Wei Sun is an assistant professor at the Hanqing Advanced Institute of Economics at Renmin University in Beijing, China.
Anthony Webb is a research economist at the Center for Retirement Research at Boston College.


Jeff David from IL posted over 2 years ago:

(r / 1 – (1 + r)^-t) × wealth

needs to be

(r / (1 – (1 + r)^-t)) × wealth

/ takes precedence over - and r/1 makes no sense.

Barry Laird from IN posted over 2 years ago:

Jeff is right with his correction

R Townsend from FL posted over 2 years ago:

I am 70 years old as of August 20th. What are the rules for me?

Charles Rotblut from IL posted over 2 years ago:


The IRS says: "You generally have to start taking withdrawals from your IRA or retirement plan account when you reach age 70½. Roth IRAs do not require withdrawals until after the death of the owner."



James Henson from IL posted over 2 years ago:

The RMD calculation would seem to work for me, and leave a little room for fluctuation in the market.

H Stringer from TX posted over 2 years ago:

I start the first RMD for my IRA in 2013 and although they will let me put it off until April of the next year it seems unwise as that forces me to take 2013's and 2014's RMD in the same year. This increase in my taxable income in 2014 I believe would be more costly than wise. I will take 2013's in 2013 and 2014's in 2014 and so on.

As I remove money from my IRA I will deposit it in a taxable account with my broker and continue to manage it and use whatever I require in any year when I require it.

I prefer to be flexible until it becomes highly necessary to use what I have left.

Richard Reem from AK posted over 2 years ago:

I have wondered about the role played by the RMD in retirement planning. Thanks for bringing up that issue.

Edwin Kkorb from IL posted over 2 years ago:

Thanks to H Stringer..I had not considered simply moving to a taxable account the RMD..good for those who don't need to spend it..

Arlyn Stanfield from WA posted over 2 years ago:

Maybe it was mentioned but I never saw it, and my ? is: Can I not have taken out of my IRA self-directed account 1% each quarter at age 70 and deposit it into a regular portfolio which has stocks and bonds etc. if I don't at that time need the cash for living expense?

Frank Boudreault from FL posted over 2 years ago:

I'm in the same boat with H Stringer of Texas.
We also have to be aware of ALL the new taxes taking place in 2013, including the 39.5% rate on Div. income. What a Country! Save all your life to enjoy retirement, and along comes the Gov. to take the Lions share. Should have spent it like the rest of the boobs.

Herbert Arnold from CT posted over 2 years ago:

Have essentially followed the modified RMD procedure since retirement - with any funds not immediately needed re-invested in stocks. Keeps the return higher.

Hampden Smith from VA posted over 2 years ago:

Good article, helpful discussion. Question: Could one simply roll the RMD into a Roth? You're paying the tax on the RMD withdrawal, which is what is required to transfer assets into a Roth. Then your rolled-over assets can increase tax free, solving Stringer's concern.

William Weisberger from NY posted over 2 years ago:

The authors refer to an "optimal drawdown strategy" with an "SEW" of 1.0, but never explain what either is or how to calculate them. Some of us are sufficiently mathematically literate to deal with such concepts. At least a reference to a relevant article would have been useful.

Thomas De temple from CO posted over 2 years ago:

I have used the RMD values along with an estimated inflation rate to determine a desired average portfolio yield such that at the end of some period, say ten RMD years, the remaining portfolio has the same purchasing power as in the start.
Instructive but obviously not very accurate because of forecasting uncertainity.

Kenneth Jeffrey from WA posted over 2 years ago:

Unfortunately, I was also a saver ‘unlike’ the boobs who live their unorganized lives’and those who shame us in D.C. I too would like to know if a RMD deduction after payment of taxes could be directly moved to a Roth.

Jean Henrich from IL posted over 2 years ago:

Thanks, Jeff. The equation has been corrected. - Jean, AAII

Michael Armstrong from FL posted over 2 years ago:

I've been using the RMD method since the first year I could. It requires a degree of conservative spending, which is a good thing, and leaves the maximum residual for a charitable bequest, unless I live too long. I'll worry about that later.

I recall reading somewhere the RMD could not be rolled into a Roth (absent any equivalent earned income), but my recaller isn't working as well as it used to.

Chuck Bloch from WA posted over 2 years ago:

I think you can convert any $$ OVER your RMD to a ROTH, BUT it should be done AFTER you have taken your RMD (--it should be a separate transaction). Both transactions are taxable events. Correct me if I'm wrong-I've done this for several years and have not heard from the IRS.

F Krasowski from CT posted over 2 years ago:

Money withdrawn from a traditional IRA is taxable. Money can be deposited into a Roth IRA, if you have earned income (or your spouse has earned income) up to the amount deposited.

Traditional IRA withdrawals are taxable income unless you have after tax contributions.

You can convert Traditional IRA money to a Roth IRA, and that is a taxable event anytime and it can not be a substitute for a Traditional IRA RMD. There is no income limit for a Tradition to Roth conversion.

C Shah from CA posted over 2 years ago:

Once RMD requirement is made one can opt for Roth conversion. But RMD and Roth are separate transactions. Say one need to withdraw $20,000 for RMD to meet IRS requirement than he/she can withdraw any additional amount for Roth conversion. In both cases he/she pays taxes on combined withdrawal.

C Wilde from VA posted over 2 years ago:

Statistically, and considering improbable (i.e. black swan) events, there is no real difference in any of the likely outcomes, although in the actual event, there will certainly be a difference. Which is entirely unpredictable.

So the real message is save, invest prudently, and pay close attention to changing conditions that may require a change in strategy, investment allocation, etc.

Unfortunately, the general level of financial education/expertise in our country is deficient.

David Dalton from TX posted over 2 years ago:

Where is the RMD "table"?

David Dalton from TX posted over 2 years ago:

Thanks. I found it.

David Frost from MA posted over 2 years ago:

Yes where is the rmd table

Valerie Kempton from AZ posted over 2 years ago:

How is the modified RMD (1.03) more efficient than just taking the RMD (1.39)? Wouldn't you be spending more, consequently running out of money sooner? Maybe I just don't understand the factor.

William Curtis from KY posted over 2 years ago:

I just picked an income level that - assuming 3% inflation rate and 6% average annual return - would provide constant inflation adjusted dollars until my age 90, then drop by 30% and last my spouse until her age 95. I added a 10% buffer for longevity risk. If a really bad year happens (like 2007-8) I cut back on income.
It is simple, and works fine.

Jean Henrich from IL posted over 2 years ago:

David, tables and worksheets for figuring RMDs can be found at the IRS website www.irs.gov. Here's the address to their RMD instruction page:
--Jean at AAII

Donald Myers from AZ posted over 2 years ago:

Part of my retirement benefit is a defined contribution plan (403b and 401a) which are subject to the RMD rules. I have been using just the RMD for more than ten years, no modification. Although there have been flucuations in the balance basically it has remained nearly the same even with the withdrawals. I find that paying attention to the investments which have been mostly buy and hold is more important than fiddling with the withdrawal rate. Of course we do have Social Security and a small defined benefit pension as well as health insurance with the latter pension. Budgeting is very important and also have a paid off mortgage.

Joseph Starnes from CA posted over 2 years ago:

Good article and excellent discussion. Thank you!

James Baker from CT posted over 2 years ago:

1. Is it really true, as has been suggested, that the difference in "SEW's" has no practical significance? If so, the 4% rule would seem to be the simplest, with some provision to cut back in down years.

2. The Roth idea is OK as long as one spouse has earned income (a job) equal to the Roth contribution ($6000 in 2003), and the account has been open for 5 years or more. Doesn't work if both spouses are retired.

3. It seems to be difficult to plan for future withdrawal amounts using the RMD method.

Robert Martin from AZ posted over 2 years ago:

Excellent article, but it raises a question.
The curve in the article gives a withdrawal rate of a little over 5% for an 80-year old.
The Treasury Department's Single Life Table from Treas. Reg. 1.401(a)(9)-9 suggests a withdrawal rate of almost 10% for age 80.
This is a huge difference, and the only apparent reason is that the Single Life table refers to a single person ,whereas the table used in the article may refer to a married person.
Can anyone address this?
Thank you!

Glen Morrill from NV posted over 2 years ago:

The article does not really address the income tax implications of the RMD method. You should prepare a spreadsheet showing the current account balance and an expected rate of return on your investments. Then apply the withdrawal factors from IRS pub 590, appendix C up until 100 years of age. This will give you a look at your taxable income each year. You will see that toward the end of life expectancy the withdrawal amounts become quite large, therefore you will pay higher tax rates in your later years.

Charles Rotblut from IL posted over 2 years ago:

The IRS discusses contributing assets to a Roth IRA from a traditional IRA in Publication 590. Here's a direct link to the page (you will have to cut and past the link into a browser):


Charles Rotblut from IL posted over 2 years ago:

One other helpful link from IRS publication 590:


Nick from fl posted over 2 years ago:

Just found out (via our friends at the IRS) that you cannot combine the year end balances of your 401k and IRA and take the total RMD from only one of them. The calculation has to be made separately for each and the RMD amount taken separately out of each one.

Ben Sabato from OH posted over 2 years ago:

Haven't read every comment, and maybe I'm missing the point, but if you have reached the age of 70-1/2 hasn't the decision automatically been made for you...RMDs? You might need additional withdrawals, but that will be the minimum. I'm assuming this article is aimed at retirees who are not yet 70-1/2, and that the RMD withdrawal method means taking withdrawals at a rate as if you had already reached RMD age. Is this correct?

Gerald Fleischli from OR posted over 2 years ago:

I, too, would be interested in the details of the "optimal drawdown strategy" against which all of these others are compared. Couldn't we just use that?

Hal Smith from CA posted over 2 years ago:

Modified RMD is an interesting concept, but the article leaves open questions.
1-what to do about spouses of different ages
2-is the total of interest,dividends,and
RMD the total spending or is pension and
S.S.added on to it.
3-is this figure pre income tax

I would be keen to see a spreadsheet showing
year by year for a retired couple with interest,dividends,S.S. and pension included,
as well as capitol gains and ending portfolio values for the life span planned for.

Jim Linnemann from MI posted over 2 years ago:

First a link:
Sun and Webb's original paper is at

The most important thing in the paper is the figures which show you the assumed withdrawal with age, and the optimal stock allocation.

Figure 2 shows the RMD method shorts you in your early retirement years but spends a lot in your 90's. Not my plan. Their optimal method is flatter but peaks in your 80's (maybe not too bad if that's peak health expenses or charity). Their spending pattern has to do with their utility function's weighting. Another thing that you learn from the text and Figure 3 is they make strange assumptions about bond returns, essentially no risk as far as I can tell (or that everyone can buy corporate bonds with no change in interest and no default risk and spend them only at maturity), and further use this to argue that the 4% rule "should" hold only bonds, which of course is completely contrary to how the 4% rule was derived in the first place.

Another annoying thing about this approach is that they pick a utility function as a benchmark (with additional discounting of the future), but then define this as what everyone should do ("optimal"), despite empirical evidence that spending patterns don't follow it and the lack of a comprehensible description of why people should follow their rule. Finally is the strange injunction that it is wrong of retirees not to raise their spending if the market goes in their favor. Prudence anyone?

There are other interesting studies at Morningstar, (search Morningstar Optimal Withdrawal Strategy) where the reference point is knowing what you can't know about the future, and the practical alternatives are compared by again using a utility function, but at least considering probability of draining the portfolio or outliving an assumed time horizon, and using more realistic equity allocations. And I can recommend reading the pages under Bogleheads Retirement Spending Models.

Personally, I'm partial to Floor and Ceiling models, where you start out at a withdrawal rate but if the market goes against you, you agree to tighten your belt to some lower level--i.e. down to essentials--until the market recovers. This sounds a great deal like the practical comments from others on this thread. It respects being the need to be able to plan by choosing a floor level at your essential expenses, and allows for more spending if things are going better. Bengen's studies suggested that with such a plan your target withdrawal could even be well above 4% of current portfolio--as long as you actually do dial back your withdrawals by up to 10% (for example) below your target when your portfolio shrinks. Vanguard also has a ceiling and floor study for retirement spending strategy.

Ronald Ferrill from SC posted over 2 years ago:

Mr. Linnemann - Thanks for your post. The addtional information will be helpful.

I am considering starting to pull money out of IRA's early because I was fired at 65 (yes, it was called a lay-off, but I know that was only to allow me to pull unemployment, since it was immediate, and I and the CFO had come to loggerheads with respect to his business plan that contained ridiculous revenue and some incorrect numbers - I refused to agree to it, and CEO and I had discussion - he wanted staff consensus, which I could not agree to), and we have significant medical bills.

I could just pull from savings, but it seems more tax efficient to start taking $$ from IRA while at a lower tax rate than potential future rates.

I would stuff the excess into investments or assets with potential for increase in value: McLaren MP12-C, 1941 Cadillac Convertible, Cabin in the woods, farm land, etc.; or maybe just give it away or spend it on living well.

Any thoughts on how to "Tax Manage" withdrawals? Younger than 70.5 yrs.

Robert Martin from AZ posted over 2 years ago:

Response to Nick
RMDs apply only to IRAs and 401(k)s.Retirement assets such as Roth IRAs,plus any other asset held for retirement (real estate, stocks , bonds, collectibles) are not subject to RMDs unless they are held in a tax-deferred retirement account.They are, however, available for drawdown.

Ararat Pogosian from CA posted over 2 years ago:

I agree with Hampden Smith from Virginia that some of us need to understand the basis and assumptions for calculating SEW and the optimum SEW in order to appreciate the results.

Avi Oren from NC posted over 2 years ago:

I happened to run into a criticism of this CRR report by a few very reliable CFPs:

The CRR study has been neglecting the research done over the last 15 years including material by such Financial Planners as Guyton, Klinger, Blanchett, Bengen, and others, where you increase spending (= withdrawals) by using guardrails approaches, onging mortality adjustments, updated Monte Carlo success rates, etc. If the CRR study would have used some of these more updated methodologies, then its conclusion would have been entirely different.

p.s. I think that at least one of the above comments mentioned this.

Dave Gilmer from WA posted over 2 years ago:

My understanding of the modern 4% rule is that you take 4% of your existing balance the first year and then increase the amount you withdrew the first year by the normal inflation factor. I do not believe the author did this, so the comparison to the normal rule is not very accurate.

If the author had taken 4% and inflated that amount by an inflation factor of 3.5% each year, you would find it almost tracks identical to the RMD withdrawals. At age 85 the 4% rule would suggest a 6.5% withdraw and the RMD would suggest 6.8%!


Kenneth Espanola from RI posted about 1 year ago:

I understand the calculations. My question deals with the optimal draw down strategy. If we are to take just the interest and dividends from the portfolio and leave the capital gains intact, what would the asset allocation look like? How would it change over time for a couple (both age 65)?

William Gaul from IL posted about 1 year ago:

Nick from Florida said you cannot combine the year end balances of your IRA's and take your total RMD from only one of them. I believe this is incorrect. I have been doing exactly this for the past 10 years. Could an IRS agent or some other expert comment on this?

William Gaul from IL posted about 1 year ago:

Nick from Florida said you cannot combine the year end balances of your IRA's and take your total RMD from only one of them. I believe this is incorrect. I have been doing exactly this for the past 10 years. Could an IRS agent or some other expert comment on this?

William Gaul from IL posted about 1 year ago:

Nick from Florida said you cannot combine the year end balances of your IRA's and take your total RMD from only one of them. I believe this is incorrect. I have been doing exactly this for the past 10 years. Could an IRS agent or some other expert comment on this?

John Curran from IL posted about 1 year ago:

I have taken my RMD for the past 8 years from several of my 403B accounts, with my scrupulous tax accountant's blessing. I believe Nick has the wrong information. #2. After taking your RMD why is it possible to transfer some of your available IRA money over to a Roth if you have no income?

William Brewer from CA posted about 1 year ago:

William from Illinois. Go back and read Nick's post again. You are misstating what it said. Nick said you cannot combine your 401(k) balance with your IRA balance and take a single distribution from one of them.

R Curry from CA posted about 1 year ago:

I wouldn't trust any method unless it had been evaluated in a Monte Carlo simulation, or unless it was very conservative (e.g., 2% of initial wealth/year)

John Flynn from FL posted about 1 year ago:

To my knowledge RMD is the only game you can play once you reach 70 1/2. The time to start converting some IRA funds to a Roth IRA is before reaching 70 in order to take it out of the reach of RMD rules. Once you have monies in a Roth you can use 4% Rule or any formula you like on the Roth Funds and still protect the balance from current taxation.

Cal M from CA posted about 1 year ago:

Is it correct that RMD only requires that you withdraw the monies from the tax-deferred account not to spend that amount? If that's the case, some portion of the RMD could be invested.

Cal M from CA posted about 1 year ago:

Is it correct that RMD only requires that you withdraw the monies from the tax-deferred account not to spend that amount? If that's the case, some portion of the RMD could be invested.

R. Olson from HI posted about 1 year ago:

Yes, you can take your RMD, pay the taxes on it, then reinvest the remainder in a taxable account.
My question is, if you take some amount in excess of your RMD (calculated to be just under that which would bump you into the next higher tax bracket) and "roll it over" into a Roth IRA, are the earnings tax free after 5 years?

If so, wouldn't that essentially be a tax-free investment?

Chuck from Illinois posted about 1 year ago:

Closing in on this decision. My current thoughts are to take a % of the year-end portfolio (say 5% or maybe even slightly higher). I know this sounds too high, but if the market drops so does my withdrawal. In theory I can never run out of money, although I run the risk of declining (in both real and actual terms. This just requires some spending discipline.


Sorry, you cannot add comments while on a mobile device or while printing.