What determines stock market prices

• Contrary to theory, the market P/E level does not primarily reflect future prospects. It reflects current conditions.

• The variables it weights heavily are not academically or economically correct, but those that make investors feel comfortable.

• High profit margins and stable, low inflation dominate this feel-good list, with stability of GDP growth (as opposed to actual growth) a distant third.


• Investors’ extreme preference for comfort, like human nature, has never changed. (Tested back to 1925.) This is unlike financial and economic conditions, which have very substantially changed in the last 20 years.

• The ebb and flow of these variables explain previous market peaks and troughs. These comfort factors, for example, have been at an extremely high average level for 20 years (as have P/Es) and remain so today. Thus today’s high priced market is the completely usual response from investors.

• Any shift back to a lower P/E regime must therefore be accompanied by a major sustained fall in margins or a sustained rise in inflation (or both).

• And, yes, I do believe these comfort variables will move to be less favorable. But probably not quickly.

Why stock market valuations are high

Introduction

In an investing and economic world in which almost everything seems to have changed in the last 20 years, one thing has remained constant: human nature. And, we can more or less prove it. At least in the case of the stock market.

The behavioral drivers of P/E ratios

Ben Inker and I designed a simple model 15 years ago to explain the shifts in P/E levels of the S&P 500. Recently we updated it. Our model does not attempt to justify the P/E levels as logical or deserved, nor does it attempt to predict future prices. It just shows what has tended to be the market’s typical response over the years to major market factors. By far, the two most important of these are profit margins, the higher the better, and inflation, where stable and lower is better, except not too low.

The market’s responses are typically quite different from what one might expect from an efficient market. Let us start with profit margins. In a rational world, stocks should sell at replacement cost. And therefore, above-average margins require a below-average P/E. An alternative way of viewing this is that above-average margins can be expected to mean revert – to move back to average – and vice versa. In real life, though, even in the past when margins were provably very mean-reverting, the market has always preferred high margins. Investors would dependably pay up for high margins, which would then decline, whacking them on the way down. Subsequently, investors would avoid markets with low average margins, which would then recover, causing them to under-perform. Over and over again. This, in short, was primitive double counting. In 1974 or 1982, for example, at cosmic market lows, very, very depressed margins would sell at equally depressed P/Es of 6x or 8x, producing a price to book (or Tobin’s Q) of one-quarter or one-third. Conversely, at market peaks like 2000, record margins were multiplied by record P/Es, producing 3x price to book. This double counting is a major market inefficiency, perhaps the major inefficiency, and the main reason why market volatility is many multiples of the stately moves in the theoretical future value of dividends. To sell at fair value – true replacement cost – profit margins must be negatively correlated with P/Es, yet in the real world they are strongly positively correlated. And always have been.

Years ago Robert Shiller pointed out that the true discounted value of a future stream of dividends was very, very stable and had been since 1882 at least. Yet we investors, transfixed apparently by short-term events and eager to double count in the manner just described, overreact in what might reasonably be called a hysterical manner. The result is about 18x more volatility for the market than is justified by the fundamentals! And apparently we are incapable of learning this from experience. So much for efficiency. As always, it seems human behavior, sometimes rational and sometimes not, trumps efficient economic behavior.

Then there’s inflation, which the market hates despite a history of stocks proving that their fundamentals are robust in the longer term in passing inflation through to consumers. Stocks, unlike bonds, are clearly real assets, so inflation should not matter. But in real life, when inflation first appears or accelerates, there is an immediate, coincident negative effect on P/E multiples. (Modigliani, one of the very few economists who seemed to understand how inefficient the market was capable of being, visited a Boston brokerage house in the depths of 1974, when the market was 7x, and made this point eloquently over lunch. He explained that because inflation was irrelevant to long-term value, the market should always be at replacement cost, or twice the then price. Which it obediently moved to in the following two years. Not a bad call. Unfortunately, he seemed to be uninterested in doing serious academic work on the stock market. Had he done so, the belief in market efficiency might have been less of an expensive and dangerous proposition for society. Counting on rational behavior – or even reasonable behavior – from an investment world that overresponds to inflation and can’t even get the sign right for the effect of profit margins can be expensively misleading. As Kindleberger said, this kind of faith in market efficiency “ignores a condition for the sake of a theory.”)

A third behavioral factor, which we first modeled 16 years ago and still has explanatory power, although much less than the first two, is the volatility of GDP growth. Notice that this is absolutely not the growth rate of GDP. I have spent a few decades with traditional portfolio managers and I can guarantee that they are made nervous by rapid GDP changes, even on the upside. Changes make them fear more bumps in the road than normal. Rather, they love stable GDP growth, with few surprises, where they can feel more in control of their own predictions. And in such a world P/Es tend to be higher when GDP growth is stable, not when it is high. (By a strange coincidence, on June 28 The Wall Street Journal presented a chart showing global GDP volatility that has this very last month as the lowest in their 44 years of data!)

During the updating last year of our behavioral P/E model, we added two new factors, which we believe provide a little further value. The first addition was an on-off switch, which causes P/Es to be a bit lower after a down quarter. Again, not a very scientific reflex in a mean-reverting world, but understandable. The second was to add the U.S. 10-year bond rate, where higher rates are negative, modestly improving the model even after the inflation component had done the heavy lifting for nominal interest rates.

The overall fit is pretty good, but it is much better than that at extremes. In the past, the model explained a record high P/E in 1929, extreme lows in the 1930s, another high in 1965, extreme lows in 1974 and 1982, and a world record high, by a wide margin, in 2000. The outlier nature of the market peak in 2000 – confirmed in our updated version – suggests something interesting: Market peaks (1929, 1972, 1987, and 2008) and market troughs (1932, 1974, 1982, 2002, and 2009) were more or less – and more rather than less – what you should have expected from investors pressured with the comfortable (and uncomfortable) news at the time, as reflected in our model. For those events, you don’t need any further bubble or bust explanations. They were normal behavioral responses to extreme data. This behavioral model seems, in that sense, like a parallel universe to all our other investment thinking on bubbles. Only 2000 stands out, fully one-third higher than “explained,” as a genuine bubble, beyond these normal factors. The reader can surely pity investors who were hung out to dry in that extra and extraordinary one-third! Even the memory still stings.

We might reasonably conclude from this finding that any large and more or less permanent decline in the market (i.e., to a new, lower trend, much more like the 1945 to 1995 period than today) would require an equally large deterioration in profit margins or increase in inflation or some combination. Without either, any large market decline would be very unusual historically and likely, I believe, to be temporary. I can conclude this point by offering my personal opinion for 2017 on the two most important factors: favorable for margins and unfavorable for inflation. If only life were easy! But, even if these guesses prove to be correct, this mixed signal does not suggest a major decline or perhaps any decline.

For the record, if you need yet another rebuttal of the Lucas/Fama and French model of economic efficiency on the part of investors, this model is it: a long-term testimonial, and a very stable one, to investor behavior that they would have to describe as inefficient by their definition. And though this investor behavior may be loosely described as rational, it is certainly economically and financially innumerate. I am happy to say that I never believed a word of their theory on the efficiency of the market, which I have always thought is better described as a behavioral jungle. But having said that, I must admit to having detracted from my usefulness as an investor by assuming that investors overall would at least respond sensibly most of the time to the data they are given. And they do not. The effectiveness and persistency of our behavioral model, almost all the components of which should not work in a resolutely sensible world, let alone an efficient one, should have persuaded me to change my thinking years ago. But, here I am, trying to explain during these last nine months or so why the general discount rate of assets has dropped by roughly two percentage points from the 1900 to 1997 average. My proposed reasons for the reduced discount rate come down to a complicated stew of factors, most of which interact with the others: higher profit margins; higher leverage; lower rates; aggressive Fed and central bank policies to push rates lower; moral hazard from the Fed that may be more important than rates – the asymmetry of helping in bad times and letting good times run; changes in the age profile of the developed world; slower population growth; lower productivity; lower GDP growth; less low-hanging technological fruit; loss of the old $16 a barrel oil; extreme income inequality; remarkable lack of progress in median and lower hourly wages; and very much enhanced corporate political and monopoly power. Phew! I truly believe we will never know for sure which factors dominate the equation, although my favorite is Fed policy and the runners-up are an aging population profile and the rising political and monopoly power of corporations. But whichever they were, they got the job done: For the last 20 years profits in the U.S. as a share of GDP and corporate revenues rose by about 30% and P/E ratios by 70% above the old normal.

Now, cutting across that previous attempt to understand these major changes in our new 20-year era, comes an entirely behavioral approach. Whether sensibly or not, investors love high margins and like stable growth even if it’s modest, and hate inflation. They felt this way from 1925 to 1997 and they felt exactly the same way in our new era of 1997 to 2017. So, behaviorally it is absolutely not a new era. It is precisely – to a 0.90 correlation – the same ole same ole. The peaks of 1929 and 1965 delivered favorable margins and inflation inputs but for a very short while in both cases. In contrast, the period of 1997 to 2017 has delivered to investors their preferred conditions almost the entire time, with only two very quick time-outs for market breaks. Can the market really be this easy to explain? Well, it has been for 92 years! And what can we investors do with this information? It tells us that if we re-enter a period of old normal profits and old normal inflation, the market’s P/E will indeed mean revert to its old average. And if we don’t re-enter such a period, the P/Es are likely to stay high. It tells us separately that if we expect a market crash, we should also expect to have a crash in margins (as we did in 2008-09) or a truly dramatic rise in sustained inflation (as we did in 1979-81) or some powerful combination. All of which is possible of course, but I think improbable, at least in the near term. This behavioral approach to explaining shifts in P/Es is certainly a much simpler equation than my previous stew-of-factors approach. But it does have some powerful similarities to my earlier arguments found in Parts 1 and 2 of “Not With A Bang But A Whimper. In both approaches, the role of profit margins is dominant. Improved margins not only move the earnings up directly, but also the P/E multiplier applied to those earnings. Inflation is also a strong secondary factor in both approaches, for low inflation, of course, drives down the interest rates, which appear to be an important ingredient in the stew.

So, where does this leave us? It suggests to me that I have in general been over-intellectualizing the working of the market for a few decades. I have had too strong a belief that investors would at least be influenced by past data in a sensible way. The market, however, appears not to care at all about the past or to learn much from it. This model for sure seems to say that for 92 years, at least, the market has with remarkable consistency been a coincident indicator of superficially appealing variables that in a strict economic sense have been inappropriate, and that have caused spectacular and unnecessary market volatility. The model is apparently a reflection of human nature and, of all factors influencing the market, human nature, as economically inefficient and unsophisticated it may be, seems the least likely to change.

Postscript No. 1: momentum and value

If the short-term behavioral variables described above dominate the short-term market level to the degree shown – a 0.90 correlation – what is the role for both momentum and value? My guess is that there is enough noise in the data for there to be room for many individual stocks to be driven in the short and intermediate term away from fair value by momentum, also a behavioral factor. More importantly for us value managers, I believe there is also enough noise in the data for individual stocks and the market to be pulled back toward replacement cost or fair value. Value (like gravity in physics) is a weak force in the short term, but very, very persistent so it can eventually work its way around the stronger coincident behavioral forces. Value was the beneficiary, after all, of a real world arbitrage: If the market priced a stock too high, management would sell stock and buy more – say, fiber optic cable – until market forces brought the price of the product, the profit margins, and the stock price down. If priced too low, management would buy stock back and reduce the underpricing and, more directly, it would withhold expansion until shortages occurred, sometimes quickly and sometimes drawn out. But more recently, increased monopoly power and other factors appear to have decreased the corporate reflex to expand in favor of stock buybacks, perhaps weakening the previously reliable game. But that of course is the question under consideration.

Postscript No. 2: a paean to changing your mind when the facts change

Descending out of the blue just before the July 4 weekend came this no doubt heaven-sent recently discovered talk by Ben Graham titled “Securities In An Insecure World,” which was given on November 15, 1963. It is one of the last he gave. After a 40-year career in which he had developed a fairly high-confidence view on how to define the value of the U.S. stock market, he was having second thoughts. Like some of us now. Here are my favorite snippets.

“In early 1955 when I testified before the Fulbright Committee the stock market was then about 400, my central value was also around 400 and the valuation of other ‘experts’ using other methods all seemed to come to about that level. The action of the stock market since then would appear to demonstrate that these methods of valuations are ultra-conservative and much too low, although they did work out extremely well through the stock market fluctuations from 1871 to about 1954, which is an exceptionally long period of time for a test. Unfortunately in this kind of work, where you are trying to determine relationships based upon past behavior, the almost invariable experience is that by the time you have had a long enough period to give you sufficient confidence in your form of measurement just then new conditions supersede and the measurement is no longer dependable for the future.” [Emphasis added. By the way, the deliberate total return from the S&P 500 from the end of 1963 until today has been 5.75% real, exactly what we at GMO assume to be the long-term, normal return.]

“My reason for thinking that we shall have these wide fluctuations – of which we had a taste in 1962, in May particularly – is that I don’t see any change in human nature vis-à-vis the stock market which is sufficient to establish more restraints in the public behavior than it showed over so many decades in the past.”

“But let me point out ‘for the record’ that it is not impossible in theory that the market’s high level alone could sooner or later precipitate a collapse without the necessity for these technical weaknesses [described above, Ed.] to show themselves. The collapse might be triggered by some untoward economic or political development. But if things do happen that way it will be the first time in market history, I believe, that we would have the end of a bull market without the excesses and abuses of the sort I have mentioned.” 

“The main need here is for the investor to select some rule which seems to be suitable for his point of view, one which will keep him out of mischief, and one, I insist, which will always maintain some interest in common stocks regardless of how high the market level goes. For if you had followed one of these older formulas which took you out of common stocks entirely at some level of the market, your disappointment would have been so great because of the ensuing advance as probably to ruin you from the standpoint of intelligent investing for the rest of your life.”

US stocks to go higher until elections

On the investment front the equation remains the same: pushing stock prices higher are the twin forces of the Fed's policy and corporate buybacks. Trying to push prices down is an impressive array of everything else: disappointing productivity, growth, and profit margins together with all our domestic and international political uncertainties. And now Brexit! It is a testimonial to the strength of those two bullish forces that they can steady the US market near its high, regardless, apparently, of what is thrown at it.

I therefore remain, on the basis of those two remarkable pillars of support, for at least one more quarter, where I have been for the last two years; despite brutal and widespread asset overpricing, there are still no signs of an equity bubble about to break, indeed cash reserves and other signs of bearishness are weirdly high.

In my opinion, the [US] economy still has some spare capacity to grow moderately for a while. All the great market declines of modern times – 1972, 2000, and 2007 – that went down at least 50 per cent were preceded by great optimism as well as high prices. We can have an ordinary bear market of 10 per cent or 20 per cent but a serious decline still seems unlikely in my opinion.

Now if we could just have a breakout rally to over 2300 on the S&P 500 and a bit of towel throwing by the bears, things could change. (2300 is our statistical definition of a bubble threshold.)

But for now I believe the best bet is still that the US market will hang in or better, at least through the election.

GMO downsizes reducing staffing levels

Boston-based Grantham has cut about 10% of its 650-person staff, according to Bloomberg. The layoffs include nine stock and bond analysts. The firm's assets have fallen about 20% in the past two years, to $99 billion, according to Morningstar.

In addition to the layoffs, several members of top management are leaving Jeremy Grantham's GMO. David Cowen, head of global equity, and Chris Fortson, head of fundamental research in global equity, will leave GMO at the end of the month. 

Jeremy Grantham Letter - Q1 2016

The tone of the market commentators back in January, when I was writing my last quarterly letter, seemed much too pessimistic on global stock markets, particularly the U.S. market, and I said so.

This relative optimism was an unusual position for me and the snap-back in these markets has validated, to a modest degree, my thinking at the time. I still believe the following: 1) that we did not then, and do not today, have the necessary conditions to say that today’s world has a bubble in any of the most important asset classes; 
2) that we are unlikely, given the beliefs and practices of the U.S. Fed, to end this cycle without a bubble in the U.S. equity market or, perish the thought, in a repeat of the U.S. housing bubble; 
3) the threshold for a bubble level for the U.S. market is about 2300 on the S&P 500, about 10% above current levels, and would normally require a substantially more bullish tone on the part of both individual and institutional investors; 
4) it continues to seem unlikely to me that this current equity cycle will top out before the election and perhaps it will last considerably longer; and 
5) the U.S. housing market, although well below 2006 highs, is nonetheless approaching a one and one-half-sigma level based on its previous history. Given the intensity of the pain we felt so recently, we might expect that such a bubble would be psychologically impossible, but the data in Exhibit 1 speaks for itself. This is a classic echo bubble – i.e., driven partly by the feeling that the substantially higher prices in 2006 (with its three-sigma bubble) somehow justify today’s merely one and one-half-sigma prices. Prices have been rising rapidly recently and at this rate will reach one and three-quarters-sigma this summer. Thus, unlikely as it may sound, in 12 to 24 months U.S. house prices – much more dangerous than inflated stock prices in my opinion – might beat the U.S. equity market in the race to cause the next financial crisis.

In the meantime, however, we continue to have the typical equity overpricing that has characterized the great majority of time since the Greenspan regime introduced the policy of generally pushing down on short-term rates. At current prices it is very close to impossible for the mass of pension funds or other institutions to realize longer-term targets of 5% a year, let alone 7%. A 60% stock/40% bond portfolio would be lucky to deliver 3% even if we were to lock in current high P/Es and above-normal margins. Prudent managers will just have to grin and bear it. The worst argument is always that extra risk has to be taken because the need to deliver higher returns is desperate. The market does not care what your targets are! And, in any case, there is a very wobbly relationship between risk and return, as the 50-year underperformance of high-risk equities to lower-risk equities attests. (And this underperformance applies to a varied definition of risk: Table 1 shows the gap between the top and bottom 20% in the U.S. equity market for volatility, beta, and fundamental quality – stability of high return and low debt.) My point over the last two years has been to emphasize how long and painful the grinning and bearing can be, and I firmly believe that investors should be prepared for considerably more pain (of overpriced assets becoming yet more overpriced) without throwing in the towel 1999-style at the worst time possible time.

Oil

As with stock prices, I am encouraged by the rise in oil prices since my unusual bullishness of last quarter. My belief remains that a multi-year clearing price for oil would be the cost of finding a material amount of new oil. This appears to be about $65 a barrel today, and costs are drifting steadily higher as the cheapest old oil is pumped. My guess is that the price of oil will indeed be as high as $100 a barrel again within five years and, perversely, I feel encouraged by the growing host of longer-term pessimists. Much as I would love as an environmentalist to have oil for transportation almost disappear in 10 years, it simply isn’t going to happen. However, there will be dramatic technological improvements in non-oil based transportation well before 10 years is out, and after 10 years… well, the oil companies will wish they had taken Ted Levitt’s advice in 1960 in his then groundbreaking “Marketing Myopia” article and defined themselves as energy and chemical companies and not defended their narrower definition of “oil companies.”

New technologies for energy and transportation

The new discoveries and engineering insights in these fields keep coming. The Grantham Foundation’s attraction to venture capital, described last quarter, plays conveniently into this and we are investing 20% of our total capital in “mission-driven” projects, mainly in venture capital. One investment is aiming at doubling the power-to-weight ratio of lithium ion batteries and another at a 200-second charge time for some types of lithium ion. Vehicle development we see includes ultra-light and ultra-streamlined electric-powered people movers, suitable for commuting and shopping in developed countries, and as a first cheap yet state-of-the-art vehicle everywhere, very fast and with a long range. (Better make hay soon, Tesla!) But, you never know. And venture investing is particularly full of disappointments. So, just in case, I am one of the 400,000 preorders for the $35,000 Tesla Model 3, likely to be the cheapest electric car per mile of range at least for a while. The mission component on some of our investments is so on target that even if they fail they represent attempts that will have deserved some of our non-profit making grants. As it is, we hope that at least one of the several game-changing projects driving these CO2-reducing technologies forward will simultaneously make our foundation a fortune!

Global warming accelerates: “so much for the pause”

Because 1998 was an outlier warm year due to a large El Niño effect in the Pacific, many subsequent years, including 2013, had lower global temperatures and led some to believe, or claim to believe, that global warming had ceased. But it turned out to be, after all, just another series like that of the S&P 500 in real terms with a little steady signal often obscured by a very great deal of noise. As it turned out, the below-trend 2013 was followed immediately by a modest new record in 2014. And then came the real test as a new powerful El Niño started to build up in 2015. Ten of the twelve months of 2015 set new all-time records, an unheard of event, and 2015 in total became a monster, not only the warmest year in recent millennia but by a record increment. Yet, the early months of 2016, still under the influence of what had become one of the most powerful El Niño effects, showed temperature increases that were even more remarkable.

This current El Niño has accelerated underlying warming caused by increasing CO2 – as all El Niños do – but this time the combined effect has been far ahead of scientific forecasts that in general remain dangerously conservative. January 2016 was the hottest January ever on the NASA series and by a new record amount. It was a full 0.22 degrees Celsius above the previous high for January. Then February became the new shocker, washing away that record by being 0.33 degrees Celsius above the previous February record. Most recently, March was once again the warmest ever March, although not quite by a record amount (see Exhibit 2). The exhibit makes the scary point clear: global temperature is not just increasing, but accelerating. The average increase from 1900 to 1958 was about 0.007 degrees Celsius per year. From 1958 to 2015 it doubled to 0.015 degrees Celsius per year, and from February 1998 to February 2016 it rose by an average of 0.025 degrees Celsius per year! Time is truly running out.

Sadly, it has become obvious that the recent talk in Paris of limiting warming to 1.5 degrees Celsius is toast, as it were. And the dreaded 2 degrees Celsius is highly unlikely to be the limit of our warming. If you line up the previous El Niño outlier of 1998 with this March 2016 El Niño (as we might do in lining up bull market highs) it gives an idea of when 2 degrees Celsius might first be broached in a future El Niño effect: just 17 years! Meanwhile, the most obvious effect to watch for in destabilizing weather patterns is an increase in record breaking, intense rainfall, such as occurred last month in Houston. Three inches an hour1 fell and kept falling hour after hour, delivering four months’ average rain in under 24 hours (unprecedented without a major hurricane), flooding major parts of the city under several feet of water.

Let me just make the point here that those who still think climate problems are off topic and not a major economic and financial issue are dead wrong. Dealing with the increasing damage from climate extremes and, just as important, the growing economic potential in activities to overcome it will increasingly dominate entrepreneurial efforts in future decades. As investors we should try to be prepared for this.

Jeremy Grantam focuses on Venture Capital

Venture capital has always been a high-risk, high-reward  proposition. Its current surge is finding an unlikely proponent in legendary investor Jeremy Grantham. 

The billionaire has a worthy cause in mind: his Grantham Foundation for the Protection of the Environment. My foundation is aspiring to an extremely large 40 percent allocation to venture capital, because that is where I see the greatest potential, and we certainly need all the money we can
make trying to protect the environment, Grantham says in his office at the Boston headquarters.

The U.K. native launched his $617 million foundation in 1997 to concentrate on Climate Change and Agriculture. In February it announced its involvement in a new $430 million fund with Boulder,
Colorado based Vision Ridge Partners and Capricorn Investment Group of New York that will focus on clean-energy companies and other sustainable assets.

Venture capital is uniquely suited for mission-driven, or impact, investing because investors can give early support to companies that share their values and help to bring those businesses and values to market, says the slender, sharply dressed Grantham.

US markets can rally higher before it becomes a full bubble

Looking to 2016, we can agree that uncertainties are above average. I must admit to feeling nervous for this year’s equity outlook in the U.S. But I am not entirely convinced. Sure, we can have a regular bear market. That is always the case. But the BIG ONE? I doubt it..

But I think the global economy and the U.S. in particular will do better than the bears believe it will because they appear to underestimate the slow-burning but huge positive of much-reduced resource prices in the U.S. and the availability of capacity both in labor and machinery.

The ability of the market to hurt eager bears some more is probably not exhausted. I still believe that, with the help of the Fed and its allies, the U.S. market will rally once again to become a fully-fledged bubble before it breaks. That is, after all, the logical outcome of a Fed policy that stimulates and overestimates some more until, finally, some strut in the complicated economic structure snaps. Good luck in 2016.