Calculating WR based on CAPE

[mistymoney]

so I was looking at this:

https://earlyretirementnow.com/2017/03/15/the-ultimate-guide-to-safe-withdrawal-rates-part-11-criteria/

specifically

A rule based on the Shiller CAPE: Calculate the Cyclically-Adjusted Earnings Yield (=1/CAPE) and use this as a proxy for expected equity returns. Then set the withdrawal rate to W = a + b*CAEY with the appropriate parameters for a and b. The first time I encountered this rule was at the cFIREsim site where they use this exact parameterization as their default values: a=1% and b=0.5.

But I can't figure out how this relates to the later graphs they have. Google says current CAPE of the SP500 is about 30. Based on this WR calculation, If the CAPE is greater than one, than the b*CAEY is always going to be less than .5, so the WR is going to be less than 1.5, but that isn't what the graphs indicate. This WR had an inital value of 3.94%.

So - what number are they using, and where is it found? Or what else have I missed?

[mistymoney]

lol! I figured it out!

the 1 should be .01 and added to the b* part then converted to percentage.

Based on the googled current CAPE of 30.81 would result in a WR of 2.62%. Applying to my current stache - I'm not ready fro prime time for sure!

[RWTL]

lol! I figured it out!

the 1 should be .01 and added to the b* part then converted to percentage.

Based on the googled current CAPE of 30.81 would result in a WR of 2.62%. Applying to my current stache - I'm not ready fro prime time for sure!

I was just going to reply!

In his safe withdraw spreadsheet, it looks like he's using a=1.75% b=0.5

I think this is the newest version https://earlyretirementnow.com/2018/08/29/google-sheet-updates-swr-series-part-28/

[mistymoney]

lol! I figured it out!

the 1 should be .01 and added to the b* part then converted to percentage.

Based on the googled current CAPE of 30.81 would result in a WR of 2.62%. Applying to my current stache - I'm not ready fro prime time for sure!

I was just going to reply!

In his safe withdraw spreadsheet, it looks like he's using a=1.75% b=0.5

I think this is the newest version https://earlyretirementnow.com/2018/08/29/google-sheet-updates-swr-series-part-28/

Thanks! I haven't delved into adjusting the variables on this, got a little stuck with the initial calculation. So much to read and absorb on ERN!

While the 2.62% result is a disappointment (and adding in .75% for 3.37% is still low after focusing on 4%) I feel like I've reached a better level of understanding on this. Which is great, and I'm feeling more confident about planning.

[RWTL]

lol! I figured it out!

the 1 should be .01 and added to the b* part then converted to percentage.

Based on the googled current CAPE of 30.81 would result in a WR of 2.62%. Applying to my current stache - I'm not ready fro prime time for sure!

I was just going to reply!

In his safe withdraw spreadsheet, it looks like he's using a=1.75% b=0.5

I think this is the newest version https://earlyretirementnow.com/2018/08/29/google-sheet-updates-swr-series-part-28/

Thanks! I haven't delved into adjusting the variables on this, got a little stuck with the initial calculation. So much to read and absorb on ERN!

While the 2.62% result is a disappointment (and adding in .75% for 3.37% is still low after focusing on 4%) I feel like I've reached a better level of understanding on this. Which is great, and I'm feeling more confident about planning.

The interesting thing about the spreadsheet is that ERN also has you add the future cash flows like pension, and Social Security.  He acknowledges that this can push the SWR to > 4% and that some people are too conservative.  Using his calculations, my SWR is 5% with 0% failure rate.  That's hard to swallow on the opposite end.  I find that I am still shooting for < 4% just in case.

[mcneally]

The part 11 link in OP says "there is some science behind" the 'a' and 'b' values, which looking the the post index is probably part 18, but I read that and don't see how the values are anything other than "whatever you feel like".
https://earlyretirementnow.com/2017/08/30/the-ultimate-guide-to-safe-withdrawal-rates-part-18-flexibility-cape-based-rules/

[mistymoney]

The part 11 link in OP says "there is some science behind" the 'a' and 'b' values, which looking the the post index is probably part 18, but I read that and don't see how the values are anything other than "whatever you feel like".
https://earlyretirementnow.com/2017/08/30/the-ultimate-guide-to-safe-withdrawal-rates-part-18-flexibility-cape-based-rules/

maybe it's his own science?

I know in simulations he's done, that using the equations produced 0 failures vs 4%.

Although, it seems to me a little opposite. If CAPE is high - market is overvalued so - sell extra and take the big vacay vs CAPE is low so more growth expected so don't sell/sell as little as you can?

[mistymoney]

lol! I figured it out!

the 1 should be .01 and added to the b* part then converted to percentage.

Based on the googled current CAPE of 30.81 would result in a WR of 2.62%. Applying to my current stache - I'm not ready fro prime time for sure!

I was just going to reply!

In his safe withdraw spreadsheet, it looks like he's using a=1.75% b=0.5

I think this is the newest version https://earlyretirementnow.com/2018/08/29/google-sheet-updates-swr-series-part-28/

Thanks! I haven't delved into adjusting the variables on this, got a little stuck with the initial calculation. So much to read and absorb on ERN!

While the 2.62% result is a disappointment (and adding in .75% for 3.37% is still low after focusing on 4%) I feel like I've reached a better level of understanding on this. Which is great, and I'm feeling more confident about planning.

The interesting thing about the spreadsheet is that ERN also has you add the future cash flows like pension, and Social Security.  He acknowledges that this can push the SWR to > 4% and that some people are too conservative.  Using his calculations, my SWR is 5% with 0% failure rate.  That's hard to swallow on the opposite end.  I find that I am still shooting for < 4% just in case.

That would be reassuring! But I have low confidence in my ability to fill it in correctly, lol!

I will try it sometime soon though!

[ChpBstrd]

The sudden increase in real interest rates to levels last seen 17 years ago makes me wish ERN would analyze the optimal AA and WR given any particular level of real rates. Unfortunately, the blog's advice is generally agnostic about economic conditions, except for his important work on valuation. Retirement at ATH's or high CAPEs has historically been very, very risky. But what do we know about retiring right after a historically massive rate hiking campaign?

In 2012 or 2013, when real rates were near zero, any bond allocation was dead weight on the portfolio, essentially forcing stocks to do more of the work. Thus a very low fixed income allocation was probably rational in those times. Like 0-10%! Plus low real yields were highly stimulative.

Today though, bonds can beat inflation plus deliver about 2-4% real returns. If real returns keep rising, it would make sense to have a bond-heavy portfolio doing the double duty of earning one's living expenses and protecting against SORR. Plus, high real yields make it harder for businesses to find growth opportunities that can clear the hurdle of higher interest costs. 60/40 and even 50/50 portfolios are in style again for good reasons, but ERN has not yet delved into the analysis.

I suspect if he did the analysis, there would be a slightly different optimal AA and WR for each kind of interest rate regime.

[mistymoney]

The sudden increase in real interest rates to levels last seen 17 years ago makes me wish ERN would analyze the optimal AA and WR given any particular level of real rates. Unfortunately, the blog's advice is generally agnostic about economic conditions, except for his important work on valuation. Retirement at ATH's or high CAPEs has historically been very, very risky. But what do we know about retiring right after a historically massive rate hiking campaign?

In 2012 or 2013, when real rates were near zero, any bond allocation was dead weight on the portfolio, essentially forcing stocks to do more of the work. Thus a very low fixed income allocation was probably rational in those times. Like 0-10%! Plus low real yields were highly stimulative.

Today though, bonds can beat inflation plus deliver about 2-4% real returns. If real returns keep rising, it would make sense to have a bond-heavy portfolio doing the double duty of earning one's living expenses and protecting against SORR. Plus, high real yields make it harder for businesses to find growth opportunities that can clear the hurdle of higher interest costs. 60/40 and even 50/50 portfolios are in style again for good reasons, but ERN has not yet delved into the analysis.

I suspect if he did the analysis, there would be a slightly different optimal AA and WR for each kind of interest rate regime.

I feel we are watching a different channel.

[Turtle]

This is something I heard about on a podcast, but it was while I was driving and I had completely forgotten.  Thank you for the reminder to look into it. 

[VanillaGorilla]

I really struggle with CAPE based arguments. Typically people say things like "the historical mean is 17, that means whenever the CAPE is above 17, equities are overvalued!". But since I was born, the CAPE has been greater than 17 for all but about a single year, and the market has returned 10.4% over that same period. If one only bought equities at sub-mean CAPE, you would have needed to do all your purchasing for a single six month period after the great financial crisis in 2009.

The mean CAPE over the last 38 years has been about 25. So does that mean any time the CAPE is > 25 we shouldn't buy equities? Because 2013 is the last time that was the case, and the market has returned 11.6% since then. This doesn't pass the smell test. In short, CAPE based investing rules seem wildly conservative, and I'd much rather just DCA into the market ignoring macroeconomic trends.

I love ERN's analyses but the CAPE based approaches is one of the (very few!) weaker arguments that he makes, imho (along with his mortgage in retirement essay). I haven't seen many compelling explanations of why the CAPE should always revert to its historical mean. So much has changed, macro-economically, over the decades. Monetary policy is evolving. The biggest companies today sell software instead of material goods. Our economy is increasingly non-material, why would the market's concept of a 'fair' price stay the same over decades?

My background is in computational science, not economics, so I certainly don't have the academic or professional credentials to seriously critique ERN's methodology, but I'm leery of the Shiller CAPE.

Ultimately, everybody seems to agree on the fundamentals: a historically completely safe withdrawal rate is 3%. 4% is somewhat risky for early retirees. Choose something in between according to your risk tolerance, the prevailing market conditions, supplemental cash flows, etc. Trying to overfit models to a tiny dataset of something as heuristic-y as the CAPE seems unlikely to be incredibly useful.

[wageslave23]

I really struggle with CAPE based arguments. Typically people say things like "the historical mean is 17, that means whenever the CAPE is above 17, equities are overvalued!". But since I was born, the CAPE has been greater than 17 for all but about a single year, and the market has returned 10.4% over that same period. If one only bought equities at sub-mean CAPE, you would have needed to do all your purchasing for a single six month period after the great financial crisis in 2009.

The mean CAPE over the last 38 years has been about 25. So does that mean any time the CAPE is > 25 we shouldn't buy equities? Because 2013 is the last time that was the case, and the market has returned 11.6% since then. This doesn't pass the smell test. In short, CAPE based investing rules seem wildly conservative, and I'd much rather just DCA into the market ignoring macroeconomic trends.

I love ERN's analyses but the CAPE based approaches is one of the (very few!) weaker arguments that he makes, imho (along with his mortgage in retirement essay). I haven't seen many compelling explanations of why the CAPE should always revert to its historical mean. So much has changed, macro-economically, over the decades. Monetary policy is evolving. The biggest companies today sell software instead of material goods. Our economy is increasingly non-material, why would the market's concept of a 'fair' price stay the same over decades?

My background is in computational science, not economics, so I certainly don't have the academic or professional credentials to seriously critique ERN's methodology, but I'm leery of the Shiller CAPE.

Ultimately, everybody seems to agree on the fundamentals: a historically completely safe withdrawal rate is 3%. 4% is somewhat risky for early retirees. Choose something in between according to your risk tolerance, the prevailing market conditions, supplemental cash flows, etc. Trying to overfit models to a tiny dataset of something as heuristic-y as the CAPE seems unlikely to be incredibly useful.

1. Nobody is saying don't buy stocks when CAPE is high. Just expect a lower than average return historically. Ie adjust wr from 4 to 3.5%. But it's still positive returns on average.

2. Present day CAPE is going to be different than historical CAPE because of changes in accounting methods.  ERN adjusts for the change in modern CAPE. 

3. It's just common sense, CAPE is a measure of how expensive stocks are compared to their earnings.  When they are expensive then they probably won't yield as much on average going forward than when they are cheaply priced. This is a broad range but has held up historically so it makes sense to build in a little extra security during times of highly expensive stocks.

4. If you bought back in 2020 when cape was high even for modern times, you would have an annualized real return of <4%. Which is much less than the historical average stock market return. So current market behavior still upholds the historical inverse relationship between CAPE and future returns.

[FIRE 20/20]

lol! I figured it out!

the 1 should be .01 and added to the b* part then converted to percentage.

Based on the googled current CAPE of 30.81 would result in a WR of 2.62%. Applying to my current stache - I'm not ready fro prime time for sure!

The post you're referring to is from 2017, 5 years before ERN finally updated to a more reasonable understanding of CAPE.  After writing "Building a Better CAPE", he looked at safe withdrawal rates and found that, "The 4% Rule Works Again! An Update on Dynamic Withdrawal Rates based on the Shiller CAPE – SWR Series Part 54".  I would comment that it never stopped working, he was just doing great analysis on incorrect data for most of the series.   

I love what he's done with his series, but until he adjusted for the persistently (and explainable!) elevated CAPE ratios of the last quarter century, everything he did was excessively pessimistic.  You don't need to use a 2.62% withdrawal rate!  Even ERN, who is one of the most pessimistic FIRE blogger out there, is now saying 4% is safe. 

I'll quote part of his conclusion:

Granted, but I still see folks applying 3% and even sub-3% withdrawal rates in today’s market. Relax, everybody, the risk of another bad market event on top of the current drawdown is low!

Likewise, with the adjusted CAPE quite close to dropping below 20 and the S&P 500 dropping more than 25% in real terms since the beginning of the year, I am also ready to announce that even in the traditional static SWR calculations, we should now safely move the withdrawal rate to 4% and above. Well, you heard it here first; the 4% Rule works again! And with a little bit of flexibility and a generous pension and Social Security benefits later in retirement, you can certainly go crazy and justify 4.5% or higher!

https://earlyretirementnow.com/2022/10/05/building-a-better-cape-ratio/
https://earlyretirementnow.com/2022/10/12/dynamic-withdrawal-rates-based-on-the-shiller-cape-swr-series-part-54/

[mcneally]

I'll quote part of his conclusion:

Granted, but I still see folks applying 3% and even sub-3% withdrawal rates in today’s market. Relax, everybody, the risk of another bad market event on top of the current drawdown is low!

Likewise, with the adjusted CAPE quite close to dropping below 20 and the S&P 500 dropping more than 25% in real terms since the beginning of the year, I am also ready to announce that even in the traditional static SWR calculations, we should now safely move the withdrawal rate to 4% and above. Well, you heard it here first; the 4% Rule works again! And with a little bit of flexibility and a generous pension and Social Security benefits later in retirement, you can certainly go crazy and justify 4.5% or higher!

https://earlyretirementnow.com/2022/10/05/building-a-better-cape-ratio/
https://earlyretirementnow.com/2022/10/12/dynamic-withdrawal-rates-based-on-the-shiller-cape-swr-series-part-54/

I certainly don't advocate a 2.62% withdrawal rate, but the CAPE is 19% higher now than when this ERN update was written in Oct 2022.

[ChpBstrd]

CAPE is sensitive to the rate of earnings growth. As an extreme example, if earnings doubled every year, CAPE would be very high, even if PE was held to a reasonable level. To illustrate:

Years ago     Hypothetical Earnings
10                $1
9                  $2
8                  $4
....                ...
1                  $512

10y Avg:       $102.30

At a PE of 20, CAPE in this example would be ((512*20)/102.3)= 100. A CAPE this high wouldn't mean the market was overpriced. If anything, 20x earnings is a low price to pay for earnings that double each year! It's the fast pace of growth that is responsible for the high CAPE in this crazy example.

So the question is, can we explain historically high CAPE with a faster than historically normal earnings growth? I found the following linear-scale chart which makes it look like the earnings growth rate accelerated over time. Perhaps someone can do the math to prove it.


For what it's worth, the PEG ratio of the S&P500 is in about the middle of its range, so we can extrapolate that the above average PE ratio reflects faster than average growth. According to multpl.com, the S&P500 earned more in the first 6 months of 2023 than it did in all of 2022, so maybe my example wasn't so crazy after all.

CAPE does offer one important reality check though: Can earnings for this stock index continue to grow so much faster than GDP, productivity, and population forever, or will such growth reach its natural limit? In that sense, a dip in CAPE might represent a cyclical buying opportunity.

For valuation purposes though, I think PEG is a better growth-adjusted valuation metric than CAPE, especially if the growth component is averaged over a few years.

[Telecaster]

Although, it seems to me a little opposite. If CAPE is high - market is overvalued so - sell extra and take the big vacay vs CAPE is low so more growth expected so don't sell/sell as little as you can?

No, because a high CAPE implies lower future returns.   If you spend your nut now and an experience low returns you run the risk of going bust.   Conversely, a low CAPE implies higher future returns, so spending more money would be safer. 

[VanillaGorilla]

CAPE is sensitive to the rate of earnings growth.

Exactly. I also wonder about the validity/utility of looking at average earnings over a decade. That might be useful for a slow moving blue chip stock but it seems barely applicable to the current crop of wildly volatile tech stocks. Tesla barely existed ten years ago.

[mistymoney]

Although, it seems to me a little opposite. If CAPE is high - market is overvalued so - sell extra and take the big vacay vs CAPE is low so more growth expected so don't sell/sell as little as you can?

No, because a high CAPE implies lower future returns.   If you spend your nut now and an experience low returns you run the risk of going bust.   Conversely, a low CAPE implies higher future returns, so spending more money would be safer.

well hopefully that big vacay isn't going to be the whole nut!

I think we understand it the same - but are extrapolating different actions.

[Telecaster]

Exactly. I also wonder about the validity/utility of looking at average earnings over a decade. That might be useful for a slow moving blue chip stock but it seems barely applicable to the current crop of wildly volatile tech stocks. Tesla barely existed ten years ago.

The theory is that you're taking into account depressed or inflated earnings due to seasonality, economic conditions, etc. as well as short term market irrationality and smoothing that out to get a better idea of the "true" value.   AFAIK, most analysts use time periods shorter than 10 years for individual stock analysis. 

While I admire all the work ERN has shared with us, he calculates his SWRs two points past the decimal.   I simply do not believe past results will be that accurate in the future.   He's implying an accuracy that in my mind does not exist.   Future data will be different so future SWRs using the same methods will be different too.   

The thing about CAPE, for me, is that I don't believe it is actionable except in very general ways.   In accumulation phase in times of high CAPE you should almost certainly just keep buying stocks.   In retirement, you've already started your WR.   Best is you can maybe decide to delay or accelerate retirement a bit.   

[wageslave23]

Exactly. I also wonder about the validity/utility of looking at average earnings over a decade. That might be useful for a slow moving blue chip stock but it seems barely applicable to the current crop of wildly volatile tech stocks. Tesla barely existed ten years ago.

The theory is that you're taking into account depressed or inflated earnings due to seasonality, economic conditions, etc. as well as short term market irrationality and smoothing that out to get a better idea of the "true" value.   AFAIK, most analysts use time periods shorter than 10 years for individual stock analysis. 

While I admire all the work ERN has shared with us, he calculates his SWRs two points past the decimal.   I simply do not believe past results will be that accurate in the future.   He's implying an accuracy that in my mind does not exist.   Future data will be different so future SWRs using the same methods will be different too.   

The thing about CAPE, for me, is that I don't believe it is actionable except in very general ways.   In accumulation phase in times of high CAPE you should almost certainly just keep buying stocks.   In retirement, you've already started your WR.   Best is you can maybe decide to delay or accelerate retirement a bit.

I 100% agree with this. My only knock on ERN is his precision.  We are only looking at historical data for guidance about the future. He gets too caught up in the trees. And your last sentence is exactly what the take away from his CAPE analysis should be - a little more caution when retiring when CAPE is high, or markets have been on a long run, etc. When things are going well, don't be over exuberant, when things look bleak, know they aren't as bad as they seem.  Otherwise it's business as usual. 

[RWTL]

Exactly. I also wonder about the validity/utility of looking at average earnings over a decade. That might be useful for a slow moving blue chip stock but it seems barely applicable to the current crop of wildly volatile tech stocks. Tesla barely existed ten years ago.

The theory is that you're taking into account depressed or inflated earnings due to seasonality, economic conditions, etc. as well as short term market irrationality and smoothing that out to get a better idea of the "true" value.   AFAIK, most analysts use time periods shorter than 10 years for individual stock analysis. 

While I admire all the work ERN has shared with us, he calculates his SWRs two points past the decimal.   I simply do not believe past results will be that accurate in the future.   He's implying an accuracy that in my mind does not exist.   Future data will be different so future SWRs using the same methods will be different too.   

The thing about CAPE, for me, is that I don't believe it is actionable except in very general ways.   In accumulation phase in times of high CAPE you should almost certainly just keep buying stocks.   In retirement, you've already started your WR.   Best is you can maybe decide to delay or accelerate retirement a bit.

I 100% agree with this. My only knock on ERN is his precision.  We are only looking at historical data for guidance about the future. He gets too caught up in the trees. And your last sentence is exactly what the take away from his CAPE analysis should be - a little more caution when retiring when CAPE is high, or markets have been on a long run, etc. When things are going well, don't be over exuberant, when things look bleak, know they aren't as bad as they seem.  Otherwise it's business as usual.

As someone who was trained in the sciences, I appreciate the precision.  Are you saying you would feel better if he "winged" it more?  I don't believe he is using precision to say you can exactly remove a certain amount.  He's using it to compare alternatives while minimizing the risk of rounding error within each step of the analysis. 

I'm a big fan of ERN, so I may be defending him a bit.  That's also a possibility.

[jeroly]

Although, it seems to me a little opposite. If CAPE is high - market is overvalued so - sell extra and take the big vacay vs CAPE is low so more growth expected so don't sell/sell as little as you can?

No, because a high CAPE implies lower future returns.   If you spend your nut now and an experience low returns you run the risk of going bust.   Conversely, a low CAPE implies higher future returns, so spending more money would be safer.

Sell extra, yes (i.e. sell high; diversify away from expensive stocks; reallocate back to baseline allocation) but don't spend extra

[Telecaster]

As someone who was trained in the sciences, I appreciate the precision.  Are you saying you would feel better if he "winged" it more?  I don't believe he is using precision to say you can exactly remove a certain amount.  He's using it to compare alternatives while minimizing the risk of rounding error within each step of the analysis. 

I'm a big fan of ERN, so I may be defending him a bit.  That's also a possibility.

I'm trained in the sciences too, and that's why I made my comment.  The number of significant digits implies that the value to the right is "known."   But I don't think it is.   For example, if ERN used daily values instead of monthly that final digit would almost certainly change.    Same with hourly vs. daily.   

And let's look at the CAPE rule:  Intercept=1.75% and slope of 0.50.  Those numbers seem oddly round.  Are we sure the "true" values aren't 1.77% and 0.48? 

As we all know SWRs are necessarily backward looking.   Which is fine, looking at history is the only tool we have.  But if we were to calculate the SWR for today, and it came out to be say, 3.71%,  we can all be virtually certain that at the end of 30 years the "true" SWR will be some different number.  Close maybe, but fantastically unlikely it would actually be that number.   

[wageslave23]

Exactly. I also wonder about the validity/utility of looking at average earnings over a decade. That might be useful for a slow moving blue chip stock but it seems barely applicable to the current crop of wildly volatile tech stocks. Tesla barely existed ten years ago.

The theory is that you're taking into account depressed or inflated earnings due to seasonality, economic conditions, etc. as well as short term market irrationality and smoothing that out to get a better idea of the "true" value.   AFAIK, most analysts use time periods shorter than 10 years for individual stock analysis. 

While I admire all the work ERN has shared with us, he calculates his SWRs two points past the decimal.   I simply do not believe past results will be that accurate in the future.   He's implying an accuracy that in my mind does not exist.   Future data will be different so future SWRs using the same methods will be different too.   

The thing about CAPE, for me, is that I don't believe it is actionable except in very general ways.   In accumulation phase in times of high CAPE you should almost certainly just keep buying stocks.   In retirement, you've already started your WR.   Best is you can maybe decide to delay or accelerate retirement a bit.

I 100% agree with this. My only knock on ERN is his precision.  We are only looking at historical data for guidance about the future. He gets too caught up in the trees. And your last sentence is exactly what the take away from his CAPE analysis should be - a little more caution when retiring when CAPE is high, or markets have been on a long run, etc. When things are going well, don't be over exuberant, when things look bleak, know they aren't as bad as they seem.  Otherwise it's business as usual.

As someone who was trained in the sciences, I appreciate the precision.  Are you saying you would feel better if he "winged" it more?  I don't believe he is using precision to say you can exactly remove a certain amount.  He's using it to compare alternatives while minimizing the risk of rounding error within each step of the analysis. 

I'm a big fan of ERN, so I may be defending him a bit.  That's also a possibility.

I compare it to trying to plan for the weather tomorrow.  The weather app can be as precise as they want to be but all I really care about is a general temp prediction.  They don't need to tell me it's going to be 54.57 degrees tomorrow.  First, I don't care, I just need to know it will be mid 50s. Second, it's only a prediction so don't add the illusion of certainty by adding decimal places. With regards to swr, just knowing that high CAPE ratios means you might want to shoot for a lower wr would be sufficient like suggesting a 3 or 3.5% wr instead of 4.

[ChpBstrd]

Exactly. I also wonder about the validity/utility of looking at average earnings over a decade. That might be useful for a slow moving blue chip stock but it seems barely applicable to the current crop of wildly volatile tech stocks. Tesla barely existed ten years ago.

The theory is that you're taking into account depressed or inflated earnings due to seasonality, economic conditions, etc. as well as short term market irrationality and smoothing that out to get a better idea of the "true" value.   AFAIK, most analysts use time periods shorter than 10 years for individual stock analysis. 

While I admire all the work ERN has shared with us, he calculates his SWRs two points past the decimal.   I simply do not believe past results will be that accurate in the future.   He's implying an accuracy that in my mind does not exist.   Future data will be different so future SWRs using the same methods will be different too.   

The thing about CAPE, for me, is that I don't believe it is actionable except in very general ways.   In accumulation phase in times of high CAPE you should almost certainly just keep buying stocks.   In retirement, you've already started your WR.   Best is you can maybe decide to delay or accelerate retirement a bit.

I 100% agree with this. My only knock on ERN is his precision.  We are only looking at historical data for guidance about the future. He gets too caught up in the trees. And your last sentence is exactly what the take away from his CAPE analysis should be - a little more caution when retiring when CAPE is high, or markets have been on a long run, etc. When things are going well, don't be over exuberant, when things look bleak, know they aren't as bad as they seem.  Otherwise it's business as usual.

As someone who was trained in the sciences, I appreciate the precision.  Are you saying you would feel better if he "winged" it more?  I don't believe he is using precision to say you can exactly remove a certain amount.  He's using it to compare alternatives while minimizing the risk of rounding error within each step of the analysis. 

I'm a big fan of ERN, so I may be defending him a bit.  That's also a possibility.

I compare it to trying to plan for the weather tomorrow.  The weather app can be as precise as they want to be but all I really care about is a general temp prediction.  They don't need to tell me it's going to be 54.57 degrees tomorrow.  First, I don't care, I just need to know it will be mid 50s. Second, it's only a prediction so don't add the illusion of certainty by adding decimal places. With regards to swr, just knowing that high CAPE ratios means you might want to shoot for a lower wr would be sufficient like suggesting a 3 or 3.5% wr instead of 4.

Perhaps we should report the precise math as it comes out - i.e. to two decimal places - but draw broader conclusions from the data.

Example 1: Instead of saying the SWR "is" something, we should clarify that it "was" or "has been" something. In this sense the precise numbers are absolutely correct because they were set in stone long ago. Also this little change of tense means we are not claiming the future will look exactly like the past (possible extrapolation fallacy, in the absence of other arguments), we are saying the past was something.

Example 2: Instead of saying the future SWR is expected to be 3.6759421%, for example, generalize and say the SWR is expected to be somewhere between 3.5% and 4%. Maybe our model spit out 3.6759421% but it's not realistic to draw such specific conclusions down to fractions of a penny. It's also slightly untruthful to say our model spit out 3.7%. The place to pad our estimate is in the conclusion-drawing step, not the math and modeling step.

I'm not accusing ERN of these specific verbal errors, and I don't have the time to review every post to call them out. But I do sometimes see these fallacies in comments, which become straw men for other comments, which necessitate a discussion about nuance, ad nauseam. The need for technical precision, fallacy avoidance, and careful drawing of conclusions is why scientists or economists take months of full time work to write a 20 page paper. But we're talking about a blog and the 2 steps above would go a long way to help bloggers retain credibility.

[wageslave23]

Exactly. I also wonder about the validity/utility of looking at average earnings over a decade. That might be useful for a slow moving blue chip stock but it seems barely applicable to the current crop of wildly volatile tech stocks. Tesla barely existed ten years ago.

The theory is that you're taking into account depressed or inflated earnings due to seasonality, economic conditions, etc. as well as short term market irrationality and smoothing that out to get a better idea of the "true" value.   AFAIK, most analysts use time periods shorter than 10 years for individual stock analysis. 

While I admire all the work ERN has shared with us, he calculates his SWRs two points past the decimal.   I simply do not believe past results will be that accurate in the future.   He's implying an accuracy that in my mind does not exist.   Future data will be different so future SWRs using the same methods will be different too.   

The thing about CAPE, for me, is that I don't believe it is actionable except in very general ways.   In accumulation phase in times of high CAPE you should almost certainly just keep buying stocks.   In retirement, you've already started your WR.   Best is you can maybe decide to delay or accelerate retirement a bit.

I 100% agree with this. My only knock on ERN is his precision.  We are only looking at historical data for guidance about the future. He gets too caught up in the trees. And your last sentence is exactly what the take away from his CAPE analysis should be - a little more caution when retiring when CAPE is high, or markets have been on a long run, etc. When things are going well, don't be over exuberant, when things look bleak, know they aren't as bad as they seem.  Otherwise it's business as usual.

As someone who was trained in the sciences, I appreciate the precision.  Are you saying you would feel better if he "winged" it more?  I don't believe he is using precision to say you can exactly remove a certain amount.  He's using it to compare alternatives while minimizing the risk of rounding error within each step of the analysis. 

I'm a big fan of ERN, so I may be defending him a bit.  That's also a possibility.

I compare it to trying to plan for the weather tomorrow.  The weather app can be as precise as they want to be but all I really care about is a general temp prediction.  They don't need to tell me it's going to be 54.57 degrees tomorrow.  First, I don't care, I just need to know it will be mid 50s. Second, it's only a prediction so don't add the illusion of certainty by adding decimal places. With regards to swr, just knowing that high CAPE ratios means you might want to shoot for a lower wr would be sufficient like suggesting a 3 or 3.5% wr instead of 4.

Perhaps we should report the precise math as it comes out - i.e. to two decimal places - but draw broader conclusions from the data.

Example 1: Instead of saying the SWR "is" something, we should clarify that it "was" or "has been" something. In this sense the precise numbers are absolutely correct because they were set in stone long ago. Also this little change of tense means we are not claiming the future will look exactly like the past (possible extrapolation fallacy, in the absence of other arguments), we are saying the past was something.

Example 2: Instead of saying the future SWR is expected to be 3.6759421%, for example, generalize and say the SWR is expected to be somewhere between 3.5% and 4%. Maybe our model spit out 3.6759421% but it's not realistic to draw such specific conclusions down to fractions of a penny. It's also slightly untruthful to say our model spit out 3.7%. The place to pad our estimate is in the conclusion-drawing step, not the math and modeling step.

I'm not accusing ERN of these specific verbal errors, and I don't have the time to review every post to call them out. But I do sometimes see these fallacies in comments, which become straw men for other comments, which necessitate a discussion about nuance, ad nauseam. The need for technical precision, fallacy avoidance, and careful drawing of conclusions is why scientists or economists take months of full time work to write a 20 page paper. But we're talking about a blog and the 2 steps above would go a long way to help bloggers retain credibility.

Yes this is it. The calculations are great. The conclusions whether by the blog or readers should be generalized. Just like a scientific study, you report the precise results but then summarize and generalize in the conclusion.

[RWTL]

@wageslave23 and @ChpBstrd I agree with both of you.  There's the science and then what you will do with that information since we all seem to agree that the future is unpredictable.  But, we have to pick a number.  Personally, I find comfort in ERN's math, and then I reduce it down further - just in case which is my own hedge against anything else that could come up. 

I also discount future Social Security cash flows by 20% - just in case the government doesn't pull a rabbit out of their hat.

I hope to be pleasantly surprised by excess returns based on all my downward rounding - rather than running out of money and eating rice and beans.