Why Market Forecasts Are So Bad; Analysts’ predictions are frequently less accurate than random guesses.

Why Market Forecasts Are So Bad

Analysts’ predictions are frequently less accurate than random guesses.

JOE LIGHT

Dec. 20, 2013 7:15 p.m. ET

The oracles known as Wall Street strategists have spoken: The S&P 500 will rise 6.4% by the end of next year. But the evidence shows you’d do just as well guessing yourself.It isn’t supposed to be that way. Strategists use complex models to forecast earnings growth, price/earnings ratios and other market-driving factors. In the 21st century, apparently, that effort has been for naught.

Take the predictions that strategists made in December 2012, attempting to forecast the S&P 500 this year.

Research firm Birinyi Associates collected such projections by 11 strategists from Wall Street’s largest firms, including Morgan Stanley MS +0.23% and Bank of America.BAC -0.95% On average, the analysts thought the S&P 500 would rise 8.2% in 2013. The S&P 500 has actually risen 27.5% this year through Friday, not including dividends—a difference of 19.3 percentage points.

To put that in perspective, let’s say you pinned a list of all the annual S&P 500 percentage changes since 1929 on a wall and blindly threw a dart at it.

More than half the time, the number you landed on would have come closer to 2013’s actual price change than the analyst consensus for the year, a Wall Street Journal analysis shows, based on data from Birinyi and historical returns.

Since 2000, analysts have also done worse than the dart thrower in 2001, 2002 and 2008. Winning 10 out of 14 times wouldn’t be so bad—were it not against such a mindless opponent.

Forecasting stock returns a year in advance is exceedingly difficult. Often, failing to foresee one big event, such as the Federal Reserve’s decision not to slow its bond-buying program in September, can mean the difference between being close or whiffing, says Tobias Levkovich, chief U.S. equity strategist at Citigroup. C +0.64%

Last December, Mr. Levkovich thought the S&P would close at 1615 this year. That made him the most accurate of strategists Birinyi tracks, but the estimate is still on track to be much too low. On Friday, the S&P 500 closed at 1818.32.

Mr. Levkovich says he thinks Congress’s fiscal progress has also been an unexpected positive for the market.

Stifel strategist Barry Bannister in December 2012 predicted the S&P 500 would close at 1600 this year.

Mr. Bannister says that strategists are “more like drag racers than grand prix drivers”—that is, they’re good at making predictions on straightaways but are usually too early or late when the market turns. He says he didn’t anticipate 2013’s “ebullience” and thinks that the market will be flat in 2014.

Investors shouldn’t expect market predictions to be right on the nose. Some strategists have not only a standard prediction but also ones for best-case and worst-case scenarios.

But in forecasting, you’ve got to set the bar somewhere.

Take weather forecasts. One easy—but very rough—way to guess this Monday’s weather would be to look at Dec. 23rds of the past. In New York’s Central Park, that day on average has been a chilly 35 degrees and seen 0.13 inches of precipitation, according to National Weather Service data from 1981 to 2010.

Of course, forecasters try to do better by using powerful computers to analyze weather patterns. If over time all that computing power doesn’t result in a better forecast, we all might as well shut down the computers and rely on those averages.

Stocks, for their part, have historical price changes for the S&P 500. Stock strategists use complex models to try to be more accurate, but if that brain power doesn’t result in a more accurate forecast, we might as well stick to historical returns.

Since 2000, strategists have failed to meet that low bar.

Say that at the end of each year, you found the median annual change in the S&P 500 since 1929 and used that as your guess for the next year’s price change. For example, in 2006, your guess would have been 9.06%—the median price change between 1929 and 2005.

That method would have beaten the strategists half the time since 2000, according to a Wall Street Journal analysis.

“It doesn’t give you any positive evidence that analysts are providing added information,” says Brad Barber, a finance professor at the University of California, Davis, who has researched the value of analyst recommendations and earnings forecasts. To be sure, Mr. Barber says that we’d need a larger sample size to prove definitively that strategists’ forecasts don’t have any value.

So what’s the problem?

For one, since 2000 the strategists as a group have never forecast a drop in stocks. The upward bias makes some sense; stocks have risen in 55 of 85 years going back to 1929.

But the optimism could also reflect Wall Street’s tendency to be forgiving of bullish strategists who end up being wrong, says Citigroup’s Mr. Levkovich.

He says a Wall Street friend once gave him some advice: “If you’re a bull and you’re wrong, you’re forgiven. If you’re a bull and you’re right, they love you. If you’re a bear and you’re right, you’re respected. If you’re a bear and you’re wrong, you’re fired.”

Mr. Levkovich, who thinks the S&P will rise to 1900 next year, says that he would make a negative forecast if the circumstances warranted it.

But the larger issue might be simpler: Forecasting the stock market accurately is extremely difficult, if not impossible, says Masako Darrough, an accounting professor at Baruch College who has researched biases in analysts’ earnings forecasts.

Rather than invest more or less in stocks based on strategist calls, she says most investors are better off sticking with their usual allocations and, for planning purposes, to assume that stocks over the long run will rise at their usual pace.

Since 1926, large-company stocks have had an average annual total return of 9.8% including dividends, according to Ibbotson Associates. MORN +1.55%

Cut strategists who must make projections some slack, but don’t bet on their accuracy either.