Post earnings announcement drift

Checked October 2026

Post-earnings drift buys companies that beat estimates and sells companies that missed, on the idea that the move keeps going for weeks. Victor Bernard and Jacob Thomas measured the drift in the Journal of Accounting Research in 1989. Backtoast ran it. A version with fixed beat and miss lines made 0.4% a year at +0.20% a trade, but failed our test on capacity and outlier dependence.

Backtoast

You pay $99 for two full tests of your own strategy, and a pass is an offer of 30% of the profit Backtoast makes trading it with its own money.

Post earnings announcement drift

The academic version ranks each quarter's earnings surprises, buys the top tenth, shorts the bottom tenth, and holds for about 60 trading days, roughly three months.

The run couldn't rank against the previous quarter's surprises, so it used fixed lines instead, buying a beat of 20% or more of the estimate and shorting a miss of 10% or more, on every US stock with no price floor. Those lines are guesses at where the top and bottom tenth start.

The options trade around the same reports is the earnings straddle.

When it works

The drift is largest in small, thinly traded stocks, where news takes longer to get into the price, but those are also the costliest to trade. Traders keep using it because it's one of the longest-studied effects in stocks, and the signal arrives on a known date each quarter. Requiring a big price move on the report day, in the same direction as the surprise, is a common refinement.

Origins

Bernard and Thomas used a forecast from each company's own past earnings, and Livnat and Mendenhall's 2006 version uses the analyst estimate, which shows a bigger drift. From 1974 to 1986 the top tenth beat the market by about 2% over the 60 days and the bottom tenth trailed it by about 2%. The run is from late 2021 to October 2026.

The run is Backtoast's, from 29 October 2026, after costs.