Rise of the Machines, Part 1: How Increased Automation in Financial Services Impacts Society
Recently, I read some articles in The Economist newspaper that talked about the increasing use of computers in asset management. While the topic is not new to those of us in the industry, I thought of our recent intern, John Simmons, who is graduating from Xavier University next month and is focused on becoming a portfolio manager. I wondered what skills does John need today to be successful in twenty years. It dawned on me that unlike college graduates of prior generations, who only needed a solid grounding in finance along with basic computer skills, John was probably going to work with machines in ways that are unknown to many of today’s portfolio managers.
I also thought about my own career and how it led me to create Joot, which seeks to take today’s cutting-edge technology to simplify and automate the increasingly complex compliance requirements of an asset manager.
After a brief discussion, John and I decided to partner on this article: Rise of the machines. Part 1 provides an overview of how pervasive technology has become in the asset management industry. Part 2 discusses the impact of this technology on portfolio management. Part 3 discusses why technology like big data, robotic process automation, and machine learning are going to change compliance forever. We hope you enjoy these thought-provoking pieces and will leave comments that will advance the discussion of how machines and humans will co-exist in the asset manager of the future. If you’d like to post a response to any of these articles, please contact me.
The Economist’s recent series of articles on how computers are increasingly relied upon in financial markets, including asset management. The first article, Masters of the universe, noted that computers now account for “35% of America’s stock market, 60% of institutional equity assets and 60% of trading activity.” A second article, March of the machines, Market, noted that on September 13, 2019, “Morningstar, a research firm, reported that in [August 2019], for the first time, the pot of passive equity assets its measures at $4.3trn, exceed that run by humans.”
For anyone focused on the current asset management industry, participants are increasingly focusing on digital enhancements because technology is scalable and can handle larger amounts of data and transactions better than humans. The scalability factors mean firms can do more with less people, which (potentially) lowers costs. When it comes to data, the same machine can work 24/7 to crunch vast amounts of information; whereas the human brain is more limited in how much data it can effectively ingest (not to mention a human brain takes decades to mature and years to be trained) and people usually need to take breaks to sleep, eat, and do something other than work.
The timeline for the digitization of financial services started a few decades ago when electronic trading displaced the old trading pits on the floors of exchanges. Today, almost all trading is done electronically. In March of the Machines, The Economist cites statistics from Deutsche Bank that says “90% of equity-futures trades and 80% of cash-equity trades are executed by algorithms without any human input.”
Next, computers were harnessed by portfolio managers to manage passive and then active strategies in vehicles such as ETFs and mutual funds. Now, those same computers, colloquially called robo-advisers, are displacing portfolio managers. Regulators have even jumped on board to use computers and algorithms to guide examinations and litigation.
The Economist articles note that while today’s computers require direction from humans, who built the models and mathematical formulas underlying these tools, advances in artificial intelligence are driving us towards a world where quantitative strategies operate under semi- or fully autonomy.
There are three levels of artificial intelligence: automated intelligence, augmented intelligence, and autonomous intelligence. Automated intelligence refers to robotic process automation, the digitizing of the mundane and boring administrative tasks. Augmented intelligence exists where most of us live -- it finds the use of robots to make humans better (e.g., chatbots, smart data, etc.). Autonomous intelligence is Skynet and the end of the world, maybe.
Underlying these technological advances are improvements in processing power and the Internet-of-Things (IoT) are fueling the growth in artificial intelligence. While it’s clear that financial services will be increasingly automated and some human-driven processes will be replaced, there are many questions around whether this trend is good for society and what role will people play in the new asset management firm.
Why digitization of finance is both good for investors and disruptive to society
In Masters of the Universe, The Economist notes that these technology advancements have lowered transaction costs (e.g., the zero-fee ETFs) for investors; technology has also provided investors with more options for managing their assets. Unfortunately, these benefits come at a price. For one, only the largest firms can afford the advanced computers, data sets, and technicians needed to constantly improve the computer models.
As a result, the financial services sector is becoming increasingly concentrated in a few firms such as Vanguard, Blackrock, Fidelity, and others. Smaller firms are being forced to abandon their own trading strategies in favor of asset allocation models designed by the asset management industry’s behemoths.
Second, as The Economist notes, there are growing concerns over financial stability as many computer models tend to trade in similar ways and can instantly trade large positions, which can suck liquidity right out of the market. In some ways, having managers with different trade horizons and strategies diversified the market to a point where catastrophic declines were nearly impossible in the short-term (although longer term economic weakness could lead to recessions and depressions). But today’s convergence around certain core strategies increases the risk of contagion in the event of a financial crisis.
Another area of concern cited by The Economist is corporate governance. Most computer models are designed to track specific economic and financial data. The models are not designed to evaluate management or the board of directors at companies. As more assets are managed by computers and their masters avoid proxy voting (or delegate to a few proxy voting vendors) there is less oversight over the people managing companies.
But technological disruption is often viewed as progress. For example, artificial intelligence is trying to eliminate human error and create returns that are satisfactory to millennial investors. Investing with robo-advisors possess a number of ramifications on this generation of investors. The industry has already seen technology implemented to speed up processes and procedures by building algorithms and code to become more efficient. In the past, the typical adviser, a human, has always overseen investment decisions, but artificial intelligence and robo-advisors are reducing the level of human oversight. Even with our current boom in technology, people still believe in relationships. And this is where the humans still excel (despite the increased use of social media and text.)
For the baby boomer generation, the advisory relationship was built upon a familiar relationship. Portfolio managers met with their clients, listened to their stories and goals, and then created an investment plan. The portfolio manager would then manage that plan while retaining responsibility for the client relationship. The emergence of artificial technology will diminish the need for relationships and trust. Thus, investors (the millennial generation) need to set expectations for their portfolios that are being managed by artificial intelligence.
The Dangers of Technology
Despite its plethora of benefits, technology also creates new dangers, and change is worrisome to many. Whether you are a small asset manager or an individual investor, the change can create uncertainty. The change will require companies large and small to adapt to a better understanding of artificial intelligence. In the case of large corporations, training and risk management will be crucial because of the cross-functionality that will be needed between compliance teams and coders. Organizations with thousands of employees will need to restructure and create a dynamic action plan to mitigate the inevitable errors that arise when implementing new products. Similarly, small companies will have to find funding from outside investors and venture capital firms who believe in the concept of robo-advisors and artificial intelligence. In both cases, implementing this technology is a large undertaking on two different spectrums.
Technology also leads to more complex integration between systems, data sets, and the vendors that manage them. The piecemeal approach of technology development means that security and operational risks may go unseen as a multitude of vendors focus only on their product, which by itself is safe but may weaken when combined with other technology that is not managed by the vendor. Unfortunately, we see this scenario play out daily when one a breach of one system allows nefarious actors to gain access to another system and the valuable data that it holds.
Finally, another worry is the concept surrounding big data. By implementing artificial intelligence, humans are turning over processing and capabilities to algorithms and codes that come up with decisions based on large amounts of data. What were to happen if there was a data breach that created fraudulent data? Using artificial intelligence that focuses on big data possess these risks.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.