It’s rare on Wall Street for a technical analyst’s core insight to be confirmed by a Nobel-prize-winning economist. The far more common situation is for academics to deny that technical analysis has any value whatsoever — and for technicians to return the favor.
Not with John Bollinger, the well-known technical analyst and creator of the technical analysis tool known as Bollinger Bands. Now is a good time to review his contributions to investment analysis, as in May Bollinger tweeted to subscribers that “After 33 years, 396 issues and roughly 1,700 hotlines I published my last advisory letter… You’ve been good traveling companions, thanks for your company along the way.”
My key takeaway from reviewing Bollinger’s contributions to investment analysis is the importance of being ruthless in the pursuit of the truth, wherever that pursuit takes you and regardless of your preconceived notions about what works and doesn’t work.
Bollinger Bands are lines that are drawn on a security’s price chart to reflect its recent volatility. The specifics of their calculation are beyond the scope of this column, but typically they are two standard deviations above and below the security’s 20-day moving average. (See accompanying chart showing Apple’s AAPL, -1.81% Bollinger Bands over the past year.) The spread between the two Bollinger Bands will be relatively narrow when recent volatility has been low (as it was for Apple last September), while much wider when the security has been highly volatile (as the stock was in December).
Bollinger’s core insight was that a security’s recent volatility tends to persist. If it’s low, for example, odds are good that its subsequent volatility will remain low for a while longer — just as you should expect subsequent volatility to be above-average if its recent volatility also has been high. It therefore is significant when a Bollinger Band is violated, and violations can be used as triggers for any of a number of trading strategies.
Prior to Bollinger coming up with this insight in the early 1980s, traders had assumed that a security’s volatility profile hardly ever changed. As Bollinger explains on his website: “At that time volatility was thought to be a static quantity, a property of a security, and that if it changed at all, it did so only in a very long-term sense, over the life of a company for example. Today we know the volatility is a dynamic quantity, indeed very dynamic.”
The Nobel Prize-winning economist whose work confirms Bollinger’s core insight is Robert Engle of New York University. He was awarded the prize in 2003 (along with the late Clive Granger) for developing a statistical understanding of why and how a security’s volatility changes over time. As Engle’s discovery is described on Wikipedia: “Previous researchers had either assumed constant volatility or had used simple devices to approximate it. Engle developed new statistical models of volatility that captured the tendency of stock prices and other financial variables to move between high volatility and low volatility periods.”
You may have seen references to Engle’s statistical model and not even known it. The model is formally known as the “generalized autoregressive conditional heteroskedasticity model,” a.k.a. GARCH, and sophisticated options traders and those who trade the CBOE’s Volatility Index VIX, +0.96% , for example, use GARCH and refer to it frequently. A simple application of GARCH is to increase your equity exposure whenever the market’s volatility over the trailing month is low, and vice versa, as I discussed in a mid-May column.
Read: How to protect your money from this stock market’s wild volatility
What does it mean to be ruthlessly empirical in your analysis of trading strategies? One helpful book I can recommend in this regard is “Evidence-Based Technical Analysis,” by David Aronson. Aronson used to teach a graduate level course at Baruch College’s Zicklin School of Business; he currently is President of Hood River Associates, a firm that employs sophisticated modeling to “enhance the profitability of quantitative investing strategies,” and the author (with Dr. Timothy Masters) of “Statistically Sound Machine Learning for Algorithmic Trading of Financial Instruments.”
Another good discussion of this topic comes from Campbell Harvey, a finance professor at Duke University’s Fuqua School of Business. In his presidential address to the American Finance Association several years ago, Harvey described how to guard against the tricks statisticians can play to make a strategy look significant when it in fact is not.
So be forewarned: The vast majority of trading strategies employed on Wall Street have not beaten the market over time, and have no claim on your hard-earned assets.
Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. His firm has not tracked Bollinger’s newsletter. Hulbert can be reached at mark@hulbertratings.com
Read: What the Dow’s breaking its 200-day moving average really means for your stocks
More: U.S. stock market forgoes $5 trillion in returns thanks to trade war, estimates Deutsche Bank