FREE NEWSLETTER
BUILDING A NEST EGG is relatively easy: Save as much as you can starting as early as you can. Invest in a diversified mix of low-cost mutual funds. Rebalance periodically. And tune out the noise.
By contrast, determining how much you can safely spend in retirement is far trickier. Consider three strategies.
First, there’s the much-discussed 4% withdrawal rate. In the first year of retirement, you spend 4% of your portfolio’s beginning-of-year value. In subsequent years, you increase that amount by the inflation rate. With that simple strategy, and a mix of 50% stocks and 50% bonds, your retirement savings should last 30 years. The problem is, you could run out of money sooner if the market performs poorly—and, of course, there’s a danger you may live longer than 30 years.
A second, even simpler approach: Spend a fixed percentage of your portfolio’s value every year. Let’s say your portfolio finishes the year at $900,000. You might spend 5% of that sum, or $45,000, the following year. With the percentage-of-portfolio strategy, you can never run out of money, because each year you’re limited to a percentage of whatever remains. But your spending could vary widely from year to year. For example, if the market falls and you finish the year with a portfolio that’s 20% smaller, you’re forced to spend 20% less the following year.
That brings me to a third way of determining how much to spend: the ceiling-and-floor method suggested by Vanguard Group. With this strategy, you again aim to spend a fixed percentage of your nest egg’s year-end value. But unlike with the second strategy, if last year was bad, there’s a limit—or floor—on how much you reduce your spending. Similarly, if the market’s performance last year was stellar, there’s a limit—or ceiling—on how much you increase your spending.
Say you start with $1 million and you’ve chosen a target of 4% spending per year, but the amount your portfolio withdrawals can change each year is limited by a 5% ceiling and a –2.5% floor. In the first year of retirement, you spend your target percentage, which is 4% or $40,000.
During that first year, your portfolio soars 20%, reaching $1.2 million at year-end. But instead of spending 4% of that sum in year two, or $48,000, you increase your spending by just 5% from the first year’s level, to $42,000. Conversely, if the market performed really badly in year one, the most you’d reduce your spending in year two is 2.5%, equal to $39,000.
Got that? To use the ceiling-and-floor approach, you’d compute three numbers at the start of each year:
You then allow yourself to spend whichever of these three numbers falls in the middle. Result: You can spend more when you’re coming off a good year and you spend less if it was a bad year, but your spending doesn’t vary too much because of the limits imposed by the ceiling and floor.
A 2010 Vanguard study showed that, with a 5% ceiling and –2.5% floor, you’re less likely to run out of money than if you adhere strictly to the standard 4% withdrawal rate. In addition, you’re less likely to end up denying yourself and leaving behind an unnecessarily large estate, which is a danger with the percentage-of-portfolio approach. A 2020 Vanguard paper looked at the strategy again, this time using a 5% ceiling and –1.5% floor, while also factoring in annual inflation adjustments.
I like the ceiling-and-floor strategy, in part because my wife and I have no children and thus we don’t feel the need to leave behind a large estate. But in practice, I don’t use any of these approaches. Why not? They’re fairly simplistic and don’t take into account major life changes, such as selling your house or moving into a continuing care retirement community. They also don’t consider tax nuances, such as whether you’re selling investments that will trigger a big tax bill or not.
Instead, before my wife and I retired, I looked at how much we were spending each year and then figured out how that would change once we retired. I use that amount as our spending target. Yes, each year, I also use the Vanguard ceiling-and-floor method to make sure we aren’t too far off base. As long as we’re close to that spending level, I can be reasonably confident that my wife and I won’t run out of money.
Brian White is retired from the University of North Carolina, where he worked as a systems programmer and then director of information technology in the computer science department. He likes hiking with his wife in a nearby forest, dancing to rocking blues music, camping with friends and stamp collecting. He also enjoys doing Volunteer Income Tax Assistance (VITA) work at the Chapel Hill senior center. Check out Brian’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.
Brian, I like your plan. I retired in my middle 50’s, but I currently don’t withdraw any from my retirement funds, plus I’m not 72 yet. However, I am receiving SS and have a pension. These two sources supply my needs, therefore, my withdrawal period will be less than the average 30 years. If only I had a crystal ball to lead me in the right direction.
My method to determine what I can withdraw from all my accounts is divide the total assets (individual accounts, IRAs and Roth IRAs) by what I think my life expectancy will be and recalculate every year. If have have $300,000 and expect to live 30 years, I can take out $10,000. If that $300,000 grows 10% that year, it would be $320,000 the following year and I would expect to live 29 more years. So for that next year, I could take out $11,034. But if I wanted to buy a new car or go on a big vacation that year, I would need to take out an extra $40,000. If the $320,000 didn’t appreciate at all that year but I took out $50,000, I would have about $270,000 and 28 years left to live so I would be able to take out only $9643 that year …but I would have a new car or would have had a great vacation when I was younger and could enjoy it more. These withdrawals would supplement any other regular retirement income I have such as Social Security and pensions.
The method is simple. If you want to take more out when you are younger, you can. If you assets appreciate or depreciate in the short term, you can see it and adjust your withdrawals. And you don’t have to pay a financial advisor to figure it out.
This is so not complicated. First, find out when you are going to die…
If you are just trying to plan your spending for this year, then there is not much effect. However, if you are planning your spending for the rest of your life, there is value in estimating what happens when you do a Roth conversion. The whole purpose of the Roth is to reduce your long term tax liability (or that of your heirs), leaving you more to spend. Therefore, it makes sense to include this in a long term spending plan.
I see what you mean. I do Roth conversions sufficient to get me up to the top of my usual tax bracket, while using cash reserves those years for spending and paying the tax on the conversions. However, if I was doing a Roth conversion big enough to greatly increase my tax liability, I would consider that to be outside the normal spending guidelines, similar to (but less expensive than) buying into a CCRC. As Roboticus Aquarius said above, the 4% is just a guideline, not a hard and fast rule. However, I personally would try to split up my conversions between multiple years so as to not push myself into a higher tax bracket on any of those years.
It seems to me that the best time to do Roth conversions is generally in early years of retirement, before you start taking Social Security and having RMDs, so you can do the conversions with relatively little tax pain. Everyone’s situation is different, though.
My way is the easiest. Live on your RMDs and reinvest the amount you don’t spend!
It helps that I waited until I was 70 to draw social security.
After reading the article and the comments, I am going to make my usual complaint: no one seems to take into account the Required Minimum Distribution. Many of us retirees have IRAs and must follow IRS rules which require us to withdraw higher and higher sums as we get older. I set up an RMD schedule and ran it out 20 years. By year 10, you are already in excess of 5%. This would seem to suggest that you may want to convert amounts to a Roth account, but be careful of IRMAA which could reduce future social security benefits.
RMDs need only be a tax event. Open a taxable brokerage account, and transfer the RMD amount from the IRA/401k into identical mutual funds in the taxable account.
My favorite is the “RMD” approach, because unlike the options above, it takes life expectancy into account. Simply put, your “minimum” spend is what the IRS would require you to withdraw if it were all in a traditional IRA, which can be calculated for any age from 0 up. If you have more than you need saved up for this minimum spend, you can spend the rest with confidence at any point along the way. These figures are in Table 1 “Single life expectancy,” and the percentage to withdraw it caculates starts at 1-2% at age 0 and increases to 100% at age 111–at which point, if you’re still thinking clearly, you might want to abandon this approach. (At age 60, it’s about 4% and at age 70 it’s about 6%.)
Except I would much prefer to have the 6% at age 60 when I can enjoy vs. at 70 when I might be too old, thus the crux
Assuming you do make the withdrawals beginning at age 60, the 4% will be of a bigger number than the 6% because of the ten years of withdrawals, unless your investment return is greater than your withdrawals, in which case you will have more confidence at 70. If you treat this as a minimum spend, you could ask yourself the question, “will I have enough left for my desired minimum spend at 70 if I go ahead and take [2% extra] this year at 60. If you’ve been a saver, your answer will likely be “Yes!” If you’ve always lived out the idea of spending more now instead of saving it for later, you may discover that your future self would kindly request you show a little restraint ;-). BTW, I’m reading a great paper right now another commenter recommended if you want to see this all laid out very technically and thoroughly https://longevity.stanford.edu/wp-content/uploads/2019/07/Viability%20SSiRS%20Final%20SCL.pdf
The 4% guideline is not a rule. It was supposed to be a starting point to the withdrawal discussion, not an end point. Also, it has never failed in all the time since 1927. That’s 63 rolling 30 year periodss of constant success There is nothing wrong with the 4% guideline, as long as you understand what it’s about. That includes the assumption of a 30 year retirement. The percentage is lower if you want to plan for 40 years… more like 3.5% as I remember. We have had low-yield environments before, and any retirees using the 4% guideline have done just fine.
Now, in reality, I’ve never heard of anyone strictly using the 4% guideline. Most retirees adjust their spending somewhat depending on returns. This can end up acting a bit like a floor and ceiling approach. There are other approaches too.
There was also a recent Standford study addressing the withdrawal problem that may prove interesting to many readers. Wade Pfau is well known for thoughtful analysis on this topic:
https://longevity.stanford.edu/wp-content/uploads/2019/07/Viability%20SSiRS%20Final%20SCL.pdf
Great topic. Those who are not fazed by algebra, statistics, and calculus, I have a recommendation for all DIY retirement. I found it to be really helpful especially dealing with the topic above with solid mathematical foundation.
https://www.amazon.com/Most-Important-Equations-Your-Retirement/dp/1118291530/ref=sr_1_1?dchild=1&keywords=seven+equations+retirement&qid=1610426958&sr=8-1
What’s missing here is any account of other sources of retirement income. If you will get Social Security income, you can adjust the required initial amount of your nest egg downward, or you can adjust the 4% estimate of retirement spending upward. An easy way to do the latter is to add 25 times your estimated annual Social Security income to your nest egg, and similarly for any pension income you expect. Then take 4% of that as your estimate of your retirement income.
Another source of retirement income, in my opinion, is imputed rent. If you will have your home paid for by retirement time, you can pretend to pay yourself what your rent would be if you had to pay it to a landlord.
What a brilliant way to do this. I have been ravenously consuming content for two year and found it weird that no-one talks about when social security kicks in and of course a large part of the disposition is not really needed!!!
So if K understand this correctly, and my soc is 4 per month I could over 1Million to to nest egg?
Thanks,
Ray
Not a bad idea, though I do it differently. I have a spreadsheet on which I estimate all my annual income and spending, including taxes. Rather than adding in amounts equivalent to my Social Security and pension income, I just compute how much of my investments I spend each year and compare that to how much of my investments I would spend using the ceiling and floor method.
Take heart. Retire spending is not always an upward sloping line into the future. Some people have theorized that retiree spending is more like a “SMILE”.
https://www.kitces.com/blog/estimating-changes-in-retirement-expenditures-and-the-retirement-spending-smile/