There is a real life version of this in the bearers cheque or Wechsel (german) or "Bill of exchange", where the wechsel can go from hand to hand, and the intermediate holders endorse the debt. If there are many names on the back of the wechsel and you know all of them (or some of them) to be solid merchants, or there is a bank guaranteeing it, it is just as good as money. The value is the principal less the interest until the maturity, less a little expenses and less the fact that the receiver always prefers money directly.
The point is that you know at least some of the intermediate holders, it would not work with complete strangers.
There is still a systemic risk, for instance if all the merchants trade at different levels of completion of the same end product, they can all go titsup at the same time. In such a scenario, the IOU's will be worth less and the money worth more.
I don't know if ripple can mimic this, but of so, I think it will be useful.
That is very much the idea with Ripple. It decentralizes exchange so that IOUs can be traded freely. Contrast with conventional exchanges who maintain total control of their IOUs (remember when MtGox removed their function for USD codes?) and nobody knows how many they've issued (on ripple this is public information. bitstamp's current gateway capitalization is $263,911 USD and 2,560 BTC).
The point about paper-checks not working when passed around through complete strangers is because you cannot trust a stranger not to counterfeit and double-spend a paper check. Issuing and trading IOUs are cryptographically signed transactions on the ripple ledger, so double-spending and counterfeiting IOUs is prevented just as it is with bitcoins on the blockchain.
I didn't think about counterfeiting, only the solidity of the issuer and endorser. It their solidity can be questioned, the IOU will be worth less, because there is some risk of getting nothing.