The Real Story About Analysts’ Estimates

By in
The Real Story About Analysts’ Estimates

“Why are Wall Street analysts’ earnings estimates almost always too high for the upcoming year but too low for the current quarter?”  I (Nick) spent 9 years as a senior equity analyst at the old First Boston (now Credit Suisse) and was on the Wall Street Journal’s “Most Accurate Earnings Estimates” list every year until I went to work for Steve Cohen.  I can, therefore, give you an informed answer to this thorny question with the following 3 points:

#1: Analysts are a perennial optimistic group.  There are, for example, just under 11,000 analyst ratings on S&P 500 companies right now.  Of those, 56 percent are “Buys” and 39 pct are “Holds”.  Just 6 percent are “Sells”.  Sometimes this positive bias proves correct; the Energy sector has the greatest percentage of “Buys” presently, at 64 percent.  But Tech is in second place, at 62 percent, and most of those have been dreadful calls this year. 

This inherent optimism spills over into analysts’ earnings estimates, and this has been true since the 1980s. Many years ago, McKinsey did a clever analysis looking at the trajectory of analysts’ earnings per share estimates through the course of a calendar year.  The chart below shows their findings.  The green squiggles track how estimates progressed through a given year.  The blue dot shows actual earnings.  The green lines almost always trend lower, showing how analysts’ estimates start too high and only come down over time.   

Analysts’ optimism has not changed since 2008, the last date on this graph.  The S&P 500 will most likely earn $220/share this year.  At mid-year, analysts had $230/share in their collective models.  They currently have $236/share for 2023, a 7 percent increase.  History says that number has to come down. 

#2: Analysts only cut those annual estimates 1 – 2 quarters at a time, and by just enough so the companies they follow can modestly beat those widely-followed expectations.  Remember: most analysts have many more “Buy” and “Hold” recommendations than “Sells”.  The first two categories are hard to defend if an analyst’s earnings estimates are consistently too high.

If you want to understand how, say, the current Q3 earnings season is going, you need to compare the percentage of companies beating estimates and the amount of those beats to prior reporting cycles.  Here is that data:

  • FactSet’s latest Earnings Insight report shows that 71 percent of companies reporting Q3 results have beaten Wall Street analysts’ estimates and overall companies are beating estimates by 2.2 percent.
  • That may sound pretty good, but these numbers are well below historical averages.  Over the last 5 and 10 years, 77 and 73 percent of companies have beaten estimates by an average of 8.7 and 6.5 percent. 
  • Q3 earnings is, therefore, not going well at all. 

#3: To use analysts’ estimates to make money trading and investing, you need to think about where their numbers will be trending 6 months in the future.  That is how far out markets look as they set prices on a daily/weekly basis.  This is why the S&P peaked in early January 2022; investors were starting to discount the cuts to 2022/2023 earnings estimates that started in June – July 2022 and continue today. 

Takeaway: to believe that US equities are at an investable low one must have strong conviction that by April-May 2023 analysts will start to increase their estimates for 2023 – 2024 corporate profits.  That implies that either they grew too pessimistic (which would be highly unusual) or US economic growth troughs in Q2 2023 and there are clear signs of a recovery going into 2H 2023.  Either/both are certainly possible, but we would like to see further estimate cuts and signs of economic slowing before we grow more bullish on US stocks. 


FactSet Earnings Insight report (analysts’ ratings/earnings beat data):

McKinsey Study of Analysts’ Earnings Estimates: