The Future of Research
by Mark Artherton – Founder & CEO of LR Investment Services
Posted on Smartkarma as an independent insight on 19th October 2017
Read more of Mark’s work by clicking here!
I recently attended the opening event for the Smartkarma UK office in Soho, London. The event was well attended and there was plenty of lively debate around the impact of technology on investment research.
The first panel discussion focused on the subscription economy. The discussion on this panel moved onto the similarities (or lack of) between investment research and other things that are now moving to be sold under a subscription model. The moderator likened car sharing to investment research. Ownership of cars will be a thing of the past and there will be a car waiting at the end of the road for the time you wish to use it. The implication of such a model is that instead of a portfolio management firm ‘owning’ investment bank waterfront coverage at a very high price, that research would be better served by a more focused subscription model where the users of research only used the research that they required at the time they required it. No need for that expensive SUV to be sat on the driveway or in the garage when the slack can be used by others whilst reducing your costs.
It was a well-received analogy, despite the differences between information and physical goods. The analogy triggered some interesting discussion amongst the panel and I would encourage readers to view the panel discussion when it is posted on the Smartkarma website. The downside of such a move for the producers of research is that the size of the pie is that much smaller. Some panelists felt that the amount spent on investment research could halve. It was stated that most the impact should be felt by the investment banks rather than independent research providers and there would be a huge market share gain from independent research. Independent Research would be in a stronger position to utilise the rise of new models to meet the subscription requirements of the consumers of research.
I would like to move this analogy further along and to sketch out the way I believe that the investment research ecosystem should evolve. This is an opinion piece and the existence of extreme legacy issues and path dependency make it likely that my vision of the future will not necessarily be the one that comes to pass and I will discuss those issues at the end of the piece.
How do you define investment research?
Investment Research can be broken down into three parts. The first part is obtaining knowledge or information. Specifically, knowledge that is available to all, not inside information. This knowledge can be disseminated widely using new technologies and is accessible to all at all times. This gives pure knowledge little intrinsic value beyond its initial cost of production. This is one of reasons why Wikipedia is free.
The second part is the filtering of that knowledge. The first stage of filtering is removing ‘fake’ knowledge. Fake news is a hot topic at the moment and likely to continue to be so. In an unfiltered, unedited internet base world the ability for anyone to pass anything off as knowledge increases dramatically. Google’s search algorithms attempt to provide some comfort for the user, by (and I am massively oversimplifying here) using other sites link to a particular source as a way to measure its veracity and validity.
The second stage of filtering is reducing the knowledge down to that which is the most pertinent to an individual’s investment process and decision-making needs. Filtering the relevant from the irrelevant in a way that fits a decision-making process. In a world where all (non-inside) knowledge is potentially available to all this filtering system is exceptionally important. This filtering has tremendous value.
The final part is the synthesis of this filtered and relevant knowledge. The synthesis has two major applications for fundamental portfolio managers. The first is its application to buying a stock in the portfolio or selling a stock from the portfolio. The second application of the synthesis is for portfolio construction, how to arrange the individual building blocks of a given portfolio. Synthesis, when combined with a circumstance based approach, has tremendous value.
Traditional investment researchers carry out all three of these functions in ways that vary in their effectiveness and use, but each will have their own specific method of filtering and synthesis which may or may not need an individual client’s needs.
To summarise, investment research is the following:
- Obtaining knowledge;
- Filtering this knowledge for quality and relevance;
- Synthesising the filtered knowledge to impact buying and selling instruments and the portfolio construction of those instruments.
The Investment Research car
Let’s now return to the analogy of the shared car. Investment research is so much more than sharing a ready-made car. Ideally, investment research is having a bespoke vehicle built to your current specification at that moment in time to achieve the complete goals of a particular journey. From choosing the type of fuel, the individual parts in the engine, the seating layout, the technology available within the vehicle, the type of terrain the vehicle will have to traverse, and the final destination. All of this needs to be carried out in a way that understands the portfolio manager’s immediate needs without the portfolio manager necessarily having to be explicit about those needs.
Imagine a scenario where a portfolio manager has at his fingertips all of the information that is relevant and synthesised to meet her individual decision-making process allowing her to make an effective decision as to Sell something from the portfolio, buy something for the portfolio, alter the portfolio construction or just to leave everything alone at the correct point in time. This information is supplied in a way that it can be digested rapidly. Quantitative and algorithmic funds are already in this scenario because numbers are easy (easier) to put into such a system. If we accept the premise that active human-led portfolio management can add value then a system as described above would add tremendous value (though some of this value is likely to be short-term and ephemeral – as you might pay hundreds of ponds to eat in a high-end restaurant). The system as described does appear to be some form of unattainable nirvana given the path dependency built into the system.
However, it is my belief that the technology tools are there to achieve this aim. From data collection all the way up to the basic level of Artificial Intelligence that has been achieved(human replicating AI is a lot further out than the futurologists would have us believe in my opinion). Human input is required at all levels and it is this human technology interaction that is very important. What is missing is the correct application of these tools – Smartkarma is a step in the right direction.
Is history holding us back?
Historical activities and path dependency often slow the evolution of systems. The vast bulk of investment research has barely changed in it format over the past 50/60 years – since Ben Graham’s seminal book ‘Security Analysis’ and the creation of the Investment Research Report as the demand from institutional investors built in the 1950s. There have been many tweaks and improvements in education and even quality, but the production model has not changed too much. Delivery was impacted by email, the internet, and smartphones. Though most delivery models are just online libraries. Consumption itself has also altered little over time.
The historical model has not lent itself well to collaboration. At best the lead analyst had a team of ‘grunts’ to do the legwork, but true collaboration across disciplines was very hard to find.
What is required is a 360-degree restructuring of the whole active management industry and the way research is produced, distributed and consumed. Cherry picking models and technology from other industries is key to this. Collaboration is also very important in this approach, getting the best to work together to give an outcome that is greater than the sum of the parts. The success of such an approach would become clear very quickly so a shift from early adopters to the mainstream could be very rapid for the company that delivers such a system.
Creative Destruction
MIFID II has provided a catalyst for buy-side firms to think about their consumption of investment research and it has also forced all the providers of investment research to evaluate their distribution models. The sell-side (investment banks) remain formidable opponents for independent research. Prior to MIFID II, inbuilt biases in the system favoured investment bank (IB) research. Some independent research providers (IRPs) were able to build businesses but true scale was rarely reached. Some IBs may exit research but most will alter their delivery of research to benefit from the changing marketplace. The bulge bracket investment banks have a lot of advantages on their side – deep pockets, extensive client relationships, and technology expertise to name but a few.
Does the buy-side need research that is produced outside of its own walls? Large buy side houses may not need it at all in the future. With more effective internal communication, the larger houses will be pressured to use the resources and the knowledge within their own walls. If the payments for research out of p&l migrates globally then justifying external research becomes harder and harder, especially when process fit is considered. The day cannot be too far away when a large buy-side house cuts off all external research. The trend of M&A to create larger and larger fund houses accelerates this possibility. A key advantage that larger houses have is the relationships with the companies. An investor relations professional can only deal with a limited number of market touch points. Large shareholders will flex their muscles and retain access. This approach of self-sufficiency does make seem to make some sense for the larger fund houses. There is a global trend to bring more capabilities in house for the larger companies. Especially for those institutions that deal in information.
There are some shortcomings to this model. Would a large house pay for a shipping analyst, for example? Someone whose knowledge is only relevant in short bursts every few years. This is also a problem for the independent research analyst who covers this sector. Large houses could avoid this sector completely. Small sectors, such as shipping, can be (more or less safely) ignored – in markets like Korea where they can move the index dial maybe this is not the case. A further option is to hire or develop specialists with complementary skills in these markets so they remain relevant through the cycle. The correct use of technology will enable individuals to cover more and more in the future, in my opinion.
For medium and small size fund houses the issue revolves around access to expertise (in this case I am not explicitly talking about expert networks as inside information issues cause problems for many). Medium and small houses need access to external research. Prior to MIFID II this would almost force smaller houses to trade more to generate commission dollars to get research access. We are all aware of the premium tiers of service at IBs and the amounts to get such service. In effect, the buy side shared the sell side’s product. Thousands upon thousands of buy-side firms sharing tens, maybe a hundred, sell-side firms resources. Large houses generated the most commission, but medium and small houses got access to this research, which was, in effect, paid for by the larger houses.
As suppliers of investment research, there are many, many high-quality individuals looking at the same things. Some will have greater knowledge and understanding in some areas and some will have a better reading of the market for making buy calls and some will have a better reading to make sell calls. Some are better at making such calls in absolute terms and some are better making them relative to a country benchmark and yet others are better making them against regional or global benchmarks. Some will fit better with one client’s style whilst others will fit better with other styles.
Oddly enough, even though they may appear to be more challenged now, the mid and small sized investment houses could be forced into the correct solution. By providing a ‘virtual’ research department that alters its composition in line with the buy-side institutions needs, a solution can be found that shares resources effectively without diluting their value. A research report that can be read by anyone is viewed as having limited value. A bespoke, collaborative piece generated by AI and HI (Human Intelligence) has much more value.
Independent research needs to decide if they are media like – low fee high readership – like a high-end FT, or if they are bespoke consultancy providers. There is space for both models in the marketplace. There are also risks in both models, with the overhead for the media model being huge, for example. Bespoke consultancy is exceptionally difficult to scale, but advances in technology are bringing scalable bespoke solutions much closer.
Does it make sense for multiple people to read the latest monetary report from the Korean Central bank? Or does it make sense for one individual to do so and summarise in a specific database friendly way? Does it make sense for multiple people to listen to the earnings call of company A or read the presentation of company B or read the IPO prospectus of company C (to be fair I am pretty sure that most people don’t have the time to read these)? Or does it make the most sense for one person to do this and summarise for a database? There is so much duplication and inefficiency in investment research.
An additional factor to consider in such a model is the benefit it could have to consumers of wealth management products. Roboadvisors are both adding new clients but also eating into the lower end of the wealth management client base. This is not the place to discuss the pros and cons of roboadvisors, but the relationship model offered to more wealthy clients would be pressured by the changes we are seeing in research delivery. The simple delivery of other’s views will not survive this revolution in the delivery of investment thought. There will always be a place for individuals to ‘sell’ the initial benefits of a given system of research, but the management of that relationship will become less lucrative.
One thing is very clear, the future of investment research will be very different to its past.
Conclusions
Being the iTunes of investment research is not enough in an industry undergoing such significant change, not when it is clear that the more effective business models such as Spotify and others are eating at the iTunes model. Differentiated models that treat information as ingredients that are supplied effectively to the synthesisers is a possible way forward.
Differentiated models are needed that treat information as Lego building blocks to be put together in a variety of different ways, collaboratively, to achieve different ends. The days of the long form report written by an individual which is then stocked in a library, put on a website, emailed (or added to iTunes) are limited. The way research is produced, delivered, and consumed will be drastically different from the model we see today. Technology will provide each individual at each point of time, a differentiated and unique build of relevant, high-quality information in a structure which allows a decision maker to rapidly make investment decisions.
The Future of Research
by Mark Artherton – Founder & CEO of LR Investment Services
Posted on Smartkarma as an independent insight on 19th October 2017
Read more of Mark’s work by clicking here!
