Parameter Changes Proposal - PPG-OMC-001 - 2021-10-28

Parameter Proposal Group: MakerDAO Open Market Committee
Authors: @Primoz @LongForWisdom, @Monet-supply, @SebVentures, @Akiva, @hexonaut, @ultraschuppi

image
Source: Risk Premiums & Competitive Rates November 2021

Table Notes:

  • Risk premiums and maximum debt ceilings assume liquidations 2.0 and are calculated based on recent market conditions
  • Negative competitive rates are mostly due to liquidity mining rewards and Compound and Aave, as well as interest and rewards earned on deposited collateral
  • Lending products are not standardized across defi and cefi lenders, and therefore rates can not be strictly compared on a like for like basis

Lending Market Overview

Rates at Cefi lending platforms remain stable from last month. Among Defi lenders, rates have trended upwards over the month, and average borrowing costs on ETH and WBTC are both roughly 1% higher.

image

Aave has seen much higher rate volatility relative to Compound, with somewhat higher borrowing rates on large cap collateral assets overall; the above chart shows net borrowing costs on stablecoins against ETH collateral at each platform. This could be accounted for by a different structure for rewards (with much higher rewards going to USDC versus DAI), or a wider variety of assets available as collateral on Aave.

We’ve also seen USDC rates remain higher on both platforms relative to DAI, which may be responsible for some of the PSM-USDC-A outflow over the past weeks. On Compound, borrowing rates have averaged around 2% higher over the past month, with some notable USDC borrow rate spikes that were not witnessed in DAI.

Abracadabra (platform of the MIM collateralized stablecoin) has also seen significant growth in the past month. The chart shows the collective debt ceilings of the MIM platform, with the vast majority being actively borrowed. This has been led by huge increase in the price of the platform’s reward token, which has been driven by origination fees charged for newly minted MIM creating a flywheel effect. It’s unclear if this growth is sustainable in the medium term but worth looking out for. Abracadabra’s list of available collateral could also help show which potential new assets are most popular with the market.

Proposed Changes

ETH-A: Increase SF from 2.0% to 2.5%

image

ETH price increased by about 40% this month alone, which drove the demand for DAI minting. DAI from ETH-A increased by 670 MM or by 40% in October, which coincided with the price increase. It is hard to tell how much additional organic demand Maker was able to attract, as the majority of DAI minting was probably mostly related to the price increase.
Largest ETH-A borrower (Nexo) minted around 476 MM and a large part of the demand was driven by a few whales alone. Maker was also able to retain its market share and top position in ETH locked between 3 largest lending platforms.
Because of the market rates increase and bullish sentiment surrounding ETH lately, we are proposing an increase in SF from 2.0% to 2.5%. The best competitive rate at Compound would still be above the proposed rate (2.8%).

image

ETH-B: Increase SF from 5.0% to 6.0%

image

ETH-B debt exposure increased by 56 MM or 95% in October, which is 2x of the collateral price increase and confirms an uptick in organic demand. The runup in ETH price potentially attracted more leverage seeking users.
We are proposing a rate increase of 1% for ETH-B in order to catch-up on market rate increases and compensate for the higher risk premium. Note that the 30% risk premium calculated is due to the large number of new users that the risk model treats as unsafe because they haven’t been tested in a market crash scenario. We also believe the ETH-B users are less sensitive to such changes. The competitive rates would be lower though, but we believe the combination of OSM delay and low LR ETH-B vaults makes this vault very attractive and unique.

ETH-C: No changes

image

ETH-C debt exposure increased by 154 MM in October and is at an all-time high. This vault, despite having a high liquidation requirement of 170%, is very attractive for its 0.5% SF as it gets hard to find such a good deal across established lending platforms.
We don’t propose any rate changes for ETH-C, since this vault is in our opinion mainly meant to stimulate new DAI minting. Furthermore, the risk premium of this vault is very low.

WBTC-A: Increase SF from 2.0 to 2.5%

image

DAI debt from WBTC-A is at ATH and closing at 750 MM, which is is the current line. We reacted to this 2 weeks ago by suggesting to adjust DC-IAM parameters. The demand we are witnessing this month has to do with the BTC price increase of 42%, but the debt exposure changed by almost 450 MM, confirming there was an increase in demand from new users. We managed to identify a few users that have migrated from Aave due to better (potentially also more predictable) rate at Maker. Below you can see that Maker finally started to fill the gap in BTC locked when compared between three largest DeFi lending venues.

image

Because of the huge demand seen lately and increase in market rates, we are proposing a marginal increase in SF of 0.5%. Rate of 2.5% would be at par compared to Aave and Compound. Also note that there is a process of releasing a low LR WBTC-B vault soon.

LINK-A: Increase SF from 1.0% to 1.5%

image

There was a bit of uptick in demand lately, competitive rates are in line and we believe that 0.5% rate increase is justified.

YFI-A: No chances (1.0% SF)

image

No changes are proposed. There was a desire to increase rates, but competitive rates are in some cases negative and we don’t want to drive the demand away.

UNI-A: No changes (1.0% SF)

image

No changes are proposed. There was a desire to increase rates, but competitive rates are in some cases negative and we don’t want to drive the demand away.

MANA-A SF: Increase DC from 5 MM to 10 MM

image

Governance decided not to offboard the MANA-A vault as the vault type got fully utilized. The current 5 MM debt ceiling was in line with the risk model in the past, but high median collateralization of current vaults of 520% does suggest we can increase the debt ceiling for another 5m. Therefore, a DC increase to 10 MM is proposed. Another increase may follow later, depending on the collateralization of vaults, but Risk suggests caution since MANA on-chain liquidity isn’t very high (30% slippage on $3 MM MANA sold).

RENBTC-A: Increase SF from 2.0% to 2.5%

image

Increasing rates to match WBTC-A.

MATIC-A: Increase DC from 10 MM to 20 MM and increase gap from 3 MM to 20 MM

image

The Polygon team deposited $55 MM worth of MATIC collateral into the MATIC-A vault yesterday and minted 3 MM of DAI. We believe they are willing to mint more DAI and we therefore propose to increase the debt ceiling to 20 MM, which is in line with risk models. We believe the vault will still reach healthy collateralization at full DC utilization (expecting more than 265% if the full amount of 20 MM DAI is borrowed).

At the same time we are also proposing to increase the gap to 20 MM, so that the Polygon team doesn’t have to make too many iterative mints because of DC-IAM limits. However, Risk proposed to decrease the gap back to the 3 MM range as soon as the debt is utilized.

In the case MATIC is proposed to be offboarded, the Risk Core Unit will propose separate parameters to offboard MATIC-A.

USDC-A, TUSD-A, PAX-A, GUSD-A

We are proposing 1% SF for USDC-A. The idea is to make users with a collateralization ratio above 101% unwind. If they won’t, Maker is able to collect additional fees. The downside of this approach is that a certain amount of vaults’ debt (around 200 MM) will go underwater, meaning Maker would accrue interests that it would later need to write-off (while sweeping vaults over to PSM-USDC-A). But the amount shouldn’t be too large (2 MM linearly per year), plus we believe the liquidation into PSM-USDC-A could happen within months.

UNIV2WBTCETH-A

image

The debt is close to fully utilized and we are proposing to increase DC to 30 MM, raise gap to 5 MM and increase SF to 2.5% to match ETH-A and WBTC-A.

Final Note

The proposed changes, if confirmed, should increase yearly revenue by about 20 MM or from 73 MM to 93 MM DAI. Most of the changes are reaction to increased rates in DeFi and CeFi, bullish market sentiment and huge DAI minting amounts observed at ETH and WBTC vaults.

Proposed SFs will get included into next week’s on-chain poll on 2021-10-31T23:00:00Z, and if passed, will be included in an executive vote on 2021-11-04T23:00:00Z.

14 Likes

Nice report @ultraschuppi

Of note working on a growth proposal that includes Polygon (MATIC), Nexo, and other Maker institutional vault players. I really want to urge people to turn down the proposal to remove MATIC from out collateral list. My growth proposal is going to want to not just expand the MATIC vault but also to include a MATIC-DAI v2 vault. I know the V2 vaults have not gotten much traction in terms of borrow. My growth proposal is going to show us a clear way to expand those.

Other point. I keep thinking we need to lighten up on the wBTC rate increases - mostly because one of the things I want to focus on is these Token-DAI v2 vaults. I wonder if we could just do a 25bps raise on the wBTC and wBTCDAI v2 for now instead of the 50bps to send a small signal Maker does want more wBTC here.

Later on we can re-evaluate going higher on those.

Even so I intend to support this proposal.

Otherwise great job and thank you for putting in the effective borrow rates using the farming data!! So appreciated as I will be referencing this for the DAI growth proposal I am working on.

2 Likes

Generally support. Couple questions:

  1. Do we have any kind of guesses about how a 100 bps vs 50 bps hike will affect ETH-B demand? I would think they would be our most demand-elastic users, but I don’t actually know.

  2. Given the reasoning of safer collateralization ratios to up the MANA DC, do we want to think about increasing that formally to mitigate some of the risk concerns? Or would that not make much difference in the end? Maybe @Risk-Core-Unit has some thoughts.

Look forward to seeing this one up for a vote. Great work.

2 Likes

Our impression is that ETH-B users are less sensitive to rate hikes than other users, since they primarily use ETH-B for high leverage. When we look at how they utilize their DAI, its less about farming and more about making ETH leverage.

We could increase LR if this is what you had in mind. And yes, this then gives us more room to increase DC further. The issue that we start facing is vault type becoming less competitive than it already is. For instance LR at Aave for MANA is only 133%.

2 Likes

Regarding WBTC-A, is it possible to increase the line significantly?

It is very good to know that WBTC locked in Maker is closing the gap as compared to competitors, I also see WBTC locked increased quite a lot after the new debt ceiling increased from 750M to 1.5B.

But there is a possibility that 1.5B is going to hit soon if there is a parabolic run, and it is really painful to see that 1.5B is the limit in that case. Even so when WBTC locked is already closing the gap compared to Compound and Aave. I certainly will be very happy to see Maker leads in WBCT locked. Maybe 3B?

Yepp, I also assume that might happen. But we just doubled the line and we should see how liquidity of wbtc is evolving to further increase the line. Keep in mind wbtc is in the end not as decentralised as Eth :wink:

1 Like

Thanks for the very prompt response.

My concern is that the executive is weekly and then there is 2-day GSM delay so that might be some delay. But as long as the team is keeping an eye on it and reacts appropriately, then it is good

1 Like

the current executive proposal includes the proposed changes after they were approved by the previous onchain poll.

please help letting this exec pass quickly. thanks!

4 Likes

the exec just passed, we are going to have the new parameters at 2021-11-10T11:18:00Z.

thanks everybody!

3 Likes

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.