Opus started with a single core tenet: to have an autonomous monetary policy.
Where did we come from and where have we been? Why did we not target collateralization ratios and what Chesterton's fences did we encounter along the way?
The first iteration of the white paper famously targeted the collateralization ratio of the system. Mere months later Dan Robinson would tweet about it in light of Terra's catastrophic collapse—in this scenario, he advocates for unique borrow rates determined by your loan-to-value ratio.
Targeting a fixed collateralization ratio ✨feels✨ right—it’s no coincidence that it’s used to determine creditworthiness and thus interest rates in the TradFi world. In this overcollateralized system, risk is priced by the controller relative to the backing of the debt asset—the closer to 100%, the higher the risk of under-collateralization, and thus premia increases to create a reliable buffer that can be used to maintain the equilibrium between buyers and sellers.
Personally, I viewed the system as a conveyor belt whose speed was determined by this controller—at the end of the moving surface were the liquidators waiting to pounce on discounted short-tail collateral assets.
However there was a glaring issue with this indiscriminate rate hike—it angered both high and low LTV CDPs alike, and in such cases we wanted to keep the highly collateralized positions happy!
To address this we inadvertently discovered an efficient way to make Dan Robinson's tweet a reality: a sigmoid base rate function with an inflection point at the exact LTV we wanted to discourage. By having a controller actuate on this point we were able to change the shape of the curve and effectively segregate desirable CDPs from undesirable high ones. At this point it was a matter of playing with the shape of the curve and eventually iterating in prod to converge to the best solution.
It still *felt* right, and given a sufficiently robust network of arbitrageurs or a redemption mechanism, it would work. However, this was around the time we decided not to implement redemptions. After all, if you're not in violation of the protocol, why should you be forcefully liquidated due to the immaturity of the markets?
So with redemptions no longer on the table and a new belief that collateralization should be determined strictly by the liquidation engine and its risk policies, we went back to the drawing board.
After some deliberation, we concluded the problem had always been matching demand for leverage with market growth, but we had been zoomed in on collateralization due to Do Kwon induced paranoia. It's the peg.
The protocol should control the peg autonomously by globally increasing the interest rates. Now we have a choice to make: dynamic interest rates or peg-driven liquidation thresholds?
“Since the LTV thresholds for each position are determined by its collateral composition, we could make the protocol more attractive by charging fixed borrow rates but controlling the global per-asset LTV thresholds to liquidate the riskiest CDPs during depeg (< peg) events."
Being able to actuate on the peg without affecting interest rates seemed very appealing, but because it affected pre-agreed-upon risk parameters, we shelved it. The idea of dynamic LTV thresholds is excellent as an addition, post-refinement!
However, it is susceptible to a simple issue: sufficiently capitalized agents can add collateral without repaying, forcing relatively low LTV positions to do so. Is this fair? Maybe, maybe not. Perhaps in the future, we will revisit it.
Instead, we opted for the tried and true interest rate. Each basket of collateral assets already has their own unique base rate, but now we had a global multiplier tied to the peg.
Under pegged? Interest rates increase to incentivize spot buys of CASH to collect cheap collateral.
Over pegged? They decrease to generate demand for borrowing.
So what happens if even 0% is too expensive? After all, CASH is not a floating peg asset that can tap into negative rates via redemption rate changes!
We have multiple solutions for this but let's talk about mitigations first.
Cross-margin positions. While stETH alone might not be the most attractive asset during a down-only market, combining it with yield-bearing USDC might just do it for you. Or perhaps a RWA.
Bonds. You heard correctly, but what do we mean? Protocol native variable rate bonds? Crypto native RWA bonds? Why not both?
And while we're at it, the protocol is able to permissionlessly mint unbacked CASH just to provide one-sided liquidity during these liquidity crunches.
The Future
While the future is uncertain, we are hoping to grow CASH to a point where we can run more autonomous money experiments!
Remember the sigmoid functions?
We want to explore bringing them back into the current peg-driven design and potentially create an additional fallback controller to address their weakness and create a better, more stable borrowing experience for low LTV CDPs—one where low LTVs pay a fixed rate and high LTVs are susceptible to the whims of the market!
Remember how we talked about giving our users the ability to choose between hedging with options or liquidations? We’re going to do that too, and a lot more!
Autonomous risk parameters. Floating peg coin? Autonomous lending markets…
What? When? We don’t know.
Some experiments require the protocol (or partner protocols!) to hit certain growth milestones and ultimately, community feedback will drive development.
We have many potential improvements on the pipeline and only time will tell which ones will be prioritized.
