"Without data, you’re just another person with an opinion." The famous quote from the American statistician Edwards Deming, who died in 1993, has never seemed more relevant than in the present day. In the world of finance, the scope of usable data to gain an edge over the market is continuously expanding. In stark contrast to traditional economic indicators, this so-called alternative data is plucked from a variety of sources. The mix includes credit card transactions, web traffic, mobile devices, IoT (Internet of Things) sensors, weather forecasts, satellite imagery, corporate flights, hospital admission rates, ESG (environmental, social and governance) data, and government contracts, not to mention of course sentiment expressed in the press and on social media, which we develop further in this report.
As alternative data comes from immense databases, machine learning and artificial intelligence are used to process it all.
Figures show that this strategy is developing fast. The number of alternative data providers has increased 20-fold in the past 30 years, states a report from the Alternative Investment Management Association (AIMA), estimating nearly 450 currently active providers as opposed to a mere 20 in 1990. In addition, more than three-quarters of investment firms use alternative data, according to Ernst & Young and associates.
A flurry of recent examples confirms this. In April, the Swiss National Bank (SNB) published a paper entitled "Nowcasting economic activity using transaction payments data". In it, the SNB explains how it assesses the value of "high-frequency payments data for nowcasting economic activity". The authors state, "Focusing on Switzerland, we predict real GDP based on an unprecedented ‘complete’ set of transaction payments data: a combination of real-time gross settlement payment system data as well as debit and credit card data. Following a strongly data-driven machine learning approach, we find payments data to bear an accurate and timely signal about economic activity." They then go on to say that "the payment models slightly outperform the benchmark models in times of crisis but are clearly inferior in ‘normal’ times.... We thus conclude that models based on payments data should become an integral part of policymakers’ decision-making."
One might start to wonder whether investors can legitimately get by without these tools
In January 2023, Spain’s central bank (Banco de España) published the working paper "A New Supply Bottlenecks Index Based on Newspaper Data" on its website. Here, the institution presents a new monthly indicator developed to measure supply bottlenecks using newspaper articles. The authors write, "The supply bottlenecks index (SBI) provides a consistent description of supply issues related to wars, natural disasters, strikes and, most recently, the COVID-19 pandemic."
One might start to wonder whether investors can legitimately get by without these tools. No, says Julien Leegenhoek, founder and CEO of the investment fund Taranis, which specialises in alternative data analysis. He believes that "data will increasingly be the differentiating factor". The Geneva-based company, launched in 2020, is unique in that it uses no economic indicators to decide on the composition of its funds. By taking this approach, the firm is practically a prototype of what finance can do off the beaten track. "We try to capture the collective point of view exclusively through market sentiment analysis and other alternative non-financial data," Leegenhoek says. It’s about applying crowd psychology to investment.
Academic experts are puzzled as to what to do with this extreme approach. "We’ll have to see over the long term," Amit Goyal, professor of finance at the University of Lausanne, suggests politely, "but I have my doubts that this type of strategy will always be relevant." That said, the initial results are promising. In early October, the company’s website proudly announced that the Taranis Market Sentiment fund had been nominated for the HFM European Performance Awards 2023 for the third consecutive year.