Robo-Analysts are making their mark on the Wall Street
An extensive Indiana University study reveals that machine-based investment recommendations are getting better
Automation has already taken over the more repetitive tasks in the global ecosystem and the next stop is more analytical positions where humans are still perceived as a better choice among the two. But with the pace at which the digital disruption is happening, this notion is probably not going to last too long. The one sector that has seen the most automation in its processes has been the financial services. Fintech revolution of the last decade has done its part to nudge things along.
Trading bots, Algorithmic trading & more recently AI trading models have already become a common occurrence in the high-frequency trading environment that we experience these days. On a typical day, 50–60% of the trading decisions are made by computers & this ratio jumps to as high as 90% with any bouts of volatility — the algorithms kick in and magnify these volatility driven moves. Basically, the machines take over!
According to a new study by Indiana University, fresh signs of Robo supremacy are emerging on Wall Street. The research analyzed 76,000 reports issued by seven different Robo-analyst firms between 2003 and 2018, as reported by Bloomberg.
Traditionally, Wall Street research departments connect investors with company executives while gathering earnings and other corporate data. The number of clients basing their investment decision based on these calls is still limited. For the study, the researchers looked at various fintech firms whose foundations were based on the idea that digital technology does a better job than humans in making investment recommendations.
“Using this type of technology to make investment recommendations or to conduct investment analyses is going to become increasingly important. Technology is one of those disruptions because you can do things probably at a lower cost and greater scale.” ~ Kenneth Merkley, Author
One of the prominent automated firms studied in the research was New Constructs — a Robo-analyst service whose clients include quant funds, consulting firms, IRAs and asset managers & can buy services within a wide price range of $10-$150 depending on the research. Other automated recommendations from firms like Minkabu, Rapid Ratings & TheStreet.com were also taken into consideration.
Researchers used the Robo-analyst recommendations against those of traditional analysts as part of the Institutional Brokers’ Estimate System. Long-term hold strategies were not considered in their comparison as they are much less actionable than an active trade.
The authors of the study are quick to point out the advantages of the Robo-generated automated recommendations, saying they are not prone to human biases or subject to conflicts of interest. While this might not lead to immediate gratifications like human analyst calls, but they end up giving substantial returns eventually.
Further findings included Robo-analysts producing a higher number of investment recommendations since they were able to decode large sets of complex data (corporate disclosures & SEC filings) much more efficiently & frequently. Automated services were also shown to produce more sell-side ratings — about a quarter fell into this category as compared to only 6% by the humans.
On the buy-side, the human bias was pretty evident producing 47% of the recommendations compared to 30% Robo-analyst recommendations. One way or the other, human analysts have been under pressure recently from the growing and improving automated competition. According to Wells Fargo Securities, Wall Street and the banking industry could lose up to 200,000 within the next decade.
Critics argue that a significant portion of the population still likes the human touch and wants to have a one to one conversation with a person they trust. While we may see both these systems work in conjunction, for now, an eventual take over by Robo-analyst which continue to evolve & improve is imminent.
Complete findings of the research were published in SSRN.