Valuation-Informed Indexing #159

by Rob Bennett

Stocks offer a stronger long-term value proposition when prices are low than they do when prices are high. That’s a nice, simple rule that always applies. There are times, though, when the realities of what the P/E10 value is telling us are more tricky than what that simple rule suggests.

Say that the P/E10 value is 18. That’s a bit higher than the fair-value P/E10 value of 15. But it is not insanely high, like the P/E10 values we saw from 1996 through 2008. A P/E10 value of 18 is a moderately high P/E level. Given that stocks offer an amazing long-term value proposition when selling at fair-value prices, they still offer a plenty-good-enough long-term value proposition when selling at moderately high prices. Stocks are worth buying when the P/E10 value is 18.

Aren’t they?

Looking only at the one factor of price, it’s true that stocks are worth buying when the P/E10 value is 18. The full reality, however, is that, as important as it is to check the P/E10 value before purchasing stocks, the P/E10 value should never be the only thing you check before buying stocks. The same P/E10 value can mean different things in different circumstances.

The P/E10 value moved upward from 8 in 1982 to 44 in 2000. There were points in the late 1980s and early 1990s when the P/E10 value that applied was 18. There was also a time in the days following the price crash of late 2008 when the P/E10 value was 18. The long-term value proposition of stocks was not even close to equal in those two very different sets of circumstances.

To understand how to make best use of the P/E10 metric, you need to understand the model (Valuation-Informed Indexing) that follows from a belief in Shiller’s “revolutionary” (Shiller’s word) finding that valuations affect long-term returns.

The premise of the model is that stock prices do not represent economic value, as is presumed in the Buy-and-Hold Model. Stock price changes are caused by changes in investor emotions, according to Valuation-Informed Indexers.

Investor emotions (and thus stock prices) do not change randomly. They change in highly predictable patterns.

In the late 1980s and early 1990s, we were on our way from a P/E10 of 8 to a P/E10 of 44. We certainly had no way of knowing at the time that we would end up at 44. We had never gone that high before. But we knew that the P/E10 level was likely to continue going up after hitting 18 even if for a few months it went below that market. Never once in U.S. history has the P/E level peaked at 18. It usually peaks at 25 or 26. So the odds were strong at the time that the long-term direction of prices was downward.

Just the opposite was the case when we hit 18 in early 2009. The P/E10 value peaked at 44 in 2000 and started its way downward. Never in U.S. history have valuation levels began working their way up again on a long-term basis without first hitting 7 or 8. So the 18 that we saw in 2009 was an 18 on the way to becoming a 7 rather than an 18 on the way to becoming a 25. That makes all the difference in the world.

Stock prices were a little high in the late 1980s and early 1990s. But the odds were that prices would be rising for a good bit of time yet. Stock prices were also a little high in 2009. But the odds then were that prices would be falling hard in years to come. The long-term value proposition of stocks was much stronger in the late 1980s than it was in 2009.

Two things throw people: (1) the pattern that always applies does not play itself out in precisely the same way; and (2) the pattern that always applies evidences itself only in the long term and never in the short term.

Say that an article appeared in the late 1980s pointing out that the P/E10 level of 18 was nothing to be concerned about, that stock prices were almost surely headed upward over the long term. And say that we then saw a drop in the P/E10 level to 15, a lower P/E10 level that remained in place for 12 months. There would have been people saying: “This shows that that stuff you sometimes hear about stock prices playing out in a predictable patten is a bunch of baloney — Prices have gone down, not up.”

There is no way to predict short-term price changes. So there is no short-term price behavior that can disprove the theory. The only thing that would disprove the theory is a showing of a case in which the pattern did not apply in the long term. In the 140 years of stock-market history available to us, there has never been such a showing.


Why does the same general pattern repeat over and over again?

It’s because stock prices are determined by the result of a battle between our Get Rich Quick impulse and our desire to invest pursuant to the dictates of common sense. Our common sense tells us that stock prices must reflect the economic realities. So, in the long run, they always do. But our Get Rich Quick impulse tells us that it sure would be nice if stock prices would double or triple for no good reason whatsoever. So, for a limited period of time (of about 10 years in length), we always permit prices to climb to levels two or three times where they would be if they reflected the true value of our stocks.

Long-term stock prices are always to a large extent predictable. But the P/E10 value by itself does not tell you all you need to know to make effective predictions. You also must take into consideration the context (whether stock prices are headed toward the highs of the particular cycle or the lows of the particular cycle) in which the P/E10 value appears.

Rob Bennett has recorded a podcast titled Passive Investing Is the Most Emotional Investing Approach of All. His bio is here.