Analysis of Agri-Food Trade Structures
To Promote Agri-Food Trade Networks
In the Islamic Countries
22
suggests a greater degree of convergence in regional diversification behaviors than the results
using out degree centrality.
Using out degree centrality, it is possible to compute an overall level of centralization of the
network. The score on this metric is 0.673, which means in essence that the centrality score of
the leading country compared with the centrality scores of the other countries is nearly 70% of
what would be observed in the most centralized network that could be constructed with these
data. The conclusion is that the global network of trade in agri-food products is therefore
relatively centralized, with countries in North America and Europe and Central Asia acting as
major global export hubs.
Table 2: Centrality Scores for Agri-Food Products, by Region, 2016
Out Degree Centrality Eigenvector Centrality
East Asia & Pacific 68.111 0.054
Europe & Central Asia 114.720 0.078
Latin America & Caribbean 61.128 0.058
Middle East & North Africa 70.667 0.074
North America 137.000 0.082
South Asia 67.875 0.065
Sub-Saharan Africa 35.957 0.052
Source: Authors’ calculations.
For the case of agricultural raw materials (Figure 10), trade is again strongly intra-regional
(predominance of red lines), although some countries also have important inter-regional links.
Regional coherence is weaker in this case than was the case for agri-food products, although
there is more evidence of a hub and spoke structure. Sub-networks are centered on Europe
(Germany), the Americas (United States), and Asia (China). There are also some countries not
connected to any major hub either directly or indirectly.
Table 3 shows that the pattern of out degree centrality increasing with per capita income is
confirmed for this product group. However, export diversification is more limited for
agricultural raw materials than for agri-food products, with the most central region, North
America, exporting to 110 countries rather than the 137 observed above. Eigenvector centrality
(column 2) confirms these observations, although the differences across regions are again less
pronounced than for out degree centrality.
Table 3: Centrality Scores for Agricultural Raw Materials, by Region, 2016
Out Degree Centrality Eigenvector Centrality
East Asia & Pacific 45.611 0.052
Europe & Central Asia 78.300 0.081
Latin America & Caribbean 35.846 0.048
Middle East & North Africa 34.286 0.071
North America 110.667 0.087
South Asia 52.625 0.067
Sub-Saharan Africa 21.826 0.040
Source: Authors’ calculations.