Since there is little real data on the topic, I'll have to prance through the meadows of made-up figures. I hope you don't mind.
The common belief is that 95% of books don't earn more than $500 in its lifetime. That's probably true. I mean, anything that stays above rank #100,000 is probably on that road, and there are millions of books in the system (on Amazon alone).
Okay, so there are 99,999 books above rank 100,000. Of those, how many are traditionally published? We may never know, and I'm not going to count. Perhaps in my youth I might have....no.
So, let's make up some random figures to make ourselves feel good.
What would you say is the ratio of trad authors to self-pub authors in that top 100,000? 1 trad to every 1000 self-pubs? Maybe even more? There are a lot more of us than them. Of that we can be sure.
The distribution of money earned per book is more on the traditional side, or so we're being told. Less authors generating books which earn more money per title than most of us will ever see. However, they get considerably lower royalty percentages. Many of the authors make much more than any of us, but it takes tremendously more book sales to achieve the same. (around 6-10 times as many, depending on several factors)
And that last bit is based on plenty of anecdotal evidence. I've seen authors talking about royalties per sale in the area of $0.30-$0.70 per copy sometimes. If I pull ~$4.00 per copy, I have to sell 1/10th the amount to make the same money. That gives me a comfortable living if I sell 300 copies of three different books in a month, as opposed to the ~9000 I would have to sell to make the same kind of money. Isn't ~3,000-5,000 copies in a week NY Times status? I've heard that less than that can land you on the list in the slower months.
So... many of us are making more than NY Times bestselling authors due to the royalty structure? On far fewer copies sold. <-- there's your headline, if it can be proven.
We don't know the amounts that traditional authors make, but we could estimate some of them if we had the sales numbers. Of course, we rarely hear about them unless they sell a huge number of copies. I'd like to compare midlist or just barely NY Times bestsellers to our own figures to see how things measure up. That, though, might be a nail in the coffin, hastening the route to the grave.
I'm going to cut it short here, but to end, I think that the ones who keep repeating the mantra ("There's no money in self-publishing.") will eventually get tired like everyone else and move on to more interesting topics. lol