Loading...

Dynamics of Increased Trade of a Large Economy

Intensive and Extensive margins effects to the Export growth

Research Paper (postgraduate) 2010 30 Pages

Economics - International Economic Relations

Excerpt

Table of contents

1. Summary

2. Brief Literature Review

3. Methodology

Decomposition in relation to exports

Econometric equations and formulation

4. Data

5. Results

Flow of exports

DC’s Group

OECD Group

Flow of imports

6. Conclusion

7. Appendixes

8. References

Dynamics of Increased Trade of a Large Economy

Contributions of intensive and extensive margins

Econometric Analysis : Section I

Author : Kerfalla Conté

17/12/2010

In this document, we have explored the extent to which the increase in exports or imports of a large economy with its trading partners appears in the intensive and extensive margins.

A country that doubles its revenue would export twice as much, and the growth of the level of its exports would arise more from the extensive margin (proliferation of new products and/or new markets). The same country would also import twice as much, and the increase of its imports would stem more from the intensive margin (increase in volume of products that are already exchanged). In addition, the contributions of intensive and extensive margins to the growth of exports vary according to the economic group or the level of development of a country. Indeed, the growth of the level of exports of a developed country that doubles its revenue comes more substantially from the intensive margin, whereas for a developing country that would double its revenue, the growth of level of exports would more likely arise from the extensive margin.

Abstract

This document explores the cross-sectional differences in export/import across the trading partners for a large economy, in terms of intensive and extensive margin, and investigates the extent to which the contribution of the margins to the cross-sectional differences evolves by year and by group of economies. Results reveal that an economy that is twice larger exports two times more and most of the variations in its export occur predominantly at the extensive margin, whereas the intensive margin accounts largely in the variations in its imports. However, a large developed country does export at the intensive margin, whereas the extensive margin account for a large part of the variations in a large (by doubling GDP PPP) for developing country’s export.

For your comments and suggestions: conte@intracen.org

The opinions and comments expressed in this document are exclusively those of the author. Consequently, they do not represent at any time the views of the ITC, its executive directors, or the countries it represents.

1. Summary

If the theoretical assessment according to which “trade of bigger countries are more intense than that of smaller countries” seems obvious, the question of ‘how the growth of its exchanges are carried out’ has been a rarely exploited problem for a long time. Academic works of these ten past years have allowed highlighting two dimensions (intensive and extensive) as a dynamic engine of increased trade for an economy that is open to international trade.

Moreover, those works allow a better understanding of the development of exports’ structure for countries having different levels of development.

In 2002, David Hummels and Peter J. Klenow were able to formalise and quantify both dimensions through which operates the growth of trade of a large economy. The aim of their research was to address the following problem: the growth of trade of an enlarging economy is characterised by:

- an increase in volume of already-exchanged products on traditional markets (intensive margin)?

- an increase of new products to exchange on traditional markets or towards new destinations (extensive margin)?

The results of the study conducted by these two authors revealed that, for a large economy, the extensive margin contributes the most (as much as two thirds) to the growth of its exports. Concerning the imports, it is the intensive margin that contributes the most to its growth.

The authors also demonstrated that large economies exported more products with high unit values.

This document’s objective is to verify, through an econometric analysis, if the conclusions of Hummels and Klenow (H-K) remain valid regardless of the year. In other words, it consisted in evaluating H-K’s conclusions in front of the implication of new rules for international trade, favouring the expansion of emergent countries.

For this reason, we have used the bilateral trade of thirteen consecutive years (1996-2009). Our results over-all confirm the conclusions of H-K (above). However, the distribution of contributions of extensive and intensive margins to the growth of exports varies according to the type of group a country belongs to. Indeed, we rediscover the conclusions of H-K while grouping all exporting countries, as well as developing countries (DC). On the other hand, concerning the OECD[1] developed countries and the fifteen-member EU, the distribution of margins has reversed on the flow of exports: the extensive margin represents the most important share of the growth of exports.

2. Brief Literature Review

It is commonly accepted in the theory of international trade that, all else being equal, “larger economies export and import much more than smaller economies”. And yet, certain divergences exist in the arguments put in advance by literature when talking of the evaluation of trade’s structure to understand how “larger economies are able to export and import much more”. Many economic models have been used to identify the characteristics of trade’s structure, by proceeding to the decomposition of total exports or imports of a certain country in an intensive margin (variation in volume of trade of a limited number of products towards the traditional partners) or extensive margin (variation of trade in number of new exchanged products or traditional products towards new markets).

In fact, in 1969, Armington starts from the idea that national differentiation highlights the intensive effect. Therefore, a country that would double its resources or it nationals national provision would have the capacity to export or import more. However, the country cannot exchange more in terms of products.

Moreover, in 1981, Krugman supported that, with the hypothesis of monopolistic competition, an economy that doubles its size could produce and export twice as much in terms of products.

In 1994, Romer’s simulations specially displayed the extensive margin of imports as imports’ fixed costs of each variety of products. In his model, he reaches the conclusion that large economies import a wide diversity of products originating from a large range of foreign suppliers.

In 2001, Head and Ries tried to distinguish the effects of increasing returns and of national differentiation of products, by supporting their study on trade between the United States and Canada. They examined if manufacturing industries produced more proportionally than what was demanded at the local level (just as in the increasing returns models) or proportionally less (just as in the differentiation models). In comparison, Hummels and Klenow have examined the implications of Head’s and Ries’ approach to the extensive margins (increasing returns) against the intensive margin (differentiation of national products), as well as their implications on the terms of trade.

In 2002, these authors reported[2] that the growth of exports of large or more productive countries stemmed more likely from the extensive margin, whereas that of their imports originating from the rest of the world was dominated by the intensive margin. Besides, they also showed that large countries exchanged more products at relatively high prices.

In addition, we know that, since 1995, the rules of trade have evolved and are still evolving with time in function of the orientations of economic policies. Such a temporal readjustment of the world economy’s structure does not impact H-K’s conclusions? This question summarizes the main motive of our work in the following document.

Thus, the aim of this study is not to establish the implications of extensive and intensive margins on the terms of trade, but rather to verify on each of the years of the period chosen the stability of the results of H-K “as to the distribution of contributions of the margins to the increase in the level of trade”.

The rest of this document is composed as follows: 3) the methodology; 4) the description of the data; 5) the econometric results; 6) Conclusion.

3. Methodology

We are using the approach developed by H-K in 2002 to characterize the exchange dynamics of economies over a period of given time. These dynamics stem from,

- a growth in volume of a limited set of products already exchanged (intensive margin)?

- an exchange of new products or existing products towards new markets (extensive margin)?

The size of an economy being defined by its weight on the world market, the decomposition of market share will be used to better identify the composition of intensive and extensive margins to the growth of trade. Market shares are used in the sole purpose of better decomposing exports into extensive and intensive margins.

Decomposition in relation to exports

At first, we break down the market share to the export of each country into two terms, intensive and extensive.

By definition, the market share to the exports of a country j is defined as follows:

illustration not visible in this excerpt

The decomposition of Abbildung in dieser Leseprobe nicht enthalten is resumed to the product of two terms such as:

illustration not visible in this excerpt

Indeed,

illustration not visible in this excerpt

The expressions (F2.1) and (F2.2) respectively represent the intensive and the extensive margins.

Remark 1

From (F2.1) and (F2.2), we notice that the intensive margin to the exports of a country measures its market share in worldwide demand, expressed by the elementary markets on which the country exports. Regarding its extensive margin to exports, it measures the weight of its elementary markets in world exports. Moreover, all else being equal, we notice that if a country j concentrated all its exports on a limited number of markets-products couples (i, s), it would have a strong intensive margin and a weak extensive margin to exports. On the other hand, if the country could progressively widen its exports to a larger number of markets-products couples, its exports would have a weak intensive margin in the favour of a more important extensive margin.

Remark 2

An alternative method exists for the calculation of intensive and extensive margins to exports such as defined in (F2.1 and F2.2). This method (also used by H-K), consists in counting up the products and markets-products couple for each country under consideration.

Econometric equations and formulation

The (F2.0) expression constitutes the base equation to determine or estimate the contributions of margins to the growth of the level of exchanges, over each group of countries and per year. The estimation method used in this section is a regression by the least squares method (LSM).

By applying the logarithmic transformation to the (F2.0) equation, we obtain an additive form:

LogAbbildung in dieser Leseprobe nicht enthalten= LogAbbildung in dieser Leseprobe nicht enthalten + LogAbbildung in dieser Leseprobe nicht enthalten (F2.0 bis)

In order to avoid an endogeneity bias, we regress each of the three variables on the GDP, calculated on the basis purchasing power parity (GDP[3] ppp). The estimator of the LSM , having a linear property and according to the equation (F2.0 bis), the marginal contribution of the GDP to the variation of market shares should equal the sum of the marginal contributions of the GDP to the variations of both margins taken individually.

Thus, the linear system to estimate is defined below, where each country represents an observation for a year and a group of selected countries.

LogAbbildung in dieser Leseprobe nicht enthalten= Abbildung in dieser Leseprobe nicht enthaltenAbbildung in dieser Leseprobe nicht enthaltenLogAbbildung in dieser Leseprobe nicht enthalten+ Abbildung in dieser Leseprobe nicht enthalten (F3.0)

LogAbbildung in dieser Leseprobe nicht enthalten = Abbildung in dieser Leseprobe nicht enthaltenAbbildung in dieser Leseprobe nicht enthaltenLogAbbildung in dieser Leseprobe nicht enthalten + Abbildung in dieser Leseprobe nicht enthalten (F3.1)

LogAbbildung in dieser Leseprobe nicht enthalten = Abbildung in dieser Leseprobe nicht enthaltenAbbildung in dieser Leseprobe nicht enthaltenLogAbbildung in dieser Leseprobe nicht enthalten + Abbildung in dieser Leseprobe nicht enthalten (F3.2)

Remark 3

The underlying assumptions to the equations above presume that:

a) The error termsAbbildung in dieser Leseprobe nicht enthalten,Abbildung in dieser Leseprobe nicht enthalten,Abbildung in dieser Leseprobe nicht enthalten, on all countries j are independent and fairly distributed, each one following a normal law.

b) We verify that the coefficients/elasticity are such as:

Abbildung in dieser Leseprobe nicht enthalten= Abbildung in dieser Leseprobe nicht enthalten+Abbildung in dieser Leseprobe nicht enthalten (F3.3)

Although exports are the only flow on which we carry most of our argumentation, the same arguments also apply on the flow of imports. We just need to replace the exports in the notations by the imports.

4. Data

To achieve the regressions defined by the equations above, we have used trade data from the database of the United Nations Statistics Division (UNSD), called COMTRADE. The GDP PPP[4] data are obtained from the IMF’s[5] database.

The period covered spans from 1996 to 2009. The classification of products’ codes for bilateral trade is that of the harmonized system of the 1996 revision (HS-1996).

Although most data-reporting countries are large economies, covering alone most worldwide trade (at least 80%), the coverage of data-reporting countries varies from one year to another and from the type of revision used. In order to benefit of a large coverage, given the reported data that is available in the harmonised system (HS), we have proceeded to a certain number of adjustments. Indeed, if among the UNSD data, some exist in a previous or posterior nomenclature to the HS-1996 revision, we proceed by doing a conversion by utilising an appropriate correspondences table. We thus avoid mixing products’ codes that would not correspond to the same product according to the chosen nomenclature.

Once the data conversion step is over, we start estimating the statistics of non-reporting countries (mirror statistics) from their corresponding trading partners’ statements, the latter having reported their trade statistics. This last step only comes as a complement to what precedes to increase the number of countries covered.

Finally, a category of countries is created in order to group a certain number of territories or entities not clearly defined elsewhere (TNA). Ultimately, we dispose of a bilateral database (in value and in volume) covering at least 180 countries and 200 destinations on at least 5000 positions with 6 figures (HS-6). Furthermore, certain countries have received a special treatment because of the modification of their country codes in the UNSD database (Montenegro and Serbia or Belgium and Luxembourg). In this case, we gather these countries or territories in one entity.

We have also conducted an analysis of the coherence of data by groups of products (HS-1996 with 2 figures) in order to estimate the inconsistencies rate between exporting countries’ statements and those reported by the importing partner countries. Particularly high inconsistencies rate have been observed in the flow of worldwide exports for the following groups: HS-27 (hydrocarbon sector) and HS-86 (transports equipment sector ‘locomotives tramways or rolling machines’).

In order to reduce the distortion of the trends from this data, we have withdrawn from the trade database the sectors HS-27 and HS-86, as well as ambiguous categories (territories TNA and products group HS-99). The removal of these entries from the base has reduced the covering of worldwide trade of our database. The evolution of worldwide trade covering rates (in %) is given below.

illustration not visible in this excerpt

5. Results

The comments of results developed in this section will concern only the flow of exports, and the main tables will solely be displayed. The remaining results and additional information can be consulted in the appendixes.

Flow of exports

Table 1 relates to the estimated coefficients (elasticity) of extensive and intensive margins from the regression of the size of economies (GDP) on each of the variables to explains, as defined by the equations (F3..). The indications of the tests’ statistics show that the regressions are overall significant at the tolerance threshold of 5% (see appendixes, table 4). The ‘GDP PPP’ explains at least 75% of the variations of the levels of markets’ share to exports of an economy. The consideration by the ‘GDP PPP’ of the variations of intensive and extensive margins to exports exceeds an average of 50% for the intensive margin and 60% for the extensive margin.

Because of the linearity property of estimators by the LSM, we verify that the sum of the coefficients of intensive and extensive margins is equal to the coefficient (total) or elasticity from the exports to ‘GDP PPP’. Afterwards, we only need to do the ratio of each margin’s coefficient to the total coefficient in order to get the growth contribution from the market’s share to exports of the intensive and extensive margins.

We notice that, irrespective of the year, the total coefficient is close to and slightly superior to the unit[6]. This suggests that a large economy has the capacity to export more on the international market. In other words, an economy that would double its size would export at least twice as much on the international market, and the increase of its exports would more likely stem from the extensive margin at the level of 60% on average.

Table 1: The set of exporting countries in the base (global level)

illustration not visible in this excerpt

The results above thus confirm the conclusion of H-K, namely:

- a country that doubles its revenue (GDP in USD) exports twice as much.

- and the increase of its exports stems more likely from the extensive margin at almost two-thirds (exporting more products and towards new markets).

To the above conclusion, however, we must note that the heterogeneity of the levels of development of countries constituting the worldwide can swamp certain imperceptible evidences at the global level (countries that are grouped in one entity). One must understand that the worldwide trade is made up of developed countries (OECD) and strong emergent economies (BRICS) on one hand, and developing countries, on the other hand, most of which are facing real difficulties to export in spite of a negligible GDP level.

In the first group, sufficiently developed economies could have almost reached the optimal diversification degree in terms of exported products, exchanging among each other similar products. In the second group, economies often exchange among each other very few products of the same type, and are often oriented towards the markets of developed or emerging countries. And that leaves a margin to be potentially exploited in the second group (table 22).

Between 1996 and 2009, the lowest level of the average number of exported products by the countries of the OECD is approximately twice as high as the maximal level reached by the DCs in average number of exported products (figure 1, figure 2). This shows a large diversification potential (at least in products) for DCs comparatively to OECD countries. It would be therefore interesting to analyze the composition of margins for these groups, taken independently.

From the results of estimations stands out a trend opposing the structure of imports of the groups of developing countries (DCs) to the OECD group. Indeed, we get the confirmation of H-K’s results on the group DCs. On the other hand, concerning the OECD group, an inversion of the margins’ importance appears. [7]

DC’s Group

The estimated coefficients reported in table 2 are all significant at the tolerance threshold of 5% (table 5 in appendixes, significance test). The variation of levels of market’s share to export is largely explained by the ‘GDP PPP’, at almost 70% on average. The variations of the extensive and intensive margins’ levels taken into account by the ‘GDP PPP’ respectively rise to 60% and 38% on average. The market’s share to exports to ‘GDP PPP’ is almost equal to the unit, irrespective of the year. This elasticity remains inferior to that obtained at the global level.

In brief, we reach the following conclusion:

- a country that would double its revenue (GDP in USD) would export twice as much.

- the increase of its exports would stem more from the extensive margin at almost two-thirds.

Table 2: The exporting countries of DCs

illustration not visible in this excerpt

Note : the conclusions on the DCs group do not change after the exclusion of BRICS countries. [8]

OECD Group

The coefficients obtained from the estimations’ result (table 3) are all significant at the risk threshold of 5% (see table 8 in appendixes). We notice that the variation of the levels of market’s share to exports is largely explained by the ‘GDP PPP’ at 75% on average.

We notice that the elasticity of the market’s share to exports to ‘GDP PPP’ is close to but remains inferior to the unit, irrespective of the year. This elasticity is inferior to the one obtained with DCs group, for in the latter, the elasticity is superior to the unit. This finding is very interesting, for it suggests that sufficiently developed countries, in relation to those that are developing, could have largely exploited their extensive margins over the years, and consequently be more inclined to specializing (re-focusing) increasingly in a certain number of products with high added-value.[9]

In short, according to the coefficients from table 3,

- an OECD country that would double its size (GDP in USD) would export more without being able to double its exports.

- the increase of its exports would stem more from the intensive margin at almost two-thirds on average.

Table 3 : The OECD countries

illustration not visible in this excerpt

Note : same results for 15-Member EU and OECD (except for Turkey, Mexico, Chili, South-Korea).

Flow of imports

As in the results obtained from the flow of exports, those of imports are all globally significant at the tolerance threshold of 5% (see appendix, tables 13 to 15). The elasticities of the market’s share to imports to ‘GDP PPP’ are close to and inferior to the unit.

The estimations’ results confirm the conclusions of Hummels-Klenow (including those obtained on the country groups OECD and DCs).

In brief, we reach the following conclusion:

- a country that would double its revenue (GDP in USD) would import almost twice as much.

- the increase of its imports would largely stem from the intensive margin at two-thirds at least on average.

6. Conclusion

The increase of the level of a country’s trade can be justified by the increase of the products’ demand initially exchanged with its trading partners or by the trading partners’ taste evolution for new varieties or new products. The intensification of these exchanges implies that, on one hand, the implicated traders in the transaction are able to produce and export/import more or able to sell or acquire products at increasingly competitive prices to their traditional partners (intensive margin). On the other hand, this intensification continues through the marketing of new product varieties to traditional partners, or through the redirection of existing or already-exchanged products towards new destinations (extensive margin).

Breaking down trade’s structure into intensive and extensive margins was initiated and elaborated by Hummels and Klenow (H-K) in 2002. These two authors have concluded from 1995 data that the increase of an increasingly large country’s exports grows to the extensive margin at almost two-thirds compared to the intensive margin, and that the imports rather increase to the intensive margin at almost two-thirds compared to the extensive margin.

In this document, we have examined the distribution of the intensive and extensive margins’ contribution to an increasingly large economy’s trade, as well as the evolution of this contribution in time.

Given the results obtained, it seemed to us that the distribution of the intensive and extensive margins' contribution of a larger economy to the growth of exports/imports does not operate in uniform or constant manner. Rather, this distribution varies according to the country group or a country’s development level, and evolves over the years.

The result we have reached is that an economy that doubled its revenue (in terms of GDP PPP) would be able to export twice as much towards the rest of the world, and that the growth of its exports would stem more from the extensive margin than the intensive margin.

However, the span of doubling an economy’s revenue remains relative to a group of economies. Thus, it was proved that the growth of exports of a country that was able to double its revenue in an OECD developed country would stem more from the intensive margin than the extensive margin. This finding that is opposite to Hummels’ and Klenow’s conclusions (2002) should be rather taken as a complement to their results instead of a counter-example to their affirmation.

Moreover, the results obtained for the flow of imports lead to the same conclusions of H-K: the proportions or contributions of the intensive margin to the growth of this flow are more important than that of the extensive margin.

7. Appendixes

Exports

Table 4: Estimations’ results for all covered exporting countries (global)

illustration not visible in this excerpt

Table 5: Estimations’ results on the group of developing countries (DCs)

illustration not visible in this excerpt

Table 6: Estimations’ results on the group of developing countries (DCs) (except for BRICS)

illustration not visible in this excerpt

Note: BRICS (Brazil, Russia, India, China and Singapore).

Table 7: Estimations’ results on the group of least developed countries (LDCs), exports

illustration not visible in this excerpt

Additional Information

illustration not visible in this excerpt

Table 8: Estimations’ results on the OECD group, exports

illustration not visible in this excerpt

Table 9: Estimations’ results on the country group 15-Memebr EU, exports

illustration not visible in this excerpt

Note: Belgium and Luxembourg grouped under Belgium.

Table 10: Main exporting countries 1996

illustration not visible in this excerpt

Table 11: Main exporting countries 2007

illustration not visible in this excerpt

Table 12: Main exporting countries 2009

illustration not visible in this excerpt

Imports

Table 13: Estimations’ results of the coefficients for all importing countries

illustration not visible in this excerpt

Table 14: Estimations’ results of the coefficients for the DCs, imports

illustration not visible in this excerpt

Table 15: Estimations’ results of the coefficients for the OECD, imports

illustration not visible in this excerpt

Table 16: Regression-All

illustration not visible in this excerpt

Note: All exporting countries

Table 17: Regression-OECD

illustration not visible in this excerpt

Table 18: Regression-DCs

illustration not visible in this excerpt

Table 19: Trade’s direction (Developed and developing countries)

illustration not visible in this excerpt

Note : the countries covered in these statistics (table 19) are those stated in list 1, selection based on the complete availability of data for the period 1996-2001-2007-2008.

List 1: Composition of country groups

illustration not visible in this excerpt

Table 20: Number of active products and per capita income, on average

illustration not visible in this excerpt

Figure 1: Average number of exported products (HS-6), developed and developing countries

illustration not visible in this excerpt

Figure 2: Average number of products (HS-6) and the level of average revenue between developed and developing countries

illustration not visible in this excerpt

List 2: Composition of the developing countries group

illustration not visible in this excerpt

List 3: Composition of the OECD countries group

illustration not visible in this excerpt

8. References

Andrew B. Bernard, J. B. Jensen, Stephen J. Redding and Peter K. Schott, “The margins of US Trade”, January 2009, CEP-LSP.

Armington, Paul (1969), "A Theory of Demand for Products Distinguished by Place of

Production," IMF Staff Papers 16, 159-176.

David Hummels, Peter J. Klenow, “Variety of and Quality of a Nation’s Trade”, January 2002, NBER.

Olivier Cadot, Celine Carrère, Vanessa Strauss-Khan, “Export Diversification, What’s behind the hump ?”, November 2009, CERDI.

Head, Keith and John Ries (2001), "Increasing Returns versus National Product Differentiation as an Explanation for the Pattern of US-Canada Trade" American Economic Review 91(4), 858-876.

Krugman, Paul R. (1981), "Intraindustry Specialization and the Gains from Trade" Journal of Political Economy 89, 959-973.

Romer, Paul M. (1994), "New Goods, Old Theory, and the Welfare Costs of Trade Restrictions" Journal of Development Economics 43, 5-38.

Schott, Peter K. (2001), "Do Rich and Poor Countries Specialize in a Different Mix of Goods?

Evidence from Product-Level U.S. Trade Data" NBER Working Paper #8492, September.

[...]


[1] Organisation for Economic Co-operation and Development.

[2] Based on the international trade flow of 1995.

[3] The estimations made on the GDP at current prices also lead to the same conclusion.

[4] Purchasing Power Parity.

[5] International Monetary Fund.

[6] This finding remains unchanged even when using the GDP at current prices in USD.

[7] See the list of countries (list 2)

[8] See list of countries (list 3)

[9] Oliver Cado, Celine Carrère, Vannessa Strauss-Khan (Diversification et niveau de développement, Nov. 2009)

Details

Pages
30
Year
2010
ISBN (Book)
9783656067672
File size
966 KB
Language
English
Catalog Number
v182772
Grade
Tags
dynamics increased trade large economy intensive extensive export

Author

Share

Previous

Title: Dynamics of Increased Trade of a Large Economy