June 21,2022
Everybody and his uncle is now looking for the entry level to the next bull market. The Goldman Sachs team recently wrote a “Recession Manual for US Equities.”
- GS: “Across 12 recessions since World War II, the S&P 500 index has contracted from peak to trough by a median of 24%. A decline of this magnitude from the S&P 500 peak of nearly 4800 in January 2022 would bring the S&P 500 to approximately 3650. The average decline of 30% would reduce the S&P 500 to 3360.”
Not to be outdone, BofA’s Michael Harnett identified 19 U.S. equity bear markets over the past 140 years: average price decline = 37.3% and average duration 289 days. That puts us at 3000 around October 19th 2022.
The reality is that no bear is really similar: different causes, different environment, different remedies, different outcomes. Like saying the average temperature in Canada is 65F. When? Where?
- GS: “Since 1948, S&P 500 earnings have dropped from peak to trough around recessions by a median of 13%. EPS have recovered by a median of 17% four quarters after troughing.”
The norm has been -10 to -15%, except in 2000-01 and 2008-09.
This is not a financial crisis à la 2008.
Could it be a 2000-01 type, given its tech-bubble-like nature, the similarly long economic cycle and a Fed also tightening?
In spite of the 9-year long cycle (averaging 3.8% GDP growth), there were no signs of inflation, stable around 2% until creeping up to 2.5% by mid-2000, right when the Fed stopped its “preemptive hikes” from 4.8% in June 1999 to 6.5%. The S&P 500 peaked in March 2000 and profits did so in September amid a strong economy. A light recession started in Q2 2001.
The S&P 500 index P/E was 24x forward EPS in March 2000 (these estimates proved to be 27% too high!) but its IT sector P/E, 29% of the index, was at 47x. In effect, ex-IT, equities were selling at 19.5x, down from 22 one year before.
The 1999 equity bubble was concentrated in technology stocks. Incredibly, the NASDAQ 100 index multiplied by 4.8x between 1998 and March 2000 to trade at 90+ times earnings. This while the S&P 500 rose 60% and the Russell 2000 only 13%. By comparison, from 2019, the NDX rose 165%, the SPX 92% and the Russell 2000 82% to their recent peak.
During 2000 the losses were understandably also concentrated in technology stocks. The NASDAQ fell by more than 75% between March 2000 and October 2002. Meanwhile, the S&P 500 lost 45% and the Russell 2000 index 36%.
But the bulk of NASDAQ’s losses occurred in 2000 (-51%) while the S&P lost only 15%. During 2001, losses expanded to the entire large cap sector. The S&P 500 cratered 38% through September 2002 even though the recession ended in November 2001 and profits bottomed out in December 2001.
In reality, this was a two-step bear market: the tech bubble explosion in 2000, a $5 trillion wipeout in a $10 trillion economy, caused a small recession in 2000.
Consumer expenditures bounced back in the first half of 2001 but the World Trade Center attacks in September 2001, the Enron/WorldCom debacles and some accounting scandals sent investors not to the sidelines but completely out of the stadium.
S&P 500 profits were down 31% YoY in December 2001 but economy-wide corporate profits declined only 12% in total during 2000 and 2001.
It thus seems appropriate to assume that, if a recession occurs, profits would likely decline by the “conventional” 10-15% from their current $218 (after Q2) level to around $191. That is unless something else happens.
- GS: “The S&P 500 forward P/E multiple has contracted by a median of 21% between its pre-recession peak and its eventual trough. During the typical recession since 1980, the index P/E multiple peaked 8 months in advance of the onset of a recession and declined by 15% between its pre-recession peak and the beginning of the recession.”
The S&P 500 forward P/E having peaked in March 2021 at 22.6, a 21% contraction would bring it to 17.8. It is now 15.6…but on forward earnings of $235. Using recessionary EPS of $191 times 17.8 = 3400. If the P/E contraction is 15%: 19.2x $191 = 3670.
Goldman’s analysis takes no account of inflation and its impact on equity valuation. In 1980-82, inflation was in double digits range, it fell in the 5% area in 1990 and 2.5% in 2000-02 and in the late 2000s. These were much different economic and financial environments.
The Rule of 20 takes inflation into account. Regardless of the cycles save the occasional bubble, the Rule of 20 P/E (actual P/E plus inflation) fluctuates between 16 and 24 with Fair Valuation at 20, where valuation upside (20%) equals valuation downside (20%).
It is now 23.2, down from 29 last December but still very expensive. Based on current estimates, EPS will reach $218 after Q2’22. At the 20 R20 P/E fair value (median of 16-24), using 6% inflation, the S&P would be at 3050. At a R20 P/E of 18 = 2600 (18 – 6 = 12 x $218).
The key is inflation: at 5% core inflation, fair value = 3270, at 4% = 3500 using $218 EPS.
These assume no recession which might take EPS close to $191 but would also likely result in lower inflation.
Using EPS of $191 (-12.5%), inflation at 3% = 3250. Inflation at 2% = 3440.
What all these numbers are saying is that current fair value would be between 2600 and 3500 depending on inflation which nobody can today realistically forecast.
They also say that recessionary conditions, bringing profits and inflation down, would narrow the range to 3270-3440, down 6-10% from current levels.
- GS: “Across the same 12 experiences since WWII, the equity market has begun to price a recession on average 7 months prior to the official start of the recession per NBER’s designation. In all but one instance, the sequence of events was the same: The market peaked prior to the recession and then bottomed prior to the end of the recession.
The 7-month average hides a range between 0 and 13 months. Equities peaked in January 2022. Q1 GDP was negative, Q2? The lead could thus be zero this time. If 7, then an August low, if 13, February 2023.
Similarly, and perhaps more useful, is the 13-month average duration from market peak to trough with a range of 1 to 30 and a median of 15. We are in the 5th month. 13 months = February. But if this is a true inflation fight, that took 21 and 20 months in 1973 and 1980 respectively = fall of 2023.
INFLATION, RECESSION, STAGFLATION…OR GOLDILOCKS
“It is difficult to make predictions, especially about the future.” (Mark Twain?, Yogi Berra?)
Yet, we’re all engaged in crystal balling, unless you abide by the Rule of 20 which only uses actual trailing earnings and inflation data, and buy and sell equities based on objective risk/reward measures. Based on current R20 metrics, this is still not the time to be over-exposed to equities.
But let’s venture some scenarios:
- The Good:
No recession in North-America as the known weak areas (housing, consumer goods, inventory liquidation, current account deficit, government disengagement, Fed QT, war, supply snags, stalled E.U., weaker China, etc.) are partially or fully offset by reshoring/nearshoring/friendshoring capex, rising farm income, higher energy investments/production/exports, goods deflation, 4-5% wage growth, etc.) Goldilocks perhaps! Not too hot, not too cold with inflation cresting shortly and easing in the second half. Seeing a slowing economy and lesser inflation, the Fed turns to neutral to see how things go. R20 fair value is around 3500 but valuations could stay higher than normal given the better environment. Risk on!
- The Bad:
Like above but a mild/short recession occurs in 2022. Profits decline 10% but inflation retreats to the 3% range as supply issues abate and energy prices decline. R20 fair value is 3330. Wait for the Fed signal (see below).
- The Ugly:
Stagflation or deep recession. Like “The Good” without most of the offsets or like “The Bad” without much disinflation. Total risk off, not knowing when and how this will end. When the Fed abdicates and morphs into its dovish creature, it will be time to re-enter.
TIMING CONSIDERATION
“Normally, the stock market struggled the most as inflation was moving up but did fantastic subsequent to the inflation peak — generally even in the event a recession materialized,” (James Paulsen, chief investment strategist, the Leuthold Group). Paulsen’s findings were based on an analysis of 17 prior “major” U.S. inflation peaks since 1940.
Here’s his tally via Axios:
Data: The Leuthold Group; Chart: Erin Davis/Axios Visuals
The little Ronald Reagan in me tends to “trust…, but verify”. It turns out that Paulsen was pretty loose with his “major” inflation spiking periods. The last 6 periods saw core inflation rather stable in the 2.0-2.5% range while in 1960, inflation peaked at 3.3% from 3.0%, not much of a spike.
But the record is quite good with 8 of the other 10 genuine peaks, having been followed by good equity markets, most very positive. Six were followed by recessions. Actually, 5 of the 6 were already in recession when inflation peaked. Only 1957 saw the recession begin 4 months later.
David Rosenberg assesses how low we can go?
Well, all the stock market has done in this first leg of the bear market is mean-revert the sky-high P/E multiple. But in bear markets, the trough multiple is 12x. And in recessions, corporate earnings decline by 23%, on average. Do the math — under the proviso that earnings and the multiple bottom simultaneously, we would be talking about 2,200 on the S&P 500. That isn’t a forecast as much as a mathematical construct. That indeed would represent a full 100% Fibonacci reversal. The next test, as an aside, is the 61.8% retracement, at 3,215 (after the latest violation of the 50% reversal level of 3,516).
A 12 trough multiple? Ed Yardeni would disagree, particularly when there is inflation. Since 1957, the average P/E at bear market lows was indeed 12.7, but with a range of 7.2 to 19.2:
Since 1942, the average is 11.7, but it’s also a useless camel that makes no bones of inflation!
As seen previously, recessions cut earnings 10 to 15%, except in special crisis like in 2000-01 and 2008-09.
So let’s do Rosie’s mathematical construct again, but using the Rule of 20 P/E. Taking inflation into account, the Rule of 20 sports a much more consistent “low low” around 15 (averaging 17.3).
But first we need an inflation rate. Given the assumed recession and based on history, core inflation could be cut in half from its current 6% rate. At 3% and a R20 P/E of 17.3 = 2700. If the Fed has its wishes, at 2% inflation = 2900.
In our risk assessment exercise, this 2700-2900 range registers as our worst case.
Since we can never know where the actual bottom will be, the Rule of 20 helps us objectively measure our risk/reward ratio. At any given time, a R20 P/E below 20 means that the valuation downside is lower than the valuation upside. At current EPS ($218 after the Q2 earnings season) and inflation (6.2%) levels, this “fair value” is 3000 (actual P/E of 20 – 6.2 = 13.8).
Briefly said, that means that, at current EPS and inflation levels, equities would only begin to provide a positive risk/reward ratio below 3000. Another way to say it is equities are still 18% overvalued given actual, known parameters.
This is a dynamic process. If analysts are right and EPS reach $224 after Q3 and inflation retreats to 5%, the R20 fair value would rise to 3360.
Inflation at 5% in the fall requires monthly core CPI prints averaging 0.3%. If 0.2% monthly, inflation would be 4.5% YoY in October and fair value would reach 3475.
THE CRITICAL FED EFFECT
Inflation is the most crucial factor to watch in the second half of 2022. Falling inflation would help the R20 fair value rise and potentially offset a mild decline in profits. More importantly, a sustained easing in inflation trends would likely appease the hawks in the FOMC and incite the Fed to turn dovish sooner than later.
Since 1957, equities performed well after 11 of 13 Fed easing episodes. The R20 P/E was above its 20 fair value in 4 of those episodes.
The Fed signal has been better than Paulsen’s inflation signal in 1957 and 2000 and proved more potent during the last 25 years of subdued inflation.
While on interest rates, David Rosenberg, a bond bull, affirms that “there is no way that we get to any bottom on the stock market absent a slide in Treasury yields (the average decline is -140 basis points).”
That only verifies since 1998 (though not in 2018). Prior bottoms occurred with much smaller declines in Treasury yields, 5 without any and 3 actually after yields rose.
CONCLUSION, P/E ANALYSIS
- The worst case is 2700-2900 if a crisis or stagflation.
- Fair value is in the 3300-3500 range.
- Watch inflation and the Fed, particularly if the economy slows measurably this summer. Slower inflation might bring the doves back and a good buying opportunity along.
PRICE TO BOOK
The S&P 500 index is now selling at 3.7x book value, down from 5.0x at its peak. The pandemic low was at 2.8x, the 2016 low at 2.7x and the 2002 low (mild recession) at 2.4x (the GFC low at 1.6x). Ex-GFC, the low range would be 2400 to 2800.
Using P/B, one always needs to also consider the ROE, the rate of return on book value.
ROE was 12.8% in 2002, 14.5% in 2016 and 15.3% at the pandemic low. There is both a growth and a compositional effect here. At $191 EPS, the ROE would be 19.0% which could justify a P/B in the 3.5x range. Assuming BV declines the “recession usual” 7% to $925, the low would be around 3250.
Another way to look at P/B is its YoY fluctuations over time. Since 1980, the S&P 500 P/B ratio declined an average 19.7% YoY at its lows. The worst drops were in 1982 (-23%), 1988 (-27%), 2003 (-23% at the market low) and 2009 (40.0% with huge financial markdowns). It is now down 21%. Worst case at -40% is 3.0x BV or 3000 but barring a crisis it would decline 27% and take the index to 3375.
(Morningstar/CPMS)
All in all, current fair value falls in the 3300-3500 range. But watch inflation…and the Fed.
Source: 
