A Sequence of Errors?

There has been much talk recently about “Sequence Risk” (a more detailed description of the opposing views can be found here and here ), as both sides ponder the Safe Withdrawal Rate (SWR) for retiring investors, and the effect of market returns on Retirement Pot longevity.

On one side of the argument lies Abraham Okusanya (of FinalytiQ ) , who argues that the early years of Portfolio Drawdown largely determine the fate of the Investor-a major loss early on can have a devastating impact on the Fund, as it requires a bigger subsequent recovery (in percentage terms) to get back to the previous level of wealth. ( A 20% portfolio decline requires a 25% gain just to get back to “break even”). This is referred to as Volatility Drag. The timing of those falls, with the greater effect felt on Portfolios if these declines occur early on in the life of the Drawdown period, is Sequencing Risk-subsequent good returns do not have the same positive effect as those returns are realised on a smaller portfolio size. The risk is therefore that negative market returns occur early on, with a disproportionate effect on the overall outcome. If the investor is simultaneously withdrawing Income, the negative effects are amplified.

On the other side of the fence sits Paul Resnik (of FinaMetrica) and Peter Worcester (an Australian Actuary), who argue that this is all rather over-done, and that an analysis of Australian historical data* suggests that there is no material difference between the returns generated by a 40% weighted Portfolio and an 80% weighted version; thus , there is little additional return for the greater exposure to Growth Assets, and no significant difference between Portfolio longevity in the”Poor” and “Worst” case scenarios in either scenario. The global correlation between International markets means that returns are consistent across Australia, the UK and the US. The reduction in equity exposure hasn’t altered the Poor and Worst returns, but HAS impacted the Best and Average Returns, which is the worst of both worlds. Therefore, Sequence Risk appears to be an unnecessary concern: the bigger concern is the opportunity cost of NOT being invested in growth assets.

There is a clear danger here of getting bogged down in details, thereby missing the wood for the trees. There can surely be no doubt that Sequence Risk does occur, but not for every client, and not to the same extent. The effect of poor initial returns will be significantly different for two Investors with the same size pension pot, having the same rate of return and risk profile if the start of withdrawal begins in October 1986, or September 2014, compared to one year later. We cannot know (ex-ante) the effect of war, epidemics,terrorism etc. on asset returns, but we do know that returns are not distributed “normally”. In this instance, risk can be seen as both a function of probability AND impact-we take out insurance for our house not because we expect (or want), to get paid out, but for fear of the high impact of the event itself on our future selves. This is “tail risk”, and here Advisors have an important role; they need the tools and resources ( which we aim to provide) to discuss and explain to Clients these risks, so that they are aware of them and are not taken by surprise- this decreases the chance that they will over-react to short-term events. If Jack Bogle’s recent warning about long term returns is in any any way accurate, then these pitfalls become crucial to avoid, since 4% p.a. returns leave very little margin for error.

One of the inherent problems for Advisors is that there are few ways of modelling the likely path of returns. Conventional cash flow modelling techniques assume linear paths in both Asset returns and Inflation, whereas we know that life is rarely that simple. Higher Inflation at the start of the retirement period implies higher base levels of withdrawal, upon which later inflation compounds, whilst Average Returns over a 20-30 year period tell us nothing about the order in which these returns occur. They thus give an illusion of certainty where there is none. The investment plan needs to be tested over as wide a range of scenarios/outcomes as possible, which is why Monte Carlo simulations have gained in popularity.

“The reason Monte Carlo simulations are being used more frequently (by financial planners), is because they do a better job explaining the potential outcomes versus time-value-of-money calculations, such as future value. The problem with a future value calculation is that it treats the outcome as certain, while in reality, and especially with the markets, nothing is certain. A Monte Carlo simulation provides a more ‘colourful’ perspective of the range of potential outcomes given the expected return and volatility of a portfolio” David Blanchett, Head of Retirement, MorningStar.

By running thousands of iterations, using specific parameters, one can determine the probability of specific outcomes (such as running out of money). Ideally, it will not only determine the chance of failure to meet the investment objectives, but also the magnitude of failure (i.e by how much). This enables corrective action to be taken, such as adjusting withdrawal rates, portfolio allocations or a combination of the two. Of course, Monte Carlo simulations are only as good as the assumptions built in, (more things CAN happen than actually do happen) and thus are only useful as a starting point for a conversation; they are by no means ideal, but like Churchill’s view on democracy, they are the best we have.

Investing involves uncertainty. The Advisor can help manage this process, but there are few viable ways of eliminating it altogether, and anyway, there would be no equity premium if there were no risk. One can mitigate these risks, by transferring some of them to others, which is what Diversification does, but it is not a silver bullet. All client’s circumstances are going to be different, and thus the Advisor, properly equipped, can perform a vital service-improving the likelihood that the client doesn’t run out of money.

* Interestingly, Challenger, the Australian Investment Manager appear to have come to the opposite conclusion using Australian data between 1979 and 2011. It would be interesting to see data relating to the Japanese experience since 1990. A prolonged Bear market would inevitably change the situation immensely- what the Western Markets (Australia, UK and the US) have experienced over the last 25 years is very different to that of Japan’s Topix Index, and may be framing the debate. 3 separate draw downs of 63%, 55% and 54% between 1989 and 2003, with a cumulative loss of 60% over that 14 year period might concentrate minds rapidly..