Master Thesis
The Growth of the BRIC Countries, their Impact on Incomes and
Jobs in Europe, Japan, and the US, and Evidence for Decoupling:
A Global Value Chain Perspective
Timon Bohn
July 5, 2013
Student number: S2504030 (Groningen)
Email: timonbohn@gmx.de
Supervisor: Prof. Dr. Marcel P. Timmer
Co-Assessor: Dr. Felicitas Nowak-Lehmann Danzinger
Degrees: Master of Science, International Economics and Business (Groningen) Master of Arts, International Economics (Göttingen)
Abstract
I assess the extent to which final demand in the BRICs has impacted incomes and jobs in Europe, Japan, and the United States (the Triad) between 1995 and 2007, and analyze whether the generation of BRIC final demand is decoupling itself from the Triad. Using a trade in value-added methodology and an input-output decomposition technique, I find that BRIC final demand induced increasing amounts of value added and jobs in the Triad as a share of Triad GDP and Triad workforce, and that services activities induced are growing in importance. The data do not support the hypothesis of decoupling, especially not for value added of high-skill content, suggesting developed countries specialize in high-skill value chain activities in their trade relations with the BRICs.
i
Table of Contents
1. Introduction ... 1
2. Review of Literature and Economic Issues... 3
2.1 North-South trade ... 3
2.2 Export statistics ... 6
2.3 Value added approach ... 10
3. Hypotheses ... 12
4. Analytical framework and data sources ... 16
4.1 Technical exposition ... 17
4.2 Application to my hypotheses ... 20
4.3 Data from the World Input-Output Database ... 24
5. The impact of final demand in the BRICs: empirical results ... 25
5.1 Exports of value added and jobs induced by BRIC demand ... 25
5.2 Sectoral decomposition of value added and jobs induced by BRIC demand ... 30
5.3 Worldwide value added embodied in BRIC demand ... 33
5.4 Worldwide compensation of labor and capital embodied in BRIC demand ... 37
6. Discussion and Conclusion ... 40
7. References ... 44
1
1. Introduction
This thesis investigates the extent to which final demand in the BRIC countries (Brazil, Russia, India, and China) induce incomes and employment in advanced countries. Final demand consists of household consumption, government expenditures, and investment. I also examine whether the generation of this final demand is becoming more or less dependent on incomes and jobs induced in advanced countries over time. The topic is especially relevant to developed countries because many of them have stagnating domestic markets and are becoming increasingly dependent on foreign demand for their production expansion. In this analysis major developed countries will be represented by the Triad, which is defined as the EU 15, Japan, and the United States. The term was coined by Kenichi Ohmae in 1985 to refer to the leading developed countries of the world (Glänzel et al., 2008). These countries still accounted for about 60% of world economic output in 2007 (WIOD database).
The rapid economic growth of the BRICs, driven in part by structural transformation and an expanding middle class within the BRICs, is transforming worldwide trade patterns and provides potentially lucrative new markets. Goldman Sachs first introduced the BRIC acronym in 2001 and now estimates that the BRICs could become as big as the G7 as early as 2032 (O’Neill 2001, 2009). In contrast to the stagnating economic weight of the Triad economies, the BRIC economies are growing fast, and from 1995 to 2007 their share in the world economy almost doubled from 7% to 13% (WIOD database). In addition, as documented by a recent report by Unido (2012), the BRIC share in worldwide trade is rising quickly. Hence the BRICs are becoming important drivers of global demand, clearly inducing more and more value added production, factor incomes, and labor worldwide. These developments offer significant opportunities but also pose new challenges for developed countries who seek to take advantage of the growing BRIC markets, such as the potential consequences for factor income distributions.
This empirical study makes three contributions to extend previous research. First, my analysis is one of the first to examine the impact of the BRICs on major developed countries using the trade in value-added methodology (Johnson and Noguera, 2012).1 This approach traces the flows in global production and division of income in global value chains using inter-country input-output tables. In a time of rapid production fragmentation, a global value chain framework improves upon international trade statistics because it captures the actual
1
2
value added in exports of a source country embodied in final demand of a destination country, in my case in BRIC final demand. Second, I use the recently released World-Input-Output Database (WIOD) for the period 1995-2007 which is accompanied by unique information on employment and factor income distributions. These datasets provide insights into how much different types of incomes and job types have been induced by foreign demand, providing a perspective that has been much less studied (Los et al., 2012). Specifically this allows me to break down value added at a sectoral level and to distinguish between capital and labor compensation for workers characterized by their skill level. This more accurately depicts the sectors and workers earning rising income shares through trade with the BRICs, which is the focus of the first part of this thesis. In contrast, public perception and even some prominent economists tend to focus on those sectors and jobs at risk from low-wage competition abroad (Lawrence and Edwards, 2013). Third, I explore whether there are signs of a decoupling of value added production induced by the BRICs from advanced countries. This has been a popular yet controversial topic in the press, and evidence for decoupling will be investigated in the second part of this thesis. Arguments for decoupling take into account the sharp rise in South-South trade, the increasing participation of developing countries in global production networks, and observations that the BRICs may be upgrading factor endowments (Unido 2012). Such indications have been well documented, especially in research examining export sophistication (Hanson 2012), and could support the idea that the BRICs are becoming less dependent on developed countries. Interest spiked with the global financial crisis of 2008, though my emphasis will be on decoupling in the years already leading up to the crisis.
3
The thesis is organized as follows. In Section 2 I summarize the literature and theoretical arguments on the impact of the South on welfare in the North. I analyze the impact of the BRICs on major developed countries using trade statistics and explain how a value added perspective provides a more accurate assessment of the impact of the BRICs on the Triad than an export-based approach. This helps me to develop hypotheses that are stated in Section 3 and will be tested using an input-output analysis. Section 4 details the methodological framework and data sources. In Section 5, I present and discuss my results. Conclusions are in Section 6. An appendix provides additional data.
2. Review of Literature and Economic Issues 2.1 North-South trade
In principle rising final demand of the BRICs should benefit developed countries insofar that it gives exporters growing markets and creates the potential for trade-induced incomes and jobs. However, trade liberalization that is required for the realization of such opportunities also exposes producers in developed countries to import competition. The emergence of the BRICs therefore may entail a reallocation of labor in developed countries that benefits certain groups and harms others. This impact is strengthened by rapid changes occurring within the BRICs, such as structural transformation, changes in factor endowments, improved production capabilities, and possibly more sophisticated consumption patterns over time (Unido 2012). Consequently, policymakers in developed countries must balance short- and medium term adjustment costs associated with an increase of trade with the BRICs against possible long-term welfare benefits implied by trade theories (Stracca 2011). These differentiating effects and political sensitivities have resulted in many heated debates in the press. This has also captured the attention of economists who have used various approaches to measure the impact of the BRICs. Therefore I begin by outlining the key predictions of North-South trade theories. In the sections that follow I discuss the insights of export statistics and lastly the benefits of value added approaches to measuring the impact of growth in the BRICs.
4
relative abundance of labor and land). Both theories imply that if advanced countries focus on capital intensive tasks in line with their comparative advantage, declines in labor-intensive industries and unskilled workers (which are relatively scarcer) will be more than offset by lower consumer and input prices, higher variety and quality of goods, and income earned in high-skill and technology industries (Lawrence and Edwards 2013). Thus, trade should be a win-win situation for all.
Many equilibrium models of North-South trade hence assume that developed countries concentrate on R&D and high-quality manufacturing while developing countries focus on labor-intensive manufacturing and primary products (Flam and Helpman 1987; Dinopoulos and Segerstrom 2007). However, not all models find that greater trade with a growing South will necessarily positively impact welfare in the North. Dinopoulos and Segerstrom (2007) develop a dynamic equilibrium model of North-South trade and determine that as trade with the South increases, the overall effect on long-run economic growth in the North is theoretically ambiguous and tends to approximate zero. The authors argue this is because the benefits of lower consumer prices are roughly balanced by the costs of reallocation from production to R&D and a permanent reduction in relative wages of northern workers. That contrasts with a number of previous models cited by the authors that imply that an increasing South will always be beneficial to long-run economic growth in the North.
In one of the few empirical studies that systematically examined the rise of the BRICs (2000 to 2011) and their impact on welfare in advanced countries, Stracca (2011) uses a panel data technique and determines there was no significant impact on per capita income growth in the North. Interestingly the author does find that advanced countries more dependent on manufacturing (including medium and high-tech) lost growth while countries more dependent on services gained growth during that time period, and that a focus on innovation in the North also did not exert a positive effect. That contrasts with the traditional view which assumes the North should specialize in capital-intensive manufacturing. The findings are similar to those by Dinopoulos and Segerstrom and of prominent economists in that they cast further doubt on the idea of an unambiguously positive impact of emerging countries in the South (like the BRICs).2 As Stracca points out, this would also be consistent with the observation that BRIC growth rates have remained high in the past decade while growth rates in developed countries have remained stagnant at a much lower level. My study differs from these studies in that the emphasis is on a more detailed analysis of actual jobs and value-added exports affected. It
2
5
takes the aforementioned theories and insights on the impact of the South (BRICs) on p.c. GDP in developed countries more or less as the starting point.
Even though many trade models and studies suggest that trade with a growing South has at worst a neutral impact on economic growth in the North, most empirical work that does focus their attention on a sectoral level tend to emphasize the losers. For instance, a recent study by Autor et al. (2012) finds a drop in manufacturing employment as exposure to imports from China rises. Further, the authors reason that due to labor market imperfections, low-skilled workers were unable to migrate to other labor markets or find employment in jobs that require a higher skill-type. A report compiled by the Conference Board of Canada (Rao, 2008) suggests that BRIC countries could threaten jobs in developed countries, but mostly for industries that depend on low-skilled labor. Thus it recommends that Canada focus on upgrading its value chain activities by emphasizing knowledge-intensive, high value added products to exploit the potential of the BRIC markets. Moreover, Blinder (2007) rank occupations according to their off-shorability and find that 29% of all US jobs are potentially off-shorable - mostly to less developed countries - within the next few decades. Surprisingly the author finds little correlation between the skill level of workers and the likelihood of off-shorability, and that high-value services can just as easily be off-shored as manufacturing (at least those that can be impersonally, e.g. electronically delivered). This casts some doubt on whether the suggestion by politicians and governmental agencies to minimize the impact of off-shoring simply by upgrading worker skills is by itself a sufficient policy recommendation.
However, other studies find that concerns about major job losses and off-shoring are exaggerated and the negativity in public perception is largely unwarranted. Lawrence and Edwards (2013), taking a more positive view, note that the scale of actual off-shoring of US jobs was rather modest and find that only some particular workers are adversely affected by trade with emerging economies. The authors conclude that North-South trade patterns are more complementary rather than competitive and that faster growth in emerging countries could rather reduce unemployment problems in the US. Along these lines, a study by Dreger and Zhang (2011) that investigates the impact of Chinese demand shocks on advanced countries using a GVAR model also discover a positive impact on output growth due to higher export opportunities. And as I will show later, the value added literature provides support that declines in manufacturing employment may be compensated by increases in supporting services (e.g. Timmer et al., 2013).
6
especially involving the relocation of labor-intensive production activities. In contrast, my study follows a more positive view as the focus is on sectors and workers benefitted by additional value-added exports to the BRICs. That is, the emphasis is on the overall (net) amount of jobs and value added, i.e. counting both gross job creation and job destruction, which is a perspective that has been much less studied. Such an analysis accepts the possibility that rising trade with the BRICs could lead to growing labor market inequalities for certain workers. For example, if growth in jobs and incomes generated for high- and medium- skilled workers is greater than the growth for low-skilled workers (and the same holds true for workers in certain sectors or subsectors), uneven and asymmetric employment effects could have distributional consequences if labor markets are not fully liberalized (Timmer et al., 2013). Labor market liberalization is one of the key factors in determining the extent to which developed countries are helped or harmed by greater trade with the South in the short run (Arnold 2002). Hence the real challenge is for policymakers to facilitate the adaption.
2.2 Export statistics
The method traditionally used to measure trade flows and industrial upgrading (proxied by export sophistication) involves gross export statistics. In the following I analyze the impact of trade with the BRICs from 1995-2007, the time period before the onset of the world economic crisis, using export statistics that I have derived from the WIOD database. By examining trade statistics I also shed light on some possible shortcomings of using an export-based approach. These weaknesses will be elaborated upon in the section that follows.
In 1995, 11% of all Triad exports (exports of the EU 15, Japan, and US) sent outside of the Triad went to the BRICs and this rose to 16% of all Triad exports in 2007. From the perspective of the BRICs, among all 41 countries included in the WIOD database (which together make up roughly 85% of world GDP), the EU 15 (24%), Japan (10%) and the US (9%) were the three trade partners that comprised the highest shares of BRIC imports in 2007. They were also the three largest markets for BRIC exporters (26% of BRIC exports went to the EU 15, 20% to the US and 7% to Japan, not including intra-BRIC trade).3 This shows that developed countries were still significant bilateral BRIC trade partners in 2007. However, in contrast to the growing dependence of the Triad on exports to the BRICs, the import shares of the Triad in all BRIC imports have declined since 1995. In 1995, the EU 15 (31%), Japan (12%), and US (11%) made up even larger shares of BRIC imports, and the BRICs also
3
7
depended relatively more on the Triad for their exports. Thus, while the Triad has become relatively more dependent on the BRICs for their exports over time, the BRICs have become relatively less dependent on the Triad for their imports.
Tables 1 and 2 show exports disaggregated by broad sector. The Triad’s exports to the BRICs were predominantly manufacturing (80%) and services (18%), while agriculture played an insignificant role (1%) in 2007. Disaggregated for the Triad, manufacturers made up 76% of exports from the EU 15 to BRIC countries, 77% of US exports to the BRICs, and 96% of exports from Japan to the BRICs in 2007. Generally, services exports become somewhat more important and manufacturing exports somewhat less important in 2007 than in 1995 (except for Japan). Russia imported the highest share of manufacturing from the Triad (91%), followed by China (81%), India (75%), and lastly Brazil (67%). The reverse trade flow shows that the BRICs were also very dependent on manufacturing in their exports to the Triad. 71% of all BRIC exports to the Triad were manufacturing, 27% services and 2% agriculture. Manufacturing was more important in BRIC exports to the US and Japan (with shares of 87% and 82% respectively) than in exports to the EU 15 (with a 57% share). However, it is important to keep in mind that export statistics only insufficiently capture trade in services because they do not account for domestic inter-linkages between services and manufacturing. Later in my analysis I will compare these sectoral export shares with sectoral shares of value-added exports that take this into consideration.
Table 1: Export composition from the EU 15, Japan, and US to the BRICs as (%) share
Source: Derived from the World Input-Output Database
1995 1.9 78.9 19.3 2007 0.9 75.5 23.6 1995 0.1 89.6 10.2 2007 0.1 95.5 4.4 1995 6.2 84.9 8.9 2007 3.6 77.4 19.0 1995 2.3 82.5 15.2 2007 1.3 80.5 18.3 Japan EU 15 USA Triad
8
Table 2: Export composition from the Triad to each BRIC as (%) share
Source: Derived from the World Input-Output Database
The differing sectoral composition of trade between each BRIC and the Triad is also a reflection of the great heterogeneity of the BRIC countries themselves, and the BRIC countries are experiencing growth and structural transformation in different ways. China’s and India’s labor composition is transitioning from agriculture to both manufacturing and services, while Russia’s and Brazil’s labor is moving mostly to services (de Vries et al., 2011). China is also the most reliant on manufacturing exports, while India and especially Brazil still depend relatively more on primary goods exports than the others. In addition, economic growth in China has been by far the most rapid in the past decade and has easily become the Triad’s most important BRIC partner. According to the WIOD, 44% of all exports from Triad countries to the BRICs went to China in 1995; by 2007 this share grew to 59%. Thus the composition of value-added exports from the Triad to the BRICs and impact of trade also likely depends on the specific BRIC country. Since China carries such a large weight, an aggregate analysis could mask the impact of the other BRICs. Though I will not go into greater detail here on BRIC heterogeneity, I will also disaggregate the individual impact of each BRIC in my analysis of value added to take possible variations into account.
Lastly I examine trade flows between the BRICs, the Triad and the rest of the world. This is to track growth in trade over time and to see whether so-called South-South trade is becoming more important relative to North-South trade. In my case North-South trade is defined as trade between the Triad and the BRICs. South-South trade is either defined as trade between BRIC countries themselves or as trade between the BRICs and non-BRIC developing countries. According to Goldman Sachs’ Jim O’Neill, “Three of the four [BRIC] countries have a BRIC counterpart as one of their top trading partners” (Keeler 2012, Global Finance magazine). This trading partner is China which has come to dominate trade between the
1995 1.1 78.1 20.8 2007 0.7 67.4 31.9 1995 2.3 86.0 11.7 2007 1.4 90.9 7.7 1995 0.5 68.3 31.1 2007 0.6 75.4 23.9 1995 3.5 87.5 9.0 2007 1.5 80.6 17.9
Year Agriculture Manufacturing Services Brazil
Russia
India
9
BRICs. And, as Hanson (2012) finds, growth in trade shares for poorer countries far exceeds the increase in their relative size (GDP) and South-South trade is increasing rapidly. If the BRICs become less dependent on developed countries over time (as could already be implied by import shares) this would indicate that BRIC final demand is decoupling from the Triad.
Exports from the Triad to the BRICs, also derived from the WIOD database, increased by about 200% from 1995-2007, and bilateral trade between the two regions increased by about 240%.4 In contrast, bilateral trade between the EU 15, US and Japan, so-called North-North trade (Hanson 2012), grew by only about 25%. This reflects the enormous amount of growth potential that the BRICs offer as an export market, at least according to export figures and compared to trade between developed countries. However, the WIOD trade statistics confirm the increasing importance of both intra-BRIC trade and South-South trade compared to North-South trade. Trade between BRIC countries increased by more than 800% from 1995-2007, with growth most rapid in the last period of the time frame. The same pattern is observed for exports from developing countries to the BRICs. Exports by developing countries, defined as all countries in the world except for the Triad, Australia, Canada, Taiwan and Korea, to the BRICs increased by about 400% from 1995-2007, and bilateral trade increased by about 460%. That clearly demonstrates that South-South trade (and especially BRIC-BRIC trade) is increasing at a much faster rate than trade between the Triad and the BRICs. Thus, it could be that developing countries are in a better position to take advantage of the opportunities that the BRICs present.
However, one must use some caution in interpreting the results as higher South-South trade figures could also reflect fragmentation and specialization in tasks, not just higher consumption and trade liberalization (Baldwin 2006; Hummels et al., 2001). Trade statistics also give support to qualifying these observations because relative increases in trade flows vastly exceeded relative increases in GDP of the respective trading partners, confirming Hanson’s (2012) finding. This was especially the case for intra-BRIC, as the combined GDP of the BRICs increased by only 155% from 1995-2007, and for trade between BRICs and developing countries, whose combined GDP only doubled. This exposes one of the weaknesses of export statistics because it is very questionable that an increase in trade flows can be interpreted as a corresponding increase in the impact of trade in terms of incomes and jobs generated in the source country. An analysis based on value added contributions of exports improves upon export statistics because it takes fragmentation into account. This will be discussed in more detail in the next section. In addition, richer countries tend to sell
4
10
products with higher unit values (Hummels and Klenow, 2005) implying that the average skill composition of products exported by the North (and thus plausibly also income earned per worker) should be higher than products exported by developing countries. The skill content of trade is another aspect I will analyze using value added methodology which will shed more insights on the type of activities countries are specializing in (in my case specifically related to value added trade with the BRICs).
2.3 Value added approach
Gross exports statistics are popular because data is widely and easily available. However, in times of internationally fragmented production, fostered by technological progress and reductions in trade barriers and communication costs (Baldwin 2006), conventional trade statistics relied on by reports and previous studies have a number of flaws. As the OECD secretary general summarizes: “there can be ‘multiple counting’….some exports may only contribute marginally to value-added if account is taken of imported intermediate inputs, while the significance of some imports in domestic firm productivity may not be evident.” (Gurria, OECD 2012).
Production networks, which have long decoupled from location of consumption, are now also becoming more fragmented geographically. One consequence of this spatial unbundling of production is that countries are increasingly competing with each other worldwide at the level of activities and stages rather than sectors (Timmer et al., 2013). In addition, as already implied by my discussion in the previous section and the quote by the OECD secretary general, gross exports may overstate net trade flows and also the domestic value added embodied in a country’s exports (Hummels et al., 2001). Finally, gross exports often do a poor job of differentiating between services and manufacturing because only the exported (manufacturing) good is counted. Many domestic services as well as agricultural products contribute to the production of manufacturing goods as intermediate inputs due to potentially significant domestic inter-industry linkages (Los et al., 2012). Therefore services are generally underrepresented in gross exports. In sum, gross export statistics can lead to imprecise, misleading interpretations and misplaced policies. Thus they are also an inadequate indicator to assess the impact of the BRICs on incomes and jobs in developed countries.
11
studies and shows that despite the increasing sophistication of exports by developing countries like China, advanced countries still capture substantial value in global value chains. For example, a famous case study of Apple’s iPod finds that US retailers and Apple designers absorb half of the retail price of €299 (Dedrick et al., 2010). Even though assembly has been off-shored to China and the iPods are credited as Chinese exports in official statistics, China captures only €15 of the total value because it contributes low value added activities and imports advanced intermediaries from Japan, Korea, and elsewhere. This shows how the technology content of a country’s exports can be misleading. Another study found a similar pattern for the Chinese export sector as a whole because many exports rely heavily on imported intermediates (Koopman et al. 2008). Other literature measures vertical linkages in global trade directly, the so-called vertical specialization. This measure takes into account the interconnectedness of production processes and captures empirically the imported input content of a country’s exports (Hummels et al., 2001). Hummels, Ishii and Yi, who use input-output tables, find that vertical specialization is increasing over time and explained 30% of the growth in exports by 14 OECD and emerging market countries between 1970 and 1990. Johnson and Noguera (2012) generalize their approach and compute the domestic content of exports in a given source country that is absorbed in the final demand of each destination country, what they call value added- (VA) exports. The authors observe that value added to export ratios (or VAX) are higher for services and agricultural contributions than for manufacturing across a large set of countries because they are often utilized as intermediate inputs in the production process. A study by Groot et al. (2011) on the impact of the BRICs on one of the Triad members, the Netherlands, found using an input-output analysis that value added trade with the BRICs, including indirect exports to the BRICs and re-exports from the BRICs, made up nearly 2% of GDP and generated nearly 100 thousand jobs. However, the authors do not make a time-series analysis, nor do they investigate incomes and jobs by skill type or sector.
12
significant difference. Thus, the Netherlands depends much less on foreign demand for their income than gross export statistics indicate. Methodologically my paper is most aligned with Los et al. (2012), which uses the same database and examines the impact of foreign demand on exports, incomes, and jobs in one of the BRICs (China). I follow the approach by Los et al. (2012), thereby extending the approach by Johnson and Noguera, because I differentiate the remuneration of labor and capital in value added and decompose value added and employment into the contributions of different sectors and production factors. In section 4 I explain the methodology used for these calculations in more detail.
3. Hypotheses
From the preceding discussion I derive the following four hypotheses. The first two hypotheses pertain to the impact of the BRICs on incomes and jobs in major developed countries, and the final two hypotheses relate to evidence for the decoupling of BRIC final demand from the Triad. Note that while I discuss developed countries in a more general sense, the hypotheses will always define developed countries as the Triad (the EU 15, Japan, and United States).
First, as the economies of the BRICs grow and fragmentation in trade increases, much evidence suggests that new opportunities are created for exporting and value-adding industries in the rest of the world. I expect that developed countries also have accrued some of the benefits of greater trade opportunities, especially at a time when export markets elsewhere (in particular to other developed countries) are stagnating. Los et al. (2012) estimate that total exports of value added, taken as a share of gross domestic product (GDP), grew from 1995-2009 in a number of large developed countries, including Japan, Germany, and the US. Furthermore, the authors find that amongst all 40 countries in the WIOD database, the share of value added exports ultimately absorbed by emerging economies is increasing relative to value added exports ultimately absorbed by mature economies. Therefore I expect that the contribution of value added exports induced by BRIC final demand, taken as a share of Triad GDP, and the jobs generated by those exports, taken as a share of the Triad workforce, are growing for developed countries, even if the impact is not as large as export statistics suggest.
13
Exports of value added to BRIC countries and employment generation as calculated to test Hypothesis 1 can be decomposed by sector of origin. In the next hypothesis I consider the sectoral composition of value-added exports from the Triad to the BRICs and the share of jobs induced by broad sector.5 I expect that as fragmentation of trade increases, developed countries will specialize more heavily in service-related activities than in manufacturing in their trade relations with the BRICs. This is especially likely for pre- and post- production activities, e.g. design, R&D, and retail, as exemplified by the Apple iPod case study mentioned earlier. Exporters and workers engaged in manufacturing industries may not benefit from increased incomes and jobs at the same rate. These are tasks the BRIC countries may have a comparative advantage in, such as assembly work in China or agriculture in Brazil and India. That could additionally benefit developed countries because service-related activities in advanced segments of global value chains offer potentially higher incomes.
This prediction contrasts sharply with the high and fairly stable manufacturing shares in gross exports outlined in section 2.2. However, it is important to look beyond manufacturing because the production of manufacturing goods often requires contributions of domestic services and agricultural activities. There can be significant domestic inter-industry linkages due to the network-like nature of production processes also within countries that are not captured by gross export statistics (Los et al., 2012). This is especially the case for higher-value and intermediary service-related activities typical in global higher-value chains, such as wholesaling, transportation, and finance (Timmer et al., 2013). I explicitly account for these indirect activities by using an input-output methodology. The 13 year time frame will also enable me to track specialization patterns over time because structural transformation is also causing the services sector to become increasingly important for income and jobs worldwide (Imbs and Wacziarg, 2004), including for value-added exports (Johnson and Noguera, 2012). Accordingly, Hypothesis 2 examines the extent to which these worldwide patterns hold true for Triad value-added exports to the BRICs and the tracks the pace of the changes.
HYPOTHESIS 2: When breaking down the value added and jobs that were induced by BRIC final demand from 1995-2007 in the Triad, the shares related to services activities are increasing faster than the shares related to manufacturing activities over time.
5
14
Hypotheses 3 and 4 stay with BRIC final demand, but rather than limiting the analysis to value added contributions in developed countries that satisfy this final demand as before (i.e. Hypotheses 1 and 2), I shift the emphasis to decoupling and thereby consider value added contributions in the rest of the world (ROW) and in the BRIC countries themselves. Throughout this analysis ROW is defined as all countries in the world except for the Triad and BRIC countries. This perspective is insightful because it explores whether the value added being generated by BRIC final demand is becoming more or less reliant on the Triad over time. If the generation of BRIC final demand is gradually detaching itself from the Triad, this would provide evidence for decoupling.
The decoupling hypothesis has received much attention in the press and suggests that developing countries (especially the BRICs) are becoming less dependent on developments in the North and more interdependent on each other. The idea enjoyed a surge of interest at the onset of the financial crisis as BRIC countries maintained high growth rates despite stagnation in the developed world, though the idea has long been controversial (Hanson 2012; Roubini, Forbes, 2009). In addition, policymakers in the BRICs have in recent years actively endorsed closer intra-BRIC ties and hope to lower dependence on the West, as exemplified by the discussion at recent BRIC summits about the creation of a BRIC development bank (England, Financial Times, 2013). My analysis differs from these very recent developments because it specifically examines evidence for decoupling already in the years leading up to the financial crisis of 2008.6
Arguments in favor of decoupling include observations that many emerging and developing countries are integrating themselves in the world economy and increasing their global value chain participation in recent years (Hanson 2012). Global trends facilitating this include the “Second Unbundling,” or the slicing up of value chain activities at the level of tasks within firms rather than at the level of firms and sectors (Baldwin 2006). Factors driving these developments include globalization, the ICT revolution that has lowered trade and coordination costs, structural transformation (especially in developing countries), and trade liberalization, all of which have made it easier than ever for less developed countries to join global value chains (Baldwin 2011). As the BRICs become wealthier and upgrade their value chain activities, they may begin to substitute imports from developed countries with their own domestic production (using domestic factor endowments) and imports from other emerging and developing countries (where labor is cheaper). In addition, I have already pointed out the
6
15
enormous surge in export activity between developing countries and BRICs, so-called South-South trade, relative to Triad-BRIC trade in Section 2.2.
Arguments against this interpretation include the possibility that trade liberalization in BRIC countries has made their final demand more dependent on foreign value added contributions in general. The reduction of trade barriers that increases opportunities for value added exports by developing countries to the BRICs also creates opportunities for value-adding exporters in the Triad. Therefore it may be the case that the value added shares of both Triad and ROW in BRIC final demand are increasing if the value added generated in the BRICs to supply their own final demand (i.e. domestic value added) is declining.
On balance, I anticipate that developing countries or the BRIC countries themselves may cut into the shares of value added income that would otherwise be induced in developed countries; thus, I predict that the Triad share of all value added embodied in BRIC final demand is on a declining trend.
HYPOTHESIS 3: From 1995-2007, BRIC countries are satisfying a declining share of their final demand through imports of value added from the Triad.
Finally I will decompose the value added content of BRIC final demand that is generated in the Triad, calculated for Hypothesis 3, into labor (characterized by skill level) and capital. This is to distinguish between factors of production embodied in BRIC final demand and to determine whether decoupling is more or less pronounced for specific production factors. The advantage of using shares of compensation for skilled workers rather than shares of skilled workers is that the decomposition of value added also includes compensation to capital. This is the residual after labor compensation is subtracted from gross value added and can make up large shares of value added.
16
remaining non-BRIC part of the world. This can be interpreted as strong evidence that developed countries are relatively abundant in high-skilled labor and developing countries are relatively abundant in low-skilled labor. Stehrer et al. (2012) also found that in value terms, developed countries tend to export relatively more skill-intensive products to developing countries. Furthermore, Timmer et al. (2013) argue that comparative advantages tend to be magnified with increasing fragmentation. This would suggest that not only are the relative shares of high-skilled labor compensation higher for the Triad, but that the absolute shares embodied in BRIC final demand may also be increasing over time if the Triad (and developing countries, too) specialize their production as fragmentation increases.
For this reason I expect decoupling to be least likely for BRIC imports from the Triad that consist of high-skill content. If supported, this may suggest that developed countries maintain favorable positions in global value chains while developing countries continue to perform less desirable (low-skill) value chain activities. Thus, I predict that developed countries are increasingly specializing the production of their added exports towards high-skill value-chain activities and that the BRICs are relying more heavily on imports from the Triad embodying high skilled labor compensation over time.
HYPOTHESIS 4: From 1995-2007, BRIC countries are satisfying an increasing share of their final demand through imports of value added that embody high-skilled labor compensation from the Triad.
4. Analytical framework and data sources
As discussed in Section 2.3., I use a trade in value added perspective (Johnson and Noguera 2012) that explicitly takes into account issues related to fragmentation and links it with socio-economic aspects of production, that is, the creation of income and jobs. The framework uses a standard decomposition methodology inspired by the input-output modeling approach first introduced by Leontief (1936). This allows me to depict the value added (and employment) content of production induced by final demand in the BRICs. I extend their approach towards a multi-country setting, e.g. similar to the international input-output modeling in Johnson and Noguera (2012). This is then applied to Groningen’s World Input-Output Database (WIOD), which contains data from 1995 to 2009 and can be downloaded for free at www.wiod.org.
17
production and/or countries. First I outline the approach implemented by Johnson and Noguera (2012) for exports of value added. This is an explicitly ex-post framework as no variables are estimated or modeled. It is based on the accounting identity that the sum of value added generated in production worldwide (global GDP) in a given year will be equal to the value of the global level expenditure, or final demand flows. This is because all final demand requires a certain amount of input by production factors to satisfy it. The sum of value added produced within all industries in a particular country gives its GDP (consumption + investment + government expenditure + exports – imports). In my case the objective is to decompose expenditure on products as reflected by final demand in BRIC countries into a stream of factor incomes and jobs generated in countries around the world.
Afterwards I generalize the approach of Johnson and Noguera (2012) in accordance with Timmer et al. (2013) and Los et al. (2012) by breaking down the values and quantities of production factors (labor and capital) induced by final demand. For example, I analyze changing job distributions for workers that are directly and indirectly induced by final demand patterns in the BRICs. This will be needed to test my hypotheses, which were described in the previous section.
4.1 Technical exposition
18
ease of notation. This leads me to the market clearing conditions which (as modeled by Timmer et al., 2013) I write as:
) , ( ) ( ) (s f s m s t y j t ij j ij i =
∑
+∑ ∑
(1)where yi(s)is the value of output in sector s of country i, fij(s) the value of goods sent from sector s in country i to satisfy final demand in any country j (home or abroad), and
) , ( ts
mij the value of goods shipped from this sector for use as an intermediary input by sector t in country j (also home or abroad).
The market clearing conditions for each of the SN goods can be modeled using matrix algebra. Let y be the stacked output vector and f the stacked final demand vector for the goods and services in each country-sector, each with a dimension of (SNx1) respectively.7 The summation of demand from each country yields global final demand, such that
) ( ) (s f s f j ij
i =
∑
. In addition, I define a global coefficients matrix of intermediate inputs Awith a dimension (SNxSN). The elements aij(s,t)=mij(s,t)/yj(t) are intermediate input coefficients and reflect the input from one country-sector (sector s in country i) that is needed for one unit of output by another country-sector (sector t in country j). This can be written as:
= NN N N N N A A A A A A A A A A L M O M M L L 2 1 2 22 21 1 12 11
where Aij is the SxS matrix with elements aij( ts, ). The matrix A captures deliveries of
intermediate inputs across all sectors and countries. Using matrix A and writing the stacked SN market clearing conditions from (1) I rewrite a system of equations as follows:
+ = N N NN N N N N N f f f y y y A A A A A A A A A y y y M M L M O M M L L M 2 1 2 1 2 1 2 22 21 1 12 11 2 1 7
19
where yi represents the output vector S in country i (such as textiles in the US), and fi the
final demand vector S for the products of country i. This can be expressed in matrix notation:
f Ay
y= + (2)
Equation (2) represents the fundamental input-output identity introduced by Leontief (1936). The model in (2) can be rewritten as (I−A)y =f , providing the following solution:
f A) (I
y= − −1 (3)
where I is an SNxSN identity matrix that has ones on the main diagonal and zeroes for all
other elements. (I− A)−1 indicates the multiplier matrix, famously known as the Leontief inverse. This represents the amount and type of intermediate inputs in all stages of production generated in the production process of one additional unit of final output, or final demand product n. In the WIOD this matrix has a dimension of 1435x1435. Now I will show that the element of a given row m and column n in this matrix reflects the total production in sector m in all production stages that is necessary to produce one unit of final output of product n. Consider
z
to be a very specific final demand vector with one element representing worldwide consumption of goods from a specific country-sector, such as Apple iPods produced in China. All remaining elements are set to zero. A final output ofz
iPods requiresintermediate inputs to be produced in the amount of Az. This is known as the direct effect. But the necessary inputs Az must be produced as well (such as the hardware), and this production requires the use of other intermediates, given by A2z, and so on. These are known
as the indirect effects. Thus,
∑
∞=0
k k
z
A , which can be rewritten as (I - A)-1z, represents the total
production necessary to satisfy an increase in final demand of one additional unit. It yields the sum of the direct effect plus all indirect effects. The Leontief inverse is multiplied by an f
column vector depicting the stacked final demand in any country for output from each country-sector.
value-20
added contributions from labor and capital, and disaggregate income and employment generated by labor of different skill types. Hence I proceed to generalize the approach of Johnson and Noguera and derive production factor requirements for any factor. This is also based on the methodology implemented by recent papers that use the WIOD database (e.g. Timmer et al., 2013; Los et al., 2012). I let pi(s) be the direct factor inputs per unit of gross
output of a production factor used in sector s in country i to produce a given product. I create a vector p with dimension SNx1 which is obtained by stacking these direct factor input coefficients for any particular factor (e.g. value added, skilled labor, all labor). They are country- and industry specific. The elements in p are direct factor inputs because they do not account for production factors embodied in intermediate inputs used. To take the latter into account, I use equation (3) to derive a SN factor requirements vector k for any final demand vector f . I pre-multiply the gross outputs needed for production of this final demand by the factor requirement vector (for example, the direct value added or labor coefficients) as follows: f A) (I p k= ˆ − −1 (4)
where a hat indicates a diagonal matrix (p = diag(p)ˆ ). Vector k reflects the total amount of direct and indirect factor inputs required for the production of each final demand level. Below, I decompose p and f into components attributable to particular countries or set of countries, and I decompose p vectors for broad factors into smaller components (e.g. labor into skill types). In this way I am able to analyze how much specific production factors in different countries contribute to the final value of particular products. All contributions can be post-multiplied with any vector of final demand levels (such as BRIC countries) to calculate the amount of a production factor that should be attributed to that specified final demand.
4.2 Application to my hypotheses
21
my hypotheses by modifying the production requirements in p and/or final demand vectors f. I will now discuss how I apply this to each of my hypotheses individually.8
In order to calculate the relative shares of income and jobs induced and test my first hypothesis, I first trace the development of changes in total income and jobs induced. Therefore I split up the final demand vector f that corresponds to the production factors of a source country i, such as Japan, into two sets of foreign markets: final demand in the BRICs (
f
BRIC), domestic final demand in the Triad (f
DOM) and the residual is foreign final demand in the rest of the world (f
ROW). This implies thatf
DOM+
f
BRIC+
f
ROW=
f
. I investigate the impact on income by using value added (million US $) to gross output (million US $) ratios of major developed countries as my p vector. This is multiplied by the Leontief inversematrix and the
f
BRIC demand vector to decompose the quantities of value added induced in thevarious countries and sectors, using the following equation which is adapted from equation (4): ROW BRIC DOM ROW 1 BRIC 1 DOM 1 ˆ ˆ ˆ k k k f A) (I p f A) (I p f A) (I p k + + = − + − + − = − − − (5)
To obtain quantities of value added induced in particular countries, one can pre-multiply these factor requirements by zero-one vectors that select countries of interest. Specifically, let
ei denote a 1435x1 vector with ones in positions where k contains factor requirements for a sector in country i, and zeros otherwise. (For example, eJapan would have ones in the 35 positions that correspond to the 35 sectors in Japan and zeros elsewhere). Then the value added in country i induced by BRIC countries can be written compactly as ei'kBRIC (summing over all sectors in country i, ' denoting a transposed vector), and the share of value added induced by BRIC countries as ei'kBRIC/ei'k. Value added induced for sets of countries can be obtained by summation (e.g. EU 15,
∑
∈ = I i i I e
e 'kBRIC 'kBRIC for all i in the EU 15), where
8
22
∑
∈ = I i i I ee is a zero-one vector that selects a set of countries I. For the impact on jobs, I use
employment (in thousands) to gross output ratios of major developed countries as my production requirements vector
p
in equation (5), with the same Leontief inverse matrix and final demand vectorf
BRIC.Now I can investigate the relative importance of BRIC final demand in generating value added in the economy compared to other final demand sources. I divide the value added contribution traced back to the BRICs as calculated above by all value added, ei'kBRIC/ei'k
where i is one of the developed countries. The ratio represents the share of a country’s GDP directly and indirectly attributable to final demand in the BRICs. Analogously I do this for employment, keeping in mind that all workers induced in a given country by domestic and foreign demand (including both foreign demand in BRICs and non BRICs) will be equivalent to a country’s labor force. Increases in the shares of value-added exports to the BRICs in GDP and jobs induced as a share of the overall size of the workforce over 13 years would confirm my Hypothesis 1 (and in Hypothesis 3 I relate this to the total final demand share of the BRICs).
As an extension of the analysis for Hypothesis 1, I will also compare the share of value added induced by the BRICs in Triad GDP, eI'kBRIC/eI'k, where I refers to the entire set of
Triad countries, with the share of value added induced by the rest of the world in Triad GDP,
k
k / '
' ROW I
I e
e . This sheds light on whether the importance of the BRICs as an export market (in terms of value added generated as a share of Triad GDP) is rising faster than the corresponding importance of ROW over time.
In my second hypothesis I examine the importance of the services sector embodied in value added and jobs induced by the BRICs. Thus I partition the vector of production requirements p by sector to derive the relative importance of different sectors:
BRIC BRIC . BRIC BRIC 1 BRIC 1 . RIC B 1 BRIC ˆ ˆ ˆ Services Manuf Agr Services Manuf Agr k k k f A) )(I p ( f A) )(I p ( f A) )(I p ( k + + = − + − + − = − − − (6)
For example, for services the pServices vector includes the direct factor input coefficients
23
broad sectors: agriculture, manufacturing, and services; the latter includes all industries in the economy except agriculture and manufacturing. (For reference, the sum across all three broad sectors is the same vector kBRIC as the second term in equation (5)). I report shares of value added and employment originating from a specific sector to total value added and employment generated from all sectors induced by final demand in the BRICs, e.g.
BRIC BRIC ' / 'kServices i k i e
e for country i or eI'kBRICServices /eI'kBRIC for a set of countries I (such as the EU 15). The hypothesis is supported if the share of services in total value-added exports and employment generated by the BRICs is rising faster than the share of the other sectors. That is, if the share BRIC BRIC
' / 'kServices i k
i e
e is rising relative to the share ei'kManuf .BRIC /ei'kBRIC, for either a
specific developed country i or sets of developed countries (using
e
Iinstead of ei).In Hypotheses 3 and 4, total final demand in BRIC countries can be traced back to the value added contributions of developed countries, value added contributions in the BRICs themselves and value added contributions of the rest of the world (ROW), by partitioning the requirements vector
p
as given by:BRIC BRIC BRIC BRIC 1 BRIC 1 RIC B 1 BRIC ˆ ˆ ˆ ROW BRIC Triad ROW BRIC TRIAD k k k f A) )(I p ( f A) )(I p ( f A) )(I p ( k + + = − + − + − = − − − (7)
where kBRIC refers to value added induced by all final demand of BRIC countries worldwide (=the total final demand flow to the BRICs). The vector pTRIAD = ˆ e TRIADp includes the value added coefficients for the EU 15, Japan, and the United States, and zeroes for other countries. Similarly, pBRIC collects value added coefficients for the BRIC countries, and pROW
the value added coefficients for all other countries. Hypothesis 3 is confirmed if the share of the first term in equation (7) in all value-added contributions induced by the BRICs,
BRIC BRIC/ '
'k ek
e Triad decreases over the 13-year time period, with e denoting a vector of ones. As an extension for Hypothesis 3 (and also 4), I distinguish the contribution of value added generated domestically in the BRICs and value added generated in other BRICs in final demand. This is to examine growth of BRIC-BRIC (i.e. South-South) trade in value added terms. I do this by performing the calculations in equation (7) for each BRIC individually and splitting the BRIC component into domestic value added and value added from other BRICs. For Brazil, for example, I write kBrazil =kTriadBrazil +kBrazilBrazil-Domestic +kRussia, InBrazil dia, China+kBrazilROW , so kBrazil
is decomposed like BRIC
k in (7) but with Brazil
BRIC
k split into Brazil
24
After this is done for all four BRIC countries, the domestic value added and value added induced by other BRICs is aggregated across all BRICs.
Hypothesis 4 disaggregates the component kTriad
BRIC on the right hand side of equation (7) by factor of production. Let pTRIAD − HS = ˆ e TRIADpHS be a vector that selects the value added coefficients of high-skilled labor for the EU 15, Japan, and the United States, and zeros for other countries and labor types, and let p ˆ TRIAD−MS, p ˆ TRIAD−LS, and p ˆ TRIAD−CAP be defined analogously for the value added of medium- and low-skilled labor and capital. Then kTriad
BRIC can be decomposed as BRIC BRIC BRIC BRIC RIC B 1 RIC B 1 RIC B 1 RIC B 1 BRIC ˆ ˆ ˆ ˆ CAP Triad LS Triad MS Triad HS Triad CAP TRIAD LS TRIAD MS TRIAD HS TRIAD Triad − − − − − − − − − − − − + + + = − + − + − + − = k k k k f A) )(I p ( f A) )(I p ( f A) )(I p ( f A) )(I p ( k (8)
Hypothesis 4 is supported if the share of the first term in all value added contributions increases over the 13-year time period, that is, e'kTriad −BRIC HS /e'kBRIC. For comparison, I will also compare this share to the shares of the other production factors in my analysis,
BRIC BRIC
' /
'k ek
e Triad −MS , e'kTriad −BRIC LS /e'kBRIC, and e'kTriad −BRIC CAP/e'kBRIC.
4.3 Data from the World Input-Output Database
The WIOD is a publicly available multi-regional database that is used to analyze the value added and factor content of trade. It covers 40 countries (the entire EU 27 plus 13 major non-European economies) with a timeframe spanning 15 years (1995-2009). These countries account for more than 85% of world GDP. The WIOD also provides a production structure for the non-covered part of the world that was estimated based on bilateral trade data and the production structure of key emerging markets. Note that my analysis only covers the years 1995-2007 because 2008-2009 data for key countries (notably China) were recently revised and released too close to the publishing of this thesis to be included, and because my focus is on the period before the financial crisis.
25
between imports by country and domestic production at the industry level that is used for intermediate or final use, both domestically by a country and foreign use. The tables trace the flows of consumption, production, and incomes within and across countries, and break down products according to their origin. The WIOD thus can be used for both inter-temporal and cross-country comparisons. In all the WIOD contains information on 35 industries and 59 product groups. See Timmer (ed.) (2012b) for more detailed information on the construction of the WIOD database.
One feature of the WIOD that sets it apart from similar databases, apart from the time-series perspective and industry level disaggregation of the supply of products, is that the WIOTs are combined with data on factor incomes and employment by skill type. This is especially important for my analysis on the impact of the BRICs on employment and income. The so-called socio-economic accounts break down value added, which is listed as a single category in the WIOD, into its components, including compensation for the production factors capital and various labor types for all 35 industries. Labor compensation and employment levels are broken down by three skill levels (low-, medium-, and high-skilled). This data builds upon data collected in the EU KLEMS project and extended to a larger set of countries. That makes it a unique database compared to the alternatives (Timmer et al., 2013). For skill level, the WIOD extends data on educational attainment in Barro and Lee (2010) by providing industry level data and relative wages by skill types. This reflects differences in remuneration of workers with different levels of education.
5. The impact of final demand in the BRICs: empirical results 5.1 Exports of value added and jobs induced by BRIC demand
Hypothesis 1 examines shares of income and jobs in the Triad economy and workforce induced by BRIC final demand. To prepare for testing Hypothesis 1, I first examine the total amount of income and jobs generated. As I do throughout this analysis, the Triad is defined as the EU 15, Japan and the United States in Section 5. Figure 1 shows that aggregate Triad exports of value added to the BRICs increased every year from 1995 to 2007. In the entire Triad, value added generated by final demand in the BRICs more than doubled from 1995 ($186 billion) to 2007 ($494 billion), and increases accelerated after 2002.9 The rate of increase over this time period was similar in each of the EU 15 countries, Japan, and the US.
9
26
Figure 1: Value added induced by all BRICs (constant US$ at 2005 prices, in millions)
Source: Own calculations based on the World Input-Output Database
Table 3 shows shares of value added exports to the BRICs in GDP for the largest Triad countries.10 Total value added generated in the Triad by BRIC final demand rose from 0.74% of Triad GDP in 1995 to 1.62% of Triad GDP in 2007. This is a significant increase and supports Hypothesis 1. In this time period, aggregate GDP of the BRIC economies grew from $2.5 trillion to $6.5 trillion (using constant US$ at 2005 prices), and the BRICs slightly less than doubled their share in world GDP (from 7% to 13%). Thus, the importance of the BRICs in terms of value added shares generated in the Triad increased even faster than their increasing weight in the global economy.
I have compared this result with the gross export share in GDP and find that gross exports exaggerate the perceived impact of BRICs. For example, in 2007 the share of induced value-added exports by the BRICs in Triad GDP (1.6%) was about 22% smaller than that of gross exports from the Triad to the BRICs (2.0%), and the discrepancy has been growing over time. The growing disconnect between value-added exports and gross-exports as a share of GDP is another reflection of the increasing fragmentation of production chains, which has also been found in prior research (Hummels et al., 2001; Stehrer et al., 2012). It supports the argument that there is a risk of overestimating the influence of foreign demand if relying only on export figures.
10
I have derived shares for all 18 Triad countries by dividing value added induced as shown in Table A1 in the appendix by GDP. This is not separately shown here.
27
Of all individual Triad countries, Finland (3.8% of GDP), Luxembourg (3.3%) and Germany (3.2%) were most dependent on BRIC final demand in 2007 while Greece (0.5%), the US (0.9%), and Spain (1.3%) were the least dependent. The low dependence of the US is not surprising given the fact that the US is less reliant on value-added exports in general than other economies of similar size (Los et al., 2012).
In sum, the part of Hypothesis 1 related to value added induced by the BRICs receives strong support because the value added generated by BRIC final demand made up growing shares of Triad GDP over time.
Table 3: Value added induced by all BRICs as (%) share of Gross domestic product
Source: Own calculations based on the World Input-Output Database
Next I compare the share of Triad GDP induced by the BRICs with the share of Triad GDP induced by foreign demand in the rest of the world (ROW). This is to analyze whether, out of all value-added exports directly or indirectly delivered from the Triad to markets outside of the Triad, the importance of the BRICs is increasing most rapidly. I define ROW as all countries except for the BRICs and the Triad throughout Section 5. The results confirm that the relative impact of the BRICs is increasing at a much faster rate. While value added induced by ROW accounted for 5.2% of Triad GDP in 1995, I find that this share increased to only 6.7% in 2007 which is a less rapid increase. This GDP share is considerably larger than the BRIC share in GDP because ROW includes not only the entire developing world but also developed countries such as Canada, Australia, Korea, Taiwan, emerging countries such as Mexico (which has close ties to the US), and the newer members of the EU.
Finally I decompose value added generated by all BRICs into the amounts generated by each of the BRICs individually. This is to investigate the relative importance of each BRIC. Figure 2 presents results that show China induces the majority of BRIC value added in the Triad with a share that rose from 43% in 1995 to 54% in 2007 (very similar to the increasing influence implied by export figures). Interestingly, there appears to be an inversed U-shaped pattern, with the importance of China peaking in 2003 before declining somewhat by 2007.
1995 1.16 0.68 0.91 1.04 0.67 0.46 0.74
2000 1.38 0.86 0.96 1.14 0.91 0.43 0.77
2005 2.38 1.06 1.14 1.57 1.94 0.63 1.22
2006 2.94 1.25 1.37 1.87 2.24 0.74 1.42
2007 3.20 1.25 1.59 2.05 2.59 0.88 1.62
28
This seems to be related to the growing influence of Russia and, to a lesser extent, India. Conversely, the shares of Brazil are on a declining trend. Still, China’s large share is reasonable considering China’s higher growth rate compared to the other BRICs. When disaggregating the Triad, I find that Japan is especially dependent on value-added exports to China.11 China accounted for 73% of all Japanese value-added exports to the BRICs in 2007, and so this is a higher share than China’s impact on the other Triad members. I interpret this as reflecting the continued importance of geography and the close ties that Japan shares with China in global value chains. In a similar way, the EU 15 is disproportionately dependent on Russia, which I believe also reflects geographical and historical ties.
Figure 2: Value added induced in Triad by each BRIC as (%) share of all BRICs, 2007
Source: Own calculations based on the World Input-Output Database
Figure 3 displays results on the impact of BRIC final demand on employment. The total number of jobs directly or indirectly serving final demand in the BRICs nearly tripled in most of the Triad from 1995-2007. In 1995, the BRICs induced 2.5 million workers in the Triad and this rose to 6.0 million in 2007. This is close to the inflation-adjusted rate of increase in BRIC GDP over the same time period.
11
