The bond is meant to reduce, not eliminate, the frequency of fraud. It also makes winding up a failing institution smoother. Yes, this impacts efficiency, but it also decreases risk. One could argue about the size of the bond, but as a mechanism it's a technique with a good history across finance.
Also, no firm is bust-proof. There are a number of human and technical errors that could take down a payment processor. Payment processors are also exposed to payment fraud. There there is good old competitive pressure.
I would willingly take the risk of default on small purchases (< $100) in return for saving 3% on the cost of the transaction. This would pay off every 50 or so transactions -- and the failure rate is likely orders of magnitude lower than that.
Consumers don't bear the brunt of the risk of a payment processor folding. The merchant with millions of accounts payable, an unsecured claim, does.
I agree with the core of the post: the system is too cautious. I disagree with the call to lower the bar for starting a payment processor to parity with a photo sharing app.
Repealing the MTA would not lower the bar "to parity with a photo sharing app". There are still a whole host of federal regulations that you have to comply with.
Perhaps this would be a more concise argument than the multi-series approach? In essence, federal regulation is more than enough and the California finish is superfluous.
Yeah, I probably should have said that explicitly. I'm trying to keep an awful lot of balls in the air right now, and sometimes one of them falls on the floor.
Well, the Moneygram fraud mentioned in the blog wasn't about small purchases - they got done for knowingly providing payment processing to scammers who were swindling people out of their life savings!
Also, no firm is bust-proof. There are a number of human and technical errors that could take down a payment processor. Payment processors are also exposed to payment fraud. There there is good old competitive pressure.