You all know I'm forever suggesting Monte Carlo analyses when addressing the retirement decision. It works. But that's not just me talking. The financial planning industry has been talking that same talk for at least several decades. A fellow named Bill Bengen initially used that calculation discipline when he concluded that a 4% annual drawdown rate resulted in high portfolio survival odds for an extended retirement period. Here is a Link that updates some of his initial thinking on this matter:https://www.nytimes.com/2015/05/09/your-money/some-new-math-for-the-4-percent-retirement-rule.html
Enjoy! Note that Mr. Bengen now feels that a more generous 4.5% drawdown rate is portfolio survival safe. With a little more attention to market conditions, more recent studies are currently suggesting that a 5% drawdown results in acceptable portfolio survival odds. However, note that these are just statistical studies with many assumptions embedded in the analyses so be cautious. Risk at the 5% drawdown schedule must be higher than at the 4% level. That's obvious. The final decision is always yours alone. It must include your comfort level; your sleeping well each and every night. Take care. Sleep well.
Think about it ... Because, for those that have reached critical mass, it works. The secret is to control spending by living within ones means. For those living in retirement that have not reached critical mass times ahead could indeed become most difficult.
Not exactly. He concluded that a 4% drawdown rate resulted in certain survival, not merely a high probability of survival: "no client enjoys less than about 35 years before his retirement money is used up." Survival is typically taken in financial publications to mean lasting 30 years.
More importantly, for the most part he used actual not statistical data. He looked at rolling 50 year periods, starting with 1926 (i.e. 1926-1976) and ending with the 50 year period 1976-2016. Monte Carlo had nothing to do with this.
You may well ask: what "actual" data did he use for years that were in his future (his paper was published in 1994)? Well here he did use statistical data. But of the simplest kind, again no Monte Carlo simulation. He merely "extrapolated the missing years at the average return rates of 10.3 percent for stocks, 5.2 percent for bonds, and 3.0 percent for inflation - a concession to the 'averaging' approach, but one that was unavoidable."
Bengen used actual returns over multiyear spans (i.e. he did not assume that year-to-year returns were random and independent). He filled in missing data by using constant annual returns (i.e. no variation of returns). Everything Monte Carlo is not.
Quotes are from Bengen's original paper, cited in the NYTimes article linked to by MJG. See Figure 1(b) in that paper for how many years a 4% drawdown rate would last if started in any year from 1926 to 1976.
Thank you all for reading my comments on Mr. Bengen and Monte Carlo retirement tools. My goal was not to tout Mr. Bengen, but much more importantly, to encourage you Guys to try a powerful Monte Carlo simulation for planning purposes.
If Bengen "concluded that a 4% drawdown rate resulted in certain survival", he was wrong. In just a few minutes I did simulations on a code, that I often recommend (Portfolio Visualizer), to estimate portfolio survival odds for drawdowns being discussed. These codes do thousands of what-if cases and almost never find a 100% survival likelihood except for uninteresting extreme cases given typical uncertainties in market annual outcomes.
I did two sets of calculations: one maximized risk by assuming a 100% US equity portfolio, and a second set that was more balanced by assuming a 50/10/40 portfolio of US equity, International equity, and US bond asset allocations. I used historical market returns in my calculations.
For drawdowns of 4.0%, 4.5%, and 5.0%, the all equity portfolio failed 14%, 18%, and 26% of the time for a 30 year test period. For the same drawdowns, the more balanced portfolio only failed to survive 3%, 6%, and 12% of the time, respectively. Diversification works as a form of portfolio survival protection. These calculations only took several minutes to complete. They demonstrate the power and usefulness of Monte Carlo simulations. Please take advantage of this resource for your own retirement planning purposes.
When doing thousands of simulations with some statistical distribution of outcomes, some failures are often projected. An investor must decide what portfolio failure rate is acceptable. In my case, my target was to reduce failure rates to under 5%. Given the uncertainties of the marketplace some risk always exists. The goal is to construct a portfolio that projects a rare failure probability given the target drawdown schedule.
Even with careful planning, crap happens. If the market rewards turn sour, the drawdown schedule should be a candidate for adjustment. Flexibility improves survival odds.
Given the Monte Carlo tools that can be easily accessed, you guys can do a better job than Mr. Bengen in terms of projecting future probable outcomes. Good luck. Take time to explore this tool and you will reduce your need for luck.
To that end, you cited one of the most well known papers on retirement planning as evidence of how well Monte Carlo works, even though it didn't use Monte Carlo. I pointed out that Bengen found zero real world return patterns where a 4% drawdown would fail (over 30 years); your response was to disparage the original work you cited approvingly.
It's enough to make one wonder whether you read the paper.
Instead of comparing and contrasting methodologies, you continue to effuse about Monte Carlo. Bergen took a different approach using using actual returns, that virtually everyone here can understand and use to draw their own conclusions.
In contrast, Monte Carlo spews out magic numbers (not unlike M* star ratings) that leave one to one's own devices to interpret. As guidance you proffer that you consider a 5% risk acceptable, but you didn't give any reasoning, rendering this fact useless. (I wonder why you used these 30 year projections at all; as I recall you've indicated an age which suggests that a 30 year horizon is, shall we say, rather optimistic.)
Even the probabilities posted are meaningless because unlike Bengen, you didn't state the assumptions you used, such as the input values for mean and standard deviations of stocks, bonds, and inflation. Nor did you even apply the same asset allocation that Bengen used.
Did you consider skew and kurtosis (the S&P 500 exhibits both)? Do you think that most people using these "push a button" tools even understand that question? (No disrespect of MFO readers is intended; many have stated that statistics is not their forte.)
The fact that a program can do thousands of computations in seconds is not so much a demonstration of the usefulness of a program as much as it is a testament to the operation of GIGO. A scalpel is a great tool in the right hands; in other hands it can be destructive.
When all one has is a hammer, everything looks like a nail.
"Did you consider skew and kurtosis?"
@msf- I thought that those were foot problems. What you mean statistics aint my forte??
Kurtosis has to do w/ outliers, lots vs few (I hope).
Not a totally unexpected attack from a trio of MFOers who often submit Ad Hominem posts directed to discredit me. That troubles me not one iota. The more they protest, the more I suspect I'm making some positive inroads.
I certainly do not apologize for my commitment to Monte Carlo analyses. I profitably used Monte Carlo analyses when working as a research engineer. When I was considering my retirement in the early 1990s, I could not find an operational Monte Carlo code dedicated to retirement planning analyses. With help from Bill Sharpe and Gene Fama,who sent market data sets and volunteered some suggestions, I wrote my own Monte Carlo code. I used that tool in making my retirement decision. It helped.
These guys protest much to much, and accuse me without knowing me or my capabilities. Their ranting polemics say much more about them than about me. They are small men!
I will continue my march. Monte Carlo is not for everyone, but MFOers should be made aware that the code is freely available and is a candidate to add to your investment toolbox. As always, it is your free and informed choice.
@MJG: that's "much too much, not much to much. You're welcome.
Ashes to ashes,
mush to mush...
Thanks to @msf for the analysis. Hope the personality aspects here don't obscure your contribution.
@CecilJK , thank you for contributing your system. I'd like to hear from others on their approach. Obviously there is no one way to do it. The ability or planned 'cushion' that would allow one to be flexible appears to be key.
As for Monte Carlo, I don't know a better, easier method that suggests if you are on tract with your savings versus retirement spending expectations. Is there one? There are no guarantees, but Monte Carlo at least supplies guidelines and gives at the very least, a ball park view.
No matter how well the withdrawal modeling system is unless an investor has reached critical mass their money is not going to last. For me, since I have reached critical mass for my lifestyle the taking of up to one half of the five year average return works well for me. I know of others before me that used it and it seemed to have worked well for them because they lived within their means. This is the primary reason I use this system plus it was handed down to me from my late father. Sometimes, as adults, we try to make things way to complex when a simple means will provide ample infromation over the more complex one. After all, I still living off of my fathers assets and I have been growing them through time. The real secret (again) is to live within one's means. And, what may be critical mass for one might not be for another.
If their are those that want to get caught up in complex modeling systems that is ok by me. I'm one that has chosen to go with the more simple approach.
I can't relate to the central question of how to survive "X" number of years on "X" number of dollars invested. Reason: I enjoy both a defined benefit pension with a partial COL rider and also a decent SS income stream. And, supplementary health insurance through retirement plan as well. Conceivably, these would provide for basic living expenses - though it would be a very "spartan" lifestyle without travel or other things that make retirement enjoyable.
In my highly atypical instance, even after taking distributions, retirement savings have roughly doubled over the nearly 20 years since retirement (albeit in nominal dollar terms only). At the same time, more than half of that has now been placed under the Roth umbrella, whereas at the time of retirement none was. Much of the reason for the increase is that the money was left largely undisturbed during the first 10 years.
As far as the article's mention that withdrawals are not linear or equal every year - I couldn't agree more. There have been years when I needed to take a larger sum - say as a sizable down payment on a new car or for unexpected home repairs - and other years when I've needed very little.
I don't envy those without a pension or other solid income stream in retirement. Not everyone would be satisfied with a somewhat spartan lifestyle either. As I look at the markets over the past 10-20 years, I'd not be eager risking a large retirement nest egg with an aggressive approach in retirement. Lots of warning signs IMHO. But, no one really knows. As I said at the start, the problem with these mathematical models is that the next 25 years could be markedly different than the last 25 - as others, notably msf, have tried to explain.
Thanks for your most recent contribution. An honest post on a controversial topic is always appreciated. The referenced article is indeed excellent. That's why I posted it.
The article is dominated by references to Wade Pfau observations. He too is a very strong advocate of Monte Carlo simulations to help arriving at retirement decisions. These codes also do yeomen service when updating after the initial decision. Change happens. As I too frequently say: forecasting is hazardous duty.
Here is a Link to a Pfau article in which he discusses some of the advantages offered by Monte Carlo retirement planning tools:
This article by Wade Pfau does an honest job at discussing both the merits and shortcomings of Monte Carlo simulators. The Pfau reference ends with the following paragraph:
"Overall, the advantages of Monte Carlo simulations likely more than make up for any deficiencies when compared to the results we obtain using historical simulations."
Since our debate over Monte Carlo codes has extracted some very emotional responses from the FMO membership, I'll take this opportunity to recommend yet another free Monte Carlo tool on the Internet. It is titled The Flexible Retirement Planner. Here is the Link to this superior tool:
I have referenced this code in earlier posts. It is user friendly. I hope you visit this site and do a few experimental calculations. I believe you will be impressed with its speed and the options available on this site. Please give it a try. You will get a quick feeling for your portfolio survival prospects, especially given your concerns over uncertain market returns in coming decades.
That uncertainty is a reasonable concern. And that's exactly the type of problem Monte Carlo,codes were designed to address. In near zero time, you can explore a range of these uncertainties, and estimate their occurrence probabilities. Your portfolio asset allocations and spending profile can be adjusted to accommodate these uncertainties.
I hope this helps.
Nothing like modesty MJG. (And a reason some of us take umbrage with some of your posts).
MikeM termed the NYT article "very good". I termed it "good" - adding a reservation regarding source. I don't have time to enjoy the additional sources you've linked. I'm sure others will.
My contribution was mainly to show how each of us has a unique circumstance and unique needs in retirement planning. I shared more than I generally do in the hopes it might help others address their own varied needs. My purpose wasn't to support or condemn Monte Carlo. In fact, in my own case, I've run no simulations (other than in the back of my head from time to time). Period.
I'll defer to @msf on the overall merit and accuracy of the calculations presented by you and/or your sources. I've found his math skills over the years both considerable and commendable. By contrast, I barely survived high school Algebra with a C, and have assiduously avoided all math classes since.
"In my own case I barely survived HS Algebra with a C - and have avoided all math classes since." Me too.
Re Algebra - Great minds think (or fail to think) alike.
Making a point of the distinction between MFOer's differing ratings of an article as good, very good, or excellent is just being picayune. At our generic interaction level, it is a distinction without a meaningful difference.
From my perspective, any candidate article is simply worth a read if any of these ratings appear. I do screen articles before recommending them to FMO members. I try to never reference junk. I recognize that MFOer's time is valuable and limited.
You need not be a math wizard to use and appreciate Monte Carlo simulators. These days, anyone can successfully operate a car without ever opening the hood. Anyone who invests today can easily operate and maneuver his way around Monte Carlo websites without fully understanding the detailed mathematics. Those sites that I recommend are user friendly by any measure.
I agree that each of us have unique circumstances and needs. That's one reason why I am reluctant to offer specific mutual fund recommendations. I rarely, if ever, do so.
I fully recognize your reluctance to use Monte Carlo analyses. You may not be comfortable with using math that you don't understand. Whatever the reason, it Macht Nichts' to me. I respect your freedom of choice and would never challenge it.
I really wish you investment success. I hope your back of the head simulations are trustworthy.
That's not what I said.
"Making a point of the distinction between differing ratings ... as good, very good, or excellent is just being picayune."
When I went to school B meant good and A meant excellent. You managed to turn 2 Bs into an A.
"Whatever the reason, it Macht Nichts' to me." I respect your freedom of choice and would never challenge it."
"I really wish you investment success."
I started investing in 1970.
Are any of guys longshoremen? Given the language, Hank's last statement suggests some possible coupling.
His closing comment quickly summarizes your great distaste for my posts. Often you misrepresent my positions, even to the extent of quoting only fractions of my statements to distort my actual meaning. That's not playing an honest game.
So your team effort is reduced to ending with a crude and vulgar proclamation that is not common on MFO. That's too, too bad. It just demonstrates the soundness of my arguments and the shallowness of yours.
Those who have followed these exchanges (forced by you), and have kept score in a general way, fully know who the winner is. And that is the sad outcome from a beligerent team effort to discredit rather modest submittals from an individual investor posting in an ad hoc manner. Consider also, that some members of that assassination team are professionals in this field. Shame on you guys. It's like the NY Yankees trying to win a ballgame from a sandlot team and losing.
I harbor no grudges. I suspect our investment philosophies and practices are not all that disparate. You seem to concentrate on me and my writing style rather than on the substance of my submittals. Why such a personal attack? Perhaps you are envious of the folks that I possibly attract with my posts and references. I am just trying to be helpful, and seek no special recognition.
Your persistence and antagonism puzzles me. Lots of things puzzle me. I have grown weary on this topic, so this will be my last posting on this matter.
"If Bengen 'concluded that a 4% drawdown rate resulted in certain survival', he was wrong" and "The article is dominated by references to Wade Pfau observations. He too is a very strong advocate of Monte Carlo simulations to help arriving at retirement decisions. "
The NYTimes article has a graphic with three other spending models by Pfau. All three show 100% survival over thirty years (worst case shows money remaining for all models). That includes a model with a constant (inflation adjusted) drawdown amount.
Yet the simple Monte Carlo tools advocated (based on mean and standard deviation inputs) intrinsically contradict this - they are built on the premise that failure is always possible (since they say that a portfolio can lose value year after year after year after ...). Does that mean that Pfau, like Bengen, was also wrong in concluding certain survival?
A problem is that by design, these simple tools are unable to conclude that survival is certain. Regardless of inputs. If you build a conclusion (failure is always possible) into a tool, you've rigged the results. You can't use these tools to "prove" that 100% success is impossible. They're unable to say anything but.
It's fine to use random number generators (aka Monte Carlo) to "run" models many times and see what outcomes might result. The problem is not in how models are used (trial and error - random numbers), but with the models themselves. Unfortunately these tools conflate the creation of the models with the Monte Carlo running of the models to generate a range of possible outcomes. Don't confuse a criticism of these tools with a criticism of Monte Carlo simulations.
These tools create simplistic models that usually assume each year's market's performance is independent and that returns are normally distributed (bell curve).
But data suggest that stock market performance is a leading indicator of business cycles. Thus stock market performance is itself cyclic (not independent from year to year) albeit with an upward bias.
"stocks as a whole move in advance of the economy" = AAII Journal, Aug 2003
As to the bond market, the trivial Monte Carlo models assume that nominal returns are independent of inflation. The Fischer hypothesis suggests the opposite.
"The Fisher hypothesis is that, in the long run, inflation and nominal interest rates move together." http://moneyterms.co.uk/fisher-effect/
The first paragraph by Pfau in his Forbes column says that the models need to include correlations - something that's antithetic to simplistic free Monte Carlo tools that assume independence of inputs in building their models.
His penultimate paragraph states simply that: "the results of Monte Carlo simulations are only as good as the input assumptions, ... Monte Carlo simulations can be easily adjusted to account for changing realities for financial markets."
It's certainly easy from a mechanical perspective to adjust the models (e.g. by changing the mean return for bonds). What's not easy at all is figuring out what adjustments to make. That gets right back to the results being "only as good as the input assumptions", or as I wrote before, GIGO.
Again quoting Pfau: "Many financial planning assumptions are based on historical returns; however, these historical returns may not be relevant in the future."
At best, even if a model is good and analysis sound, all you're going to get is a sense of whether you're saving enough (i.e. what MikeM wrote). It's of less help during retirement because, as hank noted, extraordinary events happen.
I'm wondering who the unnamed "professionals in this field" are. Or even what "this field" is. But for the record - I've never taken a statistics course in my life. I'm just an individual investor like most people here, albeit one who did once ace a course in writing and research.