Bogleheads

[DaleMaley]

Article


Spending Safely: Advisers now concede that fluctuating conditions make rigid formulas for drawing down savings unrealistic

By Lynn O'Shaughnessy

[bozo]

[bob90245]

[daryll40]

Interesting article. I suspect there are some groundbreaking findings here that may require more research.

In particular is the part where they compare two people retiring a year apart where the market tanks in between. This is the issue that I went around and around with Larry Swedroe about last year. I pointed out that it was arbitrary that someone could retire on $1M and take out $40,000 for his 4% SWR. Yet a year later after the market tanks and the portfolio is only worth $800,000 he is still taking out $40,000 plus inflation while a similar new new retiree could only take $32,000 on his $800,000. The only thing that is different is that the first guy had a greater portfolio IN THE PAST.

Larry then pointed out that "re-retiring"...which was what I was suggesting during after a GOOD year as a solution to equalize the random nature of
one's portfolio value when he retires....would just increase the risk of failure. In the end we both were right...the above scenario was arbitrary but reretiring does increase risk. We never did come to an agreement about how to reconcile the two issues.

This research seems to solve the arbitrariness by linking the SWR to market valuation. With high market valuation, you can still take an SWR of as much as 4%. But with lower valuations you could take more. If your portfolio later rose, "re-retiring" would not be an issue because although you have more to take from, the rate would go down as the valuation of the market rises. The market PE would set your SWR sort of like a golf handicap. If I retire into an overpriced market, my SWR has to be lower to compensate. But if you retire into an underpriced market, your SWR can be larger to compensate. No more problem with arbitrary numbers based on a portfolio value from last year!

Also of interest is that it seems to imply that 4% is even safer than we realized because even if you are retiring at time when the market PE is above 20, you could still use an SWR EVEN of 4%. Meaning that if you limit yourself to 4% or less, you'll still squeak by if the market starts out overvalued but will have extra safety breathing room if you retire into a not overvalued market.

I'd like to learn more. I am not sure how he determined that a 20 PE was the tipping point of overvalued vs undervalued.

Larry Swedroe....might you have any thoughts on this? I know you think about such stuff and I would value your input.

[grayfox]

[ataloss]

[daryll40]

[bob90245]

[daryll40]

Thanks for that link. I just printed it and will read it soon.

[Elb]

I've been interested in the corelation between market valuation and returns for a while, but after getting into it, I saw that the Web's leading (or at least loudest) proponent ot the theory had been ostracized. Accordingly, I assumed that the financial planning community had dismissed the idea. It now seems he was ostracized merely for being an asshole.

Anyway, I disected the linked article, and the the theory and data hold together very well. Which is unfortunate for a 65-year-old on the cusp of retirement.

However, I have been out of the market for a while, in cash in Korea, ready to head back home, so there is still a way I can avoid a crappy retirement: The S&P steadily continues on down toward 1000, and I step gently in. Could someone please message me when it's time?

Elb

[grayfox]

[mikec]

This article is really excellent. To sum it up, a SWR of 4-4.5% has been appropriate for the last 10 years or so, but in the future, if the PE10 drops below 20, then the SWR could be increased by 0.5% to 1.0% depending on low it goes. I didn't check, but I suppose that the SWR has fallen pretty close to 20 now.

And Eld, it is now time to start DCA!

[heyyou]

According to Gummy, the near 4% plus annual inflation rate would have even worked for the 30 years starting in 1928, '65, '67, and '68. No one knows if 1999 will get added to the 4% only list in 2029. Retirees that started in all the other years (the other 92% of the time periods), could have used higher WD rates. Per Gummy, the '48 and '49 retirees could WD over 14% safely for the next 30 years. See the broad variation over those 50 previous maximum SWRs for 30 year retirements?

If you are in the business of giving advice, you won't get sued for dispensing advice that is too conservative. The professionals need to be conservative. If you are just an internet poster, you have more flexibility in your suggestions since you are your only client.

Bozo, Wm Bengen's SWRs include an annual inflation adjustment so the retiree is not WDing your straight 4%. A straight % is fine if you can live on the variable income like a small business owner or a commmissioned sales person always has. A retired salaried person would be used to steadier income. The intention is to spend as much as possible up until death, without running out of money. Maintaining the initial portfolio value with your straight 4% leaves the portfolio whole at death, so your heirs will like that. Jarrett and Stringfellow showed maximum WD rates for that, their business was named Zumma. I think Jarrett was the financial advisor and Stringfellow was the computer programmer.

The Kitces article is well worth reading. The BW one was light on content, but did give exposure to some good authors.

[TimDex]

The Kitces article is excellent (though I would love to see the math and spreadsheet inclined members of this forum pick it apart).

People are wary of dealing with valuations but I don't see how you can ignore extremely high or low markets in making these determinations. Just the same, I don't see how advice people take away from this forum can be so exact. I think it's awfully hard to give anything more than a rule of thumb when you are dealing with people over the internet, and not as a client across the table.

Tim
_________________
"All man's miseries derive from not being able to sit quietly in a room alone. " -- Pascal

[richard]

From the July 3 WSJ:

"Two percent is bulletproof, 3% is probably safe, and 4% is the absolute maximum," says William J. Bernstein, an investment adviser in North Bend, Ore. If the $1 million is in a tax-deferred account, such as an Individual Retirement Account or a 401(k), he notes, "you've got to pay taxes on those assets on the way out." So you have roughly $800,000 to work with -- or $32,000 a year maximum to spend.

I'd suspect the advisors in the article are telling people what they want to hear. There's another article a few pages later in the print edition about advisors who promised high withdrawal rates to win business, and did quite well for themselves (commissions and fees) until their clients ran out of money and starting suing.

[rmark1]

The 4% of initial portfolio value + annual inflation withdrawal is based on -
Past US returns and volatilities from a roughly 60/40 fixed weight market portfolio, no costs or taxes

Once you step outside that historical record to -
Other countries market records, other asset mixes, other inflation rates, adding past costs and taxes

I suspect the 4% + annual inflation has a much lower success rate than is usually stated

[stratton]

[bob90245]

[dbr]

[ataloss]

I think originally Bengen, Trinity and REHP came up with 4% withdrawals with inflation adjustment as nearly 100% successful in the past. Good information to have but predicting the future is hard.

Clearly 4% is sometimes far too conservative. John Greaney made this point at the REHP board. Often retirees will end up with growing portfolios despite withdrawals. I am not sure I would bet the farm on a much higher withdrawal rate. Maybe wait and see and re-set withdrawals if the problem of real portfolio growth is a concern

With so many unknowns (future spending, longevity) it isn't as if some insight into the best withdrawal rate is likely to be critical, IMO.
_________________
Have Fun!

[Robert T]

.
Bernstein has a earlier article on this The Retirement Calculator from Hell – Part II. He used different portfolio returns in the two tables which show different probabilities of success for the same withdrawal rates (note the returns used in his table are real returns). I used Bernstein’s MRetire MC simulator and managed to fairly closely replicate his results.

The assumptions he seems to use is that the entire portfolio is held in a non-taxable account, and there are no expenses. I have adjusted some of the assumptions to include taxes on half the withdrawals and an annual expense ratio of 0.3% (with MCRetire allows you to do).

A few results from the table below:

- Withdrawal success rates do depend of portfolio returns (no surprise).
- A 5% withdrawal rate has similar odds of success as a 4% withdrawal rates if portfolio return were 2% higher (in the example).
- A 2% withdrawal rate does seem bullet proof (99% success rate even at lower expected returns)
- 3% is ‘probably safe’. With 80% success rate even with portfolio returns 3% lower than the historical average.
- 4% still seems okay if portfolio returns are around the historical average, but with returns 3% lower there’s close to an even chance of having enough money or running out.
- Above 4% and the odds of running out of money rise fairly quickly even with returns around the historical average.
- Finally, read Bernstein’s Part III in the series on the ‘meaningfulness’ of odds of success beyond 80%.

Assumptions:

P3 = close to 1927-2005 returns and SD for a 50:50 TSM:5-yr T-notes portfolio (the actual backtest returns were 8.2% with SD on 10.7)
P1 & P2 assumes higher returns for the 50:50 portfolio with the same SD
P4 to P6 assumes lower returns for the 50:50 portfolio with the same SD

Inflation = 3%
Income tax rate on withdrawals = 27.5%
Annual portfolio expense ratio = 0.3%
Share of portfolio held in taxable accounts = 50%
The same asset allocation is held in tax and non-taxable accounts

[gummy]

Beware very low SWR numbers.

Here's a very safe withdrawal ritual:
Put your money under a pillow.
Withdraw X% per year, increasing with inflation at 3%.
Your money lasts for 30 years.
What's X?

Now put that money in a bank paying 2% annual interest.
Withdraw that same X% per year, increasing with inflation at 3%.
How long will it last?

[ralphjones]

2%, 40years

check out this really neat site called gummy-stuff. org and...

ahh, well I guess you knew the answer all the time then

R

[bobcat2]

grayfox quotes an article as saying the following:

[gummy]

I never could understand why anybuddy would adopt some X% rule then follow it ... ignoring current market machinations.

You live in Cancun. Whether or not you go to the beach depends upon the weather.
Do you:
[1] Look out the window the see the current weather conditions?
or
[2] Look at the worst historical weather and find that Hurricane Gilbert tore up the town in 1988
... and let that event determine your comings and goings.

Tho' Monte Carlo simulations (with an assumed return distribution, Mean Return and Standard Deviation) is neat, remember:

[dbr]

It would seem evident that the "4% rule" is not a plan that a retiree should follow to manage investments, income, and expenses in retirement. Whether such a rule ever got to be a serious plan proposed by anyone is not clear, although there are many who use the analysis as a straw man in discussions.

The "rule" originated in attempts to understand the fate of portfolios in retirement and was simply a model within which one could understand the effects of asset allocation, return, volatility, inflation, and withdrawal. The context was an environment in which some advisors were brandishing excessive estimates of equity returns and ignoring the effects of volatility while withdrawing from assets.

I am of the opinion that the "rule" is still a good baseline to consider when looking for a zero order estimate of whether or not a person's accumulated assets are sufficient to support retirement. The "rule" functions as a simple and easy to calculate reality check.

[bobcat2]

I agree. The 4% SWR is a useful rule of thumb to make very crude estimates in your head to see if what someone wants to do in retirement is at all reasonable. That is the totality of the rule's, and its variants, usefulness.

Bob K
_________________
The real difficulty in changing any enterprise lies not in developing new ideas, but in escaping from the old ones. - JM Keynes
Good questions outrank easy answers. - PA Samuelson

[Oicuryy]

It is not that hard to calculate your safe withdrawal rate. It is the reciprocal of the gummy Magic Sum of your portfolio's future returns.

Ron
_________________
Money is fungible

[TimDex]

There seems to be a contradiction between the results Kitces is positing, and what Robert T comes up with...(tho' I may not be the best person to determine that.)

If you compare Robert T's results for P3, which approximates an historical 50-50 portfolio, there is a real drop in success rates as you hit 5%.

It's a little hard to compare Kitces' results, because he's using PE10 to determine results, (and he uses a 60-40 portfolio) but his results for quintile 3, (with pe's ranging from 14.7 to 17.6, in the middle of the pack, and which I assume is average), shows 4.9% as a lowest SWR, and an average SWR of 6.3%.

I would agree that a SWR of 4% is simply a useful rule of thumb, that allows a person to gage whether retirement is a possibility. Looking at Robert T's results, I still wouldn't make a bet above 4%, just the same. And for a person retiring early, 3% is a lot better.

Tim
_________________
"All man's miseries derive from not being able to sit quietly in a room alone. " -- Pascal

[gummy]

Q: Is the 4% rule useful?
A: Yes.

You're young and would like to know how much you should have in your portfolio when you retire, in 30 years.
So you imagine an income that you'd like to have from your portfolio ... in today's dollars. e.g. $30K
Then you increase that by 30 years of inflation. e.g. $30K (1.03)^30 = $73K
Then you multiply that by 25 (since 25 = 1/4%). e.g. $73K*25= $1.8M
Then, when you retire, you forget that 4% rule.

[bobcat2]

Notice how Gummy's advice is the exact opposite of the advice of financial practioners. Their advice is that before retirement use the Retirement Income Replacement Rate (RIRR) rule and plan on having 85% of your pre-retirement income in retirement. Once retired forget the RIRR rule and replace it with a SWR 4% rule or one of its variants.

Bob K
_________________
The real difficulty in changing any enterprise lies not in developing new ideas, but in escaping from the old ones. - JM Keynes
Good questions outrank easy answers. - PA Samuelson

[gummy]

Here's a quick & dirty (and FUN!) MC (fer them that like the Full Monte):
http://www.gummy-stuff.org/portfolio.xls

You press F9 every time you'd like another set of (lognormally distributed) annual returns

[nisiprius]

I'm increasingly puzzled by the SWR stuff. It seems like a quest for perpetual motion. Make the very strange but customary assumption of a thirty-year time horizon, with "success" meaning the portfolio lasts at least thirty years. In all cases, the "withdrawal rate" means initial withdrawal rate, with the assumption that is it subsequently cost-of-living adjusted.

Then the following points seem clear to me.

1) If we assume that there's no interest, no return on investment, and no inflation, then the safe rate of withdrawal is exactly 3-1/3%. No simulation needed.

2) If we assume TIPS are available, with a real rate of 1.31% or more, then the safe rate of withdrawal is 4%, with zero chance of failure. No simulation needed.

3) For all studies I've seen, if the withdrawal rate is down a bit below 4%, then the portfolio succeeds regardless of what the asset allocation is.

4) If the withdrawal rate is much above 4%, then the portfolio has a significant chance of failure. Writers sort of handwave, making an implicit assumption that a 5 or 10% failure rate is OK. That's the sort of statement that people are inclined to go along with if it's made in a matter-of-fact way that suggests that everyone accepts it. But if people were asked to name an acceptable failure rate in advance of reading anything about SWR, I wonder how many really would name a percentage that high?

5) Increasing the percentage of equities in the portfolio only reduces the failure rate when the failure rate is pretty high to begin with.

6) Although equities have an excellent chance of rewarding you in the future, you simply cannot count them before they actually hatch.

7) If you can live within 4%, you ought to be able to live within 3%.

8) If you've got a Scrooge McDuck money bin, live on "the interest on the interest," as rich families did during the Gilded Age. If you've got plenty of money, live on the income and don't invade the principal. If you've got enough money, live on 3%-then-COLAed and don't sweat it. If you've got just barely enough money or not quite enough money, look into income annuities rather than trying to predict portfolio futures.

I see no scenario where it makes sense to try to refine the SWR beyond saying "3% is OK and you can probably get away with 4% most of the time."

P. S. It's never clear to me what exactly you're supposed to do if you follow an SWR strategy, and, in at the start of year number 27, discover that a) you're in pretty good health with a life expectancy of another 6.5 years, and b) have only 10% of your initial portfolio left. There's a lot of handwaving about "course corrections," but if you're ready, willing, and able (say) to reduce expenses by 20% in year number 17, why wouldn't you just reduce them in year number 1 in the first place? If anything, you'd think your ability to control expenses would decline with declining age and health.
_________________
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

[gummy]

[bobcat2]

SWR strategies appear to be the financial practitioner's way of steering the retiree toward consumption smoothing throughout retirement. That's a laudable goal in most cases. It's just that the SWR rule of thumb strategies are not very reasonable methods to accomplish that goal.

Bob K
_________________
The real difficulty in changing any enterprise lies not in developing new ideas, but in escaping from the old ones. - JM Keynes
Good questions outrank easy answers. - PA Samuelson

[bob90245]

[Rodc]

[EyeDee]

.
Gummy & Rodc,

Considering the performance problems the Bogleheads’ servers have been experiencing, it might be helpful if you edited your posts and stopped the moving graphs at least until we get to the new server.

Thank you,
_________________
Randy

SSCA - Security (ER fund, insurance), Savings, Costs (inc. taxes), Allocation (inc. rebalancing).
Board Wiki

[heyyou]

Bengen published in the early 1990s. As others mentioned, prior to that, there was no general info about how historical returns had impacted retiree portfolios. Fifteen years later, posters are poking at the imperfections of the very first model.

I'm grateful for Bengen's pioneer work. He started with basics, then expanded from that in later papers. Have you read all of them? I think he did a lot, and others can and will add to his original ideas. Today, you could easily repeat his work. Now try that with 1991 vintage resources--a 286 or 386 computer, an early spreadsheet program, using the historical returns from a printed book, not a download. He was not a programmer, he was a financial advisor. Give the guy some credit for creativity on an original idea. Other than Gummy who is a math professional, are any of the other critics here offering better SWR programs? Much is obvious to others after some original thinker lays a foundation for them to build on.

[bob90245]

[bobcat2]

Consumption smoothing means to keep discretionary spending non-volatile from year to year. A few spending categories, such as mortgage payments and Medicare premiums, offer very little discretion. However, most spending, like how well I will eat, offers a great deal of discretion.

The point is that the portfolio AA determines the smoothness of the spending from the portfolio, not arbitrary spending rules (withdrawal rates). At one extreme complete real annuitization would provide you with a level real stream of income to support very smooth consumption throughout retirement. At the other extreme 100% investing in emerging market small cap stock funds would be expected to give you higher expected future income streams, but at the cost of an expected very bumpy consumption ride and a small, but not negligible, probability of an extremely low portfolio consumption path. It’s not the rules that dictate the consumption path, it’s instead the AA that dictates the consumption path.

A simple rule to apply is that aggressive retirement investing should be married with defensive retirement spending rates. The riskier the portfolio, the lower the spending rate. So annuititization gives the highest payout rate, followed by 100% TIPS, and bringing up the rear is 100% emerging market small cap fund portfolios. This rule protects you somewhat against near term steep portfolio drops by risky portfolios. If the high expected returns or better materialize, even the low spending rate will allow future higher spending from the risky portfolio as the expected larger size of the risky portfolio more than offsets the lower spending rate over time.

If you want a smooth spending path out of your retirement portfolio, pick a low risk portfolio. If you want a high spending path, pick a high risk portfolio, but expect a bumpy ride, and the possibility of an extremely low spending path. What you cannot do is pick a fairly risky portfolio and assume you can find some arbitrary rule that will allow for relative smooth and high spending throughout retirement. No such rule exists.

In any case the portfolio spending rate needs to be re-evaluated every year based on remaining years of life, current portfolio value, and portfolio riskiness. Except, of course, the case where you pick the 100% annuitization strategy.

Bob K
_________________
The real difficulty in changing any enterprise lies not in developing new ideas, but in escaping from the old ones. - JM Keynes
Good questions outrank easy answers. - PA Samuelson

[Cb]

[DRiP Guy]

[bob90245]

[retired at 48]

On another thread, I had suggested an OP consider a 5% withdrawal rate. I was asked by a responder to provide some links to support this with studues. Here are some additional links:


1. Retirement Calculator Smashes
Safe Withdrawal Rate Myths

[i]Excerpt: The retirement calculator reveals the safe withdrawal rate (the annual inflation-adjusted withdrawal that stands a 95 percent chance of success, presuming that stocks perform in the future at least somewhat as they always have in the past) that applies. It also reveals the withdrawal rate that is “Reasonably Safe� (80 percent chance of success), the withdrawal rate that provides for “Likely Success� (50 percent chance of success), the withdrawal rate that provides for “Likely Failure� (20 percent chance of success) and the withdrawal rate that provides for “Almost Certain Failure� (5 percent chance of success).


The retirement calculator’s default results (the results that apply until a user enters choices of his own) show the odds of success for a retirement portfolio in which the TIPS rate is 2.0 percent; the stock allocation is 80 percent… For a retirement beginning at moderate valuations (a P/E10 level of 14), the safe withdrawal rate in these circumstances is 5.4 percent.

http://www.passionsaving.com/r....lator.html


2. The Retire Early study on safe withdrawal rates.

Excerpt: The maximum "100% safe" withdrawal rate decreases as the pay out period increases. Using the PPI to index withdrawals, a 20-Year pay out period allows for a 4.78% first year, inflation adjusted withdrawal…

http://www.retireearlyhomepage.com/restud1.html


3. My latest research lifts the Safe Withdrawal Rate to 6%....

Welcome to Early Retirement Planning Insights. This site contains cutting edge research related to retirement investment planning.

At this site, we find out how different approaches affect safe withdrawal rates. Safe withdrawal rates tell us how much money retirees can withdraw from their investments with a high degree of safely. Safe withdrawal rates describe what is likely, looking forward. Safe withdrawal rates are best described in terms of probabilities and statistics.

These studies are by a well known forum participant, John Walter Russell.

http://www.early-retirement-planning-insights.com/

retired at 48

[bobcat2]

[bobcat2]

I think this idea of the probability of running out of money is itself a silly way to structure the retirement spending problem. Only a fool would continually spend at a high rate when the risky portfolio is performing well below expectations for a period of many years.

A reasonable person will downshift spending and accept a lower consumption path over the rest of life, rather than spend the portfolio down to zero and then risk spending several years living solely off SS.

So the question should be framed as to the probability of falling to an unacceptably low consumption path during retirement – Not, I am a complete idiot! – What is the probability of my exhausting all my portfolio assets on a high rigid discretionary spending path years before I die?

Bob K
_________________
The real difficulty in changing any enterprise lies not in developing new ideas, but in escaping from the old ones. - JM Keynes
Good questions outrank easy answers. - PA Samuelson

[bob90245]

[bobcat2]

If you planned to have a certain consumption path throughout retirement keeping spending relatively constant, why would you then accept keeping that spending path until you run out of money entirely, because of poor returns, and then having your spending fall off a cliff.

Who would say the following. I planned on spending $105,000 per year throughout retirement. But my portfolio performed much more poorly than I expected, so now I am spending $105,000 thru age 83 this year and after this year $18,000 per year, solely because poor portfolio returns will exhaust my portfolio this year.

Or, say the following. I planned on spending $125,000 per year thru age 80, and $90,000 per year after that because I wouldn't be as active as I got older. However, because of poor investment returns I spent $125,000 thru age 80, and $90,000 per year thru age 83, when I exhausted my portfolio due to very poor returns. Now I live on $18,000 per year from SS for the rest of my life.

Bob K
_________________
The real difficulty in changing any enterprise lies not in developing new ideas, but in escaping from the old ones. - JM Keynes
Good questions outrank easy answers. - PA Samuelson

[retired at 48]