US Corporate Earnings Estimates Are (Still) Too Low

By in
US Corporate Earnings Estimates Are (Still) Too Low

Two “Data” topics today.

#1: Our entire edge in correctly predicting Q3 US corporate earnings would handily beat expectations and stocks would rise in Q4 as a result came down to one idea: there was no way the S&P 500 would print an aggregate Q3 earnings per share result that was lower than Q2.

  • That’s what analysts were expecting at the end of September: S&P Q3 earnings of $48.93/share even though Q2 had printed an ACTUAL $52.80/share.
  • The actual Q3 result is looking like $53.47/share, 1.3 percent higher than Q2’s actual.
  • With 11 percent of the S&P 500 still to report, Q3’s actual EPS should end up closer to $54/share, or 2.3 percent better than Q2 2021.

This trend of analysts discounting current results as unsustainable continues even now; the FactSet chart below shows that analysts’ earnings expectations for Q4 2021 and Q1 2022 are below Q3’s current run rate. Specifically, the Street is looking for earnings to be down 4.4 percent in Q4 from Q3 2021, and down 2.6 percent in Q1 2022.

You’re probably wondering if seasonality might be the reason for analysts thinking Q4 and next-year-Q1 will be below Q3’s actual. Here is the seasonal earnings data for the last economic expansion (2009 – 2019, from S&P):

  • Yes, Q4 did average a small (3.4 percent) decline in S&P 500 EPS relative to Q3.
  • But … early in a recovery (2009, 2010), Q4 EPS was higher than Q3 (+8.7 pct, +1.7 pct).
  • And no, Q1 does not, on average, see an EPS decline from the prior year’s Q3. From 2009 – 2019 the average increase was 3.9 percent.

Takeaway: there’s little reason to believe Q3 2021 will be peak S&P 500 earnings, because 1) US corporations are proving their ability to hold margins as the domestic/global economy heals and 2) seasonality is simply not a factor. That means estimates have room to increase further and/or the S&P 500 still has room to outperform earnings expectations.

#2: Three points of note about Friday’s US Employment Situation report:

First, the positive revisions to August and September were almost as important as the fact that October’s headline job creation number (+531,000) was so far ahead of expectations (+450,000). August was revised upward by 32 percent to +483,000 jobs added and September by 61 percent (not a typo) to +312,000 jobs added. That means:

  • US job creation has averaged +496,000/month over the last year.
  • There are still 4.2 million fewer jobs in the US economy than in February 2020, the month before the start of the Pandemic Recession.
  • If the American economy can continue to create 500,000 jobs/month, then the gap to February 2020 will be completely closed by June-July 2022. Yes, a big “If” given a whole host of factors (participation rates, quit rates, etc.), but the last 3-month average is still +442,000/month even with all those issues as headwinds.

Takeaway: the current pace of US job creation fits with the Fed’s tapering schedule (being done by mid 2022) as well as market expectations for 1-2 rate hikes in the second half of next year. Sometimes data fits market narrative, and we have such a situation right now.

Second, labor force participation is improving for the key 25–54-year-old demographic even as overall LFP continues to stagnate. The chart below dates back to Q4 2019. As noted, total LFP has been flat at 61.x percent for over a year. Working aged adult LFP has been grinding higher, however. As a result, total LFP last month was still 1.8 points below its pre-pandemic peak but 25–54-year-old LFP was only 1.3 points below 2019 levels in October 2021.

Takeaway: demand for labor is pulling marginal workers back into the US labor force even though the total LFP doesn’t show that fact.

Lastly, while last month’s 4.6 percent unemployment rate is still not back to early 2020’s 3.5 percent, it is still quite strong by historical standards:

  • From 1970 – 1979, US unemployment was only as low as 4.6 percent in one month (October 1973). After the 1973 oil shock, the best it could do was 5.6 percent (May 1979).
  • From 1980 – 1989, the US unemployment rate never broke 5.0 percent. The low point was 5.0 pct exactly, in March 1989.
  • From 1990 – 1999, it took until 1998 before unemployment was 4.6 percent or below. It managed to hold below that level for 3 years.
  • From 2000 – 2009, unemployment was at 4.6 pct or below only from mid 2006 – late 2007.

Takeaway: judging the current US labor market by only its 2018 – 2019 (pre-Pandemic) levels misses the fact that current unemployment is already quite low by historical standards. Yes, the Fed is anchoring on early 2020 (unemployment at 3.5 pct) as its target but the long run analysis shows we can see reasonable economic growth even if unemployment takes longer than expected to decline.


FactSet Earnings Insight report: