FREE NEWSLETTER
IT’S CHALLENGING TO GO from saving during our working years to spending in retirement. Our solution: Use a modified version of the 4% rule.
Financial planner William Bengen was the first person to articulate the 4% rule. He wanted to know how much people could withdraw from their investments each year and still not run out of money. Through extensive back-testing, he found that if folks withdrew 4% in the first year, and thereafter increased this amount each year for inflation, in almost all cases they wouldn’t run out of money over a 30-year retirement.
With Bengen’s 4% rule, the amount withdrawn is driven only by the initial portfolio value and subsequent inflation. History suggests this approach should be okay despite sequence-of-return risk—the danger that the financial markets perform poorly during the years immediately after a person retires. Still, folks who retire and keep taking out an inflation-adjusted 4% do run some risk of running out of money.
I believe it’s reasonable to increase spending when the markets are doing well, while also cutting back when markets perform poorly. Bengen’s model doesn’t incorporate this. Vanguard Group developed a withdrawal method which does. I have read its guide several times and, I must confess, I still don’t understand it.
What to do? For my wife and me, I had three criteria for our retirement withdrawal strategy. First, it should be simple, something I can use when I’m 90 years old. Second, the approach should be responsive to market returns. Finally, it should be financially conservative, meaning there’s reasonable certainty we won’t run out of money before we die.
A withdrawal each year that’s simply 4% of the prior year-end balance—without the inflation adjustment in Bengen’s approach—meets these three objectives. With such a plan, a person would never run out of money. Each year, you always leave 96% of your portfolio invested. And it’s certainly simple. But a major drawback is that the amount withdrawn each year fluctuates widely because market returns are so erratic.
That’s why I nixed the idea of simply withdrawing 4% of our prior year-end balance. Instead, I’ve made two modifications, which I have found to be extremely helpful.
First, rather than using last year’s Dec. 31 balance, I apply a percentage to a three-year rolling average of year-end balances. It allows for significant spending increases only if there have been sustained market gains, while cutbacks are only required if there’s a prolonged bear market. Second, because I’m financially conservative, instead of 4%, I use 3%.
The upshot: We limit our annual withdrawals to no more than 3% of our average investment balance for the prior three year-ends. I believe this is a simple yet elegant solution. Because of a lifetime of frugal habits, we’re in the fortunate position that we could live on Social Security and a modest pension I receive. We use withdrawals from our nest egg primarily for gifts to individuals, contributions to charitable organizations and to fund a major overseas trip every few years.
Because we’re heavily invested in stocks and limit our withdrawals to 3%, I fully expect our investments to continue to grow. When the time comes, hopefully a decade or more away, we’ll spend what we need to for assisted living or nursing care, even if it exceeds the 3%.
We have just three investment accounts: a traditional IRA, a Roth IRA and a regular taxable investment account. Each January, I enter the year-end amounts for the three accounts in a simple spreadsheet. It totals the three accounts and calculates the three-year rolling average. The calculation could also be done by hand. If I pass away first, my wife can easily take over.
After a lifetime of saving, we initially found it unsettling to make withdrawals from our investments. But with this plan, we sleep well at night, enjoy the fruits of our earlier saving and remain confident we won’t run out of money.
Larry Sayler is the only person with a Wharton MBA who also graduated from Ringling Bros. and Barnum & Bailey’s Clown College. Earlier in his career, he served as CFO for three manufacturing and service organizations. For 16 years before his retirement, Larry taught accounting at a small Christian college in the Midwest. His brother Kenyon also writes for HumbleDollar. Check out Larry’s earlier articles.
Want to receive our weekly newsletter? Sign up now. How about our daily alert about the site's latest posts? Join the list.
It would be interesting to compare your three year average withdrawal to a three year “EMA” withdrawal method. The EMA would weigh your withdrawal more towards recent gains.
I like your method, Larry. Thanks for sharing.
Nice article Larry. Another withdrawal method that I found very interesting was recently posted by Mad Fitentist (The Problem with the 4% Rule (and Why You Could Retire Even Sooner) (madfientist.com). He talks about the problem with the 4% rule and gives, I think, a good solution. He basically bases one’s withdrawal percentage on the person’s discretionary income and the stock market’s performance. His method is very interesting. It may or may not be your choice of withdrawal method, but will at least get you thinking.
I solved the withdrawal volatility by separating required expenses from discretionary expenses. Required expenses stay under 2.5% of the current portfolio balance, but the withdrawal rate is always 4%. After a market crash of 55%, a 67/33 portfolio with a 4% withdrawal rate will still provide a dollar amount greater than 2.5% of the balance prior to the crash.
I also read the Vanguard withdrawal method and found it lacking. The ceiling could be infinite and still not cause a problem, so there’s no point. The floor method is also far less effective than fixed percentage, which they themselves found to be the case. I appreciate Vanguard, but their method seemed like an overcomplicate fixed percentage method. My advice: Use a fixed percentage, have a plan for variability, and just do your best.
This is a great article, and is actually a big “duh” for me. In the past I have been on the endowment committee of 2 different non-profit organizations. One of our tasks was monitoring withdrawal rate with the objective of not depleting the endowment. Current law allows a “total return” distribution, but it is up to the organization as to what base period to use. One of the charities used a rolling 8 quarter-end average, and one used a rolling 12 quarter-end average. That’s not too much harder to calculate than Mr Saylor suggests using year-end results. Looking backwards, investment totals peaked at the end of 2021 and were lower at the end of 2022. No surprise there. But the big gains in 2021 got averaged in with the previous quarters, and the big reductions in 2022 got averaged in with their previous quarters as well. This prevented taking an unusual windfall in the unusually good year, and prevents taking a big hit on the distribution in an unusually bad year, thereby somewhat stabilizing the income to the charity. So my big “duh” is, why didn’t I think of this for my own account???
Nice article Larry.
I wonder if the (admittedly minimal) work to calculate performance over the last three years, or even adjust spending based on last year’s performance, is really necessary. It seems to me that with a sufficient cash reserve, one should be able to continue withdrawing at the initially set rate even following a down year (or three) and not sell depreciated assets to do so.
I agree that withdrawing 3% of the average of the last 3 years assets is straightforward, which is helpful.
However, I wonder if that would be an appropriate withdrawal level for all ages? For example, if you were age 95 with (presumably) relatively few years remaining, withdrawing only 3% might be too low unless your objective was primarily to leave an inheritance.
What I wrote Mr. Clements about in 2016 and I still think is appropriate, albeit slightly more complicated, is to:
I call this approach “RADAR” for Risk Adjusted Drawdown Asset Rate.
Of course, each situation will have complexities such as wanting to take out a higher asset percentage before starting to draw social security, etc. However, I think the basic principle of adjusting your asset withdrawal percentage based on your estimate of how much time you (or you and your spouse) may have remaining is sound.
Does a 95-year old have relatively few years left? In most cases, yes. But there are outliers. My grandmother passed away just a month before reaching 112. At 95, she had nearly 17 years left. Thankfully, she did not run out of money and left an inheritance to my mother, her only child.
I don’t mind being conservative, knowing that I will leave an inheritance.
Fair point Larry Sayler. If your needs are fully met with a 3% withdrawal rate even when you reach an advanced age and are happy leave an inheritance, certainly there’s no reason to withdraw/spend more. My 1/potential years left “RADAR” suggestion was about how much you COULD conservatively withdraw not MUST withdraw.
By the way, your grandmother substantially “beat” the life expectancy game. With those genes, I can understand why you might also be conservative!
Unless our goal is to always ensure a substantial inheritance (even to our 70 year old heirs if we die at 95, for instance,) we may find ourselves questioning our strategy as we age. If we are going to spend or give away money, will we later wish we had done so sooner?
One alternative, which I find attractive, is to use RMD tables to calculate the amounts. Since they are required…minimum…distributions, they are inherently conservative, but change as we age, and never [completely] exhaust our reserves. Interestingly, they start out quite similar to the 3-4% figures you discuss. 4% corresponds to the RMD at about 62.5 y.o. At 72.5 y.o. the RMD is closer to 6%. If we are a 3% kind of person perhaps we could reduce the RMD rate by 1%? And perhaps a spend-it-sooner who still likes guardrails could increase the RMD by 1 or 2% (if such a person exists?)(https://www.federalregister.gov/documents/2020/11/12/2020-24723/updated-life-expectancy-and-distribution-period-tables-used-for-purposes-of-determining-minimum)
Another modification I would consider is that we put our “withdrawals” in a second basket, from which we can draw at will. Seeing this [already released] money either depleted or piling up may give additional guidance to bigger purchases, whether cars, trips, or gifts to others. I know this has an influence on us with our donor-advised fund.
The RMD table method has much to recommend it. Hard to get much simpler. Applying the factor from the tables to the entire portfolio, regardless of what kinds of accounts they are, gives you the total “withdrawal” for the year. Then all that remains is to decide what accounts to take from.
Steve, I really like the suggestion in your last paragraph. Most years we withdraw less than the 3%. Having the additional added to a second “bucket” would make me feel better when we encounter a year where we need to go above the 3%, for example, if we purchase a new car for cash.
that’s a nice sensible method, thank you. I particularly like that it doesn’t depend on fortune-telling and predicting the future..
The bogleheads method gives you solid mathematics and good clear numbers, except for the part where you have to get out your crystal ball and foretell the factor r,
r=expected return of the portfolio.
This instantly catapults us into a fantasy land.. if I could know r, I could make enough money not to need to do any of the withdrawal calculations.. ha.
I like it. 3% feels a bit “belt-and-suspenders,” but I like that too.
3.5% was mentioned; that might be my “Goldilocks” figure – not too hot and not too cold.
(I too scratched my head at that Vanguard method. 🙂 )
Nothing in finance produces endless “How many angels can dance on the head of a pin” articles more than retirement withdrawal rates.
A different mindset is needed to go from a career of accumulation to one of spending; I like your plan – it’s conservative, simple, and the surviving spouse could easily follow it. I’ve been retired 4+ years and have spent modestly from our retirement accounts but know I need to nail down a structured plan. I feel comfortable with your approach and believe it will meet our needs! Thanks for the article.
Does the <= 4% withdrawal rule or other similar withdrawal models make assumptions about social security? Does this withdrawal also make assumptions about taxes and insurance?
My 3% has to cover everything – taxes, insurance, everything. The only exception I can think of is the occasional purchase of a new vehicle. I don’t plan to do this more than once every 7 to 10 years. I would like to keep that within the 3%, but it may not be possible.
It’s very clear that you’ve carefully considered your risk tolerance and your goals in coming up with this withdrawal strategy. You primarily use your portfolio to fund charitable donations and periodic travel, but also worry about long-term care later in life.
Many will rely much more on their portfolios to fund their lifestyles. I also find it reasonable to spend less when markets are down and more when markets are up. I really like how you’ve incorporated portfolio valuations going back the last three years. But for those that need the money to live, I think retirees can go much higher than 3% as their maximum spend rate.
Appreciate the links to the Vanguard study and the former HD article by Brian White. These base an initial level of spending on the portfolio (4% for example) and apply spending increases of no more than 5% or decreases of no more than 2.5% to the previous year’s level, depending on market performance.
There are other spending strategies that allow the spending rate to go higher in response to good markets and lower in response to poor markets, but it seems to get complicated and there’s much greater fluctuation in withdrawals and flexibility required to implement.
I appreciate articles like this that discuss unconventional withdrawal strategies. In the linked article above, Adam references school endowment withdrawal strategies at the end. Yale, for example, aims to spend 5.25% of their endowment each year. How can regular folks like us approximate their approach if it appeals to us? Would love to see continued articles around withdrawal strategies and perhaps some on simplified ways to responsibly implement (if that’s possible) “high” withdrawal rates (~5%) for the more risk-taking among us.
Thanks for the article and the suggestion. As I wrote recently, I have been retired since 2000, and have yet to tap my investment accounts. That will finally change when I move to a CCRC in October, and I have been putting off deciding on my withdrawal strategy. So far I have been simply moving my RMDs from untaxed to taxed, but now I may have to spend them.
Great that you’ve landed on a well thought out process that works for you. I think that’s the key for each individual. I’ve landed on the simple process of spending a maximum of 4% of the previous end of year portfolio value with no more than a 10% reduction year over year. I also maintain a conservative 30-60-10 portfolio allocation. The numbers add up, it keeps my worries to a minimum and I enjoy life – it works for me!
I wonder about modification to your plan when the the first spouse dies. The surviving spouse will see the lower social security benefit end and the options selected on the pension may change the amount downwards or be eliminated based on the options the pension plan had and the elections made.
As informed readers will know the surviving spouse will likely be pushed into a much higher tax bracket beginning in the year after the death of the first spouse when the return filing status changes from joint to single.
My initial planning is also to not touch our Roth accounts during the period we both are alive and convert additional amounts from traditional IRAs to Roth while we are both alive with the goal of smoothing the expected current and future tax obligations.
Our current taxable investment account is a joint account as the balance is modest and the unrealized taxable gains are also modest. I expect my wife will survive me and plan to make ownership changes as our circumstances and tax rules change.
Do you have additional plans at the death of the first to die?
William, my wife is 3 years and 10 months older than me. Statistically, we should die within a month of each other! Of course, that is unlikely to happen.
We are extremely fortunate in that currently our withdrawals are used primarily for discretionary items. My pension stops if I die first. (Yes, my wife voluntarily signed off on that.) The survivor will lose the smaller Social Security check. At that time, the survivor may have to use withdrawals for living.
If a person has had a good marriage, I once read that the greatest gift a spouse can give is to be the surviving spouse. Neither my wife nor I look forward to living life alone. We truly enjoy each other’s companionship and support.
Thanks Larry.
I tried this calculation using year end 20, 21, 22. In my case the numbers weren’t significantly different , (at least in my view ) but I would like to include this method going forward as another check.
Thanks for writing an excellent article.
I’m relatively new to Humble Dollar. Do you and your brother always have your articles published on the same day?
Ken, my brother introduced me to HD. He was submitting articles for more than a year before I got up the nerve to submit my first article. Unbeknownst to us initially, this last time we both submitted articles to Jonathan the same week.
I don’t know if you have submitted anything, but I encourage you to do so. I always enjoy and appreciate the comments. Some people like what I have to say; some people disagree with my thoughts. Both types of comments are useful.
(And if you have a spouse, brother, or sister who might also contribute, even better! In addition to my brother and me, there is also a husband/wife couple who individually post articles.)
Larry, I have some articles in the pipeline. The first is scheduled to be published July 24. Taking a break from writing until I see how the first batch goes over with Humble Dollar’s readership. By the way, I was inspired to read a lot of your old articles today. I found them all interesting and appreciate the insights from your life stories.
No, that resulted from a special request to the editor….
Thank you for a very sensible withdrawal method. We are in the same situation being able to live off SS and a modest pension. For a future article it might be a good idea to provide a tutorial and link to a SS template calculating the rolling average for those of us less savvy.
Patricia,
Social Security is not involved. On a speadsheet (and it could be done on a single sheet of paper) I have the December 31, 2020 balances for our three investment accounts and the total of the three accounts. Next to that, the December 31, 2021 balances and total. Next to that, the December 31,2022 balances and total. To get the three year rolling average, I simply add those three totals and divide by three.
When we get to 2024, I will add a column for the December 31, 2023 year end balances and total. My new three year rolling average will be the sum of the year end totals for 2021, 2022, and 2023, divided by three.
Each year, I add the most recent year end balance, and delete the oldest year end balance. Doing this smooths the ups and downs.
I then multiply the three year average by .03. That is my withdrawal limit for the new year.
Hope this helps.
When I typed SS the second time I meant spreadsheet, sorry for the confusion. Excellent explanation, thanks so much!
My fault. I obviously assumed your SS meant Social Security. 🙂
Larry, how do RMDs fit into your strategy – if applicable? Do you adjust withdrawals from other accounts?
Richard, based on your comment below, it seems you and I think differently about RMDs. I have not started RMDs yet. Like you, I will use QCDs to minimize taxes on RMDs. But I don’t see why an RMD is a withdrawal. Yes, it is money I have to take out. But I don’t have to spend it. When I take out an RMD, I can simply transfer that money to a taxable account. Yes, I do have to pay taxes on the amount. That necessitates a withdrawal from our portfolio.
I hear so many different approaches to this, it’s like there is no right answer and perhaps that’s true. It seems for many the 4% rule has fallen out of favor. I read it should be 3.5% and a few say higher than 4%. One thing though, I think it was designed to work over 30 years so I wonder what happens to folks who retire well before age 65.
As you mentioned, if I was using it, I would not always adjust for inflation, but I would use a steady stream and put unneeded money in a given year into cash funds. I would use that income stream to build cash for special spending.
Like you I live off a pension and our SS so my withdrawals are limited to RMDs. Since I discovered QCDs my giving strategy changed and my RMDs are less painful tax wise.
Interesting and well thought out approach, Larry. I am not in the decumulation phase yet but like to think about what I might do when that time comes. Your plan has much to commend it.
A very good article. I’m no expert, but I think your plan seems reasonable to me. But I suspect over time the most important part of your (or any) withdrawal strategy will be the annual review and adjustment. That part of your plan seems key to dealing with whatever economic challenges fate throws at your withdrawal strategy.
It seems Larry and I are in somewhat unique positions with pensions to live on so less reliance on withdrawals. Where I see some people at risk is setting up a withdrawal strategy based on accumulated assets, but then taking a significant lump sum from those assets for whatever reason not realizing doing so may throw their whole plan off permanently.
I agree. I think that’s particularly a risk for people who aren’t comfortable playing with numbers and running calculations, but rather tend to make financial decisions more by “gut feelings”. A particular amount in their 401K may seem large unless one starts to actually do the math.