Kempen Real Estate Update: Turning up the volume, cutting through the noise
The ESG data revolution
To illustrate the growth in data – one ESG provider now offers more than 220 ESG indicators for around 11,000 companies – which equates to over 2.4 million data points in totality. This trend for more and more data seems set to continue, with another vendor expanding their data offering from approximately 1000 data point in 2020 to around 6000 indicators per fund in 2021.
However, the expansion of the ESG research market comes at a price - there are a myriad of sources, model assumptions and interpretations of the data available, which leads to an overall lack of quality and consistency in the ESG data market at large.
As a leading responsible investor, Kempen is keenly aware of the ESG data challenges. We neither doubt the difficulties faced by ESG data vendors nor disparage their work when it comes to assessing the credentials of a heterogenous set of securities with limited disclosure – it is a daunting task.
We set a high bar as to any data that enters our investment process – alpha by control requires absolute faith in the data that supports our investment decisions. That is why we have chosen to take you on a whistle-stop tour of the ESG data landscape in this Alpha REIT.
Assessing the quality of ESG data
There are many interesting aspects to ESG data quality, and a white paper to be published later this summer will go into much more depth than we can in this missive. For now, take a look at some of the most important signals that the emerging ESG data sets must be handled with care:
Consistency. A simple test of consistency across data providers is to examine the overall ESG score given to companies by each vendor. In an ideal world these would line up fairly consistently, but at present we see a wide discrepancy in the scores from one data provider to the next. We remain hopeful, that as with credit ratings today, these discrepancies will fade as the market standardises it’s methodology over the coming years.
Transparency bias. Our research leads us to conclude that the more disclosure a company offers, all else being equal, the higher their ESG scores will be. This leads to interesting and somewhat counter-intuitive results, such as British American Tobacco being ranked in the top three environmentally and socially responsible companies in the FTSE 100.1
Market cap bias. As a result of transparency bias, we note that larger market cap companies have a tendency to report more comprehensively than smaller, less well-resourced small and mid-sized companies. Our experience has shown that many smaller companies take their ESG credentials very seriously, and sometimes this appears to be missed by the data vendors.
Modelling assumptions. Models and modelling assumptions can lead to some very different outcomes when assessing real assets specific ESG data points. For example, carbon intensity for an entire real estate portfolio will regularly have to be extrapolated from data for only a small subset of that portfolio. The modelling assumptions used by different data vendors can lead to starkly different estimates for the overall portfolio. Case in point: when we compared REIT portfolio carbon emissions that had been estimated, rather than reported, we saw an average difference of 136% from one vendor to the next.
Omission. Finally, another challenge is simply that not all companies submit data to ESG data vendors.
The Kempen Real Assets Approach
Now, to the nitty gritty of our own ESG data analysis. As outlined above, a key issue we encounter is coverage – not all companies submit data to ESG data vendors. The reasons are varied, but it is worth reading the note published by Vonovia this year explaining their exclusion from the GRESB survey to understand the reasons are not always caused by a lack of resource or ill-will towards responsible investing. In those circumstances, our analysts will take a view on the company, having spoken to management and considering all the relevant disclosure, and incorporate this evaluation into our framework.
Another issue we encounter is poor quality data. No data point that enters our investment framework isn’t subject to intense scrutiny, and ESG data are no different to any other part of the machine. For example, we have seen carbon intensity data for two companies with very similar asset locations and ages, yet the modelled carbon intensity estimates are substantially different. In that event, we dig deeper to understand how this is anomaly possible, and where necessary, amend the data with a manual override.
The third hurdle is the context of a score. If a smaller market cap investment is finding it difficult to resource the necessary public disclosures or submission direct to a data vendor, but we know that the management are committed to ESG and indeed their salaries are aligned to it – we consider overriding an external score if we do not believe it reflects the facts on the ground.
All of the adjustments we make to our ESG data points are logged in our data infrastructure and we can then trace back to the exact justification for the manual override of the original data. This is key component in the transparent nature of our investment process.
The ESG outlook
We believe there are four major themes we will see develop in the arena of ESG data over the coming years, and as asset managers we will seek to support the movement in this direction where we can.
Standardisation. We envisage a future where data vendor align their requirements for survey data from companies in a standardized manner. This will allow disclosure to be a one-off event for each company on an annual basis and the efficiency gains of only engaging in one framework will be more attractive and accessible to companies across the globe.
Demand from Real Active managers. Investors, such as ourselves, will engage with the management teams of companies to press for greater disclosure and transparency about their operations and how they align with the ESG agenda.
Regulation. As the theme of ESG continues to dominate conversation amongst political leaders and financial regulators, we would not be surprised if regulation forces greater levels of disclosure on the industry.
External data validation. Technology marches ahead and it will not be long before a low orbit thermal satellite data will be able to assess the thermal efficiency of a real estate asset during cold weather. No certificates, models or assumptions need be applied by an ESG data vendor – the building is emitting the data and the capability to collect this information is being developed as we speak.
Because we set such a high bar for the data that enters our investment process, at this moment it means the majority of ESG data sets in their rawest, unclean form cannot and should not be integrated into our investment decision-making.
Conclusion
We are passionate about the importance of ESG and believe that harmonised data must be a key component in the journey towards our shared sustainable future. As part of that journey we are happy to report on ESG data sets that are available and by doing so will continue to champion the relevance of what is ultimately one of the most important issues of our times. However, for now, we remain selective about the ESG data points that underpin our investment decisions.