A couple weeks back, widely followed SA contributor Fear & Greed Trader commented on today's twelve-month trailing GAAP earnings levels, and where they stand relative to history. I think there are some important points to be made on the issue, and I disagree with F&G's conclusions, though I think his take on the issue is fairly typical of many of today's bulls.
Given that this is a response to a portion of one of his articles, I will begin by quoting him directly:
The skeptics deserve some slack on their opinion of the valuation issue, given that we haven't seen the reversal in earnings just yet. I'll concede that since the last time I said the market was inexpensive was probably in 2013. I do understand how things are different from that point in time. However, I will continue to put forth ideas and debate those that believe the market is wildly overvalued today.
The current trailing S&P 500 P/E is around 24, and that has many investors petrified. It also emboldens the skeptics, as they are convinced that the market can go no higher because of that. All anyone has to do is go back at the historical numbers to see that 24 is the long term average for the last 35 years.
I firmly believe that market participants have to factor in the current interest rate before they look at the numbers, and conclude that stocks are excessively valued and should be sold. History reveals that there were numerous occasions where the alternative of a 10-year treasury yield was competitive (5-6%) and the S&P sold for this same multiple.
The bullish side of the market story also implies that when (emphasis added) earnings rebound later this year, the PE ratio will start to be more attractive. Market surprises, as in most secular bull markets, will remain to the upside. It is prudent to add that thought into the equation."
Before I begin, let me say that F&G is distinguished from many bullish commentators in that he even brings up GAAP TTM earnings at all. Many bulls ignore it completely, which wouldn't be so bad if they didn't religiously cling to nonsensical FactSet non-GAAP one-and-two year forward earnings that never come true as their one-stop alternative.
For an excellent discourse on the GAAP vs. Non-GAAP debate, I'll direct you to the Editors' Pick piece done two weeks ago by SA Editor Marc Pentacoff, who does an excellent job establishing what the proper role of accounting statements is.
Now I will address my main of disagreement with F&G's brief analysis on this "skeptic's" concern. I will not address his interest rate argument in this piece, but let's just say that the "low rates equal high multiple" argument is weak from both a theoretic and an empirical perspective. I shall leave it there for now.
Time Frame - using a longer time period
The beginning point for F&G's analysis began in 1981. This actually marks the end of the 1968-1982 secular bear. The S&P hit a low P/E of about 6.5, which in total fairness pulls the median and mean down. (For an arithmetic mean such as the one calculated, an extreme-but-positive small value does less to drag down the mean than an extremely large value drags it up.)
If we extend out the analysis time frame using Robert Schiller's data (as displayed on multipl.com) going back to 1871, we get a long-run average P/E of 15.60x, median of 14.63x. One can see on this long-run multiple chart that the P/E really has encapsulated many of the highest historical levels over F&G's chosen time frame.
While we have to respect the reality of the high-multiple regime that we're in, we also need to remember that we call them "regimes" for a reason: they last for a while and then they change, often without much warning. (See 1981 in both the multiple graphs for a perfect example of this.)
Furthermore, this high-multiple period has itself contributed to the mean and median levels that are calculated in the longer-term period. Even in the data set that stretches back to 1871, the past thirty-five years account for 24% of the observations which result in a long-run average of 15.6x. Clearly our most recent experience has raised the prior average higher.
Mean vs. Median
Today's TTM GAAP P/E stands at 24.44x - or at least it did immediately before the Brexit vote! According to F&G's chart, over the past 35 years the mean P/E was 24.29 while the mean was 19.63. This is to say that over this time horizon, the mean P/E has been a full 23.7% higher than the median P/E!
Now providing there is any data variation (there is), in order for a mean to be a mean, the data must spend some time above the mean and some time below it. As a fun side fact, that's the reason that the standard deviation formula looks the way it does (shown). We're trying to calculate the "average distance from the average", and if we don't modify the original average equation, the average distance always sums to zero.
Interestingly, the multipl.com long-run difference between the mean P/E (15.6) vs. the median (14.63x) is only 6.6%, still meaningful but much less so than the 23.7% gap that occurred during F&G's chosen time frame.
With such a large difference between the mean and median, it may be well worth a look as to when P/E spikes have occurred over the last thirty-five years.
When was the S&P P/E ratio above its 35-year median?
- in 1987 right before Black Monday (mini bubble)
- in 1992 when our economy was coming out of a recession ("It's the Economy Stupid")
- The "Irrational Exuberance" great bubble period which began in mid 1996 and peaked from a P/E standpoint in 1999, from a price standpoint in 2000
- The recession of 2001 and earnings scandals of 2002 (Enron, Worldcom, general "bathtaking"). Of particular importance, note how the peak P/E of 2002 was actually higher than the P/E of 1999!
- The towering multiples of 2008-9. This of course had nothing to do with the SPX price being high (the index lost more than half its value), but rather the GAAP earnings falling so much lower.
- The period from 2013 to today
F&G's graphic and commentary demonstrates that today's SPX multiple is currently at the mean level since 1981. The only other times in the last 35 years that we have been at or above the median have been at the end of recessions or during bubbles.
1987's Black Monday
In 1987 the market launched higher, way beyond what the strong earnings fundamentals would suggest. Black Monday ensued, wiping out the year's heady gains in a single day. Let's call 1987 a mini bubble (still, no earnings issues): dramatic, but swiftly popped and with no major damage done to the economy or the markets.
The Early 90's Recession
In June 1989, TTM GAAP earnings peaked at $25.20/sh. By the end of 1991, they had fallen 37% to just $15.97. Note that it took thirty months and a recession for the earnings to reach their bottom level. Today's earnings have fallen 18.5% over the last nineteen months, with no economic recession.
It was in the final quarter of 1991 that the GAAP P/E hit its max - 26.9x. Earnings doubled over the next three years. Sure the P/E was high, but the market recognized that we were exiting a recession, having shed excesses, and that earnings growth can often be blistering coming out of such periods. So SPX participants bid P/E's all the way near where we are right now in anticipation of this. Also, note that this doubling was a doubling in GAAP earnings, not the operating EPS that companies so easily distort. Anyone care to wager that GAAP earnings will be double three years from now - $169/sh?
Over the three years between Q4 '91 and Q4 '94, earnings doubled while the SPX increased a mere 15% total from 417 to 480. Investors had "bought" all their earnings growth on the front end, and the actual returns were nothing special once the actual growth came. This is critical: remember that the 1991 Q4 P/E ratio was quite close to today's.
After the 1991 earnings trough, the "Roaring Nineties" was a raging price bull fully supported by a raging earnings bull. Earnings troughed at $15.97 in Q4 '91 and reached $53.70 by Q3 2000. So sure, those were heady days, but SPX earnings more than doubled the level of their prior economic peak in Q2 '89. In stark contrast, today's GAAP EPS stands just 1.75% higher than the prior peak of Q2 '07 ($86.41 now vs. $84.92 then).
Even with all that tremendous earnings growth, prices still got ahead of earnings and in Q1 '99 the GAAP P/E maxed out at 33.52. Of course, the market stormed higher for the rest of the year, but the P/E actually fell because between Q1 '99 and Q1 '00, GAAP EPS rose 33%!
The 2001 Recession
The dot-com bubble burst in early 2000. A mild recession took hold, during which GAAP earnings were cut in half in just six quarters.
This explains why the P/E rose to a staggering 46.5x - about 40% higher than the dot-com peak P/E. This level was approximately double the historical mean that F&G quotes. Of course, this P/E expansion did not involve price expansion, but contraction. It's just that earnings fell even faster.
This is huge: the SPX P/E reached its peak right as the earnings hit their trough: GAAP EPS doubled in two years, and rose an additional 74% over the following three-and-a-half years. That means that from Q4 2001 through Q2 '07, earnings rose a blistering 222% from where they were just 5.5 years earlier.
As we observed with the great 1991 earnings comeback, the scorching growth in EPS was not matched by the market itself. While EPS grew 222%, SPX rose just 33% over the same stretch of time. Investors had front-loaded all those spectacular EPS gains by paying a high P/E.
Said a little differently, it required an explosive increase in earnings to earn an annualized price return of just 5.32% over the same period.
GFC GAAP Anomaly
Finally, there was the earnings collapse of '08-09. The GAAP P/E hit 122x at its peak in June '09. All I'll say here is that the earnings spill, while severe, was also brief. It is during times like this that I do believe there is value in carefully examining the message of operating EPS vs. GAAP. The only major point that I'll add is that this extreme value pulls the mean P/E up for the entire data set.
Remember, SPX is currently at its mean P/E level, which was skewed mightily higher by the extreme values of late '08 through late '09 (observed quarterly P/Es of 61x, 116x, 122x, 84x). Just those four data points alone add about 2.5 multiple points to the 35-year S&P mean P/E!
Once again, earnings since then are up eleven-fold from the bottom level, while the market has tripled over the same time frame. I really do not believe that this period is at all relevant, other than observing that it had a very large impact on the mean P/E, so much so as to essentially force us to throw the mean away in favor of the median of 19.6 for an analysis of this time length.
All the post-GFC earnings growth that has been sustainable was concluded by Q2 2011 ($83.87 then vs. $86.40 today). That means that near all of the gains since summer 2011's SPX level of 1320 have been built on multiple expansion, with virtually none of the same gain attributable to earnings growth over the past five years of an economic expansion. There is simply no historically similar precedent for this over the period that F&G cites, not even in the roaring bull of the 1990s.
So the only times over the past 35 years on F&G's chart where P/Es have been above their median have been during mounting bubbles (1987, 1996-9, 2013-today?) or when earnings have been kicked down and the market prices in skyrocketing EPS growth. Even during the great bubble periods there was at least great earnings growth to justify higher multiples.
We now stand at 24.4x TTM GAAP EPS. We have not been in an economic recession, though if earnings tell the tale you could hardly call it "Goldilocks", as so many do. Past earnings troughs have indeed seen high P/Es - but were always followed up by gigantic earnings growth in the next 3-7 years.
So, we are left with a multiple that demands phenomenal EPS upside in order to generate merely respectable returns if history is indicative of further upside from here.
If this scenario does indeed unfold, let me be the very first to heartily praise the intrepid bulls for their keen insight and willingness to stay the course. Truly they deserve massive credit for sticking to their guns.
Of course, it has been seven years since we left the last recession. 10-year Treasuries are screaming that economic growth will be very slow. The Fed may have good reason for their caution and flip-flopping as they agonize over whether to raise rates by the second 25 basis points in 10 years. We still have a record number of the US citizenry on social assistance programs, and the labor force participation rate has hit a 37-year low.
Maybe there is a mind-bending increase in earnings right around the corner… maybe… maybe… FactSet sure tells me so.
So I ask you, what happens if P/Es are at this level and the requisite EPS growth does not materialize? What happens if in fact earnings continue to stall and/or fall?
Something to consider. Until then, please do call me a skeptic.
Disclosure: I am/we are short SPX.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: Short ES Futures w/ flexible options positioning