Sequence of returns risk?

newkeen

Re: Guyton's sequence of return risk reduction strategy

The retiree starts with 5.5% withdrawal rate of £5,500 on £100,000, but should not allow the future withdrawal rate to exceed +/- 20% of 5.5%. If the withdrawal rate crosses this threshold the amount withdrawn is reduced or increased by 10%. So if the portfolio fell from £100,000 to £81,000, a £5,500 withdrawal would be 6.8% (exceeding 5.5% +20%) so £5,500 is cut 10% to £4,950 and brings the withdrawal rate back within tolerance to 6.1%.

Makes sense.

But what if the portfolio value had fallen further such as a 2008 scenario and the £100,000 portfolio was now £70,000?
Reducing the £5,500 10% to £4,950 would still leave the withdrawal rate above the +20% of 5.5% threshold (£4,950 of £70,000 = 7% withdrawal rate).
In this scenario does Guyton's strategy require a greater than 10% cut in withdrawal down to £4,600 to bring the withdrawal rate down to the upper threshold of 6.6%?

I cannot see any reference to this scenario in any of the information.

jamesd

The changes are per year changes and if there had not been recovery by the second year decision time a second reduction would be made. No bigger first year reduction.

newkeen

Thanks jamesd,

On your thread 'Drawdown:safe withdrawal rates' your research includes (among many others) Kitces,

Under a section for improving the (drawdown) rules, one option presented is 'Hold a year of investment income in cash' What is the basis for this (there is no link to supporting research) and following your links to Kitces other research, Kitces seems unconvinced of the benefit of sacrificing equities allocation to hold cash. The suggestion is that simple rebalancing and withdrawing from a rebalanced portfolio gives better sustainable returns in almost all conditions.

https://www.kitces.com/blog/Are-Cash-Reserve-Retirement-Strategies-Really-Necessary/
https://www.kitces.com/blog/research-reveals-cash-reserve-strategies-dont-work-unless-youre-a-good-market-timer/

1 year would probably be around 4% of the portfolio, which for a retiree with a 70% equities weighting would mean 5.7% less equities. So probably not so much of a drag on performance.

But what is the case for holding 1 years cash?

jamesd

Basis was other research supporting that approach. See for example The Benefits of a Cash Reserve Strategy in Retirement Distribution Planning mentioned by Pfau and his own series on the improved results available from equity release (reverse) mortgages.

Pragmatically, some cash is needed to smooth out the income coming from investments so a year's investment income which does that isn't likely to have a large negative effect and provides significant added convenience via things like being able to set up a standing order or direct debit to pay the regular income.

newkeen

Thank you for that and providing the sources. Your other drawdown thread is very useful. I am now using cfiresim and testing figures using Guyton-Klinger as the spending plan.

I also came across further research Kitces: Dynamic Asset Allocation and Safe Withdrawal Rates https://www.kitces.com/april-2009-issue-of-the-kitces-report-dynamic-asset-allocation-and-safe-withdrawal-rates/

This suggests that the retiree can benefit when their starting withdrawal rate takes account of the current market PE10 ratio, this being a fair predictor of returns over the following period and therefore a potential flag for poor sequence of returns risk. Equally a favourable PE10 allows for a 0.5% or even 1.0% higher starting rate of withdrawal without undue risk. The rate of withdrawal can then be increased around 0.5% more using a dynamic equities allocation rebalancing according to PE10 increasing and decreasing equities allocation +/- 20% annually when the PE10 moves above/below historical valuation thresholds. The result can be a further uplift of around 10% in income for the retiree.

If correct, it suggests that safe withdrawal rates of 5% and 6% are totally sustainable. Unfortunately I don't see an option on cfiresim to run the dynamic PE10 equities allocation rules as outlined in his research, or to combine them with Guyton-Klinger.

jamesd

Right, cfiresim lacks a range of options but I mention it because it is able to at least use Guyton-Klinger.

PE10 - the ten year cyclically adjusted price/earnings ratio - does allow a range of responses, including Guyton's approach and since not being as heavily in equities means a lower loss due to a drop in equity values it's potentially very useful. I'm using it more broadly for guiding investment decisions. In my case since I'm not keen on bonds either I've moved towards P2P instead.

For those not familiar with it, PE10 has been found to be a working predictor of future performance in every market so far examined, which includes the major world markets for equities.

What you're doing is part of what I intended: that thread is a starting point and anyone interested can find way more material of interest, as you are... You might well end up with a different overall approach to the one I currently suggest, just as I'm broadly paying attention and will modify that suggestion if that seems sensible, though maybe after quite a lengthy period of consideration.

gadgetmind

jamesd wrote: »

For those not familiar with it, PE10 has been found to be a working predictor of future performance in every market so far examined, which includes the major world markets for equities.

Interestingly, I've never seen the same approach used for bonds other on very specialist bond investor sites. The recommended approach for PIs is "hold them and rebalance, because that's what you do."

simonfitba

Bit of info about using a cash buffer in retirement here...

http://myrichfuture.com/how-to-retire-early/cash-buffer-will-save-you-in-retirement/

gadgetmind

simonfitba wrote: »

Bit of info about using a cash buffer in retirement here...

Interesting that this article gives a figure of three years of cash buffer as that's what we have in place, mostly index linked. Having this 3x essentials pa as cash does mean that we're 10% cash (we're at 30x essentials in retirement pot) which some might argue is a drag on returns.

However, index linked cash is an asset that gives a lot of peace of mind, and I'm not parting with it easily! If we don't need to touch it, then it's there maintaining its purchasing power. If we do need to dip in, then belts will be tightened at the same time, and it will be replenished once things )hopefully!) perk up again.

If they don't perk up within 3-4 years, then we cut back further and consider some part time work to cover maybe 1/2 of the bills.

jamesd

Rather than three years you might consider using income units instead of accumulation units and having the distributions paid into the buffer account. That would considerably extend the duration of the cash buffer without any selling during a low.