If there's a mortgage the deed is held by the mortgagee. Or at the very least there will be a formal legal claim on the deed (in the UK it's called a lien or a charge - not quite identical but similar) which isn't satisfied until the mortgage is paid off.

Worse, there were instances in 2008 of banks stealing homes out from under people - who sometimes didn't even have a mortgage - by getting fake paperwork waved through, because it was impractical for all the competing claims on a loan to be properly validated.

https://www.salon.com/2013/08/12/your_mortgage_documents_are...

Would blockchain fix this? No. The problem wasn't tracking the transactions, it was the fact that banks and the financial industry fraudulently hid the true risk of various loan bundles.

Blockchain would only have solved this problem if the loan context - creditworthiness, payment history, and so on - was included with every loan transaction, and could easily be accessed and summarised for all the transactions in all of the loan bundles.

Which blockchain doesn't do.

Coincidentally, this is exactly the same problem that blockchain hasn't solved for FTX, Binance, Mt Gox, etc. Blockchain does not provide an objective assessment of risk, value, or corporate credibility. It's just a ledger.

Even if you can't manipulate the ledger, you can still manipulate the perceived meaning of the numbers in the ledger. Especially if you're a trusted party in the cryptoverse.

In fact it doesn't just fail to fix these problems, it creates new opportunities for spectacular frauds.

Actually yes the problem was tracking the transactions. There were literally competing spreadsheets that teams of lawyers had to argue about and reconcile by looking at emails between traders and comparing row by row. Have you ever tried to compare two spreadsheets? Lawyers know how to blackline two word docs. They don’t know how to compare two datasets programmatically. It’s part of the reason I started coding. I felt after charging $850hr to manually compare row level data across spreadsheets.

In fact counts eventually started getting harder on banks, rejecting the argument that tracking transactions was impractical and ruling in favor of homeowners and making the banks eat the loss for their own shoddy paperwork. But for many it was too late.

There is legal and practical ownership and they differ in practice.

So imagine Bank of America issues a mortgage. They don’t want to keep the risk on their balance sheet so they create a mortgage backed security with 1000 other mortgages. Then they sell that to Goldman Sachs who in turn sells it to Lehman, but not the full portfolio only a portio, say tiers 8-10 of 10. In order to “facilitate liquidity” and “market efficiency” they wouldn’t document every transaction on the local county register of deeds. They would “trade” the securities by sending each other spreadsheets and hopefully “record” the transaction with a 3rd party like MERS. But when 2008 hit it often required going to court to fight out who owned what. Worse lots of homeowners ended up being foreclosed on because their payments didn’t go to the right owner on file.

So yes, the legal title is recognized by a written record held by the government, the market implementation adds layers on top to facilitate trading the paper.

Sounds like the issue is that they aren’t properly recording their transactions, which is something that could be fairly effectively managed by a centralized database rather than the blockchain.

Or at least, I don’t see what benefit the blockchain provides over such a centralized structure.

I agree. But it could also be solved by blockchain. And that was the question.