This deliverable evaluates the case studies that are part of Task 3.8. Our main focus is on developing country impact analysis. This focus reflects the stress in the DOW on developing countries. Task 3.8 however has a slightly broader goal of analyzing the third country impacts of EU agricultural policies. According to the first deliverable of Task 3.8 (PD 3.8.1) we will also assess the competitiveness of EU agriculture vis-à-vis EU’s main competitors within Task 3.8. For completeness we devote limited space to indicating the available data for this competitiveness assessment, restricted to the GTAP component of this assessment. For mote details on the way in which the competitiveness of EU agricultural policy will be assessed we refer to PD 3.8.3, discussing the linking of CAPRI and GTAP which will be instrumental for this assessment.

The aim of this deliverable is two-fold. As outlined in the description of work (DOW) it evaluates the developing country case study for Mali. In the case of Mali detailed data are available, due to past research projects of CIRAD. Mali, however, may not necessarily be representative of all developing countries in terms of its key features, nor in terms of data availability. The focus of Task 3.8 is on developing a general methodology for analyzing impacts on developing countries. This methodology should be applicable in different countries with generally available data. Apart from describing the data available for Mali we thus also make an inventory of data available for other developing countries. This will allow us to develop a methodology for the Mali-case, as planned for in the project proposal, while assuring that the methodology may also be applied elsewhere by aligning the methodology with publicly available data for developing countries.

We start with a short discussion of the GTAP model to indicate the coverage and indicators available at the global level. An overview of the coverage of GTAP may also be important for selecting case study countries. Linking to the global level model is easier if the case study country is represented as such in the GTAP database, as opposed to being part of an aggregate. In the latter case, of which the Mali case study is an example, it is harder to link the global changes to a case study country. Coverage as a single country in the GTAP database is also important for the development of national level CGE models in case there is no more detailed Social Accounting Matrix (SAM) available. In such an instance the SAM included in GTAP may provide a consistent starting point for the modelling work.

The main part of the report is devoted to describing the different model elements for analyzing the impact on developing countries, their data requirements and the type of indicators supplied by each of the models. National level CGE models have a similar role as CAPRI for the analysis of the changes in Europe. The national CGE models will be based on a model template developed at the World Bank (MAMS). Apart from a well-developed starting point this link also provides support in terms of estimating model parameters and model development. A national Social Accounting Matrix (SAM) is required to develop a CGE model. IFPRI has made several SAMs available, and as suggested above the GTAP database provides a set of (aggregated) SAMs as well. CIRAD is in the process of completing a SAM for Mali, which will allow application of a CGE model for Mali.

Poverty is a key aspect of assessing the impact on developing countries. Micro-simulation models, linked to the national level CGE models, will be adopted for assessing the poverty impacts. Micro-simulation models rely upon household expenditure surveys, preferably in combination with census data to assure national coverage of the analysis. The World Bank provides a set of household level surveys suitable for micro-simulation analysis. Census data are publicly available only for a limited set of countries. Most countries however perform census on a more or less regular basis and these may be obtained by contacting national statistical bureaus. In the case of Mali a 1985 national level census is available, as well as a set of household surveys.

As in the case of the analyses for Europe assessing the impact on agricultural sustainability is a key point in the developing country analysis. This will be addressed through a tropical version of the FSSIM model which will yield a similar set of sustainability indictors as derived from the European FSSIM models. Data requirements for developing a FSSIM model are not only high but also entail a set of data not collected on a regular basis by international institutions. In the case of Mali sufficient agricultural data are available and are currently used to develop a FSSIM model for the cotton area.

By starting from a broader perspective on the developing country analysis we assure that the specific model components used for the Mali case study can be used in other settings as well. The model components will be developed in a modular framework to accommodate the absence of a specific dataset in different settings. Based on our inventory the farm level data required to develop a FSSIM model appears to be most limiting factor for applying the methodology elsewhere. Another concern is the variation in the years in which data are collected. This implies that even if all required dataset are available for a specific country, we will need to reconcile data from different years. Overall we may conclude that Mali will be a valuable case to study the implementation of the
methodology, posing mainly challenges in terms of linking to the GTAP model which does not distinguish Mali as a separate country. In terms of the other datasets Mali appears representative for the data availability in developing countries thus providing good testing ground for the methodology development.