Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
Support MFO
Donate through PayPal
The Magic Withdrawal Number In A Low-Interest-Rate Retirement, You'll Be Surprised: 2.8%
Not yet being retired (and it may be a long time at this rate of withdrawal) I haven't checked, but is the 2.8% recommended rate below the required minimum IRA/401K/403B withdrawal from the IRS? Could get interesting if one has to set up a second taxed investment portfolio to survive retirement and meet the drawdown requirement.
Reply to @STB65: but is the 2.8% recommended rate below the required minimum IRA/401K/403B withdrawal from the IRS
Here is the IRA table at the IRS. The short answer to your question is yes. But the purpose of the RMD is to pay the unpaid taxes so I don't think preservation of principle is a consideration to the IRS.
The first thing that comes to mind is, if 2.8% is the right number now and 4% was the right number 5 years ago, who's to say another study 5 years from now won't say that the safe withdrawal rate is 6%. It's certainly guess work trying to figure out what interest rates will be throughout retirement.
I guess the only safe way to withdraw is to wait until you have enough to start on the low end, 2-3% maybe, and be willing and able to adjust with economic times. I'm still working so we don't have to dig into savings yet, but that is my goal.
I submitted the following post to the MFO discussion group in late January recommending Monte Carlo simulations as a means to scoping the huge uncertainties with the retirement decision and an acceptable drawdown rate that generates a high likelihood of portfolio survival for the entire projected retirement period.
I believe it is applicable to the current discussion so I repost it here:
Retirement decisions are surely among the most difficult and stressful that anyone but everyone must eventually address. And the most successful retirements will be achieved the sooner that that issue is confronted; the earlier the better.
Up until just 20 years ago, that vexing problem could only be insufficiently analyzed using deterministic methods that completely failed to capture the uncertainty of future portfolio returns and its volatility.
Recently, these forecasting deficiencies could be mostly relaxed using Monte Carlo simulation calculators. I say relaxed because the uncertainties of the future marketplace can never be removed. However, Monte Carlo techniques can parametrically explore the impact of these uncertainties on portfolio survival likelihoods given various withdrawal rates. With super speed, these Monte Carlo simulations will estimate retirement wealth survival for whatever scenario the simulation user elects to study. Portfolio survival probabilities are the final output.
These calculations are completed with lightening speed on numerous websites these days. Here are Links to two such sites:
The first Link is to the MoneyChimp site. It is simplicity itself and allows for both a pre-retirement portfolio estimated return and volatility and a post-retirement projected return and volatility. It is extremely fast as it performs 1000 random simulations. Since the controlling returns are randomly selected within the machine, identically repeating the same calculation will generate a slightly different probability outcome.
The second Link is to the Flexible Retirement Planner site. It is slightly more complex than the MoneyChimp site, but it is also more detailed in its analyses. For example, it divides your portfolio into taxable, deferred tax, and tax free components. Also it permits several spending options during the drawdown phase of the retirement.
Both sites provide excellent Monte Carlo simulators. I recommend you try both of them. Do numerous “what if” scenarios to test the survivability robustness of your approaching retirement.
I am certain that these Monte Carlo analyses will better inform your savings, your retirement date, and your portfolio withdrawal rate decisions.
I resorted to Monte Carlo computations when making my retirement decisions, and believe they emboldened me and gave me some needed comfort. By way of full disclosure, I did not use the two simulators that I referenced; they were not available at my critical moment.
When dealing with uncertainty and not deterministic events, Monte Carlo methods are the proper tools to apply.
Good luck and Best Wishes.
Many other Monte Carlo simulation sites are also available on the Internet
I like and occasionally still use the Financial Engines product developed by Noble Laureate Bill Sharpe. It is mathematically a rigorous and superior Monte Carlo tool. Bill Sharpe has dedicated his work to improve the lot of us small independent investors.
One issue that I currently have with Financial Engines is that I access it through my Vanguard affiliation so when it recommends some portfolio modifications, that sponsored site tends to follow the Vanguard tradition of fewer holdings and Index-like elements. The financial linkage with Vanguard provides a potential incentive for biased recommendations and makes their proposed changes somewhat suspect. That’s just me speculating aloud without any substantial evidence of any bias.
You can gain access to the Financial Engines website by visiting Terry Savage at the following address:
By clicking on the Financial Engines box you can get a free one year subscription to that planning tool.
These Monte Carlo tools surely do not remove the uncertainty of future market returns, but at least you enter the arena better armed to deal with those uncertainties. What realistic returns you guesstimate from the marketplace will forever be a potent driver to any decision.
We run our cash flow projections to age 100, and we have found that 4-5% has been a consistently good number. Our clients who have been with us for 10-15-20 years can look at their individual numbers and see that as realistic. We use 3% inflation factor, higher than now, but in line with historic numbers. Our experience is that in lean years, clients tend to reduce their cash flow needs, but seldom increase their net, after-inflation adjusted numbers. Of course, the actual rate of withdrawal is often less of an issue the more dollars there are.
Indeed it is different strokes for different folks. But even admitting that each retiree must seek his own unique comfort zone, a few general policy rules are suitable for almost all retirees.
I have done a zillion Monte Carlo simulations using a host of Monte Carlo codes including one that I developed for my own early application needs. Results do vary a little, but some general rules of engagement can be extracted from the myriad solutions.
The most obvious generic finding is that time dominates all other considerations, both in the accumulation phase and in the projected distribution phase. Almost everyone recognizes this non-controversial factor.
A very critical factor, especially during a portfolio’s drawdown phase, is returns volatility, its standard deviation. A few misguided MFO members insist that portfolio volatility is meaningless. That position is simply plain wrong.
What is critical for portfolio survivability is compound return. A rigorous equation links compound return to both annual average return and its standard deviation. Given an annual return, a volatility increase works to reduce compound return. During its drawdown phase, high volatility operates to enhance portfolio failure (bankruptcy) rates.
That’s why many financial advisors recommend a larger commitment to bond-like products within retirement; it operates to reduce overall portfolio standard deviation.
The MRD issue is really a non-issue; it is mandated by law. Penalties if violated are just too damaging to accept. Besides it doesn’t usually impact a drawdown plan if that plan follows a 4 to 5 % withdrawal rate. The MRD schedule doesn’t demand that level of drawdown until about age 78 or 79. So MRD requirements usually don’t impact investment decisions until that advanced age.
BobC is perfectly on-target when he generalizes that Monte Carlo retirement applications typically yield an allowable 4 to 5 % drawdown schedule during retirement. It was interesting that most of his clients adjusted cash flow needs downward during lean portfolio return years. His clients show considerable wisdom.
Monte Carlo simulations can be used to demonstrate that wisdom. Portfolio survival rates are dramatically improved if some flexibility in portfolio drawdown rate is practiced as a matter of policy.
For example, most Monte Carlo codes automatically increase withdrawals as a function of COLA changes. However, if a retiree is sufficiently disciplined and motivated, he can elect to sacrifice any inflation increase when his portfolio suffers a negative year. That simple tactic works wonders to enhance portfolio survival statistics. A host of other strategies can be implemented if the portfolio is exposed to a series of bad years.
Certainly, portfolio survival becomes a critical issue when current and projected future market rewards approach the planned withdrawal rates. Major surgery is required given that dire forecast. You can identify the critical tipping points for your specific portfolio by doing “what-if” scenarios using the Monte Carlo tool.
Comments
Here is the IRA table at the IRS. The short answer to your question is yes. But the purpose of the RMD is to pay the unpaid taxes so I don't think preservation of principle is a consideration to the IRS.
http://www.irs.gov/pub/irs-tege/uniform_rmd_wksht.pdf
I guess the only safe way to withdraw is to wait until you have enough to start on the low end, 2-3% maybe, and be willing and able to adjust with economic times. I'm still working so we don't have to dig into savings yet, but that is my goal.
Anyone else have thoughts on withdrawal rates?
I submitted the following post to the MFO discussion group in late January recommending Monte Carlo simulations as a means to scoping the huge uncertainties with the retirement decision and an acceptable drawdown rate that generates a high likelihood of portfolio survival for the entire projected retirement period.
I believe it is applicable to the current discussion so I repost it here:
Retirement decisions are surely among the most difficult and stressful that anyone but everyone must eventually address. And the most successful retirements will be achieved the sooner that that issue is confronted; the earlier the better.
Up until just 20 years ago, that vexing problem could only be insufficiently analyzed using deterministic methods that completely failed to capture the uncertainty of future portfolio returns and its volatility.
Recently, these forecasting deficiencies could be mostly relaxed using Monte Carlo simulation calculators. I say relaxed because the uncertainties of the future marketplace can never be removed. However, Monte Carlo techniques can parametrically explore the impact of these uncertainties on portfolio survival likelihoods given various withdrawal rates. With super speed, these Monte Carlo simulations will estimate retirement wealth survival for whatever scenario the simulation user elects to study. Portfolio survival probabilities are the final output.
These calculations are completed with lightening speed on numerous websites these days. Here are Links to two such sites:
http://www.moneychimp.com/articles/volatility/montecarlo.htm
http://www.flexibleretirementplanner.com/wp/
The first Link is to the MoneyChimp site. It is simplicity itself and allows for both a pre-retirement portfolio estimated return and volatility and a post-retirement projected return and volatility. It is extremely fast as it performs 1000 random simulations. Since the controlling returns are randomly selected within the machine, identically repeating the same calculation will generate a slightly different probability outcome.
The second Link is to the Flexible Retirement Planner site. It is slightly more complex than the MoneyChimp site, but it is also more detailed in its analyses. For example, it divides your portfolio into taxable, deferred tax, and tax free components. Also it permits several spending options during the drawdown phase of the retirement.
Both sites provide excellent Monte Carlo simulators. I recommend you try both of them. Do numerous “what if” scenarios to test the survivability robustness of your approaching retirement.
I am certain that these Monte Carlo analyses will better inform your savings, your retirement date, and your portfolio withdrawal rate decisions.
I resorted to Monte Carlo computations when making my retirement decisions, and believe they emboldened me and gave me some needed comfort. By way of full disclosure, I did not use the two simulators that I referenced; they were not available at my critical moment.
When dealing with uncertainty and not deterministic events, Monte Carlo methods are the proper tools to apply.
Good luck and Best Wishes.
Many other Monte Carlo simulation sites are also available on the Internet
I like and occasionally still use the Financial Engines product developed by Noble Laureate Bill Sharpe. It is mathematically a rigorous and superior Monte Carlo tool. Bill Sharpe has dedicated his work to improve the lot of us small independent investors.
One issue that I currently have with Financial Engines is that I access it through my Vanguard affiliation so when it recommends some portfolio modifications, that sponsored site tends to follow the Vanguard tradition of fewer holdings and Index-like elements. The financial linkage with Vanguard provides a potential incentive for biased recommendations and makes their proposed changes somewhat suspect. That’s just me speculating aloud without any substantial evidence of any bias.
You can gain access to the Financial Engines website by visiting Terry Savage at the following address:
http://terrysavage.com/
By clicking on the Financial Engines box you can get a free one year subscription to that planning tool.
These Monte Carlo tools surely do not remove the uncertainty of future market returns, but at least you enter the arena better armed to deal with those uncertainties. What realistic returns you guesstimate from the marketplace will forever be a potent driver to any decision.
I hope this helps.
Hi Ron,
Indeed it is different strokes for different folks. But even admitting that each retiree must seek his own unique comfort zone, a few general policy rules are suitable for almost all retirees.
I have done a zillion Monte Carlo simulations using a host of Monte Carlo codes including one that I developed for my own early application needs. Results do vary a little, but some general rules of engagement can be extracted from the myriad solutions.
The most obvious generic finding is that time dominates all other considerations, both in the accumulation phase and in the projected distribution phase. Almost everyone recognizes this non-controversial factor.
A very critical factor, especially during a portfolio’s drawdown phase, is returns volatility, its standard deviation. A few misguided MFO members insist that portfolio volatility is meaningless. That position is simply plain wrong.
What is critical for portfolio survivability is compound return. A rigorous equation links compound return to both annual average return and its standard deviation. Given an annual return, a volatility increase works to reduce compound return. During its drawdown phase, high volatility operates to enhance portfolio failure (bankruptcy) rates.
That’s why many financial advisors recommend a larger commitment to bond-like products within retirement; it operates to reduce overall portfolio standard deviation.
The MRD issue is really a non-issue; it is mandated by law. Penalties if violated are just too damaging to accept. Besides it doesn’t usually impact a drawdown plan if that plan follows a 4 to 5 % withdrawal rate. The MRD schedule doesn’t demand that level of drawdown until about age 78 or 79. So MRD requirements usually don’t impact investment decisions until that advanced age.
BobC is perfectly on-target when he generalizes that Monte Carlo retirement applications typically yield an allowable 4 to 5 % drawdown schedule during retirement. It was interesting that most of his clients adjusted cash flow needs downward during lean portfolio return years. His clients show considerable wisdom.
Monte Carlo simulations can be used to demonstrate that wisdom. Portfolio survival rates are dramatically improved if some flexibility in portfolio drawdown rate is practiced as a matter of policy.
For example, most Monte Carlo codes automatically increase withdrawals as a function of COLA changes. However, if a retiree is sufficiently disciplined and motivated, he can elect to sacrifice any inflation increase when his portfolio suffers a negative year. That simple tactic works wonders to enhance portfolio survival statistics. A host of other strategies can be implemented if the portfolio is exposed to a series of bad years.
Certainly, portfolio survival becomes a critical issue when current and projected future market rewards approach the planned withdrawal rates. Major surgery is required given that dire forecast. You can identify the critical tipping points for your specific portfolio by doing “what-if” scenarios using the Monte Carlo tool.
I wish you well in your retirement.