Analysis of Agri-Food Trade Structures

To Promote Agri-Food Trade Networks

In the Islamic Countries

CONCEPTUAL FRAMEWORK AND METHODOLOGY

This section describes the conceptual framework and methodology for the report. The section

first addresses the issue of data, looking at the key databases on global agricultural trade, and

justifying the choice to use UN Comrade as the primary data source. The discussion then moves

to the methodologies applied for analyzing the data, moving from descriptive statistics to

network analysis, and a model-based derivation of indicators on comparative advantage. Finally,

there is a discussion of the methodology adopted for the country case studies in Section 4.

1.1.

Data Sources

Numerous data sources are available covering various aspects of international trade and

production of agri-food products. However, not all are suitable for the type of analysis required

by this report. A key requirement in this case is that trade flows be made available at a highly

disaggregated product level, so that flows can be matched to a product classification that is

appropriate for OIC member countries. Annex 1 contains the product classification used in this

report, and it is referred to throughout as “the Annex 1 classification””. It is based on the

Standard International Trade Classification Revision 3 (SITC Rev.3). Second, the methodologies

followed here—including network analysis, and a model based analysis of comparative

advantage—require bilateral data, i.e. flows that move from an exporter to an importer. As such,

databases that record trade with an aggregate “rest of the world” region cannot be used in this

case.

In the agri-food context, a common starting point is FAOStat. Historically, this database has

focused on production data, not trade, and included only exports and imports with an aggregate

“rest of the world” region. The current implementation, however, includes bilateral trade data,

thereby satisfying the second requirement set out above. However, the product classification

used is the Food and Agriculture Organization´s (FAO)s own classification, which is useful for

certain purposes, but which does not easily map to the SITC Rev.3, which the Annex 1

classification uses to identify the products needed for this study. The data would therefore need

to pass through a concordance to identify flows in accordance with the objective of this Study.

There would likely be some loss of breadth and detail, as the FAO database focuses on raw and

lightly transformed products, thereby excluding some products listed in Annex 1. The

conclusion is therefore that FAOStat is not an appropriate data source for this project.

The standard source for bilaterally disaggregated international trade data tracked by SITC

product is UN Comtrade. The data are typically accessed by researchers using the World Bank’s

World Integrated Trade Solution (WITS) interface. UN Comtrade contains export and import

data over a long period, as reported by countries all around the world. 54 out of 57 OIC member

countries report data to UN Comtrade, and all 57 member countries are included as partners

(those with whom flows are reported). The WITS interface makes it possible to download trade

data in a variety of product classifications, including SITC Rev. 3. From that product level

information, it is then possible to aggregate data into the groups identified in Annex 1. Data are

available on both reported exports and reported imports.

An alternative to UN Comtrade is the International Trade Center´s (ITC) TradeMap. This

database contains a wealth of information, but is destined for use by the private sector rather

than researchers. As a result, it does not facilitate the large-scale download of data across a large

number of products and countries. It also records data using the Harmonized System, but does