Robert Mundell pioneered the economics of optimal currency areas in the 1960s. His work was couched in two key assumptions:
- a central bank as currency issuer with defined economic goals (e.g. inflation, unemployment) and the power to control money supply in order to influence economic wellbeing, and
- somewhat homogenous economic regions, speak: distinct geographic areas.
Within cryptocurrency, there is a strong sentiment against #1 (although don’t underestimate the potential of central banks entering the crypto game eventually), and by design, adopters of a certain cryptocurrency don’t inhabit a defined geographical region. So if we assume that in the short and middle run, multiple cryptocurrencies will survive, the question then becomes what kind of adoption patterns we should expect, and also what adoption patterns we should consider optimal.
We could also take a look at the long run and contemplate an economy that is entirely based on cryptocurrencies, but for now the more pressing discussion is that of a mixed economy. We should expect that most everyone still receives their wages and pays their taxes in fiat money. Cryptocurrency is then restricted to industrial adoption (inter-firm settlements) and commerce (esp. buying merchandise online, but also in the future offline).
Mundell’s four criteria were labor mobility, capital mobility, risk sharing and business cycle homogeneity — criteria clearly driven by the goal to use currency as a means to reduce the negative effects of regional economic shocks. Setting aside the question who will take on that task, none of these considerations played a role in the creation of cryptocurrency. So to determine optimal currency areas in cryptofinance, we need to start from scratch.
The first question we have to ask ourselves is: will a single, global cryptocurrency suffice? In this case the current proliferation of so–called alt–coins is just the emergence of multiple candidates all jockeying in a winner–takes–all–race to become that single currency. Given the strong positive network effects inherent in currency (any currency is only worth something if we find someone who is willing to take it off our hands), the intuitive answer to this might be yes. But I will argue against it and make the case that multiple currencies cannot coexist in a socially beneficial equilibrium.
To explain this, let me split the network effects of currency into three components.
- Preferences over the institutional framework of exchange. If we could keep the network of exchange partners constant, this component will guide us towards picking our preferred currency based on technical characteristics alone. Strictly speaking, this is not a network effect.
- Global network effects. This is the scaling component — we generally prefer to have as many other participants which whom we can interact. This is also because functional transactional networks require a minimum size to be feasible.
- Local (pairwise) network effects. In a network of millions, we will only interact with a very small fraction of the participants. This is even more true for so–called indirect networks, where we will interaction with concentrated intermediaries, such as exchanges.
From the technology standards literature we know that in a world where preferences are homogeneous or generally weak and network effects are global (homogenous) and positive, a single technology standard will emerge. To a certain degree, this is also a relevant finding for cryptocurrency. But we should be aware of the key distinctions.
The standard technology adoption literature (Katz–Shapiro, Farrell–Saloner) starts from an assumption of a two-player race, from which one winner will emerge. This is very much an endgame scenario, and there are multiple real life cases where multiple standards survived, without much harm to the community. A big reason comes from the modeling assumptions these papers made.
For one, they generally assume incompatibility and forced single–adoption: you have to choose one or the other technology. In cryptocurrency, it is perfectly possible and potentially even advisable, to hold a portfolio of multiple currencies.
For two, they mostly ignore local effects in favor of global effects. This is a tricky distinction, because in theory every additional participant increases our potential exchange value. But the clearer it becomes who the actual transaction partners are, the more local effects take over from global effects as we move from “potential” to “actual”.
In a lot of cases, this assessment can already be made ex ante. If I want to build a network for bank settlements, I simple don’t care if gamers are also joining the network, and I will quickly find out that our preferences, e.g. over network security and privacy, are very much misaligned.
A third assumption which is often misunderstood is stability of preferences. This is very much a point Brian Arthur and Paul David tried to make in the mid–1980s: that as technologies evolve their usefulness to us changes, and we might end up getting locked into a network that was useful to us in the past.
Where does this leave us regarding our original question?
Currencies have no intrinsic value, they are only valuable inasmuch as we are able to exchange them for something of value to us. But cryptocurrencies come within a technological framework. Bitcoin gives us access to a pseudonymous trading network outside (or so we assumed) government scrutiny. Ethereum gives us access to a global virtual transaction machine. Hyperledger gives us access to a consortium of serious industry players, etc.
The cost of participating in all or multiple of these networks is our ability to properly balance our portfolio of cryptocurrency assets. At this point this is not even necessarily a cost, and quite certainly not a prohibitive cost. As long as this hold, we should not expect cryptocurrency–as–technology–standard follow the path of a winner–takes–all race. And we might be better off for it.