Does the Frax Decision Signal the End of Algorithmic Stablecoins?

Recently, the stablecoin Frax voted to sunset the key innovation of their protocol: partially algorithmic collateralization. Does this mean that the age of experimentation with algorithmic collateral is over?

If not, why would anyone keep trying to play with the fire that is algorithmic stabilization?

But First, Let’s Talk About Decentralization

Frax is not a decentralized stablecoin by the standards we have set for ourselves at Icewater. It is pegged to the dollar and uses dollar denominated collateral. So, the whole risk-transfer concept of a truly decentralized stablecoin is not as critical.

Let me explain that concept a little more. The first key challenge of a decentralized stablecoin is determining what stability even means. This is much harder if you can’t just peg to the dollar. The second key challenge is achieving stability. Truly decentralized assets like BTC or ETH are not stable, so how do you make them stable?

If you peg to the dollar, and use dollar-denominated collateral like USDC, you are basically avoiding both of these hard questions and simply extending traditional finance (i.e., the dollar) onto the blockchain. This is might be useful and profitable, but it’s not really a decentralized approach.

So Frax has announced that they will be 100% collateralized. If you peg to the dollar and use safe, dollar-denominated collateral, 100% might be enough. But when you are trying to create stability out of unstable collateral, it’s not enough. You need excess collateralization.

But excess collateralization becomes problematic because people can stake crypto assets. If they have to stake multiples of the amount of stablecoins they get in return, the loss of staking revenue becomes a significant opportunity cost for anyone putting up collateral. This cost can be mitigated if the protocol itself stakes some of the collateral, but this introduces additional risk.

Another Consideration: Inflation-based Seigniorage

Let me distinguish between two types of seigniorage: growth-based seigniorage and inflation-based seigniorage. Growth-based seigniorage changes the value (in Utils) of outstanding stablecoins. Inflation-based (i.e., intrinsic) seigniorage is seigniorage that is based on (or causes) inflation. That is, it isn’t associated with the number of outstanding tokens, not the total value of the tokens.

The reason this is relevant is that growth-based seigniorage should cause an increase in the value of the collateral, but intrinsic seigniorage should not. There is not a priori reason to assume that the value of volatile collateral like ETH will decrease with inflation of the dollar. So for a given amount of volatile collateral you can (potentially) continue to mint tokens at a rate sufficient to ensure a modest level of inflation and provide this revenue to shareholders (i.e., currency-as-a-service). If you have a dollar peg and dollar-denominated collateral you don’t really have intrinsic seigniorage.

What this means for a truly decentralised seigniorage shares model with inflation is that the shares have some intrinsic value based on future expected revenue (controlling for collateral volatility) which can (and should) be used to compensate for the problem of lost staking potential for the collateral. This revenue is based on the outstanding supply, as opposed to growth, so it is somewhat more stable than traditional seigniorage.

Using Shares for Excess Collateral

The upshot of all this is that shares can be used for excess collateral. That is, you can have 100% collateralization with volatile collateral like ETH, as opposed to requiring something like 200% collateral. This is a bit distinct from what Frax was initially trying to do, i.e., using shares to enable less than 100% collateralization.

As we discussed in previously, if you don’t have seigniorage revenue, using simple volatility tokens as collateral does not work as a method of increasing excess collateral. Furthermore, if your seigniorage revenue is growth-based, the value of your shares can collapse to zero almost immediately if people expect the growth to disappear, as we saw in the Terra Luna Case.

Another way to look at this is that you can use the carrot of seigniorage (including intrinsic seigniorage) to encourage the guardians (i.e., the risk-takers) of the system to add additional collateral if the value of the collateral falls (and hence, the collateralization falls below 100%).

So, my conclusion is that seigniorage-shares style algorithmic stabilization is not completely dead. It just needs to be reimagined as a way to deal with the volatility of truly decentralized external collateral, rather than as a way to replace external collateral entirely.

For a protocol like Frax, i.e., a stablecoin that is just an extension of the dollar economy, algorithmic collateral is just a PR problem. But for a fully decentralized stablecoin, it can be an important piece of the puzzle.