No. of Recommendations: 19
And before someone says valuation isn’t a concern for his wife, in an interview he did say “anyone will do fine with that”.
I would take that comment as an approximation : )
Certainly anyone who doesn't really have enough money saved to be able to afford that strategy is not going to do fine, so it certainly doesn't work for "anyone". Admittedly it's not clear what their alternatives are, other than doing their best to pick their desired trade-off between a too-low income and running out of money too soon.
As for the 4% rule being applied to a SPY type portfolio, I wouldn't recommend doing it naively. The number of years it will last is a very strong function of valuation levels on the day of retirement. Some folks who do that will go broke because their initial estimate of the sustainable withdrawal rate will be too high, because valuations were high when they started the program. I would propose the single adjustment that the SWR be set to equal the cyclically adjusted earnings yield of the index, for example using ten years of inflation adjusted earnings. It would have been 4% at the start of 1996, and again in August 2002, but would have been only 2.1% in between at the market top in March 2000. About 2.6% at the moment. *
As a fun exercise, imagine you'd tried the traditional 4% rule with 100% Berkshire stock starting at its high valuation in June 1998.
Assuming the scheme was implemented as liquidating an inflation-adjusted 1% of your original portfolio value each quarter on the quarterly anniversary of the start date, you'd be down to 40% of your original (real) portfolio size less than 11 years later. Pretty stressful if you hope to live another 20+ years. Even with a fabulous underlying investment, a 4% rule starting with unusually high valuations can give surprising results.
It did stop dropping, you'd be up to 60% of the original real portfolio value now after 27 years.
Jim
* For those who think calculating something like CAEY (similar to the inverse of CAPE) on their retirement day is too hard:
A simpler rule, which is surprisingly good given its simplicity, is just to set your SWR equal to twice the S&P dividend yield the day you retire, and adjust that dollar amount for inflation thereafter. That would have given you 4.7% SWR for someone retiring at the start of 1996, 2.2% at the 2000 market top, and 3.4% if starting in August 2002, and 2.26% today. The key thing is that the SWR is a function of market valuation on retirement day...the calculation doesn't have to be perfect. (A slight improvement would be "twice the dividend yield you'd be seeing today if S&P dividends were still at their peak". This doesn't penalize those who retire in the middle of a "dividend recession").
Still too hard? For those who hate all SWR strategies because it's too hard to do inflation adjustments, but still hold SPY or something similar:
Every quarter, spend (a) whatever you got in dividends in the last quarter, plus (b) sell stock equal to the same amount. Dividends tend to rise with inflation, and dividend yields vary pretty well with market valuation levels, so this works surprisingly well. In terms of sophistication it's miles ahead of the recommendation that everyone use 4%.
