Using Shiller PE to Time the Market

[sengsational]

There's an interesting book chapter on the topic of how to beat pure buy and hold based on using the Shiller ten-year PE. Since I know I couldn't just sit out the market for ten years or whatever it takes, I'm not a candidate for this. It's a cool concept, though.

Long, but interesting read: PE Model

 

[gcgang]

There's an interesting book chapter on the topic of how to beat pure buy and hold based on using the Shiller ten-year PE. Since I know I couldn't just sit out the market for ten years or whatever it takes, I'm not a candidate for this. It's a cool concept, though.

Long, but interesting read: PE Model

You don't necessarily have to sit out. It's a big world out there. Find those countries with favorable CAPE, invest there.

A not unusual 1/2 to 2/3 drawdown on equities would be no fun.


Sent from my iPhone using Early Retirement Forum

 

[haha]

You don't necessarily have to sit out. It's a big world out there. Find those countries with favorable CAPE, invest there.

A not unusual 1/2 to 2/3 drawdown on equities would be no fun.


Sent from my iPhone using Early Retirement Forum

There is an ETF, GVAL by Cambria Research, that follows this rule. Their algorithm is completely explained on their website.

The PE10 idea to guide investing is about as popular as poison ivy around here. I have only one complaint-it is very hard to use in a tax efficient way if much of one's money is in taxable accounts. I tried using SPY puts, but puts cost too much when the overvalue condition may go on for a very long time, as it did in the 2nd half of the '90s. I do sell outright what I think will not cost me too much in annual tax increment, and I wait to make new investments until PE10 looks better. I have also started to use etf GVAL that I mentioned above. I am a bit down on this now. Remember 1980 crash, a lot of cheap countries went to hell too, as did the overvalued US.

Many people here have argued that there is no such thing as a valuation separate from the quoted price. Others argue that using PE10 will have you out of frothy markets such as 1999. Though I disagree with point #1, #2 is absolutely true; it's simple reality.

Ha

 

[gcgang]

Hey, Ha. I just started buying GVAL, also. Probably only have 5-10% of what I intend for it. Doing in my IRRA cuz I don't think it will be very tax efficient.


Sent from my iPhone using Early Retirement Forum

 

[target2019]

GVAL looks interesting. Got a chuckle when I went to Yahoo page for GVAL. Check the last Ask: 9,000 x 100 shares.

 

[gcgang]

I'm going to go in with an ask at $8995



Sent from my iPad using Early Retirement Forum

 

[target2019]

There's an interesting book chapter on the topic of how to beat pure buy and hold based on using the Shiller ten-year PE. Since I know I couldn't just sit out the market for ten years or whatever it takes, I'm not a candidate for this. It's a cool concept, though.

Long, but interesting read: PE Model

Chapter is from Rock Breaks Scissors: A Practical Guide to Outguessing and Outwitting Almost Everybody by William Poundstone

Rock Breaks Scissors: A Practical Guide to Outguessing and Outwitting Almost Everybody: William Poundstone: 9780316228060: Amazon.com: Books

Thanks for bringing this up. I had meant to read it.

 

[Htown Harry]

Michael Kitces had an interesting blog post this week on the subject of PE 10. (With charts!)
Shiller CAPE Market Valuation: Terrible For Market Timing, But Valuable For Long-Term Retirement Planning | Kitces.com

...The bottom line, though, is simply this: while the data does suggest that market valuation tools like Shiller CAPE are a poor predictor of short-term market performance, and may be very limited as a market timing tool to improve performance, the longer-term predictive value of Shiller CAPE and its CAEP inverse suggest that it is still relevant for planning decisions where the focal point truly is on long-term returns, from setting an appropriate safe withdrawal rate (or possibly even an optimal asset allocation glidepath) to evaluating the opportunity cost of funds for lifetime income strategies like delaying Social Security, purchasing an annuity, or considering a pension lump sum.

 

[Texas Proud]

Well, did not read the 'book'... did not look like it would be easy with the way it was copied etc...


But, I do have a question.... does this method 'work' It does not appear to with the very short time frame of the ETF... it has lost 4% when the S&P is up about 10%.... BIG difference...


The reason that I ask is that if it really had good potential it would be used a lot more... there are a lot of smart people out there with money and they would be doing something like this if it really did work...


I did go look at some history.... the lowest it has been was in Dec 1920... the market DID go up a lot in the next 10 years.... but then took the biggest dive in history.... I do not know the return for the full 20 years... as Yahoo does not go back that far... but with that kind of ride... wow... (to be fair, buy and hold with that kind of ride would feel the same)....

 

[samclem]

Michael Kitces had an interesting blog post this week on the subject of PE 10. (With charts!)
Shiller CAPE Market Valuation: Terrible For Market Timing, But Valuable For Long-Term Retirement Planning | Kitces.com

Right, thanks. I knew I had seen something very similar recently: PE 10/CAPE having poor correlation to short-term equity price moves. In the cited article, Kitces links to a 2008 article about using CAPE for determining the initial withdrawal rate for a retiree.

 

[dallas27]

If the model actually works based on historical numbers, then the major players have armies of ivy league quants using it to make bets. The market incorporates the new knowledge, becomes more efficient, and therefore the model has no value in the future.

Dont burn your time looking for tricks.

Sent from my iPhone using Early Retirement Forum

 

[sengsational]

The details in the book excerpt would have only a handful of trades per trading lifetime.

So you would be in bonds for a decade and equities for a decade and a half...for example.

That is why those smart people don't do it...they'd not be able to justify spending any time on it.

Before reading it, I was very skeptical, but once I took the time to understand it, it makes sense and is very simple

If the market continue s to operate like it has been, I have no doubt that the phenomena will continue.

 

[target2019]

If you have more than one fund or stock, there are significant trades to make.
I read the beginning chapter, and skipped to the one you linked. Still working through it.

 

[ERD50]

...


But, I do have a question.... does this method 'work' ...


The reason that I ask is that if it really had good potential it would be used a lot more... there are a lot of smart people out there with money and they would be doing something like this if it really did work...

But if it only works in the long term, not the short, I think most big/smart money managers wouldn't be very interested. They need short term results.

Though I'd think it would be of interest to anyone responsible for pension funds - but again, their 'boss' will be looking for short term results to evaluate them, most likely. If this does work long term, could it actually be a niche for us DIY types? A method that won't get arbitraged away by the big/smart money?

Right, thanks. I knew I had seen something very similar recently: PE 10/CAPE having poor correlation to short-term equity price moves. In the cited article, Kitces links to a 2008 article about using CAPE for determining the initial withdrawal rate for a retiree.

Interesting. The way I 'read' FIRECalc, the safest WR will be determined by the worst case periods, and very likely those are the periods of high market valuations. Using that conservative WR during periods of low valuations leaves you with a large pot at the end.

So historically, you might use something like this to determine that it has been safe to use a higher WR. But a historical report like FIRECalc has already set the low end for WR.

Another twist on that is an approach that someone at bogleheads was promoting, which also made sense to me (though I found his explanation rather convoluted). Take the historically safe WR for the number of years you plan for (say 3.5% for example), and go ahead and re-set your withdraw amount higher whenever your portfolio increases. So if you start with $1M and a $35K withdraw, and the portfolio 5 years later is at $1.3M, you can now start taking $45.5K, and adjusting that for inflation going forward.

If you think about it, that makes perfect sense - it is what a new retiree would do in that year. It answers that old paradox about why the 'old' retiree has to stick to his initial plan, while the 'new' retiree gets to take more after a market rise. The truth is, the 'old' retiree can increase his withdraw (because he retired in a low market valuation period). The numbers back it up.

So that is, in a way, an 'adaptive' method of applying a 'market valuation' approach. Think about it.

-ERD50

 

[target2019]

A word of caution about the PE10 number. I record this number on the last day of each month, and get it from multipl.com. However, the number I record does not always agree with the historical tabular results when I look later. Now that I am looking back say 10 years for a PE10 at the site, the number at the site is slightly different from the number I downloaded in the past.

 

[samclem]

So that is, in a way, an 'adaptive' method of applying a 'market valuation' approach. Think about it.

Re: Adaptive withdrawals. I'm having a hard time figuring out how to implement it. If you sell off a smaller % (e.g. something less than 3%) if/when valuations are high (because expected share price growth is lower) and a larger % (e.g something more than 4%) if/when valuations are low, wouldn't we be "selling high?" And it would be hard to take psychologically (selling more when the portfolio is diminishing in value).

 

[Texas Proud]

But if it only works in the long term, not the short, I think most big/smart money managers wouldn't be very interested. They need short term results.

Though I'd think it would be of interest to anyone responsible for pension funds - but again, their 'boss' will be looking for short term results to evaluate them, most likely. If this does work long term, could it actually be a niche for us DIY types? A method that won't get arbitraged away by the big/smart money?


-ERD50

Well, if you read the blog post that Target1019 posted.... it really is only good for the 'middle long term'... so, it is not good for the short term and it is not good for the very long term... just in the middle...


So, is this just a coincidence as the short term becomes the very long term (there are swings you have to take into account) or does it really predict 18 years out


I do not know... I am skeptical... as an example of coincidence there is the super bowl winner method that 'predicted' the market correctly in 37 of 47 years... which from what I can read of the accuracy of PE10, the super bowl method is a better indicator... and that is for short term which means you do not have to wait 18 years to find out if you were correct...

Does Super Bowl winner predict stock returns?

 

[haha]

I adore this thread! Like I predicted, poison ivy.

Ha

 

[ERD50]

Re: Adaptive withdrawals. I'm having a hard time figuring out how to implement it. If you sell off a smaller % (e.g. something less than 3%) if/when valuations are high (because expected share price growth is lower) and a larger % (e.g something more than 4%) if/when valuations are low, wouldn't we be "selling high?" And it would be hard to take psychologically (selling more when the portfolio is diminishing in value).

It's not a higher or lower % - it's the same HSWR (Historically Safe Withdraw Rate), but just applied to the new, higher (if it goes higher) portfolio amount.

Think again about my example - if a 3.5% WR was 100% safe historically for a particular profile, then, by definition, it was safe at any time in history. So you don't need to stick to it based on some arbitrary retirement date that puts you anywhere on those cycles - it was 100% safe at anytime in history, so you can apply it (or re-apply it) at any time!

If you were to actually apply this to the FIRECalc data (not easy to do in FIRECalc, but not so hard to think about by picking a few select years as examples), you'll quickly see that by increasing the amount you withdraw in good times, you will flatten the output range for the final year. You more efficiently us the available funds, and leave less 'on the table' for heirs (you could always set aside an amount for that purpose of you want).

Remember, the 3.5% number I'm using in this example is 100% HSWR - you cannot fail taking this in the worst case. And the worst case is when you retire at a peak, so your portfolio never goes up, so you never have a chance to use this technique and take out a higher amount. That's what makes it work - you can take out higher amounts, but only when the portfolio goes higher. Those are the cases where FIRECalc leaves money on the table.

Would it be hard to take psychologically? Could be, but like so many things, our emotional issues are due to not understanding the numbers. There is no number problem with this, it really is common sense, once you get your head around it (like I said, it explains that apparent anomaly of the two retirees a few years apart in a moving market). But if it doesn't appeal to someone on an emotional level - don't do it!

Sure, the future could be worse than the past - but that is common to all withdraw systems in one way or another.

-ERD50

 

[ERD50]

I adore this thread! Like I predicted, poison ivy.

Ha

I don't think it is correct to characterize healthy skepticism and analysis as calling out the idea as being as bad as poison ivy.

I'd say that my explanation of readjusting the withdraw calculation when the market rises is a form of using the Shiller P10. The disconnect with FIRECalc is that it does not take valuations into account on the initial withdraw, but this 'adaptive withdraw' method effectively does that as time goes on.

FIRECalc assumes you retired at the absolute worst time in history for calculating a fixed WR%, at a market peak, and is appropriately conservative. But the odds are that you didn't retire at the absolute worst time - so if your portfolio rises due to this (it never rises in the worst case scenarios), then you can start taking out more. Recall that after a (inflation adjusted) portfolio rise, your actual WR% is going to be lower than when you started - you are just taking the the WR% back to where you started, which has always succeeded in the worst case. If you happen to proceed into an actual worst case path, you are still 100% safe - you 'won', you die with zero $ but never ran out. Of course, most would buffer this to account for unpredictable variations from history and/or longer LE.

This is all based on history, but then isn't P10? I think we are actually more in agreement than disagreement, it's just a different implementation approach.

-ERD50

 

[samclem]

It's not a higher or lower % - it's the same HSWR (Historically Safe Withdraw Rate), but just applied to the new, higher (if it goes higher) portfolio amount.

So, it's just the same as taking a fixed %age of each end-of-year portfolio value, something I'd planned to do anyway. I was a brilliant, adaptive, Schiller-ite all along!

 

[ERD50]

So, it's just the same as taking a fixed %age of each end-of-year portfolio value, something I'd planned to do anyway.

Close, but with a twist.

If the portfolio drops, you still can (historically) maintain your inflation adjusted WR, which means that you would be taking a higher % of portfolio.

This is exactly what happens in the 'bad case' scenarios. If we start out with a 3.5% HSWR, you'll see that some paths dropped by an inflation adjusted 40% (probably more, but use this for the example now). So the initial 3.5% is now 5.83% (3.5 / .6 remaining portfolio). But starting with a 100% HSWR, you succeed, by definition.

I was a brilliant, adaptive, Schiller-ite all along!

Yes, I think all these various ideas end up converging, if they have any merit at all. After all, you can't magically change the outcome given what you've got, there's only so much you can do within reason and the constraints of needing money to live on. I don't think different reasonable approaches can be radically different, even if they appear to be on the surface.

-ERD50

 

[haha]

This is all based on history, but then isn't P10? I think we are actually more in agreement than disagreement, it's just a different implementation approach.

-ERD50

I've read most of your posts over the years, and I agree that I rarely find many ideas at odds with those that you express. My poison ivy comment is not based on you.

I still think the thread as a whole is hilarious. PE10, or some other relatively long term valuation based approach is very likely the only way that a buy and hold approach can be improved. I would imagine that very few of us will try to implement it. I particularly enjoy the part about if it worked, many experts would be using it, and then it wouldn't work. It is clearly not the case that a lot of money is responding to this. If it were, we would not get very high or very low values. I think it may be true that the tails will be truncated, but if so that will be clear enough before the fact.

My belief is that there is no end to unwise behavior, and paying attention to value will always help. Though it will not do what cannot be done, give short term reliable signals.

Ha

 

[CaliforniaMan]

So, it's just the same as taking a fixed %age of each end-of-year portfolio value, something I'd planned to do anyway. I was a brilliant, adaptive, Schiller-ite all along!

Close, but with a twist.

If the portfolio drops, you still can (historically) maintain your inflation adjusted WR, which means that you would be taking a higher % of portfolio.

This is exactly what happens in the 'bad case' scenarios. If we start out with a 3.5% HSWR, you'll see that some paths dropped by an inflation adjusted 40% (probably more, but use this for the example now). So the initial 3.5% is now 5.83% (3.5 / .6 remaining portfolio). But starting with a 100% HSWR, you succeed, by definition.

Yes, I think all these various ideas end up converging, if they have any merit at all. After all, you can't magically change the outcome given what you've got, there's only so much you can do within reason and the constraints of needing money to live on. I don't think different reasonable approaches can be radically different, even if they appear to be on the surface.

-ERD50

Only into my 4th month now, so everything is still a bit academic for me, but the above sort of fits into what I am anticipating doing. Withdraw at a fixed SWR (say 3.5), taking that fixed percentage of the value at the end of the year, not adjusting for inflation, or decreasing (much) in a decline, but allowing the withdrawal amounts to rise with a rising (hopefully) portfolio. This seems to pretty much match ERD50's suggestion, but maybe a little more conservative, since there is no automatic increase for inflation.

 

[sengsational]

...
I still think the thread as a whole is hilarious. PE10, or some other relatively long term valuation based approach is very likely the only way that a buy and hold approach can be improved. I would imagine that very few of us will try to implement it. I particularly enjoy the part about if it worked, many experts would be using it, and then it wouldn't work. It is clearly not the case that a lot of money is responding to this. If it were, we would not get very high or very low values.
...

Very wise. I agree on all points.

The whole book "Rock Breaks Scissors" (not just the exerpted chapter) talks about outguessing things and how most everyone believes that the "hot hand" concept is real. The force that hot hand thinkers put on the market makes this very long term outguessing possible. Life would be pretty boring if all of humanity acted rationally, hehe!