A structural decomposition analysis of the air
pollution embodied in trade of China and a
comparison with India and Mexico
Yongda Zhang (
y.zhang.43@student.rug.nl)
Student No. S2667568
Faculty of Economics and Business, University of Groningen Supervisor: prof. dr. J. (Jan) Oosterhaven
Abstract
This paper examines the air pollution embodied in the international trade of China, and compares the results with Mexico and India. We use the input-output framework to gain an insight into the air pollution change embodied in trade. By using structural decomposition analysis (SDA), it is easy to analyze the driving forces for changes in air pollution and quantify how much air pollution is embodied in trade, technology and preference changes between 1999-2009. The SDA result shows that macro-economic final demand contributed most in air pollution changes, while the sectorial results are mixed. It is concluded that China, India and Mexico should be held responsible for most of their own air pollution.
1. Introduction
In recent decades, China has reached a fast economic growth with annual GDP growth rates of around 8%. The rapid processes of industrialization and urbanization relied heavily on energy consumption, especially on coal. About 52% coal consumption of the world occurred in China (BP Statistical Review of World Energy 2013) and coal accounts for 67% of the national Chinese energy mix.（BP Statistical Review of World Energy 2014）. Due to coal combustion and many other reasons， air pollution is becoming a vital environment issue. Many cities in China have suffered from air pollution since 1980s. And during the 1990s, some megacities, such as Beijing, Shanghai and Guangzhou are always listed among the top 10 most polluted cities in the world (He, et al, 2002). What’s more, air pollution in East China is increasingly severe and has constrained China’s economic and social development . Empirical research suggests that pollution costs are already quite high in China. For example, Chinese air pollution was linked to 1.23 million premature deaths in 2010, which in monetary terms is equivalent to between 9.7% and 13.2% of GDP, said the Better Growth, Better Climate report (2014). Other researchers found that the air pollution in China has generated huge socio-economic costs, which are measured in terms of consumption loss (Matus, et al, 2012). They found that during the three decades from 1975 to 2005, air pollution in China reduced annual consumption levels between 7% and 23%, in which consumption loss is measured by both net wage loss and welfare loss. And the consumption loss in absolute terms during the same period continuously increased from US$16 billion in 1975 through US$24 billion in 1990 to US$69 billion in 2005.
In 2013, China became the world’s largest trading nation in term of the sum of export and import. As the export and import becoming more and more important for China, it is worth noting about the air pollution is not only embodied in domestic consumption but also in international trade.
the emergence of air pollution problems as an urgent issue that requires international action. Nowadays more and more serious air pollution problems occur, such as acid rain and PM2.5, which is called “fine particles” and consists of all types combustion and Toxic Haze. These problems are becoming hot topics not only in environmental areas but also in economic fields.
Studies on Chinese air pollution have been conducted on many levels, ranging from air pollution problems in China’s mega cities (Chan and Yao, 2008) to government policies and future strategies for air pollution problems (He, et al, 2002). Several studies have used the input-output framework to analyze the air emission in China. However, many studies only focus on a single air pollutant, such as CO2 or SO2 (e.g. Chen and Zhang, 2010; Liang, et al, 2007; Liu and Wang, 2015). In this paper, we would like to combine the air pollutants, which we get from the WIOD, as a whole. In order to do this, we need to weight the different pollutants, because different air emissions do not have the same impact on the different consequences of air pollution. For example, CO2 emission is most responsible for the greenhouse effect, but it contributes almost nothing to local air pollution. Besides, these studies mainly focus on a single country without comparison with others.
importer in Latin America.
The plan of the paper is as follows. Section 2 provides a literature review. Section 3 introduces the setup of the model, the decomposition techniques and data sources. Section 4 analyses the results for China, and compares them with India and Mexico during the period 1999 to 2009. Finally, conclusions are presented in section 5.
2. Literature Review
This paper has some significant predecessors. Using the input-output framework and the decomposition technique, Oosterhaven (1997) analyzes the preference, technology, trade and real income changes in the European Union and explains different factors that lead to the changes. We use a similar decomposition methodology as Oosterhaven did.
For emission studies, many articles focus on air emissions embodied in trade (EET), for example, by using IO analysis and adjusted bilateral trade data, Xu, Allenby and Chen (2009) studied the EET from China to US during the period 2002 to 2007. They found that the CO!, SO! and NO! embodied in Chinese exports to the US are responsible for about 22%-30%, 20%-25% and 23%-29%, respectively, of the total Chinese emissions. Wang and Watson (2008) also quantify the emissions embodied in China’s exports. They found that in 2014, 23% of the total carbon emissions in China were linked to international trade. The phenomenon of emissions, such as NO!, SO!, shifting from developed to developing counties, is a substantial growing problem. (Kanemoto, et al, 2014)
factors behind the changes in the CO! emissions embodied in the China-US trade. The results show that the export volume was the largest driving force in CO! emission embodied in trade (Du, et al. 2011).
Others concentrate more on non-CO! greenhouse gas (GHG) emissions. For example, Zhang, Chen and et al (2015) did research on non-CO! emissions, covering CH!, N!O, etc. The result shows that the impact of international trade on China’s non-CO! GHG emission is significant, in fact, equivalent to 35.6% of the total domestic emissions. Besides, Liu and Wang (2015) reexamined the SO! emissions embodied in China’s exports. Results for 2002-2007 show that SO! embodied in exports is responsible for 15%-23% of the total domestic SO! emissions and most of them were emitted from the eastern provinces where most exports occur. We can conclude that the international trade not only includes CO! emissions but also lots of non-CO! GHG. So in our paper, instead of studying the pollutant such as SO! alone, we focus on air pollution as a whole. Further details are shown in the next section.
3. Methodology and data
3.1. Input-output framework
Fig. 1. Multi-regional Input-Output Table Intermediate purchases of
sector j in s
Final demand Total output Intermediate sales of sector i in r 𝑧!! … 𝑧!! … 𝑧!! 𝑓!! … 𝑓!! … 𝑓!! 𝑥! ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ 𝑧!! … 𝑧!" … 𝑧!" 𝑓!! … 𝑓!" … 𝑓!" 𝑥! ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ 𝑧!! … 𝑧!" … 𝑧!! 𝑓!! … 𝑓!" … 𝑓!! 𝑥! Primary inputs 𝑣! … 𝑣! … 𝑣! Total output (𝑥!)′ … (𝑥!)′ … (𝑥!)′ 𝑦! … 𝑦! … 𝑦! Air pollution 𝑒! … 𝑒! … 𝑒!
indicates the deliveries of product 𝑖 from 𝑟 (which is produced in sector 𝑖) that are sold to sector 𝑗 in 𝑠 (to be used as an intermediate input in the production of sector 𝑗). Matrix 𝑓 shows the deliveries to the final demand categories, in which element 𝑓!"!" gives the goods and services produced by sector 𝑖 in 𝑟 that are purchased by category 𝑘 in 𝑠. The column vector 𝑥 shows the total output in each sector by country. It is worth noting that the IO tables are obtained from double entry bookkeeping. As a consequence, for each sector we have that its row sum equals its column sum. (e.g. (𝑥!)′ = 𝑥!).
As originally formulated by Leontief in his breakthrough paper (Leontief, 1970), the total output, 𝑥 can be expressed as the summation of intermediary goods, 𝑍 and final demand, 𝑓1
𝑥!! = 𝑧
!"!"+ 𝑓!"!" (1)
where intermediary goods 𝑍 can be expressed as 𝑧!"!" = 𝑎 !" !"𝑥
!!. 𝐴 is the direct requirement matrix of the economy.
𝑥!!= 𝐴𝑥
!!+ 𝑓!"!" (2)
this equation yields
𝑥!! = (𝐼 − 𝐴)!!𝑓
!"!" (3)
When multiplied with an emission matrix, 𝑤, which is the air pollution coefficient, calculated as
𝑤 = 𝑒!𝑥!! (4)
the total air pollution 𝐸 can be calculated as
𝐸 = 𝑤′(𝐼 − 𝐴)!!𝑓 (5)
with 𝐸 representing the total air pollution to meet the need of final demand, 𝑓.
3.2 Decomposition analysis
Generally speaking, the change of a country’s air pollution can be attributed to a lot of factors, the overall economic growth, changes in consumption, etc.
By using the polar decomposition methodology (Dietzenbacher and Los, 1998), we can find how much the different factors contribute to air emission in China over time. The air emission in equation can be written as
𝐸 = 𝑤!𝐿𝑓 (6)
where 𝐿 ≡ (𝐼 − 𝐴)!! is the Leontief inverse. Considering the change of air emission over time, the equation above can be written as
△ 𝐸 = △ 𝑤 𝐿!𝑓!+ 𝑤! △ 𝐿 𝑓!+ 𝑤!𝐿!∆𝑓 (7) A B C
Part A represents the change in air pollution that is attributed to the changes in emission intensities Δ𝑤, which measures the influence of changing the air pollution coefficient. Part B represents the impact of the changes in the Leontief inverse ΔL. Finally, part C shows the changes in final demand ∆𝑓. However, there is one disadvantage of decomposition, called “non-uniqueness problem”, which means the results of decomposition may differ significantly by using different procedures. For 𝑛 determinants, the number of equivalent decomposition form is 𝑛!
To overcome this problem, Dietzenbacher and Los (1998) have proposed a method, which is called “polar decomposition”. Changing the first variable first, in our case, △ 𝑤, and then changing the following variable by sequence to derive the first polar decomposition. While the second polar decomposition is derived starting with the last variable, here is △ 𝑓, and changing the following by orders, which is exactly the opposite way of the first decomposition.
△ 𝐸 = 𝐸! − 𝐸! = △! !!!!! △! !!!! ! + !! △! !!!!! △! !! ! + !!!!∆!!!!!!∆!! ! (8)
Instead of the basic decomposition above, we would like to decompose the air pollution growth further by making distinction between trade and technological changes, (Oosterhaven and Linden, 1997; Oosterhaven and Hoen, 1998). The basic model is then further decomposed into
𝐸 = ŵ𝐿𝐵𝑦 = ŵ 𝐼 − 𝑇!⨂𝐴! !!(𝑇!⊗ 𝐹!)𝑦 (9)
In which,
⨂= Cell by cell multiplication;
ŵ = 1435×1435 diagonal matrix with air pollution coefficients;
𝐴!= 1435×1435 matrix, made up of 41 mutually identical 35×1435 matrices with technical coefficients, 𝑎!"•!, indicating how many products are needed from worldwide
sector 𝑖, per unit of output of sector 𝑗 in country 𝑠; 𝑇!= 1435×1435 matrix of trade coefficients, (𝑡
!"!"), indicating which fraction of this intermediate demand for worldwide products 𝑖 (exercised by sector 𝑗 in country 𝑠) is actually satisfied by supply from country 𝑟;
𝐹!= 1435×41 matrix, which is made up of 41 mutually identical 35×41 matrices with the preference coefficients, (𝑓!"•!), indicating the total need for products from
worldwide sector 𝑖, per unit of final demand of category 𝑞 in country 𝑠;2
𝑇!= 1435×41 matrix with trade coefficients, (𝑡
!"!"), indicating which fraction of this final demand for worldwide product 𝑖 is actually satisfied by sector 𝑖 from country 𝑟;
𝑦= 41 column with macroeconomic demand per category 𝑞, per country 𝑠
In order to separate the different effects between technological component and trade
component from Leontief inverse, △ 𝐿, the intercountry input-output coefficient is written as the product of cell-specific trade coefficients and technical coefficients:
𝑎!"!" = 𝑡!"!"𝑎!"•!
(10)
In which, the dot means a summation over all the countries. This is why the above definition of 𝐴! and 𝑇! refers to the worldwide sector 𝑖.
Then, the further decomposition of changes in the inter-country Leontief inverse can be calculated as:
∆𝐿 = 𝐿!− 𝐿! = 𝐿!∆(𝑇!⨂𝐴!)𝐿
! (11)
This can be proven by respectively pre- and post-multiplication of the above equation with (𝐼 − 𝐴!) and (𝐼 − 𝐴!) . By using the polar decomposition rules that we mentioned earlier, the equation is shown as:
∆𝐿 =1
2𝐿! 𝑇!!+ 𝑇!! ⨂∆𝐴!𝐿!+ 1
2𝐿!∆𝑇!⨂(𝐴!! + 𝐴!!)𝐿! (12) D E
The part D of equation (12) indicates the impact of ∆𝐴!, which means how much of the changes in Leontief-inverse are due to the actual changes in the technical coefficients. At the same time, part E shows the impact of the changes in the trade coefficients on the Leontief inverse.
Following the same logic, the final demand 𝑓 can also be further decomposed into final demand trade coefficients (𝑇!) and preference coefficients (𝐹!). The equation is shown as:
𝑏!"!" = 𝑡
!"!"𝑓!"•! (13)
In which the dot also means a summation over all countries of origin. Again, by using polar decomposition approach we use above, we can get:
∆𝐵 =1
2∆𝑇!⊗ 𝐹!!+ 𝐹!! + 1
F G
Part F shows the impact of ∆𝑇!, which means the changes in the final output trade patterns on final demand. Meanwhile, Part G shows the pure final demand preference changes on the final demand, which is one of the technical coefficients. Changes in consumers’ preferences are the majority part in the ∆𝐹!.
The final decomposition is reached by first applying procedure of equation (8) to equation of (9), which results in four components, ∆𝑤 , ∆𝐿 , ∆𝐵 and ∆𝑦 respectively. Second, we substitute the ∆𝐿 with equation (12) and ∆𝐵 with equation (14). Consequently, we get the final decomposition of ∆𝐸,
△ 𝐸 = △! !!!!!!! △! !!!!!! ! (15.1) +!!!! !!!!!!! ⨂∆!!!!!!!!!!!!! !!!!!!! ⨂∆!!!!!!!! ! (15.2) +!!!!∆!!⨂ !!!!!!! !!!!!!!!!!!∆!! ⨂ (!!!!!!!)!!!!!! ! (15.3) +!!!!∆!!⊗ !!!!!!! !!!!!!!∆!!⊗ !!!!!!! !! ! (15.4) +!!!! !! !!! !! ⊗∆!!!!!!!!! !!!!!!! ⊗∆!!!! ! (15.5) +!!!!!!∆!!!!!!!!∆! ! (15.6)
The first two elements show the result of impact from the changes in the sectorial technology, as they refer to the emission coefficient △ 𝑤 and the technical coefficient ∆𝐴! in ∆𝐿, respectively. The next two elements are related to ∆𝑇! and ∆𝑇!, which are both trade coefficients and help to explain the influence of trade patterns changes. Finally, the last two elements summarize the impact of ∆𝐹! and ∆𝑦, which indicate the changes in the final demand preferences and changes in the macro economic demand.
3.3 Data
we use the 1999 and 2009 data. The WIOT covers 27 EU countries and 13 other major countries in the world for the period from 1995 to 2011, while each country contains 35 industries. The air emission data are also from the WIOD, and they are found under the environmental accounts with the same amount of countries but for the period from 1995 to 2009. They are provided by sector and pollutant and are taken from a wide variety of sources to make sure that, for each sector, the data are at great level of detail. It contains emission data of CO!, CH!, N!O, NO!, SO!, CO, NMVOC and NH! for 35 industries in each country.
3.4 Constant price data
Most commonly, a structural decomposition analysis (SDA) is used to estimate changes of different factors over time. Because the air pollution coefficient, as one of the potential driving forces of air pollution, is measured as the air pollution per unit of output, using the data in current prices will lead to biases in the results. For example, suppose that an industry produces the same amount of products and the air pollution has remained exactly the same in 2009 as it did in 1999. Then, the air pollution coefficient should also have remained the same. However, the price of the output will have increased/decreased over those years due to inflation/deflation, and, consequently, the energy intensities calculation based on current price are decrease/increase.
In this paper, we use input-output tables that are published in monetary units. In order to convert the data of different years into the same price system (constant price), we may use the data on the IOTs in prices of both the current year and the previous year.
this paper, we are using the constant price chaining method from Oosterhaven (2015). Assuming that a value is given in current year prices (𝑧!!) and a value is given in the last year price (𝑧!!!!). Then the pure growth rate of each cell during the period 𝑡-1 to 𝑡 can be calculated as:
𝑧!!!,!− 1 =𝑧!!!!
𝑧!!!!!!− 1 (16)
In which, 𝑧!!!,!− 1 indicates the pure growth rate of cell 𝑧 from period 𝑡-1 to 𝑡. Both rates are dimensionless numbers, as the numerator and the denominator are measured in prices of the same year. Following the same logic, the growth of cell 𝑧 from 1999 to 2009 can be calculated as:
𝑧!""",!""# = !""# 𝑧!!!,!
!!!""" (17)
Thus, for any cell in the IOT, its value of 2009 in the 1999 price can be calculated as: 𝑧!""#!""" = 𝑧
!"""!"""∗ 𝑧!""",!""#3 (18)
3.5 Weight data
In Europe, ozone is responsible for almost 5% of the total health damage due to air pollution (EEA, 2014). NO!, NMVOC, CO and CH! contribute to tropospheric ozone formation. Frank de Leeuw (2002) has estimated the Tropospheric Ozone Formation Potential (TFOP) for each of these substances.
Fig. 2*
*Note. Adapted from A set of emission indicators for long-range trans boundary air pollution. Environmental Science and Policy, Volume 5, Issue 2, p. 135-145, by de Leeuw, F, 2002.
The 95% rest of the total health damage due to air pollution is caused by fine particles (PM2.5). Roughly 35%-40% of the PM2.5 exposure is caused by primary emission of fine particles (e.g. soot and dust), the other 60%-65% is formed in atmosphere. NO!, SO! NMVOC and NH! are precursors of such （secondary） particles. IIASA has developed a method to weigh such precursors in PM2.5 equivalents.
Fig. 3. Weight index in PM2.5 equivalents*
*Note. Adapted from IASA TSAP Report 15: A Flexibility Mechanism for Complying with National Emission Ceilings for Air
Pollutants, by Amann, M., & Wagner, F., 2014.
But the data for PM2.5 is not available from WIOD. We know that for a good balance between air pollutions and health, it is better to include primary particles (e.g. PM2.5) in the list of pollutants. These particles have a heavy weight in the health impacts. As most of these particles are emitted by the same sources that cause emission of greenhouse gases (especially traffic, residential heating, coal fired power plants, oil and gas recovery), there is a large potential for synergies between air pollution and health policy. But due to the limitation of data, we couldn’t do it in this paper.
Fig. 4. The weight index for different pollutant
Pollutant 𝐂𝐎𝟐 𝐂𝐇𝟒 𝐍𝟐𝐎 𝐍𝐎𝐗 𝐒𝐎𝐗 CO NMVOC 𝐍𝐇𝟑
Weight 0 0.0001 0 0.14935 0.27671 0.00786 0.07143 0.18014
There is one caveat of the weight data. Recent information of RIVM and WHO also show a direst health effect of NO! (Fischer, et al, 2015; Henschel and Chan, 2013). If this finding would be taken into account, NO! would cause almost 30% percent of the health impacts directly, PM2.5 would be responsible for around 65% of the total health damage due to air pollution and ozone will be counted less than 5%. NO! would cause health damage in both directly and indirectly via the formation of ozone and as a precursor for PM2.5.
4. Results
International trade of China, India and Mexico, which are all G20 countries together with almost 38% of the world population, has witnessed rapid growth in recent years. In this section, we will first have a look at the aggregate results of air pollution changes in these countries between 1999-2009. Then, we will examine the decomposition results and find out different contributors for air pollution changes. Finally, sectorial decomposition results will be presented in Table 3.
4.1 Aggregate air pollution change
Table 1. Air pollution from 1999 to 2009 (Unit: Kt)
Air pollution China India Mexico
Due to domestic final demand Year 1999 7742 2794 1088 Change 1999-2009 6492 1653 -203 Change in % 84% 59 % -19% EEE, emission due to foreign exports Year 1999 1724 438 219 Change 1999-2009 2819 24 -81 Change in % 164% 5% -37% EEI, emission in the rest of the world due to foreign import
Year 1999 379 232 188 Change 1999-2009 870 431 146 Change in % 230% 185% 78%
The air pollutions in China for its domestic final demand were 7.7 Mt in 1999. This amount tremendously increased to 14.3 Mt in 2009, which is almost 84%. During the same period, Chinese air pollutions embodied in exports4 (EEE) increased from 1.7 Mt to 3.8 Mt, which rocketed 164%, but in absolute value it was far less than air pollution for domestic final demand. The air pollution in China embodied in import grew from 0.38 Mt to 1.2 Mt.
Similar as China, India also had positive air pollution growth embodied in domestic final demand, exports and imports. Emission embodied in import had a big growth while emission embodied in export only had a 5.4% growth rate.
It’s worth noting that air pollution of Mexico embodied in import had a positive growth, while air pollution embodied both in its domestic final demand and exports had negative growth. In order to explain findings above, decomposition analysis is needed in the next section.
4.2 Decomposition result of total air pollution
Table 25. Contributions of drivers to embodied air pollution changes between 1999-2009 △ 𝒘 ∆𝑨𝒕 ∆𝑻𝒂 ∆𝑻𝒇 ∆𝑭𝒕 ∆𝒚 Total (Kt) CHN Domestic -98% 50% -1% 0% -18% 167% 6492 EEE -126% 43 % 90% 75 % -5% 24% 2819 EEI -39% 39% 15% 3% 2 % 81% 870 IND Domestic -31% -9% 1% -5% -12% 156% 1653 EEE -753% -363% 157% 772% -210% 497% 23.6 EEI -73% 40% 18% 27% 21% 67% 430.9 MEX Domestic -208% 17% -20% -24% -8% 143% -203 EEE -121% -15% 4% 2% -19% 49% -81 EEI -99% 23% 67% 66% -1% 50% 146
Given these decomposition results percentage number importance, we now analyze the structure decomposition results. Table 2 shows the results of structural decomposition analysis on air pollution change embodied in both domestic final demand and trade between 1999-2009. It is clear to see that the change in both emission coefficient (△ 𝑤) and macro-economic demand (∆y) contributed a lot in all these cases. It is also worth noting that, air pollution embodied in emission coefficient reduced in all these situations during this period while air pollution embodied in macro-economic demand increased a lot.
From table 1, we already know that the emission change embodied in export of India had a small 5.4% growth, however, this was because of a mix effect. Air pollution change embodied in emission coefficient (△ 𝑤) and technical coefficient (∆𝐴!) both had tremendous negative growth. This impact was partly compensated by trade coefficients (∆𝑇!) and (∆𝑇!). The positive impact of trade coefficients indicates that more intermediate input is purchased by import, and more final demands is served from trade. Besides, trade coefficient (∆𝑇!) and (∆𝑇!) also had strong impacts on air pollution change embodied in export of China and import of Mexico. Moreover, it is clear to see that technical coefficient (∆𝐴!) matters in China.
For all three countries, the negative change of emission due to changes in the emission coefficients (△ 𝑤) might partly be caused by the pollution control. For example in China, government battled to reduce smog. And stricter rules were set for heavy polluting manufacturing. Another reason is the improvement of the efficiency in the use of primary factors, such as minerals. The change in technical coefficients (∆𝐴!) had different signs. The negative change of air pollution due to changes in technical coefficients (∆𝐴!) means that the intermediate sales of pollution-intensive industries became smaller due to the improvement of technological change, and vice versa. We can see that the negative changes of emission embodied in technical coefficient were only found in domestic final demand of India and export of Mexico.
Another interesting point is that almost all changes in final demand preferences (∆𝐹!) had negative impacts, except the one embodied in import of China. This must be due to a preference shifting to high-tech or pollution-free industries. Trade coefficient change (∆𝑇!) had different signs not only among different countries but also within country for different final demands. Note that change of trade coefficient (∆𝑇!) is mainly due to the trade pattern change in the final demand. Exports of China, India and imports of Mexico are the most striking cases for the changes of emission embodied in trade coefficient (∆𝑇!).
Finally, all three countries had positive changes in air pollution embodied in the macro-economic demand (∆𝑦). It is easy to imagine that as the final demand increase, the embodied air pollution will grow.
In order to explain why some components, not the others have strong impacts, we need to have a look at the impact of these components at a more detailed sectorial level.
4.3 Sectorial decomposition result
coefficient change, trade coefficient change and macro-economic demand change. Therefore, in this section we will first focus on the decomposition result of technical coefficient change and then check the results of trade coefficient change, and finally those of macro-economic demand change. We will not discuss all these results in details. Instead, we only concentrate on some interesting points and general tendencies.
Table 3. Sectorial decomposition result of technical coefficient change (∆𝑤 + ∆𝐴!+ ∆𝐹!) between
31 -‐17626 -‐16 275 -‐2158 0 -‐94 -‐6352 -‐13 -‐442 32 -‐34626 -‐643 13 -‐2042 -‐1 -‐16 -‐7836 -‐9 -‐23 33 1693 237 55 -‐1770 -‐3 12 -‐2913 -‐1 4 34 -‐31136 -‐6390 -‐79 -‐6036 -‐703 -‐343 -‐2832 -‐35 -‐345
35 0 0 0 0 0 0 0 0 0
From table 2, it is easy to conclude that the overall emission due to technical coefficient change is negative between 1999-2009. While, in table 3 we can see the detailed impact of technical coefficient change on sectorial level and try to find out which sector contribute most to the technical coefficient change. It is easy to recognize that many sectors, such as sector 17, 23, 24, namely Electricity, Gas and Water Supply, Inland Transport and Water Transport, contribute large reductions in all six columns. This reduction effect is partly compensated by sector 1, Agriculture, Hunting, Forestry and Fishing, which has positive air pollution growth in all six columns. This may due to the lack of technological improving in reducing air pollution and the growth of intermediate input. Moreover, many other sectors such as 5, 7 and 16 also have negative air pollution growth for all columns during this period. Table 4. Sectorial decomposition result of trade coefficient change (∆𝑇!+∆𝑇!) between 1999-2009
17 31714 2144797 54858 -‐62583 109966 85547 -‐10796 45011 59270 18 116 2605 245 -‐1066 1539 152 -‐30 36 12 19 0 0 -‐11 -‐14 31 12 -‐92 0 17 20 -‐182 12982 310 -‐5 98 438 -‐123 140 257 21 40 4482 112 -‐249 436 429 -‐330 5 121 22 -‐591 4180 400 -‐2583 9033 269 -‐91 78 53 23 -‐9029 83032 11674 1730 7635 -‐5427 -‐1188 1896 2095 24 8711 429222 10562 -‐3936 7536 18050 -‐1876 3435 8869 25 -‐6066 23781 4416 -‐382 447 340 -‐255 -‐88 133 26 -‐334 2699 547 -‐10 369 198 -‐92 69 141 27 -‐364 2391 489 -‐866 1646 202 -‐40 46 143 28 87 2339 75 -‐181 250 183 -‐38 29 108 29 12 772 51 -‐176 -‐1 636 -‐9 22 20 30 -‐1128 9463 -‐166 -‐133 1627 348 -‐262 5 213 31 55 15 -‐51 0 0 27 -‐253 3 353 32 -‐173 662 17 -‐2 7 27 -‐12 11 8 33 66 687 -‐28 -‐2 5 29 0 13 22 34 -‐443 5375 303 -‐275 576 314 0 116 121 35 0 0 0 0 0 0 0 0 0
33 15510 134 108 1046 1 31 949 0 8
34 42893 1636 1425 6194 259 422 1029 8 152
35 0 0 0 0 0 0 0 0 0
Again, on the sectorial basis, table 5 shows the distribution of air pollution embodied macro-economic final demand change. Contrary to Table 3, almost every sector had a positive contribution for the air pollution growth. During the period 1999 and 2009, macro-economic demand increased almost every sector. It is interesting to see that the sector 1, agriculture, hunting, forestry and fishing, was still the main contributor. Another interesting point is seen in the impact of sector 17. This time, electricity, gas and water supply carried a big positive impact for the air pollution growth, which means a wider use of them. Other modern sectors such as sector 8 and 9, namely coke, refined petroleum and nuclear fuel, and chemicals and chemical products also had a great positive impact.
5. Conclusion
The above analysis provides a modification of the standard input-output structure decomposition analysis. By using the great structure decomposition methodology from Oosterhaven and van der Linden (1997), different effects between changes in technology and changes in trade patterns could be easily distinguish. We decomposed not only the traditional input-output coefficients, but also the final demand bridge coefficients.
In the section 4, we have discussed some of the surprising and interesting results related to the trade, technology and preference change at both aggregated level and at the sectorial level. However, it is not quite useful and efficient to summarize the wide variety of individual tables and results.
References
Markus, A., & Fabian, W. (2014). A flexibility mechanism for complying with national emission ceilings for air pollutants. Laxenburg: International Institute for Applied Systems Analysis (IIASA).
The Global Commission on the Economy and Climate (2014). Better growth better climate: The new climate economy report, synthesis report. edited by Felipe Calderon et al. Washington, D.C.: The Global Commission on the Economy and Climate.
British Petroleum Company. (2013). BP statistical review of world energy. London: British Petroleum Co.
British Petroleum Company. (2014). BP statistical review of world energy. London: British Petroleum Co.
Chan, C., & Yao, X. (2008). Air pollution in mega cities in China. Atmospheric Environment, 1-42.
Chen, G., & Zhang, B. (2010). Greenhouse gas emissions in China 2007: Inventory and input–output analysis. Energy Policy, 6180-6193.
Dietzenbacher, E., & Los, B. (1998). Structural Decomposition Techniques: Sense and Sensitivity. Economic Systems Research, 307-324.
Du, H., Guo, J., Mao, G., Smith, A., Wang, X., & Wang, Y. (2011). CO2 emissions embodied in China–US trade: Input–output analysis based on the emergy/dollar ratio. Energy Policy, 5980-5987.
Leeuw, F. (2002). A set of emission indicators for long-range transboundary air pollution. Environmental Science & Policy, 135-145.
Fischer, P., Marra, M., Ameling, C., Hoek, G., Beelen, R., Hoogh, K., . . . Houthuijs, D. (2015). Air Pollution and Mortality in Seven Million Adults: The Dutch Environmental Longitudinal Study (DUELS). Environmental Health Perspectives.
Henschel, S. and G. Chan (2013), Health risks of air pollution in Europe – HRAPIE Project, New emerging risks from air pollution – results from the survey of experts, World Health Organization Regional Office for Europe, Copenhagen, www.euro.who.int/en/health-topics/environment-and-health/air-quality/publicatio ns/2013/health-risks-of-air-pollution-in-europe-hrapie-project.-new-emerging-ris ks-to-health-from-air-pollution- results-from-the-survey-of-experts.
He, K., Huo, H., & Zhang, Q. (2002). Urban air pollution in China : Current status, characteristics, and progress. Annual Review of Energy and the Environment, 397-431.
Kanemoto, K., Moran, D., Lenzen, M., & Geschke, A. (2014). International trade undermines national emission reduction targets: New evidence from air pollution. Global Environmental Change, 52-59.
Liang, Q., Fan, Y., & Wei, Y. (2007). Multi-regional input–output model for regional energy requirements and CO2 emissions in China. Energy Policy, 1685-1700. Liu, Q., & Wang, Q. (2015). Reexamine SO2 emissions embodied in China's exports
using multiregional input–output analysis. Ecological Economics, 39-50.
Matus, K., Nam, K. M., Selin, N. E., Lamsal, L. N., Reilly, J. M., & Paltsev, S. (2012). Health damages from air pollution in China. Global Environmental Change, 22, 1, 55-66.
Oosterhaven, J., & Van der Linden. (1997). European Technology, Trade and Income Changes for 1975–85: An Intercountry Input–Output Decomposition. Economic Systems Research, 393-412.
Oosterhaven, J., & Hoen, A. (1998). Preferences, technology, trade and real income changes in the European Union. The Annals of Regional Science, 505-524.
Oosterhaven, J. (2015). A note on the constant price chaining of input-output tables. Mimeo, University of Groningen
Wang, T., & Watson, J. (2008). China's carbon emissions and international trade: Implications for post-2012 policy. Climate Policy, 577-587.
4875-4881.
Weber, C. (2009). Measuring structural change and energy use: Decomposition of the US economy from 1997 to 2002. Energy Policy, 1561-1570.
Leontief, W. (1970). Environmental Repercussions and the Economic Structure: An Input-Output Approach. The Review of Economics and Statistics, 262-262.
Xu, M., Allenby, B., & Chen, W. (2009). Energy and Air Emissions Embodied in China−U.S. Trade: Eastbound Assessment Using Adjusted Bilateral Trade Data. Environmental Science & Technology, 3378-3384.
Appendix
Table 1A Sectors and codesCode Sector Code Sector
1
Agriculture, Hunting, Forestry and
Fishing 19
Sale, Maintenance and Repair of Motor Vehicles and Motorcycles; Retail Sale of Fuel
2 Mining and Quarrying 20
Wholesale Trade and Commission Trade, Except of Motor Vehicles and Motorcycles
3 Food, Beverages and Tobacco 21
Retail Trade, Except of Motor Vehicles and Motorcycles; Repair of Household Goods
4 Textiles and Textile Products 22 Hotels and Restaurants 5 Leather, Leather and Footwear 23 Inland Transport 6 Wood and Products of Wood and Cork 24 Water Transport
7
Pulp, Paper, Paper , Printing and
Publishing 25 Air Transport
8
Coke, Refined Petroleum and Nuclear
Fuel 26
Other Supporting and Auxiliary Transport Activities; Activities of Travel Agencies
9 Chemicals and Chemical Products 27 Post and Telecommunications 10 Rubber and Plastics 28 Financial Intermediation 11 Other Non-Metallic Mineral 29 Real Estate Activities
12 Basic Metals and Fabricated Metal 30 Renting of M&Eq and Other Business Activities 13 Machinery, Nec 31 Public Admin and Defence; Compulsory Social Security 14 Electrical and Optical Equipment 32 Education
15 Transport Equipment 33 Health and Social Work
