Author Topic: Rising Equity Glide Path & very early retirement in today's market environment  (Read 6906 times)

Mr. Green

  • Magnum Stache
  • ******
  • Posts: 4485
  • Age: 40
  • Location: Wilmington, NC
There have been a couple threads on the forum (bottom) that have discussed recent research by Wade Pfau and Michael Kitces on using a Rising Equity Glide Path (REGP) in the first decade or so of retirement. I feel like I might have read another thread with REGP content in it some time ago but I could be wrong. I was hoping maybe some other here who have looked at REGP might chime in on my train of thought that follows.

I'm no longer working, though we're not drawing on our stash yet. My wife's income currently supports our expenses but she may quit at the end of the year so our transition to living off our investments is somewhat imminent. Our stash is basically all stocks and bonds, allocated 85 stocks/10 bonds/5 cash. We have some real estate that may contribute in the long-term future but for all intents and purposes we are retiring on our $1 million in investments ($40k in planned expenses). We're in our early 30's so we also hope to see two 30-year "retirement periods," though we expect Social Security income to reduce the drain on our portfolio by a healthy amount, even after expected future reductions.

Most folks in the financial world tend to agree that the US equities market looks overvalued at the moment. For the sake of this conversation I don't want to get hung up in how overvalued it could be based on CAPE, the recent changes to how they measure CAPE, etc. I tend to look at the broader strokes. When lots of professionals in any one field say X, it's usually relevant to some degree. I've been thinking about the odds of lower returns in the near future, and what that could mean for someone retiring in their 30's now-ish (us). I also consider the psychological aspect of this from the standpoint of someone who has left a career he would not go back to, which seems to be a common theme for many people.

When considering a shift to the REGP (something like 30 stocks/70 bonds at the moment of retirement) in such an early retirement, I wonder about the potential scenario where a reduced return early on due to REGP might cause a portfolio failure later in life, given the longevity of the retirement. Obviously we're talking about a scenario that's really on the margins, almost worst case scenario type stuff. Perhaps the infusion of SS into the second retirement period would make this a non-issue. Conversely, the thought occurred to me that sticking with an equities heavy allocation while young might be better in that if there is a bad sequence of returns in the first ten years, we're more able to earn income while young to shore up the portfolio for a few years, as opposed to the potential scenario described above where reduced returns with REGP early on leave us facing the prospect of work later in life.

Again, we're really on the margins here but it's just something I was thinking about. The potential difference is that the smaller losses in a bad sequence of returns with REGP might keep us comfortable drawing money from our portfolio, causing a greater draw down over the long run, instead of a pause in withdrawals during a couple bad years while we resumed working to pay for expenses.

Naturally, we have safety margins built in. We can reduce spending, etc. I don't plan to blindly take our 4% regardless of what the market does. I'm looking at this more from the thought exercise point of view, though it obviously has practical application. Maybe I'm totally overthinking it, given just how on the margins these scenarios would be. Yet that's kinda why we make asset allocation choices and study different concepts like REGP, to try and better cover the unfavorable scenarios. Thoughts?

https://forum.mrmoneymustache.com/post-fire/using-the-rising-equity-glidepath-to-reduce-sequence-of-returns-risk/

https://forum.mrmoneymustache.com/investor-alley/changing-asset-allocation-as-fire-approaches/
« Last Edit: September 07, 2017, 09:41:16 AM by Mr. Green »

Grizzly Dad

  • 5 O'Clock Shadow
  • *
  • Posts: 25
It's a really good question and something I've thought a lot about recently.

Around here, and with good reason, it's almost taken as gospel that you should always stay in the market. And this is pretty damn good advice.

However, looking at the data available would suggest that anyone expecting anything close to historic stock market returns from the present valuation levels is smoking something funky. Does that mean stocks won't do well or better than bonds or other asset classes? Who knows at this point. But CAPE, Adjusted GVA to market cap, Rev to market cap are all pretty highly correlated (>.9) to subsequent market returns over long time horizons and all show that we should be somewhat pessimistic about what to expect from the stock market over the next ten years. Over the next ten years Nominal returns are more likely to be hovering around 1-3% rather than what has been the norm historically with the likelihood of a large drawdown in the in intervening years high. We're unfortunately looking at a scenario in the current market in which stock returns are only marginally higher than the craptastic yields available in bond markets.

Does this mean it's correct to switch? Uncertain. But at this point, I would definitely expect the ride over the next 10 years to be much smoother in a portfolio weighted less heavily in US equities without the loss of huge amounts of upside.

We just did something similar in our portfolio, adjusting to a heavier international and bond weighting due to 1)Extream overvaluation in the US market 2)Reaching the other side of our traditional jobs and a desire too much less volatility than we'd have with a 80-90% US equities portfolio. I'll trade what's likely to be about 1% of nominal yield per year to avoid having a massive loss of 30-40% some year when we might need it. The equity glide path has a lot to recommend it, particularly for a very high valuation period like we face now.
« Last Edit: September 07, 2017, 12:01:40 PM by Grizzly Dad »

talltexan

  • Walrus Stache
  • *******
  • Posts: 5344
Assuming you downshift to 30% equities, can you tilt those equities toward something high-octain (like the Larry Swedroe Portfolio), like Small Cap/Value funds?

DrF

  • Bristles
  • ***
  • Posts: 464
« Last Edit: September 07, 2017, 01:32:59 PM by DrF »

Mr. Green

  • Magnum Stache
  • ******
  • Posts: 4485
  • Age: 40
  • Location: Wilmington, NC
I was doing some experimenting with cfiresim and I'm a bit surprised by the results. I know Kitces work was based on Monte Carlo simulations, and I think I may have read in one of the articles where they said the notion that a REGP scenario works out better does not hold true when only looking at historical data. I'm generally not a fan off Monte Carlo sims because it can theoretically put two 1929's back to back, which I don't think will ever happen, barring the collapse of civilization. cfiresim certainly appears to confirm that mention about historical data. Using a REGP from 30 stocks/70 bonds to 70 stocks/30 bonds over the first 15 years of a 30 year period drops the success rate by 10% compared to a 90/10 for the entire period. Even the periods starting with the stagflation of the 60's weren't improved by a REGP allocation over the first 15 years. They were just even more in the red at the end of 30 years. This reinforces my concerns about lower returns early on due to REGP causing portfolio failure later in the period. With an extreme early retirement it's not enough for the money to just last the 30 years, it needs to last and still be enough to go the distance for another 30 with some help from SS.
« Last Edit: September 07, 2017, 08:14:56 PM by Mr. Green »

Radagast

  • Magnum Stache
  • ******
  • Posts: 2538
  • One Does Not Simply Work Into Mordor
GerardC did an analysis here:
https://forum.mrmoneymustache.com/welcome-to-the-forum/cfiresim-severely-overestimates-success-rates-for-mustachians/msg1634628/#msg1634628

But I like my version (or something similar) better because I'd prefer to always have at least 10% bonds if I was living on my money:
https://forum.mrmoneymustache.com/welcome-to-the-forum/cfiresim-severely-overestimates-success-rates-for-mustachians/msg1637810/#msg1637810

You need an equity bias at all times, so do not follow Kitces. If you have less than 50% stocks you are doing something wrong in pretty much all cases. I think a reverse glide path may be a best practice, but I'll look into it more when I get closer to needing one.

AdrianC

  • Handlebar Stache
  • *****
  • Posts: 1211
  • Location: Cincinnati
It seems to me the best place to have our money is in productive assets, and if we don't want to be involved in business that means owning stocks. Keep some cash/intermediate bonds around so we don't have to sell stocks at the very bottom. Buffets advice of 90/10 feels about right to me, and so that's what we do, more or less. I do differ from Buffett in that I diversify internationally. I'm not native born American and don't quite have his super optimistic view of America.

Rising equity glide paths, small cap value, Golden Butterfly, MPT, isn't all this stuff is based on back tests? I don't really buy it.

maizefolk

  • Walrus Stache
  • *******
  • Posts: 7396
All things being equal, I agree with you that it's better to have either a portfolio failure or obviously be on a track toward failure in your 30s or 40s than in your 70s-80s. Actual failure is quite easy to define, but I'm not sure what criteria to use for "obviously on a path towards failure." If you can think of one, that'd be another good factor to incorporate into scoring different portfolios.

I did want to add that the problem with monte carlo sims isn't that they're more likely to put two really bad or two really good years next to each other, but in fact that they're not as likely to do that as we see in the actually historical data. The total return of 30 year scenarios assembled with monte carlo data will exhibit a smaller standard deviation than the standard deviation seen in real historical 30 year intervals.


Shane

  • Handlebar Stache
  • *****
  • Posts: 1665
  • Location: Midtown
Conversely, the thought occurred to me that sticking with an equities heavy allocation while young might be better in that if there is a bad sequence of returns in the first ten years, we're more able to earn income while young to shore up the portfolio for a few years, as opposed to the potential scenario described above where reduced returns with REGP early on leave us facing the prospect of work later in life.

^^^This

Because you guys are so young, I'd err on the side of having too much equities, rather than too little. Worst case scenario, you could cut back to only spending dividends from your (mostly) equities portfolio, don't sell (m)any shares while prices are low, and both get fun jobs doing something you enjoy to earn enough $ to make up the difference. Because your retirement will be so long, you NEED an extremely high % of equities. In your 30's, 40's and 50's, it will be easy for you both to get (part-time) jobs to earn enough so that you won't have to sell (m)any shares during downturns. Because of your age, I wouldn't consider anything less than 90% equities, ever.

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
It seems to me the best place to have our money is in productive assets, and if we don't want to be involved in business that means owning stocks. Keep some cash/intermediate bonds around so we don't have to sell stocks at the very bottom. Buffets advice of 90/10 feels about right to me, and so that's what we do, more or less. I do differ from Buffett in that I diversify internationally. I'm not native born American and don't quite have his super optimistic view of America.

Rising equity glide paths, small cap value, Golden Butterfly, MPT, isn't all this stuff is based on back tests? I don't really buy it.

I agree with this. All the backtesting of esoteric ideas is I think data-mining or to be more clear over-optimisation which could likely lead to suboptimal future returns of a portfolio. I think it's better to aim for the average rather than the edge. The average is more likely to be robust. The edge is more likely to provide inconsistent results.

In stating that a REGP makes some logical sense to me in that if the markets go up over the first 5-10 years then your portfolio should be safe and then you can try and get higher returns. At the same time personally I don't want the highest returns. I'm not aiming to die the richest.

Eric

  • Magnum Stache
  • ******
  • Posts: 4057
  • Location: On my bike
First, I like the idea of the REGP.  However, I wouldn't ever go as extreme as Kitces/Pfau did in their article.  You can see a benefit from starting at say 50/50 or 60/40 stocks/bonds and increasing from there.  In this way, you're sort of splitting the difference between extreme sequence of returns aversion and long term returns.

I'm about 90% sure that my plan starting next year will be to move from 60/40 to 80/20 over 5 years.  I haven't completely analyzed every past scenario on cFIREsim, and this still has problems with the late 60s like most portfolios, but the results for the 2000 retiree were promising.  And of course I'm still heavy enough in equities that the lack of a crash wouldn't really put me much behind at all.

DrF

  • Bristles
  • ***
  • Posts: 464
Big ERN also has some posts of sequence of return risk. Enjoy!

https://earlyretirementnow.com/2017/05/17/the-ultimate-guide-to-safe-withdrawal-rates-part-14-sequence-of-return-risk/
https://earlyretirementnow.com/2017/05/24/the-ultimate-guide-to-safe-withdrawal-rates-part-15-sequence-of-return-risk-part2/

Plus, you need to go to a very high bond allocation to get much/if any diversification benefit from bonds. IE, most of the risk reduction is due to having less stock allocation.
https://earlyretirementnow.com/2016/08/17/bond-diversification-is-a-myth/

Eric

  • Magnum Stache
  • ******
  • Posts: 4057
  • Location: On my bike
Big ERN also has some posts of sequence of return risk. Enjoy!

https://earlyretirementnow.com/2017/05/17/the-ultimate-guide-to-safe-withdrawal-rates-part-14-sequence-of-return-risk/
https://earlyretirementnow.com/2017/05/24/the-ultimate-guide-to-safe-withdrawal-rates-part-15-sequence-of-return-risk-part2/


Wow, that was a big nothing burger.  Two ridiculously long posts to explain what sequence of returns risk is and how it works, with the only suggestion being to use a Variable Percentage Withdrawal model.  Certainly there are other ways, as we're already discussing in this thread and others, including lower WRs, Bond Tents or other Dynamic AA, Variable withdrawals not based on the VPW model, etc.

I guess if you had never heard of the idea of Sequence of Returns Risk, that might have been a decent intro, but not sure how it fits in this thread at all.  I'm sure Mr Green is aware of the concept.

Shane

  • Handlebar Stache
  • *****
  • Posts: 1665
  • Location: Midtown
Big ERN also has some posts of sequence of return risk. Enjoy!

https://earlyretirementnow.com/2017/05/17/the-ultimate-guide-to-safe-withdrawal-rates-part-14-sequence-of-return-risk/
https://earlyretirementnow.com/2017/05/24/the-ultimate-guide-to-safe-withdrawal-rates-part-15-sequence-of-return-risk-part2/


Wow, that was a big nothing burger.  Two ridiculously long posts to explain what sequence of returns risk is and how it works, with the only suggestion being to use a Variable Percentage Withdrawal model.  Certainly there are other ways, as we're already discussing in this thread and others, including lower WRs, Bond Tents or other Dynamic AA, Variable withdrawals not based on the VPW model, etc.

I guess if you had never heard of the idea of Sequence of Returns Risk, that might have been a decent intro, but not sure how it fits in this thread at all.  I'm sure Mr Green is aware of the concept.

lol. Thanks for saving me the trouble of reading those linked posts, Eric. :)

Mr. Green

  • Magnum Stache
  • ******
  • Posts: 4485
  • Age: 40
  • Location: Wilmington, NC
I'm very attentive to sequence of returns risk, even going so far as to create a spreadsheet that shows all 30-year historical periods with total average return in 10-year increments. That data allows me to see all the historical examples where portfolios that failed and compare my returns as we move into our first 10-year period. It will be very easy to know if we need to take corrective action within 7-8 years at the latest.

Of course there are also cases where the returns during the first 10-years were what saved the next underperforming 20 so in that case REGP lowers this even more and by the time you know you're in trouble you're 18-ish years in. Fifty couple years old instead of forty couple.  Maybe not a fun time to be looking at a mandatory part-time job.

Eric

  • Magnum Stache
  • ******
  • Posts: 4057
  • Location: On my bike
I'm very attentive to sequence of returns risk, even going so far as to create a spreadsheet that shows all 30-year historical periods with total average return in 10-year increments. That data allows me to see all the historical examples where portfolios that failed and compare my returns as we move into our first 10-year period. It will be very easy to know if we need to take corrective action within 7-8 years at the latest.

It's what you posted here?  That's pretty cool.  You created that?  I like the visual.  Are the numbers nominal or real?

Of course there are also cases where the returns during the first 10-years were what saved the next underperforming 20 so in that case REGP lowers this even more and by the time you know you're in trouble you're 18-ish years in. Fifty couple years old instead of forty couple.  Maybe not a fun time to be looking at a mandatory part-time job.

I still think it's a pretty good idea, if only because the first 5-10 years are generally more likely to be the years that have the biggest influence.  And since we've been racking up returns at a pretty good clip lately, if I was a betting man (note that I am), I'd bet that the next 10 will be worse than the subsequent 10.  But as I stated above, I don't really like the idea of starting higher than 40-50% bonds, so if I'm wrong, it shouldn't put me too far behind.

If you're 18-20 years in and things are not looking great, you should be close to minimum SS age by that time, right?  That's the hidden savior of most all of those failing portfolios, especially if you have modest spending needs.

Still, it's a tough one to decide.   The fact that bond yields are pretty crappy definitely doesn't help.  Where's that crystal ball when you need it, right? 

Mr. Green

  • Magnum Stache
  • ******
  • Posts: 4485
  • Age: 40
  • Location: Wilmington, NC
I'm very attentive to sequence of returns risk, even going so far as to create a spreadsheet that shows all 30-year historical periods with total average return in 10-year increments. That data allows me to see all the historical examples where portfolios that failed and compare my returns as we move into our first 10-year period. It will be very easy to know if we need to take corrective action within 7-8 years at the latest.

It's what you posted here?  That's pretty cool.  You created that?  I like the visual.  Are the numbers nominal or real?
The numbers are real. Yes, I created it. That's just a snapshot of the spreadsheet I have. The nice thing about having that data is I can do a little forward predicting. This is why I'll know with confidence how we're doing just 7-8 years in, since I can use that data and some assumed returns for the next couple years to model hypothetical "first 10 years" data, and compare that to history. It might not give me complete clarity on whether I should consider going back to work if the percentage falls in line with a past scenario that went the distance but didn't end with favorable numbers but it helps. For instance my chart shows that, barring a 2008 type meltdown next year, someone who retired at the beginning of 2009 has already won the game. The first 10-year returns are so good that no historical scenario has posted those type of numbers and come close to failing.

I just turned 34 so 18 years of retirement makes me 52, still a decade away from the earliest SS payment assuming that doesn't go up at all by the time I get there. That's a little longer than I'd like, to be thinking of a scenario where I'm looking for some work to bridge the gap. Plus, in my own personal case my body hasn't been very nice to me over this last year. Could just be a blip but I'm reminded that not everyone reaches their 60's in stellar health, regardless of effort. Maybe by my mid-50's I'm a little worse for wear, which could compound the work issue.

I'm not hugely concerned because, as mentioned before, we are talking about scenarios on the margins, historically, but those are the types of things I consider when looking at it all. I think I've more or less convinced myself to keep my current asset allocation as a result of this thread.

DrF

  • Bristles
  • ***
  • Posts: 464
Big ERN also has some posts of sequence of return risk. Enjoy!

https://earlyretirementnow.com/2017/05/17/the-ultimate-guide-to-safe-withdrawal-rates-part-14-sequence-of-return-risk/
https://earlyretirementnow.com/2017/05/24/the-ultimate-guide-to-safe-withdrawal-rates-part-15-sequence-of-return-risk-part2/


Wow, that was a big nothing burger.  Two ridiculously long posts to explain what sequence of returns risk is and how it works, with the only suggestion being to use a Variable Percentage Withdrawal model.  Certainly there are other ways, as we're already discussing in this thread and others, including lower WRs, Bond Tents or other Dynamic AA, Variable withdrawals not based on the VPW model, etc.

I guess if you had never heard of the idea of Sequence of Returns Risk, that might have been a decent intro, but not sure how it fits in this thread at all.  I'm sure Mr Green is aware of the concept.

Haha, Big ERN must have heard you, because he just posted a new article this morning.

https://earlyretirementnow.com/2017/09/13/the-ultimate-guide-to-safe-withdrawal-rates-part-19-equity-glidepaths/#more-31662

Summary: The failsafe SWR is 3.25% with 75% equity/25% bond. The best failsafe SWR using glidepath is 3.47% going from 60%-100% equity with 0.4% glide per month.

Mr. Green

  • Magnum Stache
  • ******
  • Posts: 4485
  • Age: 40
  • Location: Wilmington, NC
Haha, Big ERN must have heard you, because he just posted a new article this morning.

https://earlyretirementnow.com/2017/09/13/the-ultimate-guide-to-safe-withdrawal-rates-part-19-equity-glidepaths/#more-31662

Summary: The failsafe SWR is 3.25% with 75% equity/25% bond. The best failsafe SWR using glidepath is 3.47% going from 60%-100% equity with 0.4% glide per month.
That's a very interesting article by ERN. His glide paths don't start nearly as bond heavy as the Kitces article, which makes more sense for someone in his 30's. I am a bit confused because he uses 60-year retirement periods but he also includes some charts that show starting years into the 1980's. A cycle starting beyond 1957 won't have 60 years worth of data. So is he running Monte Carlo sims? Just filling in the incomplete periods with random data from history? If he's running Monte Carlo sims based on historical data why reference start years? There's a lot presented there so I definitely need to reread it again.

DrF

  • Bristles
  • ***
  • Posts: 464
ERN's methodology. Didn't read through it all, so not sure if it answers all your questions.

https://earlyretirementnow.com/2017/02/01/the-ultimate-guide-to-safe-withdrawal-rates-part-8-technical-appendix/

EarlyRetirementNow

  • 5 O'Clock Shadow
  • *
  • Posts: 3
I'm ERN, the author of this study.
Post-2017 I extrapolate returns as:
a) moderate Stock bond returns for the next 10 years for stocks (3.75% real for 10 years due to low CAPE) and Treasury Yields (0.5% real return for 10 years)
b) after that long-term average returns.
Remember: SWR are mostly determined by the first few years of returns (see Part 15 of the SWR study), so it's not a crazy assumption to fill in the last few years of the simulation horizon with long-term average data.
And of course, due to the lack of volatility in the extrapolated data, consider the SWRs calculated here more of an upper bound!

ERN's methodology. Didn't read through it all, so not sure if it answers all your questions.

https://earlyretirementnow.com/2017/02/01/the-ultimate-guide-to-safe-withdrawal-rates-part-8-technical-appendix/
« Last Edit: September 13, 2017, 04:49:11 PM by EarlyRetirementNow »

EarlyRetirementNow

  • 5 O'Clock Shadow
  • *
  • Posts: 3
It's not a nothing burger. I have written extensively on the methods to reduce SoRR in my series: Lower WR (pretty much all parts), dynamic allocation (parts 13, 19), Guyton Klinger (parts 9,10). Also, variable rates based on VPW and others (part 11, 18).
Though keep in mind that variable rates a la VPW don't really help with SoRR. They may soften the impact of SoRR on the final portfolio value but SoRR still hits you through lower withdrawals along the way. So, before you call other people's extensive research a "nothing burger" make sure you have a grasp on the subject yourself.


Big ERN also has some posts of sequence of return risk. Enjoy!

https://earlyretirementnow.com/2017/05/17/the-ultimate-guide-to-safe-withdrawal-rates-part-14-sequence-of-return-risk/
https://earlyretirementnow.com/2017/05/24/the-ultimate-guide-to-safe-withdrawal-rates-part-15-sequence-of-return-risk-part2/


Wow, that was a big nothing burger.  Two ridiculously long posts to explain what sequence of returns risk is and how it works, with the only suggestion being to use a Variable Percentage Withdrawal model.  Certainly there are other ways, as we're already discussing in this thread and others, including lower WRs, Bond Tents or other Dynamic AA, Variable withdrawals not based on the VPW model, etc.

I guess if you had never heard of the idea of Sequence of Returns Risk, that might have been a decent intro, but not sure how it fits in this thread at all.  I'm sure Mr Green is aware of the concept.

Eric

  • Magnum Stache
  • ******
  • Posts: 4057
  • Location: On my bike
It's not a nothing burger. I have written extensively on the methods to reduce SoRR in my series: Lower WR (pretty much all parts), dynamic allocation (parts 13, 19), Guyton Klinger (parts 9,10). Also, variable rates based on VPW and others (part 11, 18).
Though keep in mind that variable rates a la VPW don't really help with SoRR. They may soften the impact of SoRR on the final portfolio value but SoRR still hits you through lower withdrawals along the way. So, before you call other people's extensive research a "nothing burger" make sure you have a grasp on the subject yourself.


Big ERN also has some posts of sequence of return risk. Enjoy!

https://earlyretirementnow.com/2017/05/17/the-ultimate-guide-to-safe-withdrawal-rates-part-14-sequence-of-return-risk/
https://earlyretirementnow.com/2017/05/24/the-ultimate-guide-to-safe-withdrawal-rates-part-15-sequence-of-return-risk-part2/


Wow, that was a big nothing burger.  Two ridiculously long posts to explain what sequence of returns risk is and how it works, with the only suggestion being to use a Variable Percentage Withdrawal model.  Certainly there are other ways, as we're already discussing in this thread and others, including lower WRs, Bond Tents or other Dynamic AA, Variable withdrawals not based on the VPW model, etc.

I guess if you had never heard of the idea of Sequence of Returns Risk, that might have been a decent intro, but not sure how it fits in this thread at all.  I'm sure Mr Green is aware of the concept.

I have a firm grasp on the concepts of the Sequence of Returns Risk, which is why I gained nothing from reading those articles.  The two linked posts were for beginners, and I stated that they would be a decent introduction to the topic, but were unhelpful for this thread.  Most of the posters in this thread have discussed this topic at length, so the idea that we'd need a refresher on the definition of Sequence of Returns Risk is silly.  I was not commenting on all of your posts.  I was commenting on the two linked posts.  And I stand by the statement that they are worthless in this context and it was a waste of my time to read them and don't think that DrF should've posted them here because they didn't address the questions that were being asked.

So yes, they were a nothing burger, as they offered no new insight or ideas to the people in this thread who already understand the concept.  Sorry if that hurts your feelings, but introductory articles are not all that useful to people who understand the advanced concepts already.


---------------------------------------

Edit -- however this most recent post looks a lot more interesting.  Definitely some meat on that burger. 

https://earlyretirementnow.com/2017/09/13/the-ultimate-guide-to-safe-withdrawal-rates-part-19-equity-glidepaths
« Last Edit: September 13, 2017, 06:35:40 PM by Eric »

EarlyRetirementNow

  • 5 O'Clock Shadow
  • *
  • Posts: 3
Cool! Glad you found the glidepath post more interesting.

But back to the sequence of return risk: When you say that you learned nothing from the Sequence of Return Risk posts, that's a surprise. I wrote some things that I have not seen anywhere else before I pointed them out:
a) SoRR is a zero-sum-game between savers and retirees
b) I quantified SoRR in a way I haven't seen anywhere else, see regression results of SWR on 5-year return windows. Including Newey-West heteroskedasticity-adjusted t-stats.

c) SoRR is one of the reasons why savers should not pay down their mortgage earlier than necessary.
d) I showed (in the comment section) that VPW is exposed to SoRR (which was the reason I had my doubts you're beyond beginner level on this topic, please forgive me that).
and many other fun facts...   
Where exactly have you seen those results before? If you already posted them on your own early retirement blog, where would I find that?


It's not a nothing burger. I have written extensively on the methods to reduce SoRR in my series: Lower WR (pretty much all parts), dynamic allocation (parts 13, 19), Guyton Klinger (parts 9,10). Also, variable rates based on VPW and others (part 11, 18).
Though keep in mind that variable rates a la VPW don't really help with SoRR. They may soften the impact of SoRR on the final portfolio value but SoRR still hits you through lower withdrawals along the way. So, before you call other people's extensive research a "nothing burger" make sure you have a grasp on the subject yourself.


Big ERN also has some posts of sequence of return risk. Enjoy!

https://earlyretirementnow.com/2017/05/17/the-ultimate-guide-to-safe-withdrawal-rates-part-14-sequence-of-return-risk/
https://earlyretirementnow.com/2017/05/24/the-ultimate-guide-to-safe-withdrawal-rates-part-15-sequence-of-return-risk-part2/


Wow, that was a big nothing burger.  Two ridiculously long posts to explain what sequence of returns risk is and how it works, with the only suggestion being to use a Variable Percentage Withdrawal model.  Certainly there are other ways, as we're already discussing in this thread and others, including lower WRs, Bond Tents or other Dynamic AA, Variable withdrawals not based on the VPW model, etc.

I guess if you had never heard of the idea of Sequence of Returns Risk, that might have been a decent intro, but not sure how it fits in this thread at all.  I'm sure Mr Green is aware of the concept.

I have a firm grasp on the concepts of the Sequence of Returns Risk, which is why I gained nothing from reading those articles.  The two linked posts were for beginners, and I stated that they would be a decent introduction to the topic, but were unhelpful for this thread.  Most of the posters in this thread have discussed this topic at length, so the idea that we'd need a refresher on the definition of Sequence of Returns Risk is silly.  I was not commenting on all of your posts.  I was commenting on the two linked posts.  And I stand by the statement that they are worthless in this context and it was a waste of my time to read them and don't think that DrF should've posted them here because they didn't address the questions that were being asked.

So yes, they were a nothing burger, as they offered no new insight or ideas to the people in this thread who already understand the concept.  Sorry if that hurts your feelings, but introductory articles are not all that useful to people who understand the advanced concepts already.


---------------------------------------

Edit -- however this most recent post looks a lot more interesting.  Definitely some meat on that burger. 

https://earlyretirementnow.com/2017/09/13/the-ultimate-guide-to-safe-withdrawal-rates-part-19-equity-glidepaths

DrF

  • Bristles
  • ***
  • Posts: 464
I'd like to say that I think ERN is putting out the most relevant next level analyses on SWR to date. His contributions shouldn't be taken lightly.

Eric, not everyone is a 9th level master at this stuff, so what you may deem to be primer material allows another reader to grasp the concepts.

Also, producing original content is mindblowingly hard. Please don't flippantly disregard others hard work.

Eric

  • Magnum Stache
  • ******
  • Posts: 4057
  • Location: On my bike
Cool! Glad you found the glidepath post more interesting.

But back to the sequence of return risk:

Nah, let's not.  It sort of seems like you're posting this to get me into some drawn out discussion about how your SoRR knowledge is superior to mine (especially point b), but I'm not interested.  In fact, I concede.  You may very well be more well versed than I am.  I'm less interested in a dick measuring contest and more interested in techniques to implement to hedge my risk of running out of money.

Where exactly have you seen those results before? If you already posted them on your own early retirement blog, where would I find that?

Well aren't you a cheeky little monkey?  I like that!  :)  But of course, you don't need a blog to contribute to the FIRE community.  I realize you probably haven't read much around these forums, but many blog-less people here have been doing great analytical analysis for years.  I'm probably not one of them, but at least I gained a bunch of knowledge from it.


So back to the topic at hand, your Glidepath post is much more relevant to this thread.  While I haven't run nearly as many scenarios as you did, your findings seem to echo what I found as well.  Namely that starting from a medium weight bond position (20-40%) and moving to a light to zero weight position (0-20%) over ~5 years at the start of a retirement period is a simple yet effective step towards mitigating downside risk.

Specifically this statement, which seems to be somewhat contradictory to the suggestions of Kitces in his posts on the subject:

"The very long transitions over 60 percentage points (20 to 80% and 40 to 100%) tend to be pretty consistently inferior to the other glidepaths. The 20 to 80% glidepaths are even inferior to the static 80% and 100% allocations! Apparently, the initial stock weight was too low and/or the transition took way too long (even with the accelerated slopes of 0.4% and 0.5%!)"

It actually passes the sniff test for me, since I like the idea of the Rising Equity Glidepath, but I'd never be convinced to hold only 30% stocks at any point, even if I was planning to transition off that point.

So thanks for taking that idea and fleshing it out further.  I like having this to compare my own findings against.

maizefolk

  • Walrus Stache
  • *******
  • Posts: 7396
It makes a lot of intuitive sense to me that if you start with a portfolio so bond heavy that your expected real return is <4%, that you'd be adding risk of failure rather than reducing it.

There are probably better sources but triumph of the optimists puts the long term inflation adjusted CAGR of US bonds at 1.6% and equities at 6.7%. That means at anything higher than 50% bonds you're expecting to spend down more of your stash in a year than you make up in returns. Maybe that's okay if you're just trying to last 30 years with 25 years expenses saved up and only need a small amount of capital appreciation to make it work, but for extremely long retirements a plan that expects to spend more than you earn in the early years seems foolhardy.

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
I would not underestimate the impact of a changing game board on assumptions driven by historic data. 

That said, it is the basics that work.  Live within your means.  Save.  Get your capital working for you, as productively as possible.  Pay attention and build resiliency into your plan.

I like the idea of a REGP and I think the analysis ERN does is pretty comprehensive but I don't believe that you can systematise most of these decisions related to WR's and withdrawal plans and asset allocations based on the historical data that we have.

I do think that you need to focus on the basics rather than the edges which these types of analysis focus on. I think getting to a reasonable WR, having a plan for withdrawing money, having a plan for your asset allocation and most importantly having in your words some resiliency or backup options in your plan is going to work out in reality a lot better than thinking you can come up with the perfect plan.

We also need to realise that trying to get from a 95% probability (or even lower) of success may lead to a significant number of extra years worked.

Eric

  • Magnum Stache
  • ******
  • Posts: 4057
  • Location: On my bike
I do think that you need to focus on the basics rather than the edges which these types of analysis focus on.

Porque no los dos?  Although I'm not sure I agree with the idea that your asset allocation is an edge of your plan.  That's a pretty central tenet.


I think getting to a reasonable WR, having a plan for withdrawing money, having a plan for your asset allocation and most importantly having in your words some resiliency or backup options in your plan is going to work out in reality a lot better than thinking you can come up with the perfect plan.

The REGP is simply a plan for your asset allocation.  It's not complicated at all and would take all of 10 minutes a year to implement with simple re-balancing. 


We also need to realise that trying to get from a 95% probability (or even lower) of success may lead to a significant number of extra years worked.

Except in this case, the data shows that you can gain a higher probability of success without working more or spending less.

steveo

  • Handlebar Stache
  • *****
  • Posts: 1928
Eric - my point about the edges is where you are data mining to try and come up with some better plan that inevitably is on the edge rather than a broad principle. So trying to come up with the perfect asset allocation in my opinion is unlikely to actually work out in the future. I feel the same way in relation to the REGP.

The data might show that you can increase your chances of success but that is historical data and it's doubtful that the future performance will be the same as the historical data-set.

I still like the idea of a REGP but I'm a little skeptical of implementing it in a systemised fashion. I think I'll just stick to an asset allocation that I'm comfortable with and adjust if and as required.

 

Wow, a phone plan for fifteen bucks!