Analysis of Valuation-Informed Indexing?

Question

My question. I've been reading about Valuation-Informed Indexing (VII). What is your reaction? Would you recommend this strategy? What are its potential strengths/weaknesses?

Background. For those who don't want to click through, the proposal seems to be something like this. You use an index, and adjust the ratio of stocks/bond based upon whether a simple calculation estimates that the market is overpriced or underpriced. In particular, calculate the P/E10 (the price-to-earnings ratio for the entire index, with earnings averaged over the past ten years).

• If P/E10 is below 12, invest in 90% stocks and 10% bonds.
• If P/E10 is in the range 12-21, invest in 60% stocks and 40% bonds.
• If P/E10 is above 21, invest in 30% stocks and 70% bonds.

Analysis. I saw an analysis which looked at a related strategy: instead of 12 and 21, analyze the historical data for P/E10 ratios up until now to calculate the top 25% percentile and bottom 25% percentile for historical P/E10 ratios. That analysis ran some simulations over rolling 30-year periods and compared to buy-and-hold with a fixed mix of 60% stocks and 40% bonds. Apparently, in the simulation the VII strategy beat buy-and-hold for 102 out of 110 of the rolling 30-year periods.

(My impression was that rolling periods is statistically a bit dubious, because the 110 data points are highly correlated: you only have something like 110/30 = 4 independent data points, not 110 independent data points. I don't know whether that's right.)

(Footnote: Here is some analysis looking at whether P/E10 can help predict whether bonds will beat stocks.)

Personally, I'm not sure whether to buy the VII proposal. It sounds like long-term market timing: trying to do a better job than the rest of the market at predicting, based upon a simple formula, whether the market is over-priced or under-priced. Is it plausible that this could work? I know short-term market timing has gotten a bad rap; do the reasons for avoiding short-term market timing apply to the VII proposal as well? Does the VII proposal assume the Efficient Market Hypothesis is wrong, and is that a reason to be skeptical of the VII proposal?

Summary. In short, does anyone have any reactions to, or analysis of, this proposal? Do you find the arguments and evidence credible aand convincing? Why or why not? What do you see as the strengths or weaknesses of the proposal or the evidence supporting it? Thanks in advance!

Answer 1

My reaction to this is that your observation @D.W. is spot on correct:

It sounds like long-term market timing: trying to do a better job than the rest of the market at predicting, based upon a simple formula, whether the market is over-priced or under-priced.

I read the post by the founder of Valuation Informed Indexing, Rob Bennet. Glance at the comments section. Rob clearly states that he doesn't even use his own strategy, and has not owned, nor traded, any stocks since 1996! As another commenter summarizes it, addressing Rob:

This is 2011. You’ve been 100% out of stocks — including indexes — since 1996? That’s 15 years of taking whatever the bond market, CDs or TIPS will yield (often and currently less than 2%)... I’m curious how you defend not following your own program even as you recommend it for others?

Rob basically says that stocks haven't shown the right signals for buying since 1996, so he's stuck with bonds, CD's and fixed-income instead. This is a VERY long-term horizon point of view (a bit of sarcasm edges in from me).

Answering your more general question, what do I think of this particular Price/ Earnings based ratio as a way to signal asset allocation change i.e. Valuation Informed Investing? I don't like it much.

• It is an overly simple rule,
• It is too generalized to be valid for any given stock, and
• Worst of all, might cause a conservative investor to miss out on those rare sudden inflows of return (whether due to growth, one-time special dividend, M&A or other corporate action) that are the reward for buying and holding.

Answer 2

This is Rob Bennett, the fellow who developed the Valuation-Informed Indexing strategy and the fellow who is discussed in the comment above.

The facts stated in that comment are accurate -- I went to a zero stock allocation in the Summer of 1996 because of my belief in Robert Shiller's research showing that valuations affect long-term returns. The conclusion stated, that I have said that I do not myself follow the strategy, is of course silly. If I believe in it, why wouldn't I follow it?

It's true that this is a long-term strategy. That's by design. I see that as a benefit, not a bad thing.

It's certainly true that VII presumes that the Efficient Market Theory is invalid. If I thought that the market were efficient, I would endorse Buy-and-Hold. All of the conventional investing advice of recent decades follows logically from a belief in the Efficient Market Theory. The only problem I have with that advice is that Shiller's research discredits the Efficient Market Theory.

There is no one stock allocation that everyone following a VII strategy should adopt any more than there is any one stock allocation that everyone following a Buy-and-Hold strategy should adopt. My personal circumstances have called for a zero stock allocation. But I generally recommend that the typical middle-class investor go with a 20 percent stock allocation even at times when stock prices are insanely high. You have to make adjustments for your personal financial circumstances.

It is certainly fair to say that it is strange that stock prices have remained insanely high for so long. What people are missing is that we have never before had claims that Buy-and-Hold strategies are supported by academic research. Those claims caused the biggest bull market in history and it will take some time for the widespread belief in such claims to diminish. We are in the process of seeing that happen today.

The good news is that, once there is a consensus that Buy-and-Hold can never work, we will likely have the greatest period of economic growth in U.S. history. The power of academic research has been used to support Buy-and-Hold for decades now because of the widespread belief that the market is efficient. Turn that around and investors will possess a stronger belief in the need to practice long-term market timing than they have ever possessed before. In that sort of environment, both bull markets and bear markets become logical impossibilities.

Emotional extremes in one direction beget emotional extremes in the other direction. The stock market has been more emotional in the past 16 years than it has ever been in any earlier time (this is evidenced by the wild P/E10 numbers that have applied for that entire time-period). Now that we are seeing the losses that follow from investing in highly emotional ways, we may see rational strategies becoming exceptionally popular for an exceptionally long period of time. I certainly hope so!

The comment above that this will not work for individual stocks is correct. This works only for those investing in indexes. The academic research shows that there has never yet in 140 years of data been a time when Valuation-Informed Indexing has not provided far higher long-term returns at greatly diminished risk. But VII is not a strategy designed for stock pickers. There is no reason to believe that it would work for stock pickers.

Thanks much for giving this new investing strategy some thought and consideration and for inviting comments that help investors to understand both points of view about it.

Rob