Neither Vitalik, nor Bitcoin bugs: towards a sensible cryptoeconomics, P1: Against overempasizing mechanism design
In this series of three articles, I will consider and criticize the two dominant approaches to what has been termed “cryptoeconomics”, or economic analysis applied to the crypto space: the heavily neoclassical mainstream-inspired perspective of people like Ethereum founder Vitalik Buterin and the cartoonish Austrian economics of the hardcore fans of Bitcoin. I will also show how a genuinely Austrian, realist approach to economics could offer clues as to how the central current problem of cryptocurrencies — how validators and core network developers are to be rewarded — could be solved.
In this episode, I will start with Vitalik’s approach to cryptoeconomics. What is wrong with it? In two words: mechanism design. Like the neoclassical mainstream economists whose academic output has inspired him, Vitalik treats economic phenomena like a pure technician. His latest foray into cryptoeconomics suggests that everything in the economy is a mechanism defined as “a tool that takes in inputs from multiple people, and uses these inputs as a way to determine things about its participants’ values, so as to make some kind of decision that people care about.” Vitalik further adds that “in a well-functioning mechanism, the decision made by the mechanism is both efficient — in the sense that the decision is the best possible outcome given the participants’ preferences, and incentive-compatible, meaning that people have the incentive to participate “honestly”.
The price system may seem to an engineer like Vitalik to be just such a mechanism. It aggregates the preferences of consumers and reconciles them with the capacities of producers to cater for them, giving in equilibrium a set of quantities and prices of goods and services purchased.
In reality, however, the market process through which prices form does not function remotely like its equilibrium metaphor. Prices change through individual decisions of entrepreneurs making new offers to consumers provided that the consumers accept them. No agent (à la Walrasian auctioneer) or magical machine reshuffles the mix of goods and inputs used to produced them as well as the prices every once in a while.
One could say, however, that cryptoeconomics is not about the economy-wide price system, anyway, and that the adequacy of neoclassical metaphors about it is not very important. I would argue, however, that the mechanism design approach to economics fails in other areas, too. One of those areas are the so called public goods and public bads.
Vitalik has been arguing for some time that a strong case can be made that blockchain communities underinvest into public goods such as core blockchain network protocol development. In the aforementioned recent article, he also implies that governments creating caps on supposed negative externalities is not essentially different from their enforcement of ordinary property rights.
Vitalik’s ultimate rhetorical objective is to legitimize and promote research into certain types of community-wide decisions that his opponents (largely from the Bitcoin bug camp) strongly reject. One of such key decisions is to reward validators and core protocol contributors through continued native token (e.g. BTC or ETH) issuance.
The fantasy of “public goods”, and bads
Consider three hypothetical situations:
- If the Ethereum Foundation organizes a conference at the city where Alice lives, Alice will happily purchase the ticket and attend.
- If a new Lambo cost less than $10,000, Bob would buy one without hesitation.
- If people could be credibly excluded from consuming most police services, there would be no need for funding policing through taxation, as people would voluntarily paid as much as necessary for policing.
Which two of these examples belong to the same category? At first sight, one is tempted to say 1 and 2 because there does not seem something inherently impossible about new Lambos costing less than $10,000 or the Ethereum Foundation organizing a conference at a city, unless, perhaps, we are talking about Pyongyang.
I would venture to say, however, that examples 2 and 3 are actually more similar to each other than to 1. That it is hard to exclude people living in a certain area from enjoying many of the benefits of policing is a consequence of the way the world works. In order to be effective, the police cannot verify whether every person who appeals for their services is eligible before going to the scene. Similarly, the fact that new Lambos do not cost less than $10,000 is not incidental. It is ultimately explained by the fact that the scarce inputs that go into their production are also used for other purposes that a lot of people value. At the same time, the decision at which city to organize a conference is largely a matter of free choice by the Ethereum Foundation’s leadership.
According to the mainstream theory, public goods are goods that are non-rivalrous in consumption and non-excludable. The first feature means that the quality of a good does not deteriorate with the increasing number of people consuming it. Non-excludability implies that it is hard to prevent certain people (usually, at a given location but sometimes, globally) from consuming it. According to the proponents of the theory, the combination of those features means that policing is highly likely to be underproduced.
Policing from example 3 is an imperfect example of such a good. It is imperfect because the quality of policing may well deteriorate if there are too many people that need protection in a given area. However, it seems close enough to the ideal type of a public good that the conclusions still apply. After all, there are no perfect circles or squares in the real world but one will predictably fail to put a roughly square peg of a certain size into a roughly round hole of a certain radius.
But the 3 examples we discussed in the beginning suggest that there is a different major problem with the theory of public goods. Remember the supposed implication of the public good concept that such a good is ceteris paribus likely to be underproduced in the real world. Underproduced compared to what? “Compared to the consumers’ true preferences,” a mainstream economist would respond. This would imply, however, that Lambos are also underproduced, and almost everyone would agree that this conclusion is ridiculous.
It seems odd to use the Lambos analogy to discredit the public good theory but why should the facts that Lambos require costly inputs have some sort of a scientific, ontological priority over the fact that it is hard to exclude people in a given area from benefiting from policing without paying for it? They certainly should not, if we are to avoid special pleading. The neoclassical concept of public goods is, therefore, incoherent.
Similar problem plagues the notion of purported negative externalities like CO2 emissions. To Vitalik, the government creating the caps on them is fundamentally akin to its enforcing the integrity of one’s apartment from the neighbor who would like to intrude there. However, in the real world, the vast majority of people do not treat the atmospheric concentrations of CO2 in the same way in their minds as they treat stuff that they own in the ordinary sense of the word. It is thus impossible for the government to deduce what maximum concentration of CO2 they would settle on if it were possible to have a process similar to the market provision of apartments. Just with government-provided policing, any CO2 emission cap will be an extra-economic decision by the social engineer.
Of course, the fundamental problem with the economic concepts of public goods and negative externalities does not necessarily imply that the relevant activities are not under- or overproduced compared to non-economic criteria like the ethical imperative to curb catastrophic risks. However, economically speaking, one cannot say that, for instance, protocol services will be underproduced without subsidies. In fact, the criticism of the public goods/bads theory points at the way that applied problem needs to be reframed.
In the same vein, it may well be that a certain minimum block reward may be needed to induce a given blockchain’s validators purely from the network security perspective. It just will not mean that this would reflect economic factors because it has no clear connection to consumer preferences and the allocation of scarce resources to satisfying them.
However, we do not have to be too pessimistic about the capacity of the market process to fund services provided that they are ultimately useful to consumers. The historical experience of money provision through free banking suggests just that.
The free banking analogy for rewarding core network service providers
To see what I mean, consider that it is not the first time we are facing a debate on whether a society-wide service like core protocol services needs to be subsidized. That problem arose in the 19th century with the debate on free banking and the desirability of continued new issuance.
A mainstream economist would certainly consider the maintenance of money supply to be a public good that in the absence of government involvement will tend to be underprovided. In fact, though not framed in this way, this is the founding narrative of the central banks’ control of issuance. Using the evidence from the workings of the US banking system prior to the 1913 creation of the Federal Reserve (Fed), the proponents of central banks have succeeded in convincing almost everyone who bothered to think about it that private money provision had massively failed.
However, those economists and economic historians (see, for instance, here, here, here and here) who actually bothered to study the US banking history in full detail and nuance as well as the alternative historical central-bank-free arrangements have convincingly demonstrated four things:
- Far from being a chaotic mess, a system of competing banks issuing inside money partially backed by the base money such as gold can maintain money supply without over-issuance and sudden massive deflation episodes.
- The US system prior to 1913 cannot properly be called an example of free banking because, although a central bank was absent, banking was severely restricted in all kinds of ways, first, by state governments and then by the national banking regulations aimed at funding the North’s Civil War effort.
- Other historical examples of competitive banking involving decentralized inside money issuance (especially Scorland and Canada) strongly suggest that free-banking system helped avoid money shortages while not resulting in runaway inflation episodes or even artificial investment booms à la the construction bubble of the 2000s. Most strikingly, in Canada, even during the Great Depression, not a single bank failed, despite the massive crisis emanating from its biggest export market’s collapse.
- The Fed that replaced the somewhat competitive banking system has actually probably performed worse on the key monetary policy metrics than even that highly stifled system.
How could free-banking system work without explicit mechanism design through bottom-up competition? The answer is adverse clearings and loans. Competing banks used gold reserves that people deposited with them to issue their own bank notes that they loaned out. Just like banks today they used only fractional reserves relying on the fact that it is extremely unlikely that all or even the majority of the demand depositors will demand withdrawal at the same time. Competing banks further came to accept the notes of their competitors to those who submitted those notes to them to then present those notes to the issuing banks at the clearing house. The overissuing banks would thus tend to get into trouble pretty fast.
What is the relevance of free banking to rewarding protocol service providers? The key feature of the free banking system (that even its proponents usually fail to emphasize in contrast to the adverse clearings mechanism) is that banks did not just run a monetary system. They added new money to the system as loans, as bets on investment projects that could potentially create value for consumers. This is what is currently almost totally missing from the cryptocurrency issuance design. Validators or developers are rewarded for maintaining the core blockchain networks but there is almost no connection between what they do and creating value certified by prices. The reward schedules are economically arbitrary in a way in which the earnings of competing free banks were not.
In the next instalment, I will consider the major ways Bitcoin bugs are mistaken about cryptoeconomics and why their absolutist no network currency inflation stance is wrong in light of the free-banking experience.