This column should squash, once and for all, any lingering hopes you may have that year-ahead forecasts are worth paying attention to.
I had thought that my arguments two weeks ago would have done the trick. I pointed out then that almost all forecasts are wrong, and not just by a little bit, either. And on the rare occasion when an analyst in a given December happens to make an accurate year-ahead forecast, he is no more likely to be right the subsequent December than other analysts whose forecasts the previous year were wrong.
Evidently, however, some of you continue to search for a clairvoyant genius on Wall Street whose forecasts are far more right than wrong. So let me now try a different argument: Even if, contrary to all evidence, such a genius existed, and you were fortunate enough to get his forecasts, you most likely still wouldn’t be able to use his insights to beat the market.
Consider research conducted by MIT accounting professor Chloe L. Xie. She analyzed the investment performance between 2011 and 2015 of a group of Ukrainian hackers who gained access to the servers of the major newswire firms that companies use to distribute their earnings report press releases (such as BusinessWire, PR Newswire (now part of Cision), and Marketwired (now part of GlobalNewswire)). The hackers therefore were able to find out in advance which companies would be reporting earnings that were better or worse than the analyst consensus. Over this several-year period they executed more than a thousand trades based on their advanced insight.
You’d think that these hackers would have made a killing, but you’d be wrong, according to Professor Xie. Sometimes their trades were very profitable, but not always. Overall, she found, they “performed poorly.”
Another study is perhaps even more damning. Entitled “Can executives predict how firm news affects stock price?,” the study was conducted by Darren Bernard, an accounting professor at the University of Washington; Elsa Juliani, a professor of accounting and control at INSEAD, the French business school; and Alistair Lawrence, an accounting professor at the London Business School.
The professors conducted a survey of more than 650 executives of U.S. publicly traded companies, asking them to predict how the price of their companies’ stock would react to soon-to-be-released earnings. The survey was conducted under nondisclosure agreements and assurances of anonymity.
Upon comparing the executives’ predictions to what actually happened to their companies’ stock prices, the professors found:
- “Executives incorrectly predict the direction of the stock return one-third of the time, and expectations are about six to 7 percentage points (i.e., 600 to 700 basis points) different from realized returns, on average.”
- “Executives’ expectations are more accurate than mechanical prediction models based on factors such as earnings and sales surprises about 50% of the time—no better than chance—and more accurate than a model predicting zero returns only 59% of the time.”
These results (summarized in the accompanying chart) are incredible, since the executives know a lot more about their companies than what is included in their PR-spun earnings news releases. And yet their predictions still were accurate only slightly more than what you’d expect on the basis of pure chance.
Many analysts’ forecasts of the upcoming year focus on macroeconomic developments rather than individual stocks. Perhaps analysts have a better track record when it comes to forecasting them?
Consider a study conducted a couple of years ago by Vincent Deluard, Director of Global Macro Strategy at StoneX Financial. He calculated the investment performance of a hypothetical investor who knew, in advance and with pinpoint accuracy, how much the GDP would grow in the future. He did this by assuming this investor would be fully invested in stocks if future GDP growth was higher than what it was in the then-current quarter, and otherwise 100% in cash.
Deluard found that this hypothetical advance knowledge was worth surprisingly little, if anything at all.
- In one simulation, Deluard assumed a hypothetical investor knew one quarter in advance what the final GDP growth rate would be for that quarter. That entails a lot of clairvoyance, of course, since the government’s initial estimate of a given quarter’s growth isn’t available until well after it’s over—and we have to wait even longer for the final number. Even so, according to Deluard’s calculations, this investor would have lagged a buy and hold by 1.0 annualized percentage points from 1948 through early 2018.
In a second simulation, Deluard assumed that the hypothetical investor had much greater clairvoyance, being able to predict what the final GDP number would be four quarters in advance. He found that this hypothetical investor would have performed only slightly better—beating the S&P 500
by less than 1 annualized percentage point. That is a surprisingly small reward for being a perfect forecaster. In fact, Deluard found, this investor would have done better simply using the 200-day moving average to switch between stocks and cash.
The bottom line? You should assume that year-end forecasts are worthless. Not only are they wrong far more often than not, you probably couldn’t beat the market even if those forecasts were right.
Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. He can be reached at email@example.com.