Tim Swanson thinks transactions cost 25 BTC divided by the number of transactions in a block. That's a complete misunderstanding of not just what the block reward does but of what Bitcoin even is. He has gathered some interesting data, but his analysis is unlikely to be of much use as he has no fundamental understanding of Bitcoin in the first place.
Tim Swanson is one of the few widely published that really seems to understand Bitcoin if you ask me. I think his advocacy of so-called permissioned ledgers stems from the premise that
decentralization hurts the ability of Bitcoin to integrate into the current legal/regulatory pigeon-holes. I personally don't find that compelling, but that seems to me to be the nut of his argument.
Correct me if I'm wrong.
To create a digital version of the US Dollar that retains its "cash-like" properties, David Andolfatto (VP, Federal Reserve Bank of St Louis) argues
here that miners would be needed to process transactions in order to keep "Fedcoin" free from KYC requirements:
...the e-version of the USD will probably be subject to KYC restrictions, which is unlike paper cash. To the paper cash feeling, we'd need to let the book-keeping done by disinterested third parties, like Bitcoin miners.
So here we have a Fed Official arguing that decentralized transaction processing is actually needed to fit within the existing regulatory/legal framework. The Fed can't use a "permissioned ledger" because then those granting permission would also be responsible for AML/KYC checks. This is opposite to the idea that "decentralization hurts the ability ... to integrate into the current legal/regulatory pigeon-holes."
...to keep Fedcoin free of KYC restrictions, we probably don't want the Fed involved in processing these payments.