Valuation-Informed Indexing #88
By Rob Bennett
I often make the claim that investors who make the shift from Buy-and-Hold (which calls for the investor always to stick with the same stock allocation) to Valuation-Informed Indexing (which calls or the investor to aim to keep his risk profile roughly constant and thus requires occasional allocation shifts in response to big swings in valuation levels) thereby reduce the risk of stock investing by 80 percent. That’s a pretty darn bold claim. I thought that I should explain how I came up with that figure.
Buy-and-Hold is rooted in a belief in the Efficient Market Hypothesis, which posits that investors set prices each day by responding rationally to economic or political developments. Each day is a new day in the Buy-and-Holder’s world. How stocks performed last week or last month or last year or ten years ago has nothing to do with where prices will stand at the end of today’s action. Only new developments matter. Developments that have not yet taken place obviously cannot be known to anyone. Thus, future prices cannot be known.
Buy-and-Holders naturally hope that returns will be good. And they take comfort in their knowledge that the average long-term return over the history of the U.S. market is 6.5 percent real. Most Buy-and-Holders expect to earn a return of something in that neighborhood.
But since Buy-and-Holders believe that the community of investors sets a new price each day, they can never have confidence that any particular return will apply for any particular time-period. Stock prices were falling rapidly in the early days of the financial crisis and the tenets of the Buy-and-Hold Model offered no reassurance that they would stop falling anytime soon. If each day’s prices are determined by each day’s economic or political developments, it is at least possible that prices could continue falling for a long time.
Thus, the risk attached to Buy-and-Hold investing is high.
How high? I have heard some Buy-and-Holders tell newcomers to the investing philosophy that they need to be prepared for a price drop of 50 percent to possess a realistic expectation that this strategy will work for them.
Unfortunately, it is not at all clear how these Buy-and-Hold advocates came up with the 50 percent number. My sense is that they looked at the losses suffered in the most recent extended bear market, the one experienced in the 1970s and early 1980s, and concluded that those losses are the worst that any Buy-and-Holder could ever be required to endure.
It’s not so!
Stock prices fell close to 90 percent in nominal terms following the insane bull market of the 1920s. And the level of overvaluation that brought on the Great Crash was small compared to the level of overvaluation we experienced in the late 1990s. The P/E10 value in late 1929 was 33. The P/E10 value in early 2000 was 44. Realistic Buy-and-Holders need to be prepared to suffer losses of at least 90 percent.
Now that’s risk!
The reality, of course, is that no investor can take a hit of that magnitude without selling. That’s why Buy-and-Holders try to fool themselves into believing that the 50 percent loss we saw in the bear market of the 1970s (a bear market prompted by a P/E10 level of only 25) is the worst that can happen. Buy-and-Holders are living in a fool’s paradise. They are taking on far more risk than they can bear to think about, far more risk than they can handle.
The risk being taken on by Buy-and-Holders is so great that it cannot be effectively quantified. I think it would be safe to safe that there is essentially no limit to the risk being taken on by Buy-and-Holders. I of course don’t mean to say that there are not time-periods when Buy-and-Hold performs very well. I am saying that there are other time-periods when Buy-and-Hold causes such devastating life setbacks that it is unrealistic to believe that more than a tiny number of Buy-and-Holders will be able to maintain their high stock allocations when the bad times arrive.
How about Valuation-Informed Indexers?
Valuation-Informed Indexing is rooted in a very different premise. We believe that the 6.5 percent real return applies for every year. It isn’t reflected in each year’s nominal return only because emotional investors set prices where their emotions tell them to set them. But profits are earned gradually through the application of effort over time and thus so are investment returns, according to the new model.
Since the amount earned each year and the amount reflected in the year’s nominal return often differs, Valuation-Informed Indexers believe it is possible to predict long-term returns. When nominal returns get wildly out of whack on the upside, we will be seeing years of poor returns (there is not one exception in the historical record). When nominal returns get wildly out of whack on the downside, we will be seeing years of super returns (again, there is not one exception in the historical record).
Risk is uncertainty. To the extent that an asset class provides returns that can be predicted in advance, that asset class is not risky. Since long-term returns are to a large extent predictable for Valuation-Informed Indexers, the risk taken on by those following the new model is greatly reduced.
Can the reduction in risk be quantified?
This graphic shows the correlation between the P/E10 level that applies for a broad index on the day a purchase is made and the 20-year annualized return obtained on that purchase. The R-Squared figure is 0.6335. This means that the P/E10 value that applies on the day you purchase an index fund tells you 63 percent of what you need to know to identify the 20-year return you will obtain.
So your long-term return is 37 percent unknown. That is, you have reduced the risk of stock investing by 63 percent by going with a valuation-informed strategy (Buy-and-Hold returns are 100 percent unknown. Remember, Buy-and-Holders believe that stock prices are determined by each day’s events -- future events can obviously never be known on the day you purchase your stocks).
Actually, I think the full reality is a little better than that. If you buy stocks because the P/E10 value predicts a strong long-term return (or avoid stocks because the P/E10 value predicts a poor long-term return), you are concerned only with mispredictions in a downward direction (or in an upward direction). So in most cases only half of the unknown aspect of stock returns can do you harm. That is, only half of the unknown aspect of stock returns represents risk in a practical sense.
So long-term returns are only 19 percent unknown for Valuation-Informed Indexers. Those who make the shift thereby reduce the risk of investing in stocks by roughly 80 percent.
Is this a perfect measure of the risk reduction that applies with the shift to Valuation-Informed Indexing? It is not. It is a rule of thumb. I make the claim that shifting to Valuation-Informed Indexing reduces risk by 80 percent not because I believe that to be precisely correct but because I believe the risk reduction to be very significant and because I believe that we need to make an effort to quantify this benefit to persuade investors to make the shift and because this is the