In my previous article, I talked about a potential peer-to-peer money transfer system, and the problem of how to get such a system started: Why would you be the first of your friends to join such a system? It wouldn’t be useful until you had well-connected friends to connect to the network through. I suggested that this problem might be alleviated by strapping such a money-transfer system onto pre-existing decentralized social networks.
Well, since then, I’ve read more about the recently rejuvenated Ripple, and I’ve discovered that they have a different, very clever solution to this problem. Their solution has other benefits, too, including a currency that I argue isn’t necessarily backed by debtors, as every country’s fiat currency seems to be.
Their solution is to use an internal currency, called Ripples (XRP), and maintain a single authoritative ledger (cleverly without requiring a single authoritative computer or company to say which ledger is authoritative).
If you want to pay someone via Ripple, and don’t have any friends who’ve already joined, you can still do it. What you do is you find a Ripple “gateway” who you can pay in a mutually agreeable currency; in return they send you XRP in Ripple (or, if you don’t want your own Ripple account, they might complete the rest of your transaction for you). I think the idea is that in future, you should be able to find a gateway (or even multiple competing gateways) in your country.
The XRP you have inside Ripple can then be transferred to the person you want to pay (or to their gateway), without there having to be an existing chain of trust within Ripple between the two of you; the ledger can just reduce your balance of XRP and increase theirs.
The system is also designed to make use of the chains of trust I talked about last time, maintaining records of IOUs in any number of other currencies.
But, as I mentioned at the start, there are more benefits to XRP than just facilitating payments in the absence of a chain of trust.
One is that they help to prevent denial-of-service attacks on the Ripple system. This is because each transaction entered into the ledger requires a certain number of XRP to be permanently “destroyed”. Normally, this would be only a small fraction of an XRP (and if I’m reading RippleCharts correctly, then 1 XRP is currently worth about half a US cent).
But if the servers are having trouble keeping up, they’re free to raise the transaction fee. So if someone attempts to send lots of payments back and forth around Ripple, overloading the servers, they’ll raise their fees, and the attacker will soon run out of money to pay the fees. Then, the fees will drop back to normal rates.
Even in the absence of an attack, such fees would have the effect of efficiently rationing the resources of the servers working on the Ripple network; if the fees are high at a certain time of day, and your payment can wait, then you’re free to send your payment at a cheap, low-volume time of day.
But because the fee is destroyed, rather than paid to the owners of the servers, high fees won’t encourage more people to run Ripple servers.
Another feature of XRP that I find very interesting (but which I haven’t seen mentioned on Ripple’s own site) is that this system of destroying transaction fees effectively puts a floor on the value of XRP.
You see, when Ripple becomes popular enough to make full use of all its servers, the fees should reach an equilibrium where the fee for a transaction is high enough to discourage over-use by people not willing to pay much for transaction services, but low enough to encourage use by enough people who value the services to keep the servers from being unwillingly idle. So the fee should represent the current marginal value of a Ripple transaction.
Now, if the value of XRP declines too far, then this equilibrium will ensure that the number of XRP paid in fees rises. But because the fees are destroyed, this will permanently reduce the number of XRP in existence (initially 100 billion), making them more scarce, and therefore more valuable.
So I think that the total value of all XRP in existence shouldn’t fall below the expected total of the time-discounted marginal values of all future transaction services provided by the Ripple system (as opposed to services provided by people who use the Ripple system, who are free to charge fees that they can keep).
What this means is that the value of XRP, unlike that of fiat currencies, doesn’t necessarily reflect how much value can be extracted from debtors who owe money in that currency.
I say that it doesn’t necessarily rely on debtors in this way, because, while I’ve described a mechanism that shouldn’t allow the value of XRP to fall below a certain amount, I haven’t yet discovered a mechanism that would prevent it from rising too high. Such a mechanism would protect debtors from seeing their debts balloon in value when other debtors become more productive.
But I don’t think such a mechanism exists in Ripple, and I think this is related to the absence of an inherent incentive to provide more server capacity by letting servers keep fees instead of destroying them.
So while the value of XRP might not put its debtors into class servitude the way fiat currencies do, there’s no guarantee that it won’t. Presently, I expect that there aren’t many people who have short-sold XRP, but who knows what’ll happen in future?