Beyond The 4% Rule

Home > Other > Beyond The 4% Rule > Page 3
Beyond The 4% Rule Page 3

by Abraham Okusanya


  Given the high-maintenance nature of the probability-based approach, it’s likely that the ongoing cost involved (for the platform, funds, and advice) defeats any likely benefit to the client.

  You may feel more comfortable with a safety-first approach if you’ve got a low risk appetite and are likely to wake up in a cold sweat worrying about your portfolio.

  In this case, the answer is straightforward – consider securing your baseline income through an annuity. And then if you still have any liquid funds left, consider pension drawdown for the rest (see Fig. 5: The hierarchy of retirement income needs in the safety-first philosophy). You can skip the rest of the book – unless of course, curiosity won’t let you.

  If the probability-based approach appeals, then you’ve a big job on your hands. How do you work out the sustainable withdrawal rate for your pension? What about asset allocation and charges? And for God’s sake, how do you work out how long you’re likely to live for?

  We address these questions in the rest of this book. We delve into the all the nuances around meeting retirement income needs with the probability-based philosophy. This includes an in-depth discussion of safe withdrawal strategies and the associated risks.

  Finke M, Howe J and Huston S (2011) Old Age and the Decline in Financial Literacy, Social Science Research Network, Available at SSRN: http://ssrn.com/abstract=1948627

  Cox, P., (2016) Helping consumers and providers manage defined contribution (DC) wealth in retirement

  Wade Pfau (2016), presentation slides at the FinalytiQ’s Science of Retirement Conference

  Pfau, Wade D. and Cooper, Jeremy, The Yin and Yang of Retirement Income Philosophies (November 10, 2014). Available at SSRN: http://ssrn.com/abstract=2548114 or http://dx.doi.org/10.2139/ssrn.2548114

  Bengen, William P (1994) ‘Determining Withdrawal Rates Using Historical Data’. Journal of Financial Planning 7, 4 (October): 171–180

  Bodie, Zvi and Treussard, Jonathan and Willen, Paul, The Theory of Life-Cycle Saving and Investing (May 2007). FRB of Boston Public Policy Discussion Paper No. 07-3. Available at SSRN: http://ssrn.com/abstract=1002388 or http://dx.doi.org/10.2139/ssrn.1002388

  Branning, Jason K, and M Ray Grubbs (2010) ‘Using a Hierarchy of Funds to Reach Client Goals’. Journal of Financial Planning 23, 12 (December): 31–33.

  Brancati C, Beach B, Franklin B and Jones M (2015): Understanding retirement journeys: Expectations vs reality. International Longevity Centre. November, 2015. Available online at http://www.ilcuk.org.uk/index.php/publications/publication_details/understanding_retirement_journeys_expectations_vs_reality

  CHAPTER 2

  The hidden dangers

  What’s the world’s deadliest animal to humans? Most people think of beasts with large teeth and fearsome reputations, such as the lion, rhino, wolf or the oft-cited hippo.

  It’s actually the tiny mosquito that does the most damage. It causes more deaths than virtually any other animal: responsible for about 725,000 human deaths annually. Only human beings themselves come close, with a tally of about 425,000. And what of man’s supposed best friend? Dogs kill about 25,000 people each year, almost exclusively because of rabies.

  Now compare these figures to those recorded for the so-called most dangerous animals: wolf (10), lion (100) or hippo (500). These fearsome beasts don’t even appear in the top 10 deadliest animals.

  So, what’s all this got to do with retirement income planning?

  It’s a classic example of our tendency as humans to misunderstand risk. It’s particularly relevant when you think about portfolio risk. When the new Pension Freedoms were announced by Chancellor George Osborne in 2014, many commentators focused on the risk that people could squander their savings on Lamborghinis and cruises. Other widely publicised risks include pensioners being scammed, or lured into questionable investments, be it car park schemes in China or forestry investments in Brazil.

  Fig. 6: The world’s deadliest animals

  The real risks for most retirees are more subtle. As more people go into drawdown, the real risk is that most investors don’t pay enough attention to that most silent of portfolio killers: the negative sequence of return. Clients risk being blindsided by these silent dangers while they worry about the more obvious risks, such as major market crashes. These subtle dangers are particularly relevant to retirement portfolios. This is because withdrawals amplify market risk in a way that’s obscured by the use of time-weighted returns and the averaging of long-term returns. So, many advisers and clients don’t notice until it’s too late.

  Historical evidence of sequence risk and why it matters in retirement income planning

  Sequence risk is perhaps the most significant risk to maintaining a lifetime income for people following a probability-based retirement income approach.

  When I think of sequence risk, I think of the words of the legendary Eric Morecambe, ‘I’m playing all the right notes, but not necessarily in the right order.’

  Sequence risk is the risk that the order of investment returns is going to be unfavourable. This risk exists at accumulation stage, but it’s amplified by withdrawals from a portfolio during the retirement stage. If you get good returns in the early part of retirement, you’re unlikely to run out of money with a sensible withdrawal rate of say 4% or even 5%. If you get poor or even mediocre returns in the early part of retirement, the technical term is, as I’ve said, ‘buggered!’

  I also call sequence risk ‘pound-cost ravaging’. Returns in the early period of retirement have a disproportionate effect on the overall outcome, regardless of long-term returns over the entire retirement period. And if it’s not properly managed, pound-cost ravaging can cause untold havoc.

  To illustrate this, let’s look at evidence from actual historical data over the 117 years, between 1900 and 2016, using the Dimson, Marsh and Stratton Global Investment Returns Database.

  In this research, we need to explore the relationship between historical sustainable withdrawal over a 30-year period and real returns over each of the three decades of retirement within that 30-year period. For this purpose, sustainable withdrawal rate is the percentage of the initial portfolio you could take from a portfolio and subsequently adjust for inflation, without running out of money over 30 years.

  So, for a £100,000 portfolio, a withdrawal rate of 4% gives you an annual income of £4,000 in the first year. The £4,000 is then adjusted for inflation every single year over 30 years, regardless of portfolio size during those subsequent years. The withdrawal rate is expressed as a percentage of the portfolio balance only in the first year of retirement.

  The Test

  I’ve expressed the sustainable withdrawal in monetary terms (rather than percentages) using the term sustainable income, based on a £100,000 portfolio. The sustainable income is the maximum annual inflation-adjusted income you can take if you want to run down the portfolio balances to zero at the end of 30 years.

  For this, I looked at every 30-year period between 1900 and 2016, inclusive. So, the first 30-year period runs between 1900-1929, then 1901-1930, 1902-1931… and the final 30-year period runs between 1986-2016. This gives 87 scenarios.

  The portfolio is composed of 50% UK equities and 50% UK bonds, rebalanced annually.

  I’ve not applied a fee because the point is to examine the relationship between returns and sustainable withdrawals. But when I re-tested with fees applied, the findings were the same.

  Finally, I examined the average real return in the first, second and third decades of each 30-year retirement period.

  The Result

  Fig. 7 opposite shows the annual sustainable income (inflation-adjusted) against the average real return in the first decade of retirement.

  In Fig. 7 you see a very strong correlation between return in the first decade and the sustainable income for the overall 30-year period. But look at the next chart. It shows the annual sustainable income (inflation-adjusted) against the average real return in the second and third decades of reti
rement.

  As you can see in Fig. 8, the relationship between the sustainable income over a 30-year period and the return in the second and third decade isn’t all that strong.

  Fig. 7: Sustainable withdrawal vs. average real return in the first decade

  Fig. 8: Sustainable withdrawal vs. average real return in the second and third decades

  Fig. 9: Sustainable withdrawal vs. average real return in the first, second and third decades

  Fig. 9 is the earlier two charts combined. It shows the annual sustainable income (inflation-adjusted) against the average real return in the first, second and third decades of retirement.

  This result shows a strong correlation between sustainable income over any 30-year period and the average real return in the first decade of retirement. In years where the average real return for the first decade of retirement is high, the sustainable income over the 30-year period tends to be high, and vice versa.

  Take, for example, a 30-year period starting in 1921. The average real return in the first decade was nearly 14.4%, compared to 3.9%pa in the second decade and 3.6% in the third decade of retirement. Someone starting their retirement in 1921 could have enjoyed an annual income of £12,000, adjusted for inflation over the subsequent 30 years. (A whopping withdrawal rate of 12%!) Why? Because they had a great first decade. The return in the second and third decades didn’t matter as much.

  Take another example: a 30-year period starting in 1969. The average real return in the first decade was a meagre 1.44%, compared to 10.2% in the second decade and 11.7% in the third. Never mind fabulous double-digit returns in the second and third decades of retirement, the sustainable income over the 30-year period was no more than £4,533pa (or a withdrawal rate of 4.5%).

  Fig. 10: Correlation between sustainable income and return

  Fig. 10 shows the correlation between average real return in each decade of a 30-year retirement and the sustainable withdrawal over the entire period.

  Fig. 10 shows an 83% positive correlation between the sustainable withdrawal rate over a 30-year period and the real returns in the first decade of retirement! This compares to a 26% correlation between sustainable income and average real return in the second decade and a correlation of – 33.2% (a negative correlation) to the average real return in the third decade of retirement.

  Coefficient of determination

  We can explore the relationship between withdrawal rates and returns a little further. This time, we look at the coefficient of determination (R squared) between return in each decade of retirement and the sustainable withdrawal rates for the entire 30-year period.

  R squared, or R2 is a number that indicates the proportion of the differences in the dependent variable that is predictable from the independent variable(s).

  In this case, we want to explore how much the differences in sustainable withdrawal rates can be explained or predicted by returns in the first, second and third decade of a 30-year retirement.

  As Fig. 11 below shows, 69% of the differences in sustainable withdrawal rate can be explained or predicted by the inflation-adjusted return in the first decade of a 30-year retirement. Returns in the second and third decades only explained 7% and 11%, respectively. The average return over the entire 30 years only explained 42% of the variability in sustainable withdrawal rates over the same 30-year period.

  The conclusion?

  Return in the first decade of retirement is the main driver of sustainable income over the entire 30-year period. And the way you draw income from your retirement portfolio needs to be carefully managed, particularly in the early part of retirement.

  Sequence risk is no bogeyman. Empirical data shows that it does exist. It’s vile and dangerous.

  Sequence risk is the primary reason someone drawing income from their portfolio is likely to run out of money, even at a modest withdrawal rate of 4% or 5%.

  The question is not whether capital markets will deliver a decent average return over your retirement period. It’s whether the returns will come when you need them most – in the early part of your retirement.

  Fig. 11: Coefficient of determination for sustainable withdrawal rates.

  Accordingly, it’s crucial to understand and manage the havoc sequence risk can cause to a retirement plan.

  Other than longevity risk, I can’t think of any other retirement risk that’s as potent as sequence risk. In fact, if a retiree gets a good sequence of return, longevity risk won’t be an issue for most people. Yes, that’s a bold statement, but one I can’t emphasise enough.

  It’s important that we use the right tools to model sequence risk. Traditional deterministic tools do a terrible job of this. A cash flow model assuming a net return of 3%pa will show that a withdrawal rate of 5% inflation-adjusted is sustainable over a 30-year period. In reality, that plan would have failed in over 50% of actual historical scenarios.

  We need to prepare clients for the possibility that they might experience a poor sequence of return in retirement. We need to have an action plan if that happens. And, we need to understand the withdrawal strategies that can mitigate sequence risk.

  This is a good thing. In a sense, sequence risk is a blessing in disguise; since the early part of retirement has so much impact on the overall outcome, we can put together a framework and help clients manage this stage. We can do it with confidence that the later part of retirement will be a little less stressful.

  CHAPTER 3

  What doesn’t work

  Before we delve into how best to manage the key challenges in retirement income planning, I want to explore several common practices and why they fail woefully to address the problem. In fact, some of these approaches may well be downright harmful.

  It ain’t volatility, stupid!

  The movie The Big Short opens with the quote…

  ‘It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.’ – Mark Twain

  The quote sums up the dangers of thinking you know something that isn’t actually true. There’s only one problem: Mark Twain probably didn’t say it. Oh, the irony.

  In a misguided attempt to manage sequence risk in retirement income portfolios, a great deal of effort is devoted to managing volatility. This has resulted in a vast number of volatility-managed funds and model portfolios, peddled by asset managers to advisers and clients in retirement.

  The trouble is, managing volatility in a retirement portfolio is, for the most part, a red herring. It’s a bit like bringing a knife to a gun fight.

  In a retirement income portfolio, sequence risk is your enemy. Not volatility. Sequence risk is often confused with volatility within the financial industry. I fell into the same trap in my early days of researching retirement income strategies. But now I’ve looked at the empirical data more closely, I realise that the two are related, but they’re different.

  Volatility is the day-to-day movement in your portfolio. It’s measured using standard deviation – the amount your portfolio return deviates from the average over any given time period.

  Sequence risk on the other hand relates to the order of portfolio returns.

  Volatility has very little impact on the order of returns. Two portfolios might have the same average return (mean) and volatility (standard deviation) over a given period of time. But if the order (sequence) of returns is different, then the sustainable income will be different.

  Got it? OK.

  The Test

  There’s an over 80% correlation between returns in the early part of retirement and the sustainable withdrawal rate over a 30-year retirement period. But, I find little correlation between volatility – the industry de facto measure of risk – and sustainable income.

  It doesn’t matter what period you look at. Income is as likely to be high and sustainable with a high-volatility portfolio as it is with a low-volatility portfolio.

  Fig. 12: Correlation between sustainable income and return/volatility

  Ther
e’s less than a 10% correlation between volatility in the first decade of retirement and the sustainable income over a 30-year period. And I only found a 32% correlation over the entire 30-year period.

  Sustainable withdrawal rate has little to do with volatility.

  Coefficient of determination

  Just as we did with returns, we can explore the relationship between volatility and sustainable withdrawal rate by looking at the coefficient of determination (R squared) between the two.

  Again, the sustainable withdrawal rate is the dependent variable. And R squared tells us how much of the difference in sustainable withdrawal rates is explained or predicted by volatility.

  The result shows that volatility over the first, second and third decades of retirement explains less than 2% of variability in sustainable withdrawal rate. Indeed, the average volatility over the entire 30-year period only explained around 15% of differences in sustainable withdrawal rates.

  My findings are consistent with Kenigsberg9 et al. (2014) who noted: ‘To further test whether sequence of return (SOR) risk in the first decade overshadows volatility risk in determining the maximum Sustainable Withdrawal Rate (SWR), we ran a regression analysis on historical maximum SWR (as a dependent variable) first with the first decade’s real return and then with the first decade’s volatility as possible independent variables, using a 50/40/10 portfolio and the historical returns for the 702 complete 30-year periods between January 1926 and May 2014. We found that using the first decade’s real return as the independent variable produces an R2 value of 73%, whereas the regression using the first decade’s volatility yields an R2 of only 1%. Although the use of overlapping periods may complicate this statistical approach, we think the analysis at least suggests a greater influence from sequence than from volatility.’

 

‹ Prev