Analysis of Agri-Food Trade Structures
To Promote Agri-Food Trade Networks
In the Islamic Countries
9
a node by considering its susceptibility to positive or negative external shocks. The score
represents the value of a $1 shock to the trade network at a random point, after it has fully
propagated and the system has returned to an equilibrium state. More central countries are
more subject to positive and negative economic effects coming from outside. But
mathematically, this measure is exactly equivalent to a conception of a country’s centrality in
which its score is an average of the scores of all other countries it is directly linked to. Although
more complexmathematically, the intuitive nature of this measure has made it widely applicable
in other contexts. For instance, Google uses a modified version of eigenvector centrality in its
PageRank algorithm that ranks internet searches, and it has been shown that a closely related
measure can help explain aggregate economic fluctuations based on input-output linkages
across sectors.8
The second aspect of the methodology that requires explanation is the approach to measuring
comparative advantage in agri-food sectors and tracking it over time. Traditionally, applied
trade policy researchers have used measures like Revealed Comparative Advantage (RCA).
However, they are unsatisfactory because they do not correspond to standard conceptions of
comparative advantage except in a world without sector-specific trade costs. More generally, a
measure like RCA does not have any theoretical foundation, but is presented simply as a
convenient summary statistic. However, the recent literature has shown that it is possible to
develop a measure of comparative advantage that is consistent with standard Ricardian theory,
and which can be easily implemented using standard data.9 The method relies on estimating a
standard gravity model using sectoral data, then reparametrizing the exporter-specific terms to
produce an indicator of comparative advantage that has a strong grounding in widely accepted
theory. This indicator is referred to as TRCA, to avoid confusion with the traditional measure.
The analysis of comparative advantage in this project follows that approach. Because of the large
number of parameters involved in estimation of the gravity model, it is not feasible to produce
TRCA estimates at the most disaggregated level of Annex 1: computation is too long to be
workable, and pushes the limits of what is possible in standard statistical software. The division
level of Annex 1 is therefore used, which still produces a regression with 91,150 observations,
and 4,819 independent variables, as well as 10,331 fixed effects removed by demeaning.
When analyzing trade flows at the global level, i.e. all countries together, regional groupings
established by the World Bank are used. These groups are widely used in economic analysis by
international organizations and researchers. It is appropriate to start the analysis at this general
level, because it highlights the role of different geographical areas in relation to world
agricultural trade. To provide further detail for OIC member countries, the analysis uses OIC
regional groupings in the section on trade performance of OIC member countries, as it is more
informative to distinguish between OIC and non-OIC countries at that stage.
1.3. Case Study Methodology
In addition to reviewing data, it is also important to learn from concrete experience on the
ground in each of the OIC’s three regional groupings. The countries studied here are Bangladesh
(Asia), Cameroon (Africa), and Tunisia (Arab). The rationale for choosing these three countries
is that agriculture plays a significant role in the economy of each, yet they also represent diverse
examples from the point of view of income level, climate and major crops, as well as trade
8 Acemoglu, D., V.M.Carvalho, A. Ozdaglar, and A. Tahbaz-Salehi. (2012). “The Network Origins of Aggregate Fluctuations.”
Econometrica, 80(5): 1977-2016.
9 Costinot, A., D. Donaldson, and I. Komunjer. (2012). “What Goods do Countries Trade? A Quantitative Exploration of Ricardo’s
Ideas.” Review of Economic Studies, 79(2): 581-608.