Right now, we have the government operating as both the "miners" and the auditors.
A middle-ground, then, might be one in which there is a general-purpose public blockchain where the government takes the role of the trusted third-party on-chain, thereby lowering the overall societal costs of things like unfulfilled contracts while still ensuring that my oil-reserves are what I say they are.
From the government's perspective I see this as desirable but falling prey to the principal-agent problem. We'd need virtually no civil-court judges or attorneys to prosecute white-collar crime anymore. It'd lower costs, but in doing so unemploy many of the people in the system.
But the rest of your post is pretty silly. How does a secure digital exchange protect against fraud? It doesn't.
It's that plus a whole new crop of perverse incentives which arise with any "smart contract". Especially, no smart control can reach out to the "real world" and get data, more less facts. A smart control would essentially be limited to some data-stream in on Internet. And anything like that can be subverted seven ways from Sunday. Do you really want billions riding on THAT?
(and, yes, give us some Rube Goldberg scheme of averaging six different weather stations for your result. I'm sure nothing can go wrong with any of those ideas. Yeah, World finance then gets to be hoisted on the lpad petard).
When the great recession started, it was not caused by an external shock, rather by the viral mistrust that the underwriters of the bad loans were solvent. A financial fragility problem that the blockchain can work to address.
I also don't think this is a false dichotomy. Most likely we'll end up at a place with some improvement over what we have today (I'd actually argue we're currently at this point now) but not without all of the limitations.
The required nexus between the blockchain and the tangible and legal worlds means every limitation that exists today will exist for a "smart contract". It's the same plumbing with a paint job.