Midweek Musings – the computers say sell

A short break

During my first trip to the UK since the start of the Covid-19 pandemic, so over 2 years have elapsed, I made observations as one does about a grandchild or nephew that one has not seen for some time – noting changes that parents and those seeing the kids every day don’t really see as significant.

It was a short visit so I travelled across the South of England only and stayed in various B&B, pub and holiday lodge rooms. The main changes I noted were these, all I suppose consistent with ‘macro’ changes that have been predicted:

• Every single place I went to for a meal was short of and advertising for employees. I was almost exclusively served by young women in their late teens whom I would guess were students. Service was poor at several venues because the staff were new, temporary, or there were not enough of them. I checked into and out of one place without needing any employee involvement at all.
• The large number of new electric cars of all sorts and prices was very noticeable to me as a car geek.
• Prices of routine goods were lower than I expected. Inflation is rather more severe where I live and I guess the UK inflation numbers are driven mainly by energy and housing costs. Rents in the area where I used to live were unbelievably high.
• Very many small shops that were trading 12 years ago when I left my old home are still trading in smaller town high streets, so maybe the demise of shopkeepers has been overstated. There are more coffee shops and less electronic goods retailers, I thought.

These are observations based on a short trip and pretty subjective, so I am not claiming them as gospel! Just my personal perceptions.

The march of AI

This is the main topic of my musing and this time I really am not able to propose a concrete interpretation of an observed major change in the way the investment world works, only to speculate on the implications.

My introduction to the subject

A few years back I had a lucky opportunity to listen to the Nobel prize winning finance guru, Yale Professor Robert Schiller, lecturing at Vilnius University (his grandmother was Lithuanian). He is considered one of the fathers of the behaviourial theory of finance and he was talking about his latest research.

He explained that developments in the consumption and recording of words by the likes of Google has created all sorts of new analytical opportunities, some of which provide insights into the ‘herd’ thinking of the market that the behaviourists would argue drive market valuations. For example, one can pay to find out how often a word or sequence of words occurs in a Google search across a given period, or in all global published documents.

You might for example wonder if the world at large was interesting in investing in gold and you could as a result find it useful to see what the pattern of searches has been recently for the phrase ‘gold price today’. That information can be obtained in a very granular form. This is very similar to the idea of ‘trending’ on Twitter, views on YouTube and so on.

Such data can be requested by a machine and processed with other data by the same machine to create an analytical model which might even follow through into transactional activity – asset purchases and sales.

The latest update

I learned this last week that the annual statements made verbally by the CEOs and CFOs of the largest businesses in the world, like the member companies of the S&P500 index or the FTSE 100 are immediately transcribed by AI programs on presentation (before they are even issued as printed documents) and searched algorithimically for key words and phrases, outputting instant analyis to humans or to feed into tranasctional models.

This trend towards automated analysis adds a layer of investment dealing between the traditional human, research led, objectives driven managed style and the popular low cost passive style that tracks a composite index of shares or bonds built from a set of rules and a source universe like a stock market index eg a FTSE ALL Share tracker or a Vanguard Lifestrategy fund

You may have heard me say that algorithms are never intelligent, quoting a US lawyer. This is something that can be exploited. Professor Schiller explained that the investor relations departments of large companies get to know what words or phrases are looked for by the analyst algorithms and they then plant these to their advantage in press releases and other publically available material! No doubt CEO periodic reports can be doctored the same way.

Implications

I am inclined to think that we now have 3 choices of fund management style: human and intelligent, passive and input free and finally automated and potentially gullible.

Naturally, I prefer to invest in the best of the managed funds, although finding them for the future is as much an art as a science. However, it helps to know that very large sums of money are flowing in and out of markets using the other 2 models. My current view is that this is helpful, offering a permanent underpin to values and may well explain why any systemic setback for over a decade has been speedily reversed. In other words the stock market has become a permanent sellers’ market.

But I am not sure. My intuition says that automated processes can go wrong and when they do, the blow up can be dramatic. What do you think?