Aggregation error of the material footprint: the case of the EU

ABSTRACT

The material footprint (raw material consumption) was proposed as a basis for monitoring SDGs 8.4 and 12.2. However, there is no institutionalized procedure providing globally consistent national material footprints. The OECD aims to institutionalize the material footprint through the development of one official inter-country input–output (ICIO) database applicable for its calculation. Inherent to input–output analysis is the aggregation error, which may impair the results. Therefore, in the case of the EU I analyze the aggregation error which can be expected if NACE rev2 classification is utilized for this ICIO database, and investigate the most important disaggregations, depending on the desired focus of the results. I conclude that the disaggregation level should reflect the intended purpose of the RME indicators. For their deeper analysis, and determination of strategies for their decrease, I conclude that NACE rev2 classification is inappropriate, and recommend high disaggregation and utilization of hybrid units.

KEYWORDS:

• Aggregation error
• material footprint
• raw material consumption
• RMC
• input-output analysis

Acknowledgement

I thank to Karl Schoer, Monika Dittrich, Birte Ewers, Sonja Limberger, Stephan Moll and Maaike Bouwmeester for their valuable inputs to this analysis and suggestions on the initial aggregation levels.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 For a more extensive literature review see Lenzen (2019).

2 Is should be noted that it is also the expected requirement for precision of the resulting indicators which favours Lenzen’s approach for detailed material categories rather than for the total RMC, as I assume that the precision for RMC should be considerably higher than the precision for detailed material categories (see the ‘Method’ section for more details). This results in higher requirements on the number of SPA paths which need to be explored and subsequently also the number of unique product groups which should be distinguished in order to reach this degree of precision.

3 Even though Exiobase has a higher number of product groups (200), its classification is not specifically targeted to the material footprint indicator and it does not contain some important disaggregations included in the Eurostat RME model (e.g. in Exiobase mining distinguishes 23 product groups and 13 basic metals).

4 Lenzen (2019) provides an overview of existing methods that have been applied to find minimum disaggregation requirements. Within his list, my approach is denoted as ‘intuition’.

Additional information

Funding

This work was supported by Grantová Agentura Ceské Republiky [grant number 19-26812X].