Yesterday's Profits

If 2020's forecasts are spot on, they may be true but irrelevant to long-term stock returns.  

By Nick Schmitz

Last week we wrote about how cheap stocks had gotten as a multiple of the last twelve months (LTM) of corporate earnings. In this piece is the reason we are not as focused as many on projections of the next twelve months (NTM) of earnings. It isn’t because we disagree that earnings will contract.
 
This year’s corporate earnings will likely look a lot worse than last year’s corporate earnings for most companies. One of the direst forecasts, from Citibank, has global earnings falling 50% in 2020. Does this mean that stock prices should fall 50% as well? The Chief Global Equity Strategist at Citi believes so. “Typically, stock markets fall the same as EPS [earnings per share] in a recession,” he said. His logic suggests that stocks should fall about another 20% from here.
 
Wall Street may be right that global earnings will take a 50% hit in 2020. And the prognostications of Chief Global Equity Strategists may be right that stocks will trade down 20% more in the interim to match that drop in earnings. But what’s the implication for investors?
 
GMO’s Jeremy Grantham once famously observed that “profit margins are probably the most mean-reverting series in finance,” meaning that profit declines are predictably followed by profit increases. And if a stock is worth the discounted sum of its long-term cash flows in perpetuity, the small minority of cash flows that come in the next few quarters should be largely irrelevant. In other words, the long-term earnings power of a stock is measured more by long-term averages than earnings at the bottom of a recession.
 
Stocks trading in line with earnings would thus imply an inefficient market, one that could be easily arbitraged by those who simply buy every time earnings fall significantly on the bet that they will mean revert.
 
In fact, Eugene Fama, who won the Nobel Prize for arguing that markets were efficient, found that mean reversion of earnings was remarkably predictable. Fama and Ken French found in their 2000 paper that profitability margins mean revert 38% per year. And they revert the most after they collapse the most: “An important practical implication of this result is that forecasts of earnings (e.g., by security analysts) should exploit the mean reversion in profitability,” they wrote. “In particular, negative changes in earnings and extreme changes seem to reverse faster.”
 
To check that Grantham and Fama’s theory wasn’t on the losing side of today’s conflicting advice from sell-side banks, we decided to test out of sample on aggregate Japanese corporate earnings over time.
 
Japan’s Ministry of Finance has conducted surveys of corporate earnings since shortly after WWII, and they cover about one million Japanese firms. We like the Japan test because Japan did go through 20 years of corporate profit stagnation after the asset bubble burst in 1990, removing much of the criticism that the reversion findings are a byproduct of long-term positive GDP growth.
 
So how did aggregate Japanese profits recover after extreme contractions? And if you had had the godlike ability to know how much they would contract ahead of time, would it have been a good idea to get out of the market right before the earnings came out because stock prices hadn’t declined “that” much yet?
 
Corporate profit growth in Japan has been extremely volatile historically during “brief” windows.
 
Figure 1: Aggregate Japanese Quarterly Operating Income, 1954–2019

Source: Japan’s Ministry of Finance.
 
But, following the worst 20% of all historical quarters since 1954, when profits contracted about 19% on average compared to the year before that, the one- to three-year forward growth rates were significantly higher. After profits shrank a lot, they recovered a lot. For the economy as well as for individual stocks, “negative changes in earnings and extreme changes seem to reverse faster.”
 
Figure 2: Mean Reversion of Japanese Earnings

Source: Japan’s Ministry of Finance.
 
In the extremes, following the 20 worst historical quarters for profit declines, profits typically recovered 1–2 years after the contraction. In the worst contraction, by March of 2009, quarterly profits had contracted 81% compared to March of 2008, but two years later they were only 18% lower than before that pain.
 
Figure 3: Time until Corporate Earnings Broke Even Following the 20 Worst Contractions

Source: Japan’s Ministry of Finance.
 
Unless you were concentrated in extremely high-expectation large growth stocks, we believe it was a very bad call to get out of equities because earnings were forecasted to disappoint. We tested stock returns following the 20 worst earnings contractions using the Ken French Japan data, which goes back to 1990. We compare a diversified bet on small, low-expectation stocks to a more concentrated bet on large, high-expectation stocks. These are the cumulative stock returns, including the “Year 0” down year when earnings were contracting.
 
Figure 4: Cumulative Returns Including and Following the 20 Worst Earnings Contractions

Source: Japan’s Ministry of Finance, Ken French Library 5x5 Size and Value Quintile Stock Returns. Only 2 out of 20 of these instances were during the tech bubble.
 
We found no evidence that Fama or Grantham were wrong on this general theory when applied to Japan. Profit mean reversion was particularly strong after profits had decreased the most. And stock analysts would have been wise to exploit that finding both before and after Fama’s paper.
 
Unless there is a credible model that long-term profits will not generally recover this time after an extreme contraction, we believe you should be skeptical of recommendations to sell stocks until they are down x% because this year’s earnings forecast is down x%. Even if the forecasts are spot on, they may be true and yet truly irrelevant to long-term stock returns. Firms are worth what goes on in perpetuity, not in the temporary purgatory of a recession.
 
As Fama suggested, analysts would be wise to remember this. When the head of Global Equity Strategy for a major bank is ignoring Fama’s advice on CNBC and valuations are low, we feel it’s likely a good time to buy for the long haul.
 
This time, profits may not be down because of a long-term secular trend of competitive forces, a GDP growth ceiling, technological evolution, a shock due to a lending meltdown, or the Smoot-Hawley Tariff Act. Society’s reaction to a terrifying virus caused it. But isn’t that all the more reason to be skeptical that profits won’t mean revert yet again and that today’s discount when the dust settles might be a source of returns? The cause of this contraction may be unprecedented, but margin mean reversion in extremes is not.