Valuation-Informed Indexing #178
by Rob Bennett
When stocks are priced insanely low, as they were in 1982, a regression analysis of the historical return data shows that the most likely annualized 10-year return is 15 percent real. When stocks are priced insanely high, as they were in 2000, a regression analysis of the historical return data shows that the most likely annualized 10-year return is a negative 1 percent real.
It’s hard to argue that it makes sense to go with the same stock allocation in both circumstances. So I often use those numbers to make the case for avoiding Buy-and-Hold strategies (which are rooted in the idea that it is okay for investors to remain at the same stock allocation at all times) and for following Valuation-Informed Indexing strategies (which are rooted in the idea that investors should be aiming to keep their risk profiles roughly constant, something that cannot be done without a willingness to change one’s stock allocation in response to big valuation shifts).
There’s always been a caveat, however.
The same historical return data that shows that 10-year returns change dramatically with changes in starting-point valuation levels also shows that, at the end of very long time-periods, the most likely return for stocks is always good. When prices are what they were in 2000, the most likely 30-year return is 5.01 real. The most likely 60-year return when stocks are as overpriced as they were in 2000 is 6.24. That’s not too far off from the average long-term return of 6.5 percent real.
These numbers give comfort to the Buy-and-Holders. They seem to show that, if you hold long enough, Buy-and-Hold does eventually work. Perhaps the numbers are not so hot at the end of 10 years or 15 years or 20 years. But if you are just patient enough, the Buy-and-Hold ship does eventually come in. Or at least so it once seemed to be the case.
I have often in an effort to be fair to the Buy-and-Holders told people about these numbers. I mentioned two caveats. But until recently I have believed in the basic point that Buy-and-Hold works better after the passage of a long number of years than it does after the passage of only 10 or 15 or 20 years.
The first caveat that I have pointed out is that the passage of time means something. If Investor A obtains a good return after the passage of 10 years and Investor B obtains a good return after the passage of 30 years, Investor A is following the more appealing strategy. When we invest, we are turning over the use of our money for a time in exchange for the returns we are paid to do so. We are always better off to give up the use of the money for a shorter period of time rather than a longer period of time. So, even if Buy-and-Hold strategies produce good returns at the completion of 30 years or 60 years, they are inferior to the Valuation-Informed Indexing strategies that generally produce good returns at the completion of 10 years or so.
The second caveat is that the numbers that apply at the completion of 30 years or 60 years apply in the real world only if the investor never sells his stocks. Buy-and-Holders insist that this assumption be employed in most (or all?) of the research conducted in this field. But I see little grounds for having confidence that the assumption reflects what happens in the real world.
Stock were priced at three times fair value in 2000. Prices always fall to one-half of fair value before a secular bear market comes to an end. So those who follow Buy-and-Hold strategies are likely to see a loss of close to five-sixths of their accumulated lifetime wealth before this secular bear reaches its capitulation point. Good luck sticking with your high stock allocation through a drop like that, one that will take place during the passage of many dark years for the underlying economy and that will see the publication of numerous pessimistic media accounts as the dark days grind on and on!
The numbers show that Buy-and-Holders who bought in 2000 were likely to see a 60-year annualized return of 6.24 percent real. But will any actual living investor see such a return? I have my doubts. But those are the numbers that you get from an analysis of the historical return data and so I think in fairness to the Buy-and-Holders it must be said that the 60-year return numbers are good even for stocks purchased at the worst valuation level we have ever seen in U.S. history.
What makes the numbers turn good after the passage of 30 years or 60 years?
I believe that the primary factor is that bull/bear cycles generally last about 35 years. If you purchase stocks at the top of one insane bull market, you are going to be down for a long, long time. Eventually, however, we will be at the top of the next insane bull market. The subsequent insane bull market pushes prices high enough that even those who purchased when the long-term value proposition was poor end up with good results.
I no longer believe that the 30-year and 60-year numbers are as good as I once thought they were.
We all know that the power of the compounding returns phenomenon is counter-intuitively big. Our minds have a hard time seeing how big a difference it makes to hold an investment for a long time.
I now believe that one of the factors that makes the 30-year and 60-year numbers look good is that they are the product of many years of compounding. What we may be overlooking is that even small differences in return percentages can translate into large differences in portfolio dollar values when many years of compounding are taken into account. At the end of 10 years, a strategy that yields an annualized return of 6.5 might not be all that much better than a strategy that yields an annualized return of 6.25. But at the completion of 60 years, the dollar value difference would be very large indeed.
I think that it is fair to say that Buy-and-Hold performs a bit better at the completion of 30 years than it does at the completion of 10 years. But I today lean toward thinking that the difference is not as great as I once thought it to be.