This In Practice piece gives a practitioner’s perspective on the article “ Information in the Tails of the Distribution of Analysts’ Quarterly Earnings Forecasts ” by Cameron Truong, Philip B. Shane, and Qiuhong Zhao, published in the September/October 2016 issue of the Financial Analysts Journal.
What’s the Investment Issue?
Earnings announcements rank among the most newsworthy information events affecting the value of companies’ securities. For investors, reviewing stock allocations ahead of earnings announcements can have a substantial effect on portfolio returns.
The authors consider whether investors are correctly interpreting analysts’ forecasts and whether their subsequent stock allocations are optimal. It is possible that investors do not use all the information contained in earnings forecasts and, in particular, do not pay enough attention to tail forecasts—that is, very optimistic or very pessimistic analyst forecasts.
The authors assess whether tail forecasts might offer greater investment insight than the market currently believes. They then quantify how much these tail forecasts might add to portfolio performance.
Why Do Current Earnings Surprise Benchmarks Fail to Measure Up?
Currently, analysts tend to use two benchmarks to measure earnings surprise. The first benchmark is the difference between announced earnings and earnings in the same quarter the previous year. The second benchmark is the difference between announced earnings and a consensus of analysts’ forecasts leading up to the earnings announcement.
But the consensus forecast is effectively an average, and this average masks potentially valuable outlying analyst views. These views tend to be ignored, or at least not sufficiently assessed, by the market. The authors believe that this was an oversight.
How Do the Authors Tackle the Issue?
The authors consider a third benchmark to measure earnings surprise and assess its value. This benchmark focuses on tail, or outlier, forecasts. The authors define a tail forecast as the most optimistic or most pessimistic analyst forecast for a stock’s earnings.
To understand why the authors focused on tail forecasts, let’s assume that a company’s earnings beat the consensus forecast by a relatively large amount but that the most optimistic forecast was close to actual earnings. In this case, even though there was an earnings surprise, a muted market reaction might be expected because the optimistic forecast will have been acted on by investors following the analyst who made the optimistic forecast. So, the company’s stock price will already have absorbed some of the earnings surprise.
Now assume that a company’s earnings beat the consensus by a similarly large amount but that the most optimistic forecast was close to the consensus forecast. Relative to the previous case, a stronger market reaction might be expected because the earnings surprise will not have been anticipated either by the consensus forecast or by tail forecasts. In other words, a more pronounced post-earnings-announcement drift (PEAD) might be expected.
The authors back-tested some 126,205 company-quarters from the first quarter of 1987 to the second quarter of 2012 to assess the potential effect on portfolios of including their new measure of earnings surprise. The effect on portfolios was quantified by forecast error, which is the difference between forecast earnings and actual earnings, and tail forecast error, which is the difference between the consensus forecast and the tail forecast.
What Are the Findings?
This study finds that a greater emphasis on tail forecasts captures important information that is not available via the two current measures of earnings surprise. Over both the short and longer term, for companies with large tail forecast errors, there is significantly larger PEAD.
When the most optimistic or most pessimistic forecast for a company’s earnings is close to the two current measures of earnings surprise, the market reacts more strongly in the three days following the earnings announcement. This is because when outlier forecasts are close to the consensus forecast, more weight is attached to that consensus forecast, which leads to a bigger surprise.
This study has particularly strong implications for longer-term stock performance. A portfolio with a long position in stocks in the top quintile of forecast errors and a short position in stocks in the bottom quintile of forecast errors generates a 2.41% excess return in the following quarter. But this return increases to 4.43% when the long and short positions are confined to stocks in the top and bottom quintiles of the tail forecast error.
What Are the Implications for Investors and Investment Professionals?
This study establishes a new measure of earnings surprise that could help investment professionals trade more profitably in both the short and longer term.
The authors found that the market absorbs some, but not all, of the information in tail forecasts. This finding has important implications for the profitability of momentum trading strategies in particular, given that these strategies depend to a large degree on quarterly earnings announcements.
For practitioners who aim to exploit the findings of this study, the authors advocate using their new measure in conjunction with the two current measures of earnings surprise in order to avoid understating the market reaction to earnings announcements and to maximise returns.
Finally, perhaps the overarching message is that the market should pay more attention to analysts who are prepared to stick their necks out and make out-of-consensus forecasts. Running against the herd is not common in the investment industry, so when it occurs, it should probably be taken more seriously. If analysts are prepared to assume the career risk associated with outlying views, they must strongly believe that they have a superior stock-assessment process.
In a nutshell, paying attention to well-informed lone voices might just pay off.