If you’ve read much about retirement planning, you’ve probably encountered the “4% rule,” which suggests that investors withdraw no more than 4% of their portfolio per year during the early stages of retirement.
The 4% rule is the result of various studies (most famously this one), which showed that using a withdrawal rate much higher than 4% leads to an undesirably-high probability of running out of money over a 30-year retirement — the underlying assumption being that 30 years is a conservative (on the high end) estimate of how long retirement might last.
But just how conservative is that 30-year assumption? For a 65-year-old or a 65-year-old couple, it turns out it’s quite conservative. According to a handy calculator from Vanguard (which uses data from the Society of Actuaries):
- A male age 65 has just shy of a 6% chance of living another 30 years,
- A female age 65 has a 13% chance of living another 30 years, and
- A male/female couple age 65 has an 18% chance that one of the spouses will live another 30 years.
So, for example, for a spending strategy that has a 20% historical failure rate over 30-year periods, the historical chance that an actual 65-year-old married couple would have run out of money during retirement is actually significantly below 20%. To run out of money, they’d need to have a poor-returns outcome and a living-longer-than-average outcome.
Incorporating Life Expectancy Data
Fellow blogger Wade Pfau recently ran the numbers and put together some handy charts to show the historical probability of running out of money (given various asset allocations and withdrawal rates) for a 65 year-old couple using actual life expectancies rather than an assumption that they’ll automatically live for 30 more years.
Of course, as with any study based on historical data, it’s only useful as a predictor to the extent that future returns look like past returns. And many experts believe that stock returns going forward are unlikely to be as high as those of the previous century.
Still, I find the results at least a little encouraging.
For example, according to Pfau’s research, a married couple both age 65 using a 4% withdrawal rate has a less than 10% historical probability of running out of money at most asset allocations. (Interestingly, at a 4% withdrawal rate, the failure rate is barely affected by asset allocation. Portfolios of 30% stocks, 70% stocks, and everywhere in between have nearly identical failure rates.)
Even if you bump the withdrawal rate up to 5%, the probability of running out of money doesn’t exceed 20% for most moderate asset allocations.
Admittedly, the idea that we might not live as long as we’d planned isn’t exactly an overwhelmingly positive message. But at least we might not have to be as stingy with our money as we’re often told to be.
Hi. I'm Mike Piper, the author of this blog. I'm 
Great post, Mike. Wade does raise a good question: what is an acceptable probability of failure? Personally, I think 20% is a little high—I would be more comfortable with 10% or so.
The asset allocation idea is also really interesting: stocks are more risky in the short term, but less risky in the long term, so it seems that in most scenarios these forces more or less balance.
CCP,
Yes, this sort of analysis allows one to take a different type of look at their risk tolerance. We’re no longer talking about tolerance for (temporary?) declines in portfolio value. Now we’re talking about the probability you’ll actually run out of money.
Naturally, the tolerance for this sort of thing varies from person to person–those with more flexibility to cut expenses or go back to work can probably take on a higher probability of failure because they can head-off any failure-looking scenarios before it becomes a problem.
And yeah, I thought that observation about asset allocation was very interesting. 🙂
I’m still new at this, but it seems to me that the 4% rule is woefully misguided. Retired people shouldn’t be their own actuaries, and being “unlucky” in either respect– either outliving the money or having way too much money when you die– is a really bad outcome. If they simply convert their retirement to an inflation-adjusted SPIA or a similar vehicle, then they can just spend the money that comes in and never worry about it running out. To the extent they want to leave money behind for charity or their heirs or whatever, they can just not annuitize that portion and put it in a separate account. I’m hard-pressed to see the advantage in “guesstimating” 4-5% over a guaranteed payout.
Wade continues to provide interesting analyses that I will find quite helpful when I retire in 2-4 years. I think it would be great if he could develop some kind of interactive tool where one could play various what-if scenarios for the person facing retirement. I find the whole asset allocation thing most interesting, as the received wisdom always stresses allocation as an essential component of one’s planning.
@Lance: Mike has for some time been advocating the use of SPIAs for at least part of the retirement income stream. But probably not for all, as leaving some portion un-annuitized provides for greater flexibility if additional funds are needed.
Lance: Your thoughts are very much in line with my own view (and my plan for our own personal retirement).
Still, for various reasons, many people have a strong aversion to the idea of annuitizing. Sometimes it’s the result of bad past experiences with insurance companies. Sometimes it’s a fear of putting a large amount of money into something that’s irrevocable and illiquid. Sometimes it’s an inability to tolerate the idea of betting wrong and dying earlier than average, thereby giving up what would have been a large inheritance for their heirs.
Just one question related to the typo in your post… referring to the portfolio which you mention… is that 70% stocks and 30% bonds or is it 70% bonds and 30% stocks. Thanks, Bill
Hi Bill.
It wasn’t a typo–just poor wording apparently.
I had meant that, at a 4% withdrawal rate for a couple age 65, portfolios of 70% stocks (and 30% bonds), portfolios of 30% stocks (and 70% bonds), and everywhere in between all have remarkably similar failure rates.
Hi, Mike… Oh yes, I see that now. Perhaps it wasn’t poor writing, but instead poor reading. Got it. Thanks, Bill
“To run out of money, they’d need to have a poor-returns outcome and a living-longer-than-average outcome.”
…and not do anything about it along the way.
Because we can do something about it, we don’t need a 90% or 100% probability of not going broke (over planning). An 80% chance of not-running out of money is really a 20% chance that we’ll have to do something so we don’t (under planning). Since we can’t nail it. We need to strike a reasonable balance between the likelihoods of over or under doing it. 80% or 4:1 tends to be very manageable for most people and a good balance between comfort and sacrifice.
Planning around longevity is a little trickier, but I think it’s important to weigh consequences against probabilities and not just look at odds. Would you play Russian roulette? Only a 1 in 6 chance of death? What if the gun was a 12 shooter? Less than a 10% chance you’ll blow your brains out? Low probability but huge consequences!
Who’s going to pay your bills if you end up outliving 4 out of every 5 of your peers but planned for less?
Here in Australia, we mandate the minimum amount that you must withdraw from y0ur retirement savings once you retire. This only applies to tax concessional retirement savings. The minimum amount for a retiree aged 55 is, you guessed it, 4%.
John… Here in the US, we also have a minimum withdrawal rule. I believe that it takes effect at age 72. Bill