Terminating Analysts: The Rise of the Machines
by Mark Artherton – Founder & CEO of LR Investment Services
Posted on Smartkarma as an independent insight on 3rd November 2017
Read more of Mark’s work by clicking here!
Independent Research was meant to be the future, MIFID II, scaling problems and tech trends may mean that it is the past.
“I need your clothes, your boots and your calculator”
With apologies to Terminator 2, 1991
There are many, many smart people trying to forge their own path in Independent Investment Research. For the right individual, it will always be possible to forge a lucrative career, as it is possible in journalism or book writing. As the institutional research pie shrinks the Investment Banks (IB) will fight tooth and nail with their exceptional resources to retain market share. If research can generate significant profits, then the IBs will be there. How that translates into individual pay is a different matter.
Many professional service industries are seeing the impact of AI, Machine learning, big data, etc. on their entry-level positions. The audit industry, consultants, investment banking and many others are implementing technology for ‘grunt’ level tasks. The asset management industry and the independent investment research industry will be no different.
I will touch upon two tech trends that are disrupting investment research, the first is robot writers and the second is robot primary research. There are many other technology disruptors in this space, but these are the most important in our opinion.
Robot Writers – they are here already
The world of journalism is already populated by Robots writing copy. Examples are The Washington Post’s use of Heliograf, and USA Today and Associated Press’ use of Wordsmith. Heliograf is described in a recent Wired article;
Editors create narrative templates for the stories, including key phrases that account for a variety of potential outcomes (from “Republicans retained control of the House” to “Democrats regained control of the House”), and then they hook Heliograf up to any source of structured data—in the case of the election, the data clearinghouse VoteSmart.org. The Heliograf software identifies the relevant data, matches it with the corresponding phrases in the template, merges them, and then publishes different versions across different platforms. The system can also alert reporters via Slack of any anomalies it finds in the data—for instance, wider margins than predicted—so they can investigate. “It’s just one more way to get a tip”
The purpose of Heliograf is primarily to target many different small audiences with many stories that are niche or local to grow the overall audience. The second purpose is to improve newsroom efficiency. Only large organisations can effectively lever technology to achieve this. The journalism industry is full of freelance writers but their focus has shifted to counter the rise of the bots. The Washington Post is not going to talk up the demise of traditional journalism, and the technology definitely aids those journalists with huge experience. However, it does create a block for junior reporters, those individuals who historically did most of the grunt work in an organisation. As the technology progresses it will impinge more and more on other journalistic activities. Hopefully, this will lead to a come back for investigative journalism, for example, as the technology led journalism cross subsidises quality journalism. Unfortunately, this may not be the case as the profit imperative will take the lead.
Writing a small piece on local elections or paraphrasing an earnings release is quite different to a robot writing a full in-depth report on a stock investment, but so-called ‘maintenance’ research can be automated relatively easily given the current state of technology. Junior analysts are in the firing line. This research is lowly valued but required. Time-constrained portfolio managers will always need briefings, they may give limited value to them but it does work to retain an audience. Automated report writing based upon conference calls, quarterly releases, and quarterly presentations are not that far away – a business doing this on a global level could carve an attractive niche.
Robot Researchers
The large Investment Banks are already experimenting with Robot writers to help reduce costs. More importantly, they are investing in systems such as Kensho, which provides traders with forecasts based upon its huge database (Kensho is used extensively by Goldman Sachs, for example – see below). As a standalone Independent Research Analyst, this should cause huge concern. MIFID II will push the IBs to implement research technology at an even more rapid pace, and their balance sheets and resources give them a huge edge.
Kensho – A step towards the future
Kensho is disrupting the data market and appears to work in a similar fashion to Wolfram Alpha. This is a direct attack on junior analysts and platforms such as Bloomberg, in my opinion. In fact, it is feasible to build a Bloomberg Killer today with existing technology platforms – though Bloomberg’s existing network effects and path dependency will delay its demise.
Kensho works as follows:
Questions (asked in plain English) can be typed into a simple, Google-style text box. Stuff like: Which cement stocks go up the most when a Category 3 hurricane hits Florida? (The biggest winner? Texas Industries.) Same with which Apple supplier’s share price goes up the most when the company releases a new iPad? (OmniVision, which makes the sensors in the iPad camera.) Until now, answering these types of questions required several analysts and several days. Kensho can do it in a matter of minutes.
Kensho was dreamt up outside of the IBs but the IBs were quick to recognise its use and invest. Initially, investing for traders, the shift into investment in research capabilities will give the IBs a tremendous edge.
Companies on the Smartkarma platform are already addressing some of these issues. Amareos, for example, analyses a huge amount of news articles to generate sentiment and other factors giving insight into asset behaviour. Despite some advances, I believe we are still scratching the surface in the application of technology to investment research and active investment management. Current solutions are not holistic and many fail to account for different consumption patterns in their clients. Retail investors may be hunting for the killer new process that they can follow blindly and make a fortune, professional investors are different, taking a much more evolutionary approach to changes in their process. Institutional Investors have fixed processes, and inputs need to be tailored to those processes. The technology of research delivery is lacking in this space.
The industry will continue to support human research analysts, though. There will always be space for synthesis ( I keep coming back to this concept). A combination of tech tools, internal workflow management and an in-depth understanding of client needs still needs synthesis. It is synthesis that computers will find difficult to replicate. Synthesis is the fallback position for human researchers and active investment management.
Additionally, systems may not be suited to the identifiaction of so-called ‘Black Swans’, the human mindset does throw up the occassional analyst who sheds a completely new light on a problem – Taleb, for example. As such, the role for left field thinking is likely to be occupied by Human Researchers.
The ability to service the career buy-side analyst who is looking to refine their model of warranty expenses at Samsung Electronics (5% of admin expenses), through the thematic manager who is interested in the possible future demand for 3D NAND flash, including the corporate fixed-income manager interested in the impact of future dividend policy on debt ratings, up to the global asset allocator interested in Samsung’s impact on the MSCI index performance and many others along the way is the future of investment research.
Celebrity may also have its place, the consumers of research will always read celebrity writers, though (maybe similar to the movie industry) the draw of the star analyst may be on the wane as interconnectedness reduces the luster of celebrity. Extending this further, while lower-end interaction can be handled by chatbots already, higher-end interaction will still require the human touch.
Conclusions
A combination of Kensho and Heliograph is a stealth killer for the bulk of the analyst community. The much-maligned Investment Banks have a clear edge in the area of technology and already have the client relationships. MIFID II has pushed the IBs to innovate more quickly, and I fear that Independent Research is an endangered species unless we react as a community. We need to get ahead of this wave with more collaboration.
Don’t get me wrong, the current state of AI means that we are decades away from true AI, where a human can be replaced in all ways by a machine. However, entry-level roles will cease to exist as technology does the ‘grunt’ work much more efficiently. The absence of entry-level positions will create huge problems in the future, but we are yet to see a drop off in entry-level hiring in a meaningful way. The other area that the machines will fundamentally change is the delivery and consumption of research. A revolution is underway, if you have got experience chose your path carefully.
Terminating Analysts: The Rise of the Machines
by Mark Artherton – Founder & CEO of LR Investment Services
Posted on Smartkarma as an independent insight on 3rd November 2017
Read more of Mark’s work by clicking here!
