[Signal Request] Reduce the USDC PSM `tout` to 0

Hi Everyone,

As you know we have a lot of USDC currently sitting in the PSM (~800M). While great for the peg, this also creates a lot of systemic risk. Certain community members have previously discussed reducing the tout to 0 to encourage actors to more quickly arbitrage out this USDC for DAI. By reducing the spread between tin and tout, arbitrageurs can more easily profit when DAI<USDC which will reduce the amount of USDC on our books. Therefore I would like to ask the community whether you are in favour of this change.

For more info on the PSM, click here.

Pros

  • Reduces USDC in the protocol which reduces systemic risks

Cons

  • Reduced potential income from PSM fees

Should we reduce the USDC PSM `tout` to 0?

  • Yes
  • No
  • Abstain

0 voters

Next Steps

This poll will close on 2021-05-27T16:00:00Z. If it passes with a majority, it will move to an on-chain poll the following Monday. If that then passes it will go into a future executive. If the poll ends in a “No” then no further action will be taken.

9 Likes

Hi,

A good signal request, but I voted ‘Abstain’.
I would really like to hear what Risk says regarding the tin and tout spread first.

3 Likes

This doesn’t exactly solve your problem. If you’re really worried about systematic risk, discussing raising tin might be the better first step.

Realistically it’s hard to sell 800M USDC. You cause some more upward pressure on DAI and you can hit the upper band adding more USDC exposure.

800M USDC exposure is a decent amount but as a % of DAI supply and MKR marketcap it’s not that bad? A question for MKR holders whether they want the risk.

1 Like

One of the cons, we will have a dai between 1.006 and 1.004 in curve due to the fees.

Another cons we won’t have a extra back stop under 1.

Also it doesn’t reduce the systemic risk as far as the PSM dc doesn’t change.

Idea to reduce the systemic risk would be to have IAM for the PSM with some thing like 500M and TTL 3 days.

What if we start to activate the Liquidation on tusd and USDc vault instead.

Let’s start with some stat you can find here. The PSM fees are not that big (<1M since the beginning).

USDC is a systemic risk but it’s also a stabilizer against a bank run. Such a scenario can happen if the end of the stablecoin farming season happens. The curve is 500M fueled mainly by farming. Not really different on Compound and Aave. If those positions are not delta hedged (i.e. they bought DAI on the market not having a crypto-vault), this will tank the price of DAI.

You can increase the SF of ETH-A and others but that will take some time during which the peg is in a mess.

Not saying this will happen, but it can. And breaking the peg has long-lasting effects on the customer perception.

Moreover, you are assuming USDC is trading at par which is quite accurate but not 100% accurate.

I would probably think that 0.2% could be a better step to start with.

10 Likes

I have talked about this many times and am strongly in favour of this proposal. Basically, if we’re removing USDC from the PSM, I’m all for it. Regarding some specific points raised

This is also an option but would allow DAI price to be slightly higher. I think people are less okay with that. Removing tout has no effect on the peg (since it is above $1 on average) but does remove USDC from the PSM whenever DAI dips even slightly below $1.

I talked about this here. The issue is that if we have 800M (and continue to add) during a bull market, we’re going to be quite screwed during the bear when DAI demand goes up and DAI supply goes down. Why not take every opportunity we get to remove USDC from the PSM while we can?

Could you elaborate why? The tin fees remains unchanged and would cap DAI at $1.01 wouldn’t it?

2 Likes

At first blush, I have two concerns:

First, if we want to simply reduce exposure to USDC, wouldn’t a PSM for, say, GUSD be a really easy first step at diversifying counterparty risk? Beyond that, I suspect we could enter into some kind of insurance or swap agreement with another party. Beyond that, if we feel this is an existential threat, we can set up a CU that literally becomes Circle — buy dollars, park them somewhere, collect whatever tiny yield we can.

Secondly, borrowing DAI to a big extent also creates DAI demand (for eventual repayment). This presumably creates upward pressure in a deleveraging cycle, so the act of getting more DAI into the wild ahead of that which aren’t specifically earmarked to repay a debt probably makes a more stable currency. In general, a deeper market and more DAI is probably some of the best protection we have.

I would like for us to finally decide if holding all this USDC represents an unacceptable risk, though. If it isn’t, we can move on. If it is, no time like the present to address it.

1 Like

I fixed it, from 1.006 to 1.004, basically as you have 0.004 fees, the 1inch route uses the PSM at 0% for 1.004 so the curve pool doesn’t get hit and stay at 1.004. As curve is currently the market place the dai will be between this range.

The systemic risk is not linked to what we have inside the PSM but what the DC allows as usdc inside the PSM. If we have 0 usdc in the PSM and the usdc slip to 0.5, the PSM will be filled up with 2b usdc. If we have 1b in the PSM it will be filled up to 2b.
The systemic risk is exactly the same.

So if we lowered tout to 0, USDC in PSM could probably start unwinding as soon as there is a bit of DAI sell pressure (see order book on USDC/DAI pair at Curve below). At tout of 0.4%, there needs to be about 850m sell pressure on Curve first before PSM starts unwinding.

I am not that much worried about loosing control if it starts unwinding. USDC in PSM is about 16% of all DAI supply and when we liquidated USDC-A into PSM it is going to be over 20%. I think governance could react in time if it saw USDC in PSM approaching 0 and peg to become destabilized.

There is of course a question of fees impact at unwinding. At current 0.4% tout we could collect milllions more for sure, but again I don’t see this happen any time soon, just because liquidity at Curve is so large around 1.000. You would currently basically need DAI sell pressure of 1.6bn for PSM to unwind completely, thats about 1/3 of all DAI supply.

There wouldn’t be much difference currently versus 0.4%. Still about 800m DAI bids on Curve between 0.998 - 1.000.

EDIT: The orderbook above doesn’t include curve’s fees, so numbers are slightly lower. Still, when selling 750m DAI at curve the rate incl. fees is 0.9976 (tout equivalent of 0.24%).

6 Likes

Not sure exactly how Curve works to arrive at this 1.004-1.006 range but in any case this seems reasonable to me. If we’re removing USDC from the PSM while keeping the peg there, that sound good!

Agreed, but as we unwind the PSM, the debt ceiling can then be reduced too. This would be the main advantage of getting that USDC out now.

I’m not sure if that solves the underlying issue of the DAI peg being dependent on the stability of other stablecoins.

If we do this with actual dollars, that’s centralized and if we do it with other stablecoins, we’re back to the same problem of DAI being dependent on their stability.

100% agreed. This is more a question of what do we do right now to minimize our USDC dependence. In that sense, I agree that little things like setting tout to zero help and the cost of it (lost revenue) is minimal.

About this, I think the risk right now is not representative of the risk we will face during a bear when the PSM will fill up much faster and we will have a lot less control. This is action now to avoid that problem down the road.

1 Like

@Aaron_Bartsch could you link to this somewhere in the original post for those that are less familiar with the PSM?

2 Likes

So is the problem with centralization? I thought this was about counterparty risk?

If it’s both, which is the more important concern, in the event the best tool for one isn’t good for the other?

Maybe I misunderstood your original point but what I meant was - if MakerDAO buys US dollars and parks them somewhere, we lose our “decentralizedness”. We are of course okay with issuing DAI on centralized collateral such as USDC as long as our processes remain decentralized.

The problem with too much USDC in the PSM is that it means DAI is on peg if and only if USDC is on peg. We can’t really claim to be a decentralized stablecoin if a significant chunk of what keeps it pegged is a centralized stablecoin. Not sure if everyone agrees with that but this is how I see it.

3 Likes

Since Wednesday, we’ve taken on roughly 200M USDC due to the price of ETH crashing. Would be nice to get rid of some when DAI dips below peg, as it has amidst the volatility.

On a separate note, what happens to these signal requests if the abstain option wins?

Abstain votes don’t count towards yes or no, they are ignored when determining the outcome. As it stands this signal will end successfully on the 27th and move on-chain shortly after.

That said, it would be good to get more participation, there are still only ~10 yes votes so far and less than 30 voters. I’d encourage people to campaign a bit.

2 Likes

Alright my 100DAI worth.

Lets talk about issues.

  1. USDC systemic risk. Blacklist or freezing USDC in the PSM and or regulatory changes that basically impair USDC in the PSM.

  2. What does getting rid of USDC and the PSM mean generally to the PEG because if you are arguing (1) and risk is the issue here lets talk about getting rid of the PSM and get back to decentralized PEG management schemes and the PEG floating again.

  3. General lack of work being done to get us off the USDC tit to manage DAI liquidity and decentralized PEG management.

These I see as the greatest issues. 1B of USDC exposure to (1) is a pretty big deal. I have seen ZERO discussion how to get rid of the PSM bandaid and establish a strong DAI PEG.

If curve wants to sit on the inside of tin and tout I am like - please do because I can see some serious issues if the DAI demand here reverses and eats this USDC. Can governance act fast enough to lower this. Frankly given what I have seen of governance it can’t act within anything less than a 7 day window. The better question is:

How can we lose 1-2B worth of USDC price support at ~1 on the low side of the PEG? AND the followup:
Can the SFs and DSR be raised fast enough to provide price support in such a situation?

I have felt for a long time if I could ignore (1) that Maker would be well served with USDC on the books that could be paired with DAI in LP contracts (to earn return) while providing liquidity. Also if the LPs are liquid they could be drawn out during such events and then the USDC places in multiple tiered PSMs (as I proposed long ago) when we were discussing this whole PEG situation generally. I had a crazy idea that one could then use the USDCDAI LP to mint more DAI to create the DAI sop mechanism. Having USDC in Uniswap couples it with other USDC that might be harder to blacklist/freeze btw without other legal ramifications. (this idea is a complicated topic with some hazards so I have not brought it up).

I let the analysis then now 7 months old stand and I am back to the same questions I had then and since I first came to Maker now almost 1.5 years ago.

LIQUIDITY! How will Maker sop up and release liquidity into Markets when it needs it.

We need a bloody better answer to this question than to release DAI liquidity for USDC and to sop it up using USDC in the PSM and SF, DSR.

If we are going to approach this lets talk about HOW MUCH USDC we want to unload and how much to keep and make a SECOND PSM loaded with the amount of USDC we want to unload with a tout of 0 rather than tossing it all up and hoping we don’t empty and have the DAI PEG start dropping to the low side with no means to sop up liquidity other than ‘rates’.

Central banks have mechanisms to directly sop up and extend liquidity beyond RATES in terms of direct market interventions Maker needs exactly the same mechanisms. We have tried to rely on market actors to do this and found this is NOT sufficient - perhaps it is time for Maker to step up here and consider real liquidity management vs. price fixing against USDC in addition to the basic rate tools.

If USDC is considered too dangerous to use - then we better find something else for PEG management - rapidly.

Lastly: Voted NO on the above for the basic reasons listed above. (No control over the amount of USDC that will be thrown here, and no reasoning both from systemic risk and alternative PEG management for how to manage PEG and DAI liquidity generally avoiding the USDC issue).

2 Likes

I agree that PSM tout should be 0.

  1. No fee withdrawals to USDC encourages vaultholders to generate Dai if they know they can exit to fiat without any friction. Think of this as a customer service that allows us to actually provide the value we say we’re providing.
  2. Removing the tout fee eliminates our carry risk in that spot between $0.9996 and $1.00. We don’t want all of this USDC on the books, we should use the opportunity to get rid of it in the case that someone finds an arb below $1.
  3. Freeing up this USDC with the tighter spread can likely increase fees in the long term by permitting more debt churn.
  4. When the PSM is empty, it’s an objective measure that the market has Dai under peg. We have a lot of tools to fix that, like raising stability fees and the DSR, both of which are arguably much better for the protocol overall than the marginal trade fee. Business development has told us that the DSR is one of their biggest selling points with potential Dai users, but we can’t turn it on because we’re holding large amounts of collateral that is not generating income.
  5. The PSM is a band-aid for reducing the price of Dai when it goes above peg. It works by reducing the overall ratio of collateral underlying each Dai token with a lower-quality token and is not intended to be a long-term solution to peg management. It adds a lot of risk to the system for little reward. Setting the tout to zero ensures that we re-collateralize Dai quickly when the market demands it, and that we free up the space in this and other future PSM modules so that we can weather downturns in the market when we need them.
  6. The entire system was built to correct the pitfalls of Sai, particularly the tendency of Sai to trade below peg. Unfortunately it’s design did not provide as many mechanisms on the north side of the peg. The PSM fixes that, but we really don’t need it when we’re under peg. The ability to manage rates and encourage repayment and savings of Dai under peg are part of the system design. I am in favor of letting that system work as intended.
6 Likes

Is there any reason why we can’t replace USDC with short term U.S. Treasuries? I know this might sound perverse to some in this world of crypto but Maker is obviously expanding outside of crypto. I recognize that I haven’t worked out how the DAO would hold a position in treasuries but it feels like an optimal solution. If not U.S. Treasuries, possibly a Japanese Treasury security. I do believe some stable coins like USDC have systemic risks, especially during market dislocations.
It would be great to hear a counter argument to treasuries.

I’ve been pitching this for a bit. It’s easier said than done unless the Treasury tokenizes them or you have an existing fund to partner with.

This is still the future, but it faces a lot of internal resistance within the DAO (though maybe less now that 1/3 of all DAI come from USDC) and some technical/legal complexities that need to be solved.

3 Likes

We will tokenize everything eventually. Focusing on the lowest bearing fruit first is definitely the wiser choice.

2 Likes