Drawdown: safe withdrawal rates

wneil wrote: »

Other research has highlighted that Guyton Klinger 'decision rules' is not always as good as it may first appear. The ERN website has a good analysis as part of a wider investigation into SWR

That would be because they didn't really use the Guyton-Klinger rules but still wrote as if they were. I suggest using considerable caution with anything written there.

"The way we model the dynamic rule is a simplified (decluttered) version of Guyton-Klinger
...
We discard GK's convoluted portfolio management rule. We have only two assets (stocks and bonds) and simply assume that the portfolio is rebalanced back to the target weights every month. ...
If the 12-month trailing (real) return was negative, then forego the inflation adjustment, i.e., shrink the real withdrawal by the CPI-rate that month. If the 12-month trailing return was positive, then do the CPI-adjustment. We don't use the Guyton-Klinger 6% cap on the CPI-adjustment"

Guyton-Klinger doesn't have a 6% cap on inflation increases. Guyton (2004) did but Guyton-Klinger (2006) removed it. Getting confused between the two might help to explain some of their other mistakes. Wade Pfau explained two of the changes from G to GK:

"The modern version of their rules drops Guyton’s original inflation rule, which capped spending increases at 6% when inflation exceeded that amount. It also adjusted the withdrawal rule to prevent spending freezes when the current withdrawal rate falls below the initial withdrawal rate."

Unfortunately their decision to ignore important parts of the Guyton-Klinger rules is what causes the issues their version has later.

The GK withdrawal rule really says "Withdrawals increase from year to year in accordance with the inflation rule, except that there is no increase following a year where the portfolio's total return is negative and when that year's withdrawal rate would be greater than the initial withdrawal rate". They observed that the part after and "generates about 60% fewer freezes" than not including it.

Real GK also doesn't apply the inflation rule freezing when the capital preservation rule is used and GK observed that this "increased the purchasing power retained by more than 10% without reducing the probability of success".

So their modified version caused more skipping of inflation increases and lower income because of how they changed it.

They assert and presumably built into their modified rules the claim that "the 0.2 guardrail is on top of the drop in the portfolio. If the portfolio is down by 50% and you hit the lower guardrail, the drop in the withdrawal is (1-0.5)x(1-0.2)=0.4 = 60% under the initial withdrawal". If the guard rail is tripped because drawing has increased from 5% to 6% of the current portfolio the real GK rules say cut your nominal income by 10% that year. So if you were taking 10,000, after a 50% market drop you take 9,000. You don't need to be a maths genius to recognise that 9,000 is a lot more than 4,000. If the portfolio is still 50% down the next year you'd take another 10% cut, to 8,100... and potentially for more years if there's no recovery.

So again they modified the rules to take less income than the real rules.

----

They seem to even more greatly misunderstand Guyton-Klinger, wrongly asserting that it first takes income out of the highest returning part of the portfolio, equities. It really takes it from equities first only if they are up (and if up and rebalancing cuts equities the excess goes to cash) else first from cash, then bonds, and equities after that. Though there is a bull market rule that sets aside some cash for later use.

----

That ... post isn't good and should largely be ignored:

1. It falsely asserts that the 4% rule is doomed because it doesn't use the correct equity-bond mixture which the 4% rule is for. The real 4% rule investment mixture is OK, though definitely stressed.

2. No bonds, so it's either sell equities or hold cash. Should be some bonds in there both for rebalancing and to draw on when equities are down.

3. No bonds and US shares reduce the natural yield so the cash runs out faster.

4. No bonds so it presumably isn't using Guyton's sequence of returns risk taking approach, which further improved performance/success rate.

The result is a too high two year cash suggestion when one year or GK should be used. One year topped up by the natural yield will last longer than most downturns because the yield is likely to be around 3.5-4% for UK investors. Even using 3% and drawing 6%, one year is a couple of years worth before bonds get touched.

In another discussion VCTs came up and I thought knowing which I hold might be of interest to some.

[VCT name		Value	price	shares	cost	gain%	gain£	incomePA	pays
AADV Albion Dev		941.46	0.7100	1,326	0.5279	34%	241.46	53.04	£26.52 May & Sep
AAEV Albion Enterprise	938.79	0.9150	1,026	0.6823	34%	238.75	51.30	£25.65 Feb & Aug
AATG Albion Tech & Gen	819.03	0.6900	1,187	0.5897	17%	119.06	47.48	£11.87 Jan & Jun
AAVC Albion VCT		28,415	0.6550	43,383	0.5002	31%	6715.69	2169.15	£1084.57 Jan & Jul
CRWN Crown Place	893.75	0.2860	3,125	0.2240	28%	193.75	62.50	£31.25 Mar
D4G Downing Four Gen	5,305	0.9000	5,895	0.7125	26%	1105.31	0	doesn't pay yet
EDV Elderstreet		7,516	0.5450	13,791	0.4568	19%	1216.37	413.73	2xyear, nos not to hand
KAY Kings Arms Yd	1,000	0.2000	5,000	0.1400	43%	300.00	50.00	£25 Oct

The income column is calculated based on announcements, the pays column is the most recent annual payments actually made. Gain is after taking the tax relief 30% discount and is capital only, doesn't include the dividends.

D4G and EDV buying was constrained by limited availability at the very end of the tax year of purchase, though I also wanted to diversify away from the Albion shares to reduce concentration risk.

For 2017/18 tax year I didn't buy VCTs. Instead I plan to buy EIS' in 2018/19 and apply that to the 2017/18 year, something permitted by EIS but not VCT or pension rules.