Written by Magali Delmas, Dror Etzion and Nicholas Nairn-Birch
Originally posted by Corporate knights
The ascendance of socially responsible investing as a viable financial strategy has been accompanied by a similar proliferation of ratings schemes for assessing corporate social responsibility. Simply, if investors want to allocate their funds with SRI principles in mind, they need data that will allow them to make informed decisions. Indeed, over 50 rating methodologies for assessing environmental and social performance have been developed, more than a third of them since 2005. At some point, this may have become too much of a good thing.
“We’re awash in metrics,” said Allen White, co-founder of the Global Reporting Initiative at a recent conference.
Indeed, as more ratings systems emerge, it becomes less likely for each new methodology to provide unique or supplemental information. True, different ratings use different criteria, different methodologies, different weighting schemes and, perhaps most importantly, data from different sources. More information can yield more richness and greater analytical precision, but too much information can also be distracting and confusing.
So, does more information add more value? We decided to examine the extent to which dimensions of CSR behaviour used by various rating organizations are in fact complementary or redundant. Rather than examining a specific rating scheme in isolation, we selected ratings data drawn from multiple sources – in essence, mimicking the approach of a savvy investor with a range of information sources at her disposal.
Specifically, we looked at the ratings produced by three organizations: KLD, Trucost and Sustainable Asset Management (SAM). All three are highly visible not only to investment managers and executives but also to general audiences more broadly. For example, the Newsweek Green Rankings lean on data from KLD and Trucost, while the Dow Jones Sustainability Indexes draw on data from SAM.
More importantly for our purposes, each is designed differently. KLD ratings are based on publicly available information from media providers and do not rely on data provided by the companies themselves. SAM’s methodology relies primarily upon company responses to sustainability questionnaires. Trucost, meanwhile, imputes a score for each firm based on government census and survey data, industry data and statistics, and national economic accounts, and then quantifies the environmental impacts and damage costs associated with these extractions.
In our study, we used a statistical technique called principal components analysis. It is commonly used in the social sciences to distill a large set of variables to a small set that retains most of the information. These small sets of variables, called principal components, are uncorrelated to each other, meaning that each captures a distinct aspect of the original data. Our analysis revealed that integration of the KLD, SAM and Trucost rating schemes generates two distinct components of CSR performance:
• The process dimension captures the management practices and systems that firms put into place to improve their CSR performance.
• The outcome dimension captures the impact of these practices and systems on people and the planet
The fact that there are two distinct and independent dimensions implies that processes and outcomes are much less linked than we would perhaps expect. Companies may excel at reporting, governance and the utilization of environmental management systems, yet still emit substantial amounts of pollution. Or, more cynically, they may put in place such processes “just for show” rather than actively pursuing meaningful outcomes. Moreover, process measures can be easily communicated by companies (e.g., “all our facilities are ISO 14000 certified”), and perhaps more easily fed into rating schemes.
As a second step, we examined to what degree the two dimensions were associated with market value. We found, of the two, the process dimension is more aligned with market value. Perplexing at first glance, yes, but it actually makes a lot of sense. If process measures are more abundant and can more easily be fed into ratings methodologies, they will influence market valuation.
In a perfect world, these processes would translate into expected outcomes. However, as our analysis demonstrates, processes and outcomes are distinct dimensions. Consequently, market performance corresponds to processes, but not outcomes. Put differently, markets can only respond to information available to them. So, if process measures are more easily integrated into rating schemes, then that is the information made available for investors.
Yet it is outcomes – with their tangible and material impacts on people and the planet – that ultimately matter. Our market systems, it seems, are not yet capable of embracing this fact.
The bottom line is that failure to improve the meaningfulness of CSR ratings may jeopardize the confidence that investors place in socially responsible investing, potentially undermining its credibility and raison d'être.
The conclusion is somewhat grim, but there is hope. As sophistication, transparency and market interest in CSR increase, investors are likely to become increasingly judicious in deciding which actions taken by a firm are truly oriented to real impact.