Algorithmic Stablecoins: Value Propositions and Performance Indicators Beyond Peg Maintenance

  • Commentary
  • January 22, 2021

Prior to the highly-publicized increased institutional interest in Bitcoin the blockchain industry is currently experiencing, much of the general public’s attention being paid to distributed ledger-based assets was drawn by stablecoins. The rapidly expanding asset class has drawn both praise and resistance from those assessing its potential role in the greater financial ecosystem, with proponents touting more efficient and auditable transfers of value and critics pointing to the potential pitfalls of granting largely unregulated technology firms key roles in financial infrastructure. Hitherto, such debate has almost exclusively revolved around the trajectory of collateralized stablecoins, rather than their algorithmic counterparts. Yet, in recent weeks, growing attention has been paid to the algorithmic stablecoin subsector, and Smith + Crown expects this trend to continue as general interest in stablecoins proliferates and evolves. In response to recent commentary by Flipside Crypto, this dispatch proposes a more nuanced evaluation of algorithmic stablecoins, considering value propositions and key performance indicators beyond mere peg maintenance.

Stablecoins, as they exist today, can be broadly classified as belonging to one of two categories. Collateralized stablecoins are those that are backed by pooled reserves and derive their value from the promise that they can be redeemed at any time for assets equal in value to the asset to which they are pegged. These include Tether (USDT), which is theoretically backed one-to-one by bank reserves of US dollars, and Dai, which is also pegged to the US dollar but backed by cryptoassets such as Ether and Wrapped Bitcoin held in smart contracts. Algorithmic stablecoins, however, are not backed by any external assets but rather seek to maintain their pegs through algorithmically-governed, cryptoeconomic mechanisms, nearly all of which involve the manipulation of supply in order to drive market exchange rates in a given direction. One such asset is TerraUSD, a seigniorage share-style stablecoin that uses a second, non-pegged crypto asset, LUNA, to ‘absorb’ its volatility and maintain its peg. At a high level, when the price of TerraUSD falls below its fiat peg, the protocol allows users to burn TerraUSD in exchange for newly minted Luna—at a fixed rate higher than the current market rate—and realize a profit equal to the difference between the market value of the burned TerraUSD and that of the newly minted Luna. This process is intended to increase, demand for TerraUSD while its supply contracts, moving its equilibrium price closer to its fiat peg. Conversely, when the price of Terra surpasses its fiat peg, a reciprocal procedure is enacted: newly-minted TerraUSD is exchanged for Luna at a rate preferable to that of the market, expanding supply while attempting to decrease demand for TerraUSD (as users opt to purchase LUNA instead).

Despite some criticism and controversy, the adoption of collateralized stablecoins has far outpaced that of algorithmic stablecoins. Collectively, Tether’s USD Token (USDT), Circle’s USD Coin (USDC), and MakerDAO’s Dai (DAI)—the three most popular fiat-pegged stablecoins by trading volume—operate on nine blockchains and have a combined market capitalization of $31.5B, roughly 3.5% of that of the overall cryptoasset market. In contrast, algorithmic stablecoins Ampleforth (AMPL), TerraUSD (UST), and Empty Set Dollar (ESD) have a combined market cap of $479.7M, representing just 0.04% of the total cryptoasset market capitalization. Many, including Flipside Crypto, reason that this is largely due to algorithmic stablecoins’ relatively poor ability to maintain pegs to their underlying assets. For example, while the price of Tether (USDT) failed to deviate from its $1 peg by more than $0.05 over the past twelve months, the algorithmically governed, USD-pegged stablecoin Ampleforth (AMPL) rose to a high of $2.79 during the same time period. As such, many are observing the cryptoeconomic models of new entrants such as Basis Cash and Empty Set Dollar, which implement a second stabilizing asset and a DAO, respectively, to evaluate whether they will provide more effective in demonstrating the ability of algorithmic stablecoins to maintain their pegs to an extent similar to that of collateralized stablecoins. While peg maintenance is arguably a core goal of all stablecoins, Smith + Crown believes that any evaluation of algorithmic stablecoins solely on this basis is perhaps shortsighted. Fundamental differences in economic design are such that the extent to which peg maintenance determines and indicates value in collateralized stablecoins is far greater than in algorithmic stablecoins. Namely, the number of relevant players in algorithmic stablecoin ecosystems far exceeds that in collateralized stablecoin ecosystems, wherein there exists only one relevant perspective: that of the end-user. End-users transacting in a stablecoin are reliant on peg maintenance for predictable value transfer and, as theirs is the sole perspective relevant in collateralized ecosystems, it is therein appropriate to use peg maintenance as a primary indicator of value.

However, with algorithmic stablecoins, there are a number of other relevant players from whose perspective value the value of the stablecoins’ ecosystem can be evaluated. First, to arbitrageurs and speculators exposed to the stablecoin and/or its stabilizing asset, there exists a significant opportunity for profit. As such, value theoretically lost due to peg deviation is at least partially counterbalanced by the resulting opportunities for these parties to generate relatively low-risk returns from participating in the rebalancing process, which could optimistically be considered a shift in utility rather than a net loss thereof. Further, the fact that an end user’s losses from peg deviation in such a scenario only occur should the end-user sell the stablecoin prior to successful rebalancing highlights a key metric by which an ecosystem’s value can be measured: time required to rebalance. The concept of time value of money holds that assets expendable in the present are worth more than assets expendable in the future. Assuming a rational end user whose stablecoins have fallen below their peg will wait until the peg is restored to spend them, and that the peg will indeed be restored at some point in the future, the stablecoins’ net present value to the user only declines as expected rebalancing time increases. In this scenario, the magnitude of a given deviation ultimately only determines value to the extent that it impacts the time required to rebalance, making the latter the primary indicator of value. However, it should be noted that this logic is not the only relevant factor, as the longer a stablecoin’s expected rebalancing time is, the greater risk there is of its holder incurring opportunity or other unavoidable costs whilst waiting for the peg to be restored; there exist scenarios in which the opportunity costs become so great or other costs become so immediate that the end user must liquidate prior to rebalancing, in which case deviation magnitude is likely to become equally or more consequential than rebalancing time.

All of this, of course, occurs within the context where it is assumed that the stablecoin in question will eventually return to its peg. The concept of a death spiral, where buyers prove reluctant to step in and stabilize the value of a declining stablecoin out of expectation that the price may move lower still, until eventually the price crashes to zero, represents an existential risk to these systems. That such declines have been avoided by Terra and its peers to date is clear validation that these systems appear to work, but fails to serve as definitive proof that they will continue to do so. How this impacts long-term price and value expectations for both end-users and speculators interested in stabilizing such systems, as well as long-term holders with exposure to their underlying volatility, remains an important issue.

The potential occurrence of a death spiral constitutes a relevant consideration not only from the perspective of end-users and speculators, but also from that of those who have a vested interest in the evolution of cryptoeconomic design. Measuring the relatively new algorithmic stablecoin models as ‘proofs-of-concept’ rather than solely by their current utility, there exist several additional accomplishments that can be similarly demonstrative of value. First, while collateralized stablecoins are minted proportionately to the growth of their reserves, algorithmic stablecoins often rely on the market value of a stabilizing asset for overall growth. That is, for TerraUSD to become valuable, LUNA must first have value. Thus, catalyzing ecosystem value such that a stablecoin may even be brought into existence demonstrates early participants’ perceived valuation of the network and its cryptoeconomic merit. Second, independent algorithmic stablecoin blockchains with stabilizing assets face a challenge inherent to the pursuit of their goals. For example, because LUNA theoretically absorbs the volatility of TerraUSD, the network’s staking validators, whose ability to earn rewards relies on the price and availability of Luna, experience the immediate consequences of Luna’s volatility. When the value of Luna, and thus value of staking rewards, is increasing, transaction fees are decreased and repurchased Luna is burned at a lower rater. In such networks, incentivisation of staking levels and validator participation sufficient for network security poses a cryptoeconomic challenge. As such, staking rates and validator activity may also represent key metrics by which to measure a network’s value.

Ultimately, there exist numerous determinants and indicators of value in algorithmic stablecoin ecosystems. The appropriate consideration and interpretation of these value propositions and indicators will inevitably assume forms unique to individual networks, and individual participants, and should be assessed on an individual basis. We present the above considerations not as a comprehensive framework for evaluation, but rather as demonstrative of the need for a more nuanced and complex discussion around value than perhaps currently exists.