Hello everyone, and welcome to today's webinar. My name is Lauren Smart. I'm the Chief Commercial Officer for S&P Global Sustainable1, and it is my pleasure to be moderating today's webinar, The Future of ESG Intelligence: Tackling Data Availability Gaps. Today is a particularly special day because it's Earth Day, and this also marks the one-year anniversary since the launch of Sustainable1, which is S&P Global's single source of essential sustainability intelligence. Thank you very much for joining us on this special day too. Before I introduce my stellar guests, I've got a few housekeeping items. Just bear with me for the housekeeping, and then I will be introing the guests and what we'll be talking about today. We recognize that the topic of today's webinar is of great interest to you.
We want this to be an interactive session and encourage you to submit questions for discussion. At the bottom of your screen, you'll see that there's a row of widget icons. These icons will allow you to interact with us throughout the session. I'd like to point out the Q&A widget, which can be used to submit questions to us and to the panelists, as well as the survey widget. Please take the time to fill out our short survey after the webinar. We really value your insights. Additionally, you can find any of the reports discussed today in the resource widget. The webinar is being recorded, and an on-demand version will be available shortly after we conclude. If you encounter any technical issues during the program, please try refreshing your browser.
If the issues persist, please use the Q&A widget to contact us, and a member of our technical team will assist you. Now we've got the housekeeping bit out of the way. It is my distinct pleasure to introduce my esteemed panel. I am honored to be joined today by Manjit Jus, Managing Director and Global Head of ESG Research for S&P Global. Mona Naqvi, Global Head of ESG Capital Markets for S&P Global Sustainable1 as well. Thank you very much, Mona and Manjit. Thank you for joining us today. Today we're going to be talking about the future of ESG intelligence and in particular, how to address the market pain points of ESG data availability. We're gonna share S&P Global's perspectives on these issues, and we also really look forward to your questions.
Do make use of that Q&A widget. As we all know, the ESG data market has been booming. There are now more users of ESG data than ever before, and that means that there are more use cases too, which is fantastic. As the financial community is seeking to integrate ESG into more different workflows and into different use cases, it creates some new needs that the ESG datasets that have previously been utilized need to expand and evolve to manage. In particular, we have a challenge around data availability. What do I mean by that? As many of us know, the ESG data world really started within the listed equity space, and that's where the biggest drivers were for disclosure from companies.
We have the best data, the most complete datasets, typically in large-cap listed companies. Yet, if we want to integrate ESG into every single financial decision, and we want to make sure that sustainability is central in every decision-making within capital markets, we need to consider not just large-cap listed equities. We need to consider the smaller companies. We need to consider non-listed companies. We need to consider entities that may not be companies at all. We need to be thinking about infrastructure, private equity, real estate, fixed income. Working with a bank, we need to be thinking about all the non-listed companies that might be within their loan books, as an example. We have a range of evolving use cases that require different types of datasets and some data gap-solving to address.
To expand on that a little bit further, I am going to hand over to Mona Naqvi to share a bit more about some of these challenges that we face in the market.
Thank you, Lauren, and good morning, good afternoon to all of our listeners. Thank you so much for joining. As Lauren said, we do have a rapidly expanding ESG industry, which is a very exciting thing. Of course, you know, there are always some growing pains when an industry is blowing up so quickly. We think that we've made some really interesting steps to kind of address some of these, which we're gonna talk through today. On the next slide, I just want to highlight, first of all, that there is no one size fits all approach when it comes to ESG. We must remember that like any other parameter of personal preference, ESG and sustainability considerations are but one parameter of choice in an investor's capital allocation decision model, right?
You have traditional considerations like liquidity requirements or risk appetite and time horizon. Sustainability is but just another parameter for an individual to align their investments with their own unique mix of preferences. Like any other preference in traditional finance, we wouldn't expect all of those to be the same. Across a variety of investors, you will have a big spectrum of risk appetites and time horizons, et c. Sustainability is no different. Of course, there are many different facets to sustainability. Over the years of our experience at S&P, we've kind of identified these three core buckets that we see as being foundational to this space. Oh, I'm sorry. Still on that other slide. Thank you.
We've identified these three core buckets, and although these are highly simplified schematics, just to kind of illustrate these cornerstones of the landscape, of course, any individual's preference or investment strategy could reliably span across all three and could combine them in different ways. The point here is just to illustrate that responsible investing, ESG integration, and impact investing are three distinct, in their purest forms, three distinct approaches to sustainable investing. I think a simple way to think about this is: What is the investment objective? Subject to what is a sort of second order constraint. Of course, with any investor, the goal is to ultimately make money. Sometimes there are secondary considerations or even primary ones that may come before that objective, optimize in a particular set of preferences.
With responsible investing, we all know this is the approach to sort of negatively screening or positively screening, in some cases, companies that may not align with your values. First and foremost, what we find in its purest sense, that investors in this bucket are looking to align with their values, subject to, of course, maximizing their return, but making sure that they can align with their moral opinions, that is of the utmost importance to them. As a result, the outcome or output of this for them tends to be a reputation hitting signal, right? They are still investing, but they are also telling the market that they don't agree with this particular set of business activities, and that, for them, is an important outcome of this type of investing.
We have ESG integration, and this is probably the bucket that has blown up the most in the last several years, in particular, as data has become more available. ESG integration, I think, in its purest form, is really no different from traditional investing. It is still, first and foremost, targeting the maximization of profit and returns, but it is subject to one's ability to uncover hidden risks and opportunities through the use of alternative data. Really all ESG integration is trying to tap into those hidden externalities, those currently unpriced assets and risks and opportunities lurking in the investment portfolio that may not be adequately addressed by traditional financial models. It's using alternative data, be it informed by ES G data points.
Ultimately, the goal here is to address what investors may perceive to be a systemic mispricing of assets. The outcome would be a corrective repricing that would be beneficial to the holders of the portfolio. Then we have last, but certainly not least, impact investing. Again, in its purest form, the goal of impact investing is to advance a set of real world outcomes or global goals. For example, we often see investors linking their investment results to the UN Sustainable Development Goals, for example, or aligning the economy with a net zero trajectory by 2050. These are strategies that care more about the outcome in terms of the real world impacts, although it is subject to, of course, still generating a stable financial return.
In this case, investors are very interested in understanding the real-world impacts and outcomes of their investments. This, for those impact purists among us who will know this requires that information to be measurable, intentional, additional, et c. Where am I going with all of this? As you can see, there is a really big spectrum of use cases for sustainable investing. In each of these, there are potential use cases for ESG data. It really just depends on how it gets applied. What we have is a situation where ESG data and ESG scores, in particular, as the sort of flagship data of this space, is simultaneously in some cases trying to answer many different questions.
Different data providers will have different approaches to how they answer these questions and serve these different use cases with that information. I just wanted to set the scene by highlighting it, that this discrepancy in use cases is sometimes contributing to confusion around ESG scores because how one is interpreting that data will depend very much on their vantage point depending on which of these core three pillars they are primarily focused on. On the next slide, kind of talking more a little bit about these challenges with ESG data. I think some of the primary ones to call out is that ESG scores sometimes diverge too much among data providers to be useful, or at least that is the traditional market criticism that we often hear.
It's important to kind of take a step back and understand why is that the case. We've seen already that there are different use cases out there that serve different types of investors, and of course, different data providers may be seeking to address each of those to varying degrees. Inherently in the methodologies that data providers are constructing, we are approaching these questions very differently. Another challenge is simply the input. The quality of inputs are not perfect, and we do have inconsistent data quality, as Lauren mentioned at the beginning. How different data providers deal with that missing information is also going to contribute to ESG score divergence. There are notable efforts being made on a regulatory and policy scale to help address some of that inconsistent disclosure from companies to have a more level set of inputs.
At the end of the day, ESG data and ESG scores in particular are a data provider's opinion on the relative mix of the most important issues, be it from a risk or an opportunity or even an impact perspective, depending on their own unique set of questions and philosophies that they're trying to address with their data. So ESG scores do diverge, but that actually may be intentional. These are opinions, expert opinions informed by the unique mix of preferences and questions that a unique data provider is trying to answer. Another challenge, and I guess this really does build upon the first, is that ESG data, at least in its headline score form, is relative, not absolute. By design, ESG data is assessing the risks and opportunities that companies face within their industries.
This is important because, of course, when you link ESG to the real economy and you're talking about big, unprecedented macroeconomic shifts unfolding, these are going to invariably affect different industries differently. There is a difference in the availability of technologies and efficiencies across the different industries we see. It's important to acknowledge those differences. A nice counterexample to help illustrate this is if we simply wanted to target absolute ESG score performance, meaning the absolute best irrespective of the industry, we might end up with a portfolio that's not very well diversified. An example of this could be media companies, for example, are typically less carbon intensive than an energy company. If you're going purely based on you know, who has the lowest carbon emissions, you'd end up solely investing in media companies as an example.
That may be good in terms of measuring the impact of an individual investor's strategy. You might be able to say you've reduced carbon emissions relative to the benchmark or the market as it is, but have you done very much to actually help change the behavior of those companies? Sometimes that relative aspect of ESG is important because it encourages companies within their industries to be assessed on the basis of available technologies and the business function and activities that they are actually pursuing. It encourages and incentivizes companies to do better within the world they operate within. If we solely took an absolute approach, then we may not actually help benefit companies and influence their behavior positively.
Whereas if we take a relative approach within industries, by design, we can help companies improve relative to their industry peers so that when we end up with a diversified investment portfolio that is truly reflective of the market as it is, we can actually help drive better performance standards. Last but not least, another concern sort of related to all of this is that ESG scores, at least in their headline form, are muted in the aggregate, or at least the underlying signals are. That, I think, is a really important thing to highlight, and it gets to the crux of the issue, which is that there is no single silver bullet that can single-handedly answer every question.
Naturally, an ESG score is the summation of many different questions based on the individual data provider's philosophy, and it is a useful entryway to understand at a high level what is the data provider's overall expert judgment on this. Of course, if you want to better understand the underlying signals, lots of data providers, including S&P Global, we make those underlying data, available. We actually provide up to 1,000 data points per company where available, so that you can back test and stress test the underlying performance signals associated with one, individual data point.
We should also remember that if you're an investor who has multiple sustainability objectives and you care, let's say, about climate justice, not just having a climate neutral investment portfolio, but you actually also care about the social implications of your investment choices, then you ultimately end up with an optimization problem. An optimization problem is, in an investment sense, sometimes difficult to explain because you're trying to aggregate numerous metrics all at the same time. We see in investment strategies, you do have many optimization models that are really good at perfecting the relative trade-offs between these different preferences, but at the expense of making it harder to explain. Whereas a headline ESG score, which may on the back end incorporate all of these different questions, at least makes it easier to explain headline changes in a portfolio.
For example, you could have a rules-based approach where you're screening out the bottom 25% of companies based on their ESG score ranking, and therefore you can explain decisions even if you're baking in multiple assumptions. Either way, there are trade-offs with whether you choose the individual data point or whether you go for the headline score in terms of their explanatory power for the changes driven by a strategy. That's a really important thing to remember. On the next slide, just kind of, to close out my comments here for today, I think underpinning those three bigger picture challenges are three very pointed and specific sources of potential confusion in the marketplace that have the potential to perpetuate one another in a bit of a cycle.
The first is that there are black box methodologies or at least differences in approaches from data providers when they create their ESG scores, and it isn't always perfectly clear what they are doing. Lauren has already talked about some of the inconsistencies in data availability. I've already talked about differences in use cases for ESG data. When you reconcile those two things, you end up with missing information and needing to triangulate it in some way. Which use case are you ultimately serving? What data is missing, and how do you fill those blanks? Different data providers have their own proprietary approaches to how they do that, which contributes to some of the black box methodologies because every data provider has their own secret source.
If there was perfect transparency in the market from companies and there was perfect clarity on what ESG scores were measuring, and that was universally defined for all providers, we may not have this as big of a problem, but it is, at the moment, a challenge that the industry must overcome. That is largely driven by the fact that there is, as we've talked about already, inconsistent data quality, coming from disclosures from companies being patchy. As we've talked about already, last but not least, there is a commingling of use cases. These three, can challenges or sources of market confusion somewhat perpetuate one another? If you think about it, having a black box methodology means that companies don't know how they're being ranked or scored, and so they're disclosing different metrics.
It also means that investors may be using that information differently because if you don't know how to interpret it, you will apply it in different ways. That inconsistent data quality on that second row is also contributing to this challenge because if you've got patchy missing information, obviously data providers will have to adopt different approaches to filling those blanks, which contributes to those black box methodologies. Similarly, if you've got inconsistent data quality, that also lends itself to different interpretations, again, contributing to that co-mingling of use cases. I mean, I won't go through everything, but as you can see, these have a reinforcing effect on one another.
This is why I think there is potential for a lot of confusion in the marketplace around ESG scores that really requires transparency, clarity, and quality to really overcome these challenges and help cut through and make sure that our clients know exactly what data is trying to measure, which questions it's answering, and our process and the robustness of our methodology for addressing those questions. With that, Lauren, I'll hand over back to you.
Fantastic. Thank you, Mona. Transparency, clarity, and quality. I think those are absolutely key for us, as we think about the future of ESG intelligence. I'm going to hand over to Manjit now to talk a little bit more and, about how we are incorporating those aspects into our approaches to filling data gaps. Manjit, over to you.
Thank you, Lauren. Maybe before I, you know, dive into the details, information gaps come in a wide range of shapes and sizes in the ESG space. That could be, you know, missing company disclosures in certain parts of the world, depending on the size of a company, depending on whether it's listed or not. There can be differences in expectations of what should be disclosed. You know, different stakeholders will have expectations of what should be disclosed in the public domain. You know, already there you will have data gaps because people have different expectations of what should be reported. There are also data gaps if you are, you know, doing controversy screening and you're using news.
There are parts of the world where there simply isn't a reliable set of, you know, media or news to be able to, you know, screen for certain risks. Also there may be, you know, data that simply is just so under, you know, yet to be defined that, you know, this data is something we would like to have, but simply can't get our hands on yet in the quantities that, you know, we ideally would. I think data gaps come in a lot of different shapes and sizes.
Today I wanted to, you know, walk through how we are addressing those and thinking about those specifically in context of our ESG scores as just one example of how data gaps can be filled and how we're thinking about making sure that while we're doing this gap filling, we are, you know, considering a number of factors, transparency being one of them. On the next slide, I've listed, you know, the challenge, at least from an ESG scores perspective. I think one of the challenges we as an industry face is that we are constantly being asked to cover more and more companies. You know, this was, you know, a few years ago, maybe less of an issue when, you know, we were only covering a few thousand companies.
Now, obviously, the need for ESG data has expanded so greatly across, you know, companies and in all different markets of the world, across different asset classes. Therefore, as we continuously increase the number of companies we're researching, you do get to a point where the amount of reported company disclosed information cuts off quite significantly. You know, that cut off, even though the amount of information that is being reported has been improving significantly over time, you know, that really steep cut-off tends to happen around four or five thousand companies. I think people forget that sometimes when you're researching company number 8,000, there really isn't a lot of information available in the public domain, or there may be very basic ESG disclosure.
As already alluded to before, you know, regulations, and I think increasing demand for this information from the investment community will help drive this. But, you know, realistically, I think this will take another couple of years. So that leaves, you know, actually the vast majority of most data providers universe struggling, you know, with this information gap because there is little to no information available on kind of the basic ESG factors that most providers would look at. The differences are pronounced differently depending on the size of the company, the industry sometimes even that a company can be in and the region.
We definitely see there are, you know, kind of elevated levels of maturity on ESG reporting in some industries, and some industries have managed to kind of get by with companies not really reporting much at all. That's a factor. The data gaps aren't even evenly distributed across the universe of companies that you would cover. I think, you know, again, as Lauren mentioned before, the more we get into small companies in, you know, large corporate supply chains, privately held companies, the information on ESG, but even financial information generally drops off quite significantly.
That is, you know, a huge new frontier to tackle because we are seeing, you know, increased interest and also having transparency on ESG in these companies as well. I think there is a lot of modeling used today. I mean, at S&P, we've been using, you know, models for over 20 years to fill gaps in environmental data that's being reported. Companies themselves often don't report Scope 3 emissions or don't have the full view on, you know, all the emissions that occur in their value chain. In the absence of that, you know, even companies come to us for a modeled value to kind of, you know, somehow orient themselves on, you know, what the actual value would be or to cross-check on whether, you know, the actual values they're collecting makes sense.
I think modeling is well understood as, you know, a very kind of scientific approach to try to fill information gaps. It works very well with, you know, quantitative information. It becomes increasingly complex when you're trying to model more qualitative information. I think often as users of this information become more sophisticated and aren't just using the outputs but are more interested in how this, you know, these modeling approaches are actually working, there may be a current lack of transparency on modeled approaches. You know, thinking ahead, if you're going to apply modeling as a way to fill gaps, then that needs to be done in a very transparent way. What are some opportunities?
It's not just about modeling, and I think we at least at S&P are very much focused on corporate engagement. We also believe that, you know, more transparency, more reporting benefits the entire market. The research framework that we run for over 20 years is very much focused around corporate engagement and kind of promoting companies to think about upcoming ESG topics, to start disclosing on areas that they may not feel 100% comfortable disclosing on yet. It's a process, and we've seen that the readiness of companies to disclose information to not just us, but to their stakeholders, has increased dramatically in the last few years and will continue to do so.
You know, there are other ways, and I think having more simplified metrics as a starting point for many companies would be a positive thing. A lot of the ESG reporting frameworks that exist today can be quite daunting. I know there are efforts underway to, for example, set a you know, a baseline set of ESG metrics, you know, basic things around diversity or you know, environmental metrics such as Scope 1 and Scope 2 emissions, you know, that companies can kind of ease into this idea of having to report information. Already then you have a pretty good set of basic information across a very large number of companies.
I think we are seeing a lot of uptake in terms of, you know, private equity investors, for example, starting to want to engage with their portfolio companies that are private, that may be very small on basic ESG topics because they themselves feel the need to report on, you know, the ESG profile of the companies that they're invested in. Also I think dialogue and education with, you know, the private markets and smaller companies using simplified metrics, you know, slightly modified approach is a good starting point to make sure that, you know, companies warm up to this idea that, you know, ESG disclosure certainly won't go away and that there are ways to do it, and which then, you know, for us benefits us by filling gaps.
The good news is that I think gap filling generally is very well understood in the ESG space. You know, most clients we speak to, they are using some kind of gap filling when they are looking at ESG information, whether that's provided to them by the providers of their ESG information or whether they're applying their own techniques in-house. You know, I think, you know, modeling as a way, for example, to fill gaps is not something that is new. It's not something that the users of ESG information shy away from. It's just maybe, you know, to what degree do they actually understand, you know, what is being modeled, how is something being modeled?
Really for us, I think the opportunity that we see is, you know, that, you know, if you're going to use modeling and gap filling, then be transparent about it. Alternative datasets can be a great way to fill these gaps. Maybe it isn't all about, you know, company-reported data, but we're also seeing that there are many different ways that you can use information that, for example, S&P already has on companies to start forming an opinion about, for example, their exposure to certain ESG risks or the possible negative or positive impacts that they would be having on the world through the products and the services they sell. These are things that companies sometimes themselves, you know, aren't necessarily broadly reporting, but we may have information on.
You know, that in combination then with company-reported data certainly helps to fill a lot of gaps, while not having to wait for, you know, corporate disclosures to evolve. I think the trick is how do you balance these two things, right? How do you make sure that you are gap filling, and that you are, you know, trying to fill, you know, disclosure gaps in your datasets, while at the same time trying to promote more disclosure? Because often we get asked the question, well, if you're gap filling, then what's the incentive for a company to actually disclose that information? Because we do feel that more information should be reported, you know, more broadly, that is something that we also consider. On the next slide, I have shown an example.
This is a screenshot from our S&P Capital IQ Pro platform. This I think illustrates how we're thinking about transparency, not just in applying estimations to our scores, but also in how we actually build our scores from kind of the different buckets. Here is a company score, you know, disaggregated into three buckets. There are parts of our ESG scores that require information to be available in the public domain. That is kind of that baseline measure of transparency that we expect all companies to report on. These, again, are basic metrics around corporate governance, basic metrics around, you know, environmental data like water consumption, waste, production, greenhouse gas emissions.
We know there's another kind of set of questions which are maybe slightly more advanced or go into a level of detail that maybe not all companies are reporting on today. Some companies may be reporting on it, you know, to us. Some companies may have this data internally, it's just not in their sustainability report yet. Those are maybe things that are also slightly more forward-looking in nature. Together, those kind of, you know, those two buckets create our, you know, classic, you know, ESG scores as they've been constructed for the last 20 years. Now, we know that, you know, for some companies, you know, there are limitations to what they can report.
That is sometimes there are constraints in terms of the size of the team that's responsible for ESG reporting in the company, the amount of resources they can mobilize to produce their sustainability report, or to, you know, respond to our survey each year.
you know, we also believe that, you know, to fill some of those gaps, we can also take a more light touch approach to imputation or adding, you know, imputation models to estimate some of these gaps without just saying, you know, we're gonna estimate all the gaps that we have in our scores because we do believe, you know, there are things that companies should be reporting on, and if you're not reporting on it, then, you know, we're gonna give you a zero one way or the other, which will impact your score. What we've done is we've, you know, applied estimation to a selection of questions that we have in our research framework, being quite selective about where we're applying it, so we're not just applying, you know, estimations across the board.
We have something called our disclosure analysis, which is essentially an assessment of how transparent a company is on those very foundational ESG metrics that we believe all companies should be reporting on. This is very much in line, I think, with what reporting standards are discussing today, kind of trying to coalesce around maybe not all the metrics that exist in all the standards, but what are the core set of ESG metrics that are really important, things related to climate change, things related to diversity that, you know, most people care about. We have this baseline for transparency, and that's really an important part of this because it kind of sets guardrails for how much estimation you want to apply.
A question we often get is, well, how would you make sure that you're not over kind of estimating something or that you're giving a company that hasn't disclosed information that they would get, you know, too much of a benefit by not having reported on something versus a company that has reported on something? I think that is where our disclosure analysis kicks in, and I think this is quite unique to us, is that we wanna make sure that, you know, the amount of estimation that we can apply is really being driven and controlled for by how much information a company is reporting. For example, if a company doesn't report basic, you know, greenhouse gas emission data, should we really expect them to have a very robust climate strategy?
Should we really be giving them, you know, a huge benefit of the doubt on some of these climate, you know, change questions when they don't even disclose basic environmental data? That is something that we spend a lot of time thinking about. But I think the most important thing that I really wanted to illustrate here is that we are trying to provide this in a way that you can always disaggregate the different parts. You could say, I'm looking at a score that has, you know, estimations in it, and that might be, you know, relevant, and I'll get to that on the next slide. For certain use cases and certain applications of ESG scores, it actually might be necessary.
At the same time, if you're an analyst and you wanna understand, well, you know, I'm really interested in how transparent a company was on a specific topic and why haven't they reported these metrics, you can dive into the details and kind of remove that estimation, look at the score without that. We believe that this is pretty important because I think having, you know, a single score that has some reported data and some estimated data all baked into kind of the same thing without being able to tell what is what is something that will become increasingly confusing to the people that are trying to look under the hood of these scores and understand how they're being constructed. One of the other things that we are planning to provide are confidence levels.
To each score we assign you know how confident are we in the model. That is really you know based on how much information we were able to find. You know in some cases that may be quite a significant amount of information, and in other cases, for some companies, there may only be two data points that they report in the public domain. You know based on that, you would wanna make sure that the person using the score really understands you know how that model was applied and the confidence level that results from that.
If we go to the next slide, you know, I think really in the ESG scores context, what this is trying to solve for is creating, you know, a distribution of scores across a growing universe of companies that feels a bit more normalized. If you look at ESG reporting, it's not a perfect distribution of companies. Right, this is what I was saying before. You have a very small set of companies compared to, you know, the tens of thousands of companies potentially, or hundreds of thousands of companies that you could assess in the world, that actually report on ESG information. This is likely to stay the same. You know, is to stay relatively consistent in the near term, across such a large universe of companies.
If you have a score that's based purely on, you know, what a company is disclosing, what we found as we added more companies is that you have a lot of companies and their scores clustered around the lower end of the spectrum. Companies that are getting very little points, because they're not disclosing anything. This again, if you're looking for transparency and understanding how a company is reporting, that is maybe a very useful score to have, and that is very useful information to have. If you are trying to do, you know, kind of a broad portfolio construction, and you need to be able to pick from a set of company scores that are very comparable across a very large number of companies, then you need to have a slightly better distribution.
By essentially applying estimations to fill gaps, you are creating a distribution that is slightly more normalized. It's not a perfectly normal distribution, because we also don't believe ESG is, you know, perfectly normally distributed. You will have companies that have been doing ESG for far longer than, you know, companies that are just starting today. You do have companies that are much more advanced than others. But this at least allows, you know, a set of use cases to be able to use these scores more effectively, while again, you know, giving you that transparency to go and look under the hood and see, okay, well, how transparent is the company really, and how is this estimation being applied?
Hello, everyone, and welcome to this live Q&A session dedicated to the future of ESG intelligence, tackling data availability gaps. My name is Olivier Trecco, and I'm the Head of ESG Solutions for ASEAN, Japan and Pacific at S&P Global Sustainable1. It will be my pleasure to be moderating today's live Q&A. First of all, I wanted to thank Lauren, Mona and Manjit for the very insightful discussion. Now, for the actual live Q&A session that we'll be having, I will have the pleasure to be joined by Manjit again. In this session, we'll be happy to answer all the questions that you can submit using the Q&A widget at the bottom of your screen. We do have about 30 minutes for this live Q&A, so please do come forward with your questions.
For those who want to review anything we covered, please remember this session is recorded and you will receive a copy, so you can access it on demand at your own convenience. Now, without further ado, I will start with the first question, which is for you, Manjit. How do you get reliable data from around the world and not just from developed countries? All right, Manjit, over to you.
Thanks, Olivier, and it's a pleasure to be here today to do this live Q&A. That's a really good question, and I think first of all, what we've observed is that the quality of ESG reporting and public disclosures has been increasing quite substantially, not just in, you know, in Europe and in the United States, but also really around the world. We actually see that over the last year, a lot of the real interest and uptick in ESG reporting, but also in engagement with us on our ESG scores, on our research process, has actually been driven by Latin America, by Asia. You know, we definitely see there is a global shift in terms of making more information available.
What it does mean is that, you know, you need to be prepared to accept information in different languages. You need to have analytical teams that can process that information, translate information when necessary. It is important then also to train people to understand the local context. I think also, you know, a part of this is education and then working with local markets in order to, not just, you know, promote more reporting and better disclosure and better data, but also to help companies and investors understand the importance of this. For example, you know, we've been doing a lot of engagement in Asia, for many years, for over a decade, I would say, working with local stock exchanges, working with companies.
Even before you get to the point where you have, you know, a large amount of data that you can use, you are working with them to basically educate them and learn about, you know, the local context as well that exists in those markets. We found that engagement with market participants has also, you know, helped us have a better understanding, but it has also helped to drive more transparency and more reporting, which in the end leads to better quality data that we can use in our daily work.
Thanks. Thanks for this, Manjit Jus. Actually, you mentioned data in Asia, and obviously, we're having this focus today. Could you perhaps give us a bit more information about your data coverage in the region in APAC? You know, and linked to this, you know, how do you scrub or clean the data that is being reported by companies?
Yeah. Absolutely. I mean, we again, our company coverage is global. We have been expanding coverage in Asia quite significantly. You know, some of the key markets are Japan, where we cover you know, most of the major Japanese benchmarks that exist in the market. We have also started ramping up our coverage in Mainland China. We've invested quite heavily there in building a team of analysts that again, have that local language capability which is really necessary to also understand the local context, as we see there is more and more information being reported in China, not just by the big multinational companies, but there is this big push to also have more smaller companies and mid-sized companies report, and we obviously wanna be prepared to capture that information.
The coverage in Asia is quite high and has been growing quite steadily. In terms of, you know, actual active engagement in our research process, we've seen, you know, very strong growth in countries like Thailand, historically where companies have done a lot to, you know, really improve the level of ESG reporting, and we see that other, you know, places in Southeast Asia now are also starting to follow suit. How do we scrub the data?
Without going into all the details of our process, we have a quite robust analytical framework that we've been refining for many years, which doesn't just take information that's reported, you know, either in the public domain or information that the company would give us through our direct engagement process at face value, but really looks at, okay, how does this information look compared to other pieces of information that we have on companies?
How does this information, you know, fit according to the very kind of strict guidelines that we've been developing with the aim of making sure that information is consistent, and clean to the point that when it enters our database, we can use it, you know, in a very kind of comparative way across different companies, but also within an industry to eventually create our ESG scores and other analytics. It is a multi-tiered quality control process where we're checking basically for, you know, against evidence. Is there evidence, either internal evidence or something in a company's publicly reporting that supports the information that's provided. If it's a quantitative figure, we'll do things like outlier checks to make sure that the number seems plausible.
I would also say that in my experience, when companies go through our assessment process or when they publish a sustainability report, there are generally many stakeholders that have to sign off on this information. You know, we also have a fair amount of confidence that information that is given to us has, you know, also gone through the relevant compliance processes and checks internally. There's checks happening at the company and also on our end, there are many checks happening to ensure that the information is, you know, relevant for our assessment framework, as accurate as it can possibly be, and, you know, ideally, that it fits with all the other information that other companies are reporting.
Okay, thanks. I think this links way into another question by a member of the audience, you know, about, you know, you check the quality of the data, but do you think that the emerging outcomes from ISSB, but also any other large standard-setting efforts, you know, will be useful to enhance ESG intelligence by creating more consistent, clear, and transparent methodology for reporting? You know, what's the impact you expect on quality and consistency of data in the coming months or years from these global efforts?
No, absolutely. I mean, we think initiatives like the ISSB are very important to harmonize ESG reporting because, you know, currently there are, you know, it's a patchwork of different frameworks that ask for similar things but slightly different. There's often not agreement around very obvious ESG terminology that one would expect, you know, should be similar. There are differences in how companies report greenhouse gas emissions even though, you know, we've had the Greenhouse Gas Protocol for a very long time.
You know, having a framework that brings together things that already exist and try to kind of refine them and align them at a very, very detailed level so that we're not just talking about the same concept, but we're actually using the same definition and the same guidance for companies. I do believe will help create clarity, at least on those ESG topics that are, I think, better understood, more widely reported. Some of these core concepts that, you know, we feel there shouldn't be any differentiation on, because that's where a lot of the confusion is created. You know, tied to that, the more companies use this reporting, the more regulation drives reporting.
I also believe that there will be more, you know, accountability in terms of what's reported and you know, while the existing ESG reporting frameworks today, they're not there to check, you know, what is ultimately ending up in the report. That is still the responsibility of a company or you know, a third party auditor that they have to check their report. Having you know, more consistent disclosure standards that are then also being used more widely, reported more widely, being used eventually in digital reporting formats for regulators or for investors to consume ultimately should also drive up you know, the quality of the data that's being reported.
All right. Thanks for this. Next, there's actually a great question from, you know, Alex in the audience and which I like a lot because I think it resonates with a lot of the questions and concerns we see from, you know, clients and institutions in Asia, which is the coverage of private data, which is its own particular domain. Actually, there's this one sub-question if you want is, would you allow grassroots data collection by the client themselves and aggregation to be an input to your data analytics for particular sectors or issues?
I'm not 100% sure I understood the point that the other person was trying to make around the kind of grassroots bottom-up data collection.
I guess.
What I can say. Sorry, go ahead, please.
Yeah. I say the way I think this question is, you know, is basically collection by the client themselves on their own investees, right? You have
Okay. Yeah. Got it.
Data access, which is, you know, proprietary to the investor.
Yeah. No, no, absolutely. That's actually something that we have been piloting, I would say, for the last year. You know, we have had private equity investors come to us and say, you know, "We are interested in collecting information in our investee companies." Right now they're being bombarded with all different kinds of questionnaires, some shorter, some longer. It's, you know, they're not in a proper platform. They're very difficult to kind of manage. We've actually been working with private equity investors to use the platform that we use for information collection, to use a derivative of the methodology that we use for key ESG metrics, also, obviously, listening to them and their particular needs around what ESG information they want to be able to collect this information.
Because we know that that's kind of step one, engaging these companies on the topic. As I was saying before, that's kind of part of that educational component, is important to kind of explain to them why this is important, why they should be reporting this information, and hopefully eventually get to a point where this information can be collected, you know, for a broader group of investors and not, you know, individual investors on individual companies in their particular portfolio. Absolutely. We are collecting this information today. We see different use cases. Some private equity investors are interested then in the score that we're producing as a result of this information, and others really just are more interested in that raw information that's being collected by us on their behalf.
Okay. Actually, I just wanted to complete this saying that especially this is something, like I said, we see in APAC and a lot of demand from clients to help with the collection of data on private assets. To have the, you know, to supply client with and help client with tools and infrastructure to manage, organize, and sort this data so that it can then be integrated in the analytics that Manjit is talking about. This is the whole one solution, you know, basically is something that we are currently very busy on because, you know, again, increasing the coverage for private assets is really something a focus that I think is shared by many of our clients.
I want next to take a few of the questions which are more dedicated to the imputation questions.
In the ESG, in the CSA methodology. Recently, there's been a paper which has been making the rounds, you know, on the, you know, on various dedicated specialized, journals online. A recent FTSE research, which indicated that there could be wide discrepancies between an estimation model and the actual reported figure. Would you have any indication of the actual accuracy of your imputation methodology for the CSA score?
Absolutely. I'll kind of break it out into two parts. On one hand, we are using estimation on actual data only where we believe it makes sense. Currently that is limited to the environmental data that we have. You know, emissions, Scope 3 emissions, water consumption. These are metrics that S&P, through its Trucost business, has been modeling for, you know, over two decades. That's something that we feel relatively confident with. Our environmental data is being used, you know, across the world, so we're pretty confident that, you know, we have modeling, you know, greenhouse gas emission data, water data, this type of information down. That information is flowing in to our assessment and is being used.
For the other ESG themes, we looked at, you know, are we able to model specific numbers? Are we able to model other kind of values? Especially when you get into the S dimension, it becomes a little bit trickier. We haven't tried to do that because we, I think to the point that's being made, you know, don't feel that we can do that with enough accuracy. Instead, what we've discussed today in this webinar is an imputation approach that has been applied at the score level. Rather than trying to say, you know, we've estimated the underlying information, which, yes, may then be off to kind of the actual value, what we've instead tried to model is the score.
The expected score for a company on a specific topic, not the raw underlying information, based on you know how well and how much information a company has provided on kind of an adjacent topics that they actually do disclose information on. That you know it's still modeling at a fairly granular level, but we haven't actually tried to you know remodel values. The way we've checked that is we obviously did this modeling for many companies which we do have a lot of data on and complete data on, and made sure that our models were good predictors of you know the score that a company actually would have gotten if we actually used the real data that we have.
There was obviously a lot of testing involved to make sure that the models were a good predictor of the estimated score that the company would get.
Okay. Another question from the audience is, can you actually give examples of what you would estimate and then what you could consider all companies should be reporting on? In other words, you know, what's eligible for imputation?
Roughly one third of our research framework of the questions that we collect data on for companies are data points or questions that require information to be in the public domain. Those topics are generally the, I would say, I like to call them kind of the more standard or core ESG topics, like corporate governance metrics, which you would expect to find in the company's annual report, often mandated by law. They need to report this information to their investors. Things around, you know, business ethics. Again, these are things that we expect to be on the company's website with regards to their codes of conduct. Basic environmental data. If a company isn't disclosing basic greenhouse gas emissions data or water data, then you know, that's possibly a concern.
Other kind of adjacent environmental metrics. These are things that we wouldn't normally expect to be in the public domain or reported to an organization like CDP, for example. Then on the S side, we, you know, things like a human rights policy. If that's not publicly available, then we wouldn't try to attempt to model whether that, you know, is actually publicly available or not, and whether the company has one. Generally speaking, things like policies and programs that we expect companies should be reporting in the public domain. These are never modeled or never estimated when there is missing information, with the exception of some enhancements we make on and around environmental data, as I mentioned before.
Where we do tend to impute more is on topics that are quite detailed and go into a level of granularity that you potentially don't find in the public domain. Just to use an example, to stay on human rights, you may have a question on a human rights policy which should be in the public domain, and then you may have follow on questions that go into more detail around due diligence, you know, processes and actually numbers in terms of, you know, how many audits the company did in the last year, whether they did audits. These are things that are often partially disclosed in the public domain, but not fully.
There, because we have information that is in the public domain, and we can gauge, okay, a company has a good policy on human rights, they do have some information, then we would model some of the missing questions and estimated scores there. But it's often things that are often not widely reported in the public domain or as topics discussed in corporate reporting, but don't go to the level of detail that we're asking for in our methodology.
All right. Thanks. I think that answers also some of the other questions we received on how do you do ratings or imputation when for policies which are not available in the public domain. I think that got that cut through this, so thanks a lot for that. I guess a couple of questions more. We still do have a few minutes. One is, you mentioned earlier about alternative data sets. Could you perhaps go into a bit more details on those data sets that you know you were talking about?
Yeah. Absolutely. I think increasingly as we are trying to address, you know, ESG through so many different lenses, there is a climate lens on ESG. There is, you know, looking at things in terms of risk or in terms of impact. We see that, you know, company reported data, that is one source of information. Already today in our ESG scores, we're using, you know, company reported information, and we overlay that with controversy information, basically, you know, which is how a company is actually behaving in the outside world versus what it's reporting on in its sustainability report. That's nothing new.
I think increasingly, as we are becoming kind of more focused on different, you know, components of ESG, we see that there are data sets, like, for example, you know, climate data sets, you know, weather information, where companies actually have their assets, so where their, you know, factories are located. When you take all of this information together, you can, for example, do, you know, really advanced risk analysis on a company's exposure to the physical risks of climate change. That's just one example where you may be taking satellite data, weather data, you know, meteorological data and mixing it together to come up with a view on ESG.
You could, for example, add a company's reported information around its programs and policies on climate change to overlay that to create kind of a new analytical data set, which is quite interesting. These are the types of information. I just used one example in climate that I think we are increasingly seeing are very compatible in a way with company reported ESG data. At S&P, we have lots of these data sets that aren't, you know, flagged as being ESG data sets, but when used in combination with other information, can be quite powerful.
Yeah. I think we have slightly below 10 minutes left. As we moving towards the end of our live Q&A session, you know it's time for the tricky questions, you know, or a bit more challenging. One we have is, you know, about imputation again. Wouldn't imputing hide trends? You know, so let's say you have a company that makes ESG steps would not see an improvement in its scores as it advances in its reporting and transparency. How do you solve for that issue in your methodology?
That's a really good question. That's why it was very important when we developed this approach to estimation, that it's not just simply giving a company, you know, the sector average or the sector, you know, regional average as a placeholder. We didn't want, you know, the estimation to just create a placeholder score for a company. The scores are very much driven by how much a company is reporting. This is also a very important safeguard to make sure that we're not overestimating. I think this was also, you know, when we were also speaking to clients as we were developing this, that was one of the concerns that we heard again and again.
Well, how can you ensure that you're not overestimating?" I think the general feeling was in ESG, it's better to be a bit cautious and maybe slightly more conservative and not overestimate a company's, you know, potential or expected performance on a specific topic. Having those guardrails means that actually the more a company reports and the better, not just the more reports, but the better quality that information is and the better they perform on ESG topics, the more they will benefit. I mean, a, there'll be less gaps to impute and less gaps to fill, which is a good thing. Also they will benefit from, at least using our estimation model, they will benefit from having more information that informs that model. Actually, you know, it.
We were also thinking about, you know, if you just estimate something for a company, what is the incentive for them to actually disclose this information? Because we are directly engaging with companies, and you know, direct participation, providing us more and more information, driving more transparency is absolutely key. That was an important, you know, mechanism for us as well to consider, is that, you know, it shouldn't just stop with estimation, but estimations should be used as an add-on to what you're reporting, and there should always be a clear incentive to report more and to kind of improve your performance.
I fully agree. I think that's the hierarchy, right? That we start from actual reported data with either public or privately accessed, and then estimation is the, you know, the goal to 100% coverage, which is the ultimate goal and through the most specific and accurate ways. Second, slightly technical question on imputation. Could you elaborate a bit more on why your imputed scores have a more normal distribution than the scores which are driven exclusively by disclosures?
I think this goes back to the way our methodology was built and we, you know, in our baseline scores that we've been using for a very long time, we're penalizing companies for not disclosing. That obviously results the more companies you add to your research universe that have, you know, little to no disclosure, that clusters the amount of companies that have very low scores towards, you know, one end of the spectrum. By adding estimation, you're not penalizing companies as harshly.
You're still penalizing them for not reporting on those, you know, core or standard ESG metrics that I mentioned before that should be in the public domain, where we definitely see no frameworks like the ISSB are heading to have this kind of baseline minimum requirement for what companies should be reporting on. They're still gonna get, you know, penalized if they don't report on those topics. But we also recognize that there are many other topics that we cover in our methodology that, as I said before, companies may be doing things in this area, but they may not be publishing, you know, the level of detail that we want in their public report. Is it fair to, you know, penalize them as harshly as we were in the past?
You know, that's a question we ask ourselves. What that does is that you're still, you know, rewarding companies that are reporting. You're still penalizing companies that aren't reporting on the really important topics. For the other topics, you are, you know, rather than giving a company a zero, you're giving them a value which helps to bring the overall distribution slightly more towards the right. Now, it's not a normal distribution. I think we can agree that, you know, if you look at ESG reporting and where companies are on their ESG journey, it's not a perfect distribution. Yeah, it helps the distribution, but it doesn't make it a perfectly normal distribution.
Okay. Understood. I think we have time for one last question. We have perhaps three, four minutes. One is really, you know, estimation comes, as you mentioned, with varying confidence levels, you know, depending on the quality of the input you have, and so you can have varying confidence ranges. How do you aggregate scores which have different confidence levels? Do you weigh? You know, what's the way to have a consistent aggregated overall score?
That's a great question. I think for us it's really about providing transparency at each level of the aggregation. We're still using the same aggregation and weighting schemes and weights that we use in the scores without imputation, because we need them to be, you know, comparable at the end of the day. What we've heard, speaking to, you know, the market as we've been going through this development process, is very much this need for transparency, so that, you know, an end user of the information can always go back and say, "Okay, well, I'm looking at an aggregated score," and we do have some information already today that's showing the percentage of questions, for example, that have been estimated that roll up into that score.
We are working on providing additional features and functionality around confidence levels. The feedback that we received was, you know, overwhelmingly, "Okay, that's great. You can see a score at a high level." There is an indication at the aggregate level already of confidence. Then what's really important is that I can drill down into the underlying pieces of information and really see at the individual, you know, ESG topic level, what is actually modeled and what is actually reported by the company. Rather than trying to, you know, bundle this all into, you know, a single score without any indication of, you know, what is being modeled and what is not.
That is certainly something that is very well received and we will continue to work on as you know more people are using this and providing us feedback.
Okay. Thanks so much for this, Manjit. I think that seems to be all the time we have for questions. Thank you all. First of all, thanks, Manjit, for taking the time to answer those questions. It was very insightful. I hope you know, we all had an exciting, interesting hour. We covered a lot today, so if any of you in the audience have any follow-up questions, please use the Contact Us widget you have on your screen, and of course, we'll be glad to assist. Once again, for those who want to review anything we covered, this session has been recorded and you will receive a copy shortly, so you can access it on demand at your own convenience.
In addition, last but not least, when you close out the webinar, you will be routed to our webinar survey form. Please, we'll be very happy to hear your feedback, so please take a few moments to complete. That will be all for today. Thank you very much for your attention.
Thank you very soon.