By Jack Hough
Among a vast number of clues that have been found to predict stock performance, two bode particularly poor. One just appeared and the other seems to be coming.
The first is a downward revision in earnings forecasts. Analysts, like investors, tend to be slow to change existing assumptions about stocks when new information appears. Historically, they’ve changed estimates little by little, with one revision leading to future ones in the same direction. In the past three months, analysts have reduced estimates 20% more than they’ve raised them, according to new research by Bank of America Merrill Lynch.
That’s important, because the argument that stocks are a good deal today rests in large part on shares being cheap relative to earnings. If earnings are poised for a shortfall, shares might not be so cheap, something this column argued in June (see “The Invisible Stock Bubble“).
One footnote to this first clue offers a sliver of comfort. BofA calculates that a typically downward revision cycle lasts a year and a half and drags share prices 14% lower. This one is barely half a year old and, and since it began, investors have already lost more than 14%. So perhaps the news is priced in. If not, this cycle would be worse than usual.
The second dark clue is when earnings estimates become broadly scattered. Poor earnings and stock returns tend to follow. It remains to be seen whether “estimate dispersion” will increase as some forecasts are reduced, but the early evidence suggests it will. Researchers theorize that analyst disagreement on forecasts is linked to companies clamming up on their own earnings projections, which they tend to do, as you might expect, when the projections are gloomy. Currently, the fewest companies are issuing guidance in a decade, according to Bank of America Merrill Lynch.
In this case, no news is probably not good news.