[cube13]

>An m-of-n transaction with a decent and trusted intermediary (or multiple less trusted intermediaries) could also replace bank guarantees, which is the simplest form of trade finance. There are also other opportunities in using bitcoin as proof of ownership or proof that you hold a product.

So let's replace banks with... banks?

The overhead costs that bitcoin will save are just the wire transfer fees. Beyond that, there's still lawyer time, intermediary escrow fees, etc. Those will not change.

[nadaviv]

There's one big difference here: with Bitcoin's multi-signature transactions, you can require the agreement of two of the buyer, seller and arbiter in order to release the funds. The arbiter cannot move funds on its own and must have the cooperation of one of the other parties, which means that he doesn't actually control nor hold any funds. The legal situation here is a classic binding arbitration, and not an escrow - which are very strictly regulated, require licensing in some parts of the world and have a very high costs of operations due to bonds and securities.

Dealing with arbitration services rather than with escrow makes this much simpler, drops the entry/operational costs to nearly nothing and does not even require a lawyer to assist you with the process. This lowers the barriers of entry significantly and allows to create a competitive market for arbitration services in a way that's simply not possible when an escrow is required, leading to reduced end-user costs and improved service quality.

Source: I'm the founder of Bitrated [1], a service that enables exactly those kinds of arbitration services, and received extensive legal advice on this matter from my attorney. Do note, however, that this is a new and somewhat gray territory that can be interpreted in multiple ways (and of course, IANAL/TINLA apply - ask your own lawyer before doing anything.)

(P.S. Bitrated v2, a complete rewrite I've been working on for the past 6 months, which now includes an identity & reputation management system, is about to be shortly released. If anyone is interested, you can follow @bitrated on twitter for updates. </shameless-promotion>)

[1] https://www.bitrated.com/

[nikcub]

You would replace a bank with coinbase, circle, someone else you trust or 2, 3, 4, 5 or even 10 other companies or individuals. Similar to a web of trust.

That is just to replace the trust element of a bank, not the financial function. At the moment it is nearly impossible to split trust across more than one institution - which means there is a single point of failure in the trust chain and if 2008 repeats again x% of people will lose funds. m-of-n allows you to distribute that trust, even if its just more banks signing the transactions to begin with.

Further, there are other ways to do proof of reserve with bitcoin that don't involve m-of-n.

In terms of the financing and risk assessment element, you could carve that out in tranches and sell out it on auction. At the moment the $2 trillion p.a trade finance market is limited in entry to only those who have a banking license, hence the thick fees and profits in it. You could open up the financing and terms to an open market, just like crowdsourced venture funding or project funding, or have smaller institutions that specialize in certain types of risk that they better understand (it is crazy that there are a half-dozen major banks who pretend to be able to understand and price every time of import / export finance situation anywhere in the world).