You are entitled to your opinion, so I am going to create the centralized version of this exchange first because that will prove the model. If it works in a centralized manner where the central 'authority' is merely simulating matching bids / asks, then do you agree that it would work in a blockchain?
I think as a proof-of-concept, your model (with some refinement) will work. My concern is whether or not it will be practical. As you know, a functioning market requires liquidity and I don't see any efficient way to inject liquidity into the model right now. Basically you're stuck at the fiat to crypto exchange hurdle. Using escrow and such works on a micro level but if you scale everything up to a macro view, I just don't see how you're going to be able to get all the necessary funds into the market.
Maybe I'm getting ahead of myself but it seems the main feature of your new endeavor is to eliminate the need for fiat deposits. Without an efficient way to do that, I can't see a need for all of the rest. I'm hoping in this respect you are more visionary than I am as I think this is the main issue facing Bitcoin right now (how to decentralize the exchange of fiat to BTC). I'll be watching closely and hopefully adding some productive comments along the way. Maybe even some capital if I can be convinced this is potentially a viable solution.
This is certainly a growing pain issue and adoption will be slow at first. But lets make some assumptions: assume that crypto-USD ends up tracking actual USD within a small range for 6 months. Assume that crypto-USD is actually paying a 10% APR. Now, ask yourself this... how many people will be 'interested' in trying that out? How many people will create a business out of accepting 'deposits' of USD for a small fee and selling crypto-USD? Perhaps just as a hobby? Watch as this starts catching on and people realize they can make money. All of a sudden it goes viral and you can deposit / withdraw cash with just about anyone for a small 'ATM' fee. Sure, it may take a while to catch on, but it will.
Any created business that accepts USD and issues "virtual currency" will have to obtain a money transmitter license. They'll have to comply with the same AML/KYC laws. So essentially it's like setting up gateways like Ripple, just off network. It would be difficult to do as a hobby unless you deal with very small denominations. Although maybe established gateways and Bitcoin exchanges could be enticed to be brought on board, afterall they've already forked out the cash for the licenses so they may as well dip their toe in as many markets as they can.
The dividend paying aspect is interesting and where your idea differentiates yourself. It will help to garner interest but even by your own admission, you cannot protect the longs from big market corrections. In a young and illiquid market, you're going to have large price swings. Will a 10% dividend be enough to entice a large amount of investors to see that reward as commensurate with the risk, that's the big question.
