John Hussman’s monthly letters are always brilliant, brimming with insights, based on cold, hard mathematical certainties, presented with logical beauty and eloquence. I know, however, that some find them hard to read. So, here is my “explanation” of his letter of April 11, 2021.
Dr Hussman begins by teasing apart the components that made up the stock market total returns over the last 21 years, as measured by the US S&P 500 index.
This is how the S&P 500 returned 6.8% per year for 21 years
First, let’s check out the accounting and business fundamentals. How did the actual companies develop in the real world, i.e., how much income did they generate from their clients by providing real goods and services and accepting cash transfers in return? The 500 companies in the S&P 500 index grew their revenues by 3.5% per year, during the 21 years since January 1, 2000. Not great, not terrible, kind of what one could expect of a measure of the broad economy. One per cent more people per year, one percent more units sold per person, and one per cent higher prices per year. Something like that. Slow and steady. Quite predictable.
Second, let’s look at the market’s appreciation of those fundamentals. How much was an investor prepared to pay for a company with 1 billion dollars in revenues at the start of the period vs. at the end of the period? The so called Price/revenue (or P/S, Price/Sales) multiple expanded from 2.36 to 3.09, adding 1.3% per year in returns for the shareholders.
In addition, an investor in the index also received an annual dividend of 2% per year, for a total of 3.5+1.3+2.0=6.8% annual nominal return.
Again, the companies themselves grew their actual business (sales) by 3.5% per year, the sales valuation multiple expanded by 1.3 per cent per year (the companies became 1.3% more expensive per year, 1.3% more appreciated per actual generated dollar of revenue, investors came to expect more and more from the future from a company with a given business size), and finally, the investor holding the shares received an annual dividend of 2% per year. Thus, on average, 100 dollars invested turned into 106.80 dollars a year later. Not great, not terrible. Scarily enough though, 21 years later the resulting 400 dollars held less real purchasing power then the 100 dollars did back in 2000.
An investor holding the S&P 500 for the last 21 years lost purchasing power
It wasn’t just the S&P 500 that rose in price between 2000-2020. Apartment prices, groceries, tuition fees, mobile phones, TV-sets and so on rose too. Sure, the official consumer price inflation was much lower than 6.8% per year. But, Shadowstats.com has a true inflation estimate (8-9% per year) that easily exceeds the 6.8% figure, pointing to a negative real return on the S&P 500 index over the last 21 years. Yes, you should have bought gold instead :D. The inert yellow pet lump outperformed the most dynamic economic system in the world for a period of more than two decades of unprecedented technological progress!
How S&P gave you a loss of 7% per year for a decade, before deducting the erosive effects of inflation
An even more scary example comes from the 2000-2009 period, in which S&P revenues grew by a strong 4.9% per year, but the revenue valuation multiple plunged from 2.36 to 0.68. The resulting average total annual return including dividends for a holder of the index for those 9 years was -7.0%. Yes, per year. For 9 years. And that was despite the fact that the actual companies performed very well in the real world, growing their revenues by almost 5% per year on average. Why? Simply because paying 236 dollars for every 100 dollars in revenue was just too optimistic and sooner or later there would be a reckoning when a more rational valuation regime would prevail.
To add insult to injury, the P/S multiple then dropped all the way to a pretty cheap 68 dollars per 100 dollar in sales. Investors were at that point disillusioned, licking their wounds, and out of cash to invest. Many of those who recognized the bargains on the market just couldn’t invest. And some who had leveraged bets on were forced to sell at whatever bargain basement prices the few bidders had to offer.
What about the future for stock returns? Fundamentals, valuations and dividends?
What can we expect going forward? Well, both US nominal GDP and S&P revenues have grown by less than 4% per year over 10 and 20 years. Actually, structural* GDP growth has only been 1.6% per year (*real growth at constant employment rate). Adding two per cent inflation to a similar real growth number would thus only get us to 3.6% annual nominal growth.
Let’s now assume it takes a full decade from today for the Price/Revenue multiple to retreat to its peak valuation of 1.7 in 2007, right before the house bubble burst. Peak, not bottom. Nota bene that the calculation is based on valuations only retreating to what from a historical perspective was a pretty crazy high valuation peak driven by easy money in the housing bubble and very strong growth in China. The resulting annual total nominal return for that decade would amount to 3.6%+1.4%-6.0%=-1.0%.
That’s minus 1 per cent per year on average for a full decade! Pretty raw deal, especially considering the risk of valuation excursions down toward the historical average of 1.0x revenues, rather than just to the previous 2007 bubble peak of 1.7.
Oh yeah, let’s not forget to point out that that’s the nominal annual return in a 2% official inflation scenario, so the -1% nominal is -3% real total annual return including dividends. Per year. For a decade. And if shadowstats trends are anything to go by we can expect true inflation to be around 7-10% per year… Buying stocks for the long run at super high valuations certainly doesn’t sound fun at all.
What if multiples keep trending higher?
But what if multiples just keep expanding, trending higher? Could they? They’ve gone from 0.7 to 1.0 to 1.7 to 2.36 to 3.09, so why can’t they just keep expanding? There seems to be a trend in valuations from low to higher, right? The problem lies with how the next to last investor can rationally expect to find a last investor that similarly must pay a rational price that will give him a decent enough return on his money for him want to take the deal. Even at today’s 3x revenues that’s pretty rich, let alone after an additional increase in sales multiples.
This is why there is a limit to valuation multiples, you can always choose to do something else with your money that rationally promises higher returns.
If you pay 3x revenues for something with a 10% margin you’ll only get a 3% effective annual yield on your investment. At the historical norm of 1x revenues you got a 10% earnings yield. That’s what investors have always demanded for their risk, and that’s what they’ve always got when investing in times of average valuations.
Don’t you think there will be any alternative investments in the future with an expected earnings yield of more than 3%? I think there will be.
Only reliable indicators are useful
Another way to think about valuations is this: If you have a reliable valuation multiple, one that has a predictive value on actual subsequent returns, then a higher multiple means lower expected returns, and a lower multiple means higher future returns. Otherwise it isn’t reliable. And if it isn’t reliable there is no value in using it. In today’s markets, however, many seem to expect the same or even better returns going forward than those realized recently, despite the fact that valuation multiples are triple the norm now, and that recent returns in addition benefitted from that multiple expansion…, whereas returns going forward will feel an equal drag if multiples normalize.
Paying much over 100 dollars for a 100 dollar bill has always ended badly within 20 years
The last 21 years saw a 6.8% nominal annual return as the revenue multiple increased from an eye watering bubblish 2.36 to an incredulous 3.09. Now, real growth prospects in the coming decades are likely worse, due to slowing productivity, slowing population growth and more-er ginormous mountains of debt than before.
I can’t say how the story ends, but just because the market developed from manic to crazy over 20 years, definitely does not guarantee the same stick save going forward. Higher valuations have always been resolved with low long term returns within 20 years, and always in an interesting way, i.e., through fast crashes and interim bull markets on the way to nowhere. The clock is ticking. Somehow these 3x valuations will have to be resolved – through much higher prices on everything else, including company profits, or lower stock prices.
Let me make a few closing remarks:
There is no right or wrong in the above, as concerns the relation to the real world of investing. It’s just mathematics. Fundamentals and the market’s valuation of those fundamentals dance together over deep time.
Back in 1990-91 when the Price/Revenue was 0.7, there was no way to know that it would get back to or above it’s normal level of 1.0, or when, let alone climb all the way to a historical crazy peak of 2.36 by the year 2000. You took a risk buying at 0.7, but a pretty good and calculated risk, since 1.0 was the historical norm. Also, even if valuations stagnated at 0.7 you would still get a nice return on your money from fundamental growth and dividends.
And after that, although a few good men warned that a reckoning was likely based on the starting point of 2.36, you couldn’t know that 1.0 would yet again act as a magnet and cause a valuation pull-back all the way to 0.7 times revenues again in just 9 years, by 2009. However, a reckoning sometime was no doubt baked in the cake at 2.36 (although some today claim that today’s P/S of 3.1 proves that 2.36 never was expensive or a bubble)
And again, almost like clockwork, no matter what the economy does, what goes on with technology, geopolitics etc., another decade down the road (12 years this time), the revenue multiple had shot up from a depressed 0.7, hesitated slightly around the historically pretty expensive 1.3-level, before exploding to the utterly mind-blowing 3.1x revenues as of today in April 2021.
Many now claim that it’s a new paradigm, companies will grow faster and have permanently higher profit margins, thus warranting higher P/S multiples.
But, revenue growth is mediocre at best, and dividends are low. However, profit margins actually are pretty high now, but nowhere near levels that would warrant a P/S multiple of more than 1.5 rather than 1.0, even if the new profit margin level turned out to be a permanently higher plateau. The question is if employees are content with getting a lesser piece of the pie. Company profit margins are carved from employee wages. And will those lower wage shares be sufficient to support company revenues? Somebody has to buy the products sold at higher margins, and those somebodies just got less paid.
Let’s change tack slightly and talk about what you’re buying when you’re buying a stock. What kind of wealth does that holding represent, and why does it matter if you pay 100, 200 or 300 for a company generating 100 in revenue per year?
What is wealth? Stuff and profits, not prices
Wealth consists of assets, agreements, cash flows etc. The price of an asset doesn’t really affect it’s actual wealth value. If the price of a stock with an annual profit of 10 doubles, your wealth is still the same, not 10x. The price could triple, quadruple, rise tenfold if you like, but your real wealth would still be constant, i.e., your wealth is simply your right to the annual cash flow of for example 10$. Sure, if you manage to sell the share for 1000$ rather than 100$, your wealth really is those 1000$ after the transaction has closed. But the wealth of the poor schmuck who bought your shares just changed by the difference between 1000$ and what the annual cash flow of 10$ is really worth. I’d say your wealth rose by 900$ and his fell by 900$. It was a wealth transfer from the one selling at 10x valuations to the guy buying at high valuations.
The real value of the S&P 500 most likely is a pretty stable multiple of its annual revenues. At least, the last century the market agreed that multiple on average was 1. Sometimes half that, sometimes twice that average.
A very simplified way to see the logic in value=1x revenues, is that with a 10% profit margin, sales of 100 equated to a profit of 10. If you paid 100 for annual sales of 100 and thus annual profits of 10, you got a 10/100=10% annual return on your investment.
A 10% return is a nice round number; e.g., it fits our human hands with ten digits, and we can quite effortlessly see how it’s worth our while to forego consumption, to assume some uncertainty and volatility etc. for the expected clearly visible improvement that 10% per year is. At 5% the deal feels less obvious. How many years of no fun is it really worth for maybe a 1/20th improvement on average per year? Such return prospects might instead inspire investing in education, your own business or plain present consumption. Consumption also has a time value; it’s worth more the earlier you do it. What you do in your 20s and 30s is potentially more memorable than what your spending gets you in your 70s and 80s. And you can savour the memories for much longer.
At 15% annual returns or more, the investment bargain soon becomes apparent for most. It’s time to buy with both hands. Unfortunately those opportunities only arise when most people are on their knees due to debt, stingy banks, and unemployment. Cash is king when there’s blood on the streets, because almost nobody has access to cash.
I probably lost you there, so here it comes again.
The warranted revenue multiple was 1.0 for a century, now it’s over 3!
If the world is basically the same today as it has always been: about the same growth, the same hopes and aspirations, the same natural return requirements, the same cyclicality, the same profit division between owners, governments and employees over long periods of time (the several-decades period that the stock market’s duration is) then the real value of the S&P 500 is the same revenue multiple as it has always been. Up until five years ago, the market had agreed for the full century of the existence of the Federal Reserve Bank, that that multiple was 1.0. Now it’s 3.1. Is the future of profit growth going forward the coming 50 years really that much better than it was 50 years ago? Not faster population growth, not faster productivity growth… Those trends are pointing down not up.
Either the market was insanely cheap for a century, or it’s insanely expensive now, or the real value of revenues have recently expanded threefold.
Let’s look at the evidence. Over the last century, when you bought the S&P at 1x revenues you realized 10% annual nominal returns going forward. Good, ok, not bad, not brilliant, not obviously cheap or rich, but what one might reasonably demand for the trouble. What you got was the sum total of revenue growth (about the same as GDP growth, i.e., population growth times better production methods, plus inflation) and dividends. If you paid 2x revenues you got a lot less, typically first a nerve wrecking period of sharply negative returns. And if you had the nerve to snatch some shares at 0.5x you made a killing.
Today you’re paying 3.1x revenues. 6x what the offer sometimes is in bad times. But the underlying value probably still is around 1.0. At least that’s where you can expect 10% annual returns. Here at 3.1x, even if you fantasize that 3.1x will be a permanent plateau, you’ll still only get 5% per year (fundamental growth+dividends).
And those 5% per year are in serious harm’s way if the 3.1x multiple would budge even slightly, not to mention approach more rational and historically normal levels. As a reference point, if the S&P 500 dropped by 70% tomorrow, the index would still only be about fairly valued, promising some 10% average annual returns going forward from there. Those are nominal returns, before any purchasing power erosion by inflation.
I’m not forecasting an imminent stock market crash. If valuations were timely predictors of downturns, valuations would never reach lofty levels. All rational investors have to learn and re-learn this fact. However, although I personally am almost fully invested at this extreme point in market history, I urge caution. At market extremes your actions should be correspondingly extreme. When stocks are at 3x their historical norm you should be down close to your lowest conceivable level of shares in your investment portfolio. You should similarly be closer to your maximum conceivable level of cash or alternative investments.
Max and min levels are individual. Some people consider 100% stocks their lower bound (!), others employ a 50% level and yet others all the way down to zero or below to negative numbers (net short). But whatever your theoretical minimum risk level is, you should consider going there now when valuations are at all time highs.
However, money printing complicates things. Cash is not a low risk alternative anymore. Your loss of purchasing power is 10-20% per year. Gold, stocks and house prices rise by 10-20% per year. Even consumer prices (shadowstats) rise by 10% a year.
At this point you need to protect your wealth against both a valuation compression and price inflation. There are no surefire ways out of this, definitely not for all of us. On average asset prices will fall to their real value, but a few of us might escape the worst effects by hiding our wealth in real things. Businesses are such things of course. Better than cash over long periods of time. But at 3x typical valuations the risk of an initial drop is much higher than the protection you get against a decade of inflation.
The answer lies in real, scarce assets. But what are those? How should you manage your money when money comes for free (unfortunately, however, mostly only for bankers)? That’s a story for another post (Hint: gold, Bitcoin, Ether, various DeFi projects, farm land, real estate…)
The value of the S&P index used to be 1x the sum of future dividends discounted at 10%. Now the value is 3.5x the sum of dividends at 10% discount rate