By Kyle, Investment Manager@Bing Ventures

Abstract: The flexible redemption mechanism and interest rate design of algorithmic stablecoin should come first. Similarly, the algorithm should be kept in a box and not reduced to a growth gimmick for network effects.

Algorithmic stablecoins are stablecoins that rely entirely on algorithms to regulate supply and demand to create and maintain currency values. Does the collapse of Terra/UST, the largest and most mainstream algorithmic stablecoin system on the market, mean that algorithmic stablecoin is a false proposition? How did this financial experiment turn into a crypto disaster?

Terra’s “digital central bank” business

1 UST = $1 LUNA

In 2018, A Korean developer named Do Kwon established Terra, a public chain project with a stablecoin-making mechanism, and issued a so-called algorithmic stablecoin, UST, on this chain. The algorithm mechanism to maintain the stability of UST is very simple: 1 UST= Luna worth $1. If Luna is $1, it can be exchanged for 1 UST. UST and Luna belong to the relationship of “bi-directional burning and casting”. This means that when you trade in a Luna for a UST, there is one less dollar of Luna in the market, and when you trade in a UST for a Luna, there is one less UST in the market. UST is all minted by Luna, and the upper limit of issuing coins planned by Luna is 1 billion.

Source：Tokenterminal

How can supply and demand remain stable in this process? Because of the small fluctuation of UST price, there is arbitrage space for this mechanism. Since 1 dollar of Luna is equal to 1 UST. Therefore, when UST is higher than $1 (such as $1.1), arbitrageurs will burn their Luna to exchange for UST. In this situation, you lose $1 Luna to exchange for UST, but each UST is higher than $1, you gain $0.1. Conversely, when UST is less than $1 (say, $0.9), someone will exchange UST for $1 Luna to make a profit of $0.10.

If the price is higher, some people will sell it, and if the price is lower, some people will buy it. Therefore, the price of UST is basically stable. This mechanism of stabilising value through supply and demand is similar to the way countries maintain a stable value by adjusting money supply. But there is a premise behind this: Luna must have book value and liquidity.

Rely on 20% earnings to quickly build a financial empire

Compared with ETH, public chain tokens are supported by abundant application scenarios, whereas Luna and UST have no value supported by actual application scenarios. Although Terra has incubated a large number of projects within its ecosystem, the DeFi product matrix is almost entirely built around the arbitrage mechanism of UST. It can be said that without arbitrage and nominal high returns, there is no other circulation demand other than on-chain finance.

In order to retain these speculative users and attract more, Terra established Anchor, a franchise offering UST deposits and loans. Moreover, Anchor provides an annualised rate of return of fixed deposits as high as 20%. The 20% break-even investment is very attractive, leading to many users wanting to deposit UST into Anchor to enjoy the interest.

Therefore, buying UST became the first requirement, and Luna, which was able to cast UST, naturally attracted more and more users. In addition, Terra designed a series of other measures to encourage users to own UST and improve the liquidity of Luna, such as greatly expanding the use case scenarios and scope of UST, and enabling users to borrow money to UST at low or even negative interest rates by using Luna or Ethereum as collateral.

After this operation, everyone began to trade the Luna and UST in dollars in large quantities, and the two coins that had been on the air before began to have real value support. Over the past few months, Luna’s foundation, LFG, has been buying billions of dollars’ worth of Bitcoin and several other cryptocurrencies as reserves to allay user concerns about UST decoupling. At its peak, Luna had a market cap of $40 billion, while TerraUSD was close to $20 billion. Terra has, in essence, created a mechanism to pull large amounts of real money out of the market, backed by investment houses, market-makers and exchanges. It has then used the money to buy assets as collateral, and to make itself look increasingly credible.

The beginning and end of the anchor withdrawal

The asset reserve is exhausted

The demand for UST is almost entirely driven by unsustainable high-yield financial management. In Terra’s Anchor agreement, investors were able to get a nearly 20% return on UST due to various arbitrage opportunities, so where did this amazing 20% return come from?

The logic is similar to that of a real bank: the interest paid by the lender plus the investment income, which comes from Luna and Ethereum. But the interest rates Terra uses to attract loans are so low that they yield little. Moreover, after a bad start to the year, returns on collateral are far from 20%. What if you don’t get the promised 20% interest? Terra began to make up the shortfall with funds from the foundation. From 2021 to this year, Terra has made several replenishment of its reserves, and in February founder Do Kwon voted again to inject $300 million in reserves.

Terra established the Luna Foundation Guard as a secondary reserve linked to the UST/USD algorithm, rather than relying entirely on Luna, similar to the sovereign reserve of the state. In good times, emerging markets can use accumulated trade surpluses or foreign exchange sales to build up their reserves of foreign exchange assets, such as gold, the dollar or the euro. Then, if the country later experiences an economic recession and a currency crisis, it can defend the value of its currency by selling some of its previous reserves and using the proceeds of those sales to buy back some units of its own currency from those who sold it.

But whether bitcoin is the best digital reserve asset is debatable. It has also proved that the price of bitcoin has further declined after the current crash, completely failing to play the role of stablecoin asset reserve like USDC/USDT. The fund will always run out one day, and the market is unpredictable for anyone. A decline in demand for UST can lead to a negative feedback loop that ultimately creates liquidity problems for UST and LUNA, also known as a “death spiral”, in which a large amount of capital is pulled out of the Terra ecosystem, causing LUNA’s price to collapse, and UST decoupling. In traditional finance, we can find many currency and financial crises triggered by capital flight due to lack of confidence in the country’s capital markets.

The death spiral begins

This yield opportunity supported by an arbitrage mechanism has finally come to a dry point. A large number of UST are concentrated on Anchor to earn interest, but they are not in actual circulation. It is through constantly burning Luna that UST is injected with liquidity, so as to maintain the stable price of UST. The reserve fund is burning through. According to previous calculations, Anchor’s reserve is expected to be completely exhausted in June, and a decline in deposit returns is inevitable by then.

When the deposit rate of a wealth management product suddenly drops from 20% to 5%, the user’s first choice is to withdraw the money as soon as possible. However, when a small number of large customers start to do so, and when the UST prices show signs of instability, the mass market panic will start. When users suddenly start to frantically sell the UST that has not moved before and change back to Luna, the UST price will drop Anchor. Meanwhile, the increase of Luna supply in the market will lead to the drop of Luna price. However, when Luna’s price continues to fall and the market value approaches or is less than UST, people start to panic and sell UST in large quantities. That’s when the death spiral begins, with LUNA and UST all falling into the abyss.

Source: glassnode

On May 8, Terra withdrew 150 million UST funds from the pool for the new plan, causing UST liquidity to temporarily decline. Subsequently, large enterprises started a large-scale UST selling plan, resulting in the outflow of 2 billion USD UST from Anchor that day. As a result, a large number of UST suddenly appeared in the market, and the balance between supply and demand was broken. As a result, the price of UST began to be decoupled from the Dollar, and many people changed their UST into LUNA, which also led to the increase of the supply of Luna and the plummeting price.

On May 9, Terra announced that it would use $750 million of its bitcoin reserves to help stabilise UST, but the selling didn’t end. UST’s unanchor situation fell below $0.98, and many users panicked. At that time, LUNA’s price was around $60 (down 49.5% from its previous high of $119). Over the next 36 hours, LUNA fell below $0.10 and UST traded between the extremes of $0.30 and $0.82. This disabled the protocol redemption mechanism because all users are panicking and trading a UST for $1 worth of LUNA, causing a supply explosion that drives down prices even further. On May 12th, Luna continued its plunge to almost zero and UST fell below $0.2.

On the other hand, it also created risks for bitcoin price. Terra sold a large number of bitcoins to defend its peg to UST, and the rapid selling of tens of thousands of bitcoins directly declared the market into a bear market.

LUNA’s Crime and Punishment

Blind expansion of the founding team

Either way, UST and Luna are pretty much dead as far as trading prices are concerned. Personally, DK-controlled community pools and market makers helped hasten LUNA’s demise. If you have been paying attention to Terra, you will have noticed several community initiatives last year and this year to destroy the LUNA proposal and support the casting of more UST proposals, which were passed, with the claim that these UST will be used to support ecological projects.

I guess the LUNA project party should have reached an agreement with market makers such as JUMP and Three Arrows to obtain USDT funds by packaging UST+LUNA to purchase bitcoin and other reserve assets. If the community’s actual controllers, represented by DK, had been able to maintain their LUNA/UST ratio last year, they would have been prudent (in fact, it’s hard not to doubt their intentions) to support Terra’s expansion. Although afterwards many organisations said they had been defrauded, personally I am skeptical. The veracity of the voices in this will need to be investigated by regulators. Here is a recent post-mortem report from Delphi, which helped design Terra’s economy.

Source: glassnode

In the beginning Delphi was interested in Terra’s purported project through CHAI. CHAI was a payment network in South Korea with $1.5 billion in annual transactions and more than 2.5 million users. At the end of 2021, Terra added support for CosmWasm smart contracts, allowing third parties to build applications around these stablecoins. With its first application, Mirror, users can access synthetic real world assets and gain user attention. Terra was born entirely around the need to pay for the scene.

Delphi saw this and found that Terra’s ecosystem lacked a large number of builders and critical capital support, such as Swap and lending. Therefore, they sensed an opportunity and were directly involved in the design of Terra’s new products, with the fundamental purpose of helping UST to implement more scenarios. Now their original intention is not wrong, but Anchor has gone astray after completion.

Put the cart before the horse and accelerate the crash

Anchor’s initial 20% APY is mainly used to attract large clients, which is high but reasonable for the initial stage of a DeFi project. After all, at the very beginning, Anchor’s reserve was to pay the imbalance between the return rate of UST depositors and the return rate of mortgage assets. When operating in deficit, the agreement will allocate the ANC to incentivise additional deposits of collateral assets. However, if the output of Anchor wants to keep at a high level of 20%, it obviously goes against economic common sense and ignores market fluctuations.

This means that Anchor has failed as a rebalancing mechanism of UST and LUNA, and the inclusion of BTC and other assets into reserves has essentially illustrated the failure of TERRA’s algorithm stabilisation mechanism. Perhaps these reserves could have been used to defend the peg during a brief period of decoupling, but unfortunately they collapsed without setting up a bitcoin redemption mechanism.

Imagine if Terra could have started to wind down Anchor APY last year instead of using shock therapy, and created a more diversified external collateral mechanism, which could have delayed systemic risk in the network. Delphi believes that a high level of external collateral is necessary in the long run, unfortunately it is not growing fast enough compared to the UST supply and, coupled with the decline in the value of BTC reserves, the backlog of liabilities is too large to be protected.

In my opinion, Delphi obviously ignored the basic logic of public chain development at the beginning of mechanism design, that is: the prosperity of the ecological scene promotes the growth of stablecoin demand, not the growth of stablecoin supply promotes the false prosperity of ecology.

Virtue is not worthy, there will be disaster

Pure subsidies and high returns are not desirable

If there is a product today that offers a 20% annualised return with no cap on savings, you have to understand the risk behind it. If a public ecosystem relies on revenue-chasing games to support user growth, then we need to be wary of why such projects exist. It’s understandable that the project is trying to attract traffic through these tricks in the beginning, but when your system has accumulated tens of billions of dollars in assets, you need to start thinking about how to run more soundly.

If LUNA began to accept the partial mortgage mode at the beginning of the year and gradually adjusted the Anchor yield, it would increase or decrease the UST supply at a smoother speed and increase the mainstream application of UST in the wider market, such as payment and other scenarios, I believe the huge community would not collapse so quickly.

All algorithmic stablecoins govern the minting/burning mechanism between tokens and linked stablecoins. But when stablecoins are linked to fiat currencies, it is crucial to attract people to support the peg. It’s all a confidence game. However, there is no government in this game. Therefore, perfect mechanism design is needed for decentralized algorithm stablecoin, and risk control is particularly important, which is the natural enemy of high returns. The flexible redemption mechanism and high flexible interest rate design of algorithm stablecoin should be put in the first place.

Stablecoin requires supervision and disclosure

I still believe that algorithmic stablecoins have a big future. But this does not conflict with its acceptance of strict regulations and restrictions. The endogenous expansion of stable projects is inseparable from the risk reduction brought by supervision. Algorithms tend to imply volatility, and regulation is expected to limit volatility to boxes

Therefore, in my opinion, algorithmic stablecoin projects with small circulating market value still have good prospects for registered entities in mainstream financial countries. Algorithms can effectively supplement the centralized stabilisation of overcollateralised or fully collateralised currency. Therefore, the ultimate pursuit of algorithmic stabilisers should be to maintain a reasonable fluctuation range rather than pursue absolute price anchoring. This remains an area of high interest for derivatives trading and market arbitrageurs.

I think it would be pretty boring if the value of algorithmic stablecoins were limited to anchored fiat coins. This means that cryptocurrencies surrender directly to inflation. The algorithmic stablecoin should be a better unit of account than fiat and a less risky currency than the dollar. This should be the goal of all algorithm stablecoins. So third-party oversight and disclosure is important, not only to increase confidence, but also to help users figure out which are real algorithmic stablecoins and which are “fake dollars.”

Terra’s failure does not mean that all crypto projects fail

I do not believe that Terra’s failure caused the industry’s failure. Stablecoins don’t go to the full stop, nor does the public chain. Other experimental projects related to Terra, such as Cosmos, continue to be excellent development protocols. Besides Terra, we can still expect to see the rise of other innovative projects on Cosmos.

In addition, algorithmic stablecoins remain the most challenging area for institutional innovation in cryptography. Terra has taught all algorithmic protocols a lesson in the importance of reserve support and risk control ratios. There is still room for more sparks between governance tokens and stablecoins, and of course more sustainable utility should be added in addition to arbitrage and stability.

From another perspective, the rush to success is not a measure of a public chain’s results. In this cycle, people tend to use indicators such as TVL and APY to measure the results of a public chain, while forgetting the security, scalability and interoperability of the public chain itself. Essentially, this round of LUNA/NEAR/AVAX hasn’t really built its own public chain universe, except for the high-cost subsidy of ecology and yield agriculture competition, which means the competition in this area will continue.

In the next phase, we will focus on agreements that break the DeFi narrative and move beyond the high-APY spell of the public-chain ecosystem, especially at the application layer, to move beyond reasonable business returns. Algorithm-based stablecoins can still have a long narrative within the public chain ecosystem, provided that we should be careful to embrace economic models that pursue user growth and high-interest revenue. In contrast, smaller algorithmic stablecoin projects that are not designed to anchor dollars and have more sophisticated application scenarios are more noteworthy.