Ozone peg stability mechanism

Intro
On the recent COL-5 launch party @dokwon dropped some Alpha in regards to a potential Ozone feature where part of the Ozone Luna allocation is not burned for UST but rather is exchanged for non UST stablecoins, and as incentive users can purchase discounted but locked/vested Luna.

These non-ust stable coins will be used to backstop UST in the event of a depegging where Ozone acts as a type of buyer of last resort.

So with this in mind, i want to suggest and discuss a potential additional feature that Ozone could adopt which would significantly stabilize the UST peg. After all depeggings are usually cascading during crypto flash crashes. Deep liquidity could go a long way in stabilizing the UST peg and potentially prevent a depegging.

MakerDAO Background
So the idea is inspired by MakerDAO, most people dont realize that in addition to being overcolaterized DAI has a very creative and innovative PEG stability mechanism . This mechanism basically allows protocol users to create 50x leveraged DAI/Stablecoin LP positions at low to 0 interest.

So for example a user with 10k, could use that 10k as collateral to create collateralized debt position (CDP) where a 490k DAI loan is backed by 500k in DAI/USDC Uniswap liquidity.

I suspect about 50 million DAI is backed by 50x leveraged CDPS, this creates 50 million dollars in DAI/USDC liquidity backed by only 500k. DAI seems to use a fixed oracle for determining the price of DAI/USDC, and so the CDP position only grows in value as it accrues uniswap swap fees.

A quick run down of how this works
In a single transaction the end user does the following:
Receives 500k DAI flashloan from AAVE
Uses 250k DAI to purchase 250k USDC
Deposits 250k DAI and 250k USDC in to Uniswap LP
Deposits the uniswap LP in to the respective MakerDAO vault , the MakerDAO vault requires 102% collateral so the user is able to withdraw about 490k DAI from the MakerDAO vault
The user uses the withdrawn DAI and 10k of their own collateral to pay off the flash loan.

The user now owns a makerDAO CDP 490k of DAI debt backed by 500k of Uniswap DAI/USDC LP

The user is incentivized to hold this position because they earn fees on 500k of LP

To withdraw the initial collateral and fees the user can flash loan out of the position .

Users are incentivized to open CDPS when DAI is above or below peg, when above peg the initial flash loan is in DAI, when below peg the initial flashloan is in another stable . Users are incentivized to close CDPS when DAI is below peg.

How Ozone could adopt a similer stability mechanism
Ok so here’s how i think this could work for Ozone. Since Ozone is going to be a bit heavy on UST, and potentially other stable coins , a way to allocate some of those funds could be to issue leveraged UST/other stable CDP positions , backed by low interest UST loans provided by Ozone . I would suggest 110% max LTV, and 8/9x being the “safe” threshold, 10x being the max/liquidation threshold . In this model Ozone would also collect fees and potentially interest from it’s CDPS.

Pros:
This would basically subsidize, incentivize, and magnify UST/otherstable liquidity without requiring LP rewards or other incentives. This would help provide stability to the UST peg in the form of leveraged liquidity subsidized through low interest loans.

When above peg users are incentivized to open CDPS , and when below peg Users are incentivized to close CDPS. If flash loans in other stables are possible then below peg events could also encourage users to open CDP positions. Since these are leveraged positions they provide magnified buying and selling pressure in the event of a depegging.

Unknowns/Other considerations:
Where does the other stable coin liquidity come from , i think it’s best it comes from external markets rather than Ozone it’s self , but initially other stable liquidity on the terra blockchain will take a bit to grow.

Ozone would obviously need to limit max leverage , available leverage but the factors on which that is decided would need to be modeled.

What type of UST exposure would users have in the event of a depegging? Should these CDP positions be liquidatable? Because the LP only has 50% UST exposure in the event of a depegging the CDP should still be liquid. Is higher leverage possible/safe?

I would also like to add i think a similar leveraged LP model would be possible with defi money markets but without farming or external rewards i dont think this can be sustainable.

Let me know what you think!

7 Likes

Just want to say that I think this is fascinating and I’m looking forward to what comes from the discussion. Bravo Papi

2 Likes

What if… we used another billion from the community luna pool… and simply bought stabilizing cash flow on other blockchains?

With ibc, would it possible not to support the peg with yields generated off of other blockchain rewards?

Perhaps atom and dot yields (inflationary coins) could be redirected to the peg stabilization by purchasing a basket of not only stablecoins but also autocompounding of those coins?

A stash of bitcoin as a highly liquid base layer may also be feasible since BTC exchange is now coming to cosmos.

2 Likes

Sounds like use of Olympus Pro treasury - “users can purchase discount but locked/vested Luna” - any idea if that’s the plan?

These amplified stability pools work amazingly to stabilize price, but with that, I think a black swan is also amplified. When there’s a run on a stable or even a loss of confidence, LPs will race to pull out as to not be left holding a fat bag of the depegged stable. Ive seen this happen on smaller scales and it is very ugly. Correlation to other stablecoins goes way up and so does reliance. This is a great mechanism when getting a small stablecoin off the ground but can be detrimental in some serious market weakness.

The buyer of last resort is a good feature, and can work assuming that there is some confidence left in the token… of course a discount helps sweeten that deal. The more mechanisms the better, correlation is always a concern and something we can focus on going forward.

I love your idea in general, just think we need to be careful not to make it an overweight stability mechanism. Cheers

2 Likes

What about the times when the other stable coins lose their pegs? During times of instability, all stables drift from their pegs.

The buyer of last resort in these cases are the minters/burners or the fiat-ust traders.

Right, but when LPs pull out , they will be liquidating the LP in to UST , and so for example when they un LP, the 500k in UST/Otherstable is converted in to UST/ otherstable, and let’s say UST is undervalued (ie depegged) so when you withdraw the LP you get 270k UST and 230k USDC, in this case you will need to convert the USDC in to UST and so this still creates 230k of USDC buying pressure for UST.

Essentially if USDC goes to 95 cents and UST is still at 1$ this is an arb most people would be willing to take . If both USDC and UST are 95 cents there isnt an arb. We should consider a diverse portfolio of stablecoins to spread out the risk , it’s unlikely they will all depeg at the same exact time. If all stables drift from peg the existing stability mechanism would still apply where UST can be burned for Luna. This liquidity would still dapen instability in this black swan scenario.

1 Like

I fear this would create significant UST selling pressure which is counter productive to the achieving peg stability. And these assets arnt stable,in the event of a crash this doesnt provide much stability for the peg. Compared to a basket of collateral and asset backed stable coins.

1 Like

@Papi Fascinating conversation! with respect to USDT ./ USDC depegging risk, what if we used an oracle to allow swaps at the effective value of said stablecoins instead of looking at 1 USDT = $1

2 Likes

So the idea here is that Ozone would issue debt , that debt would be used to open LP positions and the LP would be held in an Ozone debt contract (CDP) .

The swap rate would be determined by the bonding curve and demand for each asset in the LP pair.

You could perhaps use an oracle to determine the value of the loaned stable coin/collateral and then have ability to liquidate positions that go below the collateral threshold . This should derisk the protocol from other stable depeggings. Would need to be careful with short term volatility here and perhaps smooth out the oracle results.

Quick run through of how this might look:

Let’s say UST is 1.02$ and USDC is 1$ , a user could open a Ozone CDP which would loan UST, 50% of the UST would be used to purchase USDC and the UST and USDC are deposited in to LP, and the LP is held in the Ozone CDP.

The user is loaning UST against UST/USDC LP , let’s say for example the CDP minimum debt ratio is 110% (meaning the collateral is at risk for liquidation if it’s worth less than 110% of the debt value)

For sake of easy numbers let’s say the user puts up 10k in collateral and the initial debt position is 100k

100k in UST debt backed initially by 110k UST/USDC (54k UST/ 56k USDC) . Roughly , slightly more USDC than UST due to initial UST premium.

USDC drops to 90 cents. And let’s say UST is now 1$. This means the LP collateral would roughly loose 3% of it’s initial value , since the user only put up the minimum collateral the position is now below the min collateral threshold and can be liquidated . To further derisk Ozone , the protocol could self liquidate positions and perhaps offer a small incentives to contract callers.

Really great thoughts here. This is the kind of discussion that will take UST to the next level with Ozone. The mechanics of CDP merged with other stablecoin pools in theory should add the stability you propose.

That said, I agree with the direction of questions here. We would need some modeling on these ratios. Most importantly, blackswan like events of subsequent depegging after these CDP lever up by 10x or more.

I think liquidating these positions could cause further negative feedback loops if ozone facilitates and takes custody of the liquidation. Since this position would be underwater, I think taking this idea further - if there ended up being substantial liquidity in these products - we could further stabilize these CDPS with CDS. With interest that Ozones collects off these loans it could buy put barrier type strike based CDS options (.995, .99, .985 etc) from users that have additional capital looking for high short-term return that they would lock up on margin - collateral - to sell these options. Therefore ozone could exercise these as it liquidates those positions, bringing in additional capital to buy back and liquidate underwater loan without adding further selling pressure and help stabilize the peg. 1. This is probably something Sigma or Prism could help with.

The the reason I lay out the credit default swap option above is because I don’t believe ozone should continue to hold the risk of the depeg by buying back underwater CDP. The idea of creating credit default swaps brings in additional automated buying pressure. Therefore ozone would still execute the default liquidation but it would be instantly bought with a fresh influx of capital to support demand through the credit default swap.

So we assume that ozone will have some exposure to third party stable coins by default. It will fund these reserves by selling off discounted vested luna. Ozone will hold this exposure by default. This doesnt directly present risk to the peg and is somewhat of the cost of doing business unless Ozone were to only support loans in UST .

I dont think liquidations would create a negative feedback loop. Most of the risk is held by the CDP holders as their collateral is the buffer between the depegging and the CDP going underwater. The platform will self liquidate meaning when a CDP goes under the protocol will liquidate the LP , pay off the CDP debt , and take a penalty from the CDP owner. Ideally CDPs will only be liquidated in severe scenarios , day to day volatility shouldn’t cause stable coin LP positions to liquidate.

In the event of a USDC depegging and liquidation, the UST in the LP will be used to purchase USDC at the market. This creates some UST selling pressure where UST is used to buy discounted USDC . The bought USDC is repaid to OZONE. There is potential risk of lack of liquidity that should be explored as this could prevent CDPS from being closed automatically .

Love the idea of adding CDS as an additional stability/backstop mechanism but i need to chew on this a bit more to think about how this might be implemented and how it can best support this model.

1 Like

This makes perfect sense and is a key part I didn’t quite see or catch above.

Correct. And I get that. Really everything I proposed was to hedge black swan type of events so think through that lens - more feature-oriented as well because CDP has to come first. That said, I understand the customer is who bears risk though collateral, but they are also taking custody of the CDP, i.e. creating outside buying demand with high leverage. These large amounts of funds go back to Ozone via liquidation in the event of a larger peg break, flushing out some buy demand naturally. Ideally, we want that collateral to be picked from fresh buying demand that is not Ozone. At current, the way this mechanism is being proposed is that the CDP custody goes back to Ozone and then Ozone will need other CDP bids in order to issue it back out, This takes time and is also not guaranteed to have the demand needed to bring back up to peg. These possible lapses in which Ozone is exposed to the de-pegging when liquidating. In flash crashes like de-pegging events the time-lapse for fresh UST buying demand to come can have negative feedback loops without fresh automated buying in place.

I know what I laid out is complex and could also possibly be achieved by incentivizing dynamic augmented bonding curves to outside liquidators perhaps through harpoon and White Whale. For example, those liquidators could be incentivized to hold CLOB orders below the market just as laid out in the CDS put side barrier options. These liquidators providing this service are then paid through the CDP interest Ozone collects. This would work in a very similar way to institute more automated UST buying demand, which is what is needed in these black swan events.

Good thoughts. USDT is where this real risk would pop up in my mind. The stable swap invariant of the AMM curve is what also exacerbates this risk for the trade-off of tight liquidity around a narrow range. I was toying with this side of the equation yesterday but hadn’t come to this conclusion yet. Theoretically, this could be hedged somewhat by creating an MTP with USDC, DAI, BUSD. This would smooth out that risk and could be incentivized but not just selling UST.

I made an account here to respond to you after reading this entire forum thread. I feel like the large appeal for DeFi at large is that MIM and UST are decentralized stable coins. If we were to adopt a similar strategy as MakerDAO we will find ourselves in the ranks of “wrapped USDC”. Just from my perspective I have to say the full decentralized narrative is one that is very strong and should be protected at all costs. This is what makes UST so appealing to many that arrive at it, try to preserve that image and growth for UST will continue. What happens if circle is given a cease and desist for all activity involving interaction with algo and cdp stables? It would undercut the liquidity and create a huge domino for the whole UST system if all USDC involved with it were frozen.