I wish they'd give up on these pointless and empty comparisons.
Exactly. The fact that change addresses are so commonly used in bitcoin should be an indicator of how much "duplicate" transaction volume is being accounted for here.
But if there's change, then that means it's not someone shuffling money around from one pocket to another, that proves they're paying someone, right?
Well, no. It's a well known fact that people try to simulate real commerce by "splitting": sending 1 or more outputs to 2 addresses they control, in order to appear as though someone spent some money and received some change, and so now 2 people own the money that was used as the input to that transaction. There are techniques to analyse this kind of behaviour, but it's all probabilistic stuff. And there are various way to defeat that kind of analysis anyway.
So second guessing how the money is being used carries no certainty, as people are already camouflaging 1 form of behaviour as another on the blockchain. Here's a certainty; whatever the reason for the movement, this article claims 1.3 trillion was moved in 2018. And everyone who moved that BTC wanted to pay the fees to move it.