[Discussion] Utilizing USDC in the PSM

First post after lurking for so long around here, so please feel free to correct me if I have any misconceptions.

The elimination of fees to the PSM-USDC has probably made Maker the most cost effective ‘swap’ for USDC to DAI than any other platform. This has caused the USDC that backs DAI issued to be increasing at a very rapid pace. While this is good for DAI adoption, it increases our risk in case our address is blacklisted. I think we need to do more to find ways for the protocol itself to slowly derisk and reduce USDC reliance.

What are the current efforts that is trying to tackle this issue? IIUC to decrease USDC risk we need to decrease the amount of USDC in the PSM, increase the proportion of other collaterals compared to USDC, or have a system surplus big enough that problems with USDC will not matter at all.

I have seen some ideas thrown around here last year and am wondering if we can use the USDC in PSM to earn yield via mechanism similar to D3M (with Debt Ceiling limited to a certain % of USDC in the PSM) or other mechanism?

Using the USDC earned
Reducing value in PSM
What I still cannot really think of is how to use this yield in USDC to lower the PSM or increase other collateral without increasing the DAI peg. If we use the USDC to buy DAI, we will be basically neutralizing the PSM and also reducing the amount of DAI in circulation.

Protocol Expenses
We can also auction the USDC to buy ETH from the market, which can be used for the protocol expenses (CMIIW but there are some ETH required for oracle costs right?). This reduces the need to take money from the surplus buffer (allowing us to burn more MKR)

Surplus Buffer
Semi related to the previous usage, if we can use the USDC yield for some protocol expenses, we can allow the surplus buffer to continue increasing faster, as far as I understand, more DAI in surplus buffer reduces our overall risk in case a black swan event happens.

All things being equal, each additional “dai” in the Surplus Buffer increases the PSM by 1 USDC. This comes from the accounting equation. You increase equity (Surplus Buffer), you need to increase assets and the only asset item than can move by itself (market forces of the risk-free arbitrage) is the PSM.

Actually that’s not the case, removing of fees didn’t do much. You can read the full study here. What is happening is that the market is no longer willing to leverage so people are borrowing less.

There is no clear solution of what to do when you have plenty of assets but your only scalable product is crypto-borrowings. Demand for crypto-borrowing is not elastic at some point.

There are only 3 solutions I can see:

  • fighting hard to gain on-chain crypto-lending market shares. Decrease the target rate of D3M to 1% and we would be able to reduce significantly the PSM. But we would earn less due to the depressed rates (but this point is not 100% proven tho, but the community decided not to try).
  • attract institutional crypto-borrowers like Nexo that are currently not using on-chain products. The market is big but onboarding is not easy.
  • invest in the off-chain world which is not correlated. The cleanest solution is short-terms bonds , @GFXlabs proposed a similar version which is quite inferior in my view but has the advantage that someone is working on it. Other RWAs can help but provide a different tradeoff in terms of liquidity/risk/scalability.
5 Likes

Actually that’s not the case, removing of fees didn’t do much. You can read the full study here. What is happening is that the market is no longer willing to leverage so people are borrowing less.

Thank you for linking this study to me, it is very detailed. Definitely did not know of the existence of uniswap 0.01% fee tier and the high gas cost for PSM usage. I guess that makes uniswap the de facto swap for small traders, and the PSM for bigger institutions, which I think explains why increases in PSM is usually in the millions of $.

  • fighting hard to gain on-chain crypto-lending market shares. Decrease the target rate of D3M to 1% and we would be able to reduce significantly the PSM. But we would earn less due to the depressed rates (but this point is not 100% proven tho, but the community decided not to try).

If we decrease the D3M target to 1%, will that not affect existing vaults? i.e. won’t that make borrowing through D3M more lucrative than opening a vault? Did the community reject the idea of reducing rates to 1% only or decrease in general? I think I saw a parameter signal request that wants to target a lower rate.

  • invest in the off-chain world which is not correlated. The cleanest solution is short-terms bonds , @GFXlabs proposed a similar version which is quite inferior in my view but has the advantage that someone is working on it. Other RWAs can help but provide a different tradeoff in terms of liquidity/risk/scalability.

I did not want to consider off-chain world initially due to seeing how complicated it is to set up and how long it would take, hence my preferrence on on-chain solutions. But as you mentioned there seems to be a lot of things that need to be considered.

It just feels like a huge waste seeing 3.5 billion USDC sitting in the PSM not doing much.

Yes, and 1% is probably too extreme but gives the idea. There is currently a real supply of 728 MM DAI on Aave. So even if you add $1B, not everyone can migrate there. You also provide an arbitrage from people borrowing in USDC that might migrate to borrow in DAI.

The borrowing rate on Aave is the marginal borrow rate (if you borrow 1 DAI), if you borrow 100M DAI you move the needle a lot and the pricing is no longer the same.

There are plenty of second-order effects making it quite hard to understand the dynamics. And you can add recursive liquidity mining on top of that.

This actually makes me wonder what will happen when liquidity mining on Aave stops.