Driving Forces of Economic Growth in BRIC Countries

(1)

Driving Forces of Economic Growth in BRIC Countries

Author: Meixu Chen University of Groningen

Faculty of Economics and Business E-mail: m.chen.10@student.rug.nl S-number: 2473593

Supervisor: Prof. dr. J. (Jan) Oosterhaven University of Groningen

(2)

1

Abstract

This paper aims to quantify the driving forces behind the rapid economic growth in the BRIC countries. By using input-output tables from the World Input-Output Database, we look at the economic growth in terms of value added and employment in the four BRIC countries from 1995 to 2007. A structural decomposition is applied in this paper to separate the economic growth into six main driving forces, namely, technology improvement, value added or employment coefficient change, increasing trade of intermediates and final goods, changing consumer preferences and macroeconomic final demands. Moreover, a chaining technique is used to deflate the price effect over the period and give us the volume changes, which is more reliable. By looking at the growth due to domestic and foreign final demand in the BRIC countries at macro level, this paper suggests that most of the economic growth that is driven by domestic final demand can be largely explained by the high macroeconomic final demand. While the other driving forces seem to be more important than macroeconomic final demand in some BRIC countries when we look at the economic growth due to foreign final demand. At sectoral level, the economic growth of sectors are very different from each other, and the contribution of each component to their growth varies among sectors.

(3)

2

1. Introduction

The term BRIC was first coined by O’Neill (2001) to describe a group of populous emerging countries consisting of Brazil, the Russian Federation, India and China, whose economic developments account for a significant share of economic growth, demand expansion, industrial production and wealth creation in today’s world. The driving forces of BRIC countries in achieving sustained economic growth have become an interesting topic to study and they could also be the reference for the growth of other developing countries.

In recent years, following the trend of free trade and globalization, the BRIC countries have liberalized their controls on production capacity, imported capital goods, intermediate inputs and technology. These changes lead to demand expansion, especially foreign demand, facilitate easy import of better intermediates, capital goods and production technology and transform the consumption preference in these countries (Santos-Paulino, 2012). Moreover, with the rise of BRIC countries, the domestic demand of BRIC products increases rapidly as well (Yamakawa, Ahmed and Kelston, 2009). These components are often considered as the most important contributors to the fast economic growth in emerging economies. But the extent to which these driving forces contribute to the growth in developing countries still need to be studied.

(4)

3

Therefore, this paper is interested in several issues. First, at macro level, how much does each driving force contributes to the rapid economic growth in the BRIC countries? Is there a difference between the growth driven by domestic final demand and growth due to foreign final demand? Second, at sectoral level, how does each component contribute to the growth of different sectors? And how the sectoral structure changes in each country? To answer these questions, this paper decomposes the economic growth of BRIC countries in terms of value added and employment by using input-output data from year 1995 to 2007. We find that, though all BRIC countries have experienced significant growth in value added and employment, the impacts of different driving forces on the growth due to domestic final demand and foreign final demand are different and vary among countries. Moreover, by using structural decomposition analysis (SDA), the contributors of structural changes and economic growth have also been detected at sectoral level. The growth of value added and labor are clearly more significant in some sectors than the others. The contribution of each driving force to economic growth vary from sector to sector and different from the results at macro level.

The rest of this paper is organized as follows. Section 2 mentions the theoretical background for this paper. Section 3 describes the methodology we adopted to decompose the economic growth into sub-components changes, i.e. changes in trade, technology, consumer preference and macroeconomic final demand. Section 4 describes data that have been used. Section 5 gives an analysis of the empirical results from decomposition. Conclusions are presented in section 6.

2. Literature review

(5)

4

the productivity varies among sectors, he found that increasing growth in European countries was also driven by shifting employment among sectors. An analysis with a more detailed sectoral data, Jorgenson and Timmer (2010) confirmed that structural change contributes to both growth of value added and employment in advanced nations. However, growth accounting cannot explain why the sectoral structure changes. According to Dietzenbacher, De Groot and Los (2007), the factors that drive structural change cannot be addressed through the growth accounting method, because the input levels of industries are considered as exogenous variables.

Another way to study structural change is from a demand-driven perspective, which has been largely used from Leontief (1966) onward. Based on input-output tables, which detail the flows of products among industrial sectors, this perspective focuses on changes in inter-industry structures and tries to decompose structural changes into various driving components. Using this approach, Feldman, McClain and Palmer (1987) found that, in United States, most of the structural change was driven by the increasing macroeconomic final demands. Oosterhaven and Hoen (1998) further separated trade effect from other factors like changes in technology and consumer preferences, and claimed that, at country level, macroeconomic final demand growth still remains the most important component, but at sector level, the influence of changes in input coefficients, i.e. trade and technology improvements, on economic growth is quite large. More studies show that changes in domestic demand and international trade patterns drive a process of structural changes in which labor, capital and intermediate inputs are continuously relocated between sectors (Chenery, 1960; Shishido and Watanabe, 1962; Kuznets,1966; Syrquin,1976; Chenery ,1979; Chenery, Robinson and Syrquin, 1986).

(6)

5

the economic growth by using both the total hours worked by person and value added, which provides more accurate and policy relevant indicators than wage compensation and output volumes. Second, to eliminate the influences of prices and exchange rates, this paper corrects the nominal changes to volume changes by adopting the chaining technique to convert input-output tables from current price to constant prices. Third, we further separate the effects of trade, both in intermediates and in final products, from the effects of technology improvement and consumer preference changes and study how each component drives the structural changes and economic growth.

3. Methodology

In this section, we outline our approach to measure to what extend the BRIC countries’ growth depends on each driving force. The starting point is the input-output analysis which was developed by Leontief in 1930s. The input-output table allows one to trace the factor inputs needed in all stages of production of a particular final good. The structure of the IO table used in this paper is represented by Figure 1. By tracing the value added at all stages of production, it provides an ex-post accounting of the value of final demand (Los, Timmer, de Vries, 2014). Thus, we can measure the importance of foreign demand relative to domestic demand for the growth of domestic value added in a consistent framework.

The bi-country IO tables for four BRIC countries are constructed from the world input-output tables in WIOD.

Zrr Zrs frr frs xr

Zsr Zss fsr fss xs

vr’ vs’

xr’ xs’

(7)

6

As shown in Figure 1, country r refers to one of the BRIC countries, and country s denotes the aggregation of the rest 39 countries available in the database. The four matrix Zrr, Zrs, Zsr, Zss on the top left form the matrix Z.1 It is the matrix of intermediate inputs. Their elements and refer to the purchases of intermediate inputs of sector i in country r by sector j in country r and country s respectively, while and represent that the usage of intermediate inputs from sector i in country s by sector j in country r and country s respectively. The matrix of input coefficients then can be calculated as A = Zx̂-1 where x̂ is the diagonal matrix of output with the elements of vector x (xr, xs) along its main diagonal and zeros elsewhere. Vector f (frr, frs, fsr, fss) denotes the final demands in both countries. Similar to the notations in matrix Z, their elements and refer to the final demand of sector i in country r by category k in country r and country s, while and are the final demand of sector i in country s by category k in country r and country s. The standard IO model is x = Ax + f. If (I - A) is non-singular, where I is an identity matrix with ones on the diagonal and zeros elsewhere, then the relationship between all sectors of an economy is given by x = (I − A)-1f = Lf, where L ≡ (I − A)-1 is referred as the Leontief-inverse and represents the total output of each industry i to produce one unit of final demand for product from industry j. The row vector v’ is a general term to denote either value added or employment in each country. Its elements are given by vi ( , ), which represents either

gross value added or total hours worked by people of sector i in each country. Then the coefficients for value added or employment can be calculates as c’ = v’x̂-1.

Since this paper focuses on the changes of value added and employment in the BRIC countries, we need to link output to value added and employment. Then the general model becomes:

v = ĉ(I − A)-1_{f = ĉLf (1)}

1

(8)

7

where ĉ refers to the diagonal matrix with corresponding value-added coefficients or employment coefficients.

Based on input-output analysis, structural decomposition analysis (SDA) is introduced to disaggregate the total change in some aspect of that economy, for instance, value added or output, into contributions made by its various components, like technology improvement or increasing final demand (Miller and Blair, 2009). However, one disadvantage of using SDA is the so-called “non-uniqueness problem”, which is mentioned by Dietzenbacher and Los (1998). It means that the results of decomposition may differ significantly by using different decomposition methods and the number of equivalent forms increases rapidly when the number of determinants exceeds two. To overcome this problem, they proposed a method named “polar decomposition”. The first polar decomposition is derived by changing the first variable and then changing the following variable in sequence. While the second polar decomposition is derived by starting with the last variable and changing the following in orders, that are exactly the opposite of the first decomposition. And they found that the average of all decompositions can be adequately approximated by average of the two polar decomposition results. Therefore, we apply the polar decomposition to the general model and assume that there are two time periods for which input–output data are available. Using superscripts 0 and 1 for the two different years (0 earlier than 1), we can disaggregate the changes for employment and value added from period 0 to1 as:

First polar decomposition

∆v = ĉ1L1f1 - ĉ0L0f0 = ĉ1L1f1 – ĉ0L1f1 + ĉ0L1f1 – ĉ0L0f1 + ĉ0L0f1 – ĉ0L0f0

= (∆ĉ)L1f1 + ĉ0(∆L)f1 + ĉ0L0(∆f)

Second polar decomposition

∆v = ĉ1L1f1 - ĉ0L0f0 = ĉ1L1f1 – ĉ1L1f0 + ĉ1L1f0 – ĉ1L0f0 + ĉ1L0f0 – ĉ0M0f0

(9)

8

Then, the combined change of employment and value added from period 0 to1 becomes

∆v = ½ (∆ĉL0f0 + ∆ĉL1f1) + ½ (ĉ0∆Lf1+ ĉ1∆Lf0) + ½ (ĉ0L0∆f+ ĉ1L1∆f) (2)

The first term on the right hand refers to the changes in value added or employment if its coefficients have changed while final demand and the Leontief-inverse remain unchanged. In the same way, the second term refers to the contribution of changes in input coefficients and changes in final demand to the change in value added or employment respectively, while other variables remain the same. As mentioned in the introduction, this paper is interested in analyze the effects of technology change and trade on the sectoral growth in terms of value added and employment, further decomposition are required. To separate the impacts of technology improvement and consumer preference changes from trade effect, in intermediates or in final products, we need to decompose both the input coefficients A and final demand f. For this purpose, this paper follows the decomposition approach developed by Oosterhaven and Hoen (1998). This paper separates the technology component from the trade component in ∆L and writes the input coefficients as the product of cell-specific trade coefficients and technical coefficients.

= (3)

where is the summation of industry i in A0 for each country, indicating the total need for

sector i worldwide to produce one unit of sector j in country s. The changes of technical coefficients imply that, due to technology improvement, one unit of output in sector j can be now produced using less intermediate input from sector i. While changes in implies that, the fraction of demands in sector j of country s is satisfied by products i of country r. Then the changes in inter-country Leontief-inverse can be written as:

(10)

9

Where Ta is the matrix of trade coefficients . Using the polar decomposition rule mentioned before, the changes of Leontief-inverse from period 0 to1 is expressed as:

∆L = ½ L1(T0a + T1a)⊗ ∆AL0 + ½ L1∆Ta ⊗ (A0 + A1)L0 (5)

Oosterhaven and Hoen (1998) disaggregated final demand into two components as well. First is the total amount of all expenditures for final demands or macroeconomic demand y for each category k. The empirical application of Section 3 is based on the bi-country input-output tables that record 35 sectors and 6 final demand categories: household consumption, consumption expenditure by non-profit organizations serving households, gross fixed capital formation, changes in inventories and valuables and export. Since this paper does not focus on the shift among final demand categories but on the trade between different countries, we aggregate the six final demands into one category per region. The final demand matrix then becomes F (70×2) for the final demand of country r and country s.

Second is the bridge matrix B (70×2), which indicates the composition of final demand in country r and country s. The element represents the part of final demand in category k that is spend on products from sector i. To separate the effect of trade pattern changes from the impact of preference changes, the bridge coefficients can be decomposed into trade coefficients and preference coefficients (Oosterhaven and Hoen, 1998).

= (6)

(11)

10

r. Therefore, using the polar decomposition again, the changes in bridge matrix B is rewritten

as:

∆B = ½ ∆Tf ⊗ (F0+ F1) + ½ (T0f+ T1f) ⊗ ∆F (7)

Where Tf is the matrix of trade coefficients in final demand and F is the matrix for preference coefficients .Then, substituting (7) into final demand changes gives the changes

in final demand ∆f over time:

∆f = B1y1 –B0y0 = ½ ∆B(y0 + y1) + ½ (B0 + B1)∆y

= ¼ ∆Tf(F0+ F1) (y0 + y1) + ¼ (T0f+ T1f)∆F(y0 + y1)

+ ½ (B0 + B1)∆y (8)

Replacing the new expression for ∆L, ∆B, ∆f into equation (2), the general model (9) can be divided in six components (9.1-9.6):

(12)

11

After decomposing both input coefficients and final demand, it is now clear from the general model (9) that there are six driving forces, ∆ĉ, ∆A, ∆Ta, ∆Tf, ∆F, ∆y, in the changes of ∆v. The first term (9.1) indicates the impacts of changes in value added or employment coefficients. Together with the second term (9.2), they indicate the changes in value added or employment that are caused by sectoral technology improvement. The third (9.3) and fourth (9.4) terms represent the impacts from changes in trade patterns of intermediate products and final products respectively. The last two terms (9.5 and 9.6) relate to the changes in final demand preference and changes in the total final demand.

To distinguish the economic growth that is driven by domestic final demand of BRIC countries from the growth due to foreign final demand, we first calculate them separately by employing decomposition equation (9) twice and get the growth that driven by domestic and foreign final demand respectively. More specifically, the value added or employment of country r due to domestic final demand should be based on:

vr = ĉrLrBryr (10)

where ĉr (1*35) is diagonal matrix with corresponding value-added coefficients or employment coefficients of country r. Lr (35*70) is the matrix including Leontief inverse of country r (Lrr) and of country r to country s (Lrs). Br(70*1) refers to the column with bridge coefficients of country r and yr (1*1)is total final demand of country r. Following the same logic, the formula for value added or employment of country r due to foreign final demand is written as:

vr = ĉrLrBsys (11)

(13)

12

domestic consumption from the contribution due to foreign macroeconomic final demand. More specifically, the terms (9.1-9.5) remain the same, we only change the last term (9.6) by substituting ∆y with ∆yr+∆ys and the changes due to domestic macroeconomic final demand and foreign macroeconomic final demand can be separated. The results at sectoral level are shown in Table 3 and 4.

4. Data

The empirical analysis of structural change in the previous studies has been based on aggregate sector data, which may hide diverging trends at a more detailed level and thereby obscure a proper assessment of the role of structural transformation for aggregate productivity growth (de Vries, et al, 2012). Therefore this paper will use the detailed sector data from the World Input-Output Database (WIOD). It provides a global input-output matrix for 41 countries (including a rest-of-world category) over the period 1995-2011. It classifies all sectors into 35 categories (see Appendix B for the classification of sectors). This paper employs one branch of WIOD data, called the world input-output tables (WIOTs) for 40 countries (excluding a rest-of-world category) from 1995 to 2007. We do not include 2007 afterward, because the data after 2007 are less reliable due to the economic crisis. As mentioned in Section 2, this paper aggregates the data to a bi-national input-output table. Therefore, country r refers to Brazil, the Russian Federation, India or China, and country s represents the integration of the other 39 countries.

(14)

13

decompositions to eliminate the price effect and calculated the volume changes of emissions over the period.

In this paper, we convert the bi-country IOTs from current prices to constant prices by chaining technique as well. But instead of chaining the decomposition components, we chain the percentage change for each two consecutive years and then multiply the total percentage change for the whole period with the base year price (Oosterhaven, 2015). For example, for the percentage change between 2006 and 2007, we can divide the inputs of 2007 in year 2006’s price by the inputs of 2006 in current price. Since the growth of inputs is expressed in volume growth changes, the results can be calculated without concerning the effect of price changes. In this case, we can get the percentage change from 1995 to 2007 by multiplying the annual change during this period, the equation shows as the following:

(12)

Where

, indicating the annual growth and growth of cells from period 1995 to period 2007 respectively. Then the value of the IOT cells of the final year in prices of the base year can be calculated by multiplying the values of base year:

= * (13)

(15)

bi-14

country IO table of China with 12% bias. So we consider the cumulative errors of the chaining procedure in this paper are small enough that can be ignored and the chaining results are reliable.

For the employment data, we use another branch in the WIOD database, the called socio-economic accounts (SEAs), which contain annual data on employment, capital stocks, gross output and value added at current and constant prices from 1995 to 2009 for 35 industries, which are the same as the classification in WIOTs. But SEAs do not contain data in rest-of-world category. To be consistent, we exclude the data in the rest-of-rest-of-world category, and study the rest of the 40 countries in the WIOTs. Therefore, for both decomposition of value added and employment growth, we have 40 countries (see Appendix A) to analyze. Moreover, instead of using the data on total hours worked of employees, this paper measures the employment by using the data on total hours worked of all persons engaged, which refers to not only all paid employees, but also informal workers, such as self-employed workers and employers in informal firms. This may lead to a large difference since, in developing countries, self-employed workers are crucial for the productivity and output levels (McMillan and Rodrik, 2011), especially for industries like agriculture, low-skilled manufacturing, trade, business and personal services.

4. Empirical results

This section presents the empirical results of SDA. We first look at the changes of two chosen macro indicators, namely, value added and employment, from 1995 to 2007 in the BRIC countries by using graphs.

(16)

15

value added (123%), while the growth in the other two countries are less remarkable, with the growth rate of 43% in Brazil and 61% in Russia.

Figure 1. Growth of gross value added (%), 1995 – 2007

(17)

16

Figure 2 gives the growth of total hours worked by people engaged from 1995 to 2007. Compared to Figure 1, the differences among countries are clearly much less. Interestingly, the ranking of each country changes greatly. Brazil, which has the lowest value added growth in Figure 1, is ranked as the first among the four countries, with the employment growth of 27%. The growth rates of India and China are very similar with only 2% difference. Russia has the smallest employment growth from 1995 to 2007.

Table 1. Decomposition of real value added growth per country (%), 1995 – 2007

DD Total ∆ĉ ∆A ∆Ta ∆Tf ∆F ∆y

Brazil 27 -1 -1 -1 -1 0 31

Russia 59 -16 6 -4 -6 -6 85

India 102 2 -3 -5 -3 -3 113

China 143 -23 14 -2 -1 -4 159

FD Total ∆ĉ ∆A ∆Ta ∆Tf ∆F ∆y

Brazil 6 0 0 2 1 0 3

Russia 4 -2 1 0 -1 0 8

India 13 0 -1 -1 8 -1 7

China 39 -15 6 -1 30 4 15

Table 1 and 2 present more detailed SDA results of value added and employment growth in the BRIC countries respectively. 2 Each table consists of two parts. The first part (DD) is the decomposition results of the growth that is driven by domestic final demand. It is based on equation (10) and calculated with general decomposition equation (9). While the second part (FD) is the results of value added growth due to foreign final demand. It is based on equation (11) and calculated with general decomposition equation (9) as well.

In first part (DD) of Table 1, we can see that, due to domestic final demand, China has the most impressive value added growth (+143%) among the four countries. The other three countries have significant increases of value added as well, with the smallest growth of 27% in Brazil. As we can see from the second part (FD), the growth due to foreign final demand is much smaller than the growth that is driven by domestic consumption in all BRIC countries.

2

(18)

17

It implies that most of the value added growth in BRIC countries are still explained by the increased domestic final demand. Moreover, FD has a slightly different ranking of countries in Total. Russia, which is ranked in the third place of value added growth in DD, has the lowest growth in FD. It indicates that the value added growth of Russia, compared to the growth in other three BRIC countries, depends more on the domestic final demand than foreign final demand.

A closer look at the impact of each component in DD shows that the macroeconomic final demand growth (∆y) is the most important driving force of value added growth. It contributes positively to the value added growth in all countries, with the highest contribution of 159% in China. Unlike ∆y, the effects of coefficient changes are mixed and less significant. More specifically, the changes of value added coefficients (∆ĉ) have negative impact on the value added growth in all BRIC countries except India. It means the growth that contributed by weighted average of value added per output decreases in the past 12 years. Compared to its effect in Brazil and India, the effects of ∆ĉ are more significant in China and Russia. It indicates that, due to the increasing fragmentation of production, on average, there is a higher roundaboutness of production system in China and Russia. Together with their value added growth that contributed by technological change (∆A), which shows an opposite sign of ∆ĉ , they indicates that technology improvement, i.e. circumvent production techniques, in China and Russia has a larger contribution to their value added growth. For the changes in trade patterns (∆Ta, ∆Tf), both intermediates trade changes (∆Ta) and final products trade changes

(∆Tf) have negative effects on the value added growth in all four countries. It indicates that all BRIC countries have a trend of shift from domestic production to imports. The demand preference changes (∆F) further aggravate the negative trade impact on value added growth in each country. It means that, on average, consumers in four countries demand less from worldwide per unit of final demand. It contributes negatively to the value added growth from 1995 to 2007.

(19)

18

Table 2. Decomposition of growth in employment per country (%), 1995 – 2007

DD Total ∆ĉ ∆A ∆Ta ∆Tf ∆F ∆y

Brazil 21 -2 -3 -1 -1 -3 30

Russia 0 -48 2 -3 -7 -19 74

India 12 -41 -8 -2 -2 -22 87

China 10 -94 8 -1 -1 -16 115

FD Total ∆ĉ ∆A ∆Ta ∆Tf ∆F ∆y

Brazil 6 -2 0 4 1 0 3

Russia -3 -6 1 0 -1 0 4

India 2 -5 -1 0 5 -1 4

China 2 -27 3 0 17 1 8

Table 2 gives the growth of employment from 1995 to 2007. Consistent with Figure 2, the employment growth in both DD and FD show less differences among countries at the aggregated level. The total employment growth in Russia is noteworthy since it barely grows during the past 12 years. The reason can be found when we look at the contribution of each component in DD and FD respectively.

In DD, domestic final demand increases is still the largest contributor to the employment growth in all countries. It has the strongest impact in China (115%), which is followed by India (87%), Russia (74%) and Brazil (30%). Compared to macroeconomic final demand changes, the changes of other coefficients can explain relatively less of the increased employment in the BRIC countries. More specifically, the second strongest effect comes from the employment coefficient decreases (∆ĉ). It implies that the total hours worker by person engaged per output decreases, i.e. the labor productivity increases. China has the highest growth (-94%) in labor productivity. The growth of labor productivity in Russia is also considerable. Therefore, the positive impact from limited growth of domestic final demand cannot compensate the negative impact from large labor productivity growth, which led to little employment growth in Russia. Demand preference changes (∆F) negatively affect the employment growth in all countries as well, with the largest in India (-22%). The technology coefficient (∆A) and trade pattern changes (∆Ta, ∆Tf) are small and have a mixed

character.

(20)

19

has a shift from imports to domestic production, which contributes to the employment growth in China, India and Russia. Contributions from foreign macroeconomic final demand increases are less remarkable than the ones in DD.

Table 3. Decomposition of real value-added growth per sector, with disaggregated coefficient changes (%), 1995 – 2007

Sector Total ∆ĉ ∆A ∆Ta ∆Tf ∆F ∆yr ∆ys

(21)

20 India 217 21 9 -8 13 32 133 16 China 204 -21 25 2 13 6 169 11 Transport services Brazil 24 -22 8 1 0 5 30 3 Russia 45 -50 29 -12 -4 -7 79 9 India 189 -18 53 -1 6 8 133 8 China 446 -20 123 -4 40 63 215 29 Financial services Brazil 32 -7 3 0 0 2 32 2 Russia 214 21 38 -1 -3 21 135 3 India 187 -12 52 -12 7 15 128 10 China 188 36 -64 -2 28 14 162 14

Real estate services

(22)

21

Table 4. Decomposition of employment growth due to domestic demand per sector, with disaggregated coefficient changes (%), 1995 – 2007

Sector Total ∆ĉ ∆A ∆Ta ∆Tf ∆F ∆yr ∆ys

(23)

22 Financial services Brazil 11 -25 3 0 0 1 29 2 Russia 42 -95 30 -1 -2 15 93 2 India 22 -123 44 -10 5 10 90 6 China 36 -73 -56 -2 23 10 123 10

Real estate services

Brazil 40 0 2 1 0 2 32 2 Russia -10 -154 43 0 -1 19 81 2 India 103 -86 -22 -14 31 57 119 16 China 12 -127 -4 -10 8 13 128 5 Public admin Brazil 47 11 1 0 0 0 35 0 Russia 73 4 15 -1 -3 -44 100 2 India -31 -100 0 0 0 -12 81 0 China 23 -134 0 0 0 14 142 0 Education Brazil 52 29 0 0 0 -13 36 0 Russia -3 -1 0 0 -2 -82 82 0 India 58 -76 0 0 0 23 110 0 China 24 -160 -1 -1 1 37 147 1 Health Brazil 53 17 0 0 0 0 35 0 Russia 7 -17 0 0 0 -60 83 0 India 71 -2 1 0 0 -40 111 0 China 28 -192 10 1 2 53 154 1 Other community Brazil 46 16 -4 2 0 -4 32 3 Russia 23 32 -23 -1 -2 -75 89 1 India -25 -38 -41 7 9 -39 72 6 China 45 -160 32 -2 2 37 128 7 Private households Brazil 0 0 0 0 0 0 0 0 Russia 0 0 0 0 0 0 0 0 India 130 49 -26 -2 17 -29 110 12 China 0 0 0 0 0 0 0 0

(24)

23

to the sector classification (see Appendix B) in the 2002 Classification of Products by Activity (CPA). So the SDA results present in Table 3 and 4 are based on a more general sector classification. This paper will not interpret the results sector by sector, but choose some extreme differences to analyze.

For Agriculture, we first look at Table 3 of the decomposed value added growth. China has the highest growth of value added among the four countries in the past 12 years. Most of its value added growth is driven by the significant increases of macroeconomic final demand, especially domestic final demand (131%), whereas the strongest negative effect comes from the final demand preference changes (∆F). It indicates that the total need from agriculture sectors worldwide per unit of final demand in China decreases, i.e. comsumer in China demand less from agriculture sectors. It confirms the research of Yamakawa, Ahmed and Kelston (2009) that, with the increasing spending power, consumers in BRIC countries have less preference in lower value added products, such as agriculture. The impact of the other components are weak and mix character in China. Russia is the only country with a negative value added growth in agriculture sector. Because, as shown in Table 3, most effects of its coefficient changes are negative, which cannot be fully compensated by the limited growth of domestic final demand. In Table 4, we can see that, labor in agriculture sector decreases in all BRIC countries, except for India. In India, the impact of domestic final demand growth (84%) is large enough to compensate the shifts of consumer preference (∆F) and limited labor productivity growth in agriculture sector. China has an impressive improvement in labor productivity of agriculture sector and decreases in consumer preference, but the growth of macroeconomic final demand is large enough to retain most of the labor in this sector. In Russia, the total hours worked by people engaged in agriculture sector decreases by 22%, which is the highest among the BRIC countries. It mainly due to the remarkable improvement of technology, labor productivity and decreases of consumer preference. In Brazil, the impact from macroeconomic final demand on the employment growth becomes less important than the negative impact of ∆ĉ. It indicates that the negative growth of Brazilian employment in agriculture sector is due to its significant increases of labor productivity.

(25)

24

It can be explained by China’s increasing involvement of global value chain in manufacturing sector. The result in line with the research of Pei (2013) that, due to processing trade, import share of manufactured products shows a strong increase in China during past few years. Same as the effect of value added coefficients change, the effect of technology improvement becomes noteworthy as well, due to its increased importance of global value chain, i.e. circumvent production techniques. In Table 4, the changes of employment do not show a large difference among countries. Only Russia loses 20% of its employment in the sector of manufactured products. It is mainly due to the significant increases of labor productivity in manufacturing. Similar to Russia, China experienced a large increase of labor productivity in manufacturing sector, but the negative impact is fully compensated by the increases in other components, mostly by domestic final demand.

In Construction, followed by China, India shows the largest growth (223%) of value added in Table 3. As one of the fast growing sectors in India, its growth can be largely explained by the domestic macroeconomic final demand. It collaborates with the strategy that Indian governments incur huge expenditure on provision of infrastructure and construction for the welfare of their citizens (Mallick and Mahalik, 2010). Compared to India, Chinese domestic final demand growth lead to a higher growth valued added, but the value added per output decreases much more in China than in India, i.e. the import share of construction sectors in China increases more than in India. In Table 4, the employment growth in India is impressive and it is twice as much of the growth in China. According to Mallick and Mahalik (2010), the increasing employment in construction sector might be impacting its growth rate significantly in India and thereby, increasing the aggregate output in the economy. Russia also has a large decrease of employment coefficient effect, which exceed the positive impact of macroeconomic final demand growth. Therefore, Russian employment decreases slightly during the past 12 years.

(26)

25

remarkable labor productivity growth. Though India has the largest increases in labor productivity in financial services, the combined positive contribution from other components is not so impressive.

In Real estate service, Table 3 gives that India has the largest value added growth of 277% among all BRIC countries. Besides the strong impact from macroeconomic final demand of domestic consumers, the foreign macroeconomic final demand of real estate service in India also plays an important role in the value added growth. Contributions of final product trade and consumer preference changes are important factors of value added growth as well. Employment of real estate service in India, as shown in Table 4, increases from 1995 to 2007. The contributions of the seven components have similar pattern with Table 3. The employment share in Russia, again, decreases as shown in Table 4. Because the effect of final demand growth over the years is not large enough to offset the impact of labor productivity increases and keep the labor stay in the real estate service sector.

5. Conclusion

In this paper, we have decomposed and compared the growth of value added and employment in the BRIC countries from 1995 to 2007. Our analysis shows that, in general, all BRIC countries have experienced a distinctive development in value added. Employment growth in BRIC countries is also remarkable, expect for Russia.

(27)

26

At sectoral level, we further decompose the changes of macroeconomic final demand into two parts, i.e. the macroeconomic final demand of domestic and of foreign market. We find that the extent to which driving force contribute to the economic growth vary among sectors and countries. Interestingly, the growth of macroeconomic final demand is not the only determinant of value added and employment growth anymore. Impacts from other components become more significant than their impacts at macro level. For instance, in China, all sectors display a remarkable negative impact that comes from employment coefficient changes. This collaborates the findings of Bosworth and Collinss (2008) that the labor productivity, especially in manufacturing sector, shows a significant increase in China. Moreover, trade effects are found to be the important contributors too, especially for China and India.

(28)

27

References

Ark, B. V., Chen, V., Coijn, B., Jaeger, K., Overmeer, W., & Timmer, M. (2013). Recent

Changes in Europe's Competitive Landscape: How the Sources of Demand and Supply Are Shaping Up. Groningen: University of Groningen, Groningen Growth and Development

Centre.

Ark, B. V., (1995). Sectoral Growth Accounting and Structural Change in Postwar Europe. Groningen: University of Groningen, Faculty of Economics, Groningen Growth and Development Centre.

Bassaini, A., & Scarpetta, S. (2003). The Driving Forces of Economic Growth Panel Data Evidence for the OECD Countries. OECD Economic Studies. 2001, 9-56.

Chenery, HB Robinson, S., & Syrquin, M. (1986). Industrialization and growth: a

comparative study. New York, Oxford University Press.

Ciarli, T., Lorentz, A., Savona, M., & Valente, M. (2010). The Effect of Consumption and Production Structure on Growth and Distribution. A Micro to Macro Model.

Metroeconomica. 61, 180-218.

Connolly, R., & Hanson, P. (2012). Russia’s Accession to the World Trade Organization.

Eurasian Geography and Economics. 53, 479-501.

De Haan, M. (2001). A Structural Decomposition Analysis of Pollution in the Netherlands. Economic Systems Research. 13, 181-196.

De Vries, G. J., (2011). Deconstructing the BRICs: structural transformation and aggregate

productivity growth. Groningen: University of Groningen, Groningen Growth and

Development Centre.

Dietzenbacher, E., De Groot, OJ, & Los, B. (2007). Consumption Growth of Accounting. Review of Income and Wealth. 53, 422-439.

(29)

28

Dietzenbacher, E., Los, B., Steher, R., Timmer, M., & De Vries, G.J. (2013). The Construction of World Input-Output Tables in the WIOD Project. Economic Systems

Research. 25, 71-98.

Dietzenbacher, E., & Temurshoev, U. (2012). Input-output Impact Analysis in Current or Constant prices: Does it Matter? Journal of Economic Structures.1,1: 4.

Feldman, S. J., Mcclain, D., & Palmer, K. (1987). Sources of Structural Change in the United States, 1963-78: An Input-Output Perspective. The Review of Economics and Statistics. 69, 503-510.

Jorgenson, D. W., & Timmer, M. P. (2010). Structural change in advanced nations: a new

set of stylized facts. Groningen: Groningen Growth and Development Centre.

Lahr, M. L., & Dietzenbacher, E. (2001). Input-Output Analysis: Frontiers and Extensions. Palgrave, London

Leontief, W. (1966). Input-output economics . New York, Oxford University Press.

Linden, J. A. van der., & Dietzenbacher, E. (1995). The determinants of structural change in

the European Union: a new application of RAS, Groningen: University of Groningen.

Los B., Timmer M.P., & De Vries G.J. (2014). How Important are Exports for Job Growth in China? A Demand Side Analysis. Journal of Comparative Economics.

Mallick, H., & Mahalik, M. K. (2010). Constructing the Economy: The Role of Construction Sector in India’s Growth. Journal Of Real Estate Finance & Economics, 40(3), 368-384. Maurya, K.N., & Vaishampayan, J. V., (2012). Growth and Structural Changes in India’s Industrial Sector. International Journal of Engineering.6, 321-331.

McMillan, M. S., & Rodrik, D. (2011). Globalization, Structural Change and Productivity

Growth. Cambridge, Mass, National Bureau of Economic Research.

Miller, R. E., & Blair, P. D. (1985). Input-output Analysis: Foundations and

(30)

29

Moschini, G., & Moro, D. (1996). Structural Change and Demand Analysis: A Cursory Review. European Review of Agricultural Economics. 23, 239-261.

O’Neill, J. (2001). Building better global economic BRICs, Goldman Sachs Global Economics Paper No. 66.

Oosterhaven, J. (2015). A note on the constant price-chaining of input-output tables, Mimeo, University of Groningen.

Oosterhaven, J., & Hoen, A. R. (1998). Preferences, Technology, Trade and Real Income Changes in the European Union. The Annals of Regional Science.32, 505-524.

Oosterhaven, J., & Linden, J. A. van der (1997). European Technology, Trade and Income Changes for 1975–85: An Intercountry Input–Output Decomposition. Economic Systems

Research. 9, 393-412.

Pei, J., (2013). Trade, Growth, Regions, and the Environment: Input-Output Analyses of the Chinese Economy, Groningen: University of Groningen.

Robert Stehrer & Gaaitzen J. de Vries & Los, Bart & Dietzenbacher, H.W.A. & Timmer, Marcel, (2014). The World Input-Output Database: Content, Concepts and Applications, GGDC Research Memorandum GD-144, Groningen Growth and Development Centre, University of Groningen.

Santos-Paulino, A.U. (2005). Trade Liberalization and Economic Performance: Theory and Evidence for Developing Countries. World Economy 28: 783-821.

Statistics Canada. (2001). A guide to deflating the input-output accounts sources and

methods . Ottawa, Statistics Canada.

Timmer, M. P., Erumban, A. A., Los, B., Stehere, R., & De Vries, G. J. (2014). Slicing Up Global Value Chains†. Journal of Economic Perspectives. 28, 99-118.

Xu, Y., & Dietzenbacher, E. (2014). A structural decomposition analysis of the emissions embodied in trade. Ecological Economics, 101, 10-20.

(31)

30

Appendix A. List of countries in WIOD-database European Union North America Latin America Asia and Pacific

Austria Germany Netherlands Canada Brazil China

Belgium Greece Poland United States Mexico India

Bulgaria Hungary Portugal Japan

Cyprus Ireland Romania South Korea

Czech Republic Italy Slovak Republic Australia

Denmark Latvia Slovenia Taiwan

Estonia Lithuania Spain Turkey

Finland Luxembourg Sweden Indonesia

(32)

31

Appendix B. Specification of Sector ISIC

rev.3 code

Industry Name Sector Labels Sector

Classification AtB Agriculture, hunting, forestry and fishing Agriculture Agriculture

C Mining and quarrying Mining Mining

D Manufactured products

Manufacturing

15t16 Food, beverages and tobacco Food

17t18 Textiles and textile products Textiles

19 Leather, leather products and footwear Leather

20 Wood and products of wood and cork Wood

21t22 Pulp, paper, printing and publishing Pulp

23 Coke, refined petroleum and nuclear fuel Coke

24 Chemicals and chemical products Chemicals

25 Rubber and plastics Rubber

26 Other non-metallic mineral Other Non-Metallic

Mineral

27t28 Basic metals and fabricated metal Basic Metals

29 Machinery, not elsewhere classified Machinery

30t33 Electrical and optical equipment Electrical

34t35 Transport equipment Transport

36t37 Manufacturing, not elsewhere classified; recycling Manufacturing

E Electricity, gas and water supply Electricity Electricity

F Construction Construction Construction

G Wholesale and retail trade services; repair services pf motor vehicles, motorcycles and personal and household goods

Wholesale services

50 Sale and repair of motor vehicles and motorcycles; retail

sale of fuel Sale

51 Wholesale trade, except of motor vehicles and motorcycles Wholesale 52 Retail trade and repair, except of motor vehicles and

motorcycles; Retail

H Hotels and restaurants Hotels Hotels services

I Transport, storage and communication services

Transport services

60 Inland transport Inland Transport

61 Water transport Water Transport

62 Air transport Air Transport

63 Other supporting transport activities Other Supporting

64 Post and telecommunications Post

J Financial intermediation Financial Financial

services K Real estate, renting and business services

Real estate services

70 Real estate activities Real Estate

71t74 Renting of machinery & equipment and other business

activities Renting

L Public administration and defense; compulsory social

security Public Admin Public services

M Education Education Education

services

N Health and social work Health Social services

O Other community, social and personal services Other Community Personal services P Private households with employed persons Private Households Households

(33)