Skepticism About Algorithm-Based Stablecoins
A stablecoin is a native token that intends to maintain a 1:1 exchange rate against the USD at all times. One approach to achieving a 1:1 exchange rate is to put in place protocols with mechanisms (“algorithms”) claimed to keep the market value of the coin within 1±δ of 1 USD, where δ is some small number. Here, I describe the general way in which these algorithmic incentive schemes are intended to work and evaluate whether they can achieve stability.
One class of algorithmic incentive mechanisms uses positive rewards. Holders are rewarded if they hold on to their stablecoin rather than putting it on the market, which could put downward pressure the stablecoin’s price. One variant of these mechanisms is called ``parking.’’ A stablecoin holder enters into a contract with the issuer in which the holder agrees to “park” (lock) an amount of the stablecoin with the issuer for a given period of time. In return, the issuer promises the holder a reward (“interest”) on the tokens parked. The reward is paid in the form of additional new tokens received when the stablecoins are unlocked. A useful way to think about this reward mechanism is that it is similar to the role time deposits and CDs play for banks with regard to deposits. Both have the effect of reducing the supply in the market.
A second variant on this reward mechanism is that the holder ”sells” the stablecoin back to the issuer for a different, distinct token (usually called a “bond” or “seigniorage share”) that is also issued by the issuer of the stablecoin. This bond promises the payment of a given amount of stablecoin to the bond holder in the future. This future payment is principal and interest. Stablecoin supply is reduced because the stablecoin purchased by the issuer with the bonds is burned.
What can be achieved with positive rewards often can be also achieved by negative rewards. Rather than positively rewarding agents who do not put their stablecoins on the market, as in the case of parking or seigniorage shares, another class of stabilization mechanisms negatively reward agents who do not reduce or destroy their holdings should the price decline. Such mechanisms work if stablecoin holders have some ``skin in the game,’’ usually in the form of claims of revenue earned through stablecoin issuance. Holders lose some of their claims if the price does not move back toward the peg.
For negative reward mechanisms to be successful they must:
(i) get incentives right — the size of the loss to holders from misbehaving and not reducing supply must be larger than the gain from misbehaving and attempting to sell their coins before the price falls further, and
(ii) overcome the coordination problem (multi-player version of the Prisoner’s Dilemma) — the reward to an individual stablecoin holder making the necessary reduction in supply when every other holder also makes the necessary reduction in overall supply must be greater than the reward to an individual holder from holding out and keeping supply unchanged when other holders reduce supply.
My Skepticism
The stablecoins that use algorithms for stablilization are all fiat and self self-contained. They are fiat in that their stablecoin is not backed by any asset that exists outside the protocol. They are self-contained in that the asset that algorithm uses for stablization is fundamentally related to the stablecoin protocol; it is not related to an asset outside of the protocol. It is for this reason that in a previous post (“Stablecoin Protocols” Institute of Decentralized Economics) I referred to the protocols that use them as ``crypto central banks.’’ Actual government central banks change the quantity of their money by buying and selling securities denominated in the same units of the currency that they issue.
It is well known that fiat currency systems are fragile. Fiat currencies have no value in and of themselves. (In economics jargon, they are “intrinsically useless.”) Whether a fiat currency is valued depends on expectations. If people expect that in the future there will be some entity that will accept the currency in exchange for goods or assets, then it will have value today. However, if at any point people come to believe that that no entity will accept it in the future, then it will lose value immediately.
It is the fragility of fiat currencies and the fact that these algorithms are self-contained that drives my skepticism of these algorithms. The future interest offered by parking or bonds is only valuable if the stablecoin itself is valuable in the future. ``Skin in the game’’ losses are only a threat if the stablecoin is valuable, so that there is something to lose. In other words, stabilization algorithms will work if stablecoin holders believe they will and the stablecoin will be valuable in the future. Stabilization algorithms will not work if stablecoin holders do not believe they will because the stablecoin will not be valuable in the future.
It all depends on expectations, and expectations can change for good reasons or for no reason at all. Hence, my skepticism that all of these algorithm-based stablecoins will succeed.
Warren Weber, Owner, Webereconomics and former Senior Research Officer at the Federal Reserve Bank of Minneapolis.
