My previous article on the dangers of sequence of return risk in a drawdown portfolio appears to have struck a chord with readers of this blog: it prompted quite the response from many planners wanting to know the issue can be addressed.
The truth is there is no silver bullet for removing this risk, but all is not lost. The latest research in the ‘safe withdrawal rate’ (SWR) offers some clues on how to mitigate against it…
Firstly, I’d like to reiterate why the starting point for the vast majority of planners is to adopt Monte Carlo modelling in helping clients make decisions about the SWR from their portfolios.
This is a valuable exercise that helps planners work out the extent to which a client’s portfolio is vulnerable to sequencing risk in the first place. Using probability-of-failure in a Monte Carlo model, a planner can carry out a forward-looking evaluation of clients’ exposure to sequencing risk and assess how likely it is they will need to adjust their withdrawal level to avoid running out of money.
For instance, if the model suggests there is a 10% probability a plan will fail if the level of withdrawal is maintained, that plan would be considered less vulnerable than one for which a 30% probability of failure is produced. This equips a planner with the information to have a meaningful dialogue with a client about the possibility of reducing their level of withdrawal in the future.
If a plan is highly vulnerable to sequencing risk (generally, this means the probability of failure is above 20% or 30%), there are a number of approaches planners might consider for managing this in a retirement portfolio.
Of course, a starting point would be to consider if the client is prepared to live off a lower level of income from their portfolio. If this is acceptable to client, then the probability of failure may be reduced to an acceptable level.
But, from a portfolio management point of view, there are a number of approaches that advisers might want to consider when looking to manage sequencing risk in drawdown portfolios.
Cashflow Reserve Ladder
The traditional method of managing sequencing risk is to keep about two years’ worth of cashflow in cash or near-cash assets like high quality short-dated bonds. This approach means clients can avoid selling equities at the worst possible time.
This is because, in the event of a prolonged market downturn, a client can draw from the cash component of their portfolio for two years or so, and then from the bond allocation, which may last another five to ten years in a 60/40 portfolio.
By the time the client gets around to making withdrawals from the equity allocation, the market will (hopefully) have recovered.
The downside to this approach, of course, is that it screws up the asset allocation within the portfolio and only works if the portfolio isn’t rebalanced periodically back to the original allocation. This means that, as bonds within the portfolio are liquidated to pay for income, the overall allocation to equities increases. This may be something that advisers and indeed clients aren’t prepared to live with.
It also goes against the message that advisers may have preached to the clients during the accumulation stage – the importance of consistent asset allocation and rebalancing.
Rising Equity Glide Path Approach
Research by US-based financial planner Michael Kitces and Professor Wade Pfau suggests a more systematic way to managing sequencing risk, by starting with a lower equity allocation and increasing it gradually over the first decade or so in retirement.
The idea is, where a 60/40 portfolio would have normally been recommended for the client, the planner would instead start with a 40% equity allocation and increase the allocation gradually each year until it gets to 60%.
Though this approach seems counterintuitive – as the conventional view is that equity allocations should decrease as clients age – it actually works in a similar way to the cashflow reserve ladder discussed earlier.
In my view, the rising equity glidepath approach makes a lot of sense. Beginning with a lower equity allocation in the early years reduces the volatility within a portfolio at the very time when it is most vulnerable to sequencing risk.
Increasing the equity allocation over the following ten years also gives the portfolio a chance to negate the effect of reverse pound cost averaging; in other words, if equities do fall during the first decade in retirement, the rising glide path means clients benefit from buying low.
Of course one might argue that, if equities do well in those early years, lower allocation to equities will mean the client potentially loses out. But that shouldn’t be too much of a worry because the goal is to provide income in a sustainable way, not necessarily to maximise growth.
Low-Beta-High-Tilt
Another approach that planners can consider in managing sequence risk is to consider using lower equity allocations, and to tilt them towards value and small cap equities.
In their book Reducing the Risk of Black Swans: Using the Science of Investing to Capture Returns With Less Volatility, authors Larry Swedroe and Kevin Grogan suggests holding a lower equity allocation within a portfolio, but tilting it heavily towards value and small cap.
So, for instance, where a 60/40 portfolio would normally have been recommended, planners can consider recommending a 50% portfolio allocation, of half of which will be in UK and global small cap and the other half in value equities.
This idea relies on the ample evidence suggesting value and small cap equities have higher risk-adjusted returns than growth and large cap equities respectively.
But, rather than tilting your portfolio toward value and small caps to boost returns, it flips the idea on its head by doing so instead to reduce risk within the portfolio. By keeping the equity allocation low and tilting heavily toward value and small cap, the overall risk within the portfolio would be lower than what the client would otherwise have been recommended.
To put Swedroe’s research to the test, I constructed two portfolios with 60% and 50% global equity allocations respectively. ‘BetaBasic 60′ has 60% invested in the MSCI World Index and 40% in the Dimensional Global Short Dated Bond Index. ‘BetaBasic 50′ has 50% in MSCI World and 50% in the Dimensional index.
Then I constructed two ‘high tilt’ portfolios: ‘BetaTilt 50′ invests 25% in global value and 25% in global small cap, while ‘BetaTilt 40′ invests 20% in global value equities and 20% in global small cap.
With an initial investment of £100,000, I applied a cashflow of £500 per month withdrawal from each portfolio (£6,000 a year). I ran the portfolios over a period of 10, 20 and 30 years to the end of August 2014. I also ran a few scenarios assuming the client started making withdrawals from their portfolio just before the major bear markets of Sep 1992 (Black Wednesday), October 2002 and Sept 2008 (credit crunch), again running the withdrawal up until the end of August 2014.
If the theory is right, then the ‘BetaTilt 50′ portfolio should do better than ‘BetaBasic 60′ and ‘BetaTilt 40′ should do better than ‘BetaBasic 50′. Let’s see…
As it turns out, the low-beta-high-tilt portfolios did outperform in every single case, and the longer the period, the bigger the difference.
Of course, the research is backward-looking, and the usual caveats must apply. Nonetheless it relies on fundamental evidence that value and size ‘premia’ do exist in global stock markets. As long as that theory continues to hold, one can expect a low-beta-high-tilt portfolio to be an effective strategy in managing sequencing risk.
In addition, plenty more research has found that value equities tend to do better in bear markets and that indeed the value premium is generated during the periods of, loosely speaking, ‘bad times.’
In summary, I’m aware that a number of product ‘solutions’ to managing sequencing risk are out there and, in the run up to Pension Freedom Day (6 April), we can expect a plethora of new offerings to flood the market.
These products tend to have three main features in common – high cost, the use of derivatives to provide so called ‘downside protection’, and mind-boggling complexity. Advisers should be wary and question why these would work given the documented failure of products with high costs and complexity.
Interesting and robust stuff Abraham, although it amazes me that financial planners needed the issue pointing out in the first place… The thing your possible solutions ignore, though, is that it is quite possible that it will be the non-equity element of portfolios that causes the problems over the next 10 years if 10 year gilt yields of 2.1% and 20 year gilt yields of 2.7% turn out to be pure fantasyland. You won’t be surprised to hear me say it, but there has to be a case for than just a mechanistic approach to portfolio rebalancing alongside the Cashflow Reserve Ladder….
Your research is very backward looking. Wherever you go in the past you will find higher interest rates, so your short dated fund would come up with the goods i.e. some investment return.
I am not a fun of value investing, mainly for its high beta and long periods of underperformance. I would prefer a replacement in your plots with a high yield index, low volatility and quality index (or profitability as Dimensional names it). MSCI has a good study on the level of outperformance such a combination could offer.
Why I would exclude value stocks. Value investing is very dependant of the economical cycle, and also on debt and borrowing. Although interest rates are low, access to debt is pretty restricted for value stocks. We are all in high debt: Governments, companies and people – this is what has changed for Value companies.
Small stocks is something that a de cumulatiion client should always have in their portfolio, but no more than 10%.
I am not sure if in your research you deducted some fees as well, I would say around 1.5% per annum.
Another fascinating article Abraham. Great work! One can agree or disagree with your views (I mostly agree) but it certainly provides food for thought.