The Impact of Economic Growth, Industrial Transition, and Energy Intensity on Carbon Dioxide Emissions in China
Round 1
Reviewer 1 Report
The topic of the paper is very interesting. The problem of environmental protection and its determinants is very important these days (among others in the light of the sustainable development). The strength of this paper is the indication of possible application of the results by the rulers. The aim of the paper is clear and well exposed. I have few specific tips and comments to improve the paper:
- It is worth to point out possible other forms of the EKC (in the literature review). The quadratic form of EKC is not the only one using by researchers. Please explain, whether in the investigation process another forms of EKC were considered and why the quadratic form was chosen.
- Please explain whether the presence of the deterministic trend in the formation of processes were studied. It is hard to differ processes with the deterministic and stochastic trend. In this paper processes are have been recognized as difference-stationary. Please explain what are premises for this?
- Please add brackets for B-P-G test p-values in Table 6.
Please refer to the comments and make any necessary corrections.
Author Response
Thank you very much for your suggestions. We have further improved the article after more in-depth thinking. The attached document is our reply to your suggestions.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments to the author/s of the article “The impact of economic growth, urbanization and service industry development on carbon dioxide emissions in China” and the Editors
In the context of sustainability and green economy topics analyzing carbon dioxide emission are very important, but the usage of the Environmental Kuznets Curve model is not recent and, in my point of view, not relevant in our days. The Sustainability covers articles which are leading with the analysed topics and are using modern research methods. Such kind of article could be more relevent in the field of econometrics. The results of the research have revealed quite obvious ideas and economic logic. The analysis of three sectors are not novel. The proper hypothesis have not been raised and the analysed theory not lead to up-to-date questions.
Author Response
Thank you very much for sharing your opinion, which has inspired us to think more deeply. In view of your suggestions, we have also revised the article to a large extent. Attached is our response to your suggestion.
Author Response File: Author Response.pdf
Reviewer 3 Report
The article - The impact of economic growth, urbanization and service industry development on carbon dioxide emissions in China discusses and analyses the long- and short term relationship between economic growth, urbanization and service industry development and CO2 emissions in China. This study introduces Kuznets's theory with the ARDL model in the timespan from 1980 to 2019 to show environmental pressure from China's economy. The results could have implications for policymakers, economists and environmentalists. I recommend the manuscript for publication in Sustainability after some minor adjustments have been taken into account.
Introduction
Lines 35-38 „ Initiated the reform and opening-up in 1978, China's economy has 35 grown rapidly and became the world's second-largest economy in 2010 [8]. Currently, 36 China is also the largest developing country in the world. The rapid economic development is along with a large amount of CO2 emissions [9].”There is no data from your study that support this conclusion. Moreover, to build an interest to the international readers of the journal some comparative (worldwide) analysis should be added.
3.1. Data
No VIF information is found (multicollinearity among the independent variables), please clarify.
Conclusions and policy implications
Briefly describe the limitations of your study to show readers that you have considered your experiment’s weaknesses.
Please, separate the discussion from conclusions.
Author Response
Thank you very much for your recognition of our paper and your suggestions. Attached is our response to your suggestion.
Author Response File: Author Response.pdf
Reviewer 4 Report
The paper by Z. Yang et al. proposes a model of the Chinese CO2 emissions as a function of GDP per capita, urbanization level, proportion of the added value of the tertiary industry in the country’s economy, and household per capita energy consumption. The data used for the study cover the forty years between 1980 and 2019.
In my opinion, the paper has relevant flaws and thus should not be accepted for publication.
First, the paper tries to explain CO2 emission as if it is a measured, independent variable. CO2 emission, on the contrary, is simply computed on the basis of fuel used, and thus a more useful analysis would be the relation between a fuel (energy) used and urbanization and/or tertiary industry or between total energy use and the production from renewable sources.
Second, the model proposed is stationary, i.e., it assumes that the relation between, say, CO2 emission and urbanization level can be explained by the same values of the coefficients for all the period 1980 to 2019. To my knowledge, the situation of the Chinese economy and society, in general, has been changing a lot since the ‘80s and thus the assumption of an invariant model seems difficult to accept.
Third, the proposed model is unclear. In (1) it is just expressed as a generic function. In (2), it is converted into logarithms at time t, with an unusual square of the GDP term, which derivation is unclear. Then, in (3), it becomes an improper dynamic system. It computes emission at time t with some autoregressive terms (namely, emissions at preceding time steps) and other independent variables again at time t. Altogether, despite the number m of lags terms in (3) is not specified (but it seems to be m=2 in table 4 without any justification), this model appears to have a high number of coefficients (at least 15 in table 4), quite difficult to estimate on a series of 40 values. The authors then propose a number of statistical tests on the coefficients but do not present the classical performance indexes used to test the accuracy of any model.
The paper requires an accurate check of the English form since the wording is often incorrect (e.g., “the further research is desiderate to carry out. Therefore, this paper expects to fill in the gaps in this research.” Line 178; “It should be note” Line 212; “urbanization and service industry development were deliberate considered” Line 357).
More importantly, the paper is written as if addressed to econometric experts, not the readers of Sustainability. It quotes a number of econometric methods and tests without any explanation of their purpose and meaning. Even the main assumption of the EKC theory is never spelled out explicitly. Additionally, the related key result in fig. 4 is not explained nor justified.
The conclusions are all based on the results of the assumed model without any critical discussion of the model itself, as it should be done in any scientific study. As it is now, the paper seems more a student’s exercise of calibrating and analyzing a fixed model than a research work aiming at finding the most appropriate model to interpret the problem at hand.
Author Response
Thank you very much for your careful reading and your questions and suggestions. We will reply in the attached document.
Author Response File: Author Response.pdf
Reviewer 5 Report
Dear authors,
Research about which economic activities are the most polluting is the first step forward to deal with climatic change challenges nowadays. Your paper has fully met my expectations. It evaluates the impact of economic growth on carbon dioxide emissions in China by employing a proper methodology. The only issue I have with this paper is linked to the limitations of the study. These should be included in the conclusion section. For example, how sure can authors be that data about carbon emissions are accurate enough? Does the risk exist that these data can be lower than expected?
Author Response
Thank you very much for your recognition of our work. Your suggestions make us think more deeply and we have organized the reply into the attached document.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Dear authors, thank you for your improvements.
Author Response
Dear Reviewer:
Thank you very much for your recognition of our work. Your feedback has inspired us deeply, which is of great help to the improvement of our manuscript.
Reviewer 4 Report
The authors have extensively revised the paper, which is now much more readable.
However, they have not solved some questions that significantly undermine the value of the paper itself.
First, why use three independent variables to model CO2 emissions when, as already said, they are computed by multiplying the fuel used times an emission factor? In the new version, the authors have changed one variable with the energy intensity, but the product energy intensity*GDP obviously means total energy used. Now, the emission changes only if the fuel changes because of the changes in the emission factor. Is the new INDUSTRY variable a valid substitute for this change? Why not use the amount of fuel (coal, gas, oil) directly?
Second, what is the model proposed in the paper? Formula (1) is a generic function, and the assumption that it represents a nonlinear relation is indeed just an assumption. It cannot be deduced by simply writing f(something). The reference [47] quoted is a classical production function y=ax1^bx2^cx3^d that can be linearized by applying the logarithmic transformation. But what model can originate a log^2? This is not explained. Additionally, inserting such a term forces, in a way, the existence of the EKC effect. Would a model without this square term work equally well? Providing some statistical tests of the model coefficients is usually not enough to judge the model's validity. It must be tested on its ability to represent new values not used in the model calibration accurately.
My impression is that the paper provides just statistical tests on the logarithms because a correspondent model with the nonlogarithmic values does not exist. And I bet no decision-maker will ever accept a model that can be expressed only with logarithms of the variables.
I think the paper cannot be considered a valid contribution for publication if these points are not clarified.
Author Response
Dear Reviewer:
Thank you very much for your questions and suggestions, which have promoted us to think more deeply. In this regard, we have revised the manuscript and replied to your feedback in the attachment. Thank you again for your feedback on our manuscript.
Author Response File: Author Response.pdf
Round 3
Reviewer 4 Report
Dear authors, I understand your points but I still think that a model that exists only in logarithmic form (model (3) in the new manuscript has no equivalent in the original variables) is unusable and that any model must be checked against independent data. Evaluating its performance on data used for calibration does not mean much.
