The Micro-D classification

The Micro-D classification:
A new approach to identifying differentiated exports
Federico Bernini1 Julia Gonzalez2 Juan Carlos Hallak3 Alejandro Vicondoa4
This version: March, 2016
Abstract
It is common to interpret the evolution of a country’s export structure as a manifestation of
progress in its development process. Performing this exercise requires determining which
features of exported products denote higher stages in that process. We argue that exports of
differentiated products, especially when sold to developed countries, manifests the acquisition of
valuable knowledge that reflects development progress. Most existing export classifications
cannot identify with accuracy this type of exports. To address this concern, we propose a new
classification of differentiated products denoted Micro-D. This classification works at a very
fine aggregation level, especially taking into account package size as a proxy for differentiation in
resource-based sectors. We apply this classification to study the evolution of Argentina’s exports
during 1998-2011. The results point to the importance of classifying goods at a finer aggregation
level as severe misinterpretation problems arise when other classifications are used to assess
export performance.
JEL Classification: F10, F14, O14
Keywords: Differentiated products, exports, classification
We thank Paula Calvo for outstanding research assistance, Roberto Bisang for useful comments, and Facundo
Albornoz for his contribution to this paper in its early stages. The usual disclaimer applies.
1 Universidad de San Andrés, Buenos Aires, Argentina. E-mail: [email protected]
2 PhD student at University of Illinois at Urbana-Champaign. E-mail: [email protected]
3 Universidad de San Andrés, CONICET, and Secretaría de Transformación Productiva, Ministerio de Producción,
Buenos Aires, Argentina. Email: [email protected]
4 European University Institute E-mail: [email protected]
1
1. INTRODUCTION
The evolution of a country´s export structure is usually viewed as a manifestation of
progress or stagnation in its development process. To conduct this exercise, the
analyst is required to take a stand on which products – and maybe which
destinations – are more “desirable”, in the sense that they can be linked to higher
development stages. Numerous efforts have been made to construct product
classifications that capture a vertical dimension across products in terms of
technological complexity (Hartzichronoglou, 1997; Hidalgo et al., 2009; Lall, 2000),
order in which they start being exported (Feenstra and Rose, 1997), and
development level of countries that export them (Hidalgo et al., 2009). A similar but
coarser exercise is often made in country-specific academic papers and policy
reports that focus on rough indicators of export desirability such as whether
exported products are “industrial” or “non-traditional” (Gabriele, 1997; Kouzmine,
2000; Von Hesse, 1994). In this paper, we contribute to those efforts by proposing a
new, more accurate, classification of differentiated products. We argue that this
classification can better capture both development progress and prospects.
Differentiated products need not be technologically complex. However, the singular
attributes that make their physical characteristics, design, brand image, or service
reliability unique in the market also allow them to fetch a higher price and reward
higher wages. Thus, a country’s ability to export differentiated products requires the
acquisition of valuable capabilities that manifest development progress. In addition,
for developing countries like those in Latin America, exporting differentiated
products to developed countries imposes requirements in terms of export-related
2
practices that once acquired by some firms can spread throughout the sector and
even across sectors (Artopoulos et al. 2013). Therefore, growth in this type of
exports can also predict future export growth.
While a renowned classification developed by Rauch (1999) distinguishes export
products by degree of differentiation, this classification is defined at an aggregation
level – 4-digit SITC – that is too coarse to identify relevant differentiated items. This
occurs because differentiated products are often lumped together with
undifferentiated ones in broader categories. While aggregation issues permeate all
sections in the classification, they are particularly prevalent in differentiated food
(and beverage) products, which exhibit well-recognized export upgrading
opportunities in many resource-abundant Latin American countries. In this paper,
we propose a classification, denoted Micro-D that attempts to correct for these
problems.
The Micro-D classification is based on export information reported at the maximum
level of detail in export nomenclatures. One of the major advantages of working
with very disaggregated data is that food and beverage items tend to be classified by
package size, which can be used as a proxy of product differentiation. More
generally, at the finest aggregation level specific items can be more accurately
classified. On the other hand, analyzing exports at a disaggregation level beyond the
6-digit Harmonized System (HS) – which is the finest harmonized level across
countries – implies that the Micro-D classification will necessarily have countryspecific features. Nevertheless, the definitions we use are easily extrapolated to any
3
HS-based classification system since they stem from very simple rules. Here, we
apply the Micro-D classification to assess the evolution of Argentine exports during
the period 1998-2011.
We first compute growth performance in differentiated products according to the
Micro-D classification. Then, we compare it with growth in “desirable” exports
computed under three other classifications. First, we look at Manufactures of
Industrial Origin (MOI) as classified by Argentina’s National Statistical Institute
(INDEC). Second, we look at medium-high and high technologically-intensive
products as classified by Hartzichronoglou (1997). Finally, we look at differentiated
products as classified by Rauch (1999). In particular, we focus on exports of
desirable products to OECD countries, which we denote “upgraded” exports, as the
metric for assessing virtue in the recent evolution of Argentina’s export
performance. Only exports to developed countries provide a strong signal of the
acquisition of diffusible knowledge that can predict future export growth
(Artopoulos et al., 2013).
We find that the main conclusions that can be drawn from the analysis depend
dramatically on the product classification used to identify Argentina’s upgraded
exports. First, upgraded exports display the largest absolute growth during the
sample period 1998-2011 when they are computed using INDEC’s classification.
However, this result is largely driven by the misleading fact that INDEC’s MOI
include precious metals – mainly unwrought gold. These metals are undifferentiated
goods (almost commodities) that only went through basic industrial processing.
4
Second, despite being based on very different criteria upgraded exports according to
Hartzichronoglou’s (1997) and Rauch’s (1999) classifications yield very similar
results both in terms of absolute growth and composition. However, in both cases,
upgraded export growth is primarily and misleadingly explained by biodiesel exports
(61% and 57% respectively). This product is neither technologically complex nor
differentiated. As we explain later, it is tagged as such by these two classifications,
respectively, due to aggregation.
The Micro-D classification provides a substantially different account of the
evolution of upgraded exports from Argentina during the period 1998-2011.
Although it yields similar – although slightly lower – growth in total upgraded
exports, it delivers a dramatically different account of its composition. In particular,
it shows that the main sectors responsible for growth in upgraded exports are not
Chemicals or Precious Metals, as the analysis based on the other classifications
suggest, but instead Food and Beverages, which contributed 48% to total upgraded
export growth. This category includes a wide range of products widely
acknowledged to present upgrading opportunities for resource-dominated LatinAmerican countries’ exports. This finding can be viewed with optimism as this type
of export growth may indicate the acquisition of knowledge that can percolate easily
throughout a broad range export items. On the other hand, since applying the
Micro-D classification does not deliver a higher total amount of upgraded exports,
this result also indicates that other sectors have performed less successfully than
inferred from using alternate product classifications. In any event, we hope our
proposed classification will allow for a more accurate diagnosis of the sources of
5
export growth that will better guide export promotion and productive development
efforts.
The rest of the paper proceeds as follows. Section 2 justifies our methodological
approach. Section 3 describes the main features of the Micro-D classification.
Section 4 applies it to assess the recent evolution of Argentina’s exports. Section 5
concludes.
2. WHY FOCUS ON DIFFERENTIATED PRODUCTS TO
DEVELOPED COUNTRIES
It is customary to analyze the productive performance of a developing country over
a certain time period by studying the change in its export structure. In particular,
increases in the volume or share of industrial exports have been prominently
interpreted as a sign of productive development. This practice is supported by the
traditional view of economic development as an industrialization process. It is also
facilitated by the fact that national statistical institutes typically report exports
distinguishing industrial from primary products.
A finer approach does not exclusively take into account the industrial component of
a country’s export basket but instead focuses on its technological intensity.
Underlying this approach is the widespread notion that economic development is
associated with a country’s ability to produce, and thus export, technologically
complex products. Aligned with this view, Hartzichronoglou (1997) and Lall (2000)
developed two alternative export classifications according to products’ technological
6
intensity, which have been widely used in empirical studies and country reports to
assess export performance (Aggarwal, 2002; Bahar et al., 2013; Jarreau et al., 2012;
Moreira, 2007; Poncet et al., 2013; Srcholec, 2007 and Stehrer et al., 2009). 5 In
addition to capturing the stage of development of a country’s productive capabilities
at a given point in time, a country’s export composition may also predict future
export growth. According to Lall (2000): “[…technology intensive structures offer
better prospects for future growth because their products tend to grow faster in
trade: they tend to be highly income elastic, create new demand, and substitute
faster for older products.]”
A product’s technological complexity, however, may not be a good indicator of its
income elasticity, the speed at which it substitutes for older products and, more
importantly, the degree to which its design, production, and/or commercialization
requires complex knowledge. Differentiated products may have all these desirable
properties without necessarily being “technologically complex”. Due to their quality,
design, traceability, brand reputation, or customization, differentiated products
possess distinct attributes that make them less substitutable vis a vis competing
products. Thus, they command higher prices and can reward higher wages.
Although technologically complex products are often differentiated, differentiated
products need not be technologically complex. In fact, some of the most widely
recognized upgrading opportunities for Latin American countries – especially in
Hartzichronoglou (1997) classifies manufactures in high technology, medium-high technology,
medium-low technology, and low technology. Lall (2000) classifies manufactures in resource-based,
low technology, medium technology and high technology.
5
7
South America – involve precisely differentiated products that use standard
technology. At one extreme of the product spectrum, these countries have
difficulties competing in cost with low-income countries in most low-technology
products, which are usually scarcely differentiated. At the other extreme, they do not
possess the technological capabilities to compete with high-income countries in
high-technology products that are highly differentiated. In between these extremes,
however, they display a competitive potential in the production of differentiated
products, particularly those that are based on their abundant natural resources.
Despite not being technologically complex, these products are still highly valued in
international markets.
In their study of Argentine cases of export emergence, Artopoulos et al. (2011,
2013) discuss the obstacles that prevent developing-country firms from exporting
differentiated products to developed countries. They point to the need to adopt a
set of business practices which are very different from those that prevail in their
domestic markets. Firms need to adapt product designs to foreign demand
idiosyncrasies, upgrade quality, conform to foreign distributors’ way of doing
business, and engage with them so that they can be a source of information about
the evolution of foreign demand. Also, the adoption of these practices is determined
by the firm’s possession of foreign market knowledge. Adopting these practices
especially requires that the firm understands the tacit features of foreign demand in
developed countries, which tends to differ drastically from demand in developing
countries. In the same vein, González and Hallak (2013, 2016) argue that insertion
in global value chains oriented to non-mass segments of developed-country markets
8
imposes on firms less stringent but nonetheless qualitatively similar requirements
about the foreign market knowledge they need to possess.
Since exporting differentiated goods to developed countries requires foreign market
knowledge, a country’s increase in this type of exports manifests the extent to which
its firms have acquired this knowledge. Thus, growth in differentiated export to
those destinations can be regarded as a signal of increased sophistication in a
country’s productive structure and its ability to sell products abroad that command
higher prices. Furthermore, the type of knowledge underlying this kind of export
growth can be easily diffused within and across industries. 6 Thus, growth in
differentiated exports to developed countries can also be viewed as a predictor of
further export growth. Other types of export growth – particularly growth in
differentiated-good exports to other developing countries – may imply benefits to a
country in terms of increased labor demand. However, they will not necessarily
reflect the acquisition of diffusible foreign market knowledge and hence may not
predict future export growth.
Focusing on differentiated exports as a signal of sophistication and predictor of
export growth is also justified by the potential scalability of this type of exports. In
the case of differentiated products that are not dependent on natural resources, as
intermediate inputs can be sourced from abroad the volume of a country’s export is
only constrained by its availability of human and managerial capital. In the case of
differentiated products that are heavily dependent on natural resources, the
This type of diffusion process is described in Artopoulos et al (2011, 2013) in the context of
emerging new exporting sectors.
6
9
opportunities for scaling up the volume of exports are still considerable despite
limited availability of those resources. Since in most cases natural resources are
largely exported in primary forms, there is usually ample scope for transforming
them into differentiated products before selling them abroad. Thus, in practical
terms natural-resource availability does not impose in practice a binding constraint
on export growth of resource-based differentiated products.
3. A NEW CLASSIFICATION OF DIFFERENTIATED PRODUCTS
In order to assess a country’s export growth in differentiated products, we need to
be able to classify goods according to their degree of differentiation. Rauch (1999)
developed a classification that separates goods into three categories: (a)
homogeneous products are those commercialized in international exchange markets,
(b) referenced-priced products are those that display reference prices in specialized
publications, and (c) differentiated products are all remaining ones. The Rauch
classification is the best available option to identify differentiated products in
customs databases and is also the one most widely used in the literature.7
The Rauch classification, however, categorizes “products” at an aggregation level –
4-digit SITC – that is too broad to capture certain differentiated-good exports. This
problem is particularly prevalent in categories that include natural resource-based
products. For example, 4-digit SITC category 1121: “Wine of fresh grapes”, which is
Broda and Weinstein (2006) develop a method to estimate demand elasticities at the product label
using only international trade information. While these elasticities can be used as proxies for the
degree of product differentiation, they largely depend on strong assumptions about substitution
patterns, quality differences, and functional forms.
7
10
classified by Rauch as “referenced priced”, lumps together grape must, a scantly
differentiated good that has historically accounted for most Argentine exports in
this category, with boxes of bottled fine wine, a highly differentiated good that
currently accounts for most Argentine wine exports. Since achieving higher value
added through differentiation in resource-based products has long been recognized
as a promising avenue for export upgrading in Latin America, this shortcoming in
the classification has relevant implications for the analysis of these countries’
exports. To address this shortcoming, we develop an alternative classification of
differentiated products.
Our proposed classification requires export data at a very fine level of aggregation.
Since the SITC classification system does not disaggregate product categories
beyond the 4-digit level, our proposed classification resorts to highly disaggregated
data from Argentina’s customs nomenclature Sistema Informático María (SIM).
Customs nomenclatures specify some of the relevant product attributes that can
serve as proxies for differentiation only at a very fine level of aggregation. In
particular, package size, which we use as one of the main indicators of product
differentiation, is specified by the SIM only at the 12-digit level. Thus, we need to
classify goods at that fine level of aggregation. In addition to identifying package
size, an advantage of classifying goods at a fine aggregation level is that it helps
attain better precision in the classification of other salient export items. An ensuing
drawback is that, although the SIM is based on the international Harmonized
System (HS), the HS is harmonized across countries only up to the 6-digit level.
Thus, the specific classification applied here is not directly applicable to other
11
countries’ exports. For this reason, our proposed classification attempts to establish
simple and transparent guiding criteria to ensure that it can be easily adapted to the
specific nomenclatures of other countries.
The Argentine SIM nomenclature started to systematically discriminate products by
package size since 1996. 8 By contrast, other Latin American countries’ export
nomenclatures only separate items by package size more sparsely. The application of
our proposed classification will lose precision the fewer are the export items broken
down by package size. As a result, the application of this classification may lose
power when applied to the analysis of other countries’ exports. Nevertheless, by
showing the importance of distinguishing differentiated products in export statistics
for the analysis of export prowess we hope to influence future efforts in customs
data collection.
Our classification uses two broad criteria to identify differentiated products. For
food, beverages, and other agricultural-based products, the key attribute is package
size. For many other industrial products, the key attribute is the degree of
manufacturing processing. Various other criteria are applied in specific cases. The
“package size” criterion is extensively used to classify products in HS chapters 1 to
24, which consist of agricultural and food products. These products are considered
differentiated if they were exported in small packages while they are not
differentiated if they were traded in bulk or in larger containers.9 This criterion
More details of this change are provided in Appendix A.
The package size used as a threshold varies across products depending on the product characteristics
and on the level of detail provided by the SIM nomenclature. For example, fruit exports are counted
8
9
12
allows us to distinguish food exports oriented towards final consumption from
other agricultural exports that are used as intermediate inputs or are re-packaged.
The justification of this choice is that food products traded in small packages tend
to possess differentiated attributes such as a recognized brand or a unique way of
appealing to consumers. Due to the central role played by package size in our
classification, we call it the “micro-differentiated (Micro-D)” classification.
We use the package size criterion both for primary products, such as fruit and
cereals, and for manufactured agricultural products, such as chilled or frozen meat,
infusions, oil, and juice. Some food exports are classified as differentiated regardless
of package size since they present a relatively high degree of processing and are not
exported in bulk. These products include, for example, dulce de leche, candy,
champagne, preparations for infant use, bakers’ wares, pasta, breaded meat, and
products frozen using Individual Quick Freezing (IQF) processes. In the case of live
animals, they are considered differentiated if they are pure-bred or for breeding.
For a substantial fraction of the remaining chapters, we exploit the structure of the
HS classification. This classification is largely organized around a products’ primary
material, distinguishing its primary forms from products obtained from the
material’s further processing. Using this organization, we classify the former as nondifferentiated products and the latter as differentiated products. Specifically, primary
forms of minerals, chemicals, metals, plastics, rubber, leather, textiles, glass, stone,
as differentiated if they are traded in containers below 20 kilograms while meat exports are counted as
differentiated if they are traded in containers below 5 kilograms. In general, when the SIM
nomenclature distinguishes between different package sizes, we only consider the smallest one as
differentiated (although there are a few exceptions to this rule).
13
wood, and paper are non-differentiated, while manufactures made from these goods
are differentiated. The justification for this guiding rule is that among product
attributes systematically included in the description of HS (and SIM) classification
codes the degree of manufacturing processing is the one most strongly linked to the
product’s degree of differentiation.10
In many cases, we apply specific criteria to classify products as differentiated or not
differentiated based on our knowledge of products and sectors and the information
provided to us by relevant actors in the relevant industries. For example, in the case
of chemicals (chapters 28 to 38) we classify as not differentiated enzymes, casein,
propellant powders, tanning extracts of vegetable origin, raw coloring matter, and
other basic inputs of chemical industries. On the contrary, pharmaceutical products,
vaccines, perfumery, polishing or scouring preparations, and other elaborated
chemical products are considered differentiated. In many other cases, classification
is facilitated by the fact that all products included in a chapter are differentiated.
This is the case, for example, of chapters 84 and 85, which contain most types of
machinery and equipment, and chapters 86 to 89, which include the universe of
transport equipment. Appendix A provides a more detailed description of our
classification. The full classification database is provided in an online appendix
available on the authors’ websites.
In the HS classification, sometimes a chapter is entirely devoted to a particular material (e.g.
chapter 39: “Plastics and articles thereof”) and finer levels within the chapter distinguish primary
forms of this material from products that result from the material further processing. In other cases,
various materials are included in a chapter (e.g. chapter 38: “Miscellaneous chemical product”) but,
for each material, primary forms and further processing products are still distinguished at finer levels.
10
14
Our criteria might generate debatable classification decisions in some cases. For
example, while we classify certain products as differentiated because they have gone
through
substantial manufacturing they
could reasonably be
considered
homogenous. This is particularly the case of some manufactures of metal, such as
aluminum ingots, which are traded as industrial commodities in international
markets. Despite this shortcoming, we stick to the broad criterion in these cases to
preserve the simplicity and transparency of the classification criteria. Another
shortcoming stems from the fact that sometimes the maximum disaggregation level
in the nomenclature is still insufficient to distinguish differentiated products. For
example, some soybean seeds are exported due to their distinct (geneticallymodified) attributes. However, since the SIM does not have a distinct code for this
type of seed they are classified – as all other exported soybean seeds – as not
differentiated. Overall, we think these instances of misclassification are not as
prevalent to substantially affect the general conclusions we draw from the analysis.
In any event, we note that working at a more disaggregated level, as we do here,
provides flexibility to correct the classification of specific salient items in a country’s
export basket if deemed necessary. In this regard, we view our proposed
classification as a starting point in a collective effort to generate a more accurate
classification of differentiated products.
4. ASSESSING THE EVOLUTION OF ARGENTINA’S EXPORTS
As discussed in section 2, any assessment about the evolution of a country’s export
structure will largely depend on the stand we take on which products embed
15
relevant complexity and thus are more desirable as components of an export basket.
In this section, we apply our Micro-D classification to identify desirable products
and compare the results to those obtained using alternative classifications. As a
benchmark, we first look at exports to all destinations. Then, we look at exports to
OECD countries, which constitute our focus of interest.
We study the evolution of Argentina’s exports between 1998 and 2011. Those two
years mark the peaks of two macroeconomic cycles. The first peak is at the height of
the Convertibility regime that ruled the country between 1991 and 2001. The second
peak is at the height of the post-Convertibility regime that prevailed under the
Kirchners´ administration. We choose to end the analysis in 2011 because during
that year the Argentine government initiated an unabated turn towards a commercial
policy of generalized protection that imposed discretionary authorization
requirements on every import shipment, whose consequences on the country’s
international trade would make the analysis more complex. The effects of that policy
are left for future research.
Figure 1 provides a basic overview of the evolution of exports and the real exchange
rate in Argentina during the sample period. Total exports grew dramatically, from
US$ 26.4 billion in 1998 to US$ 84 billion in 2011. Although the implied growth rate
was high (218%), it was still below world trade growth during the same period
(236%). An important driver of the sharp increase in exports was the abrupt jump
of the real exchange rate after the crisis and currency devaluation in early 2002. The
dotted line in the figure shows the evolution of the multilateral real exchange rate.
16
This rate increased 140% as a result of the devaluation, remained relatively stable
until 2007, and then initiated a process of gradual appreciation that persisted
throughout the rest of the period. 11
Figure 1: Evolution of total exports and the real exchange rate
90
80
70
60
50
40
30
20
10
0
250
200
150
100
50
0
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Total (left)
Multilateral real exchange rate (right)
Total exports in US$ billions. Multilateral exchange rate: December 2001=100.
We identify “desirable” products using four alternative classifications. First, based
on the classification used by the official statistical institute in Argentina (INDEC)
we consider desirable those products included in category “Manufactures of
Industrial Origin” (MOI).12 Second, based on the most widely used classification by
technological content (TC) (Hartzichronoglou, 1997) we consider desirable those
products in categories “High Intensity” and “Medium-High Intensity”. Third, based
The real exchange rate was computed using the nominal multilateral exchange rate computed by
Argentina’s Central Bank. This nominal rate was then deflated by INDEC’s Consumer Price Index
until 2006 and by the Consumer Price Index of Santa Fe province’s statistical institute since 2007.
12 The classification used by INDEC groups products in four categories: Primary Products,
Manufactures of Agricultural Origin (MOA), Manufactures of Industrial Origin (MOI), and Fuels
and Energy. It is customary both in the academic literature and in the press to regard MOI as the
virtuous category.
11
17
alternatively on the Rauch classification and on the Micro-D classification desirable
products are those classified as “Differentiated”.
Figure 2: Desirable exports by product classification (US$ billions). All destinations.
35
30
25
20
15
10
5
0
98
99
00
01
02
INDEC
03
04
Rauch
05
06
TC
07
08
09
10
11
Micro-D
Figure 2 compares the evolution of desirable exports according to these four
classifications taking into account exports to all destinations. One salient feature of
the figure is that the evolution of differentiated exports according to the Micro-D
classification tracks very closely the evolution of INDEC’s MOI. This result is not a
priori obvious since many product categories that are considered desirable under
one classification are not so under the other. In fact, while total desirable exports in
2009 were almost the same for both classifications, only 74% of this total is
explained by the same products. For example, “Unwrought gold” is included by
INDEC under MOI but is not differentiated according to the Micro-D
classification. By contrast, “Vine of fresh grapes” is differentiated according to
Micro-D but is not a MOI. Another feature to notice is that the two upper
categories in the TC classification, combined, appear to be more stringent in the
18
identification of desirable products. During the sample period, the volume of TC
desirable exports was 25% below the volume of desirable exports as computed by
INDEC and by the Micro-D classification.
Notwithstanding differences in levels, all four classifications deliver a similar growth
result. This can be appreciated in figure 3, which exhibits desirable export growth
rates between the 1998-1999 average and the 2010-2011 average for each
classification. 13 The growth rate is highest when desirable exports are computed
using the TC classification (268%) while it is lowest under the Micro-D
classification.
Figure 3: Growth rate of desirable exports by classification. All destinations. 1998/9-2010/1.
300%
250%
200%
150%
100%
50%
0%
INDEC
TC
Rauch
Micro-D
As argued in section 2, differentiated exports to high-income countries are more
likely to manifest the acquisition of diffusible knowledge than differentiated exports
to other destinations. In particular, differentiated exports to neighboring countries
13
We use two-year averages to smooth out idiosyncratic variation in extreme years.
19
and other low-income countries do not necessarily require the capability to adapt
products and business practices to foreign market needs. Thus, we would like to
restrict the definition of desirable exports to those that have developed countries as
the destination. For this reason, we repeat the analysis restricting the set of
destinations to the 23 OECD members in 1990. This set excludes more recent
members such as Mexico, Korea or Chile. We call exports restricted to OECD
countries “upgraded” exports to distinguish them from exports to all destinations.
Upgraded exports account for a small share of desirable exports in all four
classifications. Specifically, desirable exports sent to the OECD account for 27%,
19%, 20% and 16% of the 2010-2011 average under INDEC, TC, Rauch, and the
Micro-D classifications, respectively. These shares are lower in most cases than the
share of the OECD in total exports (25%) as desirable exports are
disproportionately sent to regional partners.
The differences between the four classification criteria become larger when we
restrict the analysis to upgraded exports. Figure 4 shows that although upgraded
exports under INDEC’s and the Micro-D classifications track each other closely
during most of the sample period, in the last three years the two time series depart
dramatically. In particular, in the last sample year the Micro-D classification yields
upgraded exports of US$ 5.14 billion whereas INDEC’s upgraded exports yield US$
8.46 billion (a number 65% higher). This is almost entirely attributable to biodiesel
and precious metal exports. Both are considered upgraded by INDEC because they
are Manufactures of Industrial Origin but are not upgraded according to the Micro-
20
D classification because despite going through industrial processing they are still
commodities.
Figure 4: Upgraded exports by product classification (US$ billions). OECD.
9
8
7
6
5
4
3
2
1
0
98
99
00
01
02
INDEC
03
04
Rauch
05
06
07
TC
08
09
10
11
Micro-D
The classifications also yield different results when we look at growth rates. Figure 5
displays the growth rate of upgraded exports between the 1998-1999 average and
the 2010-2011 average according to each of the four classifications. While upgraded
exports grew by 346% and 332% under INDEC and TC, respectively, this growth
rate was substantially smaller, 179% and 126%, under the Rauch and Micro-D
classifications. Also, while upgrade exports grew at a much faster rate than did
desirable exports according to INDEC and the TC classifications, the opposite is
true according to the Rauch and Micro-D classifications.
21
Figure 5: Growth rate of upgraded exports by classification. OECD. 1998/9-2010/1
400%
350%
300%
250%
200%
150%
100%
50%
0%
INDEC
TC
Rauch
Micro-D
In order to understand why the classifications yield results that differ so much, we
divide products in twelve groups (see Appendix B for details) and identify which of
them are responsible for the main differences across classifications.
Figure 6
decomposes absolute growth in upgraded exports according to each classification.
Under INDEC’s classification, a key source of growth is “Precious Metals”, which is
mainly “Unwrought gold”. The other three classifications rightly leave these
products out of the “upgraded” set. In particular, the TC classification does not
consider them upgraded because they are not technologically intensive while the
other two classifications do not consider them upgraded because of their low degree
of differentiation.
Second, a surprising result is that the TC and Rauch
classifications deliver very similar results in terms of absolute growth sectoral
composition. In both cases, the most important sector is “Chemicals” (almost
exclusively biodiesel), which accounts for 55% of growth in this category.
22
Figure 6: Contribution to upgraded exports growth (US$ billions). OECD. 1998/9-2010/1.
6000
5000
4000
3000
2000
1000
0
INDEC
Food and Beverages
TC
Rauch
Chemicals
Precious Metals
Micro-D
Vehicles
Others
Compared to these two classifications our Micro-D classification delivers smaller
absolute growth in upgraded exports. While upgraded exports grew by 2.9 and 3.0
billion dollars, respectively, under the TC and Rauch classifications, they grew by 2.5
billion dollars under the Micro-D classification. Nevertheless, dramatic differences
arise when we consider growth composition. Two differences stand out. The first is
that while under the TC and Rauch classifications the most important growth
component is Chemicals (which represents 61% and 57% of total growth,
respectively), the contribution of this component to upgraded growth under the
Micro-D classification is very small (it only represents 8%). The second is that Food
and Beverages is the most important contributor to upgraded growth according to
the Micro-D classification (accounting for 48%) whereas it is a negligible
contributor to upgraded growth under the Rauch and TC classifications (accounting
for 1% and 8%, respectively). In fact, the Rauch classification only considers very
specific food items as differentiated while the TC classification does not even
23
include any food or beverage item in its two upper technological-content categories.
Vehicles and other products are classified as upgraded by all classifications. 14
Adding value through differentiation in food (and beverage) items sold to highincome countries is a relevant manifestation of productive transformation in
resource-abundant Latin American countries. However, the TC classification
neglects this type of transformation by focusing on technology while the Rauch
classification does not capture it due to insufficient disaggregation. The Micro-D
classification, by contrast, captures export upgrading in food products by looking at
package size in disaggregated export statistics. The opposite is true for biodiesel.
This product is considered of medium-high technological complexity under the TC
classification whereas it is differentiated under the Rauch classification. The MicroD classification, by contrast, correctly classifies biodiesel exports by Argentina as
not differentiated and thus not a contributor to upgraded exports. Here,
misclassification also comes from insufficient aggregation. Although biodiesel has
become an increasingly important item in world trade only during the 21st century,
this commodity is still included in 4-digit SITC category “Chemical products and
preparations, n.e.s.”, which also includes a number of differentiated products such
as “Dental cements and other dental fillings; bone reconstruction cements” and
“Anti-knock preparations”. Thus, even if the Rauch classification were updated it
may still misclassify biodiesel as a differentiated product. Similarly, the TC
classification includes biodiesel in the medium-high technological-content category
Qualitatively, similar differences across classifications arise when we perform this exercise
considering exports to all destinations. Vehicle exports become the main desirable export item under
all four classifications (due to the large amount of Argentine vehicle exports to Brazil) but this does
not have any effect on the differences (see Appendix figure A.1).
14
24
because of a similar aggregation problem since this classification makes the broad
simplification of considering all chemical products except pharmaceuticals in this
technological category.
Since the Micro-D classification follows the Rauch classification in focusing on
product differentiation we can assess in further detail the extent to which the former
classification improves upon the performance of the latter. To do this, we look into
three different sets of products: 15 (a) those considered differentiated by both the
Rauch and the Micro-D classifications; (b) those considered differentiated by Rauch
but not by the Micro-D classification; (c) those considered differentiated by the
Micro-D classification but not by Rauch. We identify the most representative
products within each of these categories as those that contribute the most to the
absolute volume of upgraded export growth. We display the top-five items for each
category in Table 1.
15
Products are defined at the SITC 4-digit level including only upgraded exports in each category.
25
Table 1: Upgraded exports with highest absolute export growth. 1998/9-2010/1.
Rauch and Micro-D
Description
Rauch, not Micro-D
Growth
Description
Micro-D, not Rauch
Growth
Motor vehicles
(for transport of
Description
Growth
Wine of fresh
555
Biodiesel
1503
grapes
472
goods)
Fruits, nuts and
Tubes and
277
hollow profiles
other edible
Meat of bovine
226
parts of plants
animals (fresh
243
or chilled)
Airplanes and
184
Hormones
56
Other fruits
105
Petroleum coke
34
Citrus fruit
91
22
Apples
45
aircrafts
Essential oils
88
Parts of vehicles
85
Other alloy steel
Subtotal (top five) Rauch:
in ingots
3016
Subtotal (top five) Micro-D: 2176
Total Rauch:
3091
Total Micro-D:
2469
Subtotal Rauch sums the first two columns. Subtotal Micro-D sums the first and third columns.
Values expressed in US$ millions.
In the first column, we see that both classifications coincide in a number of
manufactures that require a relatively complex production process and are
undoubtedly differentiated. The top three of those products are “Motor vehicles for
the transport of goods”; “Tubes, pipes and hollow profiles of iron or steel”; and
“Airplanes and aircrafts”. The second column highlights products identified as
26
differentiated by Rauch but not by the Micro-D classification. The salient case here
is biodiesel which we have referred to earlier.16 Argentina’s biodiesel exports grew
dramatically – from US$ 3.4 million to US$ 1.8 billion – between 1998 and 2011,
with Spain and Italy as the main OECD customers. Evidence of the high price
elasticity of this product – and hence its low differentiation – is the fact that exports
more than halved to only US$ 837 million (in 2013) after the EU imposed a 20%
countervailing duty on Argentine exports in 2012. Other products identified by
Rauch but not by the Micro-D classification as differentiated include “Fruits, nuts
and other edible parts of plants” “Hormones”, “Petroleum Coke”, and “Steel in
Ingots”. The first consists mainly of peanuts marketed in bulk. The second consists
of follicle-stimulating hormones exported to be used as inputs for medical
compounds injected in fertility treatments. The other products are relatively
undifferentiated intermediate inputs used to produce other products such as fuels,
and steel profiles.
Finally, the third column of Table 1 highlights products that our Micro-D
classification – but not Rauch – identifies as differentiated. These products include
differentiated food and beverages adapted for final consumption such as fresh or
chilled meat marketed in packages smaller than 5 kg, wine sold in bottles of less than
2 liters, and fruits sold in bags (not in bulk). Growth of wine exports was strong and
steady since 2004, with exports mainly oriented to the United States (from US$ 34
million in 2003 to US$ 322 million in 2011) and to Canada (from US$ 9 million to
US$ 84) during the same period. Exports of fresh or chilled meat to the OECD also
The relevant 4-digit HS category is “Other Chemical Products”. We refer to it as Biodiesel since
this product account for 99.99% of total exports in the category.
16
27
registered significant growth during the period increasing from an average of US$
254 in 1998-1999 to an average of US$ 535 in 2010-2011. In this case, Germany and
the Netherlands accounted for a large fraction of these exports (85% in 2010-2011).
Finally, the last three top-growing items correspond to fresh fruits (citrus, apples,
and other fruit) sold in bags below 20kg (not in bulk). Combined exports of these
three items to the OECD grew from US$ 287 in 1998-1999 to US$ 515 in 20102011.
All differentiated items by the Micro-D classification but not by Rauch are good
examples of upgrading opportunities in natural resource-based products that can be
seized by adding value through differentiation. They are also good examples of the
acquisition of knowledge about how to produce and commercialize products to be
sold in high-income countries that is diffusible and hence can generate further
export growth. As it is apparent from the analysis, improving upon the seminal
contribution of the Rauch classification is crucial for reaching a more accurate
assessment of the evolution of Argentina’s differentiated exports to OECD
countries. Based on an inaccurate classification of biodiesel as a differentiated
product, under the Rauch classification we would conclude that chemical products
were the main contributor to upgraded exports during the period 1998-2011. By
contrast, under the Micro-D classification we conclude that the major contributor
was exports of differentiated food and beverages. This implies a very different
assessment about the sources of export upgrading in Argentina during the last years
which may point to very different policy recommendations.
28
5. CONCLUSION
By proposing a classification of differentiated products that builds upon the seminal
work of Rauch (1999) we make a contribution to our ability to identify differentiated
goods that are critical for export upgrading in Latin America. In the Rauch
classification, those goods are not accounted for as differentiated due to insufficient
disaggregation. The classification we propose – the Micro-D classification –
addresses this shortcoming by identifying differentiated goods at a very fine
aggregation level. We apply the Micro-D classification to analyze the evolution of
Argentine exports during the period 1998-2011 and find outstanding differences in
the composition of export growth in differentiated products that have OECD
countries as the destination. While according to the Rauch classification growth in
this type of exports was led by Chemical Products (basically biodiesel), our Micro-D
classification indicates that the leading sector was differentiated Foods and
Beverages. This contrasting reading of export performance has implications for
international insertion policies.
Since the disaggregation requirements are beyond the level of disaggregation in
export nomenclatures currently in use in many Latin American countries, its
application in those cases may be less powerful to identify differentiated products
than illustrated in the case considered here. However, we hope this study will
contribute to build awareness in the academic community and in national statistical
institutes about the importance of keeping track of this type of exports in trade
statistics for assessing a country’s progress in upgrading its export structure.
29
6. REFERENCES
Aggarwal, Aradhna, 2007. "Liberalisation, Multinational Enterprises and Export
Performance: Evidence from Indian Manufacturing," Working Papers id:993,
eSocialSciences.
Artopoulos, Alejandro, Friel, Daniel, & Hallak, Juan C., 2011. “Lifting the Domestic
Veil: The Challenges of Exporting Differentiated goods Across the
Development Divide”. NBER Working Paper No 16947. 105, 19-35.
Artopoulos, Alejandro, Friel, Daniel, & Hallak, Juan C., 2013. “Export emergence of
differentiated goods from developing countries: Export pioneers and business
practices in Argentina”. Journal of Development Economics, 105, 19-35.
Bahar, Dany & Hausmann, Ricardo & Hidalgo, Cesar A., 2014. "Neighbors and the
evolution of the comparative advantage of nations: Evidence of international
knowledge diffusion?" Journal of International Economics, Elsevier, vol. 92(1),
pages 111-123.
Broda, Christian & David E. Weinstein, 2006. “Globalization and the Gains from
Variety,” Quarterly Journal of Economics 121 (May): 541-585.
Feenstra, Robert C., & Andrew K. Rose. 2000. “Putting Things in Order: Trade
Dynamics and Product Cycles”. The Review of Economics and Statistics 82 (3).
The MIT Press: 369–82.
30
Gabriele, Alberto, 1997 “¿Cuán no tradicionales son las exportaciones no
tradicionales? La experiencia de siete países de la Cuenca de Caribe”, Revista de
la CEPAL, N° 63 (LC/G.1986-P), Santiago de Chile, diciembre.
González, Andrea, & Hallak, Juan C. 2013. “Internacionalización de PYMES
argentinas orientadas a segmentos no masivos del mercado en países
desarrollados”. Revista Integración y Comercio (Integration and Trade Journal),
37(17), 13-23.
González, Andrea & Hallak, Juan C. 2016. “Relational linkages for insertion in nonmass global value chains: Opportunities for middle-income countries. Mimeo,
Universidad de San Andrés.
Hagedoorn, John & Cloodt, Myriam, 2003. "Measuring innovative performance: is
there an advantage in using multiple indicators?," Research Policy, Elsevier, vol.
32(8), pages 1365-1379, September.
Hatzichronoglou, Thomas, 1997. “Revision of The High-Technology Sector and
Product Classification” STI Working Paper OECD/GD(97)216, OECD Paris.
Hausmann, Ricardo, Hwang, Jason & Rodrik, Dani, 2007. “What You Export
Matters”. Journal of Economic Growth, 12(1), 1-25.
Hidalgo, Cesar & Hausmann, Ricardo, 2009. “The building blocks of economic
complexity. Proceedings of the National Academy of Sciences”, (106), 10570–
10575.
31
Jarreau, Joachim & Poncet, Sandra, 2012. "Export sophistication and economic
growth: Evidence from China," Journal of Development Economics, Elsevier,
vol. 97(2), pages 281-292.
Kouzmine, Valentine, 2000. “Exportaciones no tradicionales latinoamericanas. Un
enfoque no tradicional”, serie Comercio Internacional N° 7 (LC/L 1392-P),
Santiago de Chile, Comisión Económica para América Latina y el Caribe
(CEPAL) junio, Publicación de las Naciones Unidas, N° de venta: S.00.II.G.65.
Lall, Sanjaya, 2000. "The Technological Structure and Performance of Developing
Country Manufactured Exports, 1985-98," Oxford Development Studies, Taylor
& Francis Journals, vol. 28(3), pages 337-369.
Mesquita Moreira, Mauricio, 2007. "Fear of China: Is There a Future for
Manufacturing in Latin America?" World Development, Elsevier, vol. 35(3),
pages 355-376, March.
Poncet, Sandra & De Waldemar, Felipe, 2013. “Export upgrading and growth: the
prerequisite of domestic embeddedness”. World Development, 51, 104-118.
Rauch, James, 1999. "Networks Versus Markets in International Trade," Journal of
International Economics 48(1) (June 1999): 7-35.
Srholec, Martin, 2007. "High-Tech Exports from Developing Countries: A
Symptom of Technology Spurts or Statistical Illusion?" Review of World
Economics (Weltwirtschaftliches Archiv), Springer, vol. 143(2), pages 227-255,
July.
32
Stehrer, Robert & Woerz, Julia, 2009. "Industrial Diversity, Trade Patterns, and
Productivity Convergence," Review of Development Economics, Wiley
Blackwell, vol. 13(2), pages 356-372, 05.
Von Hesse, Milton, 1994. “Políticas públicas y la competitividad de las
exportaciones agrícolas”, Revista de la CEPAL, N° 53 (LC/G.1832-P), Santiago
de Chile, agosto.
Appendix A: Detailed description of the Micro-D classification
Before we describe the Micro-D classification in detail, we note that there were two
changes in the HS during our sample period (1998-2011). Specifically, we had to
work on the correspondences between the HS of 1996, 2002, and 2007 and on the
trade flows assigned to each tariff code that changed in order to guarantee
consistency over time. In general, the newer the nomenclature the more it
disaggregates products, so there are some cases in which tariff codes assigned to
different categories in the more recent classification are grouped together in a
previous SIM. In those cases, considering the previous SIM code as either
differentiated or not differentiated would be misleading since this code likely
includes both types of products. This problem is particularly important for meat and
fruit products in the first two years of our database (1998 and 1999) because some
trade flows were registered using the HS1988/92 nomenclature rather than the HS
1996 one. To deal with this issue, we assigned the same differentiated/non-
33
differentiated shares observed in 2000, at the 8-digit level and destination group
level, to exports reported under these troublesome codes in 1998 and 1999.
Now we proceed to describing in detail the Micro-D classification. The complete
classification for all HS 2007 product lines is available in a Stata file on the authors’
websites.

Agricultural Products, Food, and Beverages (HS 01-24): Products in these chapters are
generally classified according to “package size”. They are differentiated when
they are traded in small packages while they are not differentiated when they are
traded in bulk or in larger containers. Package size thresholds depend on the
specific product. Some food products are classified as differentiated regardless
of package size because they are not traded in bulk. This is the case, for
example, of dulce de leche, candy, champagne, preparations for infant use,
bakers’ wares, pasta, breaded meat, and products frozen using Individual Quick
Freezing (IQF) processes. Live animals are considered differentiated if they are
pure-bred or for breeding.

Minerals Products (HS 25-27): All products in these three chapters are classified as
not differentiated. These chapters primarily include raw materials and their
concentrates.

Products of the Chemical or Allied Industries (HS 28-38): Products in these chapters
are classified according to their position in the production chain. In particular,
34
raw materials are classified as not differentiated while final products are
differentiated. By chapter, the classification proceeds as follows:
o Inorganic and organic chemicals (HS 28-29): All products are classified as not
differentiated. These products are inputs in the production of other
products.
o Pharmaceutical Products (HS 30): All products are classified as differentiated.
o Fertilizers (HS 31): Only mineral and chemical fertilizers exported in packages
of 10 kg or less are differentiated. All remaining products are not
differentiated.
o Tanning or dying extracts, pigments, and other coloring matter (HS 32): Tanning
extracts and substances and coloring matters are considered not
differentiated. All remaining products (prepared pigments, paints, varnishes
and driers) are considered differentiated.
o Essential oil, perfumery, cosmetic, and toilet preparations (HS 33): All products are
classified as differentiated
o Soap, organic surface-active agents (HS 34): All products are classified as
differentiated
o Albuminoidal substances; modified starches; glues; enzymes (HS 35): Caseinates
prepared for final consumption and glues are differentiated. Al remaining
products are considered not differentiated.
o Explosives; pyrotechnic products; matches; pyrophoric alloys; certain combustible
preparations (HS 36): Propellant powders and prepared explosives are not
differentiated. Al remaining products are differentiated.
35
o Photographic or cinematographic goods (HS 37): All products are classified as
differentiated
o Miscellaneous chemical products (HS 38): Insecticides, finishing agents, pickling
preparations, anti-knock preparations, prepared rubber accelerators,
preparations and charges for fire-extinguishers, organic composite solvents,
reactions initiators, chemical doped for use in electronics, hydraulic brake
fluids, anti-freezing preparations, prepared culture media for the
development or maintenance of micro-organisms and diagnostic or
laboratory reagents on a backing are considered differentiated. All other
products are not differentiated.

Plastics and Rubber Products (HS 39-40): Primary forms such as polymers, cellulose,
silicone, and natural or synthetic rubber are classified as not differentiated.
Manufactures made from these primary inputs are differentiated (pneumatic
tires are a prominent example).

Hides, Skin, and Leather (HS 41-43): Raw hides and skins (HS 41) are not
differentiated. Articles of leather (HS 42) are differentiated. Raw, tanned or
dressed furskins are not differentiated. Articles of apparel and artificial fur are
differentiated.

Wood and Paper Products (HS 44-49):
o Wood and wood articles (HS 44): Fuel wood, wood wool, railway or tramway
sleepers of wood, wood sawn and sheets of veneering are considered not
36
differentiated. Hoopwood, wood continuously shaped, particle board,
fibreboard of wood, plywood, densified wood, wooden frames for paintings;
packing cases, casks, barrels, vats, tubs and coopers’ products of goods; and
tableware and wood marquetry are considered differentiated. Sawdust, wood
waste and wood charcoal are classified based on package size.’
o Cork and articles of cork (HS 45): Natural cork, raw, simply prepared, debarked
or roughly square is classified as not differentiated. All other products are
differentiated.
o Manufactures of straw (HS 46): All products are differentiated.
o Pulp of wood (HS 47): All products are not differentiated.
o Paper and paperboard (HS 48): Newsprint, toilet or facial tissue stock, uncoated
kraft paper and paperboard, vegetable parchment, paper and paperboard
corrugated, creped, crinkled, embossed or perforated and filter paper and
paperboards are not differentiated. Carbon paper; paper and paperboard
coated, impregnated, covered or printed; filtered blocks, wallpapers,
envelopes and letter cards, toilet paper, cartons, boxes, cases and bags,
registers, paper and paperboard labels and paper pulp are differentiated.
Uncoated or coated paper and paperboard and cigarette paper with kaolin is
classified base on package size.
o Printed books, newspaper and pictures (HS 49): All products are differentiated.

Textiles and Clothing Products (HS 50-67)
o All products in the following chapters: Silk (HS 50), man-made filaments
(HS 54), man-made staple fibres (HS 55), wadding, felt and nonwovens
37
(HS 56), carpets (HS 57), special woven fabrics (HS 58), impregnated,
coated, covered or laminated textile fabrics (HS 59), knitted or crocheted
fabrics (HS 60), articles of apparel (HS 61 and 62), other made-up
textiles articles (HS 63), footwear (HS 64), headgear and parts (HS 65)
and umbrellas (HS 66) are classified as differentiated.
o Wool (HS 51), cotton (HS 52), other vegetable textile fibers (HS 53) and
feathers (HS 67) are divided between their primary forms (as cotton or wool
not carded or combed, cotton waste, raw flax or skins and other parts of
birds), classified as not differentiated, and their elaborated forms (as yarn or
woven fabrics), classified as differentiated.

Articles of stone, ceramic and glass (HS 68-70): All products are classified as
differentiated.

Base and Precious Metals Products (HS 71-83)
o Natural or cultured pearls, precious or semi-precious stones and precious metals (HS 71):
Unwrought metals and pearls, diamonds and precious stones not worked are
considered not differentiated. Manufactured or semi-manufactured forms of
metals, precious stones or pearls are considered differentiated.
o Iron and steel (HS 72): Pig iron, ferro-alloys, ferrous products obtained by
direct reduction of iron ore, ferrous waste, granules and powders, iron and
non-alloy steel in ingots, and semi-finished products of iron are not
considered differentiated. All other products are considered differentiated.
o Articles of iron and steel (HS 73): All products are classified as differentiated.
38
o Copper and articles thereof (HS 74): Copper mattes, unrefined copper, refined
unwrought copper, copper waste, master alloys of coppers and copper
powders are not differentiated. All other products are differentiated.
o Nickel and articles thereof (HS 75): Nickel mattes, unwrought nickel, nickel
waste and nickel powders are not differentiated. All other products are
differentiated.
o Aluminum and articles thereof (HS 76): Unwrought aluminum, aluminum waste
and aluminum powders are not differentiated. All other products are
differentiated.
o Lead and articles thereof (HS 78): Plates, sheets and strips of lead; bars, rods,
profiles, tubes and pipes of lead are differentiated. All other products are not
differentiated.
o Zinc and articles thereof (HS 79): Unwrought zinc, zinc waste and zinc dust are
not differentiated. All other products are differentiated.
o Tin and articles thereof (HS 80): Unwrought tin and tin waste are not
differentiated. All other products are differentiated.
o Other base metals (HS 81): Unwrought or powder metals are not
differentiated. Elaborated forms as plates, sheets, strips and wire rods are
differentiated.
o Tools, implements, cutlery, spoons and forks (HS 82): All products are classified as
differentiated.
o Miscellaneous articles of base metal (HS 83): All products are classified as
differentiated.
39

Machinery and Appliances (HS 84-93): All products are classified as differentiated.
 Miscellaneous Manufactured Articles and Works of Art (HS 94-97): All products are
classified as differentiated, except collections and antiques because despite being
differentiated they are not reproducible.
APPENDIX B: Construction of product groups
In this appendix we describe the composition of product groups that are used in
figure 6. These groups are constructed based on the 2-digit codes of the
Harmonized System, as detailed in Table 2.
Table 2: Correspondence between product groups and 2-digits HS codes
Product Groups
2-digit HS codes
Food and Beverages
02, 03*, 04, 07, 08, 09*, 10, 11, 12, 15, 16, 17, 18, 19, 20, 21, 22
Other Agric. Products
01, 03*, 05, 06, 09*, 13, 14, 23, 24, 25*, 26*, 31*, 35, 40*, 44*, 45, 46, 99*
Textiles
41, 42, 43, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64
Chemicals
28, 29, 30, 31* 32, 33, 34, 36, 37, 38
Plastic and Rubber
39, 40*
Paper
26*, 47, 48, 49
Precious Metals
68, 69, 70,71
Metals
72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83
Machinery
84, 85, 88*, 89
Vehicles
86, 87, 88*
Other Industrial Products
25*, 44*, 65, 66, 67, 90, 91, 92, 93, 94, 95, 96, 97, 99*
Fuels
27, 99*
2-digits HS codes followed by * are not fully included in the product group
40
90% of 2-digit HS codes are fully included in a single product group. There are nine
exceptions as discussed below:

03 – Fish and crustaceans: All items are included in Food and Beverages, except Pacific
Salmon which is included in Other Agricultural Products.

09 – Coffee, tea, maté and spices: All items are included in Food and Beverages, except
Coffee not Roasted which is included in Other Agricultural Products.

25 – Salt, Sulphur, Earth & Stone, Lime & Cement: All items are included in Other
Agricultural Products, except Quicklime and Portland Cement which is included in
Other Industrial Products.

26 - Ores, slag and ash: Copper ores are included in Metals. Other minerals are included
in Other Agricultural Products. Roasted Iron Pyrites and Slags are included in Other
Industrial Products.

31 – Fertilizers: Animal and vegetable fertilizers are included in Other Agricultural
Products. The remaining items are included in Chemicals.

40 - Rubber and articles thereof: Natural rubber is included in Other Agricultural
Products. All remaining items are included in Plastic and Rubber.

44 – Wood and articles of wood: Primary forms of Wood are included in Other
Agricultural Products. Products that display some kind of elaboration are included in
Other Industrial Products.

88 - Aircraft, spacecraft: Spacecraft and multiengine airplanes are included in Vehicles.
All remaining items are included in Machinery.

99 – Others: Supplies of fuels and lubricants to ships and aircraft are included in Fuels.
Other supplies to ships and aircraft and simplified exports are included in Other
Agricultural Products. All other products are included in Other Industrial Products.
41
Figure A.1: Contribution to export growth. All destinations. 1998/9-2010/1.
20
16
12
8
4
0
INDEC
Food and Beverages
TC
Chemicals
Rauch
Precious Metals
42
Micro-D
Vehicles
Others