Differences between GTAP and TiVA databases and their Implications for Trade in Value-Added Indicators

Fragmentation of global production requires measurement of trade in value added, which accounts for the way slices of value embedded in goods or services are added at each step of increasingly international manufacturing processes within global supply chains. The quality of indicators measuring trade in value added finally depends on the quality of the underlying global Inter-Country Input-Output tables which, in turn, depend on the quality and availability of underlying national statistics and the balancing and estimation techniques used in the harmonization procedure. Although the use of statistics based on these tables has become more routine in economic research, there is neither harmonization among different global databases nor a standard in the construction of data on which trade in value added is estimated. This contribution is an attempt to systematize the potential sources of the differences between databases and to show how estimations of trade in value added are impacted. Specifically, it provides a comparison between two of the main global databases used for macroeconomic analysis of global value chains and trade in value added patterns, namely, the GTAP Data Base and the OECD-WTO TiVA database, and provides a discussion of the reasons for the deviation in the estimates of global value chains-related trade. In the first part, a description of the databases and the underlying construction methodologies is provided. In the second part, the differences between the two databases in structural economic variables are presented. Finally, major TiVA indicators are estimated based on the different databases and discrepancies among them are discussed