Subscribe to Moneycontrol Pro and get 365 bonus InterMiles! Use Code: INTERMILES
Last Updated : Nov 04, 2019 08:41 AM IST | Source: Moneycontrol.com

Right mix of technology and human touch needed for optimal investment decisions

In the immediate future, it seems more likely that computers and humans will work side-by-side

Rajeev Thakkar

You would have come across terms such as Artificial Intelligence, Machine Learning, Quantitative (or Quant) Finance, High Frequency Trading (HFT), Robo Advisory and so on.

So, what is the role of computers in finance today and in the foreseeable future? There are some who are true believers in the change and think that computers will take over investing while there are others who think that all this is hype and will soon fade away.

For a very long time now, people have been trying to use mathematics, statistics and probability, and computing power while investing in stocks, bonds, commodities, gambling and so on, to get rich. These efforts have met with varying degrees of success.


Financial and computing skills

Warren Buffett & Charlie Munger do not use computers for investing. An aspiring investor of today does not have to emulate this aspect. Buffett used Valueline’s printed books to go through the financials of listed companies. An investor today is more likely to use a computer-based screener.

Thus far, any person wanting to take up investing as a profession had to primarily study finance and the computing skills could be acquired on the job. Of late, however, this has been changing. While most of the recruits still are finance professionals, increasingly, many computer science professionals are entering the investing/trading field with or without formal training in finance.

Quant / Computers and scale

An analyst may be highly skilled and may have a great framework. However a human has only so many working hours, is relatively slow, and is prone to making mistakes in repetitive tasks. A computer has no such limitations. Let' says an analyst tracking the retail sector wants to read about discussions on Amazon in all the conference calls held by all the listed companies in a quarter. In the earlier days, such an analyst would need to hear the recordings of all the conference calls. These days, with Natural Language Processing, such a need would be served within a matter of seconds. Computers help existing investment professionals scale up their operations.

Quant / Computers and blow ups

While discussing Quants or Computers in finance, a lot of people think of Long Term Capital Management, the hedge fund run by Nobel laureates that spectacularly blew up in the late 90s. This assessment may be a bit unfair. Blow-ups happen because of leverage and lack of staying power, rather than the use of computers or lack thereof.

Another criticism relates to the flash crashes and the “fat fingers” of computers. One has to understand the concept of Garbage In, Garbage Out. If the input or the instruction fed to the computer is wrong, the output or the trade recommended by the computer will be wrong.

Quant / Computers and alternative data

Readers may have heard about the use of satellite images of parking lots to estimate the number of cars parked and hence predict the sales numbers for retailers. Similarly, there are people who look at numbers of searches on Google, social media posts for brands, Foursquare check-ins and so on to predict sales, profits etc. A lot of this would not have been possible without the use of technology. Such data is called alternative data.

Quant / Computers and repeatability

Increasing use of computers forces some amount of objectivity in the investing process. It no longer enough to say that we invest in companies with “high quality management.” One will have to breakdown the term “high quality management,” into various sub-segments such as governance, frugality and honesty. Even this is not enough, as one will need objectivity on the sub-segments. One could have a scoring system, say, for frugality. One could compare Selling and General Expenses as a percentage of sales for the company vis-à-vis its peers and score the company objectively rather than rely on heuristics such as “I once saw the company’s MD take the auto-rickshaw.” This makes the research approach process-driven rather than being decided based on anecdotes, heuristics and gut feel.

Quant / Computers and changing rules

While all the points mentioned earlier do point towards an increasing role of computers in finance and investing, some caution is required. There is an important difference between computers playing Chess or the game GO vs. investing. The main difference is that the rules in the former are more or less static. The pawn moves and the queen moves in chess will be the same across various games. However, in investing, the rules themselves keep changing quite frequently. Interest rates moving from positive to negative is one example. The attractiveness of low price-to-book stocks held true in some era. However, with asset-light businesses such as Google, Facebook, MasterCard and Visa, that metric is not very valid in investing.

Hence, in the immediate future, it seems more likely that computers and humans will work side-by-side.
First Published on Nov 4, 2019 08:41 am