The Contribution of China ’ s Outward Foreign Direct Investment ( OFDI ) to the Reduction of Global CO 2 Emissions

Under economic globalization, with the intensification of China’s reform and opening up, China’s outward foreign direct investment (OFDI) has continuously gained momentum, but CO2 emissions caused by the OFDI have not been given due attention. As one China is of the world’s leading CO2 emitters, it is necessary to conduct thorough research into the CO2 emission problem caused by China’s OFDI. Thirty-four host countries were selected as the objects of this study, including some European countries, Australia, India, Indonesia, Brazil, Canada, Japan, Korea, Mexico, Russia, and the USA. Their CO2 emissions as caused by China’s OFDI were calculated using the input-output model with non-competitive imports, the data of China’s OFDI flows, and their own energy consumption and CO2 emissions from 2000 to 2011. Then a comparative analysis was performed taking China as the comparative object. CO2 emission transfer of China’s OFDI was studied quantitatively. Finally, CO2 emissions from China’s OFDI were discussed from the perspective of industry selection and location selection. The results showed that China’s OFDI could achieve the aim of reducing global carbon emissions with reasonable industry and location selection.


Introduction
Foreign direct investment (FDI) outflow from developing countries has increased dramatically in recent years.It accounted for 33.8% of global outward FDI (OFDI) in 2014, up 7.6% from 2000 [1].China, as the world's second largest economy, has attracted much attention.Though China has been known as a destination of global investment, OFDI from China surged to USD 123.1 billion in 2014 [1], approximately at the same level of inward FDI to China.Three primary motivations behind international investments of firms, market seeking, efficiency seeking, and resource seeking, are suggested in traditional theories [2,3].In the case of emerging economies such as China, more specialized applications of the theory are needed.This has generated considerable interest in the motivations and drivers of Chinese international investment [4][5][6][7][8][9].The institutional environment and policies of the Chinese government are likely to have far-reaching and profound effects on the investment behavior of Chinese multinational corporations (MNCs) [7].Policy liberalization has a positive influence in stimulating China's OFDI [4].The "open door" policy of the late 1970s resulted in the emergence of China's OFDI from 1980s.In early 21st century, China's entry into the WTO and the "go global" strategy announced in China's long-term, innovation-oriented development plan encouraged Chinese firms to invest abroad.Promoted by the proposals of the Silk Road Economic Belt and 21st Century Maritime Silk Road (Belt & Road) in early 2010s, the countries covered by the Belt & Road initiative have the potential to create new markets for China's OFDI.More specifically, since the 1990s there have been some dramatic changes in the geographical distribution and industrial distribution of China's OFDI.In 2014, Asia received USD 85.0 billion, accounting for 69.0% of the total China's OFDI.The four major industries accounting for 71.1% were mining, wholesale and retail trade, finance, and renting and business services.Under such circumstances of policy liberalization and government encouragement, especially when Chinese OFDI is attracted to large markets [9], what are the impacts of the huge amount of China's OFDI on the host countries' environment, as well as the global environment?
The growing literature on foreign trade and the environment suggests a potentially high level of interaction between trade liberalization and carbon emissions.Currently, there are two main aspects for this study.One focuses on the carbon embodied in import and export trade, while the other is the empirical study of the impact of FDI on carbon emissions.Numerous studies concerned about the relationship between FDI and pollution in developing countries are related to the pollution haven hypothesis [10][11][12][13][14][15].As one of the largest host countries receiving inward FDI flows, more attention had been placed on China's rapidly increasing inward FDI and growing environmental pollution [15][16][17][18][19].However, there are still debates about the pollution haven hypothesis [20], and studies have been inconclusive regarding to the relationship between FDI and pollution [21].Some studies show that FDI inflows lead to an increase in carbon emissions [12,17,18,22].However, others show that FDI inflows are a contributory factor to the reduction of carbon emissions [16,[23][24][25].To resolve the theoretical ambiguity, this study turned to an empirical analysis of the pattern of foreign direct investment.Inevitably, the industrial development promoted by the inward FDI will require energy consumption and thus result in carbon emissions.The relationship between FDI and carbon emissions aside, this study's primary work is to calculate carbon emissions caused by FDI quantitatively.
The input-output analysis has been extensively used to calculate the carbon emissions embodied in the trade for many countries.The input-output analysis was theorized and developed by Leontief [26,27], and its theoretical framework and extended application were systematically detailed by Miller and Blair [28].Sectors in an economy are connected to each other by the supply-demand relationship.Given the Leontief inverse matrix, the input-output analysis can be easily utilized to calculate the total (direct and indirect) energy inputs and the associated carbon emissions of a sector, regardless of the length and complexity of the production process.Using the input-output analysis, CO 2 emissions embodied in international trade of countries have been studied, such as Austria [29], Brazil [30], China [31,32], Denmark [33], Italy [34], Japan [35], Norway [36], Spain [37], and Sweden [38].Most of the studies were focused on the estimation of CO 2 emissions embodied in international trade, and few analysts took notice of CO 2 emissions caused by capital transfer during the investment.Fu [39] identified quantitatively the volume of energy consumption and carbon emissions driven by domestic investment in China in 2007, assuming that domestic investment and foreign investment are substitutable, meaning FDI acts the same way as the domestic investment in the host countries.When the host countries attract China's OFDI to expand production and meet domestic needs, CO 2 is released in their own yards.The IPCC designed greenhouse gas inventories containing emissions within national territory from the perspective of producer.Based on this hypothesis and principle, CO 2 emissions in the host countries caused by China's OFDI can be calculated with the input-output analysis and their related data.
The impact on CO 2 emissions from host countries caused by the dramatically increasing China's OFDI has not been well addressed.In addition, a distinct and well established problem with international trade in the Kyoto Protocol is the possibility for "carbon leakage" [40].The IPCC defines carbon leakage as "the increase in emissions in Non-Annex B countries resulting from implementation of reduction in Annex B (which includes most developed countries) emissions" [41].Carbon constraint nations are likely to import from nations with lower environmental standards and as a result end up uncompetitive in an industry with pollution-concentrated products.Thus, non-carbon constrained countries gain the upper hand in pollution intensive industries relative to carbon constrained countries [42].Simultaneously, carbon leakage through international trade might not reduce global emissions as much as expected and could even raise them [43].In the process of international capital flows, the location and industry selection of foreign direct investment have the same problem due to the different environmental regulation.China has a critical role in global emissions mitigation in the post-Kyoto period as the largest emitter of CO 2 .Even as a Non-Annex B country, China has taken actions to achieve a peak of CO 2 emissions around 2030 and to lower CO 2 per unit of GDP [44].Considering that China partly transferred its redundant production capacity to other countries due to the comparative advantage of its MNCs, China's OFDI would probably bring about an issue similar to carbon leakage and increase global CO 2 emissions.Although previous empirical studies have already covered CO 2 emissions embodied in import and export trade of China, few analysts have focused on CO 2 emissions caused by China's international capital flow.Particularly, how much CO 2 emissions are caused by China's OFDI, does China's OFDI result in carbon leakage, and what contribution has China made to the global CO 2 emissions as a home country of FDI?All of these fundamental questions are quite notable.In this paper, 34 host countries were selected, and their CO 2 emissions caused by China's OFDI were calculated on the basis of the input-output model using non-competitive imports assumption from 2000 to 2011.Formulas were then built to analyze the CO 2 emission transfer caused by China's OFDI.The differences between host and home CO 2 emissions were discussed against China for the first time.Finally, based on the results of 34 host countries, the contribution of China's OFDI to global CO 2 emission reduction was evaluated.

Methodology
The foundation of input-output analysis involves input-output tables, which represents monetary transactions between supply chains in mathematical form.According to the treatment of imports, the input-output model can incorporate either the competitive or the non-competitive imports assumption [45].The competitive imports assumption treats imported products as the same as those produced domestically, while the non-competitive assumption removes imports from intermediate and final use.The standard input-output model using competitive imports assumption can be formulated as, where x is the column vector of total output, I is the unit matrix, A is the n × n matrix of direct requirement coefficient, n represents industrial sector, and y is the column vector of final demand.(I − A) −1 represents the Leontief inverse matrix.Its element α ij represents the amount of output of the industry i required directly and indirectly to produce one unit of final demand from industry j. Su and Ang [46] found that the transitions of emissions embodied in imports to those in the exports accounted for a considerable percent of total emissions if the competitive imports assumption is used.To avoid overestimating, we adopt the input-output model with non-competitive imports to calculate CO 2 emissions caused by FDI.The standard input-output model using non-competitive imports assumption can be formulated as, where A d is the matrix of domestic direct requirement coefficient, y d is the vector of domestic final demand.
CO 2 emissions from domestic final demand can be formulated from Equation (2) as, where f is the 1 × m row vector of emission factor representing CO 2 emissions per unit of energy consumption, m represents energy type, E is the m × n energy intensity matrix representing the energy consumption per unit of value of industry output, n represents industrial sector in the input-output table.
The final demand includes final consumption, gross capital formation and exports.Therefore, CO 2 emissions from domestic investment of the host country (which receives China's OFDI) also can be calculated by Equation (3).As mentioned above, it is assumed that China's OFDI acts in the same way as the domestic investment of host country.CO 2 emissions from domestic consumption of the host country due to receiving China's OFDI can be calculated by the following equation: where v is the column vector of China's OFDI flows to the host country.When the vector v is diagonalized to the matrix v, CO 2 emissions from each sector of the host country due to receiving China's OFDI can be calculated individually.
To study carbon transfer caused by China's OFDI, another assumption that the amount and distribution of OFDI flows are unaffected by related factors of the host country and home country was established.When China's OFDI is assumed to be invested in sectors of China, it has the same properties as that received by the host country.Based on this assumption, home (China invests as a home country) CO 2 emissions were calculated using input-output tables with non-competitive imports, energy consumption and emission factors replaced by corresponding data of China.Home CO 2 emissions can be formulated as, where f is the row vector of China's emission factor, E is the matrix of China's energy intensity, (I − A d ) −1 represents China's Leontief inverse matrix.
The impact of China's OFDI on global CO 2 emission reduction can be quantified by the difference between host and home CO 2 emissions with Formula (6).The similar calculation principle was used to quantify the impact of international trade on national and global CO 2 emissions [42,47].
When ∆c v > 0, the contribution of China's OFDI to global CO 2 emission reduction is positive.This indicates that China's OFDI reduces global CO 2 emissions.When ∆c v < 0, the impact of China's OFDI to global CO 2 emission reduction is negative with increasing global CO 2 emissions.When ∆c v = 0, China's OFDI makes no contribution to global CO 2 emission reduction.

Data Sources and Processing
The input-output tables with non-competitive imports of China and host countries from 2000 to 2011 are derived from World Input-Output Database (WIOD) [48].The data of CO 2 emissions and relevant energy consumption with the same industrial structure as input-output tables are provided in WIOD from 2000 to 2009.CO 2 emission factor can be calculated by dividing CO 2 emissions by relevant energy consumption.The energy consumption of China and host countries in 2010 and 2011 are derived from Energy Statistics Database (UNSD) [49].The corresponding CO 2 emissions are obtained by applying CO 2 emission factors adopted from IPCC (2006) [50] to the energy consumption.
According to WIOD data coverage, there are 35 industrial sectors and 26 energy types.The industrial sectors are combined to 12 categories, while taking account of the industrial sectors of China's OFDI flows.There are 14 types of energy that generate CO 2 emissions.Detail industrial sectors and energy types are given in Tables 1 and 2. In the case of OFDI data, the form of flow has no lag compared to stock, and therefore the OFDI flows can reflect the development and change of the current economic situation more effectively.It is more reasonable and accurate to calculate host CO 2 emissions with China's OFDI flow data.China's OFDI flows are derived from World Investment Report (UNCTAD) [1] and Statistical Bulletin of China's Outward Foreign Direct Investment (MOFCOM) [51].On the base of countries in the coverage of WIOD's data, 34 countries are chosen as host countries receiving China's OFDI.Because of the lack of data, Estonia, Lithuania, Slovenia, Portugal, and Taiwan are not included.China's OFDI flows to the 34 host countries from 2000 to 2011 are presented in Table 3.China's OFDI flows to the 34 host countries accounted for 20.4% of the total in 2011, and they consist of both developed and developing countries.It is representative to study CO 2 emissions from China's OFDI with these 34 host countries.

Host CO 2 Emissions
CO 2 emissions of these 34 host countries can be calculated by Equation ( 4), and the results are shown in Table 4. CO 2 emissions were comparatively low in most of host countries before 2005, mainly because the scale of China's OFDI was very small in those years.Host CO 2 emissions have had continuously gained momentum with the growth of China's OFDI since 2006.CO 2 emissions in some countries fluctuated from year to year due to the large difference between annual FDI flows.China's OFDI flows were short of continuation and stability.CO 2 emissions of few host countries in certain years were negative value because of the negative OFDI flow.The OFDI flow is obtained by subtracting contrary investment from FDI enterprises to domestic investors from the total foreign direct investment in the current period.FDI enterprises refer to foreign enterprises that directly owned or have 10% voting rights or equivalents controlled by domestic investors.When the contrary investment in the current period is larger than the total foreign direct investment, the OFDI flow will be negative.The top four countries in terms of host CO 2 emissions were USA, Indonesia, Australia, and Russia in 2011, emitting 1135.0,937.2, 729.4,and 601.7 kt CO 2 respectively.Moreover, India, France, Germany, and Korea also emitted over 200 kt CO 2 .During 2000 to 2011, Indonesia had the largest cumulative host CO 2 emissions with 3872.7 kt, followed by Russia emitting 3268.6 kt.The cumulative CO 2 emissions of Australia and USA also reached over 2000 kt during the period.The difference between host CO 2 emissions is significant, but it is not associated only with the amount of China's FDI flows.CO 2 emissions per unit foreign direct investment (CEPI) is calculated to further analyze the difference between host CO 2 emissions.CEPIs of the leading 13 host countries from 2000 to 2011 are shown in Figure 1.The CEPI of Indonesia stayed at a high level above the other countries during most of the period.It peaked the highest value of 10.5 kg/USD in 2001, and then dropped to 1.6 kg/USD in 2011.The CEPI of India reduced from 2.5 kg/USD in 2000 to the lowest value of 0.9 kg/USD in 2006, and then increased to 2.6 kg/USD higher than other countries in 2011.The CEPI of Russia reached a relatively high level, declining from 3.8 kg/USD to 0.8 kg/USD during the period.The cumulative OFDI flows to Indonesia, India, and Russia ranked 10th, 16th, and 6th respectively among the 34 host countries.However, their cumulative host CO 2 emissions ranked the first, 6th and second places due to their high CEPIs.CO 2 emissions per unit foreign direct investment were not only affected by energy structure, production process and technology level of host countries, but also restricted by industry selection of China's OFDI.In terms of the industrial distribution of China's OFDI in Indonesia, the FDI flows were allocated in the industry of electricity, gas and water supply.In Russia, agriculture, mining and manufacturing received most of China's OFDI.Higher energy consumption and higher emissions of primary industries led to higher CEPIs.Because the proportion of China's OFDI received by mining and manufacturing in Korea increased after 2005, its CEPI correspondingly rose in a small amplitude.The reason that why CEPIs of UK were higher than 1.0 kg/USD in previous years was also associated to the high proportion of China's OFDI received by energy industries.Consequently, the industry selection of China's OFDI has influence on host CO 2 emissions.The higher proportion of OFDI to resource intensive and energy intensive industries will certainly result in larger host CO 2 emissions.

Home CO2 Emissions
The results of home CO2 emissions calculated by Equation.( 5) are shown in Table 5.The issue of global CO2 emission reduction caused by China's OFDI can be further investigated by the comparison between home and host CO2 emissions.Most of host CO2 emissions were lower than home CO2 emissions during most of the years (if the emission is negative, it needs to be measured in absolute value).Except Bulgaria, Greece, India, Malta and Russia, cumulative host CO2 emissions in all of other countries were lower than their cumulative home CO2 emissions.The comparison results indicate that China's OFDI contributed to global CO2 emission reduction in the process of carbon transfer.Based on the definition and hypothesis of home CO2 emissions, the reason for lower host CO2 emissions in most countries than their home CO2 emissions is greatly different s in energy structure and economic technical interrelation between these host countries and China.For example, Apart from the five host countries mentioned above, mining in Australia topped other industries and received the majority of China's OFDI, and manufacturing also received a considerable portion of China's OFDI in other host countries.Nevertheless, CEPIs of the other eight host countries were lower than 1.0 kg/USD, e.g., CEPIs of France and Japan mostly maintained in the range of 0.1~0.2kg/USD.Lower CEPIs were not only related to industry selection of China's OFDI, but mainly due to the higher level of technology and management, more equitable economic technical interrelation among industries in these host countries.Therefore, the emission conduction effect of location selection during the process of China's outward foreign direct investment is not inconsiderable.

Home CO 2 Emissions
The results of home CO 2 emissions calculated by Equation ( 5) are shown in Table 5.The issue of global CO 2 emission reduction caused by China's OFDI can be further investigated by the comparison between home and host CO 2 emissions.Most of host CO 2 emissions were lower than home CO 2 emissions during most of the years (if the emission is negative, it needs to be measured in absolute value).Except Bulgaria, Greece, India, Malta and Russia, cumulative host CO 2 emissions in all of other countries were lower than their cumulative home CO 2 emissions.The comparison results indicate that China's OFDI contributed to global CO 2 emission reduction in the process of carbon transfer.Based on the definition and hypothesis of home CO 2 emissions, the reason for lower host CO 2 emissions in most countries than their home CO 2 emissions is greatly different s in energy structure and economic technical interrelation between these host countries and China.For example, though China's OFDI flows to Australia were concentrated in mining accounting for above 60%, Australia's host CO 2 emissions were just about 20% of its home CO 2 emissions.In Australia's mining industry, natural gas accounted for about 60% of its energy consumption, and the rest were mainly oil fuels.While in China's mining industry, coal accounted for above 60%.Similarly, in the energy structure of Indonesia's electricity, gas, and water supply industries, anthracite accounted for about 30% and natural gas accounted for 40%.The energy structure of the electricity industry was relatively cleaner compared to 95% of coal in China's electricity industry, and therefore the cumulative host CO 2 emissions of Indonesia were notably lower than its cumulative home CO 2 emissions.In most of the host countries, because production technology is superior, energy use is more efficient, the energy structure is dominated by fuels with lower emission (e.g., oil and natural gas), and CO 2 emission reduction target constraint is stricter.Therefore, the host CO 2 emissions from China's OFDI of most host countries were considerably lower than their home CO 2 emissions.
China's OFDI flows to Russia were allocated in agriculture, mining, and manufacturing.Russia's energy structure was reasonably cleaner than China, in which natural gas accounted for more than 60% and solid fuels like coal and coke accounted for less than 20%.However, except for 2010 and 2011, host CO 2 emissions of Russia were higher than home CO 2 emissions.The energy efficiency in Russia was so low that energy intensities of the three main industries receiving China's OFDI were more than twice those of China.The energy intensities of other industries in Russia were likewise generally higher than China.While Russia received China's OFDI, the same amount of FDI would burn more fuels through economic technical interrelation of industries, and the cleaner energy structure could not offset the higher CO 2 emission caused by the lower energy efficiency.

CO 2 emissions Transfer
Outward foreign direct investment brings about industry transfer, capital transfer, and CO 2 emissions transfer.According to the research framework of this paper, CO 2 emissions transfer caused by China's OFDI has been quantitatively analyzed by calculating host CO 2 emissions, and the contribution of China's OFDI to global CO 2 emission reduction can be evaluated by the difference between host and home CO 2 emissions using Equation (6).The cumulative host and home CO 2 emissions of 34 countries from China's OFDI are shown in Figure 2. The cumulative home CO 2 emissions were larger than host CO 2 emissions in most countries.This implies that China's OFDI flows to most host countries did not cause such serious carbon leakage, but instead made a positive contribution to global CO 2 emission reduction.Among these 34 countries, Australia reduced the largest global CO 2 emissions totally, followed by France and the USA.The cumulative home CO 2 emissions were 11,832.1,5379.7, and 4291.2 kt larger than host CO 2 emissions, respectively.In addition, the cumulative home CO 2 emissions of some developed countries such as Canada, Germany, the UK, and Luxembourg were significantly higher than their host CO 2 emissions, achieving fairly large CO 2 emission reductions.The difference between host and home CO 2 emissions was insignificant in Japan and Korea, so the contribution to global CO 2 emission reduction was negligible.Among these 34 countries, the cumulative home CO 2 emissions were 1794.3 kt larger than host CO 2 emissions in Indonesia compared to the other developing countries, while its cumulative host CO 2 emissions were the largest.For developing countries, Brazil made a CO 2 emissions reduction of 1279.6 kt among the BRICS nations (Brazil, Russia, India, China, and South Africa).In the case of India and Russia, both of their cumulative home CO 2 emissions were lower than host CO 2 emissions.China's OFDI flows to these two countries respectively increased 189.9 and 621.3 kt of global CO 2 emissions.It should be noted that Hong Kong received the largest amount of China's OFDI, and the investment mainly flowed to renting and business services, wholesale and retail trade, finance, and other service industries.Though energy intensity and CO 2 emission intensity of these industries mentioned above were relatively low, it is believed that relatively large CO 2 emission reductions could be achieved due to the huge scale effect of China's OFDI.Moreover, how to account for investment flows through tax havens and the influence on CO 2 emissions is important for a complete understanding of Chinese FDI.Nonetheless, CO 2 emissions of Hong Kong caused by China's OFDI were not calculated or analyzed in this paper by the lack of input-output tables and relevant energy consumption and CO 2 emissions.This part of work will be supplemented and improved in the further study.
industries mentioned above were relatively low, it is believed that relatively large CO2 emission reductions could be achieved due to the huge scale effect of China's OFDI.Moreover, how to account for investment flows through tax havens and the influence on CO2 emissions is important for a complete understanding of Chinese FDI.Nonetheless, CO2 emissions of Hong Kong caused by China's OFDI were not calculated or analyzed in this paper by the lack of input-output tables and relevant energy consumption and CO2 emissions.This part of work will be supplemented and improved in the further study.

Conclusions
Considering low-carbon economic development and the rapid growth of China's OFDI, CO2 emissions in host countries caused by China's OFDI were calculated using input-output analysis from 2000 to 2011.Under the hypothesis of home CO2 emissions, CO2 emission transfer caused by China's OFDI was also analyzed for the first time.An evaluation model was established to investigate the contribution of China's OFDI to the global CO2 emission reduction.Combining the results of 34 selected host countries, our results suggest that China's OFDI had a positive influence on the global CO2 emission reduction as a whole.And other conclusions are as follows: Host CO2 emissions of 34 host countries mostly showed an increasing trend with the rapid growth of China's OFDI.The cumulative host CO2 emissions of Indonesia was the largest, reaching 3872.7 kt, followed by Russia and Australia with 3268.6 and 2517.8 kt.The industry selection of China's OFDI had a distinct influence on host CO2 emissions.The higher proportion of OFDI to resource intensive and energy intensive industries would certainly result in larger host CO2 emissions.The large difference between CEPIs of 34 host countries shows that the location selection also played an important role in global CO2 emission when Chinese OFDI was mainly attracted to resources and market.
Furthermore, the difference between host and home CO2 emissions shows that China's OFDI indeed resulted in "carbon leakage" in some countries e.g., India and Russia.China's OFDI flows to these two countries increased 189.9 and 621.3 kt of global CO2 emissions during the study period.When investing abroad, China should consider the capacity of the environment in the host country.The optimization of industry selection could be an effective measure to reduce host CO2 emissions in such countries with higher CEPIs.
The comparison between host and home CO2 emissions indicates that home CO2 emissions of most host countries were larger than their host CO2 emissions: China's OFDI contributed to global CO2 emission reduction.The cumulative CO2 emission reduction achieved by China's OFDI was 40,454.2kt from 2000 to 2011, in which Australia made the largest contribution of 11,832.1 kt, followed by France of 5379.7 kt and the USA of 4291.2 kt.China's OFDI could significantly reduce global CO2 emissions while the priority of location selection is placed to those countries with higher levels of technology and management, cleaner energy structures, and more efficient energy use.

Conclusions
Considering low-carbon economic development and the rapid growth of China's OFDI, CO 2 emissions in host countries caused by China's OFDI were calculated using input-output analysis from 2000 to 2011.Under the hypothesis of home CO 2 emissions, CO 2 emission transfer caused by China's OFDI was also analyzed for the first time.An evaluation model was established to investigate the contribution of China's OFDI to the global CO 2 emission reduction.Combining the results of 34 selected host countries, our results suggest that China's OFDI had a positive influence on the global CO 2 emission reduction as a whole.And other conclusions are as follows: Host CO 2 emissions of 34 host countries mostly showed an increasing trend with the rapid growth of China's OFDI.The cumulative host CO 2 emissions of Indonesia was the largest, reaching 3872.7 kt, followed by Russia and Australia with 3268.6 and 2517.8 kt.The industry selection of China's OFDI had a distinct influence on host CO 2 emissions.The higher proportion of OFDI to resource intensive and energy intensive industries would certainly result in larger host CO 2 emissions.The large difference between CEPIs of 34 host countries shows that the location selection also played an important role in global CO 2 emission when Chinese OFDI was mainly attracted to resources and market.
Furthermore, the difference between host and home CO 2 emissions shows that China's OFDI indeed resulted in "carbon leakage" in some countries e.g., India and Russia.China's OFDI flows to these two countries increased 189.9 and 621.3 kt of global CO 2 emissions during the study period.When investing abroad, China should consider the capacity of the environment in the host country.The optimization of industry selection could be an effective measure to reduce host CO 2 emissions in such countries with higher CEPIs.
The comparison between host and home CO 2 emissions indicates that home CO 2 emissions of most host countries were larger than their host CO 2 emissions: China's OFDI contributed to global CO 2 emission reduction.The cumulative CO 2 emission reduction achieved by China's OFDI was 40,454.2kt from 2000 to 2011, in which Australia made the largest contribution of 11,832.1 kt, followed by France of 5379.7 kt and the USA of 4291.2 kt.China's OFDI could significantly reduce global CO 2 emissions while the priority of location selection is placed to those countries with higher levels of technology and management, cleaner energy structures, and more efficient energy use.

Table 1 .
Industrial sectors of China's outward foreign direct investment (OFDI) flows.

Table 2 .
Emission relevant energy types.