...And clearly, having a $1M portfolio at the peak of a bubble is not worth as much as still having $1M after the bubble has burst. (edit/add: ) and in your example, the $500,000 is worth about as much as the $1M from a year earlier, so yes, they should be able to spend about the same. I'll venture that in rough numbers, that 'same' is probably best looked at as an average of the two - about 3%.
But FIRECALC is saying that historically, the $1M portfolio survives 95% of the time at 4%. So if 95% is OK with you, and relying on history is OK with you - there ya' go. It's really the $500,000 guy that could go to 8% in this case (if you could 'trick' FIRECALC to skip the first year of downturns, as is what happens in your construct).
Yes, we all agree that there seems to be a paradox, where a guy who retired at the top of the market with $1M was told by FIRECalc he could get $40K/yr, whereas if he waited until the bubble has burst, he would be told to draw less than $40K. How to reconcile this difference?
Actually, there is no paradox. Not knowing where we are in the stock market cycle, let's say that the 4% SWR would give us a 95% chance of success over all outcomes from past market histories. This means that we have a 5% chance of failure.
Now, after our portfolio has suffered shrinkage to say half its initial size, this a posteriori knowledge now should lead us to suspect that, darn it, we may indeed fall into that unlucky 5% after all.
Another example is this (not quite the same as Katsmeow's post earlier). We are given two coins and told that if we roll 2 tails, our fate is doomed. Because there are 4 equally possible combinations, Head-Head, HT, TH, and TT, we know that our chance of being doomed is 25%.
So, we proceed to roll our first coin, and lo and behold, it comes up tail. Now, given that we already get one tail, the chance of doom for us is no longer 25%, but 50% now of getting the second tail.
Since most of us have been taught not to "time the market", we dare not theorize where we are in the market cycle. This is not bad, because people most often get it wrong, and I do not disagree with Malkiel and Bogle about this. So, when starting his retirement with $1M and drawing $40K, our retiree is told that his chance is 95%, because FIRECalc makes no assumption about where we are in the market cycle. But once our retiree has experienced the decimation of his portfolio, and it has been
known for a fact that he was unlucky to retire at the top of the market, of course his chance of failure is now greater than 5%.
On the other hand, had he delayed his retirement until now and withdrawing less than $40K, of course his chance would be a lot better.
Of course it's hard to know where we are when we are in the middle of a cycle, but I think the intro to FIRECALC should touch on this. If you just saw your portfolio rise in value because we just had a few extra good years, you should be extra conservative with the SWR %. Conversely, if we have had several bad years, you might be able to be less conservative. Human nature probably drives people to the opposite conclusion. And we can't bank on ANY of it.
Ah, again, we are told not trying to guess what phase of the market cycle we are in. It's a sin to avoid at all costs!
We are told not to rely on our heuristic reasonings, that good years tend to follow bad years and vice versa.
Of course there is no certainty in life, just lots of probabilities. There is absolutely no guarantee how long a spell of bad years will last. "You pays your money and you takes your chances".