In-depth Analysis of Ethena’s Success Factors and the Risks of a Death Spiral

In recent days, the market has been ignited by Ethena, a stablecoin protocol that can offer an annualized yield of over 30%. Many articles have already introduced the core mechanism of Ethena, so I will not elaborate here. In short, Ethena.fi issues stablecoins representing the value of Delta-neutral positions, tokenizing “Delta-neutral” arbitrage trades for ETH. Their stablecoin, USDe, also collects arbitrage profits — hence, they claim it is an internet bond that provides native internet yield. However, this scenario reminds us of the last crypto cycle’s shift from bull to bear, triggered by Terra’s algorithmic stablecoin — UST, which also rapidly absorbed deposits by subsidizing UST lenders with a 20% annualized yield through its native lending protocol, Anchor Protocol, and quickly collapsed after a bank run. Learning from this, the explosive popularity of USDe (Ethena’s issued stablecoin) has sparked widespread discussion in the crypto community, with DeFi opinion leader Andre Cronje’s skepticism drawing widespread attention. However, after reading Andre Cronje’s analysis of Ethena’s risks, I believe some of the logic is not very convincing, so I hope to discuss more deeply the reasons behind Ethena’s popularity and the risks inherent in its mechanism.

The Reason for Ethena’s Success as a CeFi Product: The Savior of Centralized Exchange Perpetual Contract Markets

To discuss why Ethena has been successful, I believe the key lies in Ethena’s potential to become the savior of centralized cryptocurrency exchange perpetual contract markets. Let’s first analyze the problems faced by the current mainstream centralized cryptocurrency exchange perpetual contract markets, which is the lack of short positions. We know that futures serve two main purposes: speculation and hedging. Since most speculators are extremely bullish on the future trend of cryptocurrencies when market sentiment is optimistic, there are significantly more people choosing to go long than short in the futures market. This situation leads to a problem where the funding rate for long positions in the perpetual contract market becomes high, increasing the cost of going long and suppressing market vitality. For centralized cryptocurrency exchanges, since the perpetual contract market is the most active, its fees are also one of the core sources of income. High funding costs will reduce the exchange’s revenue, so finding shorts for the perpetual contract market during a bull market has become a top priority for exchanges to improve competitiveness and increase income.

Here, it might be necessary to supplement some basic knowledge about the principle of perpetual contracts and the role and collection method of funding rates. Perpetual contracts are a special type of futures contract. We know that traditional futures contracts usually have delivery, and delivery involves the transfer of equivalent assets and the associated clearing and settlement, which increases the operational costs of exchanges. For long-term traders, approaching the delivery date also involves operations such as rolling over positions, and the marked price usually fluctuates more near the delivery date because as the rollover operation occurs, the market liquidity of the old target asset gradually worsens, introducing many hidden transaction costs. To reduce these costs, perpetual contracts were designed. Unlike traditional contracts, perpetual contracts have no delivery mechanism, so there is no expiration time, and users can choose to hold positions indefinitely. The key feature here is how to ensure that the price of the perpetual contract is related to the price of the native asset. In futures contracts with delivery, the source of correlation is delivery, because the delivery mechanism will transfer physical assets (or equivalent assets) according to the contract’s agreed price and quantity, so theoretically, the futures contract price will align with the spot price at delivery. However, since perpetual contracts have no delivery mechanism, to ensure correlation, an additional design was introduced in the perpetual contract mechanism, which is the funding rate.

We know that prices are determined by supply and demand. When supply exceeds demand, prices rise, and this is also true in the perpetual contract market. When there are more people going long than short, the price of the perpetual contract will be higher than the spot price, and this price difference is usually called the basis. When the basis is too large, there needs to be a mechanism that can make the basis have a reverse effect, which is the funding rate. In this design, when a positive basis occurs, that is, when the contract price is higher than the spot price, indicating that there are more longs than shorts, the longs need to pay a fee to the shorts, and the rate is proportional to the basis (not considering that the funding rate is composed of a fixed rate and a premium). This means that the larger the deviation, the higher the cost for the longs, which suppresses the motivation to go long and restores the market to a balanced state, and vice versa. Under such a design, perpetual contracts have price correlation with spot assets.

Returning to the initial analysis, we know that when the market is extremely optimistic, the funding rate for longs is very high, which suppresses the motivation to go long. It also suppresses market vitality and reduces the exchange’s revenue. Usually, to alleviate this situation, centralized exchanges need to introduce third-party market makers or become counterparties in the market (which can be found to be a common phenomenon after the disclosure of the FTX event), bringing the funding rate back to a competitive state. However, this also introduces additional risks and costs for the exchange. To hedge these costs, market makers need to go long in the spot market to hedge the risks of going short in the perpetual contract market, which is the essence of the Ethena mechanism. But because the market size is large at this time, exceeding the capital limit of a single market maker, or it brings high single-point risks to market makers or exchanges. To share these risks, or to raise more funds to smooth the basis and make their perpetual contract market funding rate more competitive, centralized exchanges need more interesting solutions to raise funds from the market. And the arrival of Ethena is just in time!

We know that the core of Ethena lies in accepting cryptocurrencies as collateral, such as BTC, ETH, stETH, etc., and going short on their corresponding perpetual contracts in centralized exchanges, achieving Delta risk neutrality, earning native returns on collateral, and the funding rate of the perpetual contract market. The stablecoin USDe issued is essentially like a warrant share of an open-ended market maker fund that does Delta risk-neutral cryptocurrency spot-futures arbitrage. Holding a share is equivalent to obtaining the dividend rights of the fund. Users can easily enter this high-threshold track through this product to earn considerable returns, and centralized exchanges also obtain more extensive short liquidity, reducing funding rates and enhancing their competitiveness.

There are two phenomena that can corroborate this view. First, this mechanism is not unique to Ethena. UXD in the Solana ecosystem actually uses this mechanism to issue its stablecoin assets. However, because it fell before connecting the liquidity of centralized exchanges, its influence did not reach expectations. In addition to the low-interest-rate environment of perpetual contracts caused by the reversal of the entire crypto cycle, the collapse of FTX had a significant impact on it. Second, a careful observation of Ethena’s investors shows that centralized exchanges account for a large proportion, which also proves their interest in this mechanism. However, while excited, we must not ignore the risks involved!

Negative Funding Rates Are Just One of the Possible Triggers for a Bank Run; the Basis Is the Key to the Death Spiral

We know that for stablecoin protocols, the tolerance for bank runs is crucial. In most discussions about the risks of Ethena, we have already understood the damage to the value of USDe collateral caused by the negative interest rate environment of the cryptocurrency futures contract market. However, this damage is usually temporary. Cross-cycle backtesting results show that usually, the negative interest rate environment does not last long and is not easy to occur. This has been fully demonstrated in the economic model audit report of Chaos Labs publicly released by Ethena. Moreover, the damage caused by negative funding rates to the value of collateral is slow because the collection of contract rates usually occurs every 8 hours. According to backtesting results, even if estimated with the most extreme -100% rate, it means that the maximum conceptual loss in any 8-hour period is 0.091%. In the past three years, negative funding rates have only occurred three times, and the average duration of negative interest rates is 3–5 weeks. For example, the negative interest rate recovery period in April 2022 lasted about three weeks, with an average level of -3.3%. June 2022 also lasted about three weeks, with an average level of -4.8%. If including the extreme funding from September 11 to 15, this period lasted 5 weeks, averaging -17.9%. Considering that the rate is positive at other times, this means that Ethena has ample opportunity to “store water on rainy days,” accumulating a certain Reserve Fund to cope with negative funding rates, reducing the erosion of collateral value by negative rates, and preventing the occurrence of a collateralization rate below 100%. Therefore, I believe that the risk of negative rates is not as great as imagined, or that some mechanisms can greatly alleviate this risk. It can be said that this is just one of the possible triggers for a bank run. Of course, if we question the significance of statistics, that is not the scope of this article.

However, this does not mean that Ethena will have smooth sailing. After reading some official or third-party analysis results, I believe we have all overlooked a fatal factor, which is the basis. This is precisely the key to Ethena’s vulnerability when facing bank runs, or the key to the death spiral. Looking back at the two very typical bank run incidents in the cryptocurrency market regarding stablecoins, the collapse of UST and the March 2023 bank run decoupling caused by the bankruptcy of Silicon Valley Bank. It can be seen that in the current development of internet technology, the spread of panic is very rapid, leading to very fast bank runs. Usually, when panic occurs, there will be a large number of redemptions within a few hours or days. This poses a challenge to the stablecoin mechanism’s tolerance for bank runs. Therefore, most stablecoin protocols will choose to allocate highly liquid assets as collateral, not blindly pursuing high returns, such as short-term U.S. Treasury bonds. In the event of a bank run, the protocol can cope by selling collateral to obtain liquidity. However, considering that Ethena’s collateral type is a combination of cryptocurrencies with price fluctuation risks and their futures contracts, this poses a great challenge to the liquidity of both markets. After Ethena’s issuance reaches a certain scale, whether the market has enough liquidity to satisfy redemption needs by unwinding the spot-futures arbitrage combination during a large-scale redemption is its main risk.

Of course, the liquidity problem of collateral is a problem that all stablecoin protocols will face. However, Ethena’s mechanism design introduces an additional negative feedback mechanism into the system, meaning it is more susceptible to the risk of a death spiral. The so-called death spiral refers to when a bank run occurs, due to some factor, the effect of panic will be amplified, leading to a wider range of bank runs. The key here is the basis, which refers to the price difference between futures contracts and spot. Ethena’s collateral design is essentially an investment strategy of shorting the basis in spot-futures arbitrage, holding spot assets and shorting equivalent futures contracts. When the basis expands positively, that is, when the price increase of the spot is lower than the price increase of the futures contract, or the price decrease of the spot is higher than the price decrease of the futures, the investment portfolio will face the risk of unrealized losses. However, when a bank run occurs, users sell USDe in large quantities in a short time, which will cause a significant price decoupling in the secondary market of USDe. To smooth out this decoupling, arbitrageurs need to actively close the open short contracts in the collateral and sell the spot collateral to obtain liquidity to repurchase USDe from the secondary market, reducing the market circulation of USDe, thereby restoring the price. However, with the closing operation, unrealized losses turn into actual losses, causing permanent loss of collateral value. USDe may be in a state of insufficient collateralization. At the same time, the closing operation will further expand the basis, because closing short futures contracts will push up their futures prices, and selling spot will suppress spot prices, which will further expand the basis. The expansion of the basis will lead to greater unrealized losses for Ethena, which will accelerate user panic, leading to a wider range of bank runs, until an irreversible result is reached.

This death spiral is not an exaggeration. Although backtesting data shows that the basis has the characteristic of mean reversion in most cases, and the market will always reach a balanced state after a period of development. However, this is not suitable as a counterargument to the above argument, because users have a very low tolerance for price fluctuations of stablecoins. For an arbitrage strategy, users may tolerate a certain degree of drawdown risk, but for stablecoins whose core functions are storage of value and medium of exchange, users’ tolerance is extremely low. Even for interest-bearing stablecoins with yield as the core selling point, the project promotion process inevitably attracts a large number of users who do not understand the complex mechanisms and participate based on literal understanding (this is also one of the core accusations currently faced by UST’s founder Do Kwon, namely fraudulent promotion). These users are the core user group that triggers bank runs and the group that ultimately suffers the most severe losses. The risk is not small.

Of course, when the liquidity of shorts in the futures market and longs in the spot market is sufficient, this negative feedback will be alleviated to a certain extent. However, considering the current issuance scale of Ethena and the storage capacity that comes with its high subsidies, we must be vigilant about this risk. After all, with Anchor offering a 20% savings subsidy, UST’s issuance volume exploded from 2.8 billion to 18 billion in just 5 months. During this time, the overall growth of the cryptocurrency futures contract market certainly could not keep up with such an increase. Therefore, there is reason to believe that Ethena’s open contract size will soon increase by an exaggerated proportion. Just imagine, when over 50% of the short position holders in the market are Ethena, its closing will face extremely high friction costs, because there are no shorts in the market that can withstand such a scale of closing in the short term. This will make the effect of basis expansion more obvious, and the death spiral will be more intense.

I hope that after the above discussion, everyone can have a clearer understanding of the risks of Ethena, maintain a respectful attitude towards risks, and not be blinded by high returns.