I-O Table Submission
Notes on procedures
Format specifications
Problem areas
What to send us?
What do we check?

GTAP Data Bases: How Much to Disaggregate?

It is generally assumed that you have better information than we do about the particular country you are supplying data for.

In general if you have any information/ideas about the correct proportions which should be used to split a sector then you should use this information to split the sector. Otherwise we will split the sector using a representative table or in the case of the agricultural sectors, data that we obtain from the FAO.

That said if you are disaggregating agriculture and all you have is a share for splitting the total sales or costs and no additional information on the different cost structures of the alternative commodities then it may be better to leave them aggregated and allow the GTAP team to disaggregate based on the FAO data collected.  The reason for this is that agricultural commodities may have significantly different cost structures, which will not be reflected in the data if you assume the same cost structures for both. An example of this is disaggregating bovine meat products (cmt) and other meat (omt). One would expect that cattle (ctl) are the primary inputs in cmt, while other animal products (oap) are the primary inputs into omt. This difference in cost structures is important for analysis.

To learn more about how we disaggregate the I-O tables see the GTAP Data Base documentation.

Representative Table
The representative table referred to above is a weighted sum of tables for primary GTAP regions (not including composite regions). The weights are such as to ensure that each region is represented in proportion to its GDP. To construct the representative table we use tables on hand from the last version of the Data Base.