On another thread, JWR asked this question by way on an article.  Earlier I had asked him what level of PE10 would support a 70% stock allocation and he replied 14.

Later, he indicated that the data shows that we haven't seen a PE14 for a period of at least a year, since 1988.

So imagine you were 45 in 1988 and are about to retire.  You missed most of the  largest bull market in our lifetimes because you were down to 20-30-40% through most of that time.  You missed the bear in 2000-2002 also.

Following PE10 valuations would have likely cost you dearly in retirement.

Worse, there can easily be another 20 years before we see PE10 back down to the 14 level.

40 years in a row?  Can you wait that out?

best,

Bill

Keep in mind that this is a 70% stock allocation in retirement while making withdrawals.

This is NOT a zero stock allocation throughout.

Notice that this is a specific outcome, rather unusual, that included a bubble.

Have fun.

John Walter Russell

Not really, John.  My example had a 45 year old in 1988.  There are 45 year olds to day too, and all ages in between.

Are you saying the PE10 is only for retired folks in the withdrawal stage?

I have no real issue with PE10 valuations for retirement for me because I don't intend to be much over 30 or at most 40% equities anyway.

"Notice that this is a specific outcome, rather unusual, that included a bubble."

That got me laughing.  Thanks.

I remember how often you have used the 1965 thru the 70's stagnation period to warn folks off TR investing.

Wasn't that a specific, and unusual outcome?

best,

Bill

Bill,

"Wasn't that a specific, and unusual outcome?"

No really,  That particular date for retirement gave very specific credibility as to the meaning that such and such a real, inflation adjusted rate of withdrawal (in this case, 5%) has only a 95% probability of success.  1966 was one case where the probability of failure should have been 100%! 8-)

People often assume that '95% success rate' is good enough.  It is good enough, in 19 out of 20 years for retirement.  The problem we all face now, whenever deciding whether or not to take that 5% rate of withdrawal, is if 2008 through 2038 will be like 1965 through 1995, or some other time period, where that specific withdrawal was good enough.

I remember how often you have used the 1965 thru the 70's stagnation period to warn folks off TR investing.

Wasn't that a specific, and unusual outcome?

Actually, NO, not for the 1965 and 1966 sequences. P/E10 was 24. The outcome was only slightly worse than "typical" at that valuation level.

But YES, a procedure that ignores probabilities can be misleading. The notion that 4% was SAFE is misleading. It survived. That was the specific historical outcome. But it was far from safe. (It had a probability of failure between 20% and 50%.)

Have fun.

John Walter Russell

Bill, this is a recent NOTE at my site.

Latest Research Findings (Price Drops)

Valuation Informed Indexing is a winner. It produces consistently good outcomes. It removes the sensitivity as to when price drops occur.

During accumulation when you are dollar cost averaging, there are two alternatives that make sense. One is to invest entirely in stocks all of the time. At year 30, the odds are about 20% that you will do spectacularly well. The odds are about 20% that you will have seriously underperformed Valuation Informed Indexing. Otherwise, the odds are that your results will be similar to Valuation Informed Indexing.

The other approach is to start with 20% stocks and 80% TIPS and wait until P/E10 falls below 15 for the first time. Then you switch to Valuation Informed Indexing. The resulting outcomes are almost identical to using Valuation Informed Indexing from the start.

Another story from my latest research is that the likelihood of P/E10’s falling below 10 is only about 50%-50%. Although it is reasonable to expect prices to fall to one half of today’s level (from P/E10=24 to P/E10=12), do not count on deeper cuts.

Have fun.

John Walter Russell

Bill,

You have good instincts in spite of yourself.

With P/E10=23, as it does today, and assuring a balance of zero or higher at Year 30:

With a 30% stock allocation, your SAFE withdrawal rate is 4.0% with TIPS (at 2%) or 4.5% with corporate bonds/preferred stock.

With a 70% stock allocation, your SAFE withdrawal rate would only be 3.6% or 3.7% (respectively).

If you were willing to take 50%-50% odds, the withdrawal rates would be around 4.6% with TIPS and 5% with corporate bonds/preferred stock.

The fact that you have a low stock allocation planned for retirement is a definite PLUS.

Have fun.

John Walter Russell

"The fact that you have a low stock allocation planned for retirement is a definite PLUS."

[/quote]

Right about here is where we always do the "yeah, but . . . "

Yeah, but if you have a high equity allocation consisting of quality companies paying a dollar dividend greater than you need to withdraw and that dollar income stream is growing at a rate much greater than expected inflation . . . . 

. . . . . then maybe you don't have to worry about P/E10 or P/E40 or P/E whatever.

I think you agreed with this, JWR, in prior posts where you recognized that a dividend strategy is different and a 100% allocation to equities even now might be just ducky.  Yes?

Regards.

Cliff 

Cliff,

"I think you agreed with this, JWR, in prior posts where you recognized that a dividend strategy is different and a 100% allocation to equities even now might be just ducky.  Yes?"

Maybehaps the P/E10 of an individual stock, or fund, is low, while that of the market is high?

Or that high yield means low price, hence low P/E?  After all, P/D and P/E are related (by the DPR), so low P/D implies high D/P, or percentage dividend yield?

So, high D/P, covering withdrawals, is less risky then low D/P (or high P/E10), which relies on share price appreciation to fund withdrawals?  A more 'risky' strategy, relying on share price appreciation?

Anyhow, all of this stuff ties together, obviously! 8-))

I think you agreed with this, JWR, in prior posts where you recognized that a dividend strategy is different and a 100% allocation to equities even now might be just ducky.  Yes?

Definitely so. I believe that dividend and income stragegies are best. Valuation Informed Indexing comes close. Fixed allocations fall far behind.

In terms of Bill's decision to stick with a fixed allocation, his choice of 30% is a good choice.

P/E10 does not go away entirely, even with dividend strategies. It helps you assess the likelihood of more favorable stock prices (and better dividend yields). It may lead you to preferred stock or corporate bonds for a few years, while hoping for better stock opportunities.

Have fun.

John Walter Russell

JWR,

"P/E10 does not go away entirely, even with dividend strategies. It helps you assess the likelihood of more favorable stock prices (and better dividend yields)."

A few questions, for my own understanding of P/E10.

As I understand it, you would use P/E10, of the market as a whole (market cap weighted) to determine the valuation (good or bad) of equities, in general.  For example, the P/E10 of the S&P500.  Based upon this, you would set your portfolio asset allocation.

1) Would/could you not compare the P/E10 of some fund, rather then the market, as a whole, to determine whether or not to invest in that particular fund?  For example, if the P/E10 of that fund was at 11, while that of the market was 22, one indicator would say light on equity, while the other would indicate exactly the opposite, for that particular fund?

2) What about the P/E10 of individual stocks (same question as 1?)

3) If P/E10 of individual stocks and funds is significant (see 1 & 2), does that mean the average investors has to pour through historical data and annual reports to come up with this number?  That is, how good is current P/E, by itself, as an indicator for an individual stock/fund?

4) How important is P/E10, per your definition, compared to the average P/E over the last 10 years?  That is, I understand the difference in the calculation, as you explained it.  I'm wondering how a simple test (of the current P/E of a stock/fund, compared to it's average over some time period, is predictive of future share price behavior).  Or even current P/E compared to P/E10, for that stock or fund.

5) Are there any guidelines, for something similar to P/E10, on the debt side of one's portfolio?

6) Finally, wouldn't the P/E, or P/E10, of of one's portfolio, rather then the market, as a whole, or any specific stock/fund, be more indicative of the future performance of one's portfolio, specifically with respect to portfolio survivability during retirement?  This gets into asset allocation.  For example, P/E10 might indicate 30% equity (valuation based), but I might choose 4 funds/stocks, whose average P/E10 might indicate 60% equity?

I realize these questions may already be answered somewhere on your website but, being retired, I've forgotton how to do any real work, including searching for answers! 8-))

1) Would/could you not compare the P/E10 of some fund, rather then the market, as a whole, to determine whether or not to invest in that particular fund?  For example, if the P/E10 of that fund was at 11, while that of the market was 22, one indicator would say light on equity, while the other would indicate exactly the opposite, for that particular fund?

2) What about the P/E10 of individual stocks (same question as 1?)

3) If P/E10 of individual stocks and funds is significant (see 1 & 2), does that mean the average investors has to pour through historical data and annual reports to come up with this number?  That is, how good is current P/E, by itself, as an indicator for an individual stock/fund?

Benjamin Graham came up with the basic idea. He recommended looking at smoothed earnings, both over 8 to 10 years and over the last three years. This was the source of Professor Shiller's idea. He applied Benjamin Graham's idea to the S&P500 as opposed to individual stocks.

So YES, when selecting individual companies to buy, look at smoothed earnings, not simply the latest year's earnings. Single year P/E jumps around considerably. Smoothed earnings do not.

The reason for looking at the last three years of earnings is to make sure that the company is not going to go out of business. [Along the same line of thought, Orygunduck looks at the last two quarters of cash flow to make sure that dividends will continue.]

An individual investor should look at the earnings in a company's annual report. He should look at more than a single year before actual purchase.

The dividend equivalents of P/E10 based on 5 to 10 years of dividend income (P/D5 and P/D10) work well. They are an alternative to P/E10.

The overall P/E10 of the market (S&P500) has a huge influence on individual stocks. Even great stocks get dragged down when the market as a whole is falling.

Have fun.

John Walter Russell

4, 5 and 6:

Remember, it is the current price divided by smoothed earnings, not the average of price to earnings ratios. [Engineering note: Go ahead, use the median earnings.]

I do not know of a debt equivalent. But smoothed earnings should help in looking at the ability to pay interest (as well as dividends). [Cash flow seems more relevant.]

I was able to look at four major slices (growth and value, large and small capitalization). I called these Gummy Slices. They all respond well to P/E10 of the market as a whole.

I believe that the merit of value investing has been proved beyond reasonable doubt. Price to earnings ratios and dividend yields are two approaches for measuring value. Single year earnings can drop. Smoothed earnings are much more reliable.

Have fun.

John Walter Russell

JWR,

Thanks.

On a more theoretical note, do you look upon the P/E10 results and the FF3F results (the value premium) as showing similar trends, or cause/effect relationships?

JWR, people do not ignore valuations, at least not Dividend/Income Investors who post on this forum.  IMO, what you are really asking is, Why do people ignore PE10?  The answer to this question is clear. 

To be useful, a market timing indicator, whether statistical or anecedotal, must perform two functions.  It must get you "out" of the market before or early in a Bear market, AND it must get you "into" the market in time to catch the Bull market phase.

What PE10 appears to do, only too well, is keep people out of the market.  PE10 fails to get investors into the market for Bull phases.  Being only lightly in equities (if in any at all) during the 90's missed out on huge potential gains during that Bull market decade.  

Investors may fail to achieve financial goals, not only because they lose money during bear markets, but because they fail to accumulate a large enough nest egg during good times to last them financially through retirement.   To achieve historical market returns it is necessary to play the game (be in the market).  Unfortunately, at times, this means there will be, temporarily, times of negative returns. 

Let us know when you have modified your market timing indicator to not only avoid the "downs", but also capture the "ups" of the market. 

Regards,
Russ

On a more theoretical note, do you look upon the P/E10 results and the FF3F results (the value premium) as showing similar trends, or cause/effect relationships?

I do not know the time frame of the FF3F model. I believe that it is very short, possibly only one or two years.

I believe that the emphasis on price to book is overdone, the high value of R-squared (and I believe that it was R-squared, not the variability that was addressed) was a result of "data dredging"--that is, it was the best of about 100 different factors. The effect is real. The reliability is less than indicated. (My guess, not knowing the details of the FF3F research.)

The FF3F research that I have seen only produces point estimates, not reliable confidence intervals along with a time frame.

I consider James O'Shaughnessy's work to be reliable (What Works on Wall Street). He uses a rolling ten year period. It supports value investing.

I believe that P/E10 is reliable when you have a time frame of 10 years or more. You need to pay attention to confidence intervals, however. They are very wide.

Have fun.

John Walter Russell

JWR, people do not ignore valuations, at least not Dividend/Income Investors who post on this forum.

I am referring to those people who advocate a fixed allocation as a result of the efficient market hypothesis and modern portfolio theory.

Typically: define risk "tolerance," maybe apply a mean-variance optimizer and then invest in index funds.

We have a few, but not many, such advocates on this forum.

Have fun.

John Walter Russell

To be useful, a market timing indicator, whether statistical or anecedotal, must perform two functions.  It must get you "out" of the market before or early in a Bear market, AND it must get you "into" the market in time to catch the Bull market phase.

I believe that an all-in or all-out timing approach is a very bad idea.

I do not recommend being out of stocks entirely. I do recommend varying allocations that depend on P/E10--stocks versus alternatives. At times, I do recommend a 100% stock allocation.

The allocation shifts that I am talking about take place over a series of years, not months and seldom (if ever) in a single year.

I favor two approaches that allow you to invest up to 100% stocks today. I favor two other approaches that allow you to be on the sidelines (partially) today.

Have fun.

John Walter Russell