“HOW MUCH CAN I withdraw from my portfolio each year?” It’s one of the most common questions that retirees ask.
In the past, I’ve talked about the 4% rule, a popular tool for addressing this question. Among the reasons it’s so popular is its simplicity: In the first year of retirement, a retiree withdraws 4% of his or her portfolio, and then that amount increases each year with inflation. If you have a $1 million portfolio, for example, you can withdraw $40,000 in the first year. It couldn’t be easier.
There is, however, a fair amount of debate about the rule. Some contend that 4% is overly generous, while others argue that it’s unnecessarily stingy. And many question the assumptions used in the original research.
Perhaps the most fundamental criticism is that the 4% rule defies human nature. Suppose you had retired in 2009, in the middle of a recession, when the stock market was depressed. If you’d set your initial withdrawal based on the value of your portfolio at that time, it would have been an artificially low number, even though that’s what the 4% rule would have dictated.
Indeed, in the 14 years since, the stock market, as measured by the S&P 500, has risen from a low of 666 to above 4,000. As a result, many retirees’ portfolios have grown substantially over the past decade. But if you’d been following the 4% rule, your withdrawals would have grown much more slowly—because withdrawals under this rule are permitted to increase only at the rate of inflation and, with the exception of the past few years, inflation has averaged some 2%.
Because of that, many view the 4% rule as more theoretical than useful. That’s why an alternative spending methodology has been gaining in popularity. It’s commonly known as the “guardrails” method.
In simple terms, guardrails are designed to be more responsive to market returns than the 4% rule, which pays no attention to market movements. Using guardrails, retirees can withdraw more from their portfolio in years when the market is strong. But in exchange for that, retirees must accept a spending cut in years when the market is particularly weak.
Because guardrails-driven spending rates are responsive to market conditions, this approach has intuitive appeal. Guardrails-based spending also typically allows for a higher initial withdrawal rate than the 4% rule.
In their research, guardrails creators Jonathan Guyton and William Klinger concluded that much higher withdrawal rates—between 5.2% and 5.6%—would be sustainable over a 40-year retirement. Instead of $40,000, or 4%, on a $1 million portfolio, guardrails would allow for between $52,000 and $56,000. This aspect of guardrails also has intuitive appeal. If retirees are willing to accept a pay cut when the market is down, they should be allowed to start out at a somewhat higher rate.
At first glance, it might seem impractical to ask retirees to cut their spending during market downturns. But proponents make two points. First, these pay cuts aren’t significant. In the standard guardrails formula, withdrawals are cut by just 10% during market downturns. Second, those who use guardrails know how the strategy works—and they know which discretionary expenses they’d trim if it’s necessary.
Another appeal of guardrails: They offer a sort of built-in early warning system. In years when the market is weak, retirees can see how close they’re getting to a potential pay cut. Assuming a $1 million starting portfolio, the lower guardrail, which would necessitate a spending cut, might be set at $800,000. If that were the case, an investor could keep his eye on the market and begin preparing for a cut if he saw his portfolio drop below, say, $900,000.
Despite these benefits, guardrails have their downsides. For starters, unlike the 4% rule, the guardrails approach is complicated. If the 4% rule is like riding a bicycle, guardrails are like a 747. To determine each year’s withdrawal rate, investors must work through a five-part framework, which includes these rules: the portfolio management rule, the inflation rule, the withdrawal rule, the capital preservation rule and the prosperity rule.
To get a sense of the complexity of guardrails, you can try this online calculator. Just set the strategy to “Guyton-Klinger.” As you’ll see, even though the calculator does the hard part, there are still about a dozen inputs.
Another downside: Retirees have to contend with much more unpredictable spending from year to year. As I noted above, guardrails will generally only require a 10% pay cut in years when the market is down. But if the market declines for multiple years in a row, as it did in 2000, 2001 and 2002, guardrails would require multiple pay cuts in a row. It’s for this reason that one critic calls guardrails a “scam” and a “horror show.”
A final criticism—one that applies to both the 4% rule and guardrails—is that they ignore an important reality: Spending isn’t linear. Research has found that most retirees’ spending follows a common pattern, with spending higher during the initial post-retirement years but generally lower later in life, as travel and other activities become harder. That’s one reason it doesn’t make sense to simply extrapolate spending from the first year of retirement. Also, retirees might have one-time expenses—a home in Florida, an RV or maybe a gift to their children—that don’t fit neatly into either of these simple spending formulas.
Where does that leave retirees? If neither the 4% rule nor the guardrails approach provides a complete solution—and the alternative, Monte Carlo analysis, is even worse—how should retirees decide on a spending rate? I have three suggestions.
First, try to build a multi-year model that incorporates both regular spending and the one-time expenses referenced above. You could do this in a spreadsheet, though I recommend financial planning software because it does a better job of estimating taxes.
Second, after building an initial model, explore variations. For example, if it looks feasible to spend $70,000 per year and to buy a $400,000 vacation home, see what it would look like if your spending were instead $100,000 or the home more expensive. This would allow you to see what general range of spending is advisable over time, making it easier to vary it from year to year within that range.
My third suggestion: Don’t accept any of these spending strategies as gospel—but don’t entirely reject them, either. William Sharpe, a recipient of the Nobel Prize in economics, has described the retirement spending puzzle as “the nastiest, hardest problem in finance.” It’s not easy, so it’s wise to attack the problem with as many tools as you can.
Each strategy has some merit. The 4% rule is a great shorthand tool because you can often do the math in your head. Guardrails, on the other hand, may be more complex, but the way it responds to changes in the market makes it more realistic. Many school endowments, it’s worth noting, use a hybrid approach: They withdraw a fixed percentage of their endowments, but that fixed percentage is often set in relation to a three-year moving average of the endowment’s value.
Adam M. Grossman is the founder of Mayport, a fixed-fee wealth management firm. Sign up for Adam’s Daily Ideas email, follow him on Twitter @AdamMGrossman and check out his earlier articles.
Do you enjoy HumbleDollar? Please support our work with a donation. Want to receive daily email alerts about new articles? Click here. How about getting our newsletter? Sign up now.
Coincidentally I just just heard about Tharp and the guardrail approach for the first time a couple days ago, and listened to this Bogleheads podcast about it with him as the guest.
The approach he advocates for this is based on using Monte Carlo simulation as the guardrails.
We have been spending down our nest egg so I can maximize my social security payment until I turn 70. I claim later this year. We are fortunate that my social security, my wife’s current social security and wife’s pension will cover all of our living expenses. Our retirement withdrawals from our self directed IRAs are for discretionary spending. We are healty, active and love to travel. We are motivated to spend now before we age out. I have seen numerous articles describing retirees spending profiles as a “smile”, higher in the beginning, tapering off when the earlier activites become more difficult and then increasing again as age related infirmities start making regular appearances.
My current plan is to withdraw 6% of our portfolio value each year starting next year until the RMD tables tell me I need to increase the percentage. In up years we get a little more. In down years we get a little less. I have allocated 30% our portfolio value into safe investments (CDs, ultra short term bonds, etc.) to defend our portfolio against market declines. That provide 5 years of withdrawals from safe investments giving our equity portion time to crash and recover. 5 years may of may not be enough protection. The equity portion of our portfolio is heavily influenced by Paul Merriman using highly diversified ETFs.
Please poke holes in this strategy. It make sense to me but I am looking for input from people smarter than me.
Maximizing SS is the biggest win in retirement planning, so job done there. In my own testing of percentage of portfolio withdrawals, I came to these general rules of thumb: 1) If your intention is to increase the balance over time by more than inflation, withdrawal 3% per year. 2) If your intention is to increase the balance (and thus withdrawals on average) by inflation, withdrawal 4% per year. 3) If your intention is to maintain the balance (i.e decrease the inflation adjusted value over time) withdrawal 5%.
If you have enough inflation adjusted income to live on, and just want to spend the portfolio down before you’re too old to use it, then I think 6% is probably fine. Just be prepared if in 15 years you’re withdrawal amounts to 6% of zero.
Where did you get the 5 year estimate for withdrawals during a market crash? Converting percentages to dollars, let’s say a $100 portfolio, if you have 30% in cash then that’s $70 stocks, $30 cash. A 50% market crash in your portfolio brings the stocks to $35. If you are taking 6% of the portfolio, the new portfolio value is $65, and a 6% withdrawal amounts to $3.9. With zero return, but a balance reduced by the prior withdrawal, the second year withdrawal would be $3.7, etc. This would actually continue for 10 years, at which point your withdrawal would be $2.1.
I wouldn’t personally spend all cash until gone, I would spend only from cash until my asset allocation returned to 70/30 through regular annual income withdrawals, which in this example would be after 4 years. But as a theoretical exercise, which I think was your intention, the result would be 10 years of cash, not 5 years.
If your actual plan is to continue spending 6% of the pre-crash balance after the crash, thus spending all cash in 5 years, then all of my above comments on withdrawal percentage are not applicable, because you’re not actually doing a percentage of portfolio withdrawal at that point.
Your strategy seems reasonable, except the 6% withdrawal rate. I fear that may be too high.
My favorite rule is the “RMD” approach. Treat your total retirement savings pool like an “IRA” and look up the required minimum withdrawal. This will give you a realistic, age-based, conservative figure to start from.
If you have good reason to believe that your life-expectancy is different than the IRS tables, you can always adjust your real age a bit to compensate–instead of a 66 year old, you might have the life expectancy of a typical 60 year old or a 70 year old.
As an added bonus, if you get to 120 years old, you can take almost everything out in a single year with this strategy ;-).
There’s a significant flaw in your reasoning when you ask “If you have good reason to believe that your life-expectancy is different than the IRS tables.” The tables are based on population averages, and it’s logically incorrect to use a population average to estimate an individual case. Insurance companies can pool individuals and get useful financial information, but that concept does not work in reverse. In order to expect to experience a population average, you would have to live 10,000 lifetimes. You only get to live one lifetime, and the age distribution for life expectancy is wide enough to make the average meaningless for an individual case. 40% of people will have a lifespan far below the average, and 40% will have a lifespan far above the average. Nobody should be estimating their lifespan based on an average that is significantly inaccurate for a given individual 80% of the time.
Ultimately the RMD method is just a version of the percentage of current portfolio balance method, only you haven’t decided on the percentages, instead deferring to an arbitrary table. This means you’ll still end up with the wild swings in income that any percentage of current portfolio balance withdrawal method would have. If you’re already going to need to deal with wild swings of income, why not also decide for yourself what the withdrawal percentage should be, so you can then plan for your split of required and discretionary expenses, as required when income will vary significantly over time.
I don’t see any flaw in the reasoning. If I read it correctly he’s saying that you start with the actuarial life expectancies used by the IRS, and you adjust up or down based on your own estimated planning age.
for instance my actuarial age is 84 now. But I know from my own family’s health and longevity that I’m not being overly optimistic to set my planning age to 95. From there it’s a matter of figuring out what age someone would have to be now to have an actuarial age of 95, and then plug that into the rmd tables to find out the withdrawal amount for this year.
Depending on which study you read, genetics has been found to account for as high as 30% (in older studies) or as low as 7% (in newer studies) of an individual’s longevity. Either way, knowing about your family history is not providing the insight you think it is.
I’m a financial modeler (mortgage backed securities, not retirement income) and decided to put many of the common withdrawal strategies through their paces in excel. In the end I settled on fixed percentage of current portfolio value. It requires planning around somewhat wild swings in income, which I do by separating required from discretionary expenses, with the former and latter each starting out as 2% of a 4% fixed percentage. This forces me to keep fixed expenses low going into retirement, but a paid off house and car can easily accomplish that. Car replacement can eventually be paid for from discretionary income, assuming I can, or even want to drive at that point. I also set aside an amount for ongoing home maintenance as part of required expenses.
The guardrail approach, as stated in the link you provided, can result in its own significant reductions in income. But the additional problem I found with it was the income cuts happen too late, due to the incremental nature of the adjustments in that method. So rather than cutting income the most at the bottom of the market, it cuts income the most 2-3 years later, when the market has often half way recovered. Using a simple fixed percentage of current value means the cuts to income happen immediately, applying the most income reduction when it is the most effective at protecting the portfolio.
Since strong portfolio survival traits require significant income flexibility, I find it best to do so with a plan that is both the most effective and extremely simple. All of the other methods I tested either hide the risk of portfolio depletion, or hide the potential income reduction amount behind a smoke screen of percentages.
Importantly, a withdrawal once or twice over 30 years that needs to be more than the fixed percentage calculation allows for, is automatically fully accounted for in all future withdrawals when using fixed percentage of current value. This is in stark contrast to the 4% rule and all other withdrawal methods I’ve tested, which all use prior withdrawal amounts to dictate future withdrawals, and where miscalculation can compound and accidentally destroy the value of the portfolio.
A number of folks have asked which retirement calculator I use. The one I use most is MoneyGuide, but it’s priced for advisors. In a recent blog post, author and CPA Mike Piper mentioned three other tools he’s aware of. Here’s the link: https://obliviousinvestor.com/social-security-and-safe-spending-rates/
This page on the Bogleheads website includes a number of additional tools, including many that are free. It’s a bit of a morass, but very comprehensive: https://www.bogleheads.org/wiki/Retirement_calculators_and_spending
I appreciate the article and the FI Calc model you referenced. Note the you have to set up withdrawals for the tax payments you will need to generate cash flow. From the Help section…How do I factor in taxes? #FI Calc does not automatically account for taxes. For this reason you must include the money to pay your taxes in your withdrawal amount.
To give a simple example, if you know that you need $40,000 per year to live off of, and you estimate that your tax bill will be $5,000 per year, then you should aim to withdraw $45,000 per year to cover both your expenses and taxes.
Don’t forget about RMDs at whatever age you need to start those.
RMDs are irrelevant to income withdrawal strategy. If an RMD is greater than your intended income withdrawal, simply transfer the excess into equivalent investments in a taxable brokerage account. RMDs define when you pay taxes, not when you spend your investments.
I generally agree with R. Quinn. It is probably best to keep your plan simple and flexible by generating a bit more income on average each year than you need. If you need every cent of 4% of your portfolio from day 1 you just may run into trouble. Under this approach you spend what you want and don’t worry about the day to day unless catastrophe strikes.
Thanks for the thoughtful piece and calculator link! Spending strategies may be the ultimate personal finance Rorschach test.
Thanks Adam. I’ve been looking at some of the guardrail research of late. Michael Kitces and Derek Tharpe (at https://www.kitces.com/?s=guard+rail&submit=&by-author=&by-category=&from-date=&to-date=) have done some interesting articles. But you are correct, they are more complicated than many retirees are comfortable with. Thanks for the calculator link, I was not aware of it.
You mention using financial planning software to help with calculations. Could you name a few specific examples? Thanks!
I break it down to 2 categories. The non-discretionary expenses (utilies, etc.) that must be paid regardless of my portfolio’s performance. No choice, nothing to calculate.
With discretionary expenses I can attempt to maintain a system that’s in line with the market’s performance.
Very interesting Adam. My view is it all needs to be as simple as possible. Average Americans are not doing calculations, especially anything complicated.
Why can’t 4% be the guide and when necessary the 4% can be lowered – spending cut and/or the inflation adjustment be skipped or modified?
How about buying an immediate annuity with a portion of assets to be used only to smooth bumps out and when not needed the payments are placed in a MM account?
I have said many times and been criticized just as many that retirement income should aim to be 100% of former base pay, not 70-80%. That is the case if for no other reason than to smooth out variations in income and spending during retirement.
I have yet to see significant changes in spending during retirement- different spending yes, but in total not significantly different. I still pay property taxes, HOA fees, mortgages were paid before retirement, more for health care related, car expenses are not less, what I previously saved in a 401k goes to travel, donations, gifts, etc.
I just think when it comes to living in retirement we overthink the process as if it is so different than before and, except for the sources of our income, I don’t think it is.
Should planning for retirement include assumptions for spending less, doing less, living a lower lifestyle? For many all that becomes a necessity, but why make that a target?
You say that what you used to put into your 401k now goes to travel and other discretionary spending. Doesn’t that mean that you are spending more on these items than you did when you were working?
Yes, more than when working
That would seem to imply that your 100% of base pay goal assumes that retirees will want to increase their discretionary spending.
Sure does. Isn’t that what retirement is about. Travel, hobbies, visiting children and grandkids, etc.