Proposal: slow minting

The idea is to allow existing or new sponsors to mint synths in a 2 step scenario, similarly to how slow withdrawals work, giving them the ability to mint new tokens while being below the GCR.

Rationale:

When synth price steadily goes down over a period of time, EMPs tend to end up being over collateralized, which adds a huge barrier of entry to both new sponsors and existing sponsors that want to increase their exposure to the synth.
As of today, YD-BTC EMP GCR is 3.81481047. A new sponsor with 1 BTC($47k) available can mint up to ~12.3k YD, when a 2.5 CR (which most people would consider safe enough) would allow them to mint ~18.8k YD. In this scenario, using slow minting they would be able to mint 50% more tokens and still sit at a CR value they are probably comfortable with.
With the current implementation of the EMP, a sponsor in this situation would need ~1.53 BTC available to mint ~18.8k and then withdraw ~0.53.
A high GCR with no way to go around it when minting hurts the potential of UMA being used as an alternative to lending for unlocking capital.

How it would work:

A first transaction would indicate the amount of synths to be minted and, after a period of time long enough for liquidators to analyze the resulting health factor of the sponsor, a second transaction would withdraw the newly minted tokens.

New sponsor:
1st tx: Lock collateral and declare the amount of synths you want to create. At this point in time, position CR = collateral / (synth price * synths to be created).
Wait for a liveness period where liquidators may liquidate the position if CR < minimum CR.
2nd tx: Mint and withdraw an amount of synths <= the initial amount declared (I don’t know if slow collateral withdrawals allow you to withdraw less than requested, but basically it should work in the same way than those).

Existing sponsor:
1st tx: Declare the amount of synths you want to create. At this point in time, position CR = collateral / (synth price * synths to be created).
Rest is the same as new sponsor.

Notes

As Sean pointed out in Discord, this feature would require liquidators to be able to liquidate without having to repay any outstanding debt, they would only need to provide a bond and a liquidation price. This removes the capital restriction in place for liquidators, which is something that might require more thoughtful consideration.

Really interesting proposal and something I think we should seriously consider.

The nice thing about people minting above the GCR is that it gives us reasonable confidence that a liquidation should always be profitable. I think there is an attack vector where someone could submit a bunch of slow minting requests with little to no backing collateral, and just rely on liquidators eventually choosing to not continue liquidations, because they are taking a loss on gas fees + locking collateral to pay final fees. Potentially could solve this by also implementing a minimum collateral threshold, similar to minSponsorTokens, but I’d have to think about this more.

1 Like

Wait, this is so confusing… I almost took out collateral last week would have been below the GCR but still above the 125% CR to mint more tokens after I saw the UMA Twitter post and then watched the how to video.
So, do not do this? Due to the fact you will get liquidated because its a cheap grab for the liquidators since they only need a bond to take all your collateral?
The UI actually tells me I can take out a certain amount and it was actually under the GCR not by much but I found that odd. It wasn’t under GCR by much but when your trusting a tool that sets you up to get ransacked that’s not cool. If I wasn’t paying attention, clicked withdraw and got liquidated Id have been pretty upset. Just checked again now. If I didn’t double check it would let me drop below the GCR. Usually when you go under it turns red and gives a warning. I just took a screenshot.

1 Like

Do you think this is worth pursuing? I started working on this at Comparing UMAprotocol:master...akhanubis:master · UMAprotocol/protocol · GitHub to get an idea of how it would work and how it would work along with some other ideas (multicollateral YD + Curve YD metapool to improve UX and capital efficiency).

More on this, I think unlocking staked/pooled collateral is the lowest hanging fruit for increasing UMA’s TVL, but I fear that people wont choose UMA for that when they see that in order to assemble a 2x collateralized “loan” they need 4x available capital (or w/e the GCR is).
Alternatives that achieve a similar end result have come up recently and it will be hard to compete vs those when UMA real LTV is much higher than the desired LTV because of high GCR.

No, withdrawing below the GCR is perfectly fine. The GCR is separate from the 125% collateralization requirement (CR) that you mentioned. The UI will stop you if you are requesting a withdraw that would put you below the CR.