ratio slightly above and below 1. In addition, Tunisia with an average of about 3500 and
a moderate increase of its exports (ratio 1.4).14
Cluster 4 (small exporters, winners): Cluster 4 groups four small meat exporters with
strong gains in export values. The ratio has a value of 3.0. Nevertheless, their coefficient
of variation is still relatively high with 0.54. Morocco is the largest exporter of this
cluster, the other countries being very small meat exporters. This is also depicted by the
"difference" which has only an amount of 1054. Examples for the cluster are Morocco
and Algeria.
Cluster 1 (Small to medium exporters, strong winners): This cluster consists of small to
medium exporters (Malaysia being the largest exporter). The cluster is characterized by
an extremely high ratio of 7.8, which implies very strong gains of these countries within
the last period. Nevertheless, the coefficient of variation is with 54 still relatively high.
Country examples of the cluster are Cote d'Ivoire and Malaysia.
Cluster 2 (medium exporters, extremely strong losers): Cluster 2 includes six medium-
sized meat exporters, mainly from Middle America, that faced extremely sharp losses of
exports. The group is characterized by a ratio of only 0.1 and a "difference" of more than
- 20000. Due to the sharp drop, in the period 2002- 2004 these countries are rather small
meat exporters Most stable, only halving its exports among this group is Costa Rica. The
other countries almost completely lost their share in exports (e.g. Honduras). The group
is as well characterized by a high coefficient of variation (0.73). However, the results of
the cluster must be interpreted very carefully, since during this period “El Nino” strongly
affected central American countries.
Cluster 3 (medium exporters, extremely strong losers): the cluster shows several
similarities to cluster 2. It groups the two countries Rep. of Korea and Zimbabwe. The
two countries were medium or even large exporters (average of about 70,000) in the first
period, but exports dropped sharply. The cluster as well shows a ratio of 0.1 and a
"difference" of nearly - 67000. The coefficient of variation is with 0.10 relatively low.
Nevertheless, it must be noted as well, that especially Zimbabwe went through a
politically very instable period which implied, that it had almost a complete breakdown
of its agricultural export structure.
Cluster 30 (large exporters, strong winners): the cluster includes the two large exporters
Chile and Mexico. Among the total group of large exporters they are the smallest ones.
However, the two countries show with a ratio of 14.1 the strongest gains of the total
sample. With 28.0 they have a relatively low coefficient of variation.
Cluster 10 (large exporter, loser): this cluster includes the only exception of the large
country exporters. Argentina is the only large country which shows losses on a very high
level. With a ratio of 0.7 and the high losses, especially in absolute terms, as it is
depicted in the "difference" Argentina stands clearly apart as a strong looser.
Cluster 20 (very large exporters, winners): Brazil, Thailand and China (in this order) are
the largest exporters. On such a high level of exports a ratio of 1.9 implies tremendous
gains in total terms.
Overall, from the cluster analysis of developing countries according to their meat
exports to OECD countries the meat sector is found to be more difficult for developing
countries participation. This became evident by three reasons: 1) the total number of
countries which participate on the market and thus have regular data on trade flows is
relatively low. 2) the market is very much dominated by only six large countries. Except
for Argentina all large countries even increased their export share. 3) from a sample of
46 countries 30 countries are found in "loser" clusters. Only 8 small to medium
countries are located in "winner" clusters. In the meat sample only 4 countries are LDC
countries. There from only Mozambique is covered in a cluster of very large winners. All
other LDC countries are found in cluster 5, the small loser cluster. The tendency of the
sector to be dominated by very few large countries seems to be even straightened by
higher food safety and quality measures.
What are finally the differences between the two sectors? Table 5 depicts in a cross-
tabulation which countries are winners or looser either in one of the sectors or even in
both. As the sample of the meat sector is smaller that of the fruit/ vegetable sector the
table includes a raw labeled "no cluster".
Table 5: Cross- tabulation of „losers“ and „winners“ in the meat and the fruit/
vegetable market
14 We find these countries remaining in this group even when going up with the number of clusters
up to nine clusters (compared to the chosen number of five clusters).
13