I recently re-read John Pfeffer’s excellent paper: An (Institutional) Investor’s Take on Cryptoassets (the full paper is linked at the bottom of his blogpost). A quick cliff notes of John’s view of protocol/utility tokens is below, though I encourage you to read his paper to get the full picture:

  • The total value (M) of a utility or protocol token (eg. ETH) can be calculated by the formula PQ/V where P is the total price of the utility provided, Q is its quantity, and V is the velocity of money (eg. how fast it moves through the system. A token’s individual value is hence M/T where T is the total of tokens.
  • For protocol/utility tokens, John argues that PQ is a function of the total compute resources required to deliver the utility itself: network, compute, etc and no more.
  • Over the long-term horizon PQ should be low as the value of all these resources (through Moore’s law, etc.) depreciates over time. Cf. the cost of a gigabyte (GB) was more than 400k in 1980 and less than 2 cents in 2016.
  • Furthermore V should be high because the currency here is digital, so moves at the speed of light. Hence PQ/V is low, essentially the marginal cost of providing the utility divided by the velocity.
  • This makes utility protocols (read: any cryptoasset that is not the ultimate non-sovereign store of value) a pretty unappealing investment. Note the value created by these protocols can still be very high, but most of that value is claimed by the users, not the creators of the protocol itself.

I’ve been thinking about the relative value of native cryptoassets (here “native”, as opposed to tokenized securities), and I come back to agreeing with John’s perspective. I think ICOs for utility tokens amount to public good (eg, charity) fundraisers, which will be the subject of a later post.

That said, I thought it might be interesting to

  1. Assume John’s model for evaluating the value of a utility protocol is accurate, namely that M = PQ/V
  2. Figure out if there’s a scenario where utility or protocol tokens (eg. Ethereum) can actually be worth a lot

To wit, in order for a utility protocol to be worth a lot, the following must be true:

  1. PQ is high, which implies (a) P is high and/or Q is high and (b) P ≠k/Q for some constant k
  2. V is low
  3. Forks are hard

One case where 1–3 are likely to be true is a universe where cryptoassets are illegal in a large enough subset of the world, and where ownership & production thereof are prosecuted brutally (eg. prison time, death sentences, etc.) In that universe:

  • P is likely to be high. Even if general compute, network, etc. costs are depreciating, the aforementioned risk in allocating these resources to a protocol would be reflected in their price.
  • V is likely to be low. For protection, cryptoasset networks would likely use multiple hops across decentralized Tor-like networks to evade prosecution, and likely layer-on privacy primitives such as zk-SNARKs which would both increase computational complexity (PQ higher) and decrease velocity.
  • Forks are likely to be hard. One reason might be that the cryptoasset community would have reason to be wary of forks as potential honeytraps by hostile government actors. In this universe, networks that have been around longer (and hence with fewer arrests or negative consequences) develop a reputation as safer, thus giving them the ability to extract rents higher than their newer forks.
  • Q is the hardest to imagine being high, but it seems more feasible to imagine a world where Q ≫ k/P. In a world where enough wealth is embodied in cryptoassets (we’re pretty close) and several large countries (eg. China) were to ban cryptoassets, this would likely hold. If P is sufficiently high (think black market prices) and Q is modestly high, PQ is high

Turns out John alludes to this on page 3 of his paper with the sentence: “use cases will be limited to dematerialised networks where the value of decentralization, censorship-resistance and trustlessness is high enough to justify the inherent inefficiency and redudancy of the consensus mechanism”

AKA: (crypto) doomsday scenarios.

To be fair, Bitcoin is also likely to be worth a lot in doomsday scenarios, so rather than hold a basket of cryptocurrencies, a better asset mix might be Bitcoin + whiskey + bullets.

What are other scenarios where PQ/V is high & forks are hard?

If you’re curious where the crypto world is going, I’m always excited for a chat.