≡ Menu

How to improve your sustainable withdrawal rate

A hamster in a wheel with the caption enough of this already.

Thus far we’ve explored why the 4% rule doesn’t work in the real world, and established a more sustainable withdrawal rate (SWR) for UK / global investors. Please read those articles if you’ve not already done so, in order to get the most out of this piece.

Now for the good bit! We’re going to talk about why you can use a higher SWR if – and only if – you’re prepared to execute a withdrawal plan that’s considerably more sophisticated than the 4% rule.

To raise our SWR we’re continuing to use the layer cake concept advocated by leading retirement researchers William Bengen and Michael Kitces.

The layer cake personalises our SWR by applying a suite of plus and minus factors.

  • The last post was all about the bad stuff. We saw how it forced my SWR down to 3%.
  • Now I’m going to layer on all the positives, and test my approach using global historical data.

Without wanting to ruin the surprise, I was pretty shocked by the results you’re about to see and I think I’ll need to be more cautious than the test suggests when the rubber really hits the road.

Even Kitces is very clear about the layer cake’s limitations:

Although the ‘layer cake’ approach of safe withdrawal rates does allow for planners to adapt a safe withdrawal rate to a client’s specific circumstances, there are several important caveats to be aware of.

The first and most significant is that many of the factors discussed here were evaluated in separate research studies, and it is not necessarily clear whether they are precisely additive.

Okay, so remember my SWR is currently bottomed out at 3%.

Let’s head for the top!

Diversification sweetener

Our baseline SWR assumes we’re invested in a Developed World 50:50 equity / bond portfolio. Yet there’s plenty of evidence that a stronger equity tilt and more diversification increases your SWR, especially over longer time horizons.

The shotgun spread of long-term equity returns means that an equity-heavy portfolio can shoot the lights out sometimes. However it also falls far short of the target on unlucky occasions. Bonds can staunch the bleeding when equities haemorrhage, yet too much bondage may also cripple you over time.

Early Retirement Now (ERN) sums up the dilemma:

Stocks have a lot of short-term risks, but in the long-term stock returns are tied to economic growth and thus, in the very long-term, real returns become less risky due to that.

Bonds have relatively little short-term risk around their trend growth rate, but their trend growth path itself has a lot of risk in stark contrast to stocks.

The answer isn’t to simply load up 80-100% in equities and hang on for dear life. Rather we can aim to alter our asset allocation as we go.

Michael McClung in his brilliant retirement portfolio book, Living Off Your Money, recommends an eve-of-retirement bond allocation of 50% for a 30-year time horizon, or 45%-35% bonds for longer stretches.

The reason for starting retirement with a heavy bond load is that you’re particularly susceptible to sequence of return risk in the closing years of accumulation and the opening decade of deaccumulation. You can protect yourself during this period with high-quality government bonds.

Kitces has shown how an ill-timed stock market crash can setback your retirement plans like an asteroid strike scuppered talking dinosaurs. He also explains how the sequence of returns in the first 10 to 15 years of your retirement can seal your fate for good or bad.

It’s important to have enough bonds to deal with these threats, and to deploy your bonds effectively.

McClung in particular has developed techniques to help you do this – see the ‘dynamic asset allocation’ section below. The trick is that you allow your bond allocation to wax and wane according to market conditions.

More generally, the historical data sampled by Kitces, Bengen, McClung and others tells us that broad diversification beyond bonds works in retirement just like it does during the accumulation phase.

They cite evidence in favour of diversifying your retirement portfolio with:

Kitces offers a diversification bonus of +0.5% SWR for significant multi-asset class diversification. It seems daft not to take it.

The Accumulator’s layer cake SWR:

3% + 0.5% diversification = 3.5%

Dynamic asset allocation

The best way to protect yourself from sequence of return risk? Live off your bonds when equities are down.

Beyond that you can further improve your portfolio’s life expectancy by using a rebalancing method that increases equities exposure when they’re seemingly cheap, and only replenishes bonds when equities have seriously outperformed.

This is dynamic asset allocation. It’s a super-charged version of ‘buy low, sell high.’ You can read about McClung’s version – called ‘Prime Harvesting’ – by downloading a free sample of his book.

Techniques such as dynamic asset allocation will test your risk tolerance because in extreme market conditions you may eat all your bonds and end up 100% in equities. Steer clear if you’re cautious.

Otherwise Kitces awards 0.2% to your SWR for dynamic allocation.

The Accumulator’s layer cake SWR:

3.5% + 0.2% dynamic asset allocation = 3.7%

Flexible spending and dynamic withdrawal rates

Here is the big SWR cherry on top. If you can cut your spending during a downturn then you gain a massive advantage over a constant inflation-adjusted withdrawal plan.

Dynamic withdrawal rates mean you adjust your spending in sympathy with your portfolio’s fortunes. Like managing a forest, being able to conserve your resources when they’re under stress is obviously more sustainable than consuming an ever greater percentage when the rot sets in.

Flexibility is key. Retirement researchers have devised all kinds of rules that allow you to spend more when the market soars, but you must also spend less when there’s trouble at mill.

McClung does a masterful job of analysing various dynamic withdrawal rates in his book, while ERN has written a sobering series exploring how much you might have to cut back using one of the better known spending systems.

In practice, cutting back is what all retirees do if their money runs short. The risk is that extra spending early in retirement may force us to spend less later if the cookie doesn’t crumble our way.

That gamble may be easier to take if you believe that retirees spend less later in life. What do you reckon? The evidence is patchy and may not apply to you. I’ve read research that concludes retirees spend more if they have it, but spend less on average because most end up with less to spend.

Kitces’ flexible spending modifier:

  • +0.5% SWR for modest spending cuts in bear markets and/or plan to decrease spending in later life.
  • +1% SWR for substantial (10%+) spending cuts in bear markets and/or plan to make significant cuts in later life.

I think Mrs Accumulator and I can handle 20% spending cuts so I plan to use Michael McClung’s EM dynamic withdrawal rules. Our State Pensions should meet near 70% of our estimated outgoings later on. I’m gonna claim the full 1% bonus!

The Accumulator’s layer cake SWR:

3.7% + 1% flexibility bonus = 4.7%

My new world portfolio SWR

My personal SWR was creamed by negative factors in the last post. It finished up at just 3%.

Now it stands at 4.7% and I’m stunned.

What does it all add up to in cold hard cash?

Well, we’d like an annual retirement income of £25,000, so our retirement wealth target at 4.7% SWR is:

(1 / 4.7) x 100 x £25,000 = £531,750

In contrast our target at 3% SWR was £833,333, which was 57% higher.

Wow. Just wow.

If I use the ‘4.7% rule’ then we’re FI already!

The sniff test

The big question is does this layer cake business pass mustard?

The research shows that your SWR changes dynamically as you shift the parameters. I can test these moving goalposts using global historical data thanks to the fantastic Timeline app.

Timeline is commercial software created by retirement researcher Abraham Okusanya. It’s aimed at financial planners who want to model portfolio withdrawal plans.

Timeline is very well designed, loaded with great features, and delivers the Holy Grail of global / UK appropriate datasets. I think it’s worth paying an IFA to run your numbers through it if they have access.1

My 4.7% SWR achieved a 99% success rating on Timeline – success means historical me didn’t run out of money in 99% of scenarios.

The bottom 10% of scenarios did require me to cut spending drastically to actually avoid running out of money though. That may not be your idea of success.

On the other hand, every path above the 10th percentile was comfy, and the best case scenario near-tripled my income for years. I used all the layer cake assumptions – good and bad – to get the result but was able to leave our State Pensions in reserve.

However that doesn’t tell me that my 4.7% rule is safe or even sustainable. We live in the future, not the past. I won’t experience the historical data in retirement, though hopefully I won’t face anything worse.

The truth is that a lower SWR is safer no matter how much kung-fu you know, so I’m not actually going to adopt a 4.7% SWR.

4% it is

So after everything we’ve been through I’m going to choose 4%.

Editor: You clutz! You total time-waster! Is this your idea of a joke?

Okay look, it’s thousands of words later and I’m as aghast as anyone, but I’m not using that 4% rule.

  • I’d have to use a 3% SWR to live with naive, constant inflation-adjusted rules.
  • But 4% with all the layer cake trimmings works with McClung’s system and it comfortably performs in Timeline.

In short, I am only happy to choose 4% because it leaves me room for manoeuvre when allied with the withdrawal techniques we’ve touched on in this series.

I also have – and must have – a Plan B.

Maybe my ability to use complex techniques will ebb through my eighties and nineties? Maybe I won’t even make it that far – a big problem given Mrs Accumulator’s interest in dynamic withdrawal rates continues to hover around zero. (“But look, they’re dynamic!”)

Plan B is to switch to a simpler, safer Floor and Upside strategy when our State Pensions kick in and annuity rates tip in our favour.

Aside from that we’ll maintain an emergency fund, there’s always the house to sell or reverse mortgage, and there are side hustles to hustle if we have to.

More than anything, digging into the research has taught me that a SWR is a very personal number. Like inside leg measurement personal. And it’s probably not even a number.

Really, it’s a floating set of coordinates that give you something to aim for. Your final destination can only be known when you arrive.

Take it steady,

The Accumulator

  1. For a fixed fee of course. []

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

{ 64 comments… add one }
  • 51 Whettam April 25, 2019, 6:52 pm

    @TA np on the typo. I think why I have enjoyed this series, McClungs book and this conversation so much, is that it reflects my journey / interest in early retirement. I was initially just really focused on the SWR and therefore what number I needed to reach to be able to retire early, however as I get nearer to my target, I have become much more interested in the practical approach to deaccumulation e.g. harvesting approach, variable spending strategies, drawdown asset allocation, etc.

    +1 @TI for your post on natural yield, I agree approach is not for everyone, but not sure why you say a “larger wodge” is required? If your numbers are right (and they look sensible to me) then its possible to use initial portfolio to buy sustainable yield with a starting yield of 3.5% – 4% so within the same range as @TA’s plan.

    As several have pointed out, one of the problems with proving that dividends have a valid role to play in drawdown is getting the data for the back testing, but it’s a shame because I think several researchers / data scientists are drawing the wrong conclusion, just because the data isn’t readily available.

    @xxdo9 that’s a 105% 😉

  • 52 xxd09 April 26, 2019, 12:18 am

    Apologies
    Old age coming sooner than expected!
    Should be 30% Equities
    xxd09

  • 53 James April 26, 2019, 1:19 am

    It can be important to understand how duration assets perform during positive and negative economic and market trends. In the papers “Improving Investment Outcomes on the 60 / 40 Portfolio” * and “A Three Asset Class Portfolio For Long Term Wealth Accumulation ” **, the use of the REIT, dividend growth, total world, and momentum stock universes are employed in the construction of portfolios that have shown higher asset trajectory growth / alpha over a 3 decade sample. Duration assets are employed during rare significant decline events. The common wisdom that duration assets must be held at all times can be a misperception of how “risk” can be experienced by an investor ( Part 2, charts 2 – 8 in ” So how does an investor use these trend indicators in investing ?” section * ). Portfolio volatility may be perceived as less of a “threat” in a market / account balance that is in a positive trend / when times are “good” ( the last 10 years for example ). Therefore,a portfolio constructed with a mix of duration assets and equity based assets may express the same volatility as a portfolio constructed with equity based assets exclusively. It is during large and emotional significant decline periods when the effects of volatilitiy may be felt, with even an investor holding a 60 / 40 portfolio feeling a “loss” aversion” bias and possibly succumbing to panic selling. Yet in a positive market trend environmnent, an investor holding a conventionally promoted 60 / 40 portfolio may be vastly shortchanging themselves of critical asset growth over time – this under the guise of “safer” or “less risk”. With the tactical piece in place, a larger “latitude” of SWR may be considered.
    . . . .
    * https://tinyurl.com/yygwgyyu
    ** https://tinyurl.com/yy4wgqb7

  • 54 MrOptimistic April 26, 2019, 7:09 am

    @Whettam. Think the larger wodge conclusion arises from an assumption that your asset selection is neutral with respect to withdrawal strategy, the dividend and coupon yield being less than the desired withdrawal rate.

  • 55 ZXSpectrum48k April 26, 2019, 7:56 am

    @TA. “You’re being grossly unfair”. Quite possible. I se value in what McClung has done. I just think at some points he can’t see the woods for the trees. He should want to quantify the sensitivity of his results to the datasets and boundary conditions. Put “error bars” around it, if you will. I think retirement researchers don’t do this because the “error bars” are of the same order of magnitude as the SWR and this might undermine the message. I disagree; it enhances the message.

    In terms of datasets, there are many available but they all cost money, some rather a lot. If nothing else, however, he could of bought a second cheap set like DMS ($4k?) to compare and extend the history for both the UK and global (ex US) returns. Taking the UK history back from the mid 20s to 1900 has quite a damaging impact on the SWRs. Many of the worst years are in that first decade.

    He also spends little time on the liability side of the equation. I think he spends just one page on inflation. Deciding the appropriate inflator is subjective but crucial in determining the SWR for early retirees. A different inflator could effectively move the WR lower by 100bp+. I see so little mention of this on FIRE blogs but the following at least raises the issue. https://medium.com/@justusjp/inflation-may-be-the-wrong-measure-for-early-retirees-8c78bd0b7283.

    As you say it’s a personal thing. I find most research too optimistic (Kitces unbearably so). Probably only ERN is toward my idea of neutral. Those “error bars” mean we can be right and wrong.

  • 56 MrOptimistic April 26, 2019, 8:30 am

    @ZX. Interesting. I returned to McClung’ s book yesterday and was reminded how ‘frilly’ it seems. All those factors, tilts and so on. I am more interested in asset allocation and associated withdrawal strategy ( what to sell, rebalancing, allocation) and find all the apparent sophistication to establish a SWR to +/-0.x % rather like whistling in the dark given the scale and range of future uncertainties and unknowns.
    I also find the natural yield approach something of a false friend giving apparent security when it does no such thing, just reduces diversity. It does however appeal psychologically in a similar way as bucket strategies.
    Since we can’t escape our own psychologies ( comfort zone) my approach is effectively a 3 bucket strategy informed, I hope, by all these other factors discussed here. I aim for an amount that matches what I need for the lifestyle, plus a bit. Opportunity losses mean nothing compared to minimising maximum loss, what it may cost in terms of lost gain is an insurance premium.

  • 57 Foxy Michael April 26, 2019, 10:27 am

    @TI Although FTSE 100 enjoys a 4% natural yield, the World is not so generous with its 2%. How do passive investors go about picking the right yield in a natural yield strategy?

    Dividends have become just another metric in the total returns toolbox. Long gone are the days where companies issued high dividends to attract shareholders. How does your SWR change when companies buyback shares instead of issuing dividends for tax reasons? What’s the tipping point?

    Psychologically, it may feel better but If you think that companies that pay dividends are a better strategy for retirement than those which don’t, then this thinking probably leaves money on the table (https://earlyretirementnow.com/2019/02/13/yield-illusion-swr-series-part-29/).

    Just to be clear, I’m not accusing dividend investing by any means. But what I’m trying to say is that dividend investing won’t give you a better SWR strategy compared to an all-equity one. You will be getting a higher income at the expense of lower price returns. I think building an income-heavy equity portfolio compared to an all-inclusive equity one falls into the mental accounting category.

    Looking forward to a natural yield article backed by numbers.

  • 58 Factor April 26, 2019, 12:11 pm

    For those incorporating an income safety “floor” in their drawdown planning, my £K income/expenditure budget for 2019/20, 16 years into retirement and based on known or past actuals, is income after tax 22 (OP after sharing on divorce 10, SP 8, ISA’d IT dividends 4) and expenditure 20 (need to spend 15, nice to spend 5), leaving “rainy day money” 2.

    N.B. The basic state pension alone = c.penury.

  • 59 Naeclue April 26, 2019, 12:26 pm

    This January was one of those rare times when I was rebalancing into equities. Since I retired it has all gone the other way. With McClung’s Prime Harvesting, he would have me never rebalance into equities and simply drawdown from bonds/cash, then replenish from equities when they grow by 20% in real terms. Just a one off case I know, but so far this year my equity portfolio is up about 10%, bonds/cash maybe 1%, so rebalancing has worked out better than not doing so, but for the previous few years I would have been better off not rebalancing out of equities.

    One thing all the recent research has shown is that the classic advice to reduce exposure to equities during drawdown has not been optimal. The question I have is, will this continue? I have been wondering whether the conclusions reached from much of the latest research may be biased by the exceptional equity returns (and sequence of returns) over the last 40 years which occurred mostly in the first 20 years. A few months ago I looked at the real returns on the MSCI World index, converted to pounds (Basically how I am invested, but excluding EM which I could not get figures for). I used inflation figures from the Barclays Equity Gilts report to convert to real returns. I found that for the 40 year period ending in December 2018, the real CAGR was 6.7%, but due to the very favourable sequence of returns, the maximum SWR, starting at the end of 1978 and ending 2018, was about 7.7%. The sustainable withdrawal rate, ending 2018 with the same inflation adjusted amount as in 1978, was 7.1%. But for the 20 year period ending 2018, the real CAGR was only 3.6% and the maximum SWR 5.1%. The sustainable withdrawal rate only 2.8% due to the poor sequence of returns from the end of the dot-com bust. A 60/40 equity/gilts fund reduced the sustainable withdrawal rate to only 2.5%.

    I think it is going to be very interesting to see how this progresses and what the research looks like in 10 years time. At the moment, it looks to me that UK 30 year MSWR’s of over 4%, however drawdown takes place, looks a little optimistic, which is why I settled on 3%. Apart form McClung’s, most research is based on US data, so maybe those investing and drawing in dollars will do better again as they have in the past, but for £ investors currently I am not optimistic.

  • 60 Naeclue April 26, 2019, 12:58 pm

    Forgot to mention, drawing down 100% equities (MSCI World) at constant, inflation adjusted 4% from the end of 1998 would have a left remaining pot size of 49% of the inflation adjusted starting value at the end of 2018, or 48% for a 60/40 fund. So both should make it through the next 10 years without running out, so maybe I am being a little too pessimistic.

  • 61 Monty Carlo April 26, 2019, 1:11 pm

    If you’re interested in rigorous financial modelling, monte carlo forecasting is a more complete look than historical backtesting.

    There is a fantastic (and free) online resource, that does both historical backtesting and monte carlo forecasting at https://www.portfoliovisualizer.com/

  • 62 The Investor April 26, 2019, 1:28 pm

    @Monty — Indeed. 🙂 We’ve covered Monte Carlo simulation before.

    E.g:
    https://monevator.com/how-to-stress-test-your-retirement-plan/

    They do come with their own problems though — or put better you have to be aware of the limitations.

    E.g.
    http://www.theretirementcafe.com/2018/07/monte-carlo-and-tales-of-fat-tails.html

  • 63 W Neil April 29, 2019, 7:04 am

    Most SWRs allow for an annual increase in line with inflation. Maintaining a level of income relative to the working population is another matter altogether. Salary and wages have some correlation to inflation and the working population is likely to be financing higher levels of debt and other expenses, so a greater proportion of retirees’ income should be disposable. However the gross income that can be generated from a conservative SWR will lag wages and income growth over time. Here’s a time series for the UK https://www.ons.gov.uk/economy/grossdomesticproductgdp/timeseries/kgq2/qna

  • 64 Adam June 14, 2019, 1:21 pm

    Nice article, thread. You have state pension, which I think is too pessimistic to think will be scrapped as we have lowest in the western world. You can equity release from your house, as an option. Not ideal but extra safety factor. May be inheritance in future.

    The 4% rule assumes preserve your principal instead of running it down. Also many retirees and early retirees work part time or have side income.

    I think you’re right to consider the upsides as well as risks otherwise you could end up with a taxable legacy, and/or working for longer than required.

Leave a Comment