Valuation-Informed Indexing #335
by Rob Bennett
My last three columns examined a recently completed research paper titled Shiller’s CAPE: Market Timing and Risk. The first column expressed concern re the paper’s finding that never before in U.S. history have we seen today’s extremely high valuation levels apply for so long a time-period. The second column looked to the research paper’s finding that overvaluation is “sticky” and argued that this is because stock prices are set by shifts in investor emotions and investor emotions are sticky. The third column explored the headline finding of the paper — that it generally does not pay for investors to engage in valuation-based market timing (Valuation-Informed Indexing). Now I would like to turn to a more fundamental question — the research paper’s assertion that, if it is not possible to use valuation-informed strategies to profit from market timing, the market is efficient.
The authors of the paper state that: “If markets are efficient, knowing CAPE should not help investors earn superior future returns by selling (buying) stocks and buying (selling) a risk-free asset when CAPE is high (low). In other words, market timing strategies using CAPE should not be profitable.”
I don’t think that that’s the test. The logic here equates market efficiency and the failure of market timing; if timing works, the market is not efficient, and, if timing doesn’t work, the market is efficient. I believe that long-term market timing always works and that the market is not efficient. So in a surface sense, I seem to buy into this logic. But I don’t think that the dichotomy being set up here is quite right. Even if the authors of the study were correct that valuation-based market timing rarely works (I am the co-author with Wade Pfau of peer-reviewed research coming to a very different conclusion), valuation-based investing strategies would still offer many benefits and a belief in market efficiency would still be a dangerous delusion.
Do you believe that an investor who held a $100,000 portfolio in 1981 when the P/E10 value was 8 possessed an asset of the same value as the asset held by an investor who held a $100,000 portfolio in 2000 when the P/E10 value was 44? I sure don’t. It is my view that the second portfolio was worth more than five times what the first portfolio was worth. Why? Because the P/E10 value of the first portfolio was one-fifth as great. The word “overvalued” means something. The nominal value of the second portfolio was insanely overstated. It seems silly to me to pretend that the value of an insanely undervalued asset is as great as the value of an insanely overvalued asset of the same nominal value.
The study tests whether it is possible to profit from valuation-based market timing by moving one’s investment funds from stocks to Treasuries at times when the P/E10 value rises to high levels. I question numerous aspects of the methodology employed in the study. But let’s leave that question to the side so that we can consider a more basic question. Does it even matter all that much? Do we really prove that the market is efficient when we prove that market timing is not possible? I don’t think so. There are many benefits to be had by taking valuations into consideration that do not in a direct sense require market timing.
One reason why market timing rarely works using the methodology employed in the study is that the return paid by Treasuries is usually very small. Short-term timing really doesn’t work. The only way that valuation-informed strategies can provide benefits is for the investors following with them to stick with them for at least 10 years. But parking one’s money in Treasuries for 10 years is a bleak prospect. Make that the contest and stocks are usually going to come out ahead (amazing enough, even the authors of this study acknowledge that this is not so when valuations reach truly insane levels, the sorts of valuation levels that have applied in the U.S. market for most of the past two decades). But what if investors can profit in other ways by being aware of how dangerous stocks become at times of high valuations? Does that not show that valuations matter and that the market is something less than efficient?
If you calculate the safe withdrawal rate using a valuation adjustment for the P/E10 value that applies on the day the retirement begins, you learn that the safe withdrawal rate was 1.6 percent in 2000 and 9.0 percent in 1982. That is another way of saying what was said up above — the investor who held a $100,000 portfolio in 1982 possessed an asset of more than five time the value of the investor who held a $100,000 portfolio in 2000. Say that you are 100 percent convinced that market timing does not work and intend to refrain from it. There is still value in knowing whether or not you have saved enough to retire, is there not? Market timing is not the only possible benefit of following valuation-informed strategies.
Buy-and-Holders are working a sleight of hand. People look at the low rates of return provided by Treasuries and they conclude that market timing is not an appealing idea. There are ways around the problems stressed by the Buy-and-Holders but most investors are sufficiently impressed by the low returns provided by Treasuries to remain cool to the market timing concept. But that’s not really the issue in dispute. The important question is — Is overvaluation real? If overvaluation is real, investors cannot even know where they stand financially without considering the effects of valuations, something that Buy-and-Holders warn them that they must never, never do because to do so would be to engage in the same sort of thought crime that causes others to try market timing.
I love long-term market timing. My research shows that investors can reduce the risks of stock investing by nearly 70 percent by tuning out the Buy-and-Hold noise and exercising price discipline when buying stocks. But engaging in successful market timing is not really the point of following a valuation-informed investing strategy. The true point is to know where you stand as an investor. If valuations matter (those who follow the peer-reviewed research have known that they do since 1981), it is not possible for any investor to know where he stands without taking valuations into consideration in all his calculations. All stock investors should be taking valuations into consideration at all times (and shunning Buy-and-Hold strategies), both those who engage in market timing and those who elect not to do so.
If the market were efficient, overvaluation could not exist. But as Shiller showed many years ago, it does.
Rob’s bio is here.