Short-termism and momentum investing

I started my career as a value investor but as people who know me for a long time (or who have followed these missives for a long time) I have become somewhat of a ‘lapsed value investor’. I still appreciate value investing but I have learned to love momentum over the years, something that so many traditional value investors refuse to do but that has helped me significantly improve my performance over the last decade. 

Momentum investing seems like the antithesis of value investing since value investing usually requires a lot of patience and a long investment horizon to exploit trend reversals while momentum investing tried to exploit the most recent trend for short-term gains. No wonder then that so many professional value investors also complain about how their clients and other investors have become so short-term oriented and impatient. There is no doubt that the rise of electronic trading and high-frequency trading (HFT) has led to a proliferation of short-term investment strategies, but it could be that this rise in short-term trading strategies has also led to the demise of value investing. It may be more than a coincidence that value investing became less successful after the financial crisis just when HFT increased its influence on market trading.

Of course, the fact that interest rates have been close to zero since the financial crisis is in my view still the most important driver for declining value investing returns, but an experiment by a group of researchers from Paris and Moscow showed that increased short-termism may have contributed to the decline of value investing as well. 

In their experiment, they created a series of bots that were equipped with machine learning algorithms to simulate real-life traders and investors. Some of these bots were chart technical investors following only price information, while others were fundamental investors following the valuation of stocks in a market. Both kinds of ‘investors’ learned from past actions through reinforcement learning algorithms (the standard learning mechanism in AI) and started with one strategy and but adjusted their investment strategy based on the profits and losses they made in the experimental market. Over time, they did more of the things that made them money in the past and less of the things that created losses.

The twist in this experiment was that all the bots also had a psychological make-up. In one of three experiments, some of the bots participating in the market were programmed to be more short-term oriented and overweight quick profits relative to long-term gains. What happened as a larger and larger share of market participants were programmed to be short-term oriented is instructive. It comes as no surprise that as a larger and larger share of bots was short-term oriented, transaction volumes in the market increased and bis ask spreads for stocks declined. The market simply became more liquid. 

Trading volume in experiment if short-term traders increase from 0% (P=0) to 100% (P=100)

Source: Lussange et al. (2019)

However, something else happened as well as more investors became more short-term oriented: Trends (both positive and negative) started to last longer. If more investors are short-term-oriented, they are looking for quick gains. And the best way to make a quick profit in markets is to bet that what has gone up will continue to go up. But if more investors act like that, this expectation becomes a self-fulfilling prophecy because more short-term investors buy stocks that have gone up and shun stocks that have gone down. The result is that trends last longer and momentum investing and trend-following becomes more profitable. And of course, since bots do more of what has made them money and less of what has lost them money, value investors slowly exit, and momentum investing starts to dominate the market. Over the last couple of years, more and more prominent value investors have closed their funds or retired. That is often seen as a sign that value investing must be due for a comeback, but if these experimental markets are a guide to real-life events, this might not be the case. Instead, value investing may be due for a permanent decline as markets become ever more short-term oriented.

Length of positive and negative trends in markets if short-term traders increase from 0% (p_0) to 100% (p_100)

Source: Lussange et al. (2019)