Financed emissions are an important gauge for financial institutions to measure and understand their exposure to the GHG emissions footprint of the various activities they finance.
Spearheaded by initiatives such as the Partnership for Carbon Accounting Financials (PCAF), the Institutional Investors Group on Climate Change (IIGCC) and the Task Force on Climate-Related Financial Disclosure (TCFD), a lot of progress has been made in defining global standards on financed emissions.
But many challenges still remain. Institutions are working through data collection and implementation of these emerging standards.
Among the most pressing unresolved issues are:
Data ambiguity and inconsistency (multiple financial- and climate-related data sources to distill into actionable shapes and forms)
Data gaps in consistently measuring GHG emissions at the asset level, over time and across various asset classes
These problems are further exacerbated by:
A lack of dedicated software & data tools for portfolio management, risk, reporting and sustainability domain experts to effectively understand performance drivers in a financed emissions portfolio (or at the stand-alone company or group structure).
An absence of solutions to automate data collection, normalisation and computation at scale to provide transparent data, in the right format, where it is needed.
Without addressing these major challenges, appropriately assessing the performance of a financed emissions portfolio would seem impossible.
In the following paragraphs, we aim to walk the reader through the financed emissions framework and highlight some of the practical complexities of computation and output analysis.
Assessing Financed Emissions
Financed emissions represent an investor’s share in an asset’s direct or indirect GHG emissions (Scope 1, 2 and 3). (1)
For the purposes of this article, we focus on corporate investment exposure, but our conclusions can be applied to all other asset classes. (2)
GHG emissions are self-reported by a company or can be estimated from proxy observations on a company’s activity. Emissions attributable to an investor can be estimated by dividing the value of their investment by the company’s expected enterprise value.
With these data points – the identifier of the corporate investment security held, the GHG emissions reported by the issuer or borrower and an attribution factor, we can calculate a company’s financed emissions or the financed emissions for a portfolio of companies.
Investors typically define their financed emissions strategy by setting a reduction target over a given time horizon for a portfolio. The targets vary by underlying financed emissions KPIs (e.g. absolute emissions versus emissions intensity), scope (1, 2 or 3), asset coverage (high- versus low-impact sectors) and asset class (corporates, commercial or residential real estate, etc.)
In order to assess a portfolio’s performance against a set target, we need to understand what the underlying drivers are for financed emissions performance. Financed emissions in a portfolio can change due to:
Change in invested amount
Change in a company’s enterprise value
Change in company GHG emissions
Making a new investment in a company
Reducing a portfolio position or divesting an exposure completely
Additionally, we know from our work on emission modelling that GHG emissions on a company level correlate with business performance, and can be very volatile. (3)
It is imperative that investors are able to understand the individual contribution of each single factor to overall performance, which requires zooming into individual assets as well as aggregating the relevant KPIs on business segments of interest, like the broader sector or activity level.
Despite all the progress and efforts, data gaps remain the most pressing issue. For simplicity we distinguish between three types of data gaps:
Absence of self-reported GHG emissions data
Time lag between carbon accounting analysis and the date at which the emissions data is publicly reported
Lack of corporate Scope 3 emissions data across different categories or altogether (which we will address in a subsequent paper)
An Introduction to Financed Emissions Analytics
To illustrate some of the challenges and demonstrate ways to tackle them, we have run our financed emissions APIs on a sample of 26 public market ESG ETFs to calculate Scope 1 + Scope 2 financed emissions.
These funds represent USD 26 billion in assets under management, with 6,858 unique holdings. We looked at quarterly and semi-annual portfolio cut-offs for the period March 2019 to December 2021. (4)
Data gaps for emissions were filled using our proprietary Scope 1 and 2 emissions estimate models.
Plotting the sum of Scope 1 + 2 financed emissions against the cut-off dates provides an insight into the performance of each fund over time. Given each fund has AUM of USD 1 billion, the numbers are comparable.
Figure 1: Scope 1 + 2 financed emissions by fund over time
We found that in almost all funds assessed, the top 10 exposures contribute disproportionately to the portfolio’s financed emissions (when measured against their total weight in terms of the portfolio’s AUM).
To prove this point, we calculated the contribution of each position in a portfolio as a proportion of the total financed emissions for each cut-off date, and aggregated these financed emission weights for the top 10 contributing companies in each portfolio against their contribution to the portfolio AUM.
Figure 2: ESG ETFs’ Scope 1 + 2 contribution relative to AUM contribution for their top 10 holdings
Figure 3 zooms into one portfolio - iShares MSCI World ESG Enhanced UCITS ETF - to see the evolution of the top 10 financed emission exposures over time. Throughout the coverage period, the top exposure keeps changing. In December 2021, Total SE appears in the top 10 exposures, while Exxon Mobil drops out for the first time.
Figure 3: Composition of top 10 exposure for iShares MSCI World ESG Enhanced
 This is in fact the lower bound of almost all funds. In the majority of cases the top 10 holdings contribute more than 50% of financed emissions.
To understand whether underlying emissions performance or portfolio weighting drives financed emissions for this sub-group, we compared how total Scope 1 + 2 emissions changed against the total weight, taking March 2020 as the base year.
Figure 4: Top 10 exposure change in underlying company emissions (Scope 1 + 2) vs change in weight
The diagram suggests that changes in financed emissions over the past two quarters were principally driven by re-weighting in the ETF positions towards companies in the top emission exposure segment, rather than change at the actual underlying companies and overall portfolio emissions performance.
Analysing the performance of financed emission portfolios becomes complex very quickly – and is compounded by the number of portfolios, the heterogeneity of underlying assets and industries, and time-series growth.
Our sample of 26 ETFs with 6,858 exposures stresses the need for broad perspectives and ways to evaluate performance, both for individual holdings and at aggregate levels, such as the portfolio as a whole or even the total asset management operation or corporate lending book.
Emissions tend to be highly concentrated in specific names and industries in a portfolio, and the same is often true for the business activities within a company. Consequently, financed emissions can change dramatically over a given period – affected by specific company-level events, without resulting in a change to the general portfolio trend.
We believe that in order to understand these dynamics, and to provide timely insights – including accurate figures for use by a company’s risk function and managers – a number of critical steps are required:
Financed emissions measurement must be seamlessly integrated into risk management and reporting workflows;
Systematic, repeatable processes are required – enabled by technology stacks and data architecture that can support analytics and reporting, providing ways to customise and iterate computations and segment relevant data chunks;
Blockages to collecting and sharing data across an institution must be addressed.
Our readers should also note that our snapshot into financed emission analytics focuses only on understanding some aspects of historic performance.
Future scenario analysis, with multi-layered pathway and energy transition dynamics, can only be successfully built when based on the robust and scalable solutions we have presented here.
Watch Arcturus.io for more to come.
1. This is merely a technical point, but emissions are always estimated, either through frameworks such as the GHG Protocol and respective emission factor tables from accepted agencies and frameworks (e.g. IPCC, EPA, EEA), or through bespoke estimate models. There are numbers of degrees of freedom between the reliability and accuracy of these models and their outcomes, but it should be assumed that companies' self-reported data is currently the most accurate way to estimate emissions.
2. Other asset classes are defined in the context of PCAF Global GHG Accounting and Reporting Standards for Financial Institutions and include residential and commercial real estate, sovereign debt, auto loans and project finance.
3. From our analysis of self-reported company emissions – covering a time period from 2015 until the present momvent – sizeable percentage changes in emissions are not unusual. That's because companies can grow or shrink through M&A and disinvestment, and is also due to the volatility of business activities.
4. For the calculation of the attribution factors, we used S&P financial data. For the emissions, we used Arcturus estimates.
5. This is in fact the lower bound of almost all funds. In the majority of cases, the top 10 holdings contribute more than 50% of financed emissions.