A tool to explain index weights

Paul: [00:00:00] Welcome to Index Ideas from FTSE Russell. I'm Paul Amery, your podcast host. In this podcast we look into how FTSE Russell indices are built and why. We explore index ideas that can help you address real-world investment challenges. As a reminder to listeners, you can't invest in an index. And so the concepts that we discuss during the podcast are not investment advice. Any reference to potential investment strategies is intended for informational and educational purposes only.

Paul: [00:00:28] In this episode of the podcast, we look at a tool that helps index designers explain why they've arrived at a particular set of weights for the components of an index. I'm joined by Hannah Layman, who is head of sustainable index research and design for FTSE Russell. Hannah, welcome to Index Ideas.

Hannah: [00:00:45] Thank you very much for having me.

Paul: [00:00:47] Hannah, why do we need a tool to explain index weights?

Hannah: [00:00:50] So when someone invests in a product that tracks an index, they're making an active choice about how much of their money or how much of their investment weight they put in each company. And a tool like the weight explainer helps clients to understand and explain the drivers of different weight positions. It allows you to visually represent why is more of my investment weight in company A than in company B?

Paul: [00:01:12] And why is the weight explainer particularly important for sustainable indices?

Hannah: [00:01:16] So traditional factor investing or style investing generally wasn't accompanied by personal value judgements. We didn't have investors coming and saying or questioning why does stock A have more weight than stock B in a quality index? Now it's very different for sustainable investing. We found that our clients are getting significant inbound questions about relative weight positions. So why does stock A have such a high weight in my ESG product? Or why does stock A have a higher weight than stock B in my green revenues product? And that's the sort of thing that the weight explainer allows you to visualise and answer.

Paul: [00:01:52] We had an episode with Andreas Schroeder our colleague on the first season of Index Ideas, where he talked about different ways of building sustainable indices. What are the drawbacks of optimisation and other black-box approaches to index construction?

Hannah: [00:02:14] So optimisation is a powerful and very well adopted mathematical technique for constructing indices. It allows you to design an index with the best possible outcome for a particular objective. The most common objective being to minimise ex-ante tracking error or variance from your underlying performance. It says effectively find a solution of a set of weights that meets all my objectives. Of all the possible solutions, find the one that minimises the chance of a difference in performance from the market. So that's obviously a very powerful tool. However, what it can mean is that you might have company A that scores better on every single metric in your set of objectives, but still receives a lower weight than company B. And when you have to go to explain that you'll have to explain that it's driven by that objective function. So you have to say because of covariance. And in that way optimisation can make the explainability of your relative weight positions quite difficult.

Paul: [00:03:10] So why does FTSE Russell's tilt-based methodology, which we use in sustainable and other indices make it easier to explain to clients, how do we arrive at particular index weights?

Hannah: [00:03:24] So as well as offering optimisation, FTSE also does single tilting, which is an alternative construction approach. You start from the benchmark weights and you use a multiplicative approach to tilt towards the factor of interest. You can then use a solver to say, how much should I tilt to achieve my specific objectives? Now you can do this across multiple tilts so you can still have multiple objectives. Each objective gets its own tilt and they're applied simultaneously. The key difference to optimisation is that within each tilt we maintain rank order. Now what that means is if company A scores better than company B on a particular factor, it has to be rewarded by that tilt. You're not allowed to penalise it for that tilt if it's scored better. It's that particular thing that makes this technique incredibly transparent, because I can break down each of those multiplicative tilts into a chart and show you exactly what's driving a stock position. And that's what we do in the weight explainer.

Paul: [00:04:20] Thank you Hannah, for explaining that. Can you give an example of how we use the weight explainer in a sustainable index, perhaps at two different companies?

Hannah: [00:04:28] The example that we give in the “weighing in on transparency”, which is the accompanying paper, the index construction is particularly complicated because trying to explain relative weight positions is even more important when you have lots of objectives that your index is trying to achieve. So in this particular case.  we looked at the EU climate transition benchmarks as an example, or the developed PAB, Paris-aligned benchmark, and we looked at the weight of Amazon and we said, okay, well, in the underlying benchmark, the weight of Amazon starts at 2.7%. But in my index it's 5%. So I'm rewarding Amazon. So what's driving that reward? And we broke down every single one of the objectives that's built into that index design and showed you how much of each of those objectives drives that overweight position. And so in the example of Amazon, you can see that the majority of that is coming from the fact that it's relatively low-emitting compared to the rest of the market. And it's also got its exposure to green revenues and a relatively good management quality score by the Transition Pathway Initiative.

Paul: [00:05:33] So Hannah, you explain that you start with the market capitalisation weights for a particular stock universe, and then you apply the tilts. Does it matter how many tilts you apply? Is it preferable to limit the number of tilts in the model?

Hannah: [00:05:46] So the number of tilts represents the number of objectives. And the more of these you add to an index design, the less companies you have to tilt towards that meet all the objectives. So you're effectively just making it harder and harder for the index to achieve everything all at once. You're wanting more out of your product. Mathematically, we call this reducing the solution space. Now, what that means is that you're going to have less companies that you can tilt towards that satisfy all these objectives. When you do that, when you add multiple tilts and particularly in the sustainable space, it's really important to balance the number of objectives with investability. So you do that by monitoring metrics such as concentration, diversification, turnover, and make sure that you're not putting too much pressure, you're not asking too much of your index.

Paul: [00:06:29] As a follow-on to that question, how do we, if we are limiting the number of tilts to take into account investability concerns, what other index design rules do we typically apply in a sustainable index?

Hannah: [00:06:42] So through tilting, you can control your exposure through constraints, and they're treated just the same way as you would set an objective. They form a tilt in the multiplicative function. So, both your targets and your constraints act as objectives that you're asking your solver to find a set of index weights that meets all of your objectives within your constraints. So all of your targets within your constraints. So those constraints might be things like, I don't want to take more than a plus or minus 2% active bet in any particular country, or I don't want to have particular style factor exposure. So every one of those constraints that you add can also be visualised on the explainer to see if it's driving particular relative weight positions.

Paul: [00:07:24] And Hannah, you've talked about adding tilts to achieve multiple objectives simultaneously. When you're looking at those objectives or those sustainability factors, is it better to have factors that are less correlated with each other or does that not matter?

Hannah: [00:07:41] This is a really good question in particular for the sustainability space, because we often have lots of objectives that are aligned with our ESG philosophy that are correlated. Now, it's very common practice to put multiple objectives into an index that are correlated. However, it does add complexity, particularly when they result in competing objectives. So, for example, you might have someone who wants to design a product that tracks an index that tilts away from companies with exposure to fossil fuels ownership. Now, the same investor might say, I also want to have the index capture companies, tilt towards companies that have exposure to green revenues. However, these are often the same companies, being diversified energy producers. And so you are both tilting towards and away from the same set of companies. You're effectively reducing that solution space even more because you're trying to find companies that both have low fossil fuels ownership and have green revenues exposure. It just makes that balance even more important in terms of understanding the different elements of Investability.

Paul: [00:08:45] So the choice of sustainability factors is really important. You should consider ones that are more independent, in other words.

Hannah: [00:08:52] Yes or you can certainly have ones that are correlated, but ones that are independent will give you, get cleaner results, I guess.

Paul: [00:08:59] Okay. Thank you for explaining that. And what feedback have you had from clients on the weight explainer?

Hannah: [00:09:04] So we've had really positive feedback from our clients on the weight explainer. In particular, it helps them help their clients. So what we find is that our clients are reporting that they were getting a huge amount of inbound queries, particularly after rebalances, about what was driving relative weight positions. And what this tool enabled them to do was really quickly both visualise and explain but also share that visualisation to be able to explain it to their own clients.

Paul: [00:09:28] Hannah, how are you using the weight explainer tool in index development?

Hannah: [00:09:32] We would use this tool in exactly the same way that our clients would use it. We use it to interrogate relative weight positions and in order to investigate the index design itself.

Paul: [00:09:41] Okay, great. And finally, where can listeners to the podcast go to learn more about the weight explainer?

Hannah: [00:09:47] We have our paper “weighing in on index transparency”, but otherwise come and talk to us. We can build you a version of the weight explainer and we'd love to have a conversation.

Paul: [00:09:55] Hannah, thank you very much for joining me. And that's it for this episode. If you've enjoyed this conversation, then please follow us and give us a positive rating or review on your podcast app of choice. If you'd like to get in touch with the show, you can do that via the email Fmt@lseg.com. But for now, from me, Paul Amery. Goodbye.