- The guiding principle for tracking multi-asset-class (MAC) portfolios’ carbon footprints is that each dollar invested in either equity or debt of an issuer finances the same amount of emissions.
- Tracking a MAC portfolio’s carbon footprint, and disentangling what drives its changes, can be done at a total portfolio level or at the sub-portfolio level when decarbonization is monitored and evaluated at the sub-portfolio level.
- At aggregate level, one can see the “net” view in terms of both additions and deletions of issuers from the total portfolio. At a sub-portfolio level, one can assess the “gross” view – additions and deletions and financing effects.
In our recent paper on portfolio-level carbon footprinting, we discussed a general framework that allows investors to attribute changes in equity portfolios’ emissions to their primary drivers. This framework can also be applied to multi-asset-class (MAC) portfolios, since it is derived from the MAC carbon-accounting standard proposed by Partnership for Carbon Accounting Financials.
For the owner of a MAC portfolio, the guiding principle is that each dollar invested in either equity or debt of an issuer finances the same amount of emissions. Therefore, we can aggregate all types of exposures for the same issuer and perform attribution at the issuer level. We have applied and demonstrated this approach in our blog post focused on tracking corporate-bond portfolios’ emissions over time.
In this post, we identify and analyze two approaches to tracking a MAC portfolio’s carbon footprint. Each approach serves different objectives: assessing changes in carbon footprint for the aggregate portfolio and for individual sub-portfolios.
Emissions attribution for the portfolio
Tracking changes in the aggregate portfolio’s carbon footprint over time provides insight into the net changes in the portfolio composition and their impact on an investor’s carbon footprint. Here we illustrate it with an example of a hypothetical MAC portfolio from Jan. 31, 2020, to Jan. 31, 2023.1 The emission data is aggregated at the issuer level.
Untangling changes in financed emissions of a MAC portfolio
The aggregate portfolio’s financed emissions declined by 38.79 megatons (Mt) overall during the three-year period.2 Looking further into this, we can note the following:
- New issuers added to the portfolio from rebalancing across equities and bonds increased its financed emissions by 28.26 Mt, while issuers divested from the portfolio lowered them by 31.90 Mt.
- Issuers consistently held in the portfolio during the entire study period decreased the portfolio’s financed emissions by 70.11 Mt.
- The increase in data coverage increased the portfolio’s financed emissions, as we were able to track emissions of more issuers.3
- The actual change of the company-level emissions of the issuers in the portfolio reduced the portfolio’s emissions by 113.15 Mt.
- Changes in the attribution factor in consistently held issuers contributed 48.86 Mt of the emission increase.4
- Changes in financing structure reduced the portfolio’s emissions by 69.59 Mt.5
- The net financing share increase contributed 507.75 Mt of the emission increase.6
- The interaction of financing structure and financing share contributed 389.29 Mt of the emission decline.7
In short, the decline of aggregate portfolio carbon footprint was mainly due to organic decarbonization at the issuer level.
Emissions attribution at the sub-portfolio level
On top of understanding the net changes in the hypothetical aggregate portfolio, it may be beneficial to track emission changes in various hypothetical sub-portfolios — e.g., by portfolio manager, asset class, region and sector. Showing attribution at the sub-portfolio level can help evaluate, compare and contrast decarbonization progress across different dimensions to inform investor actions. For example, for sub-portfolios managed by different portfolio managers, it may be necessary to look at their carbon-footprint attribution to assess each manager’s progress and disentangle the market effect from the manager’s decisions.
Below we show the attribution trees for the equity and bond sub-portfolios within the aggregate portfolio.
Attribution for the hypothetical equity portfolio
The equity sub-portfolio’s carbon footprint declined over the period. Most of the decline came from the consistently held positions, where organic decarbonization happened at the issuer level.
Attribution for the hypothetical bond portfolio
The bond sub-portfolio’s carbon footprint increased less than the equity sub-portfolio’s carbon footprint decreased. The increase of the bond sub-portfolio’s carbon footprint was mostly due to the addition of issuers to the portfolio.8 The issuers consistently held in the portfolio also contributed to a small increase in carbon footprint, mostly driven by an increase of financing share that was more pronounced than the actual decarbonization at the issuer level.
Comparison of aggregate versus sub-portfolio-level attribution
Note that the sum of similar nodes in the sub-portfolio’s carbon-footprint attribution trees does not always add up to that of the aggregate portfolio due to the attribution-tree structure, where new positions, divested positions and existing positions are reported separately.
For example, an equity manager could divest from the equity position in a company, while the fixed-income manager may decide to invest in different bonds from the same issuer. Therefore, while the additions and divestments in the equity portfolio would lead to a change in the "Divested issuers" and "New issuers" nodes, that is not necessarily the case for the bond portfolio and the aggregate portfolio. So, the sub-portfolio attributions would not sum up to the total portfolio’s attribution starting at the first layer of the attribution tree, except for the change in data coverage, as illustrated in the table below.
Summing up and breaking down
Attribution at the level of an aggregate MAC portfolio can provide insights into the net changes of an investor’s carbon footprint. Attribution at a sub-portfolio level can help assess and compare performance in each sub-portfolio individually, and potentially inform actions.
1The hypothetical portfolio is a blend of equities represented by the MSCI USA Investable Market Index and corporate bonds by the MSCI USD Investment Grade Corporate Bond Index. As of Jan. 31, 2020, the equity sub-portfolio accounted for about 90% of the total MAC portfolio‘s market value (USD 45 trillion).
2Financed emissions were calculated by multiplying an attribution factor by the emission of the issuer. In this blog post, the unit for financed emissions is CO2 megaton.
3At the end of the tracking period on Jan. 31, 2023, emission data became available for four issuers that were not covered at the beginning of Jan. 31, 2020; and emission data is missing for one issuer that had emission data on Jan. 31, 2020, but not anymore on Jan. 31, 2023. We assume zero emission for missing data in this research.
4The attribution factor is calculated as the ratio of the position-outstanding amount over total enterprise value including cash (EVIC). An investor who owns a USD 1 million position in a company valued at USD 1 billion that emits 20 megatons of CO2 a year finances 20 kilotons of CO2.
5The financing structure is a company-specific variable defined as the ratio of a company’s total equity outstanding over the EVIC. For example, an equity portfolio’s financed emissions will decrease when the overall capital market is shifting toward more with debt financing and less with equity financing when nothing else has changed.
6Financing share is defined as the shares of equities the portfolio holds relative to the total shares outstanding for the issuer, or book value of the bond to the book value of the issuer’s total debt outstanding.
7The interaction term shows the nonlinear effect where both financing share and financing structure change at the same time.
8 The “New issuers” capture the financed emissions from the new issuers added to the portfolio, which are affected by the number of issuers added, issuer-level emissions and attribution factors. In this example, the new issuers account for only a small percentage of the portfolio by par value. However, the average emission and attribution factor for the new issuers are much higher than those of the consistently held issuers. Thus, the financed-emission contribution from the new issuers is much higher than the contribution from the consistently held issuers.
Further Reading
A Framework for Attributing Changes in Portfolio Carbon Footprint