06/2024

Ethena: How long can it last?

Background

In the 233 years since the founding of the First Bank of the United States, redeemable liabilities pegged to the U.S. dollar have had a 100% failure rate when not backed by an implicit guarantee from the U.S. federal government.

USDe is a redeemable liability pegged to the U.S. dollar and issued by Ethena Labs backed by long positions in LSTs (liquid staking tokens) and offsetting short positions in ETH-USD perpetual futures. USDe is designed to function as a stable transactional dollar: a “censorship-resistant, scalable and stable form of money to gradually replace fiat stablecoins in other on-chain DeFi applications…” [Source]

Since its public launch on February 16, 2024, Ethena has grown into the 8th largest stablecoin by market cap at $2.335 billion as of Wed Apr 10th. [Source] Ethena Labs advertises an annual percentage yield (APY) of more than 24% on their homepage. [Source]

At its current growth rate, USDe will eclipse $3 billion in market cap by mid-April and may chart a similar growth trajectory to that of Terra/Luna, which reached a market cap of more than $17 billion. [Source]

There is a significant probability that USDe will experience a violent and sudden surge in redemptions in the near-to-medium term. Over an arbitrarily long time horizon, the probability of a “bank run” approaches 100%. If USDe’s market cap continues to grow at its current rate and is used as collateral in a wider variety of financial products, it poses a significant destabilizing risk to other financial applications in DeFi, especially the new breed of “restaking” applications (Eigenlayer) that require users to deposit stETH as collateral.

Overview of Existing Stablecoins

There is a long history of crypto projects that have attempted to create “crypto-native dollars;” that is: redeemable, stable $1 liabilities with no connection to the regulated U.S. banking system. However, no extant stablecoins with significant usage fit this description. Each of the largest stablecoins today have significant holdings of U.S. Treasuries and cash held in regulated, audited U.S. bank accounts as their primary reserve assets:

Tether, which represents 70% of the total stablecoin market capitalization, is backed 90% by cash and cash equivalents, nearly all of which are short-term U.S. Treasuries. [Source]
USDC, the 2nd largest stablecoin, is backed by U.S. dollar-denominated money market assets held in regulated and audited accounts at U.S. financial institutions. [Source]
DAI, the 3rd largest, is backed by centralized stablecoins like USDC and Pax Dollar (USDP) as well as other cryptocurrencies and U.S. Treasuries. Active collateral management and significant over-collateralization (~300%) are used to protect DAI’s value against a decline in the value of its cryptocurrency reserves. [Source]
Other large stablecoins, such as FD USD (First Digital), PYUSD (PayPal), TUSD (True USD), and USDM (Mountain) are backed by various combinations of U.S. dollars held in regulated bank accounts, U.S. Treasuries, and other fiat-backed stablecoins.

Timeline of Major Events

March 8th, 2023: Arthur Hayes, a widely-followed blogger and trader in the cryptocurrency sector, published a blog post titled, “Dust on Crust,” outlining a new stablecoin mechanism using a combination of Bitcoin holdings and short position in Bitcoin/USD inverse perpetual swaps to maintain a stable value relative to the U.S. dollar without the need for bank custody or fiat reserves.

July 19th, 2023: Ethena Labs announced that they had raised $6.5 million from investors including Dragonfly, Arthur Hayes, in addition to more than a dozen other notable VCs, derivatives exchanges, market makers, and cryptocurrency trading firms. In the announcement, USDe was referred to as “a scalable crypto-native stablecoin” that “is fully collateralized on-chain and maintains stability through delta-hedging ETH collateral on centralized and decentralized exchanges.”

October 10, 2023: Guy Young, founder and CEO of Ethena Labs, announced that Ethena would begin using the term “synthetic dollar” rather than “stablecoin,” to avoid being grouped with fiat-backed stablecoins like USDC and Tether (USDT). However, many similarities exist, most notably (1) a mechanism designed to maintain a $1 peg and (2) a redemption option to exchange USDe for an equal dollar amount of LSTs.

February 16th, 2024: Wachsman, a public relations firm hired by Ethena Labs, circulated a press release announcing that Ethena had raised $14 million from Dragonfly, Avon Ventures, Brevan Howard Digital, Franklin Templeton, PayPal Ventures, Arthur Hayes, and several derivatives exchange. Later, an updated version was re-circulated only including Dragonfly and four derivatives exchanges. [Source]

February 19th, 2024: Ethena officially launched USDe to the public, introducing a “Shard Campaign” rewarding users for their individual “contributions to Ethena’s ecosystem over the course of the campaign.” [Source]

February 23rd, 2024: In its first week after launching to the public, the Ethena team generated $1.05 million in revenue. Of that $1.05 million, $403K was transferred to the sUSDe Staking Module as user rewards, with the remainder sent to the Insurance Fund. However, after users berated the Ethena team for poor communication around the rewards distribution, the Ethena team decided to issue an additional $227.5K to the Staking Module as reward. [Source]

March 27th, 2024: USDe supply grew to over $1.3bn. This represents the fastest USD-dominated asset to reach >$1bn supply ever in crypto, becoming the highest-earning protocol within three weeks. [Source]

April 2nd, 2024: The ENA Airdrop takes place. Shards are replaced with Sats to onboard BTC as a backing asset in Season 2. [Source]

April 8th, 2024: Users can lock $ENA in Ethena for a minimum 7-day period, earning the highest reward rate at 30x Sats per day, with those holding at least 50% of their USDe balance in $ENA across platforms receiving a 50% boost on all USDe rewards. [Source]

April 9th, 2024: MakerDAO raises its debt ceiling to $1 billion for Dai allocations in USDe and sUSDe markets. [Source]

April 10th, 2024: Ethena integrates with exchange wallets: Binance, Bybit, OKX, Bitget @Torben just added this

April 12th, 2024: $ENA is at $1.19 with 1.43B ENA in circulation, which equals 9.5% of the total amount of coins.

April 13th, 2024: USDe depeggs to 0.97$ for 30 minutes as crypto markets crash, a first glimpse into the risk nature of the trade underlying USDe. This happened at 8 am, only a few hours after the initial bitcoin crash.

USDe Overview

USDe is a redeemable $1 issued by Ethena Labs. USDe can be acquired by a user in two ways:

Direct Mint: Users can deposit stETH collateral into a smart contract and receive an equal dollar amount of USDe, net of transaction fees.
Secondary Market: USDe can be purchased by users in the open market against a variety of other assets.

Once minted or purchased, a user can perform the following functions with USDe:

Direct Redeem: Users can deposit USDe directly with Ethena Labs and receive an equal dollar amount of stETH, net of transaction fees
Stake for sUSDe: Users can deposit USDe into a smart contract in exchange for sUSDe. Once staked, users begin earning a share of Ethena’s revenue proportional to their relative ownership in the staking pool.
Direct Transfer: Users can transfer their USDe to other users, deposit their USDe into third-party smart contracts, or trade their USDe for other assets on secondary market venues. The Ethena team has stated their desire for USDe to “act as collateral in money markets throughout DeFi.” [Source]

Once deposited into the Ethena protocol, the user’s stETH balance is transferred to an OES (“Off Exchange Settlement”) provider, a custodian that holds collateral assets off-exchange. Currently, Ethena uses Copper, Fireblocks, and Ceffu for off-exchange settlement. [Source]

The stETH balances held in OES accounts are then used as collateral to open short positions in ETH-USD perpetual futures on derivatives exchanges, including Binance, BitMEX, Deribit, ByBit, OKX, and Bitget. The short position is in equal proportion to the amount of stETH collateral, so the overall position has no exposure to the price of ETH.

Perpetuals Futures Background

Perpetual futures exchanges use a funding mechanism to keep the futures price in line with spot markets. If volume is skewed to one side of a particular market, funding rates will adjust to incentivize traders to take the opposite side.

The most common use of perpetual futures is for traders to get leveraged exposure to the long side of cryptocurrencies, which, unlike most financial assets, exhibit positive convexity. As a result, there is a significantly positive funding rate on the short side of most tokens, offering 5-25% APY to traders taking the short side. This effect is particularly strong in bull markets, where tokens are more likely to exhibit positive skew. As a result, many sophisticated traders have taken short positions in cryptocurrencies using a physical version of the underlying asset as collateral, thereby earning the positive funding rate with a net zero exposure to the underlying asset.

Ethena is effectively putting its users into this trade and passing along the revenue generated from the funding rate, while keeping a share for itself.

History has shown, that significant events can result in the crash of currencies. With the financial crisis in 2008, the Japanese Yen was heavily affected, putting the Japanese economy in deep trouble.

Currency crashes also affect future carry trades long term, as certain currencies might be negatively connotated. The Australian Dollar for example is only popular as a funding currency, after being a very attractive asset currency until its crash in 2013.

As crypto markets are much more volatile, the risk of crashing responding markets is increased.

sUSDe Overview

Each unit of sUSDe represents a single unit of USDe deposited into the sUSDe Staking Module. The price of sUSDe increases over time to reflect value accrual within the Staking Module, which is generated from ETH staking rewards and the funding rate on the short ETH-USD perpetual swap positions.

Breakdown of sUSDe Yield Generation

The mechanism through which Ethena Labs generates “yield” for sUSDe holders is broken down into two streams:

Staking Revenue: Revenue generated from the deposited LST (currently only stETH). This rate typically ranges from 2-5% APY, with a rate of 3.4% as of Feb. 26, 2024.
Short ETH-USD Funding Revenue: Revenue generated from the positive funding rate on the short ETH-USD position. This rate historically has varied between -15% and 25% depending on market conditions.

Historic Risk Profile of the Carry Trade

The historic risk profile of the carry trade has been extensively studied, particularly in developed markets and are known to do poorly in highly volatile environments. [Source]

Dollar-neutral trades tend to show positive average returns, significant negative skewness, correlation with risk factors, and notable downside risk. Conversely, a diversified dollar-carry portfolio demonstrates a greater average excess return, a higher Sharpe ratio, minimal skewness, unconditional uncorrelation with standard risk factors, and no observed downside risk. [Source]

The three Fama and French equity market risks explain the returns of the equally weighted carry trade, with no exposure to the dollar. [Source]

Returns on the carry trade tend to be higher when diversified across multiple currencies. A dynamic dollar strategy, while offering elevated returns, also exposes investors to various risks including equity market risk, bond market risk, foreign exchange (FX) risk, and volatility risk factors.

Strategies that employ spread weighting and actively rebalancing currency positions have shown potential to enhance Sharpe ratios, thereby improving risk-adjusted returns. Additionally, the choice of base currency significantly influences currency exposure and consequently heavily influences the Sharpe ratio of the carry trade.[Source]

In regions characterized by higher volatility and increased crash risk, such as emerging markets, a dollar-neutral carry strategy often outperforms its dollar-carry counterpart. Hedging a carry trade in emerging markets evaporates carry profits, as the cash risks is priced in the options. [Source]

Bank Run Dynamics

More than two centuries of financial history and more than two dozen failed stablecoin projects have demonstrated that it is impossible to create stable, redeemable dollar liabilities without implicit backing from the U.S. government.

In this case, Ethena is dependent on stETH maintaining 1:1 parity with ETH, and will eventually falter when that assumption breaks down. In the words of Ethena Labs:

“Ethena's trade is principally dependent upon the spread between the value of our collateral [stETH] & the price of the short derivatives positions [ETH-USD] not diverging in a problematic manner.”

Because stETH is not instantly redeemable for ETH, there is a natural duration mismatch that could cause the basis between stETH and ETH to widen during periods of market stress. Such a widening would create unrealized losses for Ethena on its long stETH/short ETH positions, which could become realized if users were to redeem their USDe en masse. Here is a table from Ethena outlining what this would look like in different scenarios:

Untitled

Even a small widening of the basis, such as a 2% discount on stETH relative to ETH, would create unrealized principal losses for Ethena Labs. If some users became worried and, as a result, redeemed their USDe for the underlying stETH, Ethena would be forced to close out their short position on ETH-USD perpetual futures, thereby driving the price of ETH higher and widening the basis even further, creating more unrealized losses.

This dynamic is self-reinforcing and perfectly mirrors the effect of a classical bank run, whereby a liquidity crisis can lead to a solvency crisis through mass redemption of dollar deposits.

Here is an illustrative example of how this would happen from the perspective of an individual user:

User deposits 10 stETH in exchange for 30,000 USDe.
Ethena places 10 stETH in a custody account as collateral for a 10 ETH-USD short position.
Poor market liquidity leads to a 2% basis in stETH/ETH, with ETH rising 2% and stETH staying flat.
User becomes worried about unrealized losses on USDe, redeems 30,000 USDe for 10 stETH.
Ethena closes out ETH-USD short position for a .2 ETH loss, paid for by the insurance fund.
ETH-USD price rises an additional 1% as a result of closed out short position, basis reaches 3%.
More USDe holders worried about basis risk, repeat steps 4-6.

Broader Market Impact

Beyond the Ethena protocol alone, a widening of the stETH/ETH basis would have a destabilizing effect on the broader decentralized financial ecosystem. Such a widening would lead stETH holders to attempt to redeem their stETH for ETH on a 1:1 basis directly through Lido.

However, because of the design of Ethereum staking on Lido, stETH cannot be directly redeemed for ETH until after a wait in the withdrawal queue, usually between 12 and 18 hours. In times of financial stress, the wait time for stETH withdrawals into ETH can extend to days or even weeks, as it did in October 2023.

While principal losses at Lido are unlikely, a sufficiently long redemption queue could allow a stETH/ETH basis to persist for months. A wider basis would also further incentivize stETH holders to enter the redemption queue and redeem stETH 1:1 for ETH, thereby extending the queue longer and widening the basis even further. For example, if stETH were to trade at a 20% discount to ETH, the only way for an arbitrageur to capture that 20% would be unstake with Lido directly, which likely would already have a long withdrawal queue by that point.

Furthermore, stETH functions as a short-term money market product in DeFi and is an extremely common form of collateral for borrowing.Chaos Labs reports that there are currently 1.5 million stETH tokens deposited into DeFi protocols, amounting to approximately $4.6 billion. Of this total, 71% serves as collateral within lending protocols. A widening of the stETH/ETH basis would force the liquidation of a portion of these stETH balances and lead to sales on the secondary market, further widening the basis and triggering more liquidations. [Source] Such a liquidation event could easily be triggered by Ethena, were it to reach sufficient size. A surge in interest for new “restaking” products like Eigenlayer, which offer additional yield on locked stETH, could exacerbate a sudden unwinding of stETH collateral.

MakerDAO has recently approved a governance decision to increase the debt ceiling of its direct deposit module for the stablecoin Dai to $1 billion. This decision aims to facilitate investments in Ethena USDe and its staking counterpart, sUSDe, through Morpho Labs' lending vaults. The move is part of MakerDAO's strategy to diversify its collateral exposure. [Source]

Sound Banking and the NQA Principle

A significant body of academic research is devoted to studying sound banking practices and understanding the fundamental reason why different forms of money have had different outcomes. Countless examples in traditional finance, from the Free Banking Era of the mid-1800s, to the frequent runs on demand deposits in the early 1900s, to the run on Money Market Mutual Funds in 2008 and 2020, to the run on Silicon Valley Bank in 2023 have all demonstrated how violent runs on different forms of money can be, and how they can have ripple effects that permeate throughout the broader financial sector.

Academics have honed in on several properties required for sound money, which, because of its ability to be redeemed at will, faces a unique risk that all holders demand redemption at once. Basic economics teaches that quantity and price adjust to equate supply and demand. Because money is redeemable at only one price ($1.00), any information that affects the supply and demand of money will adjust quantity instead, which, in the case of negative information (ie. the price should be $0.99), will adjust its quantity to 0. Historically, the only institutions that have been able to survive such a panic are ones with direct FDIC insurance or ones with backstops from the U.S. government. Therefore, sound money needs to be “information insensitive,” thereby satisfying the NQA (”No Questions Asked”) Principle. [Source]

This idea is intuitive when thinking about how money is used. Namely, money is information insensitive if neither party in a transaction has an incentive to produce information about its value, hence “no questions asked.” Once questions are raised about its value, money will become “information sensitive,” creating an incentive for holders to move into forms of money that remain information insensitive.

In the case of USDe, if an investor had doubts about USDe’s underlying value, thinking it may be worth $0.99 instead of $1.00, the rational decision would be to redeem USDe for $1 worth of stETH or to sell USDe for USDC/USDT on the secondary market. Eventually, if enough market participants behave in this way, USDe will become highly information sensitive. Holders, rather than confirm their $1 is still worth $1, will instead swap into other currencies, thereby triggering a run.

Because USDe is backed by a delta-neutral basis trade, rather than liabilities of the U.S. government, it fails to meet the NQA principle from the start. Once excitement cools off over the 20%+ interest rate offered on sUSDe, a widening of the stETH/ETH basis will make USDe information sensitive and trigger a self-reinforcing bank run with potentially destabilizing effects for other DeFi protocols.

Conclusion

This essay does not condemn Ethena or equate it with algorithmic stablecoins. Instead, it serves as an interrogation of the systemic effects that can unfold if the risk of an asset is underestimated as it migrates through the DeFi ecosystem and is included in an array of further products. Delta-neutral trades are not without risk; they are commonly described as "picking pennies in front of a bulldozer." It can take years to unravel, but when the Ethena trade does unravel, as it inevitably will, it will have a broad impact on the DeFi ecosystem. To reiterate, Ethena itself is not inherently bad or unreasonably risky, but just as one would not use bank deposits to post them in a hedge fund, which would also violate many of the Basel rules, we should be cautious when building DeFi applications on top of Ethena that are supposed to have different risk profiles.

What we caution against is the active misframing or poor framing by the sides of projects. USDe has little, if anything, to do with a dollar. It is an active strategy wrapped in a token, a structured product. Raising money by calling it a stablecoin can be considered reasonable since VCs should be knowledgeable enough to distinguish. But which retail might not know.

Moreover, several design and strategy choices are not well thought through. Thirty percent of the insurance pool should not be employed in a liquidity pool, shifting the liquidity in the same token one tries to insure. Second, the historical risk analysis is weak, and the community would do well to hire or appoint someone to conduct empirical and scientific risk management.

Financial innovation is valuable and necessary, but we must not disregard any risk measurement or lessons that we can learn from historical financial patterns. Looking at the recent past, we should be cautious about how we measure risk and how we communicate to the outside world. Even if the underlying layer enables trustless interaction, we still need to trust the organizations building to do their homework when it comes to financial history and financial technology.

TL