Two Surprises are better than one

A new study suggests investors who use earnings surprises the right way can double the market's average return of 10 percent a year. That's a profitpalooza: Invest $10,000 for two decades, and it's the difference between $383,000 and $67,000.

The finding is based on an oddity that researchers have written about since the 1960s, called post-earnings announcement drift, or PEAD. Simply put, stocks don't fully "price in" good earnings news immediately after it's released. They tend to jump right away, and then gradually drift higher for up to a year. PEAD has done more than perhaps any other financial phenomenon to blow holes in the notion that markets are perfectly efficient.

That said, strategies based solely on PEAD aren't as profitable as they used to be. Earnings surprises these days, for one thing, are smaller than they were decades ago. More pros are scrutinizing larger volumes of information now, and more analysts contribute to estimates. Wall Street has gotten better at figuring out how much companies will earn next quarter. (Or perhaps, companies have gotten better at hinting.) Recent studies show that investors who buy after upside-earnings surprises can expect to beat the market by around three percentage points over the following six months.

Enter Narasimhan Jegadeesh. A finance professor and market researcher, he has consulted for and sold stock-picking models to Morgan Stanley and Deephaven Capital Management, a Minnesota hedge fund. Today he teaches investing at Emory University, occasionally playing point guard in basketball games against his Ph.D. candidates. (Undergrads run too much, he says.) Earlier this year Jegadeesh, along with Joshua Livnat, who teaches accounting at New York University, published a groundbreaking paper in the Financial Analysts Journal.

The study focuses on separating high-quality surprises from lower-quality ones. A soft-drink maker might crush estimates thanks to runaway demand for its energy drink (think Hansen Natural). That's clearly good news. A tire maker may top estimates even while selling fewer tires if it steers customers toward the most expensive ones and cuts jobs (Goodyear). That's okay, but it may not send the stock on a prolonged run.

How, then, can we tell the good earnings surprises from the not-so-good ones? Look at sales. A company that's ringing the register more has better momentum than one that's merely cutting costs. Also, studies show that instances of numbers massaging are higher among companies that miss sales forecasts but beat earnings estimates.

Jegadeesh and Livnat looked at earnings announcements from 1987 to 2003 and assumed stocks were bought a day after the news hit and held for six months. The 20 percent of stocks that had the biggest positive earnings surprises beat the market by three percentage points over six months. The top 20 percent in terms of sales surprises beat by 2.6 percentage points. Zero in on companies that turn up in both groups, according to the findings, and you'll top the market's return by 5.3 percentage points over six months.