PE switching studies - proceed with caution!

[raddr]

As many of you know, I am not a fan of PE switching studies. I believe that there has been a shift over the years in the way that investors react to apparent overvaluation. This is very important in the study of SWR's because it is the first few years of retirement that are make or break for the retiree. In other words, if one experiences flat or negative returns in the first five years or so following retirement he or she faces a very significant possibility of portfolio failure down the line. A five-year period of negative returns in the middle or near the end of a retirement is not nearly so devastating as long as the initial returns were good. On the other hand, if one experiences above average market returns in the first few years of retirement than there is a very high likelihood of portfolio success as long as the portfolio withdrawal rate was reasonable to begin with.

With this in mind, I decided to look at periods when the PE10 was greater than 20 to see what the subsequent five-year real returns were. Not surprisingly, before 1980 real returns averaged close to 0% for the five years following a PE10 value of greater than 20 (Shiller's monthly data since 1871). As most of you know, the historical SWR tended to be quite low, near 4%, for a 75% or 80% stock portfolio and a 30-year payout period when the PE10 was greater than 20 to begin with. Undoubtedly, the first five years or so of poor returns played a great part in the poor performance of the retirement portfolios.

Fast forward to 1980. This is about the time that indexing became increasingly popular, stocks were more liquid, and "owning the market" became very easy and inexpensive. In return, investors became comfortable paying more for earnings relative to the pre-index days. When I looked at five-year real returns following PE10 values of greater than 20 for those years after 1980 I found very surprising results. Namely, the five-year real returns were much greater than average -- about 11.9% versus the historical norm of about 6.7%. This is shown on the following graph:





It is obvious that post-1980 a PE10 of greater than 20 did not indicate an immediate correction as was the case for much of the pre-1980 timeframe. The markedly positive five-year returns will undoubtedly lead to a higher historical SWR than would be predicted from historical studies once those numbers are known, i.e. sometime after 2010. Actually, it will take a bit longer since all of the post-1980 months with the PE10 > 20 values came after 1990. If I were a betting man, however, I'd bet the house that the historical SWR for these later years of high valuation will be much higher than was the case for the comparable high valuation years before 1980.

Given the importance of a good start in the retirement years, the data above has very important implications. Some of the PE switching studies I've seen suggest being out of the market altogether when the PE10 is greater than 20 or so. Before 1980 this would have been a good idea. Now, however, it could be a fatal mistake. History is not static and blindly following such historical studies optimized for past performance could be very hazardous to one's financial health. That said, however, for very high PE10 values (say, maybe 30 or greater) such as we saw in the late 1990s one certainly would need to be very cautious and move into asset classes that are more reasonably valued. This is what I did and it allowed me to retire at an earlier age than I would have thought possible.

My advice would be to not trust any single indicator of market value when making a huge financial decision such as drastically altering one's asset allocation, particularly if it involves getting out of the market altogether or severely minimizing one's exposure to equities. It's hard to recognize a market bubble but when an asset class is clearly overvalued I believe it is better to move to more reasonably valued asset classes rather than bail out of the market and into low yielding cash investments which would be a sure prescription for failure long-term.

[ben]

Hi Raddr!
Great analysis - and SWR discussion on a level where I can follow it -combined with some real life practical advise rather than a bunch of "we have to study/look into further" Etc. at the end.

I see myself countering the potential Early in Fire Low Returns (EFLR?) by starting of with lower withdrawal rates (say 2%-3%) and thereby countering partly any EFLR. I would then adjust up and down depending on actual growth of nest egg.

An addition could be to earn just a bit of pocket money on something - I have a bunch of ideas - but most feel too much like work .

Anybody interested in signing up for the Ben-churn and burn-investment newsletter for a measly $2400/year?

_________________
Normal; to put on clothes bought for work, go to work in car bought to get to work needed to pay for the clothes, the car and the home left empty all day in order to afford to live in it...

[caseynshan]

To my untrained eye, it appears there is still some definite correlation in your graph to high PE10 and low returns, but the definition of high, has maybe shifted from 20 to 25?

I can't remember how you figure the PE10. I tried looking at Gummy's site, but didn't spot anything... ?

[raddr]

Hi Casey,

[BenSolar]

Great post, raddr.

I second ben's thoughts.

I think it's a fair assumption that the widespread availability of low cost mutual funds, and the widespread 'knowledge' that stocks offer the highest return have raised the long term average valuation of the stock martket. Over Shiller's data I think the average PE-10 was 15.5. I wonder if the new 'average' might be as high as 20? Or if demographics problems and inflation might drive the market back down to an extended period at historic lows sometime in the next cycle down so that we look back at this period and wonder what the heck was everyone thinking.

Do you have any thoughts on what the reasonably valued asset classes are now, raddr? You've said in the past that small cap value is on the high side, but not so much as large cap. REITs have taken a bit of a whack and VGSIX is yielding about 5% which doesn't look so bad. Emerging markets have sure been on a heck of a run.

Regards,

_________________
"Do not spoil what you have by desiring what you have not; remember that what you now have was once among the things only hoped for." - Epicurus

[raddr]

Hi Ben,

[Mike]

[Dual]

Great post and I agree with your statement about moving to reasonably valued asset classes. To me, your results show that trends in the stock market can last for decades. The 1980's and 90's were, of course, a time of a historical bull market and the period preceding them from the mid 1960's to 1982 were a terrible bear market.

The lesson I get is that for LTBH to work, you have to be ready to do it for 30 years in some cases, and that is too long for retirees.

So somehow, we have to both flexible and cautious. The best asset classes change. During the 1970's, short term cash investments made the most sense, as was noted by Terhorst in his book. (TIPS would have been better if they had existed.) During the 1980's and 90's, probably an SP500 index fund was the best. Who knows what the future holds?

We have to use every tool at our disposal including measures such as PE10 and Tobin's Q. Both indicate that the SP500 and total stock market valuations are near all time highs. I am not a slice and dicer, so I guess I am going to hold cash. My preference is for a core holding of TIPS, but, in that market, prices are also astronomical so I sold about 60% of my holdings. So far, that was a good decision.

Going from a fixed strategy is uncomfortable, but I say that, like in the biological world, it's adapt or die.

Dual

[ataloss]

nice post and good use of the post 1980 data (excluded from the usual historical 30 year series) this would certainly make a reasonable person cautious about blindly applying past strategies to the future

of course an unreasonable person could invent some reason to exclude the post 1980 data (and the pre 1921 data or whatever else didn't "fit")


as gummy once said:
I can't believe that anybuddy would argue that some SWR is "95% safe", proceeding into an unknowable future**

People may think I disagree with jwr on everything. Actually I agree that a switching stragegy could increase withdrawls...I just doubt his (or my) ability to predict the optimal future strategy

_________________
Have fun.

Ataloss

[ben]

But ataloss, that is where the Ben Churn and Burn - newsletter/hedge fund/+ bonus crystal ball w. 3-D glasses comes to your rescue! Order the full package now!

_________________
Normal; to put on clothes bought for work, go to work in car bought to get to work needed to pay for the clothes, the car and the home left empty all day in order to afford to live in it...

[salaryguru]

Good stuff, raddr. Thanks for detailed analysis.

I wonder if the shift you observe occured gradually over time (increasing a little each decade) or in one large step (brought on by some technical advancement or change in regulations)?

_________________
-SG-

[raddr]

[salaryguru]

[quote="raddr"]