Safe Withdrawal Rates ... again

motivated by a discussion on the Retire Early forum
Here's an interesting observation concerning Safe Withdrawal Rates ...
>Not SWR again!
Pay attention.
The notion goes like this:
 You look at the worstcase 30year scenario in the past
 You conclude that a 4.0% withdrawal rate is Safe
meaning that your portfolio would survive 30 years if you started
withdrawals at 4.0% and increased your withdrawals if inflation was positive
 So you begin a 4.0% withdrawal today (feeling confident that it'll last for the next 30 years)
assuming that the next 30 years won't be any worse than any past 30 years
 You find that, after three years, your portfolio has dropped to 50% of what it was when you started
 Your withdrawal rate is now about 8% (since it's based upon your initial portfolio not your current portfolio ... which is half as much)
 Nevertheless, you still feel confident that you can continue with this 8% withdrawal rate (increasing with inflation) for the remaining 27 years
 Then a friend (enemy?) tells you that your 8% rate over (the remaining) 27 years is NOT safe
In fact, looking at worstcase 27year scenarios, historically speaking, your friend says that a 4.2% would survive but 8% rate would not!
 You are completely confused.
First the 4% scenario was safe (based upon the worstcase 30 years).
Now, continuing with that same 4% scenario (for the remaining 27 years) it's no longer safe
>So, do you cut back on your withdrawal rate ... from 8% to 4.2%?
Since that original 4.0% SWR was based upon historical precedent, let's look at a sample historical 30year period.
>A bad 30year period, right?
Yes. We'll consider starting in Jan, 1930 and continuing for 30 years to the end of 1959.
As indicated in Figure 1, a 4.08% withdrawal rate (increasing when inflation was positive) was "Safe" ... for 30 years.
But, after three terrible years (1930, 1931 and 1932) a $1M portfolio would have been reduced to under $300K ... the red dot.
During 1933, your withdrawal rate (as a percentage of your 1933 starting portfolio) would have been greater than 14% !!
 Figure 1

>So what's your advice?
You'll notice that this lousy 30year period (19301959) starts with three years of terrible returns but then has several great returns (and your portfolio doubles
over the next four years which explains why the 8% is okay ... for this particular time period).
>Yeah, so what's your advice?
Advice? I don't give advice. However, I will say that one should be very careful when you're retired ... and withdrawing ... and ignoring market machinations
but relying on some predetermined SWR.
>The next 27 years might be like the last 27 of 19301959 or maybe like the first 27, right?
Yes, and looking at the worstcase 27 you'd be looking at that first 27 years ...
>With those three lousy years?
Yes.
>But could that happen? I mean, that'd be six lousy years!
Anything is possible ... wouldn't you say?
>And what about the next six years?
Wait'll I check ...
