Valuation-Informed Indexing #324

by Rob Bennett

I wrote in the last column about why it is not reasonable to expect to see a perfect correlation between valuations and long-term returns. Stock prices really are determined primarily by shifts in investor emotion, as Robert Shiller showed in his “revolutionary” (Shiller’s word) 1981 research findings. But, because emotions are by definition irrational, their effects can play out in odd and unexpected ways. I don’t believe that it will ever be possible for us to make precise predictions.

Consider the circumstances that applied in 1996. Valuations were sky high. The P/E10 level reached 25, which always signals a price crash; there has not yet been a time in U.S. history when the P/E10 level hit 25 and we did not see a crash. The strategic implication is obvious, no? In those circumstances the valuation-informed investor should be lowering his stock allocation dramatically.

Not so fast.

The reason why a P/E10 level of 25 always causes a crash is that, while the natural direction of valuations is up (investors like to vote themselves raises) valuations that rise too high scare investors and cause their common-sense instinct to take control from their Get Rich Quick urge of their decision-making thought processes. A P/E10 of 25 is truly dangerous. It is indeed generally true that investors should lower their stock allocations when the P/E10 level hits 25. But the full reality is more complex.

Not all investors possess the same risk tolerance. So a valuation level that causes some investors to turn off the voice saying “stick with Buy-and-Hold” can have zero effect on others. Or it can cause an opposite effect on some investors. The jump to a P/E10 of 25 is viewed as a reward by some investors. When stock prices reach that exalted level, times have been good for those heavily invested in stocks for a long stretch. The price rise that causes some investors to begin to sell stocks can cause others to buy more heavily than others, to increase their stock allocations rather than to diminish them. Instead of crashing, prices can zoom upward.

This is what happened in 1996. Stocks really were dangerous in those days. But the years 1996, 1997, 1998 and 1999 saw some of the biggest price increases ever witnessed in the history of the U.S. market. Valuations did not begin their long march downward (a march only half completed today) until the early months of 2000. There were two waves of investor emotion hitting the market at the same time, a pullback wave that possessed the power to cause a price crash if left unimpeded and a stronger wave that more than countered the power of the first wave and instead produce big price increases for four more years.

Rarely is there only one wave of investor emotion affecting stock prices. Thus, statistical tests that assess the strength of the correlation between valuations and long-term returns understate the power of valuations to determine future prices. Robert Shiller went on the record in late 1996 with a prediction that investors who were highly invested in stocks would come to regret the choice within 10 years. The crash did not arrive until September of 2008. Shiller’s prediction was off by two years. He didn’t appreciate at the time he made the prediction that, instead of starting down immediately as prices usually do when they are already at a P/E10 level of 25, prices would first shoot up to a P/E10 level of 44 and only then begin working their way downward. He was fooled by a second wave of investor emotion that caused the bull market to continue another four years and that caused his prediction to come true two years later than he had anticipated.

The Buy-and-Holders would point to this bad prediction and conclude that Valuation-Informed Indexing is “dangerous” because it causes investors to “miss out” on gains that they would enjoy if only they stuck with the same stock allocation at all times. A surface reading of what happened in 1996 and in the years that followed does indeed support this claim.

But it is not so!

Stocks did very well from 1996 through 1999. Robert Shiller’s prediction really was off by two years. But it hardly matters. What Robert Shiller got right (that the insane level of overvaluation experienced in the late 1990s would end in tears) was far more important than what he got wrong (that the economic crisis caused by those out-of-control prices would not begin until late 2008 rather than by late 2006). Investors who followed valuation-informed strategies did NOT miss out on gains in any permanent sense. They missed out temporarily while prices climbed even higher and then avoided even bigger losses than they would have avoided had the crash come sooner and had thus been smaller and been less economically and politically destructive.

Valuation-Informed Indexing strategies often do not pay off for a good bit of time. That much is certainly true. But they always pay off (at least on a risk-adjusted basis and usually even on a nominal basis) to those patient enough to permit the various waves of investor emotion to play out. Stocks were not worth buying at the prices at which they were being sold in 1996. The insanity of the last four years of the century did not render the smart choice to protect one’s portfolio by lowering one’s stock allocation a less-than-smart one. The power of that second wave of investor emotion delayed the day of reckoning for Buy-and-Hold investors while increasing the price they paid.

There is no getting around the price that follows from following Buy-and-Hold strategies. But the price can arrive in several different forms. There are always a variety of wave scenarios that may apply in given circumstances. Investors need to consider the possibilities, note how they apply to those in their particular circumstances, make the choices that follow from those considerations, and then exercise the patience needed for the various wave scenarios to play out.

It can be frustrating and unsettling waiting for the various wave scenarios to play out. But Valuation-Informed Indexing strategies always work if given enough time. The reason why is that they MUST always work. Long-term timing is how price discipline is exercised in the stock market. It is absurd to think that there might be a market in which it were a good thing to fail to exercise price discipline. Price discipline is a positive, valuations always matter and Valuation-Informed Indexing is always superior to Buy-and-Hold in the long term.

Rob’s bio is here.