What is Ethereum for?

A big picture view of blockchains

Thanks to Transmissions11, for the correction of errors

I am beginning now, prompted by friendship, to really think about cryptocurrency and Ethereum as technologies. I am not an expert on the mechanics of Ethereum, and can’t offer any solutions to the technical problems; it may yet be helpful, however, to think through from the perspective of an economist what Ethereum is trying to do, what obstacles keep it from doing so, and what can be done about it.

i. What is Ethereum?

Ethereum is a blockchain technology which, unlike Bitcoin, is programmable. A blockchain is essentially a series of distributed ledgers, linked together with every ledger which happened in the past. These can’t be altered without everyone else involved noticing that such alteration occurred. The currency which Ethereum uses to run transactions is called Ether, or ETH: you pay a “gas fee” to pay for the cost of computing (this gets burned out of existence). Transactions resolve in blocks every 12 seconds. A block proposer (chosen at random from a commonly agreed upon source of randomness) is responsible for ordering the transactions as they come in. To be a block proposer, you must put up Ether — 32 to be specific — as collateral against fraud, malfeasance or inactivity. Everyone who runs a node — which requires no more than a regular computer — is able to verify the accuracy of the transactions being done. This is Layer 1, or L1. Layer 2 (L2) are independent blockchains with different rules which ultimately derive their security from posting its data and resolving disputes on Layer 1. The main purpose of these is to increase the number of transactions per second which can be done — because Ethereum has to be verified by a decentralized network, it can only be as fast as the lowest common denominator.

It being programmable means that you can do so much more with it than Bitcoin — you can have it run self-enforcing contracts, called “smart contracts”. For Ethereum to be successful, it is going to have to be better than a regular old contract, at least in some uses — otherwise, you would just use our normal payment systems.

ii. When is it useful?

It seems to me that the primary use-case of blockchain is for when enforcement is limited but contracts are specifiable. The “specifiability” of contracts is something which comes up a lot in theory of the firm literature — it refers to how well a contract is able to lay out all possible contingencies and cleanly resolve them. Something like an employment contract, for example, is not very specifiable — both parties want to retain flexibility in case there are unforeseen contingencies. There are many degrees of effort and quality in doing the work, and no way for any computer to easily verify it has been done properly. Ethereum is not like that — contracts, once set in motion, resolve themselves irrevocably. This makes it ill-suited for times when the provision of one or both sides of the contract is dependent on human input — if I make a contract for you to provide me with literal bricks in exchange for Ether, then someone will have to verify that the bricks were delivered in the right place at the right time. 

Something that remains entirely on the internet is extremely well-suited. The good to be delivered can be perfectly specified, and for many goods it can be hard to get enforcement. The biggest thing that comes to mind is pornography. Many porn companies have a hard time getting payment processors to work with them, primarily because of chargebacks (with some fear of legal liability mixed in). Many people get embarrassed that they have bought porn — given that payment processors have a fee of perhaps 1.5 percent, one chargeback obliterates dozens of good, paying customers. The immutability of transactions on ethereum gets rid of this. We can avoid scams by charging for time consumed, as a sigmoidal function — you have some period at a low price, to verify the quality of the goods, and then it begins paying full price.¹ Or, you can simply rely upon the regular mechanisms of reputation — people will not find it in their long term interest to scam for the same regular businesses don’t.

Ethereum can be used as a payment system for in-real-life goods, but its advantage there is mainly in working around laws. In this case, enforcement is scarce simply because it’s not legal. The drug trade springs to mind, but far larger than that are customers, largely in the developing world, who are denied access to banking services due to know-your-customer laws. These laws combat money laundering by raising the fixed cost to have a bank account, which has the drawback of excluding most of Africa. The direct cost runs to 37 billion dollars a year, and the indirect effects destroy $300 billion a year. Ethereum could conceivably replace the conventional payment channels, although you would run into great difficulties in withdrawing it. (The United States has a habit of arresting people who flout banking laws). 

Ethereum would also need to expand capacity in order to take over this role. This is a much harder problem than it would at first appear. In order for Ethereum to remain decentralized, everyone running a node has got to be able to verify the chains. If you increase the number of transactions per second, then you raise the requirements to run a node, thus making it less decentralized. As such, L1 only handles 14 transactions per second. Putting it on L2 using something like an optimistic roll up (which is basically just handling the transactions mostly off chain, and only submitting a summary, with fraud guarded against by adding a delay and a bounty for anyone who proves fraud — it relies on a trusted sequencer, hence the “optimistic”) increases the number of transactions, but makes it easier for governments to regulate — obviating one of its main use cases. Still though, it is surely possible for enough sequencers to evade government regulation to provide banking services in Africa.

That is the use which gives me the most hope, and convinces me that cryptocurrencies can be very good. There’re so many predatory, fraudulent scams and grifts in crypto that it can be easy to lose hope. What I would want to do is provide rainfall insurance to farmers in Africa on a blockchain platform. The conditions under which it would pay out (insufficient rainfall) are almost perfectly specifiable. The farmers do not have to fear that I would run off with their money, because the whole transaction will be set in motion and resolve without any of our inputs (besides satellite data, or perhaps a certification by a government agency). It’s a big fucking deal. There are enormous welfare gains out there just from allowing people to choose the crop mixture which maximizes long-run output. We must never lose sight of our ultimate goal, which is the eradication of global poverty. 

1.  I am not sure that this is the optimal pricing method, however. Grant that a particular sex worker has some degree of market power — this seems reasonable, many people want a  specific person. If the person can prevent resale, then the optimal method to sell the goods is with a two-part tariff — you charge a lump-sum fee to get in, and then a charge per unit consumed. It is an empirical question, I suppose, whether the second condition fails.