Are Safe Withdrawal Rates Really Safe?

[beee]

Discover The Little-Known, But Highly Dangerous Risks Hiding Behind The 4% Rule, And The Simple Solutions to Correct the Problem

http://financialmentor.com/retirement-planning/how-much-money-do-i-need-to-retire/safe-withdrawal-rate/13192

What do you think about this article?

[odput]

Holy crap that is a long article saying basically nothing

TL; DR - Same ridiculous bullshit claiming the 4% rule (of thumb) is a total failure since you can't know everything and even if you did I can still come up with a better doomsday scenario...better keep on working

[brooklynguy]

Holy crap that is a long article saying basically nothing

I read the whole thing and thought it was actually a pretty good plain-English summary of the history and the current state of affairs of SWR research, and explains why the 4% rule is a "rule of thumb" and not an "inviolate rule of the universe."

It seems as if the article was written for an audience of readers who are under the false impression that the 4% SWR rule says you can robotically make 4% inflation-adjusted annual withdrawals from your portfolio for the duration of your retirement without regard to the actual real-world performance of your investments or any of the other relevant circumstances of your retirement with an iron-clad guarantee against portfolio-depletion provided by the Federal Safe Withdrawal Rate Insurance Corporation and backed by the full faith and credit of the United States Government.  The article definitely constitutes worthwhile reading for that audience, but I doubt that includes many people in this forum.

My biggest problem with the article is its pessimistic undercurrent.  The author views the world through shit-brown-colored glasses, expounding on all the unrealistic assumptions behind the 4% SWR rule of thumb that have the effect of overstating its success in the real world but giving short shrift to those that have the effect of understating its success in the real world.  The short answer is that anyone with a even modicum of flexibility and safety margins built into their retirement plan should do just fine.

[a1smith]

The quick answer is that the 2nd generation work he mentions (Bengen and Trinity Study) is referenced a lot here.  The assumptions he mentions to the 2nd generation studies are correct.  The assumptions in the study are clearly spelled out and do need to be taken into consideration so there is no need to call them dubious.  The 3rd generation work seeks to build on their work.  A lot of the discussion here concerning safety factors addresses the 3rd generation concerns.  Also, there are many simulation tools, cFIREsim being one of them, that let you figure out your own SWR based on your personal assumptions which addresses the concern they have about bad assumptions.  For example, you can add pension and social security income in CFIREsim and that is not included in the Bengen or Trinity studies; they are only withdrawing money from one account.

Now, a little more discussion if you're interested.

The article questions how can everyone's SWR be 4% and that is a little misleading.  The work that came up with the 4% rate said that is the maximum safe rate when backtesting the data.  In other words, over all of the 30 year periods tested, 4% was the lowest result; all other time periods had a higher SWR.  That is well known.

The only way you are going to know your SWR exactly is to know what will happen in the future.  So, you just have to come up with your own best estimate and run with it.  On the fly course corrections are discussed a lot here.

To address the question about retirement being longer than 30 years with 4% SWR I'm reposting some work I did looking at that in cFIREsim, see below.  Basically, the conclusion is that you need to do the calculations for your own retirement period and conditions (pension, SS, asset allocation, etc) and plan on making adjustments on the fly.

So, I think 4% SWR (save 25X annual expenses) is a great rule of thumb to set your initial target and then while you are on your way to saving that much you fine tune your personal SWR estimate using what you learn.

========

In MMM's post The 4% Rule: The Easy Answer to “How Much Do I Need for Retirement?” he shows the 4.04% SAFEMAX value in Figure 2.1.  Actually, that figure is derived from William Bengen's work in 1994 (the author, Wade Pfau, extended the results to 1981 using data up to 2010), not the Trinity study.  The Trinity study occurred four years after Bengen's seminal work.  The only difference between their work is the bond indices they chose.  With Bengen, he showed 100% success at 4.15% WR; the Trinity study showed a 95% success rate at 4% WR (not 100%!!!).

If you go to the webpage MMM got that figure from, it redirects to retirementresearcher.com/the-trinity-study-and-portfolio-success-rates/.  Looking at Table 2.1, you can see that at 4% withdrawal rates you have 100% success at 30 years, but only 96% at 35 years and 85% at 40 years.

I tried a really simple experiment in cFIREsim and obtained comparable results.  I started with 2015-2045 for a 30 year retirement, with $100 portfolio, $4 spending (4%) with inflation adjust , 50/50 stocks/bonds, no fees, and everything else set to zero. That matches the study parameters. I then ran simulations for 30, 35, 40, ...., 70, and 75 year retirements. I stopped at 75 years for two reasons: most people won't be retired that long and also the data starts to be statistically insignificant because at 75 years only 70 cycles are ran, 80 years gives 65 cycles, etc.

Well, the results were so bad for 50/50 stock/bond allocation I also ran 70/30.  The results are shown in the table below.  As you can see the success rate keeps dropping and it is not clear to me that there is an asymptotic limit.  I'm not sure where the assumption that if the money lasts 30 years it will last forever came from but I consider it suspect.  If someone can post some hard calculations or a research paper that shows this I would appreciate it.

Years Retired	50/50 Success	70/30 Success
30	90.4%	93.9%
35	82.7%	90.9%
40	67.6%	84.8%
45	65.0%	80.0%
50	63.2%	81.1%
55	60.0%	82.2%
60	60.0%	82.4%
65	55.0%	80.0%
70	53.3%	78.7%
75	50.0%	77.1%

[seattlecyclone]

If you assume that the next 30 years won't be worse than every other 30-year period in recorded history, you can robotically pull out 4% every year, adjusted for inflation, and be reasonably certain that your money will last 30 years. It's certainly possible that the next 30 years actually will be the worst 30 years ever, and that's where flexibility comes into play. Be willing to spend a little less in a down year and you should be fine with a 4% withdrawal rate in the good years in any reasonably likely scenario.

[forummm]

The 4% SWR is a rule of thumb and not a law of physics. It's pretty conservative historically, but could be insufficient in the future depending on how economies do and where you invested the money. One way you can be more confident in having enough money to eat is by being flexible in the future. If the markets do poorly for a decade, you might need to cut back or get a job. Who knows.

The 4% SWR is also meant to be for 30-year withdrawal periods. If you plan to need those funds for more than that time, you might want to save more, etc. Typically you need to have a lot in stocks for a very low portfolio failure rate over longer than 30 years with a 4% withdrawal rate.

[Retire-Canada]

Well, the results were so bad for 50/50 stock/bond allocation I also ran 70/30.  The results are shown in the table below.  As you can see the success rate keeps dropping and it is not clear to me that there is an asymptotic limit.  I'm not sure where the assumption that if the money lasts 30 years it will last forever came from but I consider it suspect.  If someone can post some hard calculations or a research paper that shows this I would appreciate it.

One problem with running cFIREsim for longer periods is the number of "samples" drops due to the available historical data set and the same bad years at the start form a larger part of each of simulations you look at. I think that will skew the stats to the point that they are not really comparable.

30yr period = 115 cycles/samples
50yr period = 95 cycles/samples
75yr period = 70 cycles/samples
100yr period = 45 cycles/samples

-- Vik

[Eric]

I didn't read this one closely, but I did skim it and I've read a similar critique (both in substance and length) from this author in the past.  I agree with Brooklynguy that the he lays out all of the most pessimistic scenarios, without giving any weight to optimistic ones.  Essentially, it's the worst worst case scenario planning.  But it actually is a good primer on all the issues with a blind following of 4% withdrawals.

The main issue I have is that the article appears to be only discussing 50/50 asset allocation (at least for their future projections, although it's only mentioned once when talking about Pfau).  Most of us prefer much higher stock ratio, and therefore should expect better returns and fewer failures.  Also, there are complaints that the Trinity Study only looked at US markets, but most of us already adjust for that by including foreign stocks of 25-40% in our AAs.

In general, I feel like a lot of the people that write with this tone are actual financial planners who are giving advice to their clients.  Because of this, they feel responsible to make sure that the advice they give, that can have real world consequences for the people who take it, is 150% guaranteed to work.  It's a form of a Cover Your Ass policy.  For those of us doing our own planning, it's less necessary to make our plans completely ironclad, as we don't have to worry about accepting blame if things don't work out perfectly.  If I work a part time job for a few years, I won't be mad that a financial planner told me I'd never have to work again.  I'll be happy that I was able to escape my full time job many years earlier than otherwise would've been possible and that a little work will just increase my safety margin.

[skyrefuge]

I read the whole thing and thought it was actually a pretty good plain-English summary of the history and the current state of affairs of SWR research, and explains why the 4% rule is a "rule of thumb" and not an "inviolate rule of the universe."

Agreed. If you can ignore the pessimistic tone ("because you're more likely to pay for my book if you're scared!"), it's actually an excellent, annotated reference covering the history of SWR research. That's why it's particularly disappointing that, again, despite all his links and references to research done by people with the word "Dr." in front of their names, he makes no reference to cFIREsim (or even FIRECalc). Particularly since they address most of the items on his list of "dubious assumptions" in the 2nd Generation research.

[brooklynguy]

As you can see the success rate keeps dropping and it is not clear to me that there is an asymptotic limit.  I'm not sure where the assumption that if the money lasts 30 years it will last forever came from but I consider it suspect.  If someone can post some hard calculations or a research paper that shows this I would appreciate it.

I must have missed this post the first time around, but that particular statement by MMM is a bit of little-discussed hand-waving on his part.  I think the real point is not that a 4% WR will be sustainable over extended, supra-30-year periods, but that some minor downward adjustment to the 4% WR and/or some increased flexibility in the retirement plan can turn a "30-year SWR" into an "infinite-year SWR."

[a1smith]

Well, the results were so bad for 50/50 stock/bond allocation I also ran 70/30.  The results are shown in the table below.  As you can see the success rate keeps dropping and it is not clear to me that there is an asymptotic limit.  I'm not sure where the assumption that if the money lasts 30 years it will last forever came from but I consider it suspect.  If someone can post some hard calculations or a research paper that shows this I would appreciate it.

One problem with running cFIREsim for longer periods is the number of "samples" drops due to the available historical data set and the same bad years at the start form a larger part of each of simulations you look at. I think that will skew the stats to the point that they are not really comparable.

30yr period = 115 cycles/samples
50yr period = 95 cycles/samples
75yr period = 70 cycles/samples
100yr period = 45 cycles/samples

-- Vik

I mentioned that.

I stopped at 75 years for two reasons: most people won't be retired that long and also the data starts to be statistically insignificant because at 75 years only 70 cycles are ran, 80 years gives 65 cycles, etc.

I didn't calculate the exact amount the confidence level is being impacted; maybe stopping sooner would be necessary.

[a1smith]

As you can see the success rate keeps dropping and it is not clear to me that there is an asymptotic limit.  I'm not sure where the assumption that if the money lasts 30 years it will last forever came from but I consider it suspect.  If someone can post some hard calculations or a research paper that shows this I would appreciate it.

I must have missed this post the first time around, but that particular statement by MMM is a bit of little-discussed hand-waving on his part.  I think the real point is not that a 4% WR will be sustainable over extended, supra-30-year periods, but that some minor downward adjustment to the 4% WR and/or some increased flexibility in the retirement plan can turn a "30-year SWR" into an "infinite-year SWR."

I agree.  I mentioned calculating a personal SWR and on the fly adjustments.

Just for fun, I took the 50 year retirement scenario and increased the portfolio amount until I got a comparable % success rate as the 30 year retirement.  I kept all other parameters as mentioned above.  So, this is for someone that is FIRE about 20 years early (mid 40's).

At $116 portfolio with $4 spending (3.45% SWR) the success rate climbs to 90.53%, slightly exceeding the 90.4% success rate at 30 years. So, for a 50 year retirement about a 14% decrease in SWR is required.

As noted, this is mimicking the parameters of the Bengen and Trinity studies which only pulled from one account, has no pension or SS income streams, no adjustments with time other than inflation adjust, a 50/50 stock/bond allocation, etc.

[bdbrooks]

ValueGap.net does a decent job of looking at SWRs. I think there could be better methods for SWRs than they use. I would rather see it based off of cape ratio and real interest rates (interest - inflation). However, they definitely help with showing that retiring in 2007 and 2009 have substantially different SWRs

https://m.youtube.com/watch?v=gbtG0NWfVDs

They are geared towards financial advisors. I have an account and can run a report if anyone is interested. Basically if you are holding a significant amount of bonds, then you will want to keep

[sirdoug007]

Does anyone else see the irony in citing Pfau's research from 2010 calculating a 1.8% SWR based on valuations?

Hmm, what has the market done since 2010?  It has doubled!  Hardly the sequence of returns risk Pfau was predicting.  Goes to show you that even the experts are awful at prediction!

If you download Pfau's paper and look at the "curve fits" for valuation data the scatter is incredible.   It is so bad I don't see the point of factoring it in at all.

Be flexible and you will be way ahead.  At least that is the conclusion of the long-winded article.

The total return of SPY (S&P 500 index ETF) looks like this:

[brooklynguy]

That's why it's particularly disappointing that, again, despite all his links and references to research done by people with the word "Dr." in front of their names, he makes no reference to cFIREsim (or even FIRECalc). Particularly since they address most of the items on his list of "dubious assumptions" in the 2nd Generation research.

In this case, that may be due not to academic snobbery but to the same reason you described for the author's pessimistic slant -- his target audience is more likely to buy his book if they believe the layperson-retiree needs to rely on people like the author with impressive letters attached to their names to interpret and demystify the esoteric work of other people with even more impressive letters attached to their names than if they realize that a free, publicly-available DIY retirement planning tool like cFIREsim exists (and happens to be the superior resource to boot!).

[Middlesbrough]

Dang nabit brooklyn guy! Quit trying to teach people to fish for themselves.

[Doulos]

All percent based withdraw rate is technically infinite.
Your withdraw amounts just change.

This 4% commonly spoken of is not really a 4%, it is a fixed income adjusted for inflation, that is based off of your original stash.
If you really stick with 4% (or any other %), your income will always last for infinity.

FIRECalc cFIREsim will do that for you too.  Just change your spending to % portfolio instead.

What does this show?  It shows that being flexible really does work.

[skyrefuge]

This 4% commonly spoken of is not really a 4%, it is a fixed income adjusted for inflation, that is based off of your original stash.

Yeah. Also, this "football" thing commonly spoken of is not a ball made out of feet.

Everyone who writes about "The 4% Rule" is well-aware that that's simply a title for the method you describe. I guess it's good to mention it occasionally for complete newbs, but anyone who actually mistakenly implements "The 4% Rule" as a "4% of current assets" method (because they heard someone say that phrase once) is a dumbass.

[brooklynguy]

Does anyone else see the irony in citing Pfau's research from 2010 calculating a 1.8% SWR based on valuations?

Hmm, what has the market done since 2010?  It has doubled!  Hardly the sequence of returns risk Pfau was predicting.  Goes to show you that even the experts are awful at prediction!

The author responds to this criticism in the comments to the article.  He says:

In addition . . . I want to point out that just because the market had a nice run from 2010 through 2014 doesn't change anything.  This research is not short-term timing.  It is valid for 7-15 time frames.  Give the market time to fulfill it's expectancy.

In other words, the hapless 2010 retiree shouldn't be lulled into complacency by the astronomical growth of his portfolio over the past five years, because his retirement is still doomed to fail (he just doesn't know it yet).  I think that's pretty consistent with Pfau's own assertions.  But that's a pretty drastically pessimistic outlook -- future market returns aren't just going to be worse-than-average; they're going to be so stupendously bad that even a retiree who as-of-today is using an approximately 3.33% WR with only 25 years left on their retirement is going to completely run out of money?!

[sirdoug007]

Does anyone else see the irony in citing Pfau's research from 2010 calculating a 1.8% SWR based on valuations?

Hmm, what has the market done since 2010?  It has doubled!  Hardly the sequence of returns risk Pfau was predicting.  Goes to show you that even the experts are awful at prediction!

The author responds to this criticism in the comments to the article.  He says:

In addition . . . I want to point out that just because the market had a nice run from 2010 through 2014 doesn't change anything.  This research is not short-term timing.  It is valid for 7-15 time frames.  Give the market time to fulfill it's expectancy.

In other words, the hapless 2010 retiree shouldn't be lulled into complacency by the astronomical growth of his portfolio over the past five years, because his retirement is still doomed to fail (he just doesn't know it yet).  I think that's pretty consistent with Pfau's own assertions.  But that's a pretty drastically pessimistic outlook -- future market returns aren't just going to be worse-than-average; they're going to be so stupendously bad that even a retiree who as-of-today is using an approximately 3.33% WR with only 25 years left on their retirement is going to completely run out of money?!

That is ridiculous.  These guys need to own that they can't actually make accurate predictions of SWR based on valuations.

Look at the scatter in the source data from Pfau's research (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1699526).  You're telling me you can make accurate predictions out of that random mess?

[arebelspy]

Todd sells coaching. He makes money by being pessimistic and convincing you that you should listen to him, because he has the secret.

It gets old quickly. Or at least it did for me.  Maybe it's the right pitch for others.

[phillyvalue]

It's impossible to really pin down what a safe withdrawal rate is. Nobody knows what future stocks returns will be over the long-run, and just as importantly, the safety of a withdrawal rate depends on the person, because some people may be able to cut their spending slightly or even significantly and still be OK. Others may be relying on every penny. Obviously if you can adjust to what is going on in the market in your living standards, then there's more room for a 4% rate to work out even if returns end up being lower than expected. So, for the person who is building in vacations and other luxuries into their expected living expenses, there is flexibility and a greater probability of success than the person assuming barebones living.

But I think overall the prudent thing to do today is to note that valuations are obviously much higher than they've been in the past and the expected real, per-capita growth in the economy is lower than it has been in the past. This means that past simulations may be too generous versus what the market is going to deliver for you going forward. There's no real way to quantify all of this, but it's just something to keep in mind. Personally I would be comfortable retiring at 3% but not 4% in today's environment.

[a1smith]

Does anyone else see the irony in citing Pfau's research from 2010 calculating a 1.8% SWR based on valuations?

Hmm, what has the market done since 2010?  It has doubled!  Hardly the sequence of returns risk Pfau was predicting.  Goes to show you that even the experts are awful at prediction!

The author responds to this criticism in the comments to the article.  He says:

In addition . . . I want to point out that just because the market had a nice run from 2010 through 2014 doesn't change anything.  This research is not short-term timing.  It is valid for 7-15 time frames.  Give the market time to fulfill it's expectancy.

In other words, the hapless 2010 retiree shouldn't be lulled into complacency by the astronomical growth of his portfolio over the past five years, because his retirement is still doomed to fail (he just doesn't know it yet).  I think that's pretty consistent with Pfau's own assertions.  But that's a pretty drastically pessimistic outlook -- future market returns aren't just going to be worse-than-average; they're going to be so stupendously bad that even a retiree who as-of-today is using an approximately 3.33% WR with only 25 years left on their retirement is going to completely run out of money?!

That is ridiculous.  These guys need to own that they can't actually make accurate predictions of SWR based on valuations.

Look at the scatter in the source data from Pfau's research (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1699526).  You're telling me you can make accurate predictions out of that random mess?

I downloaded Pfau's paper and read it.  Figure 1 in the paper doesn't match the Figure 1 in your PNG file.  In the paper, Figure is a plot of SAFEMAX values over time in the US.  Can you post the link to the paper you got your Figure 1 from?

EDIT - Don't worry about it; I found it.

Predicting Sustainable Retirement Withdrawal Rates Using Valuation and Yield Measures

[happy]

If you take a step back and think about it, it should be obvious that its fundamentally impossible to reduce all the possible future world events with regard to finances, into a simple mathematical number thats going to guarantee success.  All the complex debate about methodological issues, 4% or 3.5% or 3.3%  or whatever, are really just little manhole diversions people are digging for themselves.

Bernstein ( I think, correct me if I'm wrong) said you can't get closer than 80% probability  because of uncertainty of world events.

What I take away from all this is that you just need in the ball park of 4% and then need to keep your wits about you and make adjustments as you go.

Personally I find Todds work motivating and the sell aspect doesn't bother me, but I know it does others.

[forummm]

That is ridiculous.  These guys need to own that they can't actually make accurate predictions of SWR based on valuations.

Look at the scatter in the source data from Pfau's research (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1699526).  You're telling me you can make accurate predictions out of that random mess?

The trend for returns in the 10 years after a particular CAPE value is pretty compelling. Sure, there's a bunch of variability there, but the trend is pretty clear.

[a1smith]

Does anyone else see the irony in citing Pfau's research from 2010 calculating a 1.8% SWR based on valuations?

Hmm, what has the market done since 2010?  It has doubled!  Hardly the sequence of returns risk Pfau was predicting.  Goes to show you that even the experts are awful at prediction!

The author responds to this criticism in the comments to the article.  He says:

In addition . . . I want to point out that just because the market had a nice run from 2010 through 2014 doesn't change anything.  This research is not short-term timing.  It is valid for 7-15 time frames.  Give the market time to fulfill it's expectancy.

In other words, the hapless 2010 retiree shouldn't be lulled into complacency by the astronomical growth of his portfolio over the past five years, because his retirement is still doomed to fail (he just doesn't know it yet).  I think that's pretty consistent with Pfau's own assertions.  But that's a pretty drastically pessimistic outlook -- future market returns aren't just going to be worse-than-average; they're going to be so stupendously bad that even a retiree who as-of-today is using an approximately 3.33% WR with only 25 years left on their retirement is going to completely run out of money?!

That is ridiculous.  These guys need to own that they can't actually make accurate predictions of SWR based on valuations.

Look at the scatter in the source data from Pfau's research (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1699526).  You're telling me you can make accurate predictions out of that random mess?

In the first paragraph of the introduction Pfau is using the plots' coefficients of determination (R^2) to "explain 30.3 and 20.3 percent of the variation in the subsequent 10-year real stock returns." He makes a similar statement for the third plot with respect to bonds.

I haven't seen R^2 used in that way before.  Looking at Wikipedia - Coefficient of determination it seems like that is one way of using it:

It provides a measure of how well observed outcomes are replicated by the model, as the proportion of total variation of outcomes explained by the model (pp. 187, 287)

However, they also note:

Notes on interpreting R^2

R^2 does not indicate whether:

• the independent variables are a cause of the changes in the dependent variable;
• omitted-variable bias exists;
• the correct regression was used;
• the most appropriate set of independent variables has been chosen;
• there is collinearity present in the data on the explanatory variables;
• the model might be improved by using transformed versions of the existing set of independent variables;
• there are enough data points to make a solid conclusion.

For (1), it is not even clear that the variables chosen are independent so you can't say that he can explain 30.3+20.3=50.6% of the variation of subsequent returns.
For (1), and even more important, it does not indicate causality.
For (2), since the problem is a high-order (infinite-order?) model there is certainly omitted-variable bias.
For (3), I wonder how did they pick the log and linear regressors?  Is that just what gave them the highest R^2?  In engineering, the regressor is usually chosen to reflect the physics of the problem.
For (4), I imagine they just chose dividend rate and PE ratio because that was readily available information.
For (7), since we really don't know the DOF of the problem it is hard to say if we have enough data points.  Since it is high-order my guess is no.

Maybe we should start a distributed computing project using Hadoop and MMM member PC's to crunch "big-data" and come up with a better model!! :-D

You can bet big bucks that all the big money houses are already working on this and many other big data projects.  Just one example I can think of that relates to another thread is that they can use big data to find pricing inefficiencies and buy "value" stocks. You won't find any of that work published.

[sirdoug007]

Thanks a1, I guess I copied the wrong link.  You found the paper in question.

Pfau basically finds that he can explain, at best, 75% (R^2 = .75) of the variation of the maximum withdrawal rate using a combination of 10 year earnings, 10 year dividends, and bond interest rate.  The model works best for near balanced portolios (50/50 or 60/40).  For 100% stocks it drops to about 60% and 100% bonds it drops to 55%.

His model is essentially a giant curve fit of SWR based on 10 year earnings, 10 year dividends, and bond interest rate from 1883 to 1980.

Great!  We've explained 75% of the variation in SWR right.  Well here is where I am less than impressed.

Pfau backtested his model with a subset of the data (rather than the whole dataset curve fit).  Remember we are talking about single percentage points throwing your FIRE number off by hundreds of thousands of dollars!  Well when the model runs on only a part of the data, it can underestimate SWR by as much as 2%  (see Figure 7 from the paper cut out below).

My conclusion is that valuation does have some bearing on SWRs, but the low level of accuracy makes it essentially useless to early retirees.  In the end we are talking about predicting the future of the stock and bond markets and no amount of data mining will yield a crystal ball. 

Pfau's model spits out a 1.82% SWR for 2010.  Meanwhile the stock and bond markets have done great in the first 5 years of this retirement.  It would truly take a cataclysm to derail the 2010 retiree at this point.

I think the only useful modifier of SWR is use of realtime data.  That is flexible spending plans.  cFIREsim clearly shows that being flexible and responding to realtime changes in the markets improves your chances of a successful retirement.  Trying to go beyond the present and predict the future behavior of millions of investors is an exercise in futility.

[RapmasterD]

Todd sells coaching. He makes money by being pessimistic and convincing you that you should listen to him, because he has the secret.

It gets old quickly. Or at least it did for me.  Maybe it's the right pitch for others.

+1 - And you could just stop at the first sentence. The subliminal message is, "You can't do this alone. You need my guidance."

[bdbrooks]

I think you guys should take a look at this. This is about the only place that I have found that does a good job with A) giving a SWR that doesn't result in running out of money and that B) doesn't have you withdrawing a tiny fraction of what you could have withdrawn. Check out the picture attached. The purple line is the Actual Safe Withdrawal Rate and the blue line is their estimated safe withdrawal rate. Note that there were timeframes when they correctly predicted that 10% would actually be a safe withdrawal rate (early 30's, early 40's and mid 70's). Also note that the chart stops in 1983 (because the last year that they can calculate a SWR was 30 years ago and this was produced in 2013).

https://www.youtube.com/watch?v=gbtG0NWfVDs&app=desktop

I have a subscription with them and for example withdrawing for 40 years with .5% account expenses and 100% in Equities would have an estimated 4.8% SWR...or 20% more income than the 4% rule would indicate...or 60% more than someone only taking 3%. This could cut several years off of your working life

I think most mustachians would rather aim for the most accurate (but still give reasonable safety) SWR. Then they can use their mustachian attitude and adjust things if they need to.

I know it is a decently long video, but if you significantly care about SWRs then it is worth watching.

[steveo]

^^^
Thanks for that video. I view it as getting to a certain amount and then adjusting your withdrawal amounts based on the situation at that point in time. I personally though don't agree with the idea that if you don't spend enough money its a failure. I'd rather have extra money left over than run out of money. I don't basically relate spending to happiness. In some ways for instance I'd be happy to life off a certain lower limit and forget about doing anything fancy and view that as fun.

[bdbrooks]

"Upside failure" means the person ends 30 years with over 5 times more than when they began withdrawals. Not that we are trying to increase spending just because, but it could allow you to claim financial independence sooner or give to causes that you care about without damaging your financial independence.

[LiseE]

Am I missing something here?  Isn't the balance of your asset account still earning interest which will offset the 4% withdrawal rate?

Initial Balance:  $100,000
4% SWR:      $4,000.00
Remaining Balance: $96,000
Interest earned at 4% on 96,000 = $3840

So in essence you are always replenishing what you are withdrawing with interest earned on the balance?

[forummm]

Am I missing something here?  Isn't the balance of your asset account still earning interest which will offset the 4% withdrawal rate?

Initial Balance:  $100,000
4% SWR:      $4,000.00
Remaining Balance: $96,000
Interest earned at 4% on 96,000 = $3840

So in essence you are always replenishing what you are withdrawing with interest earned on the balance?

It *hopefully* offsets it. The problem is if you retire, the market goes down (e.g. 37%), and then you have to withdraw your 4% (and having to sell a lot more shares than before it went down). This kind of a start can make it easy to run out of money faster.

[Indexer]

Am I missing something here?  Isn't the balance of your asset account still earning interest which will offset the 4% withdrawal rate?

Initial Balance:  $100,000
4% SWR:      $4,000.00
Remaining Balance: $96,000
Interest earned at 4% on 96,000 = $3840

So in essence you are always replenishing what you are withdrawing with interest earned on the balance?

You don't earn an even 4% every year. 

100k, taking out 4k a year.  I'm not including inflation just to keep thing simple, plus its only 5 years... not 30.
Example 1.
Yr1 20% growth 116k, yr2 5% 117.8k, yr3 6% 120.86k, yr4 4% 121.7k, yr 5 -50% 56.8k.

Example 2... the inverse.
Yr 1 -50% 46k, yr2 4% 43.8k, yr3 6% 42.5k, yr4 5% 40.6k, yr5 20% 44.7k.

Same returns, opposite direction equals vastly different returns.  And imagine someone in example 2 who retired 1 year later.
Yr1 4%  100k, yr2 6% 102k,  yr3 5%  103k, yr4 20%  119.7k, yr5 4%  120.5k.

[Retire-Canada]

Am I missing something here?  Isn't the balance of your asset account still earning interest which will offset the 4% withdrawal rate?

Initial Balance:  $100,000
4% SWR:      $4,000.00
Remaining Balance: $96,000
Interest earned at 4% on 96,000 = $3840

So in essence you are always replenishing what you are withdrawing with interest earned on the balance?

Yes if you can lock in an investment that returns 4% each year + inflation without fail you are set and it's that simple.

So far finding that investment vehicle has proven elusive. ;)

-- Vik

[arebelspy]

Am I missing something here?  Isn't the balance of your asset account still earning interest which will offset the 4% withdrawal rate?

Initial Balance:  $100,000
4% SWR:      $4,000.00
Remaining Balance: $96,000
Interest earned at 4% on 96,000 = $3840

So in essence you are always replenishing what you are withdrawing with interest earned on the balance?

Yes if you can lock in an investment that returns 4% each year + inflation without fail you are set and it's that simple.

So far finding that investment vehicle has proven elusive. ;)

-- Vik

Yup.  A reliable 4% investment vehicle has been readily available many times.  It's the plus inflation part that's tricky.  4-5%+ stable investmets (like CDs) tend to happen in higher inflationary environments, where if you did lock in those, you'd lose to inflation.  Unfortunately there is no TIPs available that's at 4% above inflation.  :D

[LiseE]

MMM uses 7% (adjusted for inflation) for overall market performance : http://www.mrmoneymustache.com/2011/06/06/dude-wheres-my-7-investment-return/ with that your investment balances are growing at a rate greater than you are withdrawing.

Regardless .. even in the example given with a dire scenario where the interest was -50% .. overall after 5 years that's a 10K delta compared to example #1.  It could be catastrophic for some but could also be manageable.  MMM always talks about many levels of safety margins. 

100k, taking out 4k a year.  I'm not including inflation just to keep thing simple, plus its only 5 years... not 30.
Example 1.
Yr1 20% growth 116k, yr2 5% 117.8k, yr3 6% 120.86k, yr4 4% 121.7k, yr 5 -50% 56.8k.

Example 2... the inverse.
Yr 1 -50% 46k, yr2 4% 43.8k, yr3 6% 42.5k, yr4 5% 40.6k, yr5 20% 44.7k.

[arebelspy]

MMM uses 7% (adjusted for inflation) for overall market performance : http://www.mrmoneymustache.com/2011/06/06/dude-wheres-my-7-investment-return/ with that your investment balances are growing at a rate greater than you are withdrawing.

Sure, and one could agree with that long term, but still have  a sequence of returns risk.

[theoverlook]

You don't earn an even 4% every year. 

100k, taking out 4k a year.  I'm not including inflation just to keep thing simple, plus its only 5 years... not 30.
Example 1.
Yr1 20% growth 116k, yr2 5% 117.8k, yr3 6% 120.86k, yr4 4% 121.7k, yr 5 -50% 56.8k.

Example 2... the inverse.
Yr 1 -50% 46k, yr2 4% 43.8k, yr3 6% 42.5k, yr4 5% 40.6k, yr5 20% 44.7k.

Same returns, opposite direction equals vastly different returns.  And imagine someone in example 2 who retired 1 year later.
Yr1 4%  100k, yr2 6% 102k,  yr3 5%  103k, yr4 20%  119.7k, yr5 4%  120.5k.

-50% is pretty dang pessimistic.  What are the chances of any of us seeing another -50% event in the stock market in our life times?

[forummm]

You don't earn an even 4% every year. 

100k, taking out 4k a year.  I'm not including inflation just to keep thing simple, plus its only 5 years... not 30.
Example 1.
Yr1 20% growth 116k, yr2 5% 117.8k, yr3 6% 120.86k, yr4 4% 121.7k, yr 5 -50% 56.8k.

Example 2... the inverse.
Yr 1 -50% 46k, yr2 4% 43.8k, yr3 6% 42.5k, yr4 5% 40.6k, yr5 20% 44.7k.

Same returns, opposite direction equals vastly different returns.  And imagine someone in example 2 who retired 1 year later.
Yr1 4%  100k, yr2 6% 102k,  yr3 5%  103k, yr4 20%  119.7k, yr5 4%  120.5k.

-50% is pretty dang pessimistic.  What are the chances of any of us seeing another -50% event in the stock market in our life times?

That's true. It's only happened twice in the last 16 years. Probably won't happen again for the next 70....

[beltim]

You don't earn an even 4% every year. 

100k, taking out 4k a year.  I'm not including inflation just to keep thing simple, plus its only 5 years... not 30.
Example 1.
Yr1 20% growth 116k, yr2 5% 117.8k, yr3 6% 120.86k, yr4 4% 121.7k, yr 5 -50% 56.8k.

Example 2... the inverse.
Yr 1 -50% 46k, yr2 4% 43.8k, yr3 6% 42.5k, yr4 5% 40.6k, yr5 20% 44.7k.

Same returns, opposite direction equals vastly different returns.  And imagine someone in example 2 who retired 1 year later.
Yr1 4%  100k, yr2 6% 102k,  yr3 5%  103k, yr4 20%  119.7k, yr5 4%  120.5k.

-50% is pretty dang pessimistic.  What are the chances of any of us seeing another -50% event in the stock market in our life times?

That's true. It's only happened twice in the last 16 years. Probably won't happen again for the next 70....

The sequence of returns is crazy pessimistic.  The worst 5 year return in the S&P 500 history is -18.83% (1928-1932).  The sequence of returns listed above over 5 years is -31.5%, or more than 1.6 times more pessimistic than the worst 5 year period in history.

[theoverlook]

OK, fair enough forummm, but beltim said what I meant more accurately.  The "sequence of returns" is too pessimistic.

[My Own Advisor]

I think 3% SWR is quite safe.

4% isn't bad but risky.

5% is quite risky and anything higher (6%) is nuts :)

I'm hoping for 7% long-term equity returns.  If I can get my 3-4% via dividends then another 3% growth can be used to fight inflation. 

[forummm]

The sequence of returns is crazy pessimistic.  The worst 5 year return in the S&P 500 history is -18.83% (1928-1932).  The sequence of returns listed above over 5 years is -31.5%, or more than 1.6 times more pessimistic than the worst 5 year period in history.

I think the poster was providing an example of why sequence of returns risk could lead a 4% withdrawal rate to fail. Someone can use cFIREsim to get a sense of how often their proposed retirement plans would have failed using real sequence of returns historical data.

[Indexer]

Am I missing something here?  Isn't the balance of your asset account still earning interest which will offset the 4% withdrawal rate?

Initial Balance:  $100,000
4% SWR:      $4,000.00
Remaining Balance: $96,000
Interest earned at 4% on 96,000 = $3840

So in essence you are always replenishing what you are withdrawing with interest earned on the balance?

You don't earn an even 4% every year. 

100k, taking out 4k a year.  I'm not including inflation just to keep thing simple, plus its only 5 years... not 30.
Example 1.
Yr1 20% growth 116k, yr2 5% 117.8k, yr3 6% 120.86k, yr4 4% 121.7k, yr 5 -50% 56.8k.

Example 2... the inverse.
Yr 1 -50% 46k, yr2 4% 43.8k, yr3 6% 42.5k, yr4 5% 40.6k, yr5 20% 44.7k.

Same returns, opposite direction equals vastly different returns.  And imagine someone in example 2 who retired 1 year later.
Yr1 4%  100k, yr2 6% 102k,  yr3 5%  103k, yr4 20%  119.7k, yr5 4%  120.5k.

-50% is pretty dang pessimistic.  What are the chances of any of us seeing another -50% event in the stock market in our life times?

That's true. It's only happened twice in the last 16 years. Probably won't happen again for the next 70....

The sequence of returns is crazy pessimistic.  The worst 5 year return in the S&P 500 history is -18.83% (1928-1932).  The sequence of returns listed above over 5 years is -31.5%, or more than 1.6 times more pessimistic than the worst 5 year period in history.

I wasn't making market predictions.  I was just explaining sequence of returns for a poster that didn't understand how it works.

Examples 1 & 2 both experience markets down about 31% over that time period.  However example 2 has far less money because the crash happened in year 1, not year 5.  The point of the exercise was to show how sequence of returns works.  Mission accomplished!

I figured 5 years worth of extreme data would prove the point just as easily as 30 years of more realistic data.  ;)

[beltim]

The sequence of returns is crazy pessimistic.  The worst 5 year return in the S&P 500 history is -18.83% (1928-1932).  The sequence of returns listed above over 5 years is -31.5%, or more than 1.6 times more pessimistic than the worst 5 year period in history.

I think the poster was providing an example of why sequence of returns risk could lead a 4% withdrawal rate to fail. Someone can use cFIREsim to get a sense of how often their proposed retirement plans would have failed using real sequence of returns historical data.

That's fine, but I was responding to your post. Sure, markets may have fallen 50% (though not in a calendar year), but they've never fallen 31.5% over 5 years.

[arebelspy]

The sequence of returns is crazy pessimistic.  The worst 5 year return in the S&P 500 history is -18.83% (1928-1932).  The sequence of returns listed above over 5 years is -31.5%, or more than 1.6 times more pessimistic than the worst 5 year period in history.

I think the poster was providing an example of why sequence of returns risk could lead a 4% withdrawal rate to fail. Someone can use cFIREsim to get a sense of how often their proposed retirement plans would have failed using real sequence of returns historical data.

That's fine, but I was responding to your post. Sure, markets may have fallen 50% (though not in a calendar year), but they've never fallen 31.5% over 5 years.

This is fine, but I think his point was just illustrative, that sequence of returns matters.  He could make it so it ended with a 7% CAGR after 30 years, and show vastly different outcomes.  The point wasn't that those returns are realistic, but that when returns comes matters.  It was certainly exaggerated to show that point.

[beltim]

The sequence of returns is crazy pessimistic.  The worst 5 year return in the S&P 500 history is -18.83% (1928-1932).  The sequence of returns listed above over 5 years is -31.5%, or more than 1.6 times more pessimistic than the worst 5 year period in history.

I think the poster was providing an example of why sequence of returns risk could lead a 4% withdrawal rate to fail. Someone can use cFIREsim to get a sense of how often their proposed retirement plans would have failed using real sequence of returns historical data.

That's fine, but I was responding to your post. Sure, markets may have fallen 50% (though not in a calendar year), but they've never fallen 31.5% over 5 years.

This is fine, but I think his point was just illustrative, that sequence of returns matters.  He could make it so it ended with a 7% CAGR after 30 years, and show vastly different outcomes.  The point wasn't that those returns are realistic, but that when returns comes matters.  It was certainly exaggerated to show that point.

I'm not disagreeing with that.  My response was specifically to forummm about the pessimism.  That's all. It's not important to the larger point.

[TomTX]

Am I missing something here?  Isn't the balance of your asset account still earning interest which will offset the 4% withdrawal rate?

Initial Balance:  $100,000
4% SWR:      $4,000.00
Remaining Balance: $96,000
Interest earned at 4% on 96,000 = $3840

So in essence you are always replenishing what you are withdrawing with interest earned on the balance?

The big thing I see "missing" is using the term "interest" which implies a guaranteed return.

That's thinking like all of your money is in a savings account or CDs. If you are in the discussion of a SWR for early retirement, you either have to invest a big chunk of your money in something other than fixed income (Stocks, rental real estate), or have a SWR so low that you won't retire early.

[Filliteracy]

Discover The Little-Known, But Highly Dangerous Risks Hiding Behind The 4% Rule, And The Simple Solutions to Correct the Problem

http://financialmentor.com/retirement-planning/how-much-money-do-i-need-to-retire/safe-withdrawal-rate/13192

What do you think about this article?

Am I the only one who finds the "each paragraph is one sentence long" style of writing seriously annoying? Honestly, if you aren't able to write two or more sentence consecutively, you're literally just yelling out random free-standing statements, and avoiding having to actually build an logical supported argument.