How much of the climate problem does the investment industry own?

In this note I will document how I came to the conclusion that the investment industry ‘owns’ approximately 25% of annual global greenhouse gas emissions. This feels like a necessary first step towards addressing the climate problem. However, it is not a given that ownership of a problem naturally leads to the owner solving the problem. We will briefly refer to this in closing.

As a passing comment, I was surprised at how hard it was to make sense of an apparent abundance of data on emissions. Some datasets favour carbon dioxide (CO2) emissions, others greenhouse gas (ghg) emissions; each dataset appears to define sectors differently. As a consequence I developed a rule of thumb (CO2 is 70-75% of total ghg) and developed my own, high-level, sectors to suit the current thought experiment.

Finding total emissions data on the internet is straightforward, but the data comes with a lag. It takes a little more reading around to find estimates of current emissions rates. For our purposes, we do not need a high level of accuracy so we will state that current CO2 emissions are slightly below 40bn tonnes per annum[1], and ghg emissions around 52bn tonnes[2][3].

Finding the investment industry’s ‘ownership’ of emissions requires a bit more work, and some assumptions – which can clearly be challenged. Arguably, a more accurate way to quantify the investment industry’s emissions would be bottom-up – aggregating all the emissions from all the assets owned. However there are substantial data challenges with this approach, including the problem of cross ownership of shares. I am not looking for high accuracy at this stage, and am willing to tolerate some rough justice implied by simplifying assumptions. Consequently, I assume:

  1. The investment industry owns the entirety of all listed companies (they actually own a large subset, so at this stage we are overcounting)
  2. Corporate bonds are only issued by listed companies (this assumption allows us to ignore corporate bonds; we do not know in which direction the inaccuracy of this assumption would affect the results)
  3. Lending money to sovereigns (buying their bonds) does not make the investment industry responsible for public sector emissions
  4. Allocations to real estate and private equity are relatively small and therefore the emissions can be ‘covered’ by the overcounting within assumption #1.

In short, this list of assumptions allows us to proxy the investment industry’s emissions by simply considering a global equity index, for which aggregate data exists. If we consider the MSCI All Country World Index, then current ghg emissions (scope 1 and 2) are currently 6.4bn tonnes. I make one final, heroic, assumption that the scope 3 emissions (largely attributable to the use of sold products) are the same size as the scope 1 and 2 emissions. From informal conversations with industry peers it appears that the range of estimates for the size of scope 3 emissions is wide, from the lowest being around 50% of scope 1 and 2 to the highest being in excess of 100%. I have assumed something at the upper end of the range. It follows that public investor-owned companies produce around 12.8bn tonnes of annual ghg emissions and are therefore responsible for 25% of all emissions (12.8/ 52 = 24.6%).

Relying on assumptions is more comfortable the more confident we can be that they are reasonable. To this end, I looked for evidence to corroborate this result – and I (re)discovered a CDP report from 2017[4]. Using data for 2015, CDP attributed 30.6bn tonnes of ghg emissions to 224 fossil fuel extraction companies (“operational and product GHG emissions” from page 10 of their report). This is approximately 60% of total emissions (30/50). In essence they have attributed back emissions from all other sectors (ie scope 3 activity). This is very pragmatic in terms of simplifying the number of companies to engage with, but is it reasonable?

We will check the reasonableness two ways. First, we will see if we can get close to the 25% number derived from the MSCI ACW Index, and second, we will see if we can satisfactorily explain the ‘missing’ 40% of emissions.

Helpfully, CDP provide further information. They state that of the 30.6bn tonnes, 30% come from public investor-owned companies, 11% from private investor-owned companies and 59% from state-owned companies. [Note, this is 2015 data so pre-Saudi Aramco’s IPO which would shift the proportions slightly]. For now we will assume all of the 11% private sector is attributable to institutional investors, but this is likely an overstatement. It follows that the investment industry is “directly” responsible for about 25% of annual emissions [(30% + 11%) of 60%]. Tada, as some might be tempted to say.

For the second test I must continue with a relaxed attitude to accuracy as I will need to cast around and combine other data sources. The results of my research are shown in the table below.

For my part, I am satisfied that the missing 40% of emissions is sufficiently explainable that I think the CDP approach is very reasonable. It also throws a new light on the problem, which then suggests new solutions. I develop this in a separate thought piece[5].

An aside on accuracy

I have mentioned a couple of times already that I am not aiming for accuracy. This is largely because of the uncertainty within the data we are dealing with. To illustrate, consider the wildfires row in the table above. The linked article suggests the average annual release of CO2 was 8bn tonnes. However, other effects including regrowth offset some of this, leading to their estimate of 5-10%. My note in the table is intended to convey that the emissions for 2018 would be higher, and for 2019 and 2020 higher still. To this point, Prof Ian Goldin suggests that wildfires in 2018 released 32bn tonnes of CO2[6]. This is materially different(!). So much so, that I come back to it in the postscript. For now I note that different sources of emissions data are not easy to reconcile and so I have aimed for a good-enough level of accuracy.

Does owning a problem lead to solving the problem?

My purpose in this note was to attribute a proportion of the climate problem to the investment industry. I am satisfied that 25% is a reasonable attribution. What the investment industry does with this conclusion is far from certain. A number of considerations apply here:

  • What capacity do industry organisations have to contribute to a solution? (ability)
  • What should be the extent of the contribution – minimum, fair share, generous? (extent)
  • Do industry organisations have a moral incentive to contribute? (intrinsic motivation)
  • Is the solution likely to be profitable, reducing fiduciary duty concerns? (extrinsic motivation)

The question I would like to pick up here is the extent of contribution. Is the minimum contribution to do nothing, and leave the problem for governments and investee companies to sort? Is the investment industry’s fair share solving 25% of the problem? Or, given that wildfires and melting permafrost are not going to amend their ways and provide their fair share of the solution, is it a higher number? And is being generous even possible when bound by the requirements of fiduciary duty? All of these questions imply autonomy, but that is not a given. The inevitable policy response could introduce compulsion, and if that is combined with cynicism regarding the realism of required actions we could find ourselves in a pretty toxic industry culture. Better, in my opinion, to get out ahead and start on some meaningful actions while they remain voluntary. The Institute and its working groups will continue to grapple with difficult issues such as these.

Postscript | is it already too late?

I was shocked to come across Prof Goldin’s estimate for wildfire emissions quoted above. My understanding of the emissions data, such as it is, allows me no other conclusion that the ghg annual emissions data of 52bn tonnes we have being using are wrong – by about 50%. If wildfires are now producing 32bn tonnes, or more, each year rather than the historical level of 8bn, then we need to add 24bn tonnes to our annual figures. We should probably brace for some bad news from the IPCC later this year. We might be told that the remaining carbon budget associated with 1.5C of warming is now lower, by a substantial margin.


[1] ourworldindata.org states that 2018 CO2 emissions were 36.58bn tonnes (link here)

[2] ourworldindata.org carries an article quoting 52.3 bn tonnes, which we think relates to 2017 as the source quoted is a 2018 paper. The expectation for 2020 from exponentialroadmap.org is 54.2bn tonnes (version 1.5 updated March 2020. link here)

[3] We follow standard convention for these measurements in that ghg emissions are measured in tonnes of ‘CO2e’ which is shorthand for carbon dioxide equivalent

[4] CDP Carbon Majors Report 2017

[5] See our note The answer is more primary investment; a lot more Thinking Ahead Institute 2020

[6] Terra Incognita: 100 Maps to Survive the Next 100 Years, Ian Goldin and Robert Muggah, Penguin, 2020