org/10.3390/app112210832
Comments # 3: 3. Research Methods
The authors should first clarify the purpose of the two methodologies applied, both the Super-SBM model and the Lasso econometric technique. They should contextualize it in the paper.
Comments #4: 3.1 Super-SBM Model Considering Undesirable Output
In the first case, the authors use the Super-SBM model, which contains undesirable output, to measure the efficiency of regional green innovation. That is, the authors should indicate that with this methodology the dependent variable of the model is defined from the coefficient ρ, because this is not said at any point in the entire paper.
The authors also do not refer to the software used to implement the Super-SBM model.
I consider this part to be very important because it is where the efficiency of regional green innovation is defined from the coefficient ρ.
To better clarify that starting from the decision making units (DMUs) that represent the regions, a set of indicators (input/output) are evaluated. Subsequently, indicate which indicators are part of the input and which are part of the output (desired and undesired).
In short, the authors should better clarify the purpose of this method, so that people unfamiliar with this methodology can understand it.
In addition, the CCR model and the BCC model are mentioned, but there is no mention of these techniques in the literature. It should be indicated what the methods are, whether they are similar techniques to SBM or how they differ.
Comments #5: 3.2. Lasso Regression Model
The authors do not clarify that the Lasso technique is precisely the one they are going to use for the selection of variables that will determine the efficiency of regional green innovation. This should be made clear in this section.
The authors do not indicate the software used for its application.
Comments #6: 4. Research Design
The authors should rename section 4 as follows:
4. Research Design
4.1. Data Source
4.2. Explained Variable: index selection
4.3. Core Explanatory Variable
4.4. Control Variables
The selection of the index (former section 4.2.1 and now new section 4.2) actually corresponds to the dependent variable, whose determination is carried out by applying the Super-SBM method, but at no time is reference made to section 3.1 relating it to equation (1). For example, the inputs x correspond to the indicators:...the desired output yg is formed by the following indicators...and the undesired output yb is defined by ... The abbreviation of the dependent variable (EFF?) should be indicated. In table 5 the variable EFF appears, is it the dependent variable? what does EFF mean?
In the old section 4.2.2 and now new section 4.3, ESG represents an explanatory variable for the efficiency of regional green innovation and not an algorithm. In line 271 instead of “and the algorithm is as follows:” it should read “and the variable is defined as follows”. In line 273 instead of “In this algorithm, ESGit represents” it should read “ESGit represents”.
Equation (7) has errors. It should be indicated mit in the equation (7) and line 275. Line 273: ni (not ni).
In the new section 4.4, what the authors call “other explanatory variables” are in fact “control variables” that are introduced in the econometric model to correct for possible estimation biases that could affect the coefficient of the main factor (ESG), due to the high covariances between the main factor and the rest of the explanatory variables. I recommend that the authors indicate this and cite econometric references on this problem of covariances and the need for control variables.
Comments # 7: 5. Empirical analysis
The authors should begin by analyzing the variable under study, which is the efficiency of regional green innovation, whose abbreviation is not clear in the text, and then the most important explanatory factor, which would be the ESG variable.
Comments # 8: 5.3. HT Unit Root Test
The authors speak of the HT unit root test applicable when dealing with a short panel. First, the panel is not short, because there are 11 cross-sectional units and 10 temporal units. They speak of a short panel when the number of cross-sectional units is much higher than the number of longitudinal units. On the other hand, they talk about the HT contrast, but do not include any reference to this test. They should include this reference, as well as the full name of the contrast they have actually applied, indicating the statistic and distribution of the statistic.
Comments # 9: 5.4. Lasso Regression
The authors comment that most of the variables in the model are affected by multicollinearity, according to Table 6. However, this is not true. The value VIF<10 or 1/VIF<0.1 (see Gujarati, D., Porter, D (2009): Basic Econometrics. Fifth (5th) Edition, McGraw-Hill). Actually, there are only two variables with severe multicollinearity: Urbanization Level and Economic Development Level. The authors should correct this comment in the text and indicate that the multicollinearity is not severe.
Comments # 10: 5.4. Lasso Regression
Table 7 does not explain the meaning of ID. I understand that it is the number of regressions performed, i.e. a total of 100. The authors should clarify this.
Comments # 11: 5.5. Least Squares and Stepwise Regression
The authors present in Table 9 the results of Least Square method, Stepwise regression and Lasso regression method. Another related technique is Ridge Regression. They should include the results of Ridge Regression in Table 9 for comparison.
Table 9 should indicate whether the coefficients are significant at 1%, 5% or 10%. That is, indicate the p-value with the different methods.
Comments # 12: 6. Conclusion and Suggestion
Authors should include in the conclusions the comment to Ridge Regression.
They should indicate that the methodology applied does not take into account the heterogeneity that is present in a panel of data and that can lead to significant estimation biases due to the fact that only Least Square Method corrected with penalties is applied.
Author Response
Thanks for the comments. All the responses are shown in the next file.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThe article presents a methodologically sound analysis of the efficiency of green innovation, contributing relevant information to the literature, specifically on regional development. It is adequately structured, but could benefit from a greater diversity of international references in the theoretical framework.
Some policy implications could be developed, as well as an in-depth and clear discussion of the limitations inherent to the specific regionality studied.
Comments on the Quality of English LanguageNA
Author Response
Thanks for the comments. All the responses are shown in the next file.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for Authorsplease add more references for this study
Author Response
Thanks for the pointing out. We all agree with these points. Therefore we have changed the references in this article. You can find it in the revised article as followed.
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors have taken the suggested comments into account and the manuscript has improved sufficiently to justify its publication in Sustainability.
Author Response
Thanks for the pointing out. We all agree with these points. Therefore we have changed them in the revised article. Please see the revised article.
