Set forth below is the text of a comment that I recently posted to the discussion thread for another blog entry at this site:
Isn’t it fraud to say 4% won’t work? After all, you have admitted that predictions fail.
No predictions are made in the calculation of the safe withdrawal rate. The return scenarios that have applied in the historical record are considered and the withdrawal rate that barely works in the event that the worst-case return scenario happens to pop up starting from the valuation level that applies on the day the retirement begins is identified as the safe withdrawal rate.
The mistake that the Buy-and-Holders made was to consider only the one particular return scenario that happened to pop up in the historical record. For example, the highest valuation level we experienced prior to the late 1990s was the “33” that applied just prior to the onset of the crash of 1929. In that case, we saw a lucky return scenario pop up and a 4 percent withdrawal barely survived. However, the 4 percent withdrawal survived at that valuation level only in 50 percent of the return scenarios that we have seen in the historical record. There was a 50 percent chance that a retirement beginning in 1929 that used a 4 percent withdrawal rate would fail. A retirement with a 50 percent chance of failing is obviously not safe.
I don’t need to make any predictions to say that. I just look at the historical record. No predictions are needed because I choose the worst-case scenario. It is usually the case that a withdrawal rate not identified as “safe” has some chance of working out. I could predict that a lucky return scenario would pop up and that an unsafe withdrawal rate would in fact work out. But that’s not safe withdrawal rate analysis. The literature defines the concept as calling for examination of what happens in a worst-case scenario. It is because the investor is finding out what happens in a worst-case scenario that he views the withdrawal rate identified as a safe one.
I hope that helps a bit, Anonymous.