Heck, if ownership of other real-world assets was defined onchain, instead of paper, you could post that as collateral. For instance, if the ownership of your house was officially registered onchain, then you could post it as collateral. Because whoever loans you, would know that he would get the official ownership of the asset. Which is for all intents and purposes the same thing that happens when we mortgage our house. Except that with our current systems we involve a lot of lawyers, banks, and notaries, that to a large degree would be dis-intermediated into code. To be noted that, of course, you can refuse to leave the house even if you lose control of its ownerhship, but these happens with both paper ownership and onchain ownership. Both would require the enforcement of private property rights by some authority holding the sanctioned right to forcefully remove you.
I don't quite see how it could be a significant share of workload that would go to code. The parts you can dispatch there just are not that significant in costwaise when running a financial organization.
Furthermore, even the part that could be dispatched to code, won't. See DAO in the early days of etherium.
To be clear, those products, lending money without fiat collateral are much, much more significant feature of modern financial system than payment settlement. Until crypto solves those somehow, any comparison is meaningless. Like trying to compare wheels with cars.
Services exist to do what you want, but like in fiat institutions the checking process has to be performed "off-chain" in lieu of collateral. If your follow up question is "well then what's the point of crypto?", consider that getting a fiat loan at a bank with collateral still requires human checks in order to ensure that, for example, the collateral exists. In crypto, the collateral check is performed automatically and autonomously, and is therefore immediately more efficient.
Last year for the ETHOnline hackathon, I made an app that allows anyone to request a loan in any erc20 token with (optional) collateral in any set of erc20s. In addition, the borrower identity is used to increase confidence: a linked Lens profile and/or verified passport using Coinpassport. The principal is then crowdfunded. [2]
While making this, I learned that Aave was originally called ETHLend and followed a similar model of requesting individual loans but this approach makes it difficult to attract liquidity so when they developed over-collateralized pools, the protocol became much more popular.
It's going to take better identity tools than we currently have. Of course, undercollatorized lending is the basis of traditional banking so replicating it on chain is a massive win for the ecosystem.
There's a few projects working on credit scores on chain: Sprectral [3] and Cred protocol. See this article about using Cred protocol [4]
[1] https://jumpcrypto.com/writing/paradigms-for-on-chain-credit...
[2] https://ethglobal.com/showcase/lwned-75x0f
[3] https://www.spectral.finance/
[4] https://medium.com/chelo-finance/chelo-tech-guide-2-uncollat...
My occasional lazy-ass status check: what's the status of non-crypto-collateralized loans at the moment? Can I get a crypto loan to buy a car or house nowadays without having to post crypto as collateral?
(If the answer is yes, I'd be curious to know what aspect of present day crypto makes it competitive to establishment finance.)