Commentary by Dr. James McCann, Chief Scientist, Plotinus Asset Management
No television business report would be complete without showing graphs, jiggling like a Debussy prelude across the screen, underpinned with a chyron trailing coded prices at the speed of light.
This appeals to our visual processing and we can quickly tell when a stock has been losing favor by the red line jaggedly oscillating but steadily trending downward.
Testing investment models using old historical data is fine, but that was then, and this is now.
Technical analysis foregoes a deep-dive into the complex cash flows of a company in favor of following a bouncing ball across the screen. The attractions are obvious, the information needed is easily acquired and understood. No need to read through the company reports and peruse the last six shareholder meetings: all that information has been priced in by teams of diligent analysts, at least in theory. Trend following takes the view that the stock price possesses a momentum (and thus an average rate-of-change) underlying the randomness. Indeed, momentum describes the inertia of the stock price, as if it followed Newton’s first law of mechanical motion. Financial momentum can arise as buy (or sell) orders are drip fed into the market. So even as the price increases, the buy orders continue to flow. The stock motion is driven by lack of liquidity in this case. A close cousin of trend following is ‘reversion to the mean’ in which a fictive force pulls the price towards a “target level.”
Trend following as an investment strategy might seem an affront to those focused on commercial fundamentals and those wedded to the efficiency of the market. It seems purely focused on signal processing, rather than economics. However, in various modifications, these methods have proved their value where it counts: in profits! The methods have been refined to the extent that those ubiquitous online charts now come with a smorgasbord of exponentially-weighted moving averages. Implicitly, the extrapolation into the future supposes correlation with the past. That correlation very much depends on what is meant by short-, medium- and long-term. It can be seconds or months. The concept applies just as well to low-latency trading as a sleepy mutual fund. All one requires is partial autocorrelation wherever and whenever it may be found.
The standard investment disclaimer that ‘future returns cannot be predicted based on historical returns’ would seem to say that there is not even a trustworthy trend in the returns of funds year-to-year! If only someone would do a study to see whether this was true. Enter James Choi and Kevin Zhao from Yale (2020). They found that star managers seemed to lose their star dust rather quickly. And this is recent “trend.” So, the old methods that worked last year are not working this year. One of the reasons suggested was that stocks have lost their mojo and that trend following was falling out of fashion. Our analysis supports this idea: Testing investment models using old historical data is fine, but that was then, and this is now. Markets are always moving and changing, and nothing stays still. Star managers know when to change the menu and the recipe, whether they be analysts or technicians. Otherwise, the active manager will be supplanted by her peers, or worse by a passive fund in the affection of investors!
The title quotation is usually attributed to that giant of 20th century science, the Danish Physicist Neils Bohr. But some prefer to credit it to “Yogi” Berra, the peerless catcher for the New York Yankees. Whoever said it was a wise person. ■
February 2022
© 2022 Plotinus Asset Management. All rights reserved.
Unauthorized use and/or duplication of any material on this site without written permission is prohibited.
Image Credit: Valdore at Can Stock Photo.
