A comparison of Multi-Regional Input-Output databases regarding transaction structure and supply chain analysis

Abstract

Abstract Multi-Regional Input-Output (MRIO) databases might be used to derive information of the regional and sectoral distribution of otherwise unknown supply chains (SC). There are several MRIO databases available that were created using different approaches. This paper evaluates MRIO databases with regard to (1) their comparability and consistency in general and (2) the best way to apply them to enable SC analysis. Four MRIO databases were analyzed: Eora, EXIOBASE, GTAP and WIOD. They were compared based on (A) transaction structure and (B) SC analysis results. (A) was done by transforming the databases to a unified form and determining the relative differences of their transaction values. For (B), a SC analysis based on cumulated intermediate consumption (CIC) and one based on structural path analysis (SPA) were carried out and compared with the respective real life SC. The results showed that around 80% of the total transaction volume in MRIO databases was intra-regional transfers. Inter-regional transfers varied strongly between databases but overall similarity between databases was high when considering the total transaction volume. This was supported by SC analysis based on CIC. The SC analysis by production tier varied significantly between databases with only 50–63% of simulated suppliers being from the same region or sector. The regional distribution of value added also showed clear differences between databases. The indication by the case study showed that the analysis based on CIC achieved more overall conformity with the real SC but both ways of analysis only capture a fraction (14%) correctly. These results led to the conclusion that the preferred mode of MRIO analysis for supporting subsequent SC assessments would be based on CIC but that neither mode would be recommended for detailed SC analysis as shown for the example of Social LCA. Future research should try to look at comparing disaggregated MRIO databases, analyze the creation processes of MRIO databases and clarify why the share of intra-regional transactions was consistent across databases but surprisingly high.