The CO₂lumnist: Data providers and the greenhouse gas pie

The recently published IPCC report shows that we are globally emitting around 59 gigatons of CO₂e each year. As an responsible asset manager we try to reduce the carbon footprint of our portfolios by excluding, or reducing the weight in, companies that have dirtier emission profiles. To be able to implement this, we need of course company emission data for calculating the carbon footprint of our investment portfolios. There are several data providers that deliver this for company emissions.

As a spin-off from an internship project, I started to look at how much of this 59 gigaton of CO₂e is actually covered by these data providers’ emissions data. That is what this month’s memo is all about. I use Trucost data for the analysis, but using other providers’ data will result in similar conclusions. I will only look at scope 1 data for this analysis to exclude double counting, as scope 2 and 3 are indirect emissions.

The total aggregate of scope 1 for the whole Trucost universe is equal to 23 gigatons of CO₂e. This implies that the companies in the Trucost universe cover roughly 40% of all global direct emissions. This is mainly caused by Trucost (and data providers in general) focussing on the larger corporate institutions, and hence missing smaller company, governments related, and residential emissions in their universe.

The figure above splits both the IPCC global and the aggregated Trucost emissions into the five primary IPCC industry sectors.¹ The energy sectors emits, unsurprisingly, the most of the five sectors with more than 20 gigatons of CO₂e. All the energy companies in the Trucost dataset together emit almost 15 gigatons of CO₂e, which is a bit less than 75% of the global emissions from the energy sector. This implies that for the most polluting sector a rather large part of the emissions is covered by emission data available to us.

If we look at the other sectors a different pictures emerges. For industry still around half of the global emissions are covered, but for the other three categories, the percentage of coverage becomes very low. For Agricultural, Farming and Other Land Use (AFOLU), the lowest coverage of all is found. Most farms or agricultural companies are privately owned and are usually pretty small, which explains why the coverage is so low for AFOLU.² For transport and buildings, I expect a substantial part to be emitted by smaller private companies as well, but here also commuting and residential buildings will be significant.

I think it is important for us as asset managers to have a rough impression on which part of the global greenhouse gas pie we invest in. I believe the 40% from the example above serves this purpose pretty well, but we also invest in companies that are not in the Trucost universe, and there are companies in this universe that we do not invest in. Depending on the net effect, this 40% will be somewhat higher or lower.

Footnotes

¹There is no direct mapping available from GICS sectors, which I use for Trucost data, to IPCC sectors. Therefore, I developed a manual mapping to group the Trucost company emissions by IPCC sector classification.
²As illustrative example: Friesland-Campina is one the AFOLU companies in the small orange bar for Trucost.