-----------------------------
- We find one very powerful conclusion: that smoothing the returns on individual assets by simple absolute momentum or trend following techniques is a potent tool to enhance withdrawal rates, often by as much as 50% per annum!
- of even greater social relevance is that it removes the ‘left-tail’ of unfortunate withdrawal rate experiences, i.e. the bad luck of a poor sequence of returns early in decumulation. We show that diversifying assets over time by switching between an asset and cash in a systematic way
- Much of the retirement literature to-date, in exploring the relationship between sustainable withdrawal, longevity and portfolio investment returns, makes the simplifying assumption of constant year-by-year returns, for example, see Estrada (2017b), footnote 9. This approach is clearly wrong in omitting the role of path dependency.
- diversification across asset classes is no longer enough - quite simply, the world is too volatile.
- We show that diversifying assets over time by switching between an asset and cash in a systematic way is potentially more important for the retirement income experience than diversifying one’s portfolio across asset classes.
- can we do better than using (long only) bonds and equities in the portfolio? Crucially, what is the benefit, if any, of switching to cash at times in an effort to reduce drawdowns and smooth returns? Will such a systematic smoothing rule such as trend following offer a superior decumulation journey? The answer is a most persuasive ‘yes’.
- It is now increasingly recognised that trend following (otherwise known as Absolute Momentum or time-series momentum, see Moskowitz et al, 2012 and Hurst et al, 2017), has delivered ‘strong positive returns and realized a low correlation to traditional asset class returns for more than a century’, Hurst et al, (2017).
- The robustness of this performance over returns since 1800 has been recently demonstrated by Baltussen et al (2019). They also show that the statistical significance of this performance is robust to the development of large numbers of potential factors and data snooping.
- In addition to raising average returns above a long-only, holding strategy, an Absolute Momentum strategy also typically improves a portfolio’s risk adjusted returns (Hurst et al, 2017 and Faber, 2007).
- whilst bonds have less volatile returns, they remain subject to maximum drawdowns of a similar order of magnitude to equities. The maximum (real) drawdown for portfolios with 80%, 90% and 100% bonds are, respectively, 49%, 53% and 67%! This serves to undermine the core argument in favour of TDFs.
- Trend following, or Absolute Momentum, thus allows investors to take on a greater weighting in stocks for the same risk or a reduced level of risk for the same asset allocation…This is potentially a powerful practical result in that diversifying between equities and bonds would seem to offer an inferior investment solution than sticking solely with equities and moving in and out of the asset into cash when the trend signals so indicate [emphasis in the original]
- It would thus appear that the withdrawal experience in general is enhanced by trend following portfolio adjustment, both for the representative ‘median’ experience, and, importantly, the lower bound experiences. [emphasis added]
- Interestingly, the fabled ‘4%’ Sustainable Withdrawal Rate over a 30-year decumulation period of Bengen (1994) and others is not supported by results in Table 2, with a minimum withdrawal rate of 3.57% for a representative portfolio; this compares to 4.43% for the smoothed version, comfortably above the 4% threshold
- It is avoiding these low PWR outcomes that is the premise for glidepath strategies. If one adopts trend following, however, the probability of observing the poor PWRs is substantially reduced such that there is little reason to transition to anything more conservative than a 60-40 TF portfolio. Trend Following (or similar smoothing of returns) is the key to reducing Sequence Risk and enhancing the decumulation experience.
- In anticipation of our proposal later in this paper and elsewhere, we would advocate subjecting all asset classes to smoothing by trend following (see Faber, 2007, Clare et al, 2016), rather than just adding a CTA component to an otherwise long only conventional target date fund allocation.
- It is clear from this discussion that there is no consensus as yet on de-risking both towards and during retirement.
- The application of an absolute momentum or trend following overlay on the assets in a drawdown retirement portfolio reduces maximum drawdowns across asset classes, countries and historical periods (see Clare et al, 2017 and 2019, Faber 2007), reducing sequence risk and potentially improving the retirement journey, but why not use derivatives to achieve similar ends? But as AQR emphasise, “Unfortunately, in the typical use case, put options are quite ineffective at reducing drawdowns versus the simple alternative of statically reducing exposure to the underlying asset.” (Israelov, 2017, p 1,)
- hanging portfolio allocation between the risky asset and cash (so-called “divestment”) will give a better result than buying put protection, unless the drawdown coincides with the option expiry cycle. In fact the paper suggests that investing 40% in equity and 60% in cash has given similar returns as the protected puts strategy but with under half the volatility and much improved peak-to -trough drawdown experience.
- Ilmanen (2016)…refers to index put buying as protection for equity portfolios as (looking at historical data) “roughly a minus one Sharpe strategy.”
- Trend following has a clear positive Sharpe ratio, and it has done well in most of the historical bear markets over the past hundred years. Direct hedging is costly and only delivers value when combined with the ability to predict short-term market crashes and unwind the positions quickly after a crash.
- there is an industry-wide aversion by advisers to recommending holding large proportions in cash even when the evidence is stacked up in its favour: they feel they can hardly charge fees for advocating cash holdings whereas in fact this form of timely diversification may well be the only sensible game in town.
- We find that the withdrawal experience from conventional equity/bond portfolios can be substantially enhanced by applying these smoothing techniques, effectively creating a much higher lower bound for withdrawal rates in the 20th century (4.4% pa versus 3.6% p.a. for the popular 60-40 allocation with withdrawals lasting 30 years), and also improving withdrawal rates for the “average”experience (6% v 6.4%, see Table 2). In all cases strategies targeting a set volatility (as is popular with retirement consultants) allows far higher equity proportions and consequent withdrawal rates for a given volatility level (or “bucket”).
- Our findings for varying decumulation period lengths suggests that unless one is very risk averse it would only appear sensible to apply conventional glidepath strategies towards the last few years of decumulation, and then only move towards a 50-50 equity-bond allocation.
- Israelov (2017) emphasises the possible advantages of devoting a percentage of the portfolio to trend following assets which are robust to large equity drawdowns and are largely uncorrelated with conventional assets. They have in mind a CTA “hedge” fund, for example, and possibly allocating to it up to 15% of wealth. This certainly improves the experience in reducing drawdowns (and hence Sequence Risk) but begs the question as to why stop at 15% - why not consider the advantages of trend adjustment for the whole portfolio? [RH aside: I asked of myself this very question in a prior post but I also, after that post, came to the conclusion that the possibility of extremely long non-performance intervals for trend following would make it very difficult for managing retirement portfolios where one might be planning based on a sub-20 year “intermediate” horizon (say from 70-80 where 80 might be an optimal age for annuitization). Over very long horizons I agreed that trend in isolation looks like a good deal, basically it looks like a long term negative-premium (positive return) put which would be nice thing for a consumption portfolio]
No comments:
Post a Comment