Chinese CO2 emission flows have reversed since the global financial crisis

This study seeks to estimate the carbon implications of recent changes in China’s economic development patterns and role in global trade in the post-financial-crisis era. We utilised the latest socioeconomic datasets to compile China’s 2012 multiregional input-output (MRIO) table. Environmentally extended input-output analysis and structural decomposition analysis (SDA) were applied to investigate the driving forces behind changes in CO2 emissions embodied in China’s domestic and foreign trade from 2007 to 2012. Here we show that emission flow patterns have changed greatly in both domestic and foreign trade since the financial crisis. Some economically less developed regions, such as Southwest China, have shifted from being a net emission exporter to being a net emission importer. In terms of foreign trade, emissions embodied in China’s exports declined from 2007 to 2012 mainly due to changes in production structure and efficiency gains, while developing countries became the major destination of China’s export emissions.

applying the structural decomposition analysis (SDA). See the review of SDA studies on energy and emissions in Su and Ang (2012;Energy Economics 34, pp.177-188; for studies published before 2010) and Wang et al. (2017;Energy Policy 107, 585-599; for studies published in [2010][2011][2012][2013][2014][2015]. The authors are suggested to conduct the SDA to provide the insights into the embodied emission changes in 2007-2010 and 2010-2012. Some specific comments: Lines 30-31, "shift from an investment-driven economy to a consumption-driven economy": Recent work by Su and Ang (2017;Energy Economics 65, 137-147)'s Table 2 shows that the investment in China accounts for 43.5% (=2664.0/6080.7) of its total emissions in 2007 and 52.7% (=4312/8185.4) of its total emissions in 2012. I don't think this statement is correct.
Lines 59-60: The authors should read the work by Su and Thomson (2016;Energy Economics 59, 414-422) Table 5&6. It cannot conclude that the embodied have peaked. Previous studies have mapped and quantified the material flows and GHGE generation consumption and production relationship between China and the rest of the world. This study goes further: it untangles the differences in production and consumption based accounting methods in the regions of China over a time series (2007)(2008)(2009)(2010)(2011)(2012), during the GFC), providing valuable insights. Specifically on the domestic and international outsourcing of emissions. It shows that the outsourcing of emissions is not just a Developed->Developing world problem, but also a problem between sub regions of developing countries, and also between developing countries. This work is rigorous, original and very publishable. It should attract scholarly (and media attention), it will influence thinking in the field (and hopefully broader policy debate around subregional development with regards to carbon intensive industries). Given the methods and appendices provided, this study findings should be reproducible. Some brief comments: Line 124-151 The functional unit of choice for this paper is g of c02e per yuan, per year or per capita. However these expressions of carbon relationship may be masking material and physical flows, as there is little discussion of the types of goods exported (east vs west), ie the choice of functional units may hide a decoupling the relationship of high cost goods vs high carbon intensive goods. With a 30 sector IOT (with uncertain data) there is little chance for further exploration of the exact goods traded but this would be an interesting further investigation. (although I also understand there is little room to go deeply into this relationship in a nature comms article) 437-451 impact exponents and inter regional cooperation. I would like to see a worked through example of this method and results as a separate article using the data set in this paper, I understand the space limitations of a NCC article and feel this additional article could be sent to ESR or similar. 478-492 Could the authors provide further explanation as to why they chose to aggregate 30 by 30 rather than dis-aggregating to 57 by 57? As this would have provided more detail of some of the trade linkages.
This manuscript adopted a multi-regional input-output model to analyze the interregional carbon emission flows in China after the financial crisis (the 'new normal' era). The results are clearly presented and discussed. The main message that emission flow patterns have changed greatly as the (interregional and international) trade patterns changed is straightforward but solid. Some prospective discussion was also provided based on these results.

Response:
Many thanks for your support and valuable comments. We have carefully revised the manuscript following your comments.
My main concern on its suitability in Nature Communications is the novelty and significance of this manuscript. Method wise, the past decade has seen a great amount of work on the development and use of multi-regional input-output analysis for consumption-based emission accounting (including the authors' own work). A potential innovation in this manuscript is its consideration of interregional trade within China; however, previous papers have included this development and its application in understanding China's inter-provincial carbon leakage, for example, Feng, K., Davis, S.J., Sun, L., Li, X., Guan, D., Liu, W., Liu, Z., Hubacek, K., 2013. Outsourcing CO2 within China. Proc. Natl. Acad. Sci. 100, 11654-11659 (By the way, this paper of the authors' own was NOT cited in the manuscript). Of course, the manuscript has used data of

Response:
Many thanks for your comments. During the first round submission, we made clear to the editorial team that our study is based on a widely used approach that provides robust results and that the contributions are the new dataset that was developed and the relevant policy analysis and new insights that can be achieved with the new dataset. We also made this point clearer in revised manuscript.
We utilized the latest socioeconomic datasets to compile 2012 China's multiregional input-output (MRIO) table. This is a new data contribution to the academic field. All MRIO tables developed and used in this study are provided as Supplementary Data for this submission, which will be made publicly available via data repository after the publishing process.
We estimated the carbon implications of recent changes in China's economic development patterns and role in global trade in the post-financial-crisis era (i.e. 2007-2012). Compared with previous analysis, we discovered an interesting reversal of emission flows between Chinese regions and some important changes globally. In addition, as suggested by Reviewer 2, we added a structural decomposition analysis (SDA), based on the latest available data, to analyze driving forces behind these emission changes. Seven factors were included in the analysis: emission coefficient, energy mix, energy efficiency, production structure, consumption structure, consumption per capita, and population. The results show that the shift of Southwest China from being a net emission exporter to being a net emission importer was mainly due to the rapid growth in consumption in these regions. The decline of China's export emissions from 2007 to 2012 was mainly due to production structure changes and efficiency gains. The SDA results on emissions embodied in China's domestic and foreign trade are shown in Figure 2 and Figure 4c, respectively.
The two papers you mentioned have now been added in the revised manuscript (reference 6 and 17).
Some other minor comments: -I think territory-based emission is more suitable than production-based emission. Response: Many thanks. All "production-based emission" in the manuscript has been replaced by "territory-based emission".
-Line 343-345: Do you think the increase trend from 2010 to 2012 will continue and thus bring the emissions back to the pre-crisis levels in the future?

Response:
We think that China's export emissions will not return to the 2007 level for two arguments. First, China's export volume has been decreasing since 2012. Second, China's production structure is becoming less carbon emission intensive.
-Ending paragraph: While I agree with the direction that carbon leakage in south-south trade deserves more attention, I personally find these sentences as concluding remarks a bit too strong (the majority of carbon leakage would still occur between developed and developing countries in the near future).

Response:
We followed your suggestions and deleted related statements in the ending paragraph. We agree with you that the majority of carbon leakage would still occur between developed and developing countries in the near future.

Response to Reviewer #2:
The authors constructed the multi-country/multi-region I-O tables by linking Chinese multi-region I-O tables and global multi-country I-O tables (GTAP dataset) in 2012 to study the embodied emission flows in China. They provide the latest findings regarding to the embodiment in interregional and international trade of 30 regions in China after global financial crisis. The authors further compared the 2012 results with results for year 2007 and 2010. This is the positive contribution of this paper. There are also some parts need further improvements, including missing important references in the literature, some unclear parts in the data treatment, and results comparisons. The major and specific comments are given below:

Response:
Many thanks for your support and valuable comments. We have carefully revised the manuscript following your comments.
The unique contribution of this paper comes from the linking the Chinese regional I

Response:
We have followed your suggestion and applied SDA to estimate the driving forces of embodied emission changes in 2007-2010 and 2010-2012. Seven factors were considered, including emission coefficient, energy mix, energy efficiency, production structure, consumption structure, consumption per capita, and population. The SDA results are shown in Figure 2 and Figure 4c. We added an introduction on SDA approach in the methods section. The two review papers were added in the revised manuscript (line 547-549). Some specific comments: Lines 30-31, "shift from an investment-driven economy to a consumption-driven economy": Recent work by Su and Ang (2017; Energy Economics 65, 137-147)'s Table 2 shows that the investment in China accounts for 43.5% (=2664.0/6080.7) of its total emissions in 2007 and 52.7% (=4312/8185.4) of its total emissions in 2012. I don't think this statement is correct.

Response:
We have deleted the statement related to the "shift from an investment-driven economy to a consumption-driven economy". The useful references you provided have been added (reference 4 ).

Response:
The definitions of share of value added embodied in exports are different. Su and Ang (2017) discussed the value added embodied in different final demand categories, including rural consumption, urban consumption, government consumption, gross fixed capital formation, inventory change, and exports. In our study, we discussed the domestic and foreign value added in exports. The definitions in Su and Ang (2017) and our study are shown in the following equations.  Table 5&6. It cannot conclude that the embodied have peaked.

Response:
We have deleted the statement on the peak of emissions embodied in exports. The change trends on emissions embodied in China's exports in our study are consistent with those in Su and Thomson (2016). We compared our results with those of Su and Thomson (2016) (line 250-252).

Response:
We have made emission data consistent between Table S1 and excel files. In our original submission, the data in excel files included emissions from economic sectors and residential energy consumption, while Table S1 only included emissions from economic sectors. That's why the overall value was different. In the new submission, we deleted the emissions from residential energy consumption in the excel files. Lines 480-481: Please provide the matching table between the 57-sector classification in GTAP and 30-sector classification in Chinese multi-region IO tables.

Response:
We added Table S5 to show the concordance of sectors for GTAP and China MRIO.  Table S2 for the detailed definition of world regions).
The assumptions and data for the Rest of the World (ROW) are obtained from the GTAP database. Introduction of the regional disaggregation can be found on the GTAP website (https://www.gtap.agecon.purdue.edu/databases/regions.asp?Version=9.211).
The ROW includes: Antarctica, Bouvet Island, British Indian Ocean Territory, and French Southern Territories.

Response to Reviewer #3:
Previous studies have mapped and quantified the material flows and GHGE generation consumption and production relationship between China and the rest of the world. This study goes further: it untangles the differences in production and consumption based accounting methods in the regions of China over a time series (2007)(2008)(2009)(2010)(2011)(2012), during the GFC), providing valuable insights. Specifically on the domestic and international outsourcing of emissions. It shows that the outsourcing of emissions is not just a Developed->Developing world problem, but also a problem between sub regions of developing countries, and also between developing countries. This work is rigorous, original and very publishable. It should attract scholarly (and media attention), it will influence thinking in the field (and hopefully broader policy debate around sub-regional development with regards to carbon intensive industries). Given the methods and appendices provided, this study findings should be reproducible.

Response:
Many thanks for your support and valuable comments. We have carefully revised the manuscript following your comments.

Some brief comments:
Line 124-151 The functional unit of choice for this paper is g of c02e per yuan, per year or per capita. However these expressions of carbon relationship may be masking material and physical flows, as there is little discussion of the types of goods exported (east vs west), ie the choice of functional units may hide a decoupling the relationship of high cost goods vs high carbon intensive goods. With a 30 sector IOT (with uncertain data) there is little chance for further exploration of the exact goods traded but this would be an interesting further investigation. (although I also understand there is little room to go deeply into this relationship in a nature comms article) Response: Many thanks for your comments. We agree with you and are aware that the exact goods traded among Chinese regions (east vs west) are an interesting topic. For this research, we could guarantee the method and data robustness of a 30x30 sector MRIO. This is also comparable with earlier MRIO format and compilation processes. We will consider further investigation on different types of goods traded among Chinese regions in future research, but currently this is beyond the scope of the presented study. Most trade studies based on MRIO share this limitation due to available MRIOs and data but at the same have the advantage of comparability of results. 437-451 impact exponents and inter regional cooperation. I would like to see a worked through example of this method and results as a separate article using the data set in this paper, I understand the space limitations of a NCC article and feel this additional article could be sent to ESR or similar.

Response:
478-492 Could the authors provide further explanation as to why they chose to aggregate 30 by 30 rather than dis-aggregating to 57 by 57? As this would have provided more detail of some of the trade linkages.

Response:
We are aware that 57-sector MRIO would provide more detail of trade linkages. For this research, we could guarantee the method and data robustness of a 30x30 sector MRIO. This is also comparable with earlier MRIO format and compilation processes. The main purpose of this paper is to compare China's emission flows from 2007 to 2012. So we aggregate 57 sectors into 30 sectors to make the 2012 MRIO model comparable with those for 2010 and 2007. We will consider a further disaggregation for IO sectors, but currently this is beyond the scope of the presented study.