≡ Menu

Why the 4% rule doesn’t work

The 4% rule is about as safe as a bomb, a lightning strike, a virus, a rocketThe 4% rule went viral because it was billed as simple and safe (*coughs a noise that rhymes with bullwhip*).

Unfortunately, the 4% rule is not safe.

Nor is it simple, once you put the nuance back.

The story is seductive – that you can withdraw 4% a year from your portfolio and never run out of money.

This is often known as a safe withdrawal rate (SWR). Unfortunately the 4% version is about as reliable as that other withdrawal method you’ve heard of.

Got a portfolio of £1 million? The 4% rule claims you can safely withdraw £40,000 in year one, adjust that amount by inflation in year two, and so on, every year until the happy hereafter.

The 4% rule also gives us the rule of 25. Want to live the life of Reilly on £40,000 a year?

£40,000 x 25 = £1 million

That’s the sum you need to amass before you can hit the beach.

Simple as that. At a stroke of the calculator anyone grappling with a defined contribution pension can treat it like it’s one of those turnkey defined benefit, gold-plated jobs!

If only.

The 4% rule: the things they forgot to tell you

Where to begin?

The 4% rule doesn’t include taxes. The £40,000 figure above is gross income. You’ll need to live on less after tax.

Worse: the investment growth assumptions that underwrite the rule assume no capital gains, dividend or interest taxes. If your investments aren’t completely shielded from tax then you’ll need to lower your SWR.

The 4% rule doesn’t include investment costs. Fund charges and platform fees chip away at your annual returns and leech the SWR. Financial planner and researcher Michael Kitces explains that the answer isn’t as simple as deducting your portfolio’s total OCF from the SWR either.

(Another thing to note: the 4% rule reinvests dividends. If yours are spent or taxed then fuhgeddaboudit.)

The 4% rule applies to 30-year retirements. If you live longer than 30 years then the failure rate creeps up unless your SWR goes down.

Financial planner William Bengen, whose research inspired the 4% rule, recommended a 3% SWR to see you through 50 years or more.

The 4% rule uses US historical returns. Bengen’s original portfolio comprised:

  • 50% US equities
  • 50% US intermediate government bonds.

Bengen then used historical annual returns from 1926 onwards to discover that an initial withdrawal of 4% would have enabled retirees to live out the next 30 years on a constant, inflation-adjusted income, without running out of money, come hell or Great Depression.

That’s nice, but remember the US enjoyed super-powered investment returns during the period studied. Other developed countries did not fair so well. Retirement researcher Wade Pfau calculates:

  • The UK’s SWR as 3.36%
  • Germany’s as 1.01%
  • Japan’s as 0.27%.
  • Even the global portfolio only made 3.45%

Apply the 4% rule to Japan and your money ran out one third of the time. In the worst case, your money evaporated in just three years!

Pfau and others even doubt that Americans can rely on future returns being so kind.

Known safe withdrawal rates will fall if a future sequence of returns is worse than anything currently stinking up the historical record.

What can you do with that information? Well, some researchers have worked on the link between current asset valuations and SWR. You’re advised to choose a more conservative SWR when valuations are high, while you can live a little when valuations are low.

Incidentally, the 4% rule even fails in the US when you use a different dataset. Many retirement researchers argue that the sample sizes are too small anyway.

The 4% rule applies to a specific asset allocation. Change the 50-50 US equities and intermediate government bonds split and you’re playing a different game. Bond heavy portfolios (say over 65% bonds) have historically sustained lower SWRs, especially over longer time horizons.

Sticking with allocation, UK investors shouldn’t use US SWRs – but you should appreciate that UK SWRs aren’t appropriate either if you’ve got a globally diversified portfolio.

Bengen and others have shown that diversifying into certain risk factors can improve your SWR. What about other assets such as REITs or gold? Will they improve your chances? The future is uncertain.

The 4% rule’s definition of success is probably not yours. Some SWR studies apply a sneaky ‘success’ rate. They count a SWR as sustainable if it only failed 5% or 10% of the time. The famed Trinity study did this. I think this is acceptable, but you may not. Either way it’s not ‘safe’.

Failure itself is defined as people running out of money before they run out of time. You spent your last dollar as you expired on the final stroke of midnight, December 31st, on the 30th year of your retirement? You’re a success baby!

This definition of failure keeps things simple but it’s not realistic. Most people aren’t oblivious to plummeting portfolios. They won’t fling themselves off the cliff edge like an Olympic lemming. Many will slow down their spending before it becomes unsustainable. People also cut spending in scenarios where the situation looks dire but hindsight tells us things ultimately worked out just fine.

Sadly, you don’t know which it is at the time. People cutback early because they can’t predict if they are history in the making, or whether they’re living through just another close shave for the 4% rule. In other words, living the rule can be pretty scary without a Plan B.

The 4% rule is inflexible. What if you need to spend more than your SWR allows? I don’t mean you have the occasional bad year. I mean something changes that proves your original income estimate to be off-base. Maybe you have unforeseen health costs, or a newly dependent family member. Perhaps there’s no obvious lifestyle creep but your personal inflation rate constantly outstrips headline inflation. Within five to ten years you’re spending way more than planned.

A high SWR like 4% gives you little room for manoeuvre when high spending meets poor returns. Everybody needs a more flexible plan than the basic 4% package suggests.

What if you spend too little? The selling point of a SWR is that it’s supposed to survive the nightmare scenarios. If life turns out better for you – and most of the time it does – then you could have spent more before you were gonged off. All the other caveats notwithstanding, Kitces shows that the 4% rule typically does leave large sums of spare lolly on the table.

Now that’s a good problem to have, especially for your heirs. Whereas, if you definitely want to leave something for your heirs, well, that’s not the 4% rule’s bag. It assumes capital depletion is A-OK. If it’s not then you’re into the expensive world of capital preservation.

So, you can spend too much or too little! Which is it? Well, naive application of the 4% rule can lead to either. It’s a rule of thumb not a strategy.

None of this is meant to impugn Bengen’s original research. It was groundbreaking and he clearly flagged his assumptions back in 1994. The 4% rule has taken on a life of its own, whereas Bengen’s work was only meant to be part of the puzzle. What’s often missed is the advances made in retirement research since.

You can devise a strategy from the wider body of knowledge. See McClung and also our post on how to work out a more sustainable withdrawal rate.

Step one is understanding that living off your money is as much art as science. And step two is knowing that the 4% rule does not work as popularly advertised.

Take it steady,

The Accumulator

Bonus appendix: 4% rule maths

Year 1 income: Withdraw 4% of your starting portfolio value.

500,000 x 0.04 = £20,000 annual income

Year 2 income: Adjust last year’s income by year 1’s inflation rate (e.g. 3%):

£20,000 x 1.03 = £20,600

The 4% SWR only applies to your first withdrawal. Every year after you withdraw the same income as year 1, adjusted for inflation, regardless of the percentage that removes from your portfolio.

Year 3 income: Adjust last year’s income by year 2’s inflation rate (e.g. 2%):

£20,600 x 1.02 = £21,012

And so on. Until the end. Which this is. At least it feels like it.

Receive my articles for free in your inbox. Type your email and press submit:

{ 70 comments… add one }
  • 51 The Accumulator March 23, 2019, 12:36 pm

    Quality thread! Sorry I’ve missed it. It’s been quite the week. I agree with all who’ve said the 4% rule is a good tool for:

    1. Realising there even *is* a number.
    2. Getting you started.

    The rule has the power to completely shift your mindset and then your path in life. The important next stage is discovering that the 4% rule number definitely won’t be your number, and hey, there might not be a ‘number’ afterall, as the deaccumulators who are actually living it seem to agree.
    I agree I’d rather run a floor and upside strategy and that’s the reality for most of us thanks to the State Pension. At some point I’ll look seriously into annuitising to nail down more of the floor. Do any of the greybeards out there have insight into optimal age to get a good deal?

    @ John in Co – You and I agree. You’ve summarised the point of the article in a nutshell: to use the ‘4% rule’ you must understand it’s limitations. Where you and I disagree is what the term ‘4% rule’ has come to mean. I think you’re saying the ‘4% rule’ is the original research with all its caveats. I’m saying the ‘4% rule’ is the unnuanced sugar pill spread by “wet-beind-the-ear-reporters” and quick-win bloggers. William Bengen’s paper didn’t use the term ‘4% rule’. I have complete respect for Bengen’s research, but in my view the term ‘4% rule’ is shorthand for the naive, popularised version.

    @ TI – I respect the muddy, hand-waving approach the more I look into this, but I need something more systematic because I intend to live closer to the wire i.e. My plan relies on spending down the capital because I’m not prepared to grind out more years in accumulation until I can comfortably live off the income. I can see from your comments that you’ve already acknowledged this p-o-v; just can’t seem to stop my fingers from typing this anyway 🙂

    @ Justus – Nicely taken out of context. Your quote is from my point that Bengen used US historical returns which are rather better than most other countries. His starting point was a 50:50 portfolio. To warn readers that they need to look at other research showing the international situation, it really wasn’t necessary to burn everyone’s time with another 300 words explaining that Bengen also used *US historical returns* to explore every other possibility under the sun. I’ve linked to his research so anyone can see his parameters for themselves.

    In my second point about specific asset allocation, I simply try to warn readers that different asset allocations change the game. Bengen recommended 50-75% equities, but debate rages to this day on that score and it can never be definitively settled given it relies on trade-offs and uncertain futures. The salient point for me in this regard, is that virtually all of the research I’ve read shows that being too heavy in bonds knackers your SWR in most scenarios.

    @ Mr Optimistic – you seem to be in a great position. Happy retirement.

    @ Factor – thank you for laying out. Again, that’s a position I’d be happy to be in.

    @ Mathmo – *cof cof* that’s TA not TI.

    +1 for various comments from Brod, Whettam, Vanguardfan.

    @ Lemsip – Your comment about US and UK investors having access to same investment choices – I could interpret that as saying UK investors can therefore happily adopt popular US SWRs (I don’t think you meant this, I just want to clear up any potential for misunderstanding.) Wade Pfau’s international research is intended to show that not even US investors should rely on historical US SWRs.

    @ Chris – yes, there is lots of research on life cycle spending. The general trend for the average bod seems to show a decline in spending as people reach old old age. However, are you the average bod? Some research shows that retirees with plenty of resources maintain high levels of spending throughout their lives. There’s a strong implication that retirees with limited resources spend less as they age simply because they have less to spend.

  • 52 Mathmo March 23, 2019, 1:14 pm

    Urk, sorry TA

  • 53 MrOptimistic March 23, 2019, 5:47 pm

    @TA. Re annuities. Obviously depends on the state of the annuity market, your health, spouse etc but the general advice I got was start to look at annuities again at about age 75. There will be a scientific way if doing it but life is too short.
    Suspect life expectancy predictions are going to be shown to be too optimistic which hopefully will mean that life offices will be a bit more generous ( and transfer values for dB schemes will come down).

  • 54 SemiPassive March 23, 2019, 7:18 pm

    I’m still targeting an average 4% yield through a natural income portfolio of investment trusts and ETFs across global and UK equity income, commercial property and infrastructure (inc solar & wind energy) and bonds of various types.

    While still in the accumulation phase for at least another 6 or 7 years, I’m effectively doing a dry run of the strategy and will be checking how the total yield compares each year to see if it will be consistent enough and rise with inflation. And to see what you actually get compared to the predicted yield.

    Aware of the dangers of chasing yield, but its a pretty diversified portfolio across nearly 20 ETFs and ITs. All of them are distributing, so I can roll up the cash to make new purchases maybe a couple of times a year to keep trading costs down.
    I can’t see a better way to get around sequence of return risks/selling in a bear market when I actually start taking the cash as income. Its like an automatic SWR.
    You can always shift strategy to sell units in later retirement if it looks like you’re going to end up dying too rich.

  • 55 The Accumulator March 23, 2019, 7:25 pm

    Effectively it’s a lower SWR strategy. If your dividends are cut and you need the money then you can spend down some capital. All things being equal a lower SWR means less chance of running out of money, more chance of leaving a generous legacy to the cat’s home. Beyond that there’s just the chance of a high dividend strategy returning less than total market because of higher fees and less diversification.

  • 56 The Accumulator March 23, 2019, 7:28 pm

    @ NickC – thank you for the link to the calculator. Very handy. It says I won’t go broke even if I live to a 100. As long as I enjoy those historic US investment returns 😉

  • 57 The Accumulator March 23, 2019, 7:32 pm

    Also instructive to see how more likely I am to be dead than any other outcome.

  • 58 Kraggash March 23, 2019, 9:36 pm

    Whenever I look at the long term performance of a fund, those offering higher dividends than an ETF of that sector provide much less total return. Almost as though companies make better use of retained profits for investment, rather than paying them out as divis.
    TL:DR I think you reduce total return by chasing higher dividends.

  • 59 Gentleman's Family Finances March 24, 2019, 12:06 pm

    Whatever swr you use, retirement and early retirement is a leap of faith into the unknown.
    Our situation is that offcial retirement in our 60s looks more certain but the next 25 less so.
    Ways round that? Answers on a postcard.

  • 60 Getting Minted March 25, 2019, 11:24 pm

    As mentioned above (#21) I’m guided more by natural yield than a 4% plus inflation safe withdrawal rate but having now reviewed the numbers my spending over the last five years has stayed below both guidelines. Capital growth has however been modest after the market falls last year. I aim to remain flexible on both spending and investing according to future circumstances and will continue to use natural yield as a guideline.

  • 61 Dawn April 2, 2019, 6:44 pm

    Maybe 4% swr wont work for US market in the future as the US has been ‘leading the way’ up to this point in time. Maybe its time for the rest of the world to start ‘leading the way’ with better returns, Maybe EM will emerge and provide fantastic returns going forward. If this is so, going forward, the 4% swr will be adequate.
    Jim Collins over in the US does suggest US investers may need to diversify their portfolios globally in the future.

  • 62 The Accumulator April 6, 2019, 1:57 pm

    Hi Dawn, the market doesn’t pay you because you need it. Hopefully you have a Plan B.

    More generally, good piece here on dividend / natural yield strategies: http://www.theretirementcafe.com/2019/04/a-good-many-retirees-seem-to-be.html

  • 63 Vanguardfan April 6, 2019, 3:41 pm

    Good link. I’ve always believed the ‘spend only dividends’ strategy to be mainly about psychology. Problem is, what’s the chances of the SWR being the same as the dividend yield? And if you design your portfolio yield to fit your predetermined withdrawal rate, this is surely going to drive you towards an active strategy that is certainly going to involve more risk than a Lars type buy the world portfolio. Letting the withdrawal tail wag the asset allocation dog if you will!

  • 64 Mathmo April 6, 2019, 4:26 pm

    @Vanguardfan — very much my view. I’m not sure I want my investment / withdrawal decisions driven by the CFO’s distribution preference or the prevailing tax regime.

  • 65 Kraggash April 6, 2019, 5:05 pm

    I imagine a genuine SWR should be (must be?) very close to the rate you would get by buying an annuity. Both are pretty much risk free, and you generally have to increase risk to increase returns. Stocks are inherently risky, and if you try to eliminate risk by calculating a SWR, then you will be reducing the returns (in this case the money you take out of the fund).

  • 66 Mathmo April 6, 2019, 5:13 pm

    SWR has a greater than zero expected wealth at the end, and annuity does not. While residual values tend to have fairly low effects on valuations that far out, some non-trivial probability residuals are really very large.

    You might also wonder if the insurer needs to make a profit / charge a fee on top of the underlying economic risk and compare whether that is replicated in the SWR portfolio (wrapper charges? tax? sub-scale admin?).

  • 67 Getting Minted April 7, 2019, 6:02 pm

    I don’t agree with Retirement Café’s negative view on a spend the dividends only approach. Looking at five years of actual drawdown history from a predominantly UK equity portfolio, income returns have been reassuringly stable with an average yield of 4.05%, a standard deviation of 0.33%, and an actual range from 3.58% to 4.42%. Capital returns have an average of 0.97%, a standard deviation of 9.33%, and an actual range from -10.69% to +11.43%. Total returns have an average of 5.03%, a standard deviation of 9.43%, and an actual range from -6.45% to +15.60%. Relying on the dividends enables me to be more relaxed about the capital fluctuations. Not selling the shares enables me to benefit from future capital growth and from dividend increases. Spending about 12% less than the natural yield allows me to reinvest a little. My current natural yield is now above what the SWR would be and above the best flat rate annuities for my age.

  • 68 Kraggash April 7, 2019, 6:15 pm

    @Mathmo – the ‘expected wealth’ at the end is irrelevent. Your dead. You could not spend it when you are alive, as you still need the SWR cash. Anyhow, an endowment has money left in it at the end as well – for the insurance company. So in both cases, it is not YOU who benefits from it.

    In both cases, we are talking about SWR after costs (fund/platform versus insurer charges),

  • 69 Mathmo April 7, 2019, 7:22 pm

    It’s a very short-term view to say that I don’t benefit merely because of my inevitable mortality event.

    I do plenty of things which I will never see come to fruition, but derive satisfaction from the knowledge of doing the right thing.

    Despite this, there’ s a greater fallacy at the base of assuming there’s the same economic activity going on behind the swr and the annuity. The statistical nature of one vs the single probability of the other and the economic rent of the insurer, together with the underlying investment constraints mean they are cousins, not siblings.

  • 70 Kraggash April 7, 2019, 9:07 pm

    Indeed, Mathmo, but you are constraining the statistical nature of the one by imposing the S part of SWR. I.e. Safe = must not run out under any circumstances, or length of withdrawal period would bring the SWR down close to the deterministic return of the latter. Indeed, if not, why would annuity providers not seek to obtain the higher returns of the ‘statistical method’ if both approaches were equally safe?

Leave a Comment