Annual stocks forecasts are worse than wrong
Usually cheery predictions often wide of the mark.
Don’t expect Wall Street investment houses to predict a stock market decline in 2017. That’s just not what they usually do.
Instead, they engage in an annual ritual that is underway: Every December as the holidays approach, Wall Street gurus examine the stock market, and nearly all declare that stocks will rise in the forthcoming calendar year.
The forecasts are still coming in for 2017, but preliminary tallies suggest that — no surprise — strategists are bullish, probably mildly so. Through last year, since 2000, the consensus has always been bullish, holding that the market would rise, on average, about 9.5 percent a year, according to calculations by Bespoke Investment Group. In reality, it rose only 3.9 percent a year, on average, in that period.
So the cheery predictions have often been wrong. Does it really matter? After all, the stock market actually rises most of the time.
Bullish predictions encourage investors to pour fresh money into the markets, helping asset management companies to enjoy rising profits. Even if returns don’t match the expectations set by forecasters, memories are short and money is being made.
But here’s the rub: The stock market sometimes falls, and from time to time, it absolutely tanks. Since the start of 2000, the Standard & Poor’s 500 index has ended in negative territory in five calendar years (2000, 2001, 2002, 2008 and 2015) and has been virtually flat once (in 2011). But while a handful of individual forecast- ers have, from time to time, predicted mildly negative years for stocks, the Wall Street consensus in every single year since 2000 has predicted a rising market.
Consider the calamity of 2008. If you had money in stocks that year, you would probably remember. The S&P 500 fell 38.5 percent in the course of those 12 months. It would have been very useful to have received advance warning that stocks were about to plummet, but the Wall Street consensus did not ring out an alarm. On the contrary, the forecast for 2008 was unusually bullish, calling for a rise of 11.1 percent. Wall Street missed the mark by 49 percentage points that year.
How bad is the industry’s track record in making predictions? I had assumed that the annual forecasts were essentially worthless — no better than flipping a coin. But Salil Mehta, an independent statistician who has blogged about the topic, tells me I’ve been too kind. The forecasters, as a group, are much worse than that.
“It’s not easy to be as bad as they are,” he said in an interview. “They are much worse than random chance alone would predict.” (Mehta was formerly direc tor of research and analytics for the U.S. Treasury’s Troubled Asset Relief Program and for the federal Pension Benefit Guaranty Corp.)
Af t e r e x a mi n i n g f o r e - casts by major investment houses going back to 1998, Mehta found that 8 percent of individual analyst predictions called for a small market decline in subsequent years. But those predictions of decline were worse than random: In the years when the market did fall, 9 percent of forec asts — never enough to counter the bullish consensus — predicted that it would happen, essentially the same as in a year in which the market rose.
Yet, as Mehta found, the typical negative individual forecast called for only a 5 percent decline, so the size of forecast errors, like the 49-percentage-point error of 2008, was larger than would be expec ted from mere chance. On a statistical basis, the forecasters were “actively adding negative value” — essentially destroying value by issuing spurious numbers.
If I were to simply program my computer to predict a market increase of, say 5 percent each calendar year, without fail, Mehta said, my forec asts would be better than the consensus version — which is to say, they would be lacking in value. The forecasts are worse than that.
Oddly, 2016 is likely to look like a reasonably good year for the consensus forecasts, which last December called for the S&P 500 to end 2016 a little above 2,200. That is where the market is hovering now. But even this year is no time for gloating: Early in 2016, the market declined s h a r p l y a n d f o r e c a s t e r s shi f ted t heir predic t i ons lower. Then they adjusted upward after the market rebounded. The consensus forecast this year looks good only if you rely on the December version and discard the other two.
Individual forecasts somet i me s s w i n g w i d e l y y e t remain way off the mark. In December 2015, for example, Federated Investors predi c te d t hat t he S & P 5 00 would rise 18.2 percent this year. By February, it sharply downshifted, forecasting a net loss of 10.5 percent. But facing a rising market, by September, it became mildly bullish, predicting a net 2.7 percent gain in 2016. So far, the first estimate seems to be the best one: The stock market is up more than 10 percent for the calendar year.
Part of the problem is that calendar years are arbitrary periods when it comes to the markets, which are notoriously difficult, if not impossible, to time accurately and BUSINESS STAFF BUSINESS EDITOR STAFF WRITERS consistently. Will the market rise 6 percent in the next 12 months? No one knows.
While such forecasts are fruitless, there are others that are less precisely timed, more nuanced and sometimes worth pondering.
Back in 2012, I wrote about one such forecast. Seth J. Masters, chief investment officer of Bernstein Private Wealth Management, predicted then that the Dow Jones industrial average would reach 20,000 within the decade, an increase of roughly 50 percent. The forecast was made during a period of deep gloom. The market had been flat in 2011, many ordinary investors were on the sidelines, and most analysts had tempered their customary bullishness.
But Masters looked at the relatively low stock market prices and at the extremely low level of interest rates engendered by the Federal Reserve and other central banks, and concluded that the probability of a long-term increase in stock values was very high. That wasn’t a very controversial conclusion. But it contained a message: He was urging investors to take a little risk and hold stocks for the long run. The market is trading very close to 20,000 now, and I spoke to him again last week.
“If anything, we weren’t bullish enough then,” he said. “We said it would probably happen by 2022, and we are already there.”
Mas t e r s wa s s h e e p i s h a b o u t maki n g e ve n t h a t forecast, pointing out that he doesn’t issue annual calendar year predictions. “I don’t know how to do that accurately,” he said.
His outlook is less bullish now. “We’ve just had a great bull market,” he said. “That’s not going to keep happening. At this point, I’d predict that we will cross Dow 20,000 many times: The market won’t just go up from here.”
Stock prices have surged since the election and are much higher than they were in 2012, interest rates are rising, and political uncertainty is mounting. He said that the greatest likelihood over the next five or 10 years is that stock returns will be lower than they have been lately, and that bond returns will be constrained too.
“It may be a more difficult time for investors,” Masters said. But he wouldn’t issue stock market price targets. “We can analyze trends, we can give some probabilities — we can’t really predict the future.”