As a follow-up to the R Squared blog on the impact of automation on bank branches, let's take a look at how the broader financial industry is changing due to a transformative technology: machine learning.
While machine learning (ML) has existed since the 1950s, only in recent years has the hardware progressed enough to make ML applications relevant—and potentially necessary—in business. This year, an index that tracks stocks that are poised to benefit from increased adoption of robotics and machine learning is up over 40%.
You may already be aware of ML's impressive strides across several applications, such as Netflix and YouTube recommendations, medical diagnosis through image pattern recognition, and sentiment analysis of textual data.
However, let’s see how ML may affect a few aspects of the financial industry.
Rise of the Robo-Advisors
The Department of Labor's impending Fiduciary Rule will require financial professionals to act in the best interests of their clients (one would hope!), rather than just providing "suitable" investments for a client's needs; they will also have to clearly explain the compensation they receive. As such, since the rule could meaningfully put pressure on the industry's commissions, wealth management firms might rely more on automated advisory capabilities.
Moreover, independent robo-advisor services, such as Betterment, are gaining traction in the smaller segment of the industry. And while human touch is still highly valued among net worth individuals, ML could help human advisors with portfolio analysis on the back-end of their services.
Could there be pitfalls to ML in wealth management? What if all the advisory algorithms are trained on the same data, and thus learn the same rules and patterns in order to maximize each client's risk-adjusted returns? At worst, this could conceivably cause a crowding of strategies, exacerbate the prices of over-valued and under-valued securities, and possibly lead to a flash-crash.
Some investment banks have indicated they are implementing artificial intelligence (AI) capabilities into their trading services. Mizuho is reportedly going to offer clients AI equity trading services in certain parts of Asia that will help clients trade large volumes at a lower cost. And, JPMorgan has implemented machine learning to help predict when companies may need to raise capital through a secondary offering. Again, we see similar pitfalls to relying on ML for such services: can the intelligent algorithms avoid crowding into similar strategies?
Clients themselves are also increasingly utilizing ML in their investment strategies. According to Eurekahedge, an index of hedge funds using AI saw an 8.44% annualized return since 2010, compared to more traditional quant fund indexes (1.62% and 2.62%) and a broad hedge fund index (4.27%). So, brokers that can service clients with AI capabilities should be poised to do well in an era of ML.
Who Will Benefit from Machine Learning?
We have only touched on two segments that ML has impacted. There have already been applications of ML in risk and credit management, fraud detection, security, and call centers. Even financial regulatory agencies are considering using AI to enforce their respective rules.
Given all this, who will be the winners from this technology in five years? Intuitively, it seems the large players would win since they have access to big data and the capacity to invest in the right people and resources. This makes sense, given the high costs of acquiring and storing large volumes of data and the high demand for scarce talent like data scientists. In 2012, some data scientists out of Harvard earned starting salaries of $300,000; we can only imagine that amount has increased significantly since then.
That said, we have seen large companies get complacent and become victims of their own success. The internet age demonstrated that—see Netflix vs. Blockbuster, or Amazon vs. brick-and-mortar retail. Innovation does not run in the blood of every manager. As such, we don't take it as a given that all of today's large financial firms will thrive in the next five years.
According to a PwC survey last year, two-thirds of US financial firms who responded said they are limited from utilizing ML due to resources, regulations, and operations. With these structural barriers to innovation in place, smaller and more nimble players might exploit ML to a greater extent than their larger peers and thus create a new landscape for the financial services industry.
Interested in reading more RSQ market insights?
Visit our blog library
or download Our Latest market commentary.
As international equity investors, the team at R Squared Capital Management (former team at Julius Baer / Artio Global) utilizes fundamental and macro analysis in our quest to correctly identify structural tailwinds and headwinds at the geographic, sector and company levels.
FROM THE DESK OF DAEIL CHA
Daeil Cha is a Partner and Analyst at R Squared Capital Management.
Prior to joining R Squared, Daeil was an Analyst at Suffolk Capital Management.
Daeil received an MBA from Columbia University and a Bachelor of Arts in Psychology, with a focus on Neuroscience, from Princeton University.
To view other RSQ team member bios, click here.