Analyzing India's intra-industry trade in post-liberalization era.

Srivastava, Amit ; Medury, Yaj

Abstract

Intra-Industry Trade (IIT) has added a new dimension in understanding the phenomenon of international trade and now it has become an integral part of the world trade. The present paper analyses the patterns of India's IIT with respect to India's trade with World. The paper focuses on three aspects of India's IIT, these are, the effect of liberalization on IIT; contribution of export and import on IIT, and finding out the overall nature of India's IIT. The result of the study shows that the liberalization process has not affected the degree of liT significantly; over-the-years, contribution of import in IIT has surpassed the contribution of export; and India's overall IIT is vertical in nature but interestingly its contribution has decreased a little bit.

Keywords: Intra-Industry Trade, Vertical and Horizontal Intra-Industry Trade, Import-Weighted Average Tariff

JEL Classification: F12, F14

I. INTRODUCTION

International trade is a trade between different countries and countries engage themselves because they benefit from doing so. Different theories have been proposed to explain the reasons for international trade, some of them are--Absolute Advantage theory by Adam Smith and Comparative Advantage by David Ricardo. Both of these theories helped in understanding international trade in a better way. Later on Eli Hecksher and Bertil Ohlin pointed out the importance of factor endowments in deciding the nature of a country's trade. They combined the concept of factor abundance with the idea that different goods require different factors of production and a country will have comparative advantage in the goods whose production involves intensive use of country's abundant factor. However, the focus of international trade was limited for measuring (and understanding) the benefit of a country from trade, and it was assumed that export and import will depend on the overall efficiency of the country for producing the goods. If the country is relatively efficient in producing a good (by whatever reason) it will export, otherwise import. The underlying assumption over here was--a firm in the same industry will produce identical goods and consumers will not be able to distinguish between the products of different firms. A country, therefore, either only exports goods within the same industry, or only imports these goods, but not simultaneously export and imports goods within the same industry. But later on it was realized that many countries involved in simultaneous exports and imports of very similar goods and services and this was termed as intra-industry trade (IIT).

Intra-industry trade was first observed in 1960s while analyzing the nature of international trade of the Benelux Customs Union and thereafter 6 founding members of European Economic Community. Later on it was realized that goods and services produced by different firms (of same country or different countries) are not identical, thus, these goods and services are close, but not perfect, substitutes, and this requires a change in the demand structure of an economy. However, specially after different works of Grubel and Lloyd in 1970s, a systematic and coherent investigation of intraindustry trade theory, measurement, empirical analysis and policy aspects started (Greenaway and Milner--1987).

This evidence of IIT was contrary to the classical trade theory and was seen as the starting point of the renewal of international trade theory, the theoretical base in understanding this phenomenon being considerably enlarged since then (Lionel Fontagne--2005). Product differentiation forms the basis of this type of trade as similar products that differ either in terms of quality or in terms of attributes (Burange and Chaddha--2008). Until mid-1980s, researches were specially concerned with the theoretical explanation of intra-industry trade and thereafter the focus has been changed from theory to measurement, empirical analysis and policy aspects. Later on it was realized that intra-industry trade is a substantial part of total trade, specially bilateral trade between countries with similar relative factor endowments (Gullstrand--2002). Conventionally it was believed that intra-industry trade occurs mainly among the developed countries, however, Reinert (2003), worked on Asian countries (developing) and concluded that the phenomenon is emerging even in Asian countries.

With continuous research on IIT, since early 1980s, it has been possible to disentangle IIT into two categories--horizontal intra-industry trade (HIIT) and vertical intraindustry trade (VIIT). In case of HIIT, products differ in their attributes but do not differ in quality or price. Producers in the industry are assumed to use same factors of production and same techniques. On the other hand, in VIIT, products can differ in terms of quality which leads to difference in prices also. Here producers use different factors of production and different techniques.

This paper focuses on India's IIT with the world between 1988-2008. Some studies have already been conducted regarding India's IIT but these studies have done at high aggregation level, i.e., at HS-4 digit level, for example, Veeramani (2003) and Burange & Chaddha (2008). Our work is based on somewhat more disaggregated level, i.e., HS6 digit level. The present paper tests few hypotheses with special reference to India, both at aggregated and disaggregated level, these are; first--IIT increases with liberalizing economy, second--in India's IIT, contribution of export is more than import, and third--India's overall IIT with world is vertical in nature. To study the effect of liberalization on IIT in a better way, we shall conduct two more tests. In first test, we shall try to find out the correlation between changing tariff rate with total IIT as well as different components of IIT; and in second test the pattern of IIT, HIIT and VIIT of different product groups will be discussed.

The paper is organized as follows: Section II provides a brief review of theoretical literature. Section III deals with the methodology of analyzing IIT and disentangling it into HIIT and VIIT, Section IV deals with data analysis, Section V deals with Results and Section VI will conclude the paper and will also discuss about the limitations and prospects of future work of the study.

II. LITERATURE SURVEY

Historically, intra-industry trade (IIT) was not even thought of and reasons for international trade was considered as differences in factor endowments or differences in technology. In simple words, countries produced and exported products which matched with the characteristics of the country while import the products which were least matched with the nation's characteristics. Therefore it was considered that international trade between countries will be driven by comparative advantages of countries. Greater the differences in national endowments, greater the trade volumes. Therefore, international trade was considered as inter-industry trade, i.e., trade will be taken place between different industries for export and import. Ricardian and Hecksher-Ohhn theories contributed significantly in understanding this form of international trade.

Till the time when international trade was considered to be the inter-industry trade, there was no problem in understanding and explaining the trade pattern. But later on, with the revelation of intra-industry trade in the early 1960s, the perspective of defining and explaining the reasons for international trade had changed. The research got initiative in this direction specially after Leontief Paradox, in 1950s, in which he proposed that "although US is most capital abundant country in the world but exported labor-intensive commodities and imported capital-intensive commodities". This finding surprised the world because it was against the existing concept and propelled the research in this direction. Later on, Linder, in 1961, observed the interesting phenomenon that was "a country with high and similar per capita income has similar demand patterns; therefore the result was exchange of similar but differentiated products". This was termed as Linder's Hypothesis and it changed the perspective of analyzing international trade, earlier it was analyzed/described on the basis of supply based theories like factor endowments and technology etc, but thereafter the analysis was based on the demand structure of a country. Although, Linder hypothesized that nations with similar demands would develop similar industries but he did not use the term "intra-industry trade". Later on his work was extended by Hanink (1988) and he empirically proved that the fundamental relationships expressed by Linder were valid.

Specifically the term "intra-industry trade" was first used by Balassa (1966) and initial empirical work was done by Grubel (1967) where he studied the nature of IIT was dominating in the trade liberalization period of EEC (European Economic Community), but even then the concept had not drawn enough attention until its measurement technique was developed by Grubel and Lloyd (1971). They developed an index (popularly called as GL index) to measure the degree of IIT and proved that IIT is a "pure phenomenon". They suggested that the main reason of IIT was specialization in narrow product ranges, joint production unmatched by complementarities in demand, trade across borders in high transportation costs industries. According to them, oligopoly market is the most important reason for IIT.

After then extensive research have been done on IIT to understand its nature, determinants and measurement related issues and this was required also because the traditional trade theories were unable to explain the reason of IIT. These works led to differentiate the trade of products into two categories-horizontal and vertical. Horizontal product differentiation means that many varieties of the products are available but these varieties do not differ in terms of quality significantly; while vertical product differentiation means that many varieties of the products are available and these varieties differ in terms of quality in a meaningful and significant way.

Initial works on product differentiation and IIT was mainly based on horizontal product differentiation. The famous "love-of-variety" approach was proposed by Dixit and Stiglitz (1977), which was based on the monopolistic environment and they concluded that consumers would prefer variety to one product only. Almost in the same line of work, Lancaster (1980), assuming perfect monopolistic market, said that consumers will prefer his "most-preferred goods" or "ideal goods" if different specifications of the same product is available before him and he concluded that IIT is most likely to occur in similar economies rather than different ones. Different works of Krugman (1979, 1980, 1981), based on Chamberlinian monopolistic competition under increasing return, concluded that international trade is very much possible in economies of similar factor endowments and similar countries, in fact, will trade more.

Later on studies regarding vertical product differentiation had also been conducted. Falvey (1981) worked on the model of vertical IIT. He showed that vertical IIT and inter-industry trade can co-exist and increasing return to scale is not an essential condition for trade. He also showed that trade would vary inversely with the level of trade restrictions. Some more refined work was done by Falvey and Kierzkowski (1987) where they showed that the share of vertical IIT will be correlated with the average market size of the two countries and it will be greater, the greater the difference in the capital-labour endowment of the two countries. Shaked and Sutton (1983), and Sutton (1986) worked on vertically differentiated goods but their work was based on "oligopoly" market rather than monopolistic environment. In their work they concluded that for vertical differentiation condition, the finiteness condition (i.e., only few firms will survive) is necessary and sufficient. Here the conditions will be applicable in those industries where the main burden of quality improvement is on R&D and other fixed costs while unit variable costs will rise only slowly with increase in quality, as well as, consumers are willing to pay for quality improvements. Davis (1991) worked on economies of scale and its relations with IIT and found that increasing returns are not necessary to explain IIT and it is very much possible with constant return to scale. He developed a simple model that can give a unified account of intra-industry trade and inter-industry trade, and the striking feature of the model was that intra-industry trade attains a maximum at a point where countries have identical factor endowment ratios.

As far as the measurement issues of intra-industry trade are concerned, the first notable work was done by Grubel and Lloyd (1971), they developed an index for measuring the degree of IIT. They worked on Australian trade data (mainly two sectors--iron and steel, and petroleum) and found that as the aggregation level increases, degree of IIT also increases. To understand the importance of IIT, for the first time, they measured the average IIT as a percentage of total commodity trade and they found that in most of the cases the average was quite high, both for countries as well as for industries. Empirically, they also found that the level of Australian IIT is maximum with the countries of similar factor endowments (New Zealand and South Africa). Later on Greenaway and Milner (1981, 1983) worked on the measurement issues of IIT using UK's trade data. In their first work they focused on the problems regarding measurement of IIT at the industry level that whether and how to appropriately adjust the industry indices for the effects of overall trade imbalance and they concluded that even-if macrodisequilibrium is identifiable, it cannot be generalized at industry level because IIT is the product of industry characteristics and not of macro-economic adjustment process. In their second work, they focused on the problem of categorical aggregation in the measurement of IIT. They discussed different ways to identify and check the problem.

Regarding the factors influencing the extent of intra-industry trade, an elaborated work was done by Balassa and Bauwens (1987). They worked on a multi-country and multi-industry framework specially for manufactured goods where product differentiation predominates. They selected few important determinants of IIT, as discussed in the then existing literature, and categorized them into two categories--industry characteristics and country characteristics, the latter was further divided into common country characteristics and specific country characteristics. Their study was based on 38 countries and 152 industries and they found that in most of the cases the results were same as expected theoretically. In another empirical work, Bergstrand (1990) also tried to find out the effects of different determinants on the degree of IIT. He worked on the trade data of OECD countries and basically horizontally differentiated products, and found that in most of the cases the determinants were having same sign as expected theoretically. In his work, for the first time, he tried to "uncouple" the effects of higher average capital-labour ratio and higher average per capita income. He proved that for some manufacturing industries, a higher average capital-labour endowment ratio for two countries can lower the share of IIT, even though a higher average per capita income can raise it.

Although different dimensions of IIT had been discussed till then, but there were no formal criteria to make distinction between horizontally differentiated products with vertically differentiated products. Abd-el-Rahman (1991) was the first one to disentangle horizontal product differentiation with vertical one. He said that if the difference between export and import unit value is maximum upto 15 per cent, the product will be said to be horizontally differentiated; and if this difference exceeds the 15 per cent limit then it would be considered as vertically differentiated products. Later on some more works have been done to emphasize the importance of disentangling horizontal IIT (HIIT) with vertical IIT (VIIT) like Greenaway, Hine and Milner (1994, 1995), Gullstrand (2002), and Andresen (2003). Fontagne, Freudenberg and Gaulier (2005) had done a systematic decomposition of world trade into three trade types: inter-industry, intra-industry in horizontally versus vertically differentiated products, and found that increase in IIT at world level was due to two way trade of vertically differentiated products, so according to them VIIT was the major contributor in the IIT and world trade. However, a recent work by Cabral, Falvey and Milner (2008) tested the hypothesis about the relation between endowment differences and the share of HIIT, VIIT and NT (net trade). They found that the prediction for HIIT was quite conventional--larger endowment differences would reduce the level of HIIT. While, the result of VIIT was bit surprising, theoretically it should grow with increase in endowment differences but it grew as long as the endowment differences remain small. The effects of larger specific factor endowment differences depend on whether the specific factor was used by the industry, if not, then VIIT declines for larger endowment differences; if so, then the share of VIIT increases (decreases) if the trading partner has an ever larger (smaller) endowment. Intra-industry trade was considered to be significant because its share in world trade was found to be increasing. Brulhart (2008) calculated GL index for world for both SITC 3-digit and 5-digit and found that it was about 20% and 7% respectively for the year 1962 while it increased to 44% and 27% respectively in the year 2006. The trend has increased significantly specially after 1990. Therefore we cannot overlook the contribution of IIT in total trade.

Some impressive works have been done specially with Indian context, few of these are mentioned here. Veeramani (2001) compared India's IIT for manufactured and primary commodities, between 1987 to 1999. He worked on 4-digit ITC data and measured the degree of IIT by using GL-index. He found that both, GL index and export growth of primary commodities, were low while both were high for manufactured items. He also found that GL index was higher for those industries where export was more than import. Another interesting finding was India's IIT was increasing with more dissimilar countries and therefore he concluded that India's IIT is vertical in nature.

In another study by Veeramani (2003), where he analyzed the effects of liberalization on India's IIT, he found that liberalization not only promoted IIT but also helped in trade expansion, i.e., both export and import increased and domestic industries were not go out of business. He further speculated that, in future, IIT will gain even greater significance in India's international trade.

Burange and Chaddha (2008) have also conducted a study on India's IIT for 19 years period from 1987-88 to 2005-06.They worked on 4-digit ITC level and calculated GL index for both multilateral IIT as well as IIT with respect to 7-country groups (defined by World Atlas).They found that at multilateral level, India's IIT grew from 23.48 to 32.09 between 1987-88 to 2005-06 respectively with an 1.86 per cent of ACGR (annual compounded growth rate). Their findings were somewhat similar to that of Veeramani (2003) that India's IIT is morewith dissimilar economies, they found it is maximum with Europe and Asia but interesting thing was that they found that as far as growth of India's IIT is concerned it is maximum with Middle East and African countries, which were more similar in nature. They also analyzed the contribution of different industries in India's IIT. According to them, Natural Pearls, Machinery and Transport Equipments stifled the growth of IIT while Manufactures contributes maximum along with Mineral Products and Base Metals.

Although all these works contributes significantly in understanding the nature of India's IIT but the major limitations of these works (related to India) are two folds: first, they have been studied at high aggregation level (4-digit) and second, they have not separated horizontal and vertical intra-industry trade specifically. This paper is taking care of these limitations, as already mentioned in Section I, i.e., Introduction.

III. METHODOLOGY

(a) Measuring Intra-industry Trade

Although there are different measures of calculating intra-industry trade are there in literature, but still the most widely used is the Grubel-Lloyd (GL) index. We also prefer to use GL index to any other, because of its wide acceptability. Grubel and Lloyd (1971) were among the first economists to measure the extent of intra-industry trade. They measured IIT as the percentage of country's total trade (exports plus imports) which was matched or balanced, that is exports equal imports. For individual product group or industry i, the formula for measuring IIT (GL index) is as follows:

[IIT.sub.i] = [X.sub.i] + [M.sub.i] - [absolute value of [X.sub.i] - [M.sub.i]]/([X.sub.i] + [M.sub.i]) x 100

OR

[IIT.sub.i] = 1 [absolute value of [X.sub.i] - [M.sub.i]]/([X.sub.i] + [M.sub.i]) x 100 (1)

where [X.sub.i] and [M.sub.i] represents the exports and imports of product group i respectively and HT is the degree of intra-industry trade for product group or industry i. In this case, closure the value to 100, greater will be the degree of IIT, closure the value to 0 represents higher inter-industry trade.

To obtain the average level of IIT for a country (IIT) they proposed a weighted measure of IIT which can be written as:

[IIT.sub.j] = [n.summation over (i=1)] ([X.sub.i] + [M.sub.i]) - [n.summation over (i=1)] [absolute value of [X.sub.i] --[M.sub.i]]/[n.summation over (i=1)] ([X.sub.i] + [M.sub.i]) (2)

where n is the number of industries at a chosen level of aggregation.

One problem with [IIT.sub.j] (2) is that it makes no allowance for any imbalance in a country's total trade and is a downward biased measure of IIT in case of an imbalance. The greater the imbalance lowers the value of IIT. To avoid this problem, they .proposed an alternative formula, which adjusts for any trade imbalance, as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

The above equation (3) is called as an adjusted measure at the chosen level of aggregation.

(b) Disentangling Horizontal and Vertical Intra-industry Trade

Total IIT can be disentangled into horizontal IIT (HIIT) and vertical IIT (VIIT), as follows:

IIT = HIIT + VIIT (4)

While to measure whether the IIT is horizontal or vertical, following equation is used:

1 - [alpha] [less than or equal to] [UV.sub.exp]/[UV.sub.imp] [less than or equal to] 1 + [alpha] (5)

where [UV.sub.exp] and [UV.sub.imp] are unit-value of export and import respectively. There is no imp obvious value for a (dispersion factor), but in most of the research 0.15 or 0.25 value is used. In this paper, we prefer to use 0.15, which is the most frequently used value. Therefore, assuming the value of a to be 0.15, if the above condition (equation 5) holds valid, the IIT will be considered as HIIT, otherwise VIIT. Moreover, according to Greenway et al (1994 and 1995), when relative the unit value of exports relative to the unit value of imports lies within the range of [+ or -] 15 %, the IIT will be considered as HIIT, while when the relative unit values are outside the range, the IIT will be considered as VIIT. Thus, IIT will be considered to be VIIT, when it satisfies the following condition:

[UV.sub.exp]/[UV.sub.imp] < 1 - [alpha] or [UV.sub.exp]/[UV.sub.imp] < 1 + [alpha] (6)

In addition to this, VIIT can further be classified into low quality (LQVIIT) and high quality (HQVIIT). The first part of equation 6 represents the condition of LQVIIT while second part represents HQVIIT.

Apart from this, unit values of exports and imports are calculated as:

[UV.sub.exp] = [Value.sub.exp]/[Units.sub.exp] and [UV.sub.imp] = [Value.sub.imp]/[Units.sub.imp] (7)

where values are measured in terms of dollars and units are in terms of selling unit like kg, tones etc. It is mentioned precisely in the next section of the paper.

(c) Calculating Import-weighted Average Tariff

To study the effects of changing tariff rate on intra-industry trade, we have calculated the import-weighted average tariff (IWAT). Here we have not taken into consideration of simple average tariff because although it measures the overall degree of protection but it does not adjust for significance of different products in the trade profile, so a high tariff on an insignificant product may overstate the degree or protection. On the other hand, weighted average tariff (here import-weighted average tariff) takes into the account of each product in the import profile of the economy under consideration; however its major limitation is, it understate the level of protection because a very heavily protected product will be imported less and therefore receive a small weight, but in the paper it is considered as a tool to understand the pattern of overall tariff change, specially after liberalization. The paper takes into account overall IWAT because the IIT considered over here is on total trade basis. To calculate IWAT, the formula used is:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)

where 4 and 6 stands for 4-digit and 6-digit HS-level, j stands for a country, w stands for weighted average of tariff, m represents the import value and n is the total number of products imported at that level. First part of Equation 8 is import-weighted average tariff at 4-digit while second part represent it at 6-digit. Now this equation will be used to study the pattern of changing tariff level with the advent of liberalization process and then to correlate the changing tariff level with IIT.

IV. DATA ANALYSIS

The present paper is based on the detailed study of India's intra-industry trade (IIT) with world. The data have been taken from UNCOMTRADE available on the website of WITS. The data taken is both 4-digit and 6-digit HS-classification (Harmonized System classification 1988-92), covering the period from 1988-2008. As mentioned in the Section I of the paper, the objective of the paper is to tests few hypotheses regarding India's intra-industry trade, both at aggregated (4-digit) and disaggregated (6-digit) level. These are mentioned again here:

First: Degree of India's IIT (both at aggregated (4-digit) and disaggregated (6-digit) level) increases with the advent of liberalization process.

Second: In India's IIT (both at aggregated (4-digit) and disaggregated (6-digit) level), contribution of export is more than import.

Third: India's overall IIT, (both at aggregated (4-digit) and disaggregated (6-digit) level), with world is vertical in nature.

After testing these hypotheses, we shall conduct two more studies to have better understanding of the effect of liberalization on IIT. In the first study, we shall discuss the effect changing tariff level on IIT and try to find out that is there any correlation between them, and in the second study we shall discuss the pattern of IIT, HIIT and VIIT with different product groups.

Now we shall test all the three above mentioned hypotheses one by one. To test the first hypothesis, first we have taken out the total trade (export plus import) figure of India for the last 21 years for the above mentioned period, 1988-2008, both at 4-digit and at 6-digit. Although our objective is to test the variation of GL-index specially after the start of liberalization process in India, i.e., 1991 onwards, but we have taken the data from 1988 just to compare the situation. From the total trade figure, we have separated out those 4-digit and 6-digit group in which intra-industry trade have taken place. All those group, both at 4-digit and 6-digit level, in which intra-industry trade have not taken place at all have not been included in the calculation of GL-index, similarly if in any group intra-industry trade have not been taken in a particular year, then for that year that particular group have not been included in the calculation of GL-index. Now GL-index, for the total trade, have been calculated for both 4-digit and 6-digit level. To calculate the index for the total trade, we have used both unadjusted IIT (equation 2) and adjusted IIT (equation 3). As have already been discussed in Section II of the paper, the value of unadjusted is somewhat lower that the value of adjusted GL-index. Then we plot our result on graph to understand the trend in better way.

[FIGURE 1 OMITTED]

Figure 1 shows the pattern of GL-index at 4-digit and Figure 2 shows the pattern for 6-digit. According to Fig 1, although the percentage contribution of 4-digit intraindustry trade to total trade have increased significantly from 1988 (78%) to 2008 (96.9%), but till the year 2000 the pattern was not stable while from the year 2001 onwards it has increased significantly but the pattern is almost stable. This shows the contribution of intra-industry trade is very high (more than 95 per cent) in the total trade, which signifies the importance of IIT. On the other hand, if we look at the pattern of 4-digit GL-index, both unadjusted and adjusted, we find that both values have decreased in 2008 as compared to their values in 1988. The unadjusted 4-digit GL-index value have decreased from 35.1% to 31.1% from 1988 to 2000 respectively while for the same time period the adjusted 4-digit GL-index value have decreased from 44.8% to 42.1% respectively. However both the values have increased a little bit after the year 2001.

Similarly, it is clear from the Fig 2 that percentage contribution of 6-digit intraindustry trade have increased significantly from 53.1% in the year 1988 to 96% in the year 2008. Apart from this, the value of 6-digit GL-index, both unadjusted and adjusted, have also increased, like for unadjusted the value have increased from 20.3% in 1988 to 25.2% in 2008 while for adjusted it has increased from 29.8% to 34% respectively for the same time period. Moreover, it is also clear from these two figures that as the aggregation level decreases, the value of GL-index will also decrease.

[FIGURE 2 OMITTED]

Therefore looking at the Fig 1 and 2, we can say that although in absolute sense, the contribution of intra-industry trade have increased significantly in the total trade of India but as far as the degree of intra-industry trade is concerned, it has not increased much, rather it has decreased a little at 4-digit level and slightly increased for 6-digit aggregation level. Thus if we look at our first hypothesis, we can say that it is not true for 4-digit aggregation level but fits with 6-digit aggregation level.

Now we shall test our second hypothesis that contribution of export is more in India's intra-industry trade, both at 4-digit and 6-digit aggregation level. To test this, we have taken out the total export and total import figure of only those groups (4-digit and 6-digit both), where IIT have been taken place. Then we divided the total export and total import figure with India's total trade figure (total export plus total import) and then we plot the result on graph to understand the pattern. Figure 3 and 4 shows the contribution of export and import in India's IIT at 4-digit and at 6-digit aggregation level respectively.

[FIGURE 3 OMITTED]

Figure 3 shows that the pattern of contribution of export and import in the India's total intra-industry trade at 4-digit aggregation level. The figure shows that in 1988, the percentage contribution of export was less than that of import and the situation persist till 1992. From 1993 onwards till 2000, its contribution has increased, but not significantly, and finally from the year 2001 onwards, the contribution of import in total intra-industry trade has surpassed the contribution of export.

[FIGURE 4 OMITTED]

Now we shall test our third hypothesis that India's IIT is vertical in nature. To disentangle intra-industry trade into horizontal (HIIT) and vertical (VIIT), we have used equation 5, discussed in Section III.B, and dispersion factor ([alpha]) used is 0.15. VIIT is further categorized in high quality (HQVIIT) and low quality (LQVIIT) for which we have used equation 6. Further, equation 7 is used to calculate unit value of exports and unit value of imports. It is important to mention here that all the data have been taken from WITS web site, although most of the information was available but few information was missing (mostly unit values) and therefore we were not able to calculate all the values.

Figure 5 and 6 shows the level of HIIT and VIIT, as well as LQVIIT and HQVIIT for the period of 1988 to 2008, at 4-digit and 6-digit level respectively. From figure 5 it is clear that, at 4-digit level, VIIT dominates HIIT but the interesting thing is that, over the period HIIT have increased from 9.4 per cent (out of the total products in which IIT is observed) in 1988 to 13.6 per cent in the year of 2008, while for the same time period VIIT have decreased from 90.6 per cent to 86.4 per cent respectively. HIIT was maximum in the year 2006 and its value was 18.4 per cent, after then its value decreased; just opposite to that VIIT was minimum in that year and its value was 86.4 per cent after then its value increased. Apart from this, the figure also shows that over the years the LQVIIT have decreased while HQVIIT increased. In 1988, out of the total products in which IIT was observed, in 64.8 per cent of products LQVIIT was seen while in 25.8 per cent of products, HQVIIT was observed (remaining 9.4 per cent, obviously HIIT was seen). But in the year 2008, the percentage of LQVIIT has decreased to 54.9 per cent and HQVIIT has increased to 31.5 per cent (remaining 13.6 per cent was HIIT).

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

Almost same type of pattern is seen at somewhat higher disaggregation level, 6digit, which is shown in figure 6. Here also VIIT dominates over HIIT, as well as, contribution of LQVIIT is more than HQVIIT. But over the periods, LQVIIT has decreased and HQVIIT has increased. The figure also shows that, although VIIT is more, but over the years it has deceased little bit as compared to HIIT. Therefore, both figure 5 and 6 proves that, in general, India's IIT is vertical in nature.

Now we shall study the changing pattern of tariff rate and its effect on IIT. The extent of liberalization process can also be studied by looking at the pattern of tariff rate Figure 7 shows the pattern of changing import-weighted average tariff of India since the beginning of liberalization process. It is not continuous because of unavailability of data. Theoretically, liberalization of an economy means decrease in the tariff rate and looking at Figure 7, we can say that overall tariff rate have decreased drastically after beginning of liberalization process. In 1992, it was 27.83% and 27.82% while in 2008 it became 6.27% and 6.12% respectively for 4-digit and 6-digit. Although it increased a little bit in 1999 but thereafter its value decreased continuously. Therefore we can say that liberalization process have led to decrease in the tariff rate. On decreasing the tariff rate, import should increase and hence ideally IIT should also increase.

So here we can form this hypothesis that on decreasing the tariff rate IIT should increase, i.e., tariff rate and HT should be negatively correlated. Here we shall test this hypothesis by calculating the correlation coefficient between tariff rate and total IIT, as well as different components of IIT as discussed in the paper. Since the paper is based on the total IIT of India with respect to World, therefore the import-weighted average tariff rate have also been taken with respect to World and on total basis, i.e., it includes all the commodities and all the countries.

Here it is pertinent to mention that since tariff data was not available for all the twenty-one years (1988-2008) discussed in the paper, therefore the analysis was done on the basis of available data on tariff rate (nine years only) but this data is sufficient to give information desired here. Since our study is based on both 4-digit and 6-digit level therefore we have calculated import-weighted average tariff rate at both of these two levels, but interestingly, as shown in the Figure 7, the IWAT at both of these two levels are almost same. Even then, just to keep things separate for 4-digit and 6-digit level, we have calculated the correlation coefficient between tariff rate and IIT separately for 4-digit and 6-digit, the result of correlation study is shown in the Table 1.

[FIGURE 7 OMITTED]

The results are quite interesting. Looking at Table 1, we can say that at 4-digit, unadjusted total IIT has not increased even after decrease in the tariff rate while adjusted 4-digit IIT has increased but not significantly because its correlation coefficient is very low. At the same level of aggregation, it is actually horizontal IIT which increased with the advent of liberalization and not VIIT. The trend is almost similar at 6-digit aggregation level, which shows that, all the three - unadjusted, adjusted and HIIT, have increased with decreasing tariff rate while VIIT has decreased, but for total IIT, correlation coefficient is not very strong.

But if we interpret the result of these correlation coefficients with the help of tvalue, it adds another dimension in our analysis. Since "tabular t-value" at significance level of p=0.05 and for 7 degree of freedom (because total number of observation, i.e., years, is 9) is 2.37, therefore if we compare absolute value of"calculated t-value" with "t-tabular" we shall find that the only correlation coefficient which is significant is 6digit HIIT and 6-digit VIIT. Hence, we can say, with 95 per cent confidence, that on decreasing the tariff rate 6-digit HIIT increases while 6-digit VIIT decreases. On the other hand, t-calculated of 4-digit HIIT and 4-digit VIIT, although not significant, but very close and if we decrease our confidence level from 95 per cent to 90 per cent (or significance level p--0.1), then it will become significant because at that level value ofttabular is 1.895. Therefore we can say that correlation coefficient between tariff rate and, 4- and 6-digit HIIT & VIIT is significant while it is not significant for 4- and 6-digit total IIT, both adjusted and unadjusted. These results support our earlier findings.

Another study was conducted to get deeper understanding of the nature of IIT, HIIT and VIIT on different product categories. Here the total products, according to ITC HS classification, is categorized into three categories - Group A consists of those products where product differentiation is very less, it is basically from Section 1 to Section 5 or Chapter 1 to 27 of ITC HS Classification; Group B consists of those products where product differentiation is moderate, it is from Section 6 to Section 15, i.e., Chapter 28 to 83 (of ITC HS Classification); and Group C consists of those products where high product differentiation is possible, it ranges from Section 16 to Section 21 or Chapter 84 to 98 (of ITC HS Classification). We have calculated adjusted GL-IIT, HIIT and VIIT for all three groups; the result is shown in Table 2.

According to Table 2, 4-digit and 6-digit adjusted GL-IIT is same as expected theoretically that IIT of Group A products should be lower and Group C products should be higher. Interestingly, in some cases IIT at 6-digit level is higher than 4-digit level, this is contradictory to the theoretical concepts. Another very interesting result is shown in the Table 2 which is again contradicting the theory that although Group A product is basically primary products and Group C is manufactured, so HIIT should be minimum in the former and should be maximum in the latter case, but the actual result is just opposite. Although this result is not supporting the theory, but it is matching with our earlier findings that India's IIT is mainly vertical in nature.

V. RESULTS AND DISCUSSION

After analyzing the data in the last section, we can conclude our study on India's intraindustry trade during 1988 to 2008, in the following few points, these are:

--Amount wise, contribution of intra-industry trade in total trade have increased; this shows the importance of intra-industry trade for India.

--Value of GL-index, for India's total trade, has decreased for 4-digit aggregation level, but it has increased slightly for 6-digit aggregation level; this shows that even with the advent of liberalization, degree of intra-industry trade has not increased significantly as expected, so this finding does not support our first hypothesis that with the advent of liberalization process, India's IIT has increased.

--Value of GL-index decreases as the aggregation level decreases, i.e., 4-digit to 6-digit: this shows the expected trend that as data is going to be more disaggregated, value of GL-index should decrease.

--Contribution of imports in total IIT has increased over the years while for exports, it has decreased: this finding does not support our second hypothesis that contribution of export is more in India's IIT.

--Basically India's IIT is vertical in nature but over the years VIIT has decreased while HIIT has increased little bit; this result although supports our third hypothesis that India's IIT is basically vertical in nature, but give some new insights into it.

--In vertical IIT, although low-quality VIIT (LQVIIT) is dominating over highquality VIIT (HQVIIT), but interestingly over the years contribution of LQVIIT has decreased while HQVIIT has increased; this may be because of the fact that India's trade is more with dissimilar countries.

--Apart from these results, our last two studies also supports that the degree of IIT has not increased, as significantly as expected, even after liberalization. Moreover although India's IIT is mainly vertical in nature, but in reality VIIT decreases with decrease in tariff rate, i.e., with the advent of liberalization process. Even in manufactured products, where product differentiation is high, VIIT dominates over HIIT.

If we look at the results of the three hypotheses and two studies and try to conclude, then we can say that in general the nature and pattern of India's IIT is in consonance with the theory. The deviations in our result can be explained with the fact that India is a developing country and most of the works on IIT, discussed in the paper, have been done with the trade data of developed countries. For example, Grubel (1967) used the data of EEC, Grubel and Lloyd (1971) used Australian trade data, Greenaway and Milner (1981, 1983) used UK's trade data, Bergstrand (1990) used OECD and Abd-elRahman (1991) used French trade data; and the level of economic development of a country will also have a significant impact on IIT of the country. As Linder stated that (Hanink 1988) "although the trade between countries with similar income level is no different than intraregional trade but the high level of trade between similar but poor countries is unlikely. Poor countries have domestic economies that are not sufficient to generate any significant number of potential exports, nor are poor domestic economies large enough to appeal many foreign producers". We found that although the percentage of IIT in total trade is high but the degree of IIT is moderate, this is because India is a developing country. This has also been showed by Havrylyshyn and Civan (1985), and Clark and Stanley (1999) that the degree of IIT of developing countries will be lower than that of developed one. We also found that India's IIT is basically vertical in nature this is because most of the India's foreign trade is with developed countries or high income countries (USA, EU, China etc.), and we have already discussed that larger the difference in factor-endowment, larger will be VIIT or inter-industry trade. In the last study, we found that even in manufactured products, HIIT is not high and in-fact VIIT dominates HIIT, this finding can also be justified with the fact that India is a developing country and therefore demand for horizontally differentiated goods will be lower than demand for goods differs in quality and price significantly, i.e., vertically differentiated. This is because higher demand for horizontally differentiated goods itself depict that consumers of the country love different varieties and this will be possible only when per capita income of the country is high. Since India is a developing country, as well as income distribution is not even, therefore variety- seeking consumers are less in number which resulted into lower HIIT. Thus we can explain the reason why India's intraindustry trade is basically vertical in nature.

VI. CONCLUSION

In this paper, we have tried to analyze India's intra-industry trade with world from different perspectives and interestingly, the results of the two out of the three hypotheses we have considered here, are different from some previous works done for India (as discussed in the paper). The nature and pattern of India's intra-industry trade matches with the theoretical underpinnings of nature and pattern of IIT of a developing country. The scope of the paper can further be expanded by analyzing cross-country-as well as cross-sector trend of India's IIT along with the tariff rate for different countries/sectors. This will provide some finer details of India's IIT. Another limitation of the paper is that it calculates IIT on total basis while it may be possible that the product is getting imported from one country and getting exported to other but this thing cannot be reflected in the paper so it may leads to different result. We hope that our work provokes some further research in this direction.

References

Abd-el-Rahman, K. (1991), "Firms' Competitive and National Comparative Advantages as Joint Determinants of Trade Composition", Review of World Economics, 127, pp. 83-97.

Andresen, M. A. (2003), "Empirical Intra-Industry Trade: What We Know and What We Need to Know", available on: "http://www.sfu.ca/~andresen/ papers/Empirical_IIT_Lit_Review.pdf", dt 5th September 2009.

Balassa, B. (1966), "Tariff Reductions and Trade in Manufactures among Industrial Countries", American Economic Review, Vol. 56, June, pp. 466-473.

Balassa, B. and Bauwens, L. (1987), "Intra-Industry Specialisation in a Multi-Country and MultiIndustry Framework", The Economic Journal, 97, December, pp. 923-939.

Bergstrand, J. H. (1990), "The Hecksher-Ohlin-Samuelson Model, The Linder Hypothesis and the Determinants of Bilateral Intra-Industry Trade", The Economic Journal, Vol. 100, No. 403, December, pp. 1216-1229.

Borkakoti, J. (1998), International Trade: Causes and Consequences, Palgrave Macmillan.

Brulhart, M. (2008), "An Account of Global Intra-Industry Trade, 1962-2006", Research Paper Series Globalisation, Productivity and Technology, Research Paper No. 2008/08, The University of Nottingham (available on "http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1103442" dt. 5th September 2009).

Burange, L. G. and Chaddha, S. (2008), "Growth in India's Intra-Industry Trade", Working Paper UDE24/2/2008, Department of Economics, University of Mumbai, March 2008.

Cabral, M., Falvey, R. and Milner, C. (2008), "Endowment Differences and the Composition of IntraIndustry Trade", Research Paper Series, Research Paper No. 2008/36, The University of Nottingham.

Clark, D. P. and Stanley, D. L. (1999), "Determinants of Intra-Industry Trade Between Developing Countries and the United States", Journal of Economic Development, Vol. 24, No. 2, December, pp. 79-95.

Davis, D. (1991 ), "Explaining The Volume of Intraindustry Trade: Are Increasing Returns Necessaary?", International Finance Discussion Paper, No. 411, September.

Dixit, A. K. and Stiglitz, J. E. (1977), "Monopolistic Competition and Optimum Product Diversity", American Economic Review, Vol. 67, No. 3, June, pp. 297-308.

Falvey, R. E. (1981), "Commercial Policy and Intra-Industry Trade", Journal of International Economics, 11, pp. 495-511.

Falvey, R. E. and Kierzkowski, H. (1987), "Product Quality, Intra-industry Trade and (Ira)perfect Competition", in H. Kierzkowsi ed. Protection and Competiton in International Trade, Basil Blackwell, New York.

Fontagne, L., Freudenberg, M. and Gaulier, G. (2005), "Disentangling Horizontal and Vertical IntraIndustry Trade", CEPII Working Paper, No 2005-10 July.

Greenaway D. and Milner, C. (1981), "Trade Imbalance Effects in the Measurement of Intra-Industry Trade", Weltwirtschaftliches Archiv, 117, pp. 756-762.

--.(1983), "On the Measurement of Intra-Industry Trade", The Economic Journal, Vol. 93, No. 372, December, pp. 900-908.

Greenaway, D., Hine, R. and Milner, C. (1994), "Country Specific Factors and the Pattern of Horizontal and Vertical Intra-Industry Trade in the UK", Weltwirtschaftliches Archiv, 130, 77-100.

--(1995), 'Vertical and Horizontal Intra-Industry Trade: A Cross Industry Analysis for the United Kingdom", The Economic Journal, 105 (November), pp. 1505-18.

Grimwade, Nigel (2000), International Trade: New Patterns of Trade, Production and Investment, Routledge.

Grubel, H. G. (1967), "Intra-Industry Specialization and the Pattern of Trade', The Canadian Journal of Economics and Political Science, Vol. 33, No. 3, August, pp. 374-388.

--and Lloyd, P. J. (1971), "The Empirical Measurement of Intra-Industry Trade", The Economic Record, 47, December, pp. 494-517.

Gullstrand, J. (2002), "Does the Choice of Intra-Industry Trade Measure Matter?", (taken from website "http://www.snee.org/ publikationer_show.asp?id=99" on 5th September, 2009; Later on this paper was published as: "Does the measurement of Intra-Industry Trade Matter?", Weltwirtschaftliches Archiv, 138, 2002, pp. 317-39).

Hanink, D. M. (1988), "An Extended Linder Model of International Trade", Economic Geography, Vol. 64, No. 4, September, pp. 322-334.

Havrylyshyn, O. and Civan, E. (1985), "Intra-industry Trade among Developing Countries", Journal of Development Economics, Vol. 18, Issues 2-3, August, pp. 253-271.

Kinnerup, K. (2005), "Country-Specific Determinants of Intra-Industry Trade- The Case of France", Department of Economics, School of Economics and Management, Lund University, August 2005 (available on: http://www.lu.se/ o.o.i.s?id=19464&postid=1334655, dated: 5th September 2009).

Krugman, P.R. (1979), "Increasing Returns, Monopolistic Competition and International Trade", Journal of International Economics, 9, pp. 469-479.

--(1980), "Scale Economies, Product Differentiation and Pattern of Trade", The American Economic Review, Vol. 70, No. 5, pp. 950-959.

--(1981), "Intraindustry Specialization and Gains from Trade", Journal of Political Economy, Vol. 89, No. 5, pp. 959-973.

Lancaster, K. (1980), "Intra-Industry Trade under Perfect Monopolistic Competition", Journal of International Economics, 10, pp. 151-175.

Marrewijk, Charles van (2002), International Trade and the World Economy, Oxford University Press.

Mikic, M. and Gilbert, J. (2008), Trade Statistics in Policy Making -A Handbook of Commonly Used Trade Indices and Indicators, Economic and Social Commission for Asia and The Pacific, United Nations Publication (available on: http//:www.unescap.org/tid/publication/tipub2491_pdf, dated 21st November 2009).

Shaked, A. and Sutton, J. (1983), "Natural Oligopolies", Econometrica, Vol. 51, No.5, (September), pp. 1469-1483.

Sutton, J. (1986), "Vertical Product Differentiation: Some Basic Themes", The American Economic Review, Vol. 76, No. 2, Papers and Proceedings of Ninety Eighth Annual Meeting of the American Economic Association (May 1986), pp. 393-398.

Veeramani, C., (2001) "India's Intra-Industry Trade Under Economic Liberalisation: Trends and Country Specific Factors", Working Paper No. 313, Centre for Development Studies, India (March).

Veeramani, C. (2003), "Liberalisation, Industry-Specific Factors and Intra-Industry Trade in India", Working Paper No. 97, ICRIER, India, March 2008.

We are deeply indebted to Dr. Biswajit Nag (IIFt, New Delhi, India) for his valuable suggestions. We received a great deal of help from him for preparing the article.

AMIT SRIVASTAVA AND YAJ MEDURY

Jaypee University of Information Technology, Solan, India Table 1 Correlation Coefficient between Import-Weighted Average Tariff Rate and level of IIT. both at 4-digit and 6-digit GL-IIT Adjusted GL-IIT 4-digit 0.398(1.148) -0.148(-0.395) 6-digit -0.415(-1.201) -0.377(-1.074) HILT VIIT 4-digit -0.628(-2.137) 0.628(2.137) 6-digit -0.724(-2.774) 0.724(2.774) Source: same as figure 7; (Figures in the brackets shows "calculated t-value" of the concerned correlation coefficient) Table 2 Adjusted GL-ITT, HIIT and VIIT for Different Product Groups Adjusted GL IIT index 4-digit level 6-digit level Year Gr A Gr B Gr C Gr A Gr B Gr C 1988 27.4 45.7 77.7 32.4 18.1 59.0 1989 33.9 45.5 76.5 48.2 18.8 63.0 1990 26.1 39.2 75.5 32.8 17.5 60.5 1991 21.2 40.1 63.7 27.5 16.8 46.5 1992 23.7 42.6 64.1 28.2 11.7 55.5 1993 22.8 47.1 61.2 26.1 13.8 51.6 1994 20.2 36.2 60.2 20.1 18.2 51.6 1995 18.0 35.7 65.8 22.2 17.8 56.1 1996 16.4 40.7 66.8 16.1 16.6 55.9 1997 14.5 36.9 72.1 13.5 17.9 61.5 1998 10.5 45.8 73.4 10.7 18.9 60.5 1999 12.7 52.1 74.5 11.9 19.3 62.7 2000 31.8 58.4 74.2 33.5 22.9 63.8 2001 19.5 43.1 76.1 18.6 22.9 65.4 2002 24.5 51.5 74.8 23.2 26.0 62.8 2003 23.6 49.4 68.8 22.2 24.3 58.5 2004 27.1 46.4 71.7 25.8 24.1 61.5 2005 28.0 44.9 72.0 26.1 28.5 62.9 2006 24.1 42.2 75.5 23.3 25.0 66.6 2007 25.0 39.7 81.0 24.2 23.7 70.2 2008 26.7 41.2 78.6 25.9 28.4 70.3 Horizontal IIT (percentage) 4-digit level 6-digit level Year Gr A Gr B Gr C Gr A Gr B Gr C 1988 6.0 11.3 6.6 8.6 13.1 3.6 1989 9.8 12.6 5.1 7.8 11.7 4.6 1990 10.7 12.9 4.6 10.3 11.3 4.1 1991 9.7 11.5 4.0 7.6 11.2 3.0 1992 8.7 11.2 3.4 7.0 10.2 3.9 1993 11.1 9.6 7.6 8.8 11.6 3.9 1994 9.1 12.6 4.7 8.9 12.5 4.0 1995 10.2 12.4 5.1 10.2 12.7 5.3 1996 10.6 13.0 5.6 9.3 12.8 6.0 1997 11.5 15.0 6.0 13.5 13.6 6.6 1998 13.3 14.3 6.3 11.5 14.2 6.0 1999 10.4 13.3 7.9 9.3 14.2 6.4 2000 8.1 12.2 5.3 7.5 12.7 6.3 2001 17.6 12.5 11.1 12.3 11.5 6.5 2002 12.1 12.4 6.0 11.8 14.0 6.9 2003 10.5 13.1 7.2 11.3 18.6 7.1 2004 13.3 16.3 12.9 10.2 17.2 15.2 2005 12.4 19.9 12.2 11.7 18.1 15.4 2006 17.2 20.5 10.6 14.3 21.7 9.0 2007 14.1 16.3 9.9 12.3 19.3 7.4 2008 9.1 17.0 6.5 12.3 17.8 7.4 Vertical IIT (percentage) 4-digit level 6-digit level Year Gr A Gr B Gr C Gr A Gr B Gr C 1988 94.0 88.7 93.4 91.4 86.9 96.4 1989 90.2 87.4 94.9 92.2 88.3 95.4 1990 89.3 87.1 95.4 89.7 88.7 95.9 1991 90.3 88.5 96.0 92.4 88.8 97.0 1992 91.3 88.8 96.6 93.0 89.8 96.1 1993 88.9 90.4 92.4 91.2 88.4 96.1 1994 90.9 87.4 95.3 91.1 87.5 96.0 1995 89.8 87.6 94.9 89.8 87.3 94.7 1996 89.4 87.0 94.4 90.7 87.2 94.0 1997 88.5 85.0 94.0 86.5 86.4 93.4 1998 86.7 85.7 93.7 88.5 85.8 94.0 1999 89.6 86.7 92.1 90.7 85.8 93.6 2000 91.9 87.8 94.7 92.5 87.3 93.7 2001 82.4 87.5 88.9 87.7 88.5 93.5 2002 87.9 87.6 94.0 88.2 86.0 93.1 2003 89.5 86.9 92.8 88.7 81.4 92.9 2004 86.7 83.7 87.1 89.8 82.8 84.8 2005 87.6 80.1 87.8 88.3 81.9 84.6 2006 82.8 79.5 89.4 85.7 78.3 91.0 2007 85.9 83.7 90.1 87.7 80.7 92.6 2008 90.9 83.0 93.5 87.7 82.2 92.6 Source: same as figure 1