This piece was originally published in Barron's Wall Street's Best Minds column, November 19, 2013. The S&P 500 closed that day at 1,787.87, 130% of which is ~2325 vs yesterday's close of 1950.82 :
By Jeremy Grantham
November 19, 2013
Economics is a very soft science, but it has delusions of hardness or what has been called physics envy. One of my few economic heroes, Kenneth Boulding, said that while mathematics had indeed introduced rigor into economics, it unfortunately also brought mortis.
Later in his career he felt that economics had lost sight of its job to be useful to society, having lost its way in a maze of econometric formulas, which placed elegance over accuracy.
At the top of the list of economic theories based on clearly false assumptions is that of Rational Expectations, in which humans are assumed to be machines programmed with rational responses. Although we all know – even economists – that this assumption does not fit the real world, it does allow for relatively simple conclusions, whereas the assumption of complicated, inconsistent, and emotional humanity does not. The folly of Rational Expectations resulted in five, six, or seven decades of economic mainstream work being largely thrown away. It did leave us, though, with perhaps the most laughable of all assumption-based theories, the Efficient Market Hypothesis (EMH).
We are told that investment bubbles have not occurred and, indeed, could never occur, by the iron law of the unproven assumptions used by the proponents of the EMH. Yet, in front of our eyes there have appeared in the last 25 years at least four of the great investment bubbles in all of investment history.
First, there was the bubble in Japanese stocks, which peaked in 1989 at 65 times earnings (on their accounting) having never peaked at over 25 times previously, to be followed by a loss of almost 90% in the MSCI Japan index! Second, we had the Japanese land bubble peaking a little later in 1991.
This was probably the biggest bubble in history and was certainly far worse than the Tulip Bubble and the South Sea Bubble. And, yes, the land under the Emperor's Palace, valued at property prices in downtown Tokyo, really was equal to the value of the land in the state of California. Seems efficient to me … California is so big and unwieldy. Next, we had by far the largest U.S. equity bubble in 2000, which peaked at 35 times earnings compared to a peak of 21 times in 1929, yet had had previous growth rates less than half of those in 1928 and 1929.
Finally, we had what I described in 2007 as the first truly global bubble. It covered all global stocks, fine arts and collectibles, and almost all of the real-estate markets. The last of these was led by the U.S. housing market, which, having benefited from its great diversity, had historically been remarkably stable until Greenspan got his hands on it.
Compared to previous ultra-stable data, this measured as a 3.5 sigma event, which, according to the EMH assumption of perfect randomness, should have occurred only once every 10,000 years. Yet, encouraged (brainwashed might be a more accurate description) by the EMH, Bernanke (and Yellen) could not, or would not, even recognize the risk, to our very substantial cost.The edifice of unproven and totally inaccurate assumptions represented by the EMH was not defended as having useful output, but was defended, especially by the high priests of the theory, as being the most accurate reflection of reality that could be rolled out.
Thus, 12 reasons were given as to why the 22% drop in one day (over one-third in a day and a half) in the 1987 crash was a rational economic response to a suddenly changed world. (My then partner, Dick Mayo, typically immersed in market trading 10 hours a day, was indeed braced that day for remarkable events.
He, like other professionals, had paid no attention to any of these reasons, but had focused on the new, unproven so not appear on the list of 12!) Similarly, Japan at 65 times earnings was justified by counting errors and low interest rates, and the Tech Bubble by Greenspan's "permanently" increased productivity that appeared everywhere except in the data, which showed a subsequently reduced level of productivity. I must admit that I have never heard an EMH rationale for the Emperor's Palace, but you get my point. It has also been suggested that Fama's work led to indexing. Not really.
When we offered indexing at Batterymarch in 1971 we did so because we knew it was a zero-sum game. That for us was a complete and sufficient reason for indexing: active managers summed to market returns less large fees and commissions while indexers summed to market returns less small fees. To prove our belief, we simultaneously ran an active portfolio that ended its first eight years – a random number selected to coincide with my stay there – up 7% a year relative to the S&P. Batterymarch more or less shared the indexing business in its first few years with Wells Fargo, with the considerable propaganda skills of Dean LeBaron, our senior partner, more or less offsetting their huge size. They however did talk about the market's efficiency, which, particularly back then, only existed in the minds of a few professors and apparently one or two academically inclined Wells Fargoans. To nail home this point, Jack Bogle's Vanguard Index Fund in 1975 was, like us, also emphatically based on the concept of a zero-sum game and the certainty it offered that most players would underperform. None of this efficient market nonsense was detrimental to us value managers so I should find time to thank all those involved for producing and passionately promoting the idea. During the 1970s and 1980s I am convinced it helped reduce the number of quantitatively-talented individuals entering the money management business.
Why waste your Ph.D in particle physics on an efficient market? The field was left for an extra 20 years or so to very ordinary, not particularly quantitatively-minded individuals. And very nice it was too. And, by the way, the proponents of the EMH not only promoted their theory, but via the academic establishment the high priests badgered academic researchers into leaving, resigning themselves to nontenure, or getting religion, as it were. Into this morass of false assumptions there did come a ray of light back in 1981, when Robert Shiller proposed a simple test of market inefficiency. He assumed total clairvoyance and asked the question: What were markets worth back in, say, 1880, 1915, 1961, etc., if you knew both the long-term market return, or discount rate (in the 6% to 7% range after inflation), and, more importantly, you also knew the complete and accurate future stream of dividends? Even on its own the dividend stream really is fairly smooth, but because the market is worth the sum of all future discounted dividends, it becomes remarkably smooth so two-thirds of the time the "fair value model," as we'll call it, is within ±1% of its long-term trend. The trend turns out to be about a rather modest 1.5% real. (The relatively stable series of GDP is put in here for reference. It is within ±4.5% two-thirds of the time.) The red series is what we emotional and career-protective investors do to this stable world. The S&P 500 is within ±19% of its trend two-thirds of the time! This almost ridiculous volatility, 19 times more than is really justified by the underlying fundamentals, is in my opinion caused mainly by individual investors driven by behavioral factors that result in herding: non-experts simply feel more comfortable in a herd and, indeed, "prudent" investing has long been legally defined as doing what others do. For professional investors it is caused by the need to report upwards to decreasing levels of market familiarity – the top decision-makers usually look and feel very much like individual investors (with some very notable exceptions) and impose their will on the institutional investors, whose number one imperative is to keep their jobs. Being wrong on your own, as Keynes describes so eloquently in Chapter 12 of The General Theory, is the cardinal crime for an investment manager. To avoid this, the professionals try very hard to ensure that if they are going to run off any cliff they will: a) have a lot of company; and b) that most of the company will be one step ahead. In short, the management of career risk results in very destructive herding. It also produces a great deal of extrapolation, also designed to protect their careers. After all, if you make a forecast, Lord knows, you can be wrong. So, instead you use extrapolation, which Keynes said is a convention we adopt even though we know from personal experience that it is not applicable in the real world. (Keynes was 47 years ahead of all other economists in understanding markets. And counting, for they are still nowhere near catching up.) Going back to Shiller's simple experiment: was it enough to cause a pause in the march of market efficiency? Not for a second. It was batted away like a bothersome fl y for reasons that I cannot explain with a straight face.
This brings up an interesting question. What would it take to become as volatile as the market? I believe there are only two choices here, one bad and one good. The bad one is typically self-referential: that investors change the discount rate minute by minute. In this way we would have to believe that they did not panic in 1987 as portfolio investors all tried to squeeze through the small door of the burning theatre, but that they merely reassessed the distant future that morning as very much less attractive than the day before! Whatever the ridiculous move in the market, the math will "solve" with an equally ridiculous change in the assumed discount rate. The second theory, though, is compatible with everything I know about the market (or what I accepted from Keynes): that extrapolation dominates the workings of the market.Let me briefly give you my prime evidence. I will spare you one of my favorite exhibits in the interest of saving space. What it would show is that the 30-year U.S. government bond peaked in 1982 at a 16% yield, because inflation had spiked for a second to 13% (even though Paul Volcker was already on the anti-inflation warpath). Yes, you might expect the T-Bill to be 14% or so, which it was. But a 30-year bond!
To extrapolate a full 13% inflation – a complete outlier event, by its very nature bound, kill or cure, to be temporary – for a full 30 years! More recently, of course, we extrapolate currently very low inflation for 30 years. My case rests....MUCH MOREI want to highlight this bit:
...Timing Bear MarketsMy personal view is that the Greenspan-Bernanke regime of excessive stimulus, now administered by Yellen, will proceed as usual, and that the path of least resistance, for the market will be up. I believe that it would take a severe economic shock to outweigh the effect of the Fed's relentless pushing of the market. Look at the market's continued advance despite almost universal disappointment in economic growth. Only Japan was a modest pleasant surprise at 0.7% ahead of forecast and the U.K. and Switzerland scraped home by the skin of their teeth.
Everyone else fell short.There have been few such occasions when such broad disappointment with economic growth still allowed the U.S. and most other major economies to make material upward moves in their stock markets. It is yet another testimonial to the global reach of the Fed's stimulus of equities (as was the very substantial decline in emerging market equities on just talk of tapering! In equities there are few signs yet of a traditional bubble. In the U.S. individuals are not yet consistent buyers of mutual funds.
Over lunch I am still looking at Patriots' highlights and not the CNBC talking heads recommending Pumatech or whatever they were in 1999. There are no wonderful and influential theories as to why the P/E structure should be much higher today as there were in Japan in 1989 or in the U.S. in 2000, with Greenspan's theory of the internet driving away the dark clouds of ignorance and ushering in an era of permanently higher P/Es. (There is only Jeremy Siegel doing his usual, apparently inexhaustible thing of explaining why the market is actually cheap: in 2000 we tangled over the market's P/E of 30 to 35, which, with arcane and ingenious adjustments, for him did not portend disaster. This time it is unprecedented margins, usually the most dependably mean reverting of all financial series, which are apparently now normal.) By June this year, markets felt relatively quiet and under the surface there was still a considerable undertow of risk aversion in the institutions.
The Russell 2000 and the GMO High Quality universe were both just level with the S&P, all up 16%. Normally we would have expected the Russell to outperform handsomely. However, since then speculation has perked up so that today, the broad U.S. market is up 20% and the Russell 2000 is a more typical six points ahead while stocks in the GMO High Quality universe are several points behind. We have also had a sharp and unexpected uptick in parts of the IPO market in the U.S., so I would think that we are probably in the slow build-up to something interesting – a badly overpriced market and bubble conditions.
My personal guess is that the U.S. market, especially the non-blue chips, will work its way higher, perhaps by 20% to 30% in the next year or, more likely, two years, with the rest of the world including emerging market equities covering even more ground in at least a partial catch-up.And then we will have the third in the series of serious market busts since 1999 and presumably Greenspan, Bernanke, Yellen, et al. will rest happy, for surely they must expect something like this outcome given their experience. And we the people, of course, will get what we deserve. We acclaimed the original perpetrator of this ill-fated plan – Greenspan – to be the great Maestro, in a general orgy of boot licking....
