Extensive and Intensive Margins of India's Exports: Comparison with China

Transcription

1 WP Extensive and Intensive Margins of 's Exorts: Comarison with C Veeramani and Prachi Guta Indira Gandhi Institute of Develoment Research, Mumbai March 2014 htt://

2 Extensive and Intensive Margins of 's Exorts: Comarison with C Veeramani and Prachi Guta Indira Gandhi Institute of Develoment Research (IGIDR) General Arun Kumar Vaidya Marg Goregaon (E), Mumbai , INDIA (corresonding author): Veeramani@igidr.ac.in Abstract Should 's exort romotion olicies be targeted at accelerating exort growth at the extensive(new trading relationshis) or at the intensive margin (increase in trade of existing relationshis)? To hel answer this question, we undertake a comarative study of exorts from and by analysing the role of extensive and intensive margins in the exort market enetration of the two countries during We further decomose intensive margin into quantity and rice margins. As far as extensive margin is concerned, our results show that the ga between the two countries is getting narrower as is clearly catching u with. By contrast, lags significantly behind in terms of intensive margin due to an abysmally low and stagnant quantity margin. Intensification, rather than diversification, has been the crucial driving force of 's exort success. 's exorts of caital-intensive roducts erformed better comared to labour intensive roducts. The lacklustre erformance in labour-intensive exorts is entirely due to a lack of deth in's market resence even as it exanded the range of its roducts and markets. Our analysis suggests that can rea rich dividends by adoting olicies aimed at accelerating exort growth at the intensive margin. Contrary to the general ercetion, there exist a great otential for to exand and intensify its exort relationshis with the traditional develoed country artners. Keywords: Exorts, Extensive Margin, Intensive Margin,, JEL Code: F10, F14, F15 Acknowledgements: We are thankful to Prema-Chandra Athukorala for constructive comments. An earlier version was resented at the IGIDR Silver Jubilee Conference on Trade Industry held during October 22-24, We thank the articiants for comments.

3 Extensive and Intensive Margins of s Exorts: Comarison with C. Veeramani and Prachi Guta Abstract Should s exort romotion olicies be targeted at accelerating exort growth at the extensive (new trading relationshis) or at the intensive margin (increase in trade of existing relationshis)? To hel answer this question, we undertake a comarative study of exorts from and by analysing the role of extensive and intensive margins in the exort market enetration of the two countries during We further decomose intensive margin into quantity and rice margins. As far as extensive margin is concerned, our results show that the ga between the two countries is getting narrower as is clearly catching u with. By contrast, lags significantly behind in terms of intensive margin due to an abysmally low and stagnant quantity margin. Intensification, rather than diversification, has been the crucial driving force of s exort success. s exorts of caital-intensive roducts erformed better comared to labour intensive roducts. The lacklustre erformance in labour-intensive exorts is entirely due to a lack of deth in s market resence even as it exanded the range of its roducts and markets. Our analysis suggests that can rea rich dividends by adoting olicies aimed at accelerating exort growth at the intensive margin. Contrary to the general ercetion, there exist a great otential for to exand and intensify its exort relationshis with the traditional develoed country artners. 1. Introduction Since the early 1990s, has undertaken imortant reforms with a view to transforming its economic system from an inward looking lanned economy to one that is more market oriented. Trade and exchange rate olicies have been liberalised and restructured in order to remove the anti-exort bias endemic to imort substitution olicies. The focus of the exort olicy, by and large, shifted from roduct-secific incentives to more generalised incentives based rimarily on the exchange rate. In keeing with the argument that a market determined 1

4 exchange rate would make exorting activities inherently more attractive, adoted full current account convertibility in August The quantitative restrictions (QRs) on imorting caital goods and intermediates were mostly dismantled in 1992, although the ban on imorting consumer goods continued, with some excetions, until the late 1990s. Alongside the removal of QRs, customs duties in the manufacturing industries were gradually reduced. Following the new tariff reductions introduced in the March 2007 budget, has emerged as one of the world s low rotection and oen industrial economies (Pursell et al 2007). Did n exorters resond ositively to changes in the incentive structure engendered by the reforms? A revious study, focusing on merchandise exorts, showed that the first decade of reforms (from to ) was characterised by a relatively low growth rate of dollar exort earnings at 8% er year, while the second decade ( to ) stands aart for its strong growth rate of 21% a year (Veeramani, 2012). Data for the more recent years, however, indicate that exort growth has started to slow down. The value of exorts declined to about $300 billion in from $306 billion in , registering a negative growth of about 2% 2. Overall, desite the recent slowdown, since 2002, has recorded a suerior exort erformance comared to its own ast record. It must be noted, however, that s merchandise imorts have been growing faster than exorts throughout the ost-reform eriod resulting in increasing merchandise trade deficit. During the eriod to , while merchandise exorts recorded a growth rate of about 20% er annum, merchandise imorts grew faster at the rate of 23%. The surluses in services trade and rivate transfers have heled to artially offset the growing deficit in the merchandise trade account. In , for examle, the merchandise account shows a huge deficit of $195 billion, of which about $107 billion was offset by the invisibles earning, leaving a high current account deficit 1 The government introduced a major downward adjustment in the ruee exchange rate against the major international currencies in July In February 1992, a dual exchange rate system was introduced, which allowed exorters to sell 60% of their foreign exchange earnings at the free market rate and 40% to the government at the official lower rate. In Aril 1993, a further move towards the deregulation of the external sector took lace when the government adoted full convertibility on the trade account by unifying the official exchange rate with the market one. These stes culminated in adoting full current account convertibility in August The average annual growth rate for the eriod to is about 20% er annum. The latest available data shows that recorded an exort of $243 billion during Aril-January , registering a growth rate of just 5.7% on a like-to-like basis. 2

5 of $88 billion, or 4.8% of GDP. Faced with the situation of unsustainable current account deficit, the government has resorted to ad hoc olicy measures, such as utting restriction on gold imorts, to deal with the roblem. The long term solution to the roblem of unsustainable current account deficit lies in ensuring that exort growth kees ace with the growth of imorts. The crucial question is what tye of olicy interventions would hel achieve faster exort growth. The answer to this question, taking a cue from some recent studies, crucially hinges on whether olicy makers should target exort growth rimarily at the extensive margin (new trading relationshis) or at the intensive margin (increase in trade of existing relationshis). The intensive margin of a country s exort growth is attributable to its ersistent exort relationshis that is, exorts of already exorted roducts (old roducts) to already existing market destination for those roducts (old markets). Note that intensive margin growth can arise as a result of rice growth, quantity growth or both. The extensive margin refers to changes in the value of exorts due to diversification of old roducts to new market destinations and /or due to the exorts of new roducts. Trade models differ in terms of the emhasis laced on different margins as channels of exort growth 3. Traditional trade theory based on erfect cometition considers industry as the unit of analysis while firms within an industry are assumed to be identical and roduce homogenous roducts. Since roducts are not differentiated, horizontally or vertically, there is no extensive margin and exort growth comes from quantity exansion alone. Theoretical analyses of intra-industry trade usually rest on the assumtion of horizontal (different varieties are of a similar quality and same rice) or vertical (varieties are of different qualities and rices) roduct differentiation. Thus, in horizontal models, ioneered by Krugman (1979), exorts grow along the extensive margin that is, exansion of variety. In vertical models, as found in Flam and Helman (1987), exorts can grow along the rice and quantiy margins as a result of imrovement in quality. Finally, extensive margin is an imortant channel of exort growth in trade models with heterogeneous firms and fixed exorting costs (Metilz, 2003). In these models, exosure to trade will induce only the more roductive firms to enter the exort market, as entry into these markets is costly and can only be afforded by 3 See Hummels and Klenow (2005) for a detailed discussion on the imortance of different exort margins in different trade models. 3

6 the more efficient firms. Exorts along the extensive margin can grow with falling trade costs as new firms with horizontally differentiated variety enter the exort market. As to the emirical literature, the relative role of extensive and intensive margin in the growth of trade has been debated. Using a cross section data of 126 exorting countries in 1995, Hummels and Klenow (2005) found that larger economies exort substantially more and that extensive margin accounts for 62 ercent of the greater exorts of larger economies. Similarly, Evenett and Venables (2002) found extensive margin to be quite imortant for growth in develoing country exorts. A number of more recent studies, however, conclude that intensive margin lays the dominant role (Helman et al 2008, Amiti and Freund 2010, Felbermayr and Kohler 2006, Eaton et al, 2008, and Besedes and Prusa, 2011). Besedes and Prusa (2011, 371) notes that a country's oor exort erformance is not because it struggles to start new relationshis, but mainly because it lags behind the better erforming countries in terms of survival and deeening of existing exort relationshis. Clearly, a roer understanding of trade growth along the different margins, as oosed to the usual focus on aggregate trade flows, would better inform olicies. In order to decide whether exort romotion olicies be targeted at accelerating exort growth at the intensive or at the extensive margin, we need to know the relative role that the two margins have layed in determining s ast exort growth. In addition, to rovide a comarative ersective, it would be instructive to know the relative imortance of the two margins in other countries which have already become successful exorters. is a natural choice for comarison given its sectacular exort growth during the recent decades and its similarities with in terms of size, stage of develoment and relative resource endowments. In this aer, we decomose s exort erformance during into changes at the intensive and extensive margins. Intensive margin has been further decomosed into rice and quantity margins. To ut the analysis in roer ersective, s erformance at the two margins has been comared and contrasted with that of for the same eriod. The decomosition analysis has been undertaken for aggregate merchandise trade as well as for disaggregated sectoral level. Disentangling the rice and quantity comonents of intensive margin is imortant from a olicy ersective. For if exort growth comes mainly from quantity exansion, it may imly that the country must use increasing amounts of its resources caital, labour and natural 4

7 resources to sustain the growth. Countries in the early stages of develoment, with large endowment of surlus labour and resources, may exerience quantity, rather than rice, driven exort growth. On the other hand, rice driven growth lays a critical role in sustaining exort growth of advanced economies. If exort growth is mainly driven by rice growth and if rice reflects roduct quality, it may imly that the country has to invest more in human caital and technological innovation in order to sustain growth. The rest of the aer is structured as follows. Section 2 exlains the decomosition methodology and the database used. Section 3 discusses the decomosition results for while section 4 rovides the comarison with. Finally, Section 5 rovides the concluding remarks. 2. Decomosition Methodology Based uon the method roosed by Hummels and Klenow (2005), we analyze the structure of exorts from the given country i ( and, in our case) in year t to a destination grou D (which consists of several artner countries j). We suose that country i cometes with the rest of the world (r) in the markets of the destination grou D. Exort enetration of country i relative to r is denoted as S it that is, ratio of total exorts from i and r: where S it X X it rt j Nijt ji Nrjt X it = value of total exorts from country i to destination grou D in year t X rt = value of total exorts from r to destination grou D in year t x ijt = value of exorts from country i to artner j in roduct and year t x rjt = value of exorts from country r to artner j in roduct and year t x x ijt rjt N ijt = the set of artner-roduct airs where country i records exort relationshis (i.e., the set where x 0 ) N rjt ijt = the set of artner-roduct airs where r records exort relationshis (i.e., the set where x 0 ) rjt (1) 5

8 It is clear that X it (X rt ) is equal to the sum of exorts across all artner-roduct airs where country i (r) record exort relationshis. The exort enetration ratio (S it ) can be exressed as the multilicative roduct of extensive and intensive margins. The intuition behind this decomosition is that S it deends on: (i) the relative number of exort relationshis by i and r (extensive margin), and (ii) the relative values of exorts within the common set where both i and j record exort relationshis (intensive margin). For examle, i s exorts could be lower than r because the former records fewer exort relationshis than the latter that is, N ijt rjt N and/or because the value of exorts from i is lower than that of r within the common set of artner-roduct airs. Intensive margin for the exorting country i for the year t can be exressed as: IM X it it (2) xrjt ji Nijt The denominator of IM it reresents the total exorts from r in those artner-roduct airs where country i records exort relationshis in the given year t. Therefore, intensive margin is the ratio of country i s exorts to the total exorts of r within the common set of artner-roduct airs. The value of IM it is always ositive and can be below or above 1. For the case when Nijt is a subset of N rjt, the extensive margin for the exorting country i is defined as 4 : EM it ji Nijt X rt x rjt (3) The denominator of EM it reresents the total exorts from r while the numerator is the sum of exorts from r across the set of artner-roduct airs where country i records exort relationshis. It is clear that EM it is a measure of the fraction of r s exorts that occur in those artner-roduct airs where country i reorts ositive exort values. For examle, country i will have a lower extensive margin, but higher intensive margin, if its exort relationshis are fewer (that is, its exorts are concentrated in a small number of roducts and 4 The assumtion that N ijt is a subset of N rjt simly means that exort relationshis recorded by country i ( or ) is a sub set of the exort relationshis recorded by r. This is indeed the case in our data. 6

9 in a few artners). On the other hand, the country will have higher extensive margin and lower intensive margin, if it sreads its exorts thinly over many roducts and artners. It may be noted that the numerator of EM it is equal to the denominator of IM it. Therefore, the multilicative roduct of the two margins rovides S it, which is the ratio of total exorts from country i to total exorts from r. rjt X ji it Nijt Xit it it it xrjt X rt X rt ji Nijt IM EM S x (4) It is clear that the intensive margin catures the deth of a country s exort rofile, while extensive margin catures the breadth. The intensive margin (IM it ) can be further decomosed into a rice index (P it ) and a quantity index (Q it ). This decomosition makes use of the fact that IM it catures changes in the value of exorts due to changes in quantities and/or rices. IM it = P it Q it The rice index measures the weighted average ratio of i s rices to r s rices at each sixdigit level roduct, where the weights are the shares of each roduct category in total exorts of the common set [see Hummels and Klenow (2005) for more details]. P it N ijt uv uv ijt rjt w ijt where, uv ijt and uv rjt are rices (roxied by unit value) of roduct exorted by i and r resectively to artner j and wijt is the logarithmic mean of s ijt (the share of roduct exorted from i to j in i s total exorts) and s rjt (the share of roduct in r s exorts to j, where N ijt ): 7

10 s ijt j x ijt xijt N ijt, s rjt ji x rjt N ijt x rjt w ijt s ijt ln s ijt s rjt ln s rjt s ijt ln s ijt s rjt ln s rjt 2.1. Data In order to measure intensive and extensive margins, we make use of finely disaggregated [at the 6-digit level of Harmonised System (HS)] trade data from UN-COMTRADE accessed using the WITS software. The database contains extensive information, on a bilateral basis, on value, quantity and number of exort relationshis. An exort relationshi has been identified if 0 _ that is, if country i reorts ositive exort value to artner j in roduct x ijt (at the HS 6-digit level) and year t. Unit values (value of exorts divided by quantity), required for the decomosition of intensive margin into rice and quantity comonents, have been comuted at the 6-digit level 5. We use exort data reorted by, and rest of the world covering the eriod Exorts by rest of the world reresents the sum of exorts reorted by all countries (excluding country i) during the eriod. It may be noted that the number of countries that have reorted data to the UN vary from year to year. In order to make sure that our estimates are strictly comarable over time, our definition of rest of the world should comrise a homogenous set of countries that have consistently reorted data during the entire eriod. We note that 79 countries (including and ) have reorted data for all years during A small number of 6-digit HS codes, for which data on quantity are not available, have been excluded from the analysis. 6 Trade data, according to the HS classification, is available for from However, started reorting the HS based data only from

11 It must be noted that our calculation of extensive margin is sensitive to the level of data aggregation. In general, calculation based on relatively aggregate roduct categories would underestimate extensive margin by incorrectly relegating variety differences within categories into intensive margin. In this resect, we seek to do the best feasible job, given the availability of comarable data, by using the most disaggregated (6-digit HS level) data for and on a bilateral basis. 3. Decomosition Results for Table 1 reorts the exort enetration rate (S it ), the extensive margin (EM it ), the intensive margin (IM it ), rice index (P it ) and quantity index (Q it ) comuted for the eriod for aggregate merchandise exorts. s aggregate exort enetration rate increased from 0.7% in 1995 to about 2% in 2011 with a growth rate of 7.5% er year. The relative imortance of the two margins in driving the growth of exort enetration can be measured by decomosing S it into EM it and IM it. For examle, in 1995, S it (0.007) is the roduct of an extensive margin of and an intensive margin of The value of EM it suggests that the artner-roduct airs where had an exort resence (i.e., x 0 ) accounted for a quarter of r s exorts. Put differently, in 1995, only 25% of r s exorts faced some degree of direct cometition from. During the subsequent years, however, s extensive margin has increased significantly reaching 53% in 2000 and 67% in 2011, registering a growth rate of 5.6% er annum. Clearly, the breadth of s market resence has increased over the years, and by 2011 as much as 67% of r s exorts did face a direct cometition from. ijt In contrast, the deth of s market resence did not change much as reflected in the values of IM it which increased marginally from in 1995 to in 2011, with an average annual growth rate of 2%. These estimates suggest that the value of s exorts amounted to 2.7% that of r s within the common set in 1995, which has increased to 3.3% in The increase in extensive margin notwithstanding, s exort enetration rates (S it ) remained constant during the second half of the 1990s (Figure 1A). The rise in extensive margin, during this eriod, did not translate into an increase in exort enetration due to an 9

12 offsetting decline in intensive margin. Thus, the values of IM it and EM it suggest that s stagnant exort enetration during the second half of the 1990s was not due to a lack of diversification but mainly due to a lack of intensification and secialisation. Since the early 2000s, however, s exort enetration increased slowly but steadily, excet for a decline in Until about the mid-2000s, this growth has been almost entirely driven by extensive margin while intensive margin remained mostly unchanged. Thereafter, however, both the margins have been resonsible for driving s overall exort growth 7. It may be summed u that while the ositive contribution of diversification continued unabated throughout the eriod, exort growth stemming from intensification began, albeit in a small way, only since the mid-2000s. Now we turn to the decomosition of intensive margin into rice (P it ) and quantity (Q it ) margins (see Table 1 and Figure 2A). Our results show that the quantity margin remained constant while the rice margin showed a small increase (Figure 2A). The value of rice index (P it ) has always been below 1, which means that n roducts are generally cheaer than those from the rest of the world. For the whole eriod, the rice index recorded marginally higher growth (1.2% er annum) than the quantity index (0.9% er annum). Having discussed the general trends at the aggregate level, we now turn to discuss the results from the decomosition analysis carried out searately for different commodity grous. Table 2 shows the results for three broad commodity categories manufacturing, mineral fuels and other roducts 8. Overall, manufacturing, the major commodity category, mimics the trends and atterns observed at the aggregate level. The growth of intensive margin has been even more sluggish for manufactured exorts rimarily due to a negative growth along the quantity margin. Yet, s manufactured exorts erformed reasonably well in terms of overall exort enetration on account of extensive margin growth. 7 The decline in the exort enetration rate in 2010 was caused by extensive margin which declined from 65% in 2009 to 48% in This categorisation is done as follows. First, using a concordance table (available in WITS), we have matched the 6-digit HS codes with the corresonding Standard International Trade Classification (SITC) codes. Then, using data at the 6-digit HS level, we have decomosed the category-wise exort enetration rates (S it ) into the various comonents. Following the usual ractice, the commodity categories are defined as: manufacturing (SITC 5 to 8 less 68 and 667), mineral Fuels (SITC 3) and other roducts (SITC 0 to 2 lus 4, 68 and 667). SITC 9 (Commodities and transactions not elsewhere classified) has been excluded. 10

13 There has been an excetionally high exort growth of refined etroleum roducts from since the early 2000s (Veeramani, 2012). This is reflected in s exort enetration rate in mineral fuels, which increased noticeably from almost zero in 1995 to 4% in The high growth in mineral fuels can be attributed mainly to extensive margin growth though rice index also contributed ositively. A further disaggregated rofile of these broad commodity categories is resented in Aendix Table A1 and A2. Within the category of non-manufactured roducts, recorded imressive growth of exort enetration in Non-ferrous metals (SITC 68), which can be attributed almost entirely to growth along the quantity margin. Other nonmanufactured roduct grous, which recorded a relatively high growth of exort enetration are Crude materials, inedible, excet fuels (SITC 2) followed by Beverages and tobacco (SITC 1). Extensive margin has been the driving force of exort growth in both these roduct grous. Turning back to manufactured exorts, exort enetration rates in Machinery and transort equiment (SITC 7) increased much faster than average from a meagre value of little above 0.1% during the second half of the 1990s to 0.9% in This has been driven by the fast growth of intensive and extensive margins, with a growth rate of about 6% er annum along each of the margins. Yet, s overall exort enetration rate in SITC 7 is altry comared to a hefty 26% of in s traditional labour-intensive roducts, groued under Manufactured materials and Miscellaneous manufactured articles, showed oor erformance comared to caital-intensive grous such as machinery, transort equiment and chemicals. This lacklustre erformance of labour-intensive roduct grous is entirely due to a negative growth rate along the volume margin offsetting the gains from extensive margin growth. A number of studies have noted a general bias in s manufacturing secialisation attern in favour of caital and skill intensive industries and against unskilled labour-intensive industries 9. This is an anomaly given the fact that the country s true comarative advantage lies in unskilled labour-intensive activities. In order to shed further light on this, we comute 9 See for examle, Kochhar et al (2006), Panagariya (2007), Krueger (2010), Veeramani (2012, 2013). 11

14 the margins for commodity grous classified according to the factor intensity of roduction in different industries 10. The results are reorted in Table It is evident that the average annual growth rate of exort enetration has been the fastest in natural-resource intensive roducts (13.5%) followed by human caital-intensive roducts (10.7%) and technology-intensive roducts (9.9%). The growth rate has been the lowest in unskilled labour-intensive roducts (2.6%). The high growth in natural resource industries has been mainly driven by nonferrous metals (SITC 68), which recorded a hefty growth rate of 16.6% er annum (see Table A1). In the case of technology and human-caital intensive roducts, the growth rates of intensive margin (about 4.5% in each grou) has been noticeably higher than that for aggregate exorts (about 2%) while the growth rates of extensive margins are not significantly different from that for aggregate exorts. The high intensive margin growth in both these categories can be attributed to volume growth. In the case of unskilled labour-intensive roducts, extensive margin recorded an imressive growth rate of about 7% er annum comared to 5.6% for aggregate exorts. This suerior erformance along the extensive margin, however, did not translate into a high exort enetration rate due to an offsetting negative growth rate (-6.2%) along the quantity margin. The high growth of s extensive margin in unskilled labour intensive industries imlies that the country did succeed in establishing exort relationshis over many roducts and artners but the low and falling intensive margin means that its market resence has become increasingly thin or shallow. It is clear that s low exort enetration in unskilled labourintensive industries can be attributed to lack of intensification rather than lack of diversification. It has been argued that s labour laws create severe exit barriers and discourage large firms from choosing labour-intensive activities and technologies. 10 The Heckscher-Ohlin model, the workhorse model of international trade, ostulates that the comarative advantage of a country is closely related to its relative resource (factor) endowments. According to this model, a country will secialize in and exort roducts that are intensive in the use of the factor that is abundant in that country. Thus, being a labour abundant country, trade liberalisation is exected to generate faster growth of labour-intensive, rather than caital-intensive, exorts. 11 Using the factor intensity classification of the International Trade Centre (ITC), adated by Hinlooen and van Marrewijk (2008), we classify the traded roducts into five secific categories: natural resource-intensive, unskilled labour-intensive, human caital-intensive and technology-intensive. This classification is available at: (htt://www2.econ.uu.nl/users/marrewijk/eta/intensity.htm) (Viewed on 19 Setember, 2013). A total number of 240 items, at the 3-digit SITC level, have been groued into five categories (number of items in each category in arentheses): rimary (83), natural-resource intensive (21), unskilled-labour intensive (26), human caitalintensive (43), technology-intensive (62), and unclassified (5). 12

15 The bias in s incentive structure against labour-intensive manufacturing has a bearing on the geograhical attern of exorts from the country 12. Arguably, s roduct secialisation atterns rovide it with a comarative advantage in relatively oorer markets (such as Africa) but at the cost of losing market shares in the richer countries. Products from with high technology and skill content are unlikely to make inroads into the quality conscious richer country markets. These roducts, however, may enjoy a cometitive advantage in the relatively oorer country markets. At the same time, rich country markets rovide a huge otential for labour-intensive exorts from develoing countries including. Thus, secialisation out of traditional labour-intensive roducts imlies a general loss of s exort otential to advanced country markets 13. In the ast, the traditional develoed country markets (comrising Australia & New Zealand, Euroe, Jaan and North America) accounted for the major share of s exort basket. But their dominance has been steadily declining over the last two decades. The aggregate share of these markets in s merchandise exorts declined from about 63% in 1993 to 35% in 2010 (Veeramani, 2012). The remaining grou of countries (which include South & Central America, Caribbean and the various regions of Asia and Africa), account for nearly two-third of s merchandise exorts in The share of the high income OECD countries in s total manufacturing exorts declined sharly from 58% in 2000 to 41% in There are several reasons to believe that the general incentive structure in is biased against labourintensive manufacturing. In this context, the role of s labour laws is a highly controversial issue. Many argue that s rigid labour laws are rimarily resonsible for the lack of dynamism in labour-intensive manufacturing (see. Kochhar et al, 2006; Panagariya, 2007; and Krueger, 2010). Another grou of scholars, however, question this argument (see Nagaraj, 2011 and Bhattacharjea, 2006). Though, there is no unanimity of oinion in this regard, a growing number of emirical studies suggest that the role of labour laws cannot be ignored (see, for examle, Hasan et al, 2007 and Aghion et al, 2008). Other constraints that stand in the way of labour-intensive manufacturing include inadequate suly of hysical infrastructure (esecially ower, road and orts) and a highly inefficient and cumbersome land acquisition rocedure. Faced with ower shortages, caital and skill-intensive industries, such as automobiles and harmaceuticals, might be in a osition to rely on the high-cost internal sources of ower. This otion, however, is not affordable to firms in the labour-intensive segments that generally oerate with low margin. Similarly, one may argue that land acquisition rocedures create a bias against large scale labour-intensive manufacturing. 13 An illustrative examle will make this oint clearer. s exorts of assenger motor vehicles a caital and skill intensive roduct grou - increased remarkably from $151 million in 2002 to $4511 million in 2010, registering a growth rate of 44% a year. Low & middle income countries are the major destinations for these exorts from. In 2010, the high-income countries accounted for only 8% of the n exorts of assenger motor vehicles while Sub-Saharan Africa accounted for 11%. By contrast, the high-income countries accounted for 58% of s total exorts of HS 6105 ( men s or boy s shirts, knitted or crocheted ) - a traditional labour intensive grou while the Sub-Saharan Africa accounted for just 1%. 13

16 In order to reflect further on the changing market destination of s exort, we carry out the decomosition analysis for two broad grous of markets high income OECD versus other countries (see Table 4). It is clear that s exort enetration to the high-income OECD countries grew much slower (5.5% er annum) comared to other market destinations (8.7% er annum). The slow ace of growth in high-income OECD can be attributed rimarily to the negative growth rate of quantity margin. The different anels in Figure 3 deict a disaggregated rofile, across market grous, for the three major factor intensity based commodity grous. These anel charts deict several interesting contrasts with, a detailed discussion of which is rovided in the next section. 4. Comarison with s overall merchandise exort enetration rate has increased dramatically from 4% in 1995 to above 18% in 2011, registering a growth rate of 11.5% er year (Table 1). Turning to the decomosition results, s extensive margin in aggregate merchandise exorts had increased from 39% in 1995 to 75% in While always recorded a higher EM it value than, the latter recorded a higher growth rate (5.6% er annum) than the former (4.9% er annum). Thus, is catching u with in terms of roduct and geograhic breadth of exort markets. Clearly, differences along the extensive margin growth cannot account for the vast ga between and in terms of overall exort enetration rate. It is beyond doubt that s erformance has been mainly driven by growth along intensive margin, which had increased from about 10% in 1995 to as high as about 25% in 2011, with an imressive growth rate of 6.3% er annum (comared to 2% er annum for ) 14. s exort enetration rates grew somewhat slowly during the second half of the 1990s while the ost-2000 eriod stands aart for the excetionally high growth of exort enetration rates (see Figure 1B). It is clear that this high growth during the ost 2000 eriod has been brought about mainly by intensive margin. Extensive margin also contributed, to a lesser degree, during the first-half of the 2000s while intensive margin was entirely resonsible for maintaining the growth during the second-half. 14 While comaring these growth rates, it is imortant to kee in mind the difference in the values of IM it for the two countries in the beginning of the eriod. started off with a high value of 10% comared to just 2.7% for. Desite the high base effect, s growth rate (6.3%) is more than three times as high as s (2%). 14

17 Overall, it is clear that intensification, rather than diversification, has been the crucial driving force of s exort success. By contrast, s low exort enetration is due to a lack of intensification rather than lack of diversification. Having established that s exort exansion took lace rimarily along the intensive margin, the next ertinent question is whether the intensive margin growth has been driven by rice increases or increases in exort volumes. The rice (P it ) and quantity (Q it ) comonents of s intensive margin are shown in Table 1 and in Figure 2B. It is evident that s intensive margin growth has been mainly driven by volume, which grew at 4.6% er annum during the entire eriod of and at 9.2% er annum during the sub-eriod of In contrast, as noted above, s Q it index for aggregate merchandise exorts remained almost unchanged during Values of s rice index (P it ) have always been below 1, which means that Chinese roducts are generally cheaer than those of the rest of the world. Earlier, we noted a similar attern for vis-à-vis rest of the world. Interestingly, comarison between and reveals that the former record higher P it values than the latter 16. In other words, on an average, while both n and Chinese roducts get lower rices than those exorted by rest of the world, Chinese roducts are generally sold cheaer than n roducts. But, s relatively high rice margin has not translated into high intensive margin due to an abysmally low and stagnant quantity margin. While s P it values are lower than that of, it must be noted that s quantity enetration over the years has been achieved without exerting any downward ressure on the relative rices of its roducts 17. In fact, rices contributed ositively, albeit to a small extent, to s overall exort growth. For, as can be seen from Table 1, s rices relative to the rest of the world have increased at 1.6% er annum comared to the annual growth rate of 1.2% for. Clearly, s exort quantity exansion has not exerted a downward ressure on its rices. 15 This result is similar to Bingzhan (2011) who, using a different decomosition method, shows that 's exort growth, between 2001 and 2007, was mainly driven by quantity growth accounting for about 70% of its overall exort growth. 16 Using finely disaggregated (10-digit level) U.S imort data, another study showed that in a large majority of the cases, the 10-digit level exort unit values of are significantly higher than that of (Veeramani and Saini, 2011). This is consistent with our finding that the P it values are higher for than for. 17 Amiti and Freund (2010), however, noted that, between 1997 and 2005, the average rices of goods exorted from to the United States fell by an average of 1.5 ercent er year. 15

18 That s exort volumes have grown desite the increase in its relative rice may suggest that non-rice factors, such as roduct quality and roduct differentiation, might have layed a role in bringing about the quantity exansion. Quantity exansion can accomany relative rice increase if there has been an imrovement in the quality of roducts being exorted from. Additionally, exort variety growth within 6-digit roduct categories, which has not been taken into account in our calculation of extensive margin, can contribute to volume increase without a corresonding change in relative rice 18. In general, the simultaneous increase of quantity and rices could be driven by several factors such as changes in roduct quality, variety growth, roductivity imrovements in, declining rofit margins and exchange rate movements 19. A detailed analysis of the relationshi and interlay between these factors and their relative imortance is beyond the scoe of the resent aer. What factors exlain the fact that the P it values have been much higher for as comared to? Detailed and careful comarison of manufacturing industries in and by Van Ark et al (2010) shows that unit labour costs (nominal cost of labour required to roduce one unit of outut) was slightly higher for relative to for the year Thus, relative cost differences is not the key factor that exlains the difference in the values of P it between and. It is also unlikely that the P it values reflect differences in the quality of roducts being exorted from the two countries; if suerior quality had been the reason for s higher P it, it would have resulted in greater quantity enetration by. It is more likely that differences in the P it values between and closely reflect certain fundamental differences in the nature of intra-roduct secialisation in the two countries. It is now well recognised that countries engage in roduction and trade by secialising in distinct varieties and rocess within a roduct. The roduct varieties sourced from different countries could be differentiated on the basis of quality, factor content, and 18 Use of more disaggregated data is unlikely to alter our results qualitatively. For examle, Amiti and Freund (2010), using finely disaggregated data (8-digit level Chinese exort data and 10-digit level U.S. imort data) have established the rimacy of intensive margin for. 19 The simultaneous increase of quantity and rice imly that s raid exort quantity growth might not have adversely affected its terms of trade that is rice of exorts in relation to imorts. 20 The comarison is available for two years, 1995 and recorded a higher unit labour cost than both in 1995 and 2002 though the difference has declined significantly during recorded a noticeable decline in its unit labour cost as a result of a significant growth of labour roductivity during this eriod. Desite its higher labour roductivity, s unit labour cost remained higher due to its relatively higher level of labour comensation. Van Ark et al (2010) also noted that before the early 2000s recoded higher levels of labour roductivity than reflecting the high caital intensity in s manufacturing roduction. 16

19 other attributes. For examle, Scott (2004) notes that the United States increasingly sources the same roduct from both high-and low-wage countries and that the unit values of imorted roducts vary widely even within finely detailed roduct categories 21. It was also noted that, in general, caital and skill-abundant countries exort varieties that command higher rices while the varieties exorted from labour-abundant countries receive lower rices. Further, a variety s unit value increases with an increase in the caital intensity of the roduction technique used to roduce it. As mentioned earlier, s secialisation atterns exhibit a general bias in favour of roducts/rocesses that are caital and skill intensive. Thus, if unit values of varieties are ositively correlated with the caital and skill intensities of their roduction, s relatively higher P it values is a reflection of its intra-roduct secialisation in caital and skill-intensive varieties. A higher rice that results from distorted 22 secialisation, however, does not translate into higher volume of exorts as evident from the stagnant values of s Q it. By contrast, s relatively lower P it values reflect its high degree of intra-roduct secialisation in labour intensive varieties and rocesses. The values of s Q it index recorded a raid growth since the attern of its secialisation has been in alignment with its comarative advantage. Driven by its high level of secialisation, has been able to cature a significant world market share in labour-intensive varieties (Amiti and Freund, 2010) 23. At the same time, the high-income OECD countries have resonded to the cometition from by secialising in more sohisticated varieties - that is, varieties that embody higher level of technology, skill and caital (Schott, 2008). In order to exlain the observed contrasts between and, it is imortant to closely look at the decomosition results for different commodity grous. The high growth of 21 For examle, Schott (2004, 647) notes that men s cotton shirts from Jaan are roughly 30 times as exensive as the identically classified variety originating in the Philiines. 22 This attern of secialization is distorted as it is inconsistent with s comarative advantage in labourintensive activities given its relative factor endowments. Based on a comarative analysis of relative resource endowments (hysical caital, human caital and arable land), Veeramani (2013) notes that both and are abundantly endowed with unskilled labour while hysical caital, skilled labour and land are relatively scarce in both the countries. For the more recent years, s relative endowment of unskilled labour is significantly higher than that of. It is beyond doubt that the true comarative advantages of both the countries, more so for, rest in varieties and rocesses that intensively use unskilled labour rather than hysical caital and skilled labour. 23 Amiti and Freund (2010, 54-55), based on a detailed analysis of s exorts during , observes that the skill content of s manufacturing exorts remained unchanged once rocessing trade is excluded. When examining the skill content of s total manufacturing exorts, it looks like there has been an increase over the samle eriod. However, it turns out that this is mainly due to the increased skill content of imorted inuts that are then assembled for exort a ractice known as rocessing trade. 17

20 s overall exort enetration has been almost entirely driven by manufactured roducts (Table 2). s exort enetration in manufactured roducts grew at the rate of 13.5% er year mainly on account of an imressive growth along the intensive margin. The high growth along the intensive margin, in turn, has been mainly brought about by a volume growth of about 5% er annum. By contrast, s manufactured exorts recorded a negative growth rate along the quantity margin (-0.8%). It may be reiterated that a s high volume growth of manufactured exorts has been achieved without any decline in rices in fact, the rice index recorded an increase at the rate of 2.1% er annum. s suerior exort erformance is reflected across all roduct grous within manufacturing (Table A2). Machinery and transort equiment (SITC 7) recorded the fastest growth of exort enetration from about 2% in 1995 to as high as 26% in 2011 with a growth rate of about 20% er annum. This is followed by SITC 6 (11.5% er annum), SITC 8 (9.1% er annum) and SITC 5 (8.8% er annum). For all roduct grous, intensive margin growth has been brought about by volume growth and without any discernible decline in the rice index. In the case of SITC 8, the traditional labour-intensive grou, s exort enetration increased remarkably from about 12% in 1995 to as high as 48% in For this roduct grou, the value of s intensive margin for the latest year is a hefty 50% comared to just 3% for, affirming the increasingly dominant role that the former lays in the world market for labour-intensive roducts 24. Turning to the factor-intensity based grous, Table 3 shows that s average annual growth rate of exort enetration has been the fastest in technology-intensive roducts (18.4%) followed by human caital-intensive roducts (12.8%) and unskilled labourintensive roducts (10.8%). The grou of rimary roducts exerienced near stagnation in terms of exort enetration. Perhas, the most striking asect of s exort erformance is the henomenal increase of its exort enetration in unskilled labour-intensive industries from 15% in 1995 to little less than 70% in For the latest year, the value of s intensive margin for this category is as high as 73% comared to a altry 5% for. For 24 In the case of non-manufactured roduct grous, n exorts generally grew faster than s. s exort enetration rate in mineral fuels declined from about 4% in 1995 to 2% in 2011, due to a large negative growth (-7.2 er annum) along the quantity margin. Excet for Non-ferrous metals (SITC 68), and Food and live animals (SITC 0), s exort enetration in other non-manufactured roduct grous recorded either negative growth or stagnation. 18

21 this category, s extensive margin also is highly imressive with an EM it value of 0.94 in 2011 comared to s 0.73 for the same year. s high exort enetration rates in technology and human caital-intensive roducts are driven by intensive margin, which grew faster than extensive margin in both the roduct categories. In the case of technology-intensive roducts, in articular, s intensive margin grew at a remarkably faster rate of 12% er annum comared to the extensive margin growth rate of 5.7%. While both volume and rice contributed ositively to the intensive margin growth in technology and human caital intensive industries, the relative contribution of the former has been much higher. The attern of growth in unskilled labour-intensive industries reveal some contrast in the sense that the growth of extensive margin (8.6% er annum) in this category has been higher than that of intensive margin (2% er annum). Further, intensive margin growth has rimarily been a result of rice increase rather than quantity increase. Overall, the results for s unskilled labour-intensive industries indicate that while intensive margin layed a crucial rule in increasing its exort enetration in the ast, maintained its high growth over the last few years by diversifying into new markets and roducts (as evident from the high extensive margin growth) and by imroving quality of the existing roducts (as evident from the growth of rice index). Thus, s henomenal exort success in unskilled labour intensive industries stemmed from its ability to continually exand the breadth as well the deth of its market resence. In contrast, s lacklustre erformance in this category is entirely due to a lack of deth in its market resence even as it could exand the range of its roducts and markets. In contrast to, s exort comosition shows a strong bias in favour of labourintensive roduct grous. It is imortant to note that the above analysis underestimates the imortance of labour-intensive exorts from. This bias arises due to s significant resence in global roduction networks and fragmentation based trade in a range of manufactured roducts (Athukorala, 2012; Veeramani, 2013). s exort romotion olicies since the 1990s relied heavily on a strategy of integrating its domestic industries with the global roduction networks 25. In articular, based on imorted arts and comonents, 25 Global roduction networks refer to the links between a lead or a key firm and its suliers in different countries (Weiss, 2011). In certain industries, such as electronics and automobiles, technology makes it ossible to sub-divide the roduction rocess into discrete stages. In such industries, the fragmentation of roduction rocess into smaller and more secialised comonents allows firms to locate arts of roduction in countries where intensively used resources are available at lower costs 19