I also think this focus on pay later doesn't necessarily work for those new to the system (i.e. everyone that matters) who could still be apprehensive and hesitant to commit to a transaction.
The use of reputation isn't new. Spare Change had a system of reputation for those who reversed transactions. They let merchants block those with low reputations. A fun anecdote is that they needed to make a reverse-reverse transaction for those that wanted the reputation back.
But let me bring it back to merchant acquisition. Amazon and Google are both having a hard time with this too. People don't want to spend development resources on a new payment processor unless they really, really need to.
That's why solutions like http://wepay.com are awesome, because they have a payment product designed for a sizable niche that really could use their services.
(Well, technically, it really fixes the merchant's problem with that user more than the user's problem -- which is one potential stumbling block in user adoption.)
This doesn't alleviate the problem with nickel/dime scale micropayments, which is that users hate them with a passion unmatched by a thousand burning suns. That was true back when Clay Shirkey mentioned it in 2000 and is true today. The mental transaction cost makes every single microtransaction a significant barrier to conversion.
This is one reason aside from credit card fees why people who do micropayments successfully tend to use a dual currency model where the micropayment currency (bought for cash money) is as easy to spend as the currency which you can earn from ingame actions. (i.e. one-click no auth no confirm purchases, like buying something from the vendor in WoW.)