This paper outlines an approach to disaggregating labor payments in the GTAP, global economic data base. The split between low- and high-skilled labor is based on occupational data. High-skilled labor is assumed to consist of managers, administrators, professionals, and para-professionals.

Data are gathered on this occupational split, by sector, in fifteen different economies, and these are mapped to GTAP sectors. Regression analysis shows a systematic relationship between GDP per capita and the national stock of tertiary and secondary educated labor on the one hand, and the sectoral labor payments split on the other.

This model is used to predict labor splits, by sector, in the remaining GTAP regions. The results are evaluated in terms of the implied economywide skilled -unskilled labor payment ratio. Overall, the results seem promising enough to warrant inclusion in the GTAP, version 4 data base.

Date of this Version

September 1998