Surprise! Behavior Has Not Changed

Barron’s latest cover story, Wall Street’s Real Earnings Surprises, suggests that positive earnings surprises are not being rewarded with share price jumps like they used too. It may be true that the one-day response rate has declined, but since this touches close to home, we would like to highlight how we look at this behavior and see if there is reason for caution.

At 361, our equity strategies are managed in large part based on research findings that there is a positive return associated with earnings surprises (and even more so to earnings revisions, but we’ll save that for another post). Our quantitative models capture these earnings surprise events and use them as inputs into our stock ranking process every day. We believe investors are still paying attention to these events over the longer term.

Our main thesis for equities, is that price moves reflect changing expectations regarding a company’s prospects. Being able to anticipate positive behavior in the form of earnings surprises and, more importantly, analysts’ earnings revisions, should lead to positive returns over time. From our perspective, there is no reason to sound the alarm. We track this data daily, but our lens is focused on a longer timeframe with respect to its validity.  Our process is based on research performed on a monthly basis, which captures the post-earnings-announcement drift (PEAD) referenced in the story.

Another key difference between Barron’s observations on market behavior and what our methodology seeks to take advantage of, is that they are looking at the average return of all S&P constituents. Our models rank each stock based on the relative magnitude of surprise, and measure average returns of each decile. The “spread”, or difference between the best 10% (decile 1) and worst 10% (decile 10), is an indicator of the efficacy of the metric. The chart below shows a slight downtrend in the spread’s peaks over the past two years, but does not signal that investors are completely shrugging off positive surprises. Admittedly, this picture is a short timeframe (our research spans a longer history).


Source: S&P Global ClariFi, 361 Capital.  Decile 1 and Decile 10 returns are an equal-weighted monthly average.

Here is the same chart for the small cap universe.


Source: S&P Global ClariFi, 361 Capital.  Decile 1 and Decile 10 returns are an equal-weighted monthly average.

We utilize this measure by gaining “tail” exposure. That is, we seek to select stocks that rank higher on this measure (among others). That way, even if the “average” of all stocks is declining, the highest ranked may still be offering alpha opportunities.

This is not an attempt to dismiss the Barron’s story, but more of a means to give a quick glimpse on our thinking around this phenomenon, and to demonstrate that how you measure this matters when judging its efficacy.  In fact, of the five stocks identified in Barron’s screen, three are owned in 361 equity portfolios. A fourth is highly ranked in our behavior model, but is unattractive on a valuation basis, and so it has been passed over. For a deeper dive into our longer-term research on investor behavior and how we implement it, read our white paper. The attractiveness of various market factors ebb and flow each period, though we would contend that human behavior is evolving slowly and thus, not in danger of losing its predictive ability.

Read our related blog, How are Analysts Feeling About Small Cap Stocks?