A powerful AI platform originally developed with the goal of putting Deep Neural Networks to work on European public health policy issues has steadily gained a reputation for being one of the most compelling technologies available to financial markets participants.
Founded in 2012, Irithmics’ deep roots in public health epidemiology made an application of the platform’s AI to the information asymmetries that characterize the financial markets a natural development, particularly given the finance industry background of its CEO Grant Fuller.
In 2018 Irithmics partnered with the London Stock Exchange to provide issuers with differentiated market intelligence and investor analytics. The platform’s AI analyses hundreds of thousands of portfolios from 250,000 global institutional investors and funds.
Taking modern portfolio theory as an analytical starting point in terms of expected returns, volatility and correlations, and augmenting this with the views (expectations) of fund managers, Irithmics’ deep neural network approach maps investor behavior, identifying and establishing patterns. Eighty unique characteristics are used to classify and describe behaviors and dynamics of investors.
Irithmics is not concerned with the specific structure of any individual portfolio, but like John Snow deciphering the Cholera code through a Victorian water pump, the platform is concerned with what the behavior and dynamics of an entire population reveals.
It is the ability to analyze large numbers of portfolios and discover underlying patterns and trends that influence their structure, management and performance that delivers the most valuable insights. By tracking changes in investors’ holdings and global macroeconomic data and evaluating these changes through the lens of deep neural networks the platform enables finance professionals to establish a deeper understanding of market dynamics and participants’ expectations.
The results of this reverse-optimization are a powerful exercise in de-obfuscation.
Amongst a multitude of analytics, Irithmics is able to produce 90-day Supply & Demand Pressure (SDP) forecasts generated by the AI’s ongoing assessment of global institutional investment strategies and dynamics. At the end of 2019 Fuller’s team queried the platform based on FTSE 100 data up to 31st December 2019. The 90-day forecast to be generated would cover the period up to and including what is commonly interpreted as a massive pandemic driven sell-off that began on 20th February 2020.
Fuller explains that Irithmics’ AI had revealed indications that a significant proportion of FTSE 100 corporates would experience a sell-off in February/March 2020, and these were already present in the data even though the sample predated any knowledge of COVID-19 whatsoever. Unfortunately for conspiracy theorists, this is not an occasion to rejoice. “One of the hypotheses as to why the FTSE 100, and many other indices, fell so sharply is that it coincided with strategies naturally looking to reduce their exposures to FTSE 100 companies. That was anticipated on the 31st of December 2019. Irithmics’ understanding of strategies was highlighting many were likely to reduce their exposures to FTSE 100 companies around February and March. My personal hypothesis is that when COVID-19 began to unfold, investors began seeing the market turbulence and uncertainty. Knowing their strategies were to reduce exposures anyway just thought, ‘Oh, you know, we were going to do this in a weeks’ time anyway. The market is beginning to fall quite a lot, so, why don’t we simply bring forward our plans to reduce our exposures and reallocate the capital sooner?’ Of course, when this happens with enough people, the market crashes really quite significantly.”
Although Irithmics supplies data, such as the SDP data feed, to asset managers, it can’t be pegged as a traditional alternative data provider. A focus on the information asymmetry which exists in markets brings Irithmics and asset managers together, but Irithmics isn’t about alpha discovery, Asset Management often is.
Irithmics is not collecting, storing and commoditizing unique or rarefied ‘raw’ data sets in the traditional alternative or ‘exhaust’ data sense. They are not in the business of providing satellite imagery, consumer transactions, shipping container receipts, geolocations etc.
“Irithmics is mapping how institutional investors change their portfolios, not on the basis of an individual manager but as a collective population of managers. We have no interest in the individual manager or fund at all. What we’re trying to do is understand how the entire ecosystem of strategies that are at play in the market at any one time interact, affect the allocation of capital and risk, and ultimately asset prices.”
“Irithmics is attempting to describe what everybody else thinks is going to happen and how everyone else is positioning themselves in the market. Irithmics is not trying to find alpha or predict risk with its data sets in and of itself, instead we’re analyzing the behavior of others and decoding where their behavior suggests there is alpha or risk.”
Nevertheless, Irithmics recognizes that with SDP data, for example, alpha can be identified within its derived datasets, or as the result of a ‘blended’ approach incorporating other datasets. Irithmics is listed as a provider on alternative data platforms like Eagle Alpha and Neudata, Irithmics’ potential is becoming apparent to a growing number of large asset managers aiming to combine outputs from Irithmics analysis with their own quantitative strategies.
“At the end of the day,” Fuller says, “all of the data that Irithmics is using within its algorithms is publicly available data. It can all be easily sourced, and asset managers almost certainly already have.
“Irithmics provides a way to understand how others are behaving in response to those data. These subtle changes in behavior, in how others are allocating and reallocating capital as they receive, analyze and absorb new data are helping asset manager and corporates better anticipate investor reactions.”
Learn more about Irithmics AI-forecasted supply and demand pressure by downloading ‘Decoding the Impact of Institutional Investors on Market Supply and Demand’ describing the use of artificial intelligence to expose patterns in the allocation behaviors of institutional investors with respect to shares in Barclays and IAG.
Logged-in members can download the article by clicking the link under all the “Related Posts” below. If there isn’t a link then you aren’t logged in! To log in or register visit here.