Will AI Disrupt the Index Industry? | The Financial Technologist

When thinking about industries that have been hotbeds of innovation over the last couple decades, the index industry is usually not one that comes to mind. Born from a need for publishers to measure the movements of broad markets, few could have predicted that financial indices would go on to underpin one of the biggest disruptions to the money management industry: the shift to passive investing. Today, millions of indices exist and cover every conceivable corner of the market – across asset classes, investment approaches and financial exposures. The resulting building blocks, and associated customized solutions, have enabled index users to build more targeted and efficient portfolios. In a 2022 Opimas report, it was estimated that global spending on indices would reach $6.3 billion, with annual growth at 13%. It is no wonder then, that this is an industry that now employs top quant, data and technology talent and is fiercely competitive. However, no industry is immune to disruption, and winners and losers from both growth and hype in Artificial Intelligence are already emerging. If we take for instance, the STOXX Global Artificial Intelligence index, which measures companies positioned to benefit from a shift to AI, it has gained roughly 50% in 2023 through mid June, led by Nvidia, which is up roughly 200%. In this article, we peek into our crystal balls to see where AI will take the industry. To assess the disruptive impact of AI, we first need to be specific about what these words mean. For a start, let’s distinguish AI from AGI – Artificial General Intelligence, i.e. human level reasoning– to remove considerations of extinction level events. Instead, we focus on a definition that the writer Ted Chiang termed ‘applied statistics,’ e.g. the enablement of computers to analyze data, identify patterns, and make decisions or predictions that can improve with usage. We already know of the tremendous progress being made in this field from machine learning to neural networks to natural language processing, evidenced by the many GPT models being released today. As much as machine learning is a subset of what we consider AI, disruption is also a subset of innovation. Innovation is ultimately a discipline of problem solving, and can happen across a broad spectrum - successful innovation strategies are cognizant of this. While there is no industry-standard classification of this spectrum, one can think of innovation as happening on a scale between incremental innovation - improvements to products, processes and services - up to disruptive innovation - where a new technology or business model upends an entire industry. With the terms defined, we can revisit the question: Will AI disrupt the index industry? Our view is that the proliferation of AI technology – as we defined above - will more likely be a productivity driver than a force for disruption in the index industry. Large index providers have historically enjoyed a wide moat derived from product ecosystems, but competitive forces are being fueled by innovation. We believe AI is a technology that will not threaten those moats by itself, but can lead to advancements up and down the value chain for both more established and newer entrants. As a few examples, when it comes to who creates strategies – proficiency in programming may be less of a prerequisite to being an index developer. With regards to what those strategies track – Natural Language Processing + Machine Learning can extract more signals from the plethora of available data, allowing for more specificity in how and when to implement highly targeted exposures. This is a continuation of a trend that has been well underway in the world of Thematic Indices, for example the STOXX Global Metaverse index, which uses patent data as part of the component selection process. When it comes to how customers interact with index data – AI can understand and potentially explain what users care about, to better serve their needs. Many of these examples are problems being worked on today by index providers, but recent advancements will add fuel to the fires. Naturally, these advancements pose a management problem as well. Like with other industries, how to acquire the requisite talent and computing power, as well as how to prioritize projects, will be key determinants of a successful strategy. Similarly, ethical usage that avoids spurious results and ensures index strategies remain deterministic (reproducible) will need to be a focus. There is also Amara’s law – the tendency to overestimate short-term impact and underestimate long run impact – to contend with. Thankfully, these are not new problems, but with the AI transformation, we may have a few more tools in our arsenal to attack them. Read more articles like this in The Financial Technologist. Download your free copy here.