The nastiest, hardest problem in finance
Nobel Prize-winning professor William Sharpe once called it the "nastiest hardest problem in finance". He was talking about retirement planning, with the three future unknowns:
1. longevity;
2. inflation;
3. investment returns
making for a complex answer to what, on the face of it, seems like a simple question:
“Will I/we run out of money before I/we die?”
Financial theorist and neurologist William Bernstein had this to say:
"No one in his right mind would walk into the cockpit of an airplane and try to fly it, or into an operating theater and open a belly. And yet they think nothing of managing their retirement assets. I've done all three, and I'm here to tell you that managing money is, in its most critical elements even more demanding than the first two.!"
Pensions Freedoms
The challenge of creating a sustainable retirement income is a relatively recent one for U.K. retirees. In the 2014 Budget, Chancellor George Osborne announced radical pension freedoms, and these came into force in 2015, allowing retirees to flexibly access their Defined Contribution pensions from the age of 55 without tax penalties. Before pension freedoms, most retirees purchased an annuity, which was becoming less attractive with falling annuity rates in recent years.
These freedoms haven't always led to good client outcomes, with more than £30m lost to pension scammers between 2017 and 2020. Furthermore, according to FCA data, 40% of regular withdrawals were taken at an annual rate of over 8% of the pot value, which many consider unsustainable.
The '4% rule'
A popular 'rule of thumb' is that you can safely take an inflation-adjusted 4% from your pensions and investments each year (far lower than the 8% mentioned above) without running out of money. How might this rule apply to a typical retiree? Let's start by looking at the work of the man who undertook the original research!
William "Bill" Bengen
After studying aeronautics and astronautics at MIT, Bill Bengen worked in his family's bottling business before studying for his CFP and launching his financial planning business. In 1994, Bill published a landmark paper in the Journal of Financial Planning named 'Determining withdrawal rates using historical data'.
Bill analysed historical market returns from 1926 and determined the highest withdrawal rate as a percentage of the initial investment pot that could be withdrawn each year and adjusted for inflation in subsequent years without running out of money over a thirty-year retirement.
An example is shown below:
Starting pot: £1,000,000
Withdrawal in year one: £40,000
Inflation at the start of year two: 2%
Withdrawal in year two: £40,800 (£40,000 * 1.02)
Inflation at the start of year three: 3%
Withdrawal in year three: £42,024 (£40,800*1.03)
Repeat every year.
Bengen named this the Safe Withdrawal Rate (SWR) and found that the SWR was around 4% for his dataset.
Challenges
It's worth noting that Bengen never claimed that the 4% SWR was a rule, something he believed was invented by the media, according to an interview with Michael Kitces conducted in 2020. We will explore reasons why this 'rule' should not be taken as gospel, breaking our analysis down into three parts:
Data challenges of the 4% rule (this blog post)
To help us further investigate the '4% rule', we will use example clients David and Samantha Smith.
David Smith is 60 and has recently retired from ABC Chemicals. His wife Samantha is 55 and is also retired. Together, they have a retirement portfolio of £1,000,000 and are looking for a sustainable income of £40,000 per annum.
To simplify the example and help illustrate the key points, we will make the following initial assumptions:
Neither David nor Samantha will receive a state pension;
Taxation and taxation optimisations are ignored;
They do not plan to gift to their children or leave a legacy;
They are not expecting any inheritances;
They do not want to plan for potential care home fees;
They are not planning to downsize;
They are keeping expenditure assumptions simple. For example, they have chosen not to differentiate between early and late retirement spending;
They are not planning on purchasing a secure income (e.g., annuity) at any stage.
They have a 30-year retirement horizon (i.e. David dies at 90 and Samantha at 85).
They are paying no fees.
(Don't worry; we will address these points as we move through this series of blog posts.)
David and Samantha will be following the strategy outlined above:
Starting pot: £1,000,000
Withdrawal in year one: £40,000
Inflation at the start of year two: 2%
Withdrawal in year two: £40,800 (£40,000 * 1.02)
Inflation at the start of year three: 3%
Withdrawal in year three: £42,024 (£40,800*1.03)
We will be using a tool called Timeline, which looks at around 100 years of historical data and is used by many financial planners specialising in retirement planning. Timeline uses historical market returns across broad asset classes and inflation data to provide a range of outcomes. For example, had (older versions of) David and Samantha retired in September 1962, Timeline shows whether they would have run out of money before September 1992, given their planned portfolio, withdrawals and retirement horizon (30 years). However, the outcome would have been slightly different if they had retired a month later (due to differing inflation and market returns in the two non-overlapping months - September 1962 and October 1992). By analysing hundreds of such historical outcomes (twelve per year multiplied by the number of available years), we can obtain valuable insights into the future potential outcomes which David and Samantha might reasonably expect. We have used this tool for many years, and readers of Noel’s book, 'Planning for Retirement: Your Guide to Financial Freedom' will recognise the concepts and diagrams (and David and Samantha!).
Challenge One: US data – 4% “rule” doesn’t necessarily hold for the UK
Bengen used U.S. historical data for his original work, with a split of large-cap stocks (shares) and intermediate-term bonds. Bengen looked at outcomes from 0% to 100% stocks in 25% increments and found the sweet spot to be between 50% and 75% stocks.
For our examples, we will use a portfolio of 50% developed market stocks and 50% global bonds and historical U.K. inflation. As a reminder, we will use a 30-year retirement horizon (i.e. David dies at 90 and Samantha at 85).
As can be seen from the screenshot below, the news is less favourable for U.K. retirees compared to their U.S. equivalents, with David and Samantha only able to take out a maximum initial starting withdrawal of £33,700 (3.37%) in the first year (see the red bar in the screenshot below) without risk of running out of money further down the road vs around 4% for the U.S. equivalent.
Put another way, taking out 4% could see the money running out after around 21 years in the historical worst-case example, nine years short of the thirty-year retirement horizon.
(It’s worth mentioning that readers of Noel’s book may see slight differences in SWR rates and worst historical times vs. the numbers here. For example, the book has the worst case SWR of 3.42% occurring in the late 1960s, whereas, in this example, we have 3.37% around 1915 during World War One. Timeline has periodically altered/improved their data, which mainly explains the differences, but it’s worth emphasising that retirement planning is not an exact science, something that we will come back to (again and again!).
So why the difference between our UK and US investors? It all comes down to the underlying asset class returns and inflation in the scenarios. For example, the UK had a particularly bad time with inflation during the 1970s and the First World War.
Furthermore, Bengen's dataset started in 1926, whereas Timeline has data from 1915 (the worst historical time to retire in this example).
An example of how changing the asset class can impact sustainable withdrawal rates was when Bengen published subsequent research in 1996, adding small-cap stocks to the original two asset classes (large-cap stocks and intermediate-term bonds). He found an increase in the safe withdrawal rate (up to 4.3%) but cautioned that this might not be the case in the future.
Challenge Two: Asset allocation of 50% stocks and 50% bonds
In challenge one, we used a portfolio of 50% developed market stocks and 50% global bonds. Real-world portfolios may differ from this. For example, if we increase the equity content of the portfolio from 50% to 60%, the SWR increases from 3.37% to 3.5%, and the money lasts for an extra few years.
Conversely, if we reduce the stock content to 30% (and therefore, our portfolio contains 70% bonds), our SWR is now down to 3.09%
and the worst case has the money running out after around 18 years.
Note that it's not just stock/bond split in the portfolio that can impact our SWR, but also diversification. If we use 100% US market stocks for our portfolio rather than the broader developed markets, the safe withdrawal rate is reduced to 3.1%.
This emphasises the importance of diversification (both in terms of asset class and geography) and not having all of your eggs in one basket. The current bestselling tracker funds on a popular retail investment platform indicate this might not be the case.
Challenge Three: Limited historical data
Bengen published his paper in 1994, and his analysis covered 1926 to 1976, just 50 years. Bengen had data up to 1992, so it’s worth pointing out that Bengen wouldn’t have known how, for example, how the 1970 retiree would have fared as he didn’t have a full 30 years of retirement data at this point.
Almost thirty years have passed since Bengen’s original work, and Timeline has data going back to 1915, meaning that we now have nearly 80 years of retirement data available (Timeline currently goes to the end of 2021, so our "latest" retiree for a 30-year retirement would be someone finishing in 1991). While 80 years of retirement data might seem a lot, it is not even three of our 30-year retirement examples. We may experience far worse market conditions than in our historical data. For example, Morningstar research suggests that a sustainable withdrawal rate closer to 3% may be more appropriate due to elevated valuations. (Of course, safe withdrawal rates can only be determined in hindsight, so it's always worth bearing this in mind when reading predictions!).
As historical market data continues to improve in terms of quality (cleaner data with fewer errors) and quantity (historical data going back further in time), it can cause us to question our beliefs. For example, recent research by Edward McQuarrie has questioned the belief that stocks always outperform bonds over the long term. This again emphasises the point we covered in challenge one that the data we use for our retirement modelling may well be revised in the future.
Challenge Four: Starting stock market valuations are not taken into account
Picture the scene: David and Samantha retired in 2023 and started drawing down their investment pot at the rate of 3.3% (£33,000), using a portfolio of 50% stocks and 50% bonds. Based on historical data, they are reasonably confident they can continue to take inflation-adjusted withdrawals from their portfolio over the next 30 years without running out of money.
Fast forward two years to 2025. Another couple, let's call them Mark and Rebecca Jones, have identical investments and expenditure plans and are planning their retirement. Their investment pot, which was at the same £1m as the Smiths in 2023, is now at £850,000 after the market slump of 2024. They now feel they can only spend around £28,000 (3.3%*£850,000) without running out of money due to the lower starting investment balance
.
This paradox was described by leading U.S. financial planner Michael Kitces in a 2008 publication. Bengen built upon this research in a 2020 article, suggesting higher initial withdrawal rates (albeit on a smaller starting pot) should be considered due to lower (better) asset valuations (as you might expect after market falls).
Conclusion
Bengen's work was instrumental in putting a framework around retirement withdrawal planning. Examples of common thinking at the time included:
The average portfolio return is around 7% per year. Therefore, I can withdraw 6-7% a year without fear of running out of money.
In retirement, you cannot afford to take the risk of being invested in the stock market and should instead be invested 100% in bonds.
Bill changed all of this, and even though retirement research has evolved in the years since his initial publication, his work is still considered key in helping us build sustainable retirement portfolios. However, as we have covered in this article, thought must be given to how Bill's analysis applies to your individual retirement plan.
In part two of this series, we look at the data challenges of the 4% rule.
Want to find out more
If you want help with building a robust retirement plan, please get in touch.
About us
The team at Pyrford Financial Planning are highly qualified Independent Financial Advisers based in Weybridge, Surrey. We specialise in retirement planning and provide financial advice on pensions, investments, and inheritance tax.
Our office telephone number is 01932 645150.
Our office address is No 5, The Heights, Weybridge KT13 0NY.
Please note: This blog is for general information only and does not constitute advice. The information is aimed at retail clients only.