I'm going to take issue with Grangaard's statement that only 3 ten-year periods of stock return were negative. By my numbers, there were 8 ten-year periods when REAL "stock" returns were negative, 8 out of 68, from 1926 to 2002. There were 30 ten-year periods (out the 68 ten-year periods) when the REAL "stock" returns did not double your money. And then there were 38 ten-year periods when the REAL "stock" returns did double or more than double your money.
Remember to think in terms of Total Returns, not annualized or average returns. So, if someone gives you 6%/year for 10 years, that's 179% in total returns. There were 23, out of 68, ten-year periods, from 1926-2002, when the total REAL stock returns were less than this. 23/68 = 34%. 66% is not really the vast majority of 10-year time periods, is it?
I used the Wilshire 5000 as a proxy for "stocks". If you use the S&P 500, you get almost the same thing. With the S&P 500, there were 7-ten year periods with negative real returns, 33-ten year periods, out of 68, when the real stock returns did not double your money, and 25-ten year periods, out of 68, when the real total returns were less than 179%.
As for having more total money than you started with 10 years prior (in REAL terms so you can keep your withdrawals inflation adjusted), you pretty much have to double your stock money, in real terms, if your portfolio is 50/50. This is assuming you spend the 50% bonds in the previous 10 years. With equity returns most likely going to be lower in the future, I seriously doubt that this is likely to happen. So, it's probably good to keep your withdrawal % down to 4% and under.
Now, having said that, I think the "bucket" idea is a very good one. Though, I do have a problem with the way the portfolio is rebalanced. Assuming you let the stocks ride for 10 years, and don't rebalance, at the end of the tenth year, you have 0% in cash/bonds and 100% in stocks. As time progresses towards the tenth year, your portfolio becomes more and more concentrated in stocks. That doesn't really make me feel very comfortable.
I'd rather use a modified version of the "buckets" and withdraw from the bond/cash portion of the portfolio when stocks are not doing well (like a bear market), and then take withdrawals from the stock portion when the stock market is doing well. You can still protect your stocks (and not reverse- DCA) in bear markets this way. Frank Armstrong explains it better:
http://www.investorsolutions.com/Premier/premiercontent/files/Investing during retirement.pdf
Did the Grangaard book take the strategy and extend it to subsequent 10-year periods? For example, start in in years 1926, 27, 28,...,1972, 1973:
(1) Spend the bond portion totally within each ten years, then
(2) Replenish the bond portion from the stock portion, then
(3) Let the remaining stock portion grow for the next ten years, then
(4) Replenish the bond portion from the stock portion, then
(5) Let the remaining stock portion grow for the next ten years,
(6) and so on for 30 or 40 years for each starting year.
I tried a simulation spreadsheet for a 5% withdrawal rate, and got pretty bad results after 30 years of retirement. For 30 years, I got 24 successes and 24 failures. For 40 years, it was about half and half again. 4% did better, but still 13 failures (ran out of money) at 30 years.
- Alec