Collateral Optimization (C-OP)
Last updated
Last updated
UNION’s collateral optimization product (C-OP) solves the problem of inefficient collateralization for assets in DeFi borrow protocols (e.g., lending platforms). As a decentralized American Put option, C-OP can provide lending platforms a known liquidation value at the time of borrow, allowing less collateral to be locked. C-OP can also be used as a stand-alone downside protection instrument or for taking on a directional position.
The source code for C-OP can be found here.
The problem of over-collateralization (OC) and capital efficiency limitations in DeFi is an area that UNION identified in its whitepaper. In discussions with various protocols, providing more efficient use of collateral has consistently gotten heavy interest. This interest, paired with the wide and relevant applications of DeFi option primitive we create, made Collateral Optimization an easy first for UNION’s product build.
DeFi’s composable nature unlocks collateral in a way that is not characteristic of Traditional Finance (TradFi).
For example, Alice can deposit ETH as collateral on Maker, mint Dai, send that Dai to the Dai/USDC Curve Pool, earn incentives in CRV, and stake CRV to the governance contract to acquire more CRV — eventually repaying her initial collateralized debt position (CDP) position on Maker with rewards/interest from liquidating her assets received via her clever maneuvering through the DeFi jungle.
However, as we’ve mentioned before, fluid collateral of volatile and sometimes illiquid crypto-assets comes with some explicit trade-offs. In particular, reduced capital efficiency of lending protocols in DeFi due to the need to over-collateralize positions to manage the risk of liquidation adequately. High OC requirements, such as the standard 150 percent OC ratio deployed on popular lending protocols, have detrimental ripple effects across yields, risk, and the composability of DeFi.
To address this problem, we introduce Collateral Optimization (C-OP): UNION’s collateral protection product that can be embedded directly within a lending protocol and furnishes as a composable “Lego Block” for other lending protocols.
To address this problem, we introduce Collateral Optimization (C-OP), UNION’s collateral protection product that can be embedded directly into a lending protocol as a composable DeFi Lego Block.
At a high level, C-OP enables lending platforms to optimize capital efficiency via options. Specifically, we use American-style Put options priced via a binomial model (explained later in this piece) as our base instrument.
The use of derivatives, including options, to reduce risk in portfolios and enhance capital liquidity in the form of lowering margin requirements, has been long practiced in TradFi. With the maturation of DeFi lending and appearances of derivatives in DeFi, this is a perfect time for DeFi to leverage the financial methods of TradFi.
For example, rather than putting up 150 percent collateral for an ETH position on Compound, Alice can use C-OP to put up 125 percent collateral. The remaining 25 percent covered by C-OP is the cost of a put option. Critically, the risk profile for the lending protocol in both cases is the same (assuming efficient markets). Should the ETH price spiral downwards, the lending platform exercises the option on the collateral, creating the same risk profile as if 150 percent OC was initially deposited.
There are various parameters that must be accounted for to determine the exact amount of borrowing collateral freed to the borrower, including the strike of the put option relative to current underlying assets and the percentage of underlying assets relative to protection collateral in the borrower’s account.
At a more granular level, C-OP operates as a contract-by-request (CBR) pool similar to Hegic, where buyers of protection (uUNN tokens) are buyers of puts and sellers (LPs) of protection (pUNN tokens) are sellers of puts. The pool is a constant put writer.
An American put option is the right to sell an asset at a specific strike price before its expiration date. American options differ from their European counterparts in that American options can be exercised at any point prior to their expiration. In contrast, European options can only be exercised on their expiry date.
American put options are excellent instruments for mitigating downside price risk. For example, if the price of Tesla stock is $110, and Alice (a large holder of Tesla stock) wants to mitigate downside price risk, she can buy American put options with a strike of $100 and an expiry 3-months out to cover any downside price risk. Should the price of Tesla fall to $90 a month later, Alice can exercise her option and sell her Tesla shares at $100.
UNION specifically needs an American-style option, because the need to meet liquidation requirements due to collateral sell-off can occur at any time. As such, lending protocols cannot be constrained to a set maturity term, such as those offered by the simpler-to-price European-style option.
The Black-Scholes model, which is a mainstay for pricing options and utilized frequently by TradFi, DeFi, and even CEXes, has well-known limitations that are especially relevant for American-style Put options. The two primary limitations are: 1) inability to account for exercise payoffs prior to maturity and 2) assumption of underlying asset price distribution (normal distribution) that underestimate down-side moves.
To better capture the risk, C-OP utilizes the binomial options model, which relies on an iterative numerical method over an epoch of time. Due to its time-step simulation approach, we can set the percent movement up/down at each time step, allowing us to do things like price with changing volatility over the life of an option or weigh the probability of up/down movements heavier.
We will calibrate these different settings to best account for risk in the market as product development continues.
In short, the binomial model holds these particular advantages for crypto assets:
Black-Scholes does not price American options accurately because it explicitly assumes exercise only at maturity.
Black-Scholes assumes a normal distribution of returns (log-normal distribution of prices), which isn’t as accurate for instruments like equities and crypto-assets.
Black-Scholes assumes constant volatility over the life of the option.
The binomial model allows us to adjust the upside/downside movement probabilities with more precision.
The Black-Scholes model tends to misprice instruments like crypto assets, under-valuing downside put pricing. Comparatively,
Pricing of options for C-OP is determined by:
The price of the option.
The bonding curve.
To value non-linear outcomes of options, we will use the delta-gamma approximation for option pricing to get Value at Risk, or VaR. Options are settled in stablecoins. Prices of options will be periodically calibrated against other options, such as those on liquid central order-book exchanges. Of note, CEX vs. DEX derivatives will vary in liquidity and settlement costs. The bonding curve accounts for the pool solvency via a Minimum Capital Requirement (MCR) metric. As demand increases, the pool towards MCR, and the price should increase and accelerate. If the point of MCR is reached, no further protection can be written so that the pool’s solvency is protected.
C-OP protection tokens, uUNN, are ERC-721 tokens. Each protection purchased is a unique token.
C-OP liquidity pool tokens, pUNN are ERC-20 tokens. Each pool will have their own set of pUNN tokens and pUNN tokens are fungible within the same pool.
Buyers and sellers pay and receive premiums in stablecoins, respectively, even if the underlying collateral is a volatile asset like WBTC or ETH. For sellers (aka LPs), this means the C-OP does not have maturity or term consideration for pUNN. For buyers, since the option is American-style, exercising the option is entirely manual — meaning that, even if an option is in the money, meaning profitable at the time of evaluation, if the purchaser does not exercise the option before the expiration date, then the option expires worthless.
Pool premiums paid by protection buyers are split amongst protection writers (LPs), the Pool Reserve, and the Foundation Reserve. Protection writers (pUNN holders) accrue yield and can withdraw their pUNN tokens at any point, as long as the pool’s liquidity does not fall below the MCR.
So, if Alice decides to deposit WBTC/ETH into a lending protocol with the C-OP integrated, she will be prompted to add C-OP for a small premium — the price of the put option. She will enter the amount to protect, select the term length, tolerance for price slippage, and will proceed to confirm the transaction via her wallet (e.g., MetaMask). Alice will subsequently be issued uUNN tokens (protection buyer), which she can pledge to the lending protocol as collateral, reducing her OC ratio and allowing Alice to retain a portion of her pledged capital as opposed to a complete loss or liquidation.
