When Is A Safe Withdrawal Rate Too Safe?

Like it? Share it!Share on FacebookTweet about this on TwitterEmail this to someoneShare on LinkedInPin on Pinterest

Today’s post comes from my friend DJ, who blogs over at MyMoneyDesign.  He was kind enough to fill in for me during my vacation last week.   We spent our vacation working on the relocation move from Good To Great, and I’m working on a post about the Lessons Learned (“7 Days To A Great Retirement”), be on the lookout for it!

So…today’s your lucky day!  We’re getting a treat with a blogger of DJ’s caliber guest posting on The Retirement Manifesto!  DJ is a great personal finance blogger, and I think you’ll enjoy his take on “Safe” Withdrawal Rates.  I love differing points of view and encourage you to think for yourself about critical retirement decisions such as what withdrawal rate you’re comfortable using for your retirement.

(For the record, The Retirement Manifesto is using a 3.5% max withdrawal rate for our retirement plan, with a Bengen “Floor” and “Ceiling” range as mentioned by DJ in his article below).

With that, Take It Away, DJ!

When Is Your Safe Withdrawal Rate Too Safe?

One of the reasons I think so many people become frustrated with saving for retirement is because they think they should save up a lot more than they really need.

I tend to blame major media outlets for this misunderstanding.  It’s not uncommon to hop on your favorite news site and see headlines that are trying to convince the average person that they will need a nest egg of at least $2 million dollars or more before they can even think about retirement.

Do You Really Need $2M To Retire? Why such a high number? Click To Tweet

More often than not, the root of these out-of-reach retirement savings targets goes back to using a safe withdrawal rate that is WAY TOO conservative.  3 percent … sometimes 2 percent …

The truth: Rates that low are totally unnecessary.  You can do better … and for a lot less!

The simple math behind retirement planning is that your target nest egg amount is calculated by taking your desired retirement income and dividing it the withdrawal rate.   This means that your withdrawal rate and savings goal have an inverse relationship.  When one goes up, the other goes down.  That’s information that we can totally use to our advantage!

While you’ll definitely want to pick a withdrawal rate that is reasonably “safe”, we also don’t want to over-do-it by choosing one that is unnecessarily low.  If we do, we’ll end up paying for it in “sweat equity”!

I strongly believe it pays (… literally …) to know the facts behind the safe withdrawal rate that you plan to use for your retirement plan.  So with that, here’s a few things I think will help you to better optimize your savings goals.

The 4 Percent Rule Was a “Floor”

Probably the most popular safe withdrawal rate out there is the “4 Percent Rule”.

For anyone who isn’t familiar with this, the 4 Percent Rule is a study that says you can safely withdraw 4 percent of your initial retirement nest egg value every year (with inflation adjustment) for at least 33 years.  It was first suggested in an article from 1994 by Bill Bengen using real market data from the past 68 years.  It was later validated in 1998 by a similar article called The Trinity Study.

One of the big misunderstandings about the 4 Percent Rule is that it is not an average but rather a “floor”.  This means that out of all the historical market data that was analyzed, Bengen  found that 4 percent was the minimum rate that worked every time for at least 33 years; no matter what year you retired.  In fact, if you take a closer look at Bengen’s data, you’ll see that in many instances a withdrawal rate of 4 percent actually worked for 50 or more years!

In other words: The 4 Percent Rule was in fact very safe.


Skepticism of Safe Withdrawal Rates

“But that was then,” critics will say.  “What about now?”

Skeptics of the 4 Percent Rule will often argue that this past decade of lower than usual bond yields (which seems to not be changing very quickly any time soon) will lead to significantly different outcomes than we’ve seen in the past.

More so, there is also a belief that the overall average annual return of the markets will be lower over the next few years.  According to data compiled by NYU, the average annual return of the S&P 500 since 1928 has been between 9 and 10%.

In light of these arguments, you’ll often see headlines and talking heads in the media encouraging more conservative withdrawal rates of 3 percent or lower.

Should we be worried?

Not exactly.  In a recent article by financial guru Michael Kitces, he re-ran his own simulation of the 4 percent rule using modern data and found that “less than 10% of the time does the retiree ever finish with less than the starting principal”.   In addition, he also stated that “over two-thirds of the time the retiree finishes the 30-year time horizon still having more-than-double their starting principal”.

So this begs the question: Are we being too overly cautious?

The ramifications of this are not trivial.  If we are, then it means we could be over-saving when a smaller nest egg might do just fine.

So how do we help ourselves to know for sure?


Useful Portfolio Longevity Strategies

Fortunately, since the introduction of the 4 Percent Rule, there have been a number of great theories to come forward.


  • Shiller CAPE. Kitces found that trends in withdrawal rates were not random and actually related to something called the Shiller cyclically adjusted price to earnings ratio or CAPE.  He concluded that a safe withdrawal rate as high as 5.5 percent could be used if you retire when stocks are under-valued (i.e. when the Shiller CAPE less than or equal to 12).
  • Bengen’s Floor and Ceiling Method. In later publications by Bengen, he proposed a more dynamic approach for making retirement withdraws where your rate fluctuated between self-imposed “floor and ceiling” values based on the current market performance.  The idea behind this new method was to allow retirees to safely enjoy more of their savings while taking helpful precautions to scale back in tough times.
  • Percentage of Remaining Balance. This method has the retiree taking a percentage of whatever portfolio value remains rather than some fixed, inflation-adjusted amount year after year.  Though your withdrawals would fluctuate more rapidly, with this approach, you can in theory never run out of money!
  • Then there’s an approach I suggested on my blog where you reduce the amount of inflation adjustment that you pay yourself each year. In one study we conducted, I found that by reducing your inflation adjustment by just 1.5 percent, you could nearly simulate the results of a portfolio that was using a 0.5 percent lower withdrawal rate.

Think about what an impact any one of these this could have on your savings target, and how many years it could knock off your plan.


Let’s say that you’d like to have a retirement income of $48,000 per year ($4,000 per month).  You’re saving $12,000 per year between you, your spouse, and your employer’s contributions.  You expect your savings will grow at an average annualized rate of 8 percent.  Since we want this example to be adjusted for inflation (3 percent), we’ll reduce our growth rate to 5 percent.

Using standard conventional wisdom with the 4 percent rule, we’d expect that we’d need to target is a nest egg savings of $48,000 / 0.04 = $1,200,000 to make this work.

But let’s say the Shiller CAPE is low and reveals that stocks appear to be under-valued for the next decade or so.  Knowing this, we can reasonably accept a riskier 5 percent withdrawal rate and this will drop our savings target to $48,000 / 0.05 = $960,000; a reduction of $240,000!

As a matter of fact, we could use any one of these strategies to reduce our savings AND / OR increase the likelihood that we will not ever run out of money!



Don’t fall for scare-tactic headlines that will have you saving for decades longer than you really need to.  While none of us can control what’s going to happen in the future, we can use the past to make reasonable approximations of how our finances might behave going forward.  The 4 Percent Rule already errs on the side of caution providing a value that serves as a floor.  But even if you want more safety and security than that, you can optimize beyond this using the knowledge that more recent studies provide.  That’s valuable information that can get you closer than ever to achieving financial independence!

Author bio: DJ is the author of the book “How Much Money Do I Really Need to Retire & Achieve Financial Independence?” and writer for My Money Design, a blog that details his journey to early retirement and financial freedom.  Connect with him via Facebook or Twitter.


Like it? Share it!Share on FacebookTweet about this on TwitterEmail this to someoneShare on LinkedInPin on Pinterest


  1. Hey DJ. Thanks for this very intriguing post. Mrs Grooy and I based our retirement on a 3% withdrawal rate. I always suspected that this rate might be too conservative.

    1. Thanks! Yes, a SWR of 3% is a quite conservative mathematically. However, there is a human element to this discussion that we all must also recognize. Ultimately the SWR you choose will come down to your tolerance for risk. I’m sure at 3% you will find you have just about 100% chance of success. If this makes you feel better and you can swing it, then go for it.

  2. How do taxes play into this? If I am in the 25% tax bracket, I would have to take out more than $48,000 per year to have a monthly net of $4000 or am I getting something wrong? That’s what worries me when I look ahead to my withdrawals. If I have $1,000,000 in savings to draw from, and I take out 3.5% ($35000), I have to multiply that amount by 75% to determine how much I actually have (in pocket) to spend on expenses and fun things. Yes? No? Please help me understand all of this. $35000 x .75 = $26250 annually. Divide that by 12 months and I only have $2,817.50 to work with.

    1. Dell, GREAT question, and you’re spot on. You absolutely have to account for taxes in your Withdrawal Rate calculation. Your withdrawal rate must include the funds you’ll require to pay your taxes. If the majority of your retirement savings are in IRA/401(k) Before-Tax accounts, the tax burden could be significant.

      I sense that this is an under-appreciated aspect of retirement cash flow planning, and appreciate you highlighting the point.

      1. If I wait until my full retirement age (66) to begin taking my Social Security benefit, I may be able to manage on my monthly Soc Sec combined with my $2817.50 savings withdrawal. However, I am starting to read up on Social Security and how it works and I understand that even Social Security will be taxed. It also seems that because I have been a saver all of my life, my Social Security taxes will be higher than others. Very scary.

        1. Much like a pension, it all gets taxed. It is important to account for taxes in all of these scenarios. That is why there is a push by some to convert traditional 401(k)s into Roths when your income decreases (say if you go part time) but before you retire. Same goes for other taxable accounts.

    2. When talking about SWR’s, taxes are always assumed after the fact. This has been the case in almost every professional SWR paper I have ever read because when it comes to taxes there are so many different directions and assumptions that could be made. For example: What bracket you fall into? Are you using tax-exempt resources like a Roth? Is some of this money from capital gains and qualified dividends (which falls into different tax brackets)? Each will make an impact on the answer.

      However, I will still give you an example to help clarify the issue. Remember that in the US, Federal taxes work according to a marginal tax bracket system. This means we do not simply multiply our income by 25%. We have to multiply each level of our income by the tax bracket that it falls into. More on that here: https://taxfoundation.org/2017-tax-brackets/

      If we use the example where you take out $48,000 per year and we assume you are married filing jointly, remember that you first get to subtract a standard deduction of $12,700 and two personal exemptions of $4,050 each.

      $48,000 – $20,800 = $27,200 of taxable income.

      Going by the 2017 tax brackets:
      10% x $18,650 = $1,865
      15% x $8,550 = $1,283
      Total Federal taxes = $3,148, or an effective tax rate of 6.6%. This works out to only $262 /month.

      (You can of course redo the example with whatever numbers better describe your real situation).

      Again, this scenario could potentially be even further improved if some of this money were to come from Roth sources, capital gains, or qualified dividends.

      1. Hi DJ (and Fritz),

        Great post! I also found your blog post on the impact of reducing the inflation adjustment (linked above) to be very useful. So many of us in the ER community consider a higher WR early on, considering later SS or pension income will help sustain longer-term income needs. Your demonstration/walk-through was beautiful! This is a very helpful insight about how much an impact this approach can make on designing a plan.

        Nature Lover

        1. Thanks! I appreciate the positive feedback. Yes, the inflation adjustment impact on withdrawal rates was one of the most interesting takeaways I got from the Trinity Study. Although it is certainly not realistic to assume a retiree would never adjust for inflation, it is still very helpful to know that it could be used (even in modest amounts) to either achieve a higher safe withdrawal rate or increase the likelihood that you will not run out of savings. It never hurts to have one more trick you can use at your disposal!

      2. Thanks for the info and the article. I find retirement a little intimidating and I appreciate all the information I can get.

  3. Honestly I think your being a bit aggressive here. I agree the four percent rule works for thirty year periods, but the odds are significantly lower over sixty years. Why? The four percent rule checks your odds of surviving thirty years. It does not check the odds of keeping your assets in tact. Ie after thirty years you again need to pass another thirty year check so you must have met the four percent rule again. As such I would recommend 3.25 for long retirements. After thirty years of retirement going back to work will be less likely. Alternately having alternate income streams in retirement can change this significantly. Social security is one such income to consider. So I guess it depends. If your in your 50s I’d trust the four percent rule or maybe even more aggressive. If your forty I’d be prepared to go higher. This of course assumes you want to retire early, which is also not necessarily true.

    1. FTF, I tend to agree with you, and I’m using 3.5% for my SWR (I’ll be retiring at Age 55), with a floor and ceiling range. I think it’s good to consider multiple points of view on this critical topic, then decide for yourself what you’re comfortable with for your personal situation. Clearly, the younger you are when you retire, the more conservative you should be with your SWR. Thanks for stopping by!

    2. I do agree that for early retirees there is more risk to the standard 4 Percent Rule. A typical early retiree in his 40’s or 50’s will need his money to last 40-50 years; well beyond the original assumption that one would only be retired for 30 years.

      However, don’t forget that Bengen’s 1994 paper demonstrated that a SWR of 3.5% would leave you with money for 50 years or more for every rolling period analyzed. You can also loosely recreate these same results for yourself using FIRECalc.

      I also would not necessarily turn away a SWR of 4.0%. As I mentioned in the article, Kitces published in 2015 that “less than 10% of the time does the retiree ever finish with less than the starting principal”. That’s a 90% success rate.

      This of course all gets back to each individual’s level of comfort and tolerance for risk. If 90% odds sound good, then a 4.0% SWR will do fine. But if you need your confidence level to be as close to 100% as possible, then 3.5% might be a better fit.

    1. Hey 10!. Wow, we’re getting specific now, aren’t we! I’ll have to check out your post, thanks for sharing your SWR. We’re similar, given that I’m using 3.5% with a floor/ceiling.

      Curious, how old will you be when you retire? (or, if you’ve already retired, how old were you?).

    2. Thanks for sharing! I’ll have to give it a thorough read later.

      I just briefly looked it over, and it appears this conclusion is for a 50-year term, correct?

      Also I noticed that it uses a 100% equity index. Did you try it with various levels of bond inclusion? I only ask because both the Bengen paper and Trinity Study concluded 100% equities to be inefficient. The sequence of returns risk was too high and produced counterproductive results.

      1. MMD, You are right about the case but in the article I clearly state the objective of that exercise. Insisting on 100% probability of success with 100% equities forces the back-data calculator to choose the absolute lowest SWR, what I call failure edge. That was found to be 3.27%. So, the result should be used only for that purpose, and no other inference to be drawn. Including even 10% bonds or cash gets you better on the efficient frontier curve but the downside is lower median terminal value, which is obviously highest for 100% equities case. So, using a small bond allocation will actually increase the SWR in this case for the reason you mention but that wasn’t the objective of that exercise – it was to get to the worst case SWR under current environment of low bond yields (which is why bonds were even removed from the simulation). Coincidentally, my ultra conservative SWR from this exercise matches what ERN arrived at after extensive simulations and analysis.

        Fritz, to answer your question, I haven’t decided that yet! I am enjoying the ride for now.

    1. Unfortunately RMD’s are a tricky topic. As I’m sure you’re aware, if you don’t take the RMD, then you will faced with a hefty penalty that will definitely eat away at your savings unnecessarily. Therefore, even if your RMD is higher than a SWR, the government leaves you little choice but to take the RMD. However, that doesn’t mean you have to spend it. This money could easily be diverted into your taxable investments so that you can grow your capital gains and qualified dividends. As long as your income stays below the 25% tax bracket, these assets would continue to be available as you need them tax-free.

    2. Hey David, nice to see you leaving a comment on my site, hope all is well with you. DJ beat me to the punch in answering your question about RMD’s, but I agree with his point. RMD’s are a separate topic, and the level that you’re required to distribute is mandatory. If it’s in excess of your SWR, simply reinvest the excess in your taxable accounts, as suggested by DJ.

  4. MMD and Fritz.

    Good post that adds to the growing awareness of the 4% myth. Like Full Time Finance and TFR, I sit on the low side of the fence. My wife and I retire next year at age 51 and 45 respectively with two young boys ( they will be 11 and 9 when we pull the plug). Our projected SWR will be 2.5% and we can make that 2% if necessary. That low WR is driven in part by a modest, but not insignificant pension, that I get from my (soon to be) former company. That will serve as income floor (starting at age 51) in form of a lifetime annuity once I transfer the current cash balance to that form of payment. One could also argue that we have “over-saved” but we prefer to be conservative and fully embrace the thought of a higher WR if market tail-winds are at our back. Pleasantly surprised and all that. By the way, we factor in kids college funds in our net worth but NOT in our SWR calculations.

    Early Retirement Now has a great series on the SWR and the impact of retirement age, desire to leave a legacy fund, SS projections and asset allocation risk tolerance all play into it. If you are early 40’s, want significant % bonds in your portfolio and are projected to have little SS income, you better plan to take a haircut on the WR. Think of a haircut of 0.5-0.75% or else trouble awaits…….

    1. Great point on Early Retirement Now’s series on Withdrawal Rates. The best, most comprehensive, series I’ve seen on the topic. A 14-part series!! (Click here to have a look). It was because of his series that we decided on a 3.5% SWR, with downside to 3.0% if required.

      Congrats on being able to have a SWR of 2.5% with such an early retirement. You’re crushing it! I’m curious, is your pension inflation adjusted, or fixed? I also have a pension, but it’s fixed. Therefore, I’m looking at the investment side of my equation to offset inflation over the years. It’s a factor, and something you may want to think about if you haven’t already. Given that you’re “over-saved”, you likely have a natural hedge. Also, I agree with leaving the kids college $$ out of the SWR calc.

      Again, well done!

      1. Like your plan, mine is fixed also. I don’t enter any COLA or inflation adjusted numbers when I run the retirement calculators. Would be a bit sweeter if we were both inflation adjusted wouldn’t it….?

        Still, I can’t complain about having a pension where my company made all the contributions over the years. Has eased the pain a little after nearly twenty years hard labor…! :>)

    2. That’s a remarkable achievement! To be able to retire by age 51 and only need to draw on 2.5% of your income means you guys should be rock solid financially.

      I’m curious: If you didn’t have the pension, do you think you would still gravitate between 2 and 2.5%? Or do you think you would be more open to a higher rate? I only ask because I believe you are in a more unique situation and most people reading this will not be eligible for a pension.

      1. Without pension, We would be looking at a SWR between 2.5-3%. Still conservative. We will be watching the spend like hawks in the early years and will be flexible as the markets dictate.

  5. Thanks for this post. At age 52, I am planning on starting with a withdrawal rate of 2%. I am open to considering a 3-4% withdrawal rate after I am older than 60. I just tend to be conservative.

    1. 2% is what I would consider to be very conservative. But then again I cannot blame you. Being ultra-conservative early on introduces another extremely important concern: The first 10 years are what will make or break your retirement savings. This is due mostly to the phenomenon called “sequence of returns risk”. All it takes is a couple of “bad” years early on to completely devastate your nest egg. If you’d like to see a simple example, then please check out this guest post about it here:

      If there is a time to be as tight with your money as possible, statistically the first decade of your retirement is definitely it.

  6. Past performance is NO indication of future results. The problem with the (lack of) strategy is not the rear view mirror, linear, theoretical math. It is the lack of options which are the only method to deal with future uncertainty: volatility, inflation, taxes, fees, interest rates, lawsuits, health, healthcare costs, longevity, sickness, death, and other unknowns. In any event 30,40, or 50k in income per million of assets is pretty weak and depressing. There needs to be a more efficient strategy to allow for the option of safely taking higher withdrawal rates with contingencies, flexibility, and options.

  7. I have written extensively on the issue of SWR on my blog and I’m a skeptic of the 4% rule. It has a failure rate that’s too high conditional on today’s CAPE value. And we can come up with new ways of repackaging the same data, but that won’t change the facts. For example, the Kitces 90% rate of maintaining purchasing power? Make sure you read the fine print: that’s in nominal dollars. Lots of people who preserved only the nominal value of their initial portfolio after 30 years will run out of money over the next 20 years of retirement. The actual probability of preserving your initial portfolio after 30 years with a 4% WR and a 60/40 portfolio is only around 55%, not 90%, and that’s unconditional of the CAPE. With a CAPE between 20 and 30, this probability drops to 30% and for a CAPE>0 it’s 0%. A far cry from that 90% estimate from Kitces that’s floating around.

    I computed simulations conditional on the CAPE between 20 and 30 (today: 29). Failure rates of around 30-40+% after 50-60 years:
    (see part 3 of the SWR series)
    OK, admittedly, this assumes a 50% final value, but even with asset depletion, the failure rates were 25-40%. Not a pretty picture!

    But I agree that we don’t have to go all the way down to 3% SWR or below. 3.5% should likely be all right. 3.25% would be extremely safe. I find the 3% and sub-3% SWR a little bit too conservative. But, hey, the risk is ending up with too much money. Not the worst outcome. 🙂

      1. Dr ERN,

        Glad you are here to keep the community on the straight and narrow on all things related to SWR!

        And really cool to hear your voice recently on the ChooseFI podcast. Very clear, straight to the point – just like your writing.

    1. Thanks for the input. I see more and more that your series is being referenced throughout FI articles, and its definitely something I will need to review in the future. I find it interesting that your last comment recommending a SWR of 3-3.5% fits right in line with Bengen’s conclusion in his 1994 article: 3-3.5% is pretty much the threshold between nearly-guaranteed financial success and starting to encroach on taking some level of risk.

  8. This was a great post! So many quote 4% but almost no one goes into the nuances, bless you. I’m good with 3% because I saved a lot, but most people aren’t so fortunate. I hate to think of the needless pain misinterpreting safe withdrawal studies has caused.

  9. With my retirement still far away (10+ years) I stick to the 4pct for now. When we get closer and closer, I will work out what rate might please me more.
    FOr me, it is a possibility that I keep “working”, even post FI. Ideally, work covers my basic spending, so that the stash the first years only needs to pay for holidays and extra’s. It then can grow bigger so that by the time I stop working, my effective rate is way lower.

    Why would I do this? With a 50pct tax on any income above 38K and no pre tax contributions in Belgium, it is hard to save a lot…

    Thx fot his view, I like post like this (just like the ERN series)

  10. Great post, and food for thought! It seems one can justify a SWR anywhere from 1-6% by fudging expected returns, inflation, SS, etc. In the end it comes down to what makes you “feel good,” which makes total sense. We are talking about the rest of our lives!

    My comfort zone usually hovers around 3.5%.

    1. Dr. C. SWR is, indeed, a very personal decision. I just hope folks are knowledgeable when they make it. I suspect most of the readers of this blog are savvy and understand the implications, but worry about the many folks who don’t.

      We’re aligned on the 3.5%, same target as mine! Let’s hope we’re right….

  11. Very interesting! I’m hoping to hold off on withdrawal for a few years, but I could see drawing 5% because we have a lot of other income streams. Especially if it were for discretionary spending and not fixed bills. If I just drew from stocks, I might not stack my bills up to 5%.

    1. Ms. Montana, I like the way you’re approaching life, with your family sabbatical in Montana. Curious if you’ve had to touch your principal while you’ve been on sabbatical, or if your “other income streams” have been sufficient to cover your living expenses? With a house full of kids, I’m sure your expenses will go down in retirement!

  12. Jon Guyton and Michael Kitces (and maybe Wade Pfau, as well) have pointed out that withdrawal rates of 5% or 6% or even higher are certainly possible. All that seems to be necessary is the ability to reduce spending when markets decline or under-perform.

    At age 66, my biggest fear is not running out of money. Rather, it is that I might reach age 85 and discover that I could have spent tens of thousands more each year than I actually did.

    Level spending throughout a 25 to 35 year retirement is neither likely nor desired by most people. A healthy and energetic 65 year old receives much greater utility from an extra $10,000 annual discretionary spending than does a worn out 90 year old.

    Advocating a 4% safe withdrawal rate – based on an incorrect assumption that retirees will desire level real spending until the end of life – is far too conservative and does a great disservice to retirees who have accumulated adequate nest eggs.

Leave a Reply

Your email address will not be published. Required fields are marked *