Blazing a trail: ESG integration and big data
Underpinning all AXA IM Rosenberg Equities investment strategies is a set of beliefs that we think contribute to a better investment outcome for our clients:
…future earnings and fundamentals drive equity returns
…factors link to fundamentals
…ESG information can lead to fundamental insights
…a relentless focus on the details improves investment outcomes
…innovative use of technology benefits clients
Few asset managers put environmental, social and governance (ESG) beliefs at the heart of their investment philosophy, but our commitment to ESG is such that we have included ESG considerations across all of our strategies, some US$22bn of assets, and intend to provide reporting on ESG KPIs1 for all our portfolios. We believe this is a clear commitment to a more sustainable form of equity investing and puts us in the industry vanguard.
The road to integration
AXA IM Rosenberg Equities has a long experience of managing socially responsible portfolios for clients and has been managing responsible investment (RI) strategies since the 1990s. Initially, the process consisted of responding to clients’ needs through divestment from companies involved in controversial activities or norms-based negative screenings.
Over the past years, responsible investment has garnered increasing attention from investors around the globe recognising that company ESG features can impact portfolio performance not only in terms of risks but also have positive long term implications for returns. Indeed, our own research on links between solid governance and company’s return on equity (ROE) appear to support a strong economic argument for a broader integration of ESG rather than applying an exclusion-based policy.
As a result, in 2014 Rosenberg Equities actively integrated ESG into a portfolio managed under our Sustainable Equity strategy, choosing to consider companies’ holistic ESG attributes rather than simply applying a divestment policy.
AXA IM’s well-established RI data framework allows ESG criteria to be incorporated into our quantitative investing models relatively easily. ESG data simply becomes another piece of information to be incorporated into the investment process and taken into consideration when portfolios are constructed. Optimisation techniques allow a portfolio to be constructed that retains the risk/return attributes we seek while improving the overall ESG profile. Our approach also allows for specific impact targets to be achieved, for example an explicit level of carbon exposure or governance considerations such as the percentage of woman on company boards, but we take account of ESG considerations in our strategies in multiple ways:
- We believe the management of water and energy resources will profoundly affect our economic and environmental future. As result, we seek to tilt our portfolios to have a lower carbon footprint and water intensity compared to the benchmark index.
- We believe that companies that have a strong corporate governance deliver stronger earnings quality. As such we aim to achieve an improved measure of governance in our portfolios compared to the benchmark index.
- We expect our portfolio to display an improved overall ESG profile compared to the benchmark. Our model will naturally prefer companies with higher ESG scores.
We generally prefer to consider ESG actively by favouring companies with higher ESG scores, all else equal, and tilting away from those for which the opposite is true. However, where applicable we will wholly divest from a number of specific sectors:
- Tobacco companies: Companies included under tobacco in the Global Industry Classification Standard (GICS) are divested as we believe regulatory and profitability risk to be under-represented
- Controversial weapons: Companies involved in anti-personnel landmines and cluster bombs. Also including chemical and biochemical arms and nuclear weapons
- Coal: Exclude companies that generate more than 50% of their revenues from the coal industry
- Palm oil: Exclude companies which do not commit to responsible and sustainable palm oil production (RSPO) standards
Finally, we actively screen for the most severe controversy stocks, which we believe present a risk to future operations due to poor business practices. We begin with the ‘Category 5’ controversies list by Sustainalytics, a leading provider of ESG and corporate governance research and ratings, but also consider other stocks that may have been flagged by our investment staff for targeted review. Stocks that fall into this bucket are reviewed by our internal controversy committee and, unless there is compelling evidence that their controversial issues will be addressed in the near future, and/or unless the company is currently a target of our overall corporate engagement initiatives, the companies are removed from portfolios.
The above outlines how we have captured ESG information in our investment process. However, ESG data itself will continue to change and evolve, so we are exploring alternative data sets to classify the commitment of companies to ESG.
Our testing of alternative data included analysing ESG commitment using financial reports. We looked at the ESG vocabulary used in companies’ annual reports using a Natural Language Processing (NLP) technique. The results showed the median and lower/upper quartile of word counts for each sector in the S&P 500 Index. From this data we found the following:
- Almost all industries have seen a material increase in ESG discussions
- Energy and basic industries’ ESG mentions remain (perhaps surprisingly) low
Our NLP modelling is in relatively early stages and further work needs to be done before implementation into the investment process, but together with our traditional analysis, we have found ESG information is:
- Economic in nature
- Not well represented in traditional investment information (e.g. financial statements)
- Not typically required by regulators or standardised
- Slow moving… for the most part
- Long-horizon… for the most part
We believe that ESG information is complementary to traditional fundamental information. It is potentially material, meaning that it may change our opinion of the fundamental ‘worth’ of a stock. Taking ESG factors into account is nothing more than considering extra-financial information while evaluating the attractiveness of an asset. Using a quantitative investment process means we can do this consistently and methodically across a wide range of assets.
1KPI – key performance indicator