Decomposition of export growth into diversification patterns

Research Paper (postgraduate), 2010
27 Pages

Free online reading

Export Growth and Diversification Patterns Intensive versus Extensive Margins

Internal working paper

Kerfalla Conté-ITC/DMD/MAR


This paper explores the extent to which products and product-markets diversifications contribute to export growth. It has used the bilateral trade at a very detailed level across countries, between 1996 and 2008. The results suggest that the export growth comes mainly from the intensive margin, i.e. product lines that had been already exported to the existing established trading partners. The decline in export, either due to the fall or the extinction in the existing established trade relationships, occurs more substantially on export growth for developed countries than for developing countries. And the appearance of new exports (extensive margin), either new products or markets, represents a vital source in export earnings for the developing countries, especially for the least developed countries (LDCs).

Please send any comments or suggestions to:

The findings and interpretations expressed in this document are entirely those of author. They do not necessarily represent the view of the ITC, its Executive Directors, or the countries they represent.

Export Diversification Margins

Export diversification is a process during which a country expands its export portfolio for a sustainable increase of the exports earnings. A diversified country is meant to be less vulnerable to adverse terms of trade shocks by stabilizing exports revenues (Ghos and Ostry, 1994). An export diversification can take two main forms, (i) product diversification and (ii) market diversification, and each form can be driven through two other channels: expansion of a set of new items (extensive margin) or upgrading the quantity of the existing items (intensive margin).

In this section, we investigate on how to segregate the export patterns into the extensive and intensive margins and the extent to which they contribute to the total change in exports.


The ideas explored in this paper are partly based on the Balza and Pineda (2008)[1] methodology. With respect to the definition above, we define what is meant by extensive and intensive margin. The extensive margin is the growth of exports due to the expansion in new product lines. Whereas the intensive margin refers to the growth of exports in product lines that have been already exported (old or continuing exporting products).

We call traditional-product lines, products that have been exported from a period of time and continue to be exported (continuing products). Death-product lines are those exported at the beginning of the period but are no longer exported (old products). New product-lines refer to the products that were not exported at the beginning of the period but started to be exported at the end of the period. A traditional destination is that to which at least one product line had been sent at the beginning and ending periods. Death destination refers to product lines that were sent at the beginning of the period but not at the end of the period. Furthermore, new destination is that to which no products were exported at the beginning of the period but at the end of the period. Finally, we call existing products/destinations, the traditional or death products/destinations.

How to calculate the extensive and intensive margins?

Following the definition above, the intensive margin is calculated by identifying the set of product-destinations that have been exported at the beginning and the end of periods (existing products to existing destinations). The extensive margin corresponds to the sum of new products to existing destinations, existing products to new destinations, and new products to new destinations (see table 1).

By decomposing the existing lines into the traditional and death lines (see table 1), the intensive margin can be generated as sum of “traditional products to traditional destinations”, “traditional products to death destinations”, “death products to traditional destinations”, and “death products to death destinations”. On the other hand, the extensive margin is generated as sum of “new products to new destinations”, “new products to traditional destinations”, “new products to death destinations”, “traditional products to new destinations”, and “death products to new destinations”.

Table 1: Export patterns’ margins (Extensive and Intensive)

illustration not visible in this excerpt

Note: Extensive margin (Em) = TPNM + DPNM + NPNM + NPTM + NPDM. Intensive margin (Im) = TPTM + TPDM + DPTM + DPDM (T = Traditional, P = Products, N = New, M = Markets, D = Death).

Data and Coverage

Based on the Harmonized System nomenclature[2], we used the bilateral trade data retrieved from the UNSD[3] (COMTRADE). Data are on the annual basis from 1996 to 2008, combining reporting data and converted and mirror estimates[4] statistics as all the countries do not report their own trade data to UNSD. No selecting criteria has been applied to individual countries, that is, some adjustment had been made to group some countries into one entity such as “Serbia and Montenegro”; “Belgium and Luxembourg”; “territories not else classified” which compiles small territories. Our dataset is based on around 200 countries, and the reporting countries’ statistics represent at least 85 per cent of total trade. To calculate the exports changes, we have split the period 1996-2008 into the initial period, from 1996 to 2000, and the final period, from 2004 to 2008. Then, we calculated dummy variables to identify whether or not a product line has been exported within the initial period and/or within the final period. The change is the difference between the values at the beginning of the period (1996-2000) and the values at the end of the period (2004-2008), and the contributions are ratios of a specific change to the overall changes for an individual country.


There are different ways to define new or existing products (using the number of times a product has been exported within the period, or using the average exports of the period of times). According to the strategy used, the product patterns (composition) can be influenced. On the other hand, the market diversification does not fit into the product classification point of view. Therefore, using the same notion can lack the spread of export diversification due to the geographical expansion. In addition to that, the level of product classification (HS 6-digits) seems to limit the scope of product lines that a country can export, which means the extensive margin is systematically underestimated. At the aggregate level, the composition of the selected countries within the groups can influence the total changes, affecting along the proportion of the extensive and the intensive margins


Between the periods 1996-2000 and 2004-2008, world average export values increased significantly by 121%, moving from $US 5’686 billion at the beginning of the period to $US 12’570 billion at the end of the period. At the individual level, the overall highest increase in exports was experienced in China followed by Germany, U.S.A, Netherlands and Japan. Among the leading performers, several developing countries have demonstrated a strong dynamism in exports such as Russia, Mexico, Emirates, Brazil, India, Malaysia, Thailand or Turkey. On the other hand, the largest decrease in average export values within the periods was recorded exclusively within the developing countries such as Dominica, Netherland Antilles, Guinea-Bissau, Tonga, Palau, and Gambia (see table 2, and for a full list of countries refer to annex table 8).

Table 2: Change in average export values (top five performers and under-performers)

illustration not visible in this excerpt

Note: Author’s calculation based on UN-COMTRADE. “Dped” stands for developed countries, “Dpng” corresponds to developing countries, and “Trcs” denotes transition economies. Export values are in thousands of $US.

Notice that the majority of the countries that have largely increased their exports (in average values), have reduced their exports patterns in terms of number of product lines exported within the periods considered. That is, the purpose in this section is to decompose the product lines to establish the extent to which export growth was generated by introducing new exporting product lines, or by intensifying the volume of the existing exporting products, or by the disappearance of the formerly exported products.

Following the types of products as defined above, the export growth in general seems to be largely dominated by the increase in existing product lines which are essentially driven by the traditional product lines. Among the developed countries (OECD minus Korea, Mexico, Chile, Israel), the contribution of the traditional product lines to the export growth is up to 90%. However, even though the growth is still driven by the traditional product lines, the configuration changes when it comes to the group of developing countries. In this group, there are two entities: i) countries of which exports patterns are dominated by the number of traditional product lines such as China followed by South Africa, India, Emirates, Mexico, Turkey, Malaysia, Russia, Thailand, Indonesia, etc and ii) countries of which exports structure shows a significant number of the new products lines such as Nigeria, Guam, Paraguay, Lesotho, Ethiopia, Iraq, Burundi, Maldives (see table 3, and for a more detailed view refer to annex table 9 and 10).

Furthermore, the introduction of the new product lines has produced a limited positive effect in the export growth. Within the developed countries, the contribution of the new products in the export overall changes has not exceeded 5%, with the exception of Finland, Cyprus, Spain, Norway and Canada. Among the developing countries, the introduction of the new lines has not played a significant effect in the export growth, but instead, the contribution of new product lines exceeded 30% of export changes for some countries (see table 4).

In general, the extinction of exports due to the death product lines has not harmed the dynamism of the exports for all groups of countries. Among the developing countries, the death product lines were more frequent, exceeding 10% of export extinction for some countries. Also, the significant drop in export growth due the death product lines was largely recorded for Palau and Guinea-Bissau.

Table 3: Leading dynamic exporters within the developing countries.

illustration not visible in this excerpt

Table 4: Countries that new product lines contributed with at least one third to export growth

illustration not visible in this excerpt

How are the exports changes segregated between the intensive and extensive margins? Do they differ for individual countries in different groups of economy?

From the entire groups of countries, the overall exports changes are predominantly due to the intensive margin which is around 98%. This means that the world export growth has occurred through the increase in exports on the existing established product-destinations. This result is largely confirmed among the developed countries for all regions, where the contribution of the intensive margin to the exports changes exceeds 90% (see table 5). It is important to note that the extensive margin within the developed countries occurs more importantly for Asian countries (Japan, Korea and Singapore) compared to the rest of the regions, where the extensive effect in terms of geographical diversification represents a significant proportion than the discovery of new product lines. Clearly, the developed Asian economies seem to add new export earnings by expanding their existing products to a large number of new destinations, which added $US 27 billion on average as export earnings, whereas the rest of developed economies (E.U., U.S. and Canada, Australia and New-Zeeland) would create new value of export by diversifying the product lines on their existing destinations. With the exception of the Asian group, the developed economies are keen to utilize their knowledge from their local market-destinations to sell more product lines, the latter adding $US 49 billion in export earnings on average.

Table 5: Extensive and Intensive margins’ contributions to export growth

illustration not visible in this excerpt

On the other hand, even though we observe the same conclusion for the developing countries concerning the predominance of the intensive margin, there is a persistent dissimilarity among the individual countries: countries of which intensive margin’s contribution to the export changes exceeds 50%, and those having less than 50% contributing in their own export changes (see annex table 13). In the latter, LDCs have experienced significantly new exporting measures, either by exporting new product lines to existing market-destinations and/or to new destinations. Moreover, the extensive margin was the exclusive source for growth of some of these countries, contributing to more than 100% of the increase in export, because the value of the export of their existing trading relationships declined with a negative intensive margin (see table 6). In that situation, these countries have to sell less of their existing products to their existing market-destinations, but some of them succeed in growth by diversifying their portfolio (offering more products to more market-destinations) such as Eritrea (that sold $US 17 million for its existing products on the existing market-destinations and added $US 22 million from the new exports earnings) while others did not succeed at all such as Gambia (that sold $US 17 million for its existing products on the existing market-destinations and lost $US 7 million on exploring new exports).

Table 6: Countries for which extensive margin counts for more than half of export changes (sample)

illustration not visible in this excerpt

Moreover, even though the increase in exports was generated by the intensive margin for developing countries, the contribution of the extensive margin to the export growth is more important than it was for developed countries. Indeed, the growth occurred in the extensive margin was around 9% for the African region, and up to 20% for the pacific islands. However, in these two regions, the diversification strategies were different (see table 7). For instance, Africa created new exports earning by significantly exporting new products to its existing market-destinations, which added on average $US 18 billion to export earnings. Whereas the Pacific Islands succeeded by exploiting more of their trading relationships: expanding market-destinations for their traditional products and enlarging the product lines to their traditional destinations which added $US 579 million in new export earnings on average. In other words, this means that developing countries have tried somehow to develop new export earnings, by using their knowledge, cultivated on their traditional market-destinations to expand or sell more product lines on. And their attempts to create new exports by exploring new products to new destination or selling traditional products to new destinations were quite moderate (geographical diversification).

Table 7: Decomposition of export growth by extensive and intensive margins across regions

illustration not visible in this excerpt

Concluding remarks

In this document, we have investigated in a very detailed manner how to explain export growth through the diversification patterns across countries. In our findings, it occurs that export growth is mainly generated by the growth in existing product-destinations relationships (intensive margin), for all the groups of countries. While the export growth from product-markets destinations development (extensive margin) counts for a moderate proportion, especially for developed countries. These results confirm those released by Newfarmer and Brenton in 2007, also by Armugo-Pacheco and Pierola in 2008.

Despite the fact that the extensive margin generates less export earnings, developed countries are more likely to use their innovations and experiences learnt from their traditional market-destinations to export a large number of new product lines. The experience learning from the established exporting destinations is somehow used more significantly within developing countries, especially LDCs, for which the appearance of new exports represents a non-negligible proportion in export growth. However, within this group, the growth into product diversification (new products to existing or new destinations) has been more substantial than the geographic diversification (existing or new products to new destinations), which may imply high entry costs in terms of promoting and targeting new buyers. Here, our findings differ from what was concluded by Newfarmer and Brenton and Amurgo-Pacheco. They found that within developing countries, the growth in geographic diversification was more important than the discovery of new export products.

Additionally, the extinction of the export due the disappearance or death of product lines is more frequent and more important within the developed countries’ export growth than the one recorded among developing countries. Nevertheless, this phenomenon could be essential for countries to enhance their export growth if they succeed to cope with constrains that generate these failures.


Table 8: Ranking in countries’ export growth (change in average values).Abbildung in dieser Leseprobe nicht enthalten

illustration not visible in this excerpt

Table 9: Product composition across countries, those having exported more new lines than traditional

illustration not visible in this excerpt

Table 10: Product compositions across countries, those having exported more traditional lines than new.

illustration not visible in this excerpt

Table 11: Export changes due the type of product composition, developed countries.

illustration not visible in this excerpt

Table 12: Export changes due the type of product composition, developing countries

illustration not visible in this excerpt

Table 13: The contribution to export growth (extensive and intensive margins).

illustration not visible in this excerpt

Table 14: The contribution to export growth for the extensive and intensive margins (detailed), developed countries.

illustration not visible in this excerpt

Table 15: The contribution to export growth for the extensive and intensive margins (detailed), developing countries.

illustration not visible in this excerpt


Alberto Amurgo-Pacheco and Martha Denisse Pierola. Patterns of Export Diversification in Developing Countries (Intensive and Extensive Margins). World Bank Policy Research, Working Paper WPS4473, January 2008.

Lenin Balza, Maria Caballero, Leonardo Ortega y José Pineda. Market diversification and exports growth in Latin America. Universidad Central de Venezuela, January 2008.

Paul Brenton, Richard Newfarmer, Peter Walkenhorst. Export Diversification: A Policy Portfolio Approach. September 2007.

David Hummels and Peter Klenow. The Variety and Quality of a Nation’s Exports, 2005.

Ghosh, Astish R. and Ostry, Jonathan. Export Instability and External Balance in Developing Countries, IMF Working Paper No. 94/8, International monetary Fund, Washington DC, 1994

illustration not visible in this excerpt


[1] Used Evenett and Venables’ (2002) approach.

[2] HS at 6 digits in its revision 1996.

[3] UN Statistics Division.

[4] Statistics from the reporting partners’ declarations.

26 of 27 pages


Decomposition of export growth into diversification patterns
Technical-paper/ Private research
Catalog Number
ISBN (Book)
File size
1001 KB
Document produced to analyse mecanism of export growth in the developing economies. I submitted to my section for discussions (Market Analysis Section - ITC)
Quote paper
Kerfalla Conte (Author), 2010, Decomposition of export growth into diversification patterns, Munich, GRIN Verlag,


  • No comments yet.
Read the ebook
Title: Decomposition of export growth into diversification patterns

Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free