The relationship between Bitcoin and Ripple might be similar. Bitcoin is a special type of commodity money, like gold. A limited quantity of Bitcoins will ever exist – 21 million, of which 8.1 million have already been mined. Like gold, Bitcoin involves no counter-party risk: owning it doesn’t require the bearer to trust another person. Bitcoin transactions are also irreversible and don’t depend on a centralised authority to authorise or remember them. Once the thing, or bits have changed owner, it’s permanent. Bitcoin went through a large speculative bubble–perhaps inevitably–garnered by hype and media attention. Its critics dismiss it as flawed, yet it is still around and, at the time of writing, a Bitcoin is worth six dollars.
Ripple is a different kind of system: a proposal to create a peer-to-peer credit network, in which anyone can create their own money as IOUs, subject to other people’s willingness to accept them. The creation of new money as IOUs is made possible by people’s willingness to believe that the issuer will accept them back in exchange for goods and services at a later date. Alternatively, Ripple users can pay people who don’t explicitly trust their IOUs, by routing payments through a chain of intermediary users who do. If Alice wants to pay Paul, who doesn’t know her, she can still send him an IOU via John, who is trusted by both of them. This idea forms the basis of what could potentially be a digital, peer-to-peer credit system. Rather than merely providing us with a new way of lending old money to each other without banks, Ripple would allow its users to create their own.
At the moment, Bitcoin is entering its third year. It has received generous media attention, however its main users are mostly hardcore hackers and libertarians. It hasn’t yet caught on to the mainstream, for a number of reasons, including usability, scarcity, legal ambiguity and the lack of goods and services currently on offer for Bitcoin. It is interesting to see that despite this, Bitcoin has not yet collapsed. The network has stayed true.
Ripple has not yet successfully scaled beyond the level it needs to in order to become useful as a means of payment, rather than just an IOU tracking tool for friends. Part of the reason is that is suffers from unusual bootstrapping difficulties: its network only becomes really useful after reaching a certain size and density. Before that point, the network effects are not that strong. You might have a number of friends on a Ripple system, but it’s unlikely they’re the kind of people you need to do business with. Yet in order to get big enough to enable impersonal transactions to happen, lots of people need to join it.
Perhaps Bitcoin and Ripple can help each other, much like paper and gold complemented each other in the gold standard arrangements before they collapsed in 1971. However, this time, the technologies can be fundamentally peer-to-peer. A synthesis of Ripple and Bitcoin would create a sophisticated monetary system fully outside the control of any government or corporation. Such a system could provide a stable source of liquidity to people and businesses around the world, without having to rely on the monetary policy of governments, or lending habits of banks.
How exactly might this work?
Bitcoin as reserve currency
The most obvious starting point is to create a peer-to-peer credit network based on genuinely decentralised technology. Much like Bitcoin, such a network would process and record transactions in a totally peer-to-peer manner, without the need for a central book-keeper. However, in the case of Ripple, individuals would be creating their own money as IOUs, rather than mining it. A decentralised design would mean that each node in the network would know about every node it was connected to through a trust relationship. The software would work as a protocol, like email, rather than as a hosted service.
This proposal already exists as the theoretical latter stage of Ripple’s development. However, it’s possible to imagine something further still. It would be interesting if Ripple’s international unit of account became Bitcoin. Self-issued credit currencies in the Middle Ages were often denominated in gold or silver bullion. Most of the time, notes were redeemed in an equivalent amount of goods or services, as gold and silver were hoarded in temples and monasteries. The metals were used to periodically settle accounts, particularly large ones. The gold standard which followed later allowed international trade to take place thanks to a similar principle: currencies were pegged to a quantity of gold, but were often redeemed in goods and services. Gold only changed hands when deficits needed to be settled between countries. This, in turn, imposed discipline to keep control of their balance of payments, lest they run out of gold.
A similar option might be available to a decentralised Ripple network. Users could hold bitcoins in ‘reserve’ accounts, with the ratio of credit issued to bitcoins held exposed to other users of the network. Such transparency would allow everyone to see how leveraged any given user was. People would accept credit from other members who didn’t necessarily produce anything they wanted, because they would be able to convert credit into Bitcoin on request. Bitcoin would provide a level of security and confidence in the credit system. Such a set up seems to have worked for gold and paper for centuries.
There is nothing wrong with users extending credit in excess of their Bitcoin reserves: credit is denominated in Bitcoin, but redeemable in all kinds of goods and services, as well as Bitcoin, if necessary. What matters is that users have confidence in each other’s credit, whether they intend to redeem it for Bitcoin or not. With the ability to create credit denominated and backed by Bitcoin, users are no longer artificially limited by its supply. Credit can expand and contract according to confidence, but still stay within credible limits.
The last ingredient to this hybrid model is a trust metric, which would allow users to share information about their confidence in users with their neighbours. In the informal gift economy, the checks and balances of reputation accounting ensure that free-riders are eventually outed, while people who do their fair share (or more) win trust and respect. Similarly, a good trust metric will make credit-worthy users well regarded, further extending their ability to create money. People who don’t honour their promises should see their trust ratings drop, discouraging others in the network from accepting their credit.
In principle, this is how many trust measurement systems work: on eBay, judgements of individuals about a person’s trustworthiness are aggregated into points. This rating system makes eBay work: it gives people who don’t know each other sufficient reason to engage in time-delayed transactions, where fraud would otherwise be easy to get away with. Reputation currencies create confidence.
However, an eBay-ratings model would almost certainly be gamed in a monetary system (and already is to an extent on eBay). How do you design a trust measurement system that can’t be? It sounds like an unlikely task, but ideas about how it can be achieved exist.
With a Bayesian approach, trust can be measured perspectivally. If Alice trusts Bob, and Bob trusts Jane, Alice is more likely to trust Jane as well, though perhaps not as much. This dynamic is already at work informally in social networks. Jane has no intrinsic trust score floating above her head; rather, different people will trust her to different degrees, depending on their relationship to her. A Ripple trust metric can work in the same way.
When people connect on Ripple, they declare a trust relationship which is quantifiable, because it is expressed in terms of credit limits. If Jane trusts Bob with 50 BTC and Adam with 25 BTC, it’s possible to infer that she trusts Bob’s credit twice as much as Adam, and so on. These decision can then be used to weight calculations about who Jane can trust, based on who Bob or Adam trusts. A perspectival trust metric, as I’ve argued in this post, might be the missing ingredient which helps Ripple to scale beyond personal trust relationships, as it needs to.
Could it happen?
Perhaps the way money has worked in the past–as a combination of commodity and promise–is the key to creating a successful monetary system. A Bitcoin-Ripple mash-up would combine the scarcity of Bitcoin with the power of credit creation. All the technology to build such a hybrid model already exists, after all, so it is perhaps only a matter of time until someone creates it.
Philip Coggan, Paper Promises (2011), ch.3 “Going for Gold”
Satoshi Nakamoto, Bitcoin: A Peer-to-Peer Electronic Cash System (2009)