Delta-neutral DeFi strategies overview
It’s hard to imagine TradFi without delta-hedging: the options market is central to all derivatives trading volumes. Options market-makers usually dynamically delta-hedge their books with many other techniques that require much more sophisticated math. The main idea of traditional market makers’ business is market neutrality: they don’t bet on directional movements, earning on the bid-ask spread, exchange fees refund, and fees for serving clients’ orders.
This post is about DeFi, so let’s find out how we can employ some option-world tricks here to create safer yields without being too deep in financial math.
When talking about providing liquidity, everyone mentions Impermanent Loss — the difference between the value of the LP position and the portfolio of tokens that are held proportionally equal to the liquidity pool’s ratio at the moment of opening position. The problem is that not so many people are ready to take risks associated with HOLDing tokens, especially in the bear market: you usually sit with stablecoins, waiting for the dip to accumulate a position. What does IL say to more risk-averse people who never want to become a holder, but to control risks and earn good returns in DeFi?
Comparing LP position with HODL portfolio is reasonable because of the risks LPs are exposed to:
- Regular directional market risk (typical “long” position’s risk: if each asset dollar price rises by 100%, position value will double, if prices fall by 50%, position value will decrease by half)
- Risk of pool assets relative price change (amounts of tokens in LP position change with its relative price, causing us to have more asset, which price falls and less asset, which price rises: that’s Impermanent Loss). So LP’s position is a portfolio of HODLer, which proportion changes if it’s constituents rise and fall with different amplitudes.
But you can make profits by providing liquidity while having no exposure to absolute prices of assets: hedging helps us to eliminate directional risk, but earn more that can offer lendings or stableswaps, taking exposure on pool assets correlation, which causes IL.
Further, we’ll examine some ways to safely farm pairs of volatile tokens by adding a “short” exposure part to positions in popular DeFi primitives. We’ll focus on fully on-chain strategies, but you can replicate them replacing any short piece with a short position on perpetual futures anywhere on a centralized exchange.
*Let’s consider the case where we want to get USD-notioned returns. As an initial deposit, we have some USDC (or another stablecoin, whose value almost equals 1$).
1️⃣ Hedged staking
That section needed to help people not-in-finance-at-all get accustomed to the most straightforward example of using hedging in DeFi yield strategies. So if you’re familiar with long and short exposure notions and their interaction, you can move straight to the following sections.
There are a lot of liquid staking protocols. They wrap locked underlying assets into liquid ones to increase capital efficiency, examples are stETH, rETH, and stMATIC. As wrapped assets usually can’t be unstaked immediately, corresponding pairs are being created on DEXes (stETH/ETH, MATIC/stMATIC). Such pools can pay good returns, having nearly 0 Impermanent Loss. Still, the position in that pool has long exposure on the underlying price (returns are proportional to price change). If ETH price falls by 50%, stETH’ll likely fall by 50%, too as the LP position value.
Let’s consider having a position in lending protocol: imagine depositing USDC as collateral, borrowing ETH with 60% LTV, and swapping borrowed ETH back into USDC. If ETH price falls by 50%, we’ll need twice less USDC to buy borrowed ETH amount, increasing the initial deposit by 30%. If ETH price rises, we’ll suffer losses and, at some moment, will be liquidated. That is a plain short position (returns are proportional to minus price change).
We can combine mentioned above blocks to gain a return that has no sense to how the underlying price changes:
1. Deposit USDC into a lending protocol
2. Borrow ETH (more or less depending on how active you can control borrowing position liquidation)
3. Deposit ETH into staking/associated pool
USD values of 2 and 3 parts change oppositely, so their sum’s value stays constant, while we’re getting only farming/staking rewards.
It’s needed to mention that because LTV doesn’t equal 100%, not all USDC used for this strategy gain farming rewards. Effectively APR’ll be lower than displayed on the DEX page by (100-LTV)%. Although we have to add lending APR (that we get by providing USDC as collateral) and subtract borrowing APR (that we pay for borrowing). Usually, they’re relatively low and nearly eliminate each other.
Return curves of hedged staking. Staking, lending rewards, and borrowing fee are not displayed.
Strategy summary:
- Almost no risks, except one associated with staking contract safety
- Small returns, as in stablecoin lending
2️⃣ Hedged UniV2 farming
By having only the LP position opened, we take exposure on asset prices relative to USD (usual price risk) and on assets price relative to each other (Impermanent Loss). The first part of the risk is typical “long” position risk: if each asset dollar price rises by 100%, the position value will double, if prices fall by 50%, the position value will decrease by half.
We can eliminate the first type of risk by borrowing assets against USDC as collateral before putting them into the liquidity pool. When opening a position, we’ll have 𝑥 and 𝑦 in the pool (long) and equal amounts borrowed (short).
The second type of risk is realized when assets relative price changes: amounts of tokens in the LP position change too, causing us to be effectively long on the asset, which price falls, and to be short on the asset, which price rises: that’s Impermanent Loss. Now we should only compare returns of a pool with expected IL, unaware in what direction prices move. To control Impermanent Loss, we can rebalance the strategy: withdraw part of the pool position to repay debt or borrow more for amounts in the pool equal to amounts borrowed.
Let’s consider the case of farming ETH-BAL pair that we’ll further denote as x-y. Suppose the total USD value we want to invest in such a strategy is V0. In that case, we first have to put this USDC into Aave as collateral, then borrow ETH and BAL worth this amount of USDC:
In the LP position we’ll have the following:
That is precisely the same amount of x and y borrowed.
To cover all possible ETH and BAL price changes, we’ll have to look at a 3D plot. We’ll only look at the case when ETH price increased by 10% since position opening and see how the position value’ll vary with BAL price change:
Strategy summary:
- Isolated Impermanent Loss risk, which is small in correlated pairs, but be aware of small-cap alts, their volatility can eliminate even the most enormous rewards.
- Significant expected return: relatively safe pairs like CRV/ETH or COMP/ETH can offer more than 45% APR.
3️⃣ Hedged Balancer-like farming (generalized CPMM)
We can think of univ2-type pools mentioned above as AMMs, that rebalance the underlying liquidity pool so that its dollar value is split equally between its constituents. But there is a generalization, where the proportion in which the value is divided, can differ from 50/50. An example of a protocol that offers such pools is Balancer. The popular proportion there is 80/20. The BAL/ETH pool is constructed like that. What is important is that such pools have lower IL, so investing there can be a lot safer ceteris paribus.
So having V0 USDC to invest, we should borrow ETH & BAL worth:
And deposit it into the pool, having
Let’s look at 80BAL/20ETH hedging pool Returns curve in case when ETH price increases by 10% since opening.
Strategy summary:
- Decreased IL, but position usually contains more alt tokens in pairs like 80BAL/20ETH, which can be not liquid enough to enter and exit pools instantly.
- Big expected return and decreased IL can bring returns up to 60% in pairs like 80BAL/20ETH.
4️⃣ Leveraged delta-neutral farming
Many LYF (leveraged yield farming protocols) like Tulip, Alpaca, Alpha Homora, and Tarot can create a position similar to those mentioned above but with higher leverage. All You need to do is select appropriate pair and borrow needed amounts of assets when opening a position. If we’re talking about USD-notioned return, we should select pair of Volatile asset + Stablecoin, then with 2’nd leverage, you need to borrow 100% of the Volatile asset, with 3’rd leverage — 75% of the volatile asset and 25% of the stablecoin.
- 2’nd leverage: borrowing 100% of volatile asset, you’ll have all risky assets that are put in the pool being borrowed, resulting in delta neutrality, as we discussed above.
- 3’rd leverage: depositing 1 $ of USDC, we’ll borrow 1.5$ worth of volatile asset and 0.5$ of stablecoin. So at the moment of position opening, we’ll have 1.5$ worth of asset and 1.5$ worth of stablecoin with 1.5$ worth of asset being borrowed, causing the position to be delta-neutral.
The returns curve of such a position looks like the previous one’s, but multiplied by 2 or 3 (depending on selected leverage), which means multiplied farming rewards, but although multiplied IL.
Let’s look at how’ll SOL/USDC delta-neutral farming position behave depending on SOL price change:
Strategy summary:
- Huge expected returns: can frequently see >100% APRs in Tulip farms.
- Amplified IL can increase losses in a highly volatile market. Although, smart contract risk is increased due to the use of nested protocols.
5️⃣ Delta-neutral concentrated liquidity and delta-neutral leveraged concentrated liquidity.
These strategies are now just emerging; however they work very similarly to Hedged Farming and Leveraged delta-neutral farming positions, but here you have to control the range where provided liquidity is active. These strategies are fraught with much more risk, as most active concentrate liquidity providers are unprofitable. If you understand, how it works, you likely don’t need such shallow guides :)
💫 What Cetra does with all this cool stuff
While those strategies are great to use, there are a lot of details, that were omitted in such an overview: the problem of liquidations management is really big: you should actively control your position, moving capital between borrowing and farming part to remain delta-neutral and not being liquidated.
Doing it manually you will likely face a large divergence of calculations and real swaps, as prices will change while entering the position. That’s why we built Cetra.
We are building a protocol to bring the adoption of Delta-Neutral instruments into DeFi. Put your money on autopilot with automated vaults, managing mentioned strategies. So you can be calm about liquidations and the risk of losing too much money on unexpected movements.
Cetra is implementing all of the above strategies to create the most significant delta-neutral protocol in DeFi.
