What should be the ideal C/R (Collateral Ratio) ratios for the different collaterals used on Comdex's Synthetic App?

[Ismail]

The ideal way moving forward was to have a single, and common C/R ratio of 150% for all the collaterals used to mint cAssets. Still, we concluded that there should be a dynamic C/R for every cAsset-collateral pair on further research.

We have defined a Risk parameter module where the collateral and the minted cAsset pairs are accessed on specific lengths to determine the C/R ratio accordingly.

The parameters are set for the collateral and the cAsset minted. The riskier the cAsset-Collateral ratio, the higher the C/R ratio for the pair.

The risk parameters for the cAssets currently are as follows(Values calculated for the last 90 days):

1. Intraday worst drawdown for the last 90 days
2. Volatility ( Standard deviation of the daily logarithmic value )
3. Average 24 hr volume
4. Minimum 7 day average 24 hr volume

Let’s break it down to understand it:
A token like $ATOM would score higher in the above parameters as it has more liquidity and good volumes as compared to something which is more volatile and has a bit less liquidity like $CMDX, the C/R ratio for keeping $ATOM as collateral would be less than that of keeping $CMDX as collateral as per the parameter scores. So, People would be able to borrow more capital efficiently for $ATOM compared to $CMDX as per the parameterized scoring mechanism.

The same would be applicable for the cAssets minted. Each one would be parameterized and given a score to how much C/R ratio would be needed to mint these assets.

This would enable a fluid C/R float as per the riskiness of an asset and protect user funds from unnecessary liquidation.

Open points to discuss upon:

1. Should we adopt a dynamic C/R ratio model?
2. What parameters other than those mentioned above should be considered while defining the C/R ratio?
3. Any possible drawbacks while implementing the dynamic C/R ratio model?
4. Time frames to be considered for backtesting the data?

I would like to hear your opinions & ideas and further discuss the topic.

[lyptis]

Comdex has to be competitive. It must predict what future platforms will offer and provide an edge. Leaving 50% of collateral as a buffer is user-friendly for those not experienced with liquidation mechanisms, but it essentially means a third of the value deposited is sitting idle.
We have to be honest about what users want.
Maximum usage of funds with less worries.

We also have to be honest about where the “safety” for users comes from. The value users save by not getting liquidated easily is directly proportional to the opportunity cost of the collateral buffer.

But if the history of Defi shows us anything, it’s that the best path to safety for users is well-designed information sources, directly available in-app, wherein users can easily understand risks, and will be notified aggressively before making wrong moves.
UI design can accommodate different risk profiles and the protocol itself can remain open and appetizing for all kinds of participants.

I would suggest 120% as a flat rate for all assets excluding UST which would have 110%.
These would be levels you can only get to through a series of clicks and acceptance of risk. The UI would have a suggested level for each asset, as you suggest.
Creating a modular system from the start allows many more possibilities down the road.
i.e. if comdex has competitive advantage over other synthetic platforms, it will integrate firmly into other defi apps first, propelling TVL and strengthening the peg of cAssets.

Having a flat rate also allows baskets of different assets to be used as collateral quite easily, a feature that is very attractive. Users can be exposed to ATOM and LUNA, while stabilizing their mint with a bit of UST, and utilizing cAsset functions in the same app.

Looking forward to launch,
Cheers

[RedRabbit33]

Correct me if I am wrong to think of this issue like this, but I have broken it up into two separate parts: minting asset and collateral asset multiplier.

1. Minting Assets: I like the idea of a dynamic C/R as different commodities have different risk profiles and I think having dynamic C/Rs should be a goal. It does however add a layer of complexity that I think many will have questions about. For me, it comes down to who is Comdex’s primary target audience. If Comdex’s primary target audience of users is already familiar with synthetics and commodity markets a dynamic C/R makes a lot of sense.

However, if Comdex is targeting a larger market audience that is generally unfamiliar with synthetics and commodity markets I believe that a fully dynamic C/R might not be the best initial step forward as it might be harder to attract users and capital. Rather, I think it might be easier for many to understand if Comdex started with a system where different commodity groups had different C/Rs. For example, the C/R for precious metals (i.e. gold and silver) would be different from industrial metals (i.e. copper and tin), and agricultural plant (i.e. corn and soy ) different from agricultural animal (i.e. beef and pork). I think it would be easier to explain and understand why Comdex has different C/Rs for different asset groups.

Personally, I prefer a phased or incremental approach that allows C/Rs to slowly become more dynamic over time as it would reduce risk and provide more opportunity for input on how a dynamic C/R can best be calculated.

I am supportive of Comdex starting off with different C/Rs for different classes of commodities though. I believe that 150% is a good ceiling, and that there are good reasons that certain classes don’t need such a high C/Rs, for example precious metals might only require 130% and industrial metals 135% but energy commodities might require 140% and agriculture might need to be 150%. Someone familiar with the commodities market can probably come up with an appropriate C/R range for each class.

2. Collateral Asset Multiplier: I personally am against a fully dynamic collateral asset multiplier. I believe that like property, there are different classes of tokens that have different value as collateral regardless of price. For example, using a $10k vehicle as collateral has different value than using $10k in jewelry or $10k in 1 year US Treasury bills.

Similar to having different C/Rs for different classes of commodities, I think different classes of assets should have different multipliers. For example stablecoins should have a multiplier ranging from 1-1.1, ‘Blue Chip’ or Tier 1 coins and tokens should have a multiplier ranging from 1.1-1.2, Tier 2 a range of 1.2-1.3, and Tier 3 a range of 1.3-1.5. I think there some objective criteria that can be used to originally place a coin/token in a certain tier, such as market cap, length of time on market, 90 day average trading volume, 90 day average price volatility, etc. Such a discussion may be premature if only a few assets will be initially accepted as collateral such as UST, ATOM, CMDX, XPRT, and PSTAKE. If more assets will be initially accepted or quickly added such as ION, OSMO, PSTAKE, JUNO, and SCRT, I think having predetermined tiers with a set of criteria may be helpful. It would probably be helpful to have a predefined mechanism for assets to change tiers automatically or by community vote.

[Ismail]

Thank you for the detailed suggestions!

Having a flat rate for tokens with a low market cap and low volumes would trigger heavy losses, resulting in a liquidation spiral. This is the reason for a dynamic C/R ratio based on the token.These parameters would be adjustable through further governance, and more parameters can be added if the users demand.
The commodities, as well as the collateral, will both be adjusted to an optimal point which is best for an amateur user who has low-level knowledge of synthetics and derivatives and also capital efficient or the people familiar with it.

Having flat rates, as @lyptis suggests, is a double-edged sword where if we have a flat rate, we compromise on user security, and if we don’t, the protocol would be rigid and not capital efficient, so we wanted to devise a mechanism where we have both of these in mind and treat every cAsset-collateral pair as differently with dynamic adjustment.

For instance, let’s assume a user mints cGold with $UST. These two assets can be termed as relatively safe, and the C/R would be relatively low as the risk ratio of cGold/UST is low and can be understood as a safe bet.
Now, if someone mints cOil with $UST we know the collateral is relatively safe, but as cOil is a more volatile commodity, the risk ratio would be higher, so a user would need to add more $UST collateral as compared to the above cGold trade.
The riskiest bet is let’s take, for example, let’s assume a low market token $ABC, First the governance votes to get $ABC as collateral to mint cAssets. Now the asset is onboarded but has a very low-risk parameter score, so this will require the user to post more collateral. In case someone decides to mint cOil, which is already a volatile commodity with a more volatile token $ABC, the user has to post the most collateral, as the collateralized and the borrowed asset both are unstable.

So @RedRabbit33, as per what I understand from point 2 that is a collateral multiplier is already implemented where the risk parameters are already set so that each Crypto asset that is put as collateral is evaluated.
The idea of changing the tier or risk parameters could be done every three months where the parameters are reevaluated, and the community can vote to bring those changes on-chain.

In my personal opinion:
There is one more dynamic that can be explored here; we set a low barrier C/R, somewhere from 120% to 140%. When the user chooses the collateral and the borrowed asset, they would be alerted of all these risk parameters and what would be their ideal or the protocol recommended C/R based on the RIsk Parameters framework. The user has a choice to mint below the protocol recommended C/R if they want to take that risk.
Furthermore, as @lyptis suggested, a basket of collateral assets is also being worked upon but would not be live in the first version of the app and would be something we would bring in the later stages.

Appreciate your views