Maker Trilemma - PEG, Decentralization, DAI 1:1 collateral backing - USDC Discussion

I honestly can’t find a good topic catagory (apologies @LongForWisdom )

I pulled this from the tail end of another post. I believe the topic is even more important currently.

We can make it a formal poll but I am going to refrain unless people consider this an important discussion and we can frame a poll appropriately.

The issue that seems to be cropping up in different ways is the fact that USDC is now backing 60% of all DAI and this is leading various people to have significant concerns about protocol risk of the majority of DAI being backed by a centralized stablecoin.

This came to the fore at least in my mind due to post from @hexonaut regarding G-UNI as well as risk requesting some significant SF changes to try to get USDC off the books.


There appears to be a Maker Trilemma that can’t easily be satisfied.

DAI PEG to 1.
DAI fully backed 1:1

The question as a poll would be for everyone to apply a ranked choice voting of order of importance of the above to the above.

The reasoning for the trilemma is as follows:

Practically it appears we have to give up on 1 of the above to satisfy the other two.

Discussion of the points in the Trilemma

  • DAI backing 1:1 - MakerDAO possibly could manage the PEG if we basically mint unbacked DAI or allow the DAI 1:1 collateral value backing to slide. There are many ways we can do PEG management (there were a number of viable approaches some technical some managerial) if we give up on 1:1 DAI collateral backing. (NOTE: In conjunction with this I honestly would like to see us with better control over system operations so that we can shut down components vs. ESing the entire system. These are pretty heavy discussions should giving up on 1:1 DAI collateral value backing be chosen by polling).

  • Decentralization - I think a real discussion of what this means in terms of allowed collateral backing, as well as governance is appropriate. In reality I think we are talking about the system being anti-fragile but I think many people realize that Maker is NOT anti-fragile in any sense of the word, but I believe many people would like to move towards a system that has greater anti-fragility if possible. There already has been a lot of writing and thinking on this topic but when it comes to polling this needs to be spelled out in concrete ways for people

  • DAI PEG how close to 1. There has been this clamor about DAI being off PEG before the PSM - so much maker chose to take on increasing USDC to maintain this ‘hard PEG’ via the PSM. I honestly think people don’t realize how much risk has been taken on by Maker with the PSM to maintain the PEG. I brought up this idea of using USDC to mint DAI during the week of Black Thursday as an ‘emergency’ measure, not as something Maker should take on permanently. But once this was done people ran with it. Rather than approach anything with ‘moderation’ Maker pretty much has gone ‘hog wild’ not just with this but everything else. I personally found it ironic talking to people who were concerned why is Maker so slow to do anything, when in reality MakerDAO does a lot of things faster than fast. When we really need to pause and think we just dive headlong into something that gets political support (for good or ill) and hope it works or we will clean up the mess later.

Well folks the time to clean up the PSM and USDC exposure mess is now upon us. imo based on my reading of how legislation is working here we have perhaps a minimum of 1 year and a maximum of 3 years to deal with this situation.

We also need the community to speak up with a general poll.

Everyone can and should speak so everyone can answer the poll.
I really want to hear from thousands of people not just MKR holders, Maker participants but DAI holders, vault owners (past and present), and anyone else.

I also really want to understand how other protocols think that they will avoid protocol issues if legislation comes that basically forces stablecoin issuers to blacklist their protocols effectively freezing funds. Is this a real issue or not?

We keep dancing around how to deal with, for lack of a better term, what I will call the Maker Trilemma but in the end we have to decide IS this a true trilemma, or not.

Until we frame the real issues here we basically are pot shotting solutions and probably digging deeper holes for the protocol and avoiding the real issues which could broadside the entire DeFI space at some point.


I think this is hard to answer because the “decentralization” one is a suitcase word people pack a lot of different things into. Decentralized governance? Decentralized processes? Decentralized collateral? Decentralized ???

If you’re not talking decentralized collateral, then I’m not sure there’s a trilemma at all. And even if you are, that’s a trilemma now, but not necessarily in the future.

But since I know you asked for an opinion and not another esoteric question, I’ll bite. Peg first, collateral second, any definition of decentralization third.

Why? Maker’s business collapses completely without a hard peg on DAI. It is perhaps endangered, and certainly made less anti-fragile (a word I hate — the opposite of fragile or gracile is robust) by under collateralization. No matter what form of decentralization you like to put in your personal dictionary, Maker can still both survive and flourish, albeit with varying amounts of problems along the way.

As always, I reserve the right to change my mind in the face of persuasive words and reasonings.


IMHO the trilemma is as in academic economics:

Sovereign IR
Decentralization (or absence of capital controls).

With that trilemma I actually have a lot to say…

1 Like

This is coming up because of USDC honestly.

So you think it is preferrable to load up the protocol with centralized stablecoins and that this is a non-issue? I am just curious.

Expand on IR? Probably should be obvious to me but it isn’t.

Is the term PEG or more like ‘stable currency value’ as most currencies do not hard PEG to another currency or value ( like gold/silver). Last check prevailing economic theory is that hard PEG’ed currencies have too many problems. Which looks to be one of the problems Maker is encountering with the DAI hard collateral backing value.

As to general feelings on decentralization. I consider this to be the most ill defined piece of the Trilemma proposed above. But I think we had better think about this carefully and rapidly before Maker DAI becomes fully a USDC wrapper and legislation stomps on DAI through USDC basically rendering the PSM useless and worse Maker with 3B frozen USDC on the books and no way to unload it.

I honestly think a revisit to discussions I am sure have happened in the past with Maker as well as Bitcoin etc. is appropriate and appreciate responses.

If this is a non-issue be happy to just pass over it.

Somehow based on the whole G-UNI and risk rates proposal this seems like it is an issue that needs to be discussed.

Do Treasury notes count? Because I’m all about ballast in our balance sheet so that anything risky is dwarfed by the US government owing us money every 6 months (that’s the Treasury payment schedule). The pitch I’m making in my proposed MPCU is 5% of the Treasury market in 5 years — that’s $1 trillion. Dunno if that counts as centralized stablecoin or not :man_shrugging:

…like bank deposits? Or accounts receivable? Or a dollar-denominated debt?

Decentralization is lack of capital controls

IMHO the entire issue is completely anomalous in the context of the 21st century. Generally a financial institution would have places to deploy its capital safely. Here, exclusively because of the delays in onboarding RWA, Maker cannot achieve a proper balance sheet and has to resort to extreme measures. I am fully dedicated to making sure the onboarding comes sooner rather than later.


Thanks for the link.

I felt that RWA was the cleanest way out of the dilemma but if I am reading right here the issue is also related to ‘rates’ which ironically brings me back to a suggestion I had regarding having the vault SF be related to the debt facility utilization. Though on first thought not at all sure this solves the basic issues observed in the past.

I think given this came from USDC becoming a problem on the books is probably the most appropriate definition of decentralization.

I am not at all sure that RWA can still get us out of the basic Trilemma problem though so we are going to have to make a decision about what the Trilemma is for Maker and which one we are going to give on because if I am reading this right - we are going to have to give on one of them to get the other two the question is what are the three and which one will we have to give on.

Thank you again for the post and discussion.


I believe once RWA can be onboarded - the answer is to allow fluctuations in the DSR to keep stability - and for clients to use DSR derivatives to manage the risk.

In fact that has been my full time project for two years :grinning:.

It doesnt address the issues brought up in the last thread but it addresses other parts of the Trilemma generally.


I think the issue is whether our collateral can come under ‘capital controls’ or not. Or whether we can apply capital controls. I am having to think and potentially reframe this Trilemma based on really good feed back from @Akiva on the Impossible trinity in the context of crypto. In the loosest sense Maker is acting as a sovereign but using backing collateral to back money printing, and then forcing rates against dynamically moving market rates.

One of the key problems people still are not really considering or addressing is that this PEG problem got worse when ‘farming’ came on-line mostly due to the fact that returns on capital skyrocketed and there was no counter acting balance for cost of capital to counteract the liquidity drains. We still are seeing this today with now 3B on the books.

I don’t think either of you were here when I was asking if the PEG was a 50 or 150M problem vs a 500 or 1.5B problem. It is now a 3B problem and we are looking at lowering the cost of capital vs. raising it. I am not saying raising this cost is the solution but I also pointed out that it might behoove us to raise the SF to see if DAI demand would actually decrease as a test. At the time (and probably even today) I was shot down in flames with that idea. If you think about it if we had floated SF rates higher we might have been able to reduce DAI demand. Instead we dropped them and saw DAI demand increase dramatically.


Appreciate the effort.

I think we do really need a better framing of the issues here. I feel like there are converging circumstances that may just squeeze Maker into really bad situations. It was my perception that having a ton of USDC on the books ready for RWA to convert DAI loans into USDC would be ideal. RWA is not going anywhere as far as I can tell. The structures complex and probably are going to have capital control issues (legislation/legal) so that we are not going to be able to fully utilize this as a possible solution.

Honestly does it make sense to look at putting bonds on the books here or are we just digging a deeper hole on potential capital or (collateral) controls.

1 Like

That we require not just overcollateralization but also repayment only in our own token will always increase pressure on the peg unless market participants disbelieve we can approximate the peg at all. Each loan that creates current supply creates an even greater future demand.

That phase of market skepticism is behind us, so it’s important to realize that when a vault mints DAI, that is creating future demand so that the debt can be repaid. There have to be inflationary pressures to counteract this in the long run for the peg to hold.

Right now, the only way to do that is through the PSM. This is a feature, not a bug — our zero-yield debt with a face value of $1 is valued at >$1 in the market.

No one wants to print unbacked DAI, so we’re left with wrapping other $1 assets like USDC or bridging to the real world. Now that we have a simple, transparent, mechanical, low-risk, and low-effort solution to bridging DAI with something as safe and steady as Treasuries or even fiat, there’s nothing stopping us from literally doing what the Fed does when it needs to increase money supply. The PSM can then be used just for fine tuning and the heavy lifting of peg management rolled into real-world instruments in Maker’s own portfolio.

Ok while risking looking like a fool here because I am just thinking out loud vs. researching and considering carefully…

Not clear to me that each loan demands more supply because all DAI paid as interest basically has to buy back MKR so it returns to the market via this mechanism to be DAI neutral. Not doing this creates a deflationary environment but the amount of demand here is so high holding back 50M DAI in the surplus isn’t the constraint.

Also another point. If you believe in this cycle at some point you have to give up on 1:1 DAI backing to satisfy the required inflation because at some point there won’t be sufficient collateral value to provide the required inflation (pretty much the situation before USDC). The US Treasury btw has been making $1T platinum coins to back the $Ts being minted - is this really hard backed? I mean if you never intend on paying back any debt and inflating the currency away you are going to basically break money at some point which has happened repeatedly historically btw.

I agree right at the moment USDC is how we are printing DAI and allow it to remain backed. At some point that can break and this is why we are talking about this. Blacklist USDC in Maker and now what happens?

Actually there is something stopping us - we are requiring DAI to be backed 1:1 with collateral (in fact it is worse than this it is like >1.13:1). If we could make a $1T platinum coin we could mint $870B DAI which brings me back to the 1:1 collateral backing. Just because you can service debt fees doesn’t mean you have the assets to back the debt, not now nor in the future. But to economists as long as you serviced your debt - it is presumed good for all time. Like just because I paid my mortgage today I am guaranteed to pay it not just tomorrow but until it is paid off or rolled.

So what I am hearing is that Maker is just another sovereign entity - with additional restrictions. In the loosest sense we are a constrained sovereign that may also be facing capital or collateral controls.

We are at the experiment phase and laying out foundations to build on. I’m the first to be pissed by the slowness of things.

I would refer to the accounting post I made last year or the Crypto-banking article I wrote.

People seem fine by having the $3B USDC invested in T-bonds. USDC is no different than investing those $3B in t-bills (maybe Paxos or Gemini are closer to that). I personally think T-Bonds are too risky at this stage at scale and t-bills are yielding 0% (after operational costs). So we are planing to do exactly the same thing except it’s better because we would do it.

The PEG is needed for DAI to be a useful means of exchange.
Solvency (DAI 1:1 backing) is needed otherwise you end up as Iron Finance.
Decentralization needs to be defined and built. It’s the early days so most decentralization providers are still small startups that are still in the process of building the products (we are just a bit older trying to figure out what crypto-banking means). Not to say that the regulator didn’t even started the work.


From what I can see due to lack of any interest in the thread looks like people are generally fine with 3B USDC backing DAI.

If this is true then I can only conclude that radical rate changes like that proposed in

are NOT required.

Even so in the end perhaps 10 MKR whale wallets will continue to decide these issues.

1 Like

Frankly, of the 3, I think 1:1 backing is the least important of these supposed principles.

I’m seeing general sentiment towards Dai change for the worse, and it’s mostly about the USDC backing. If Dai basically works as a USDC wrapper, then why use Dai instead of USDC when you’re stacking not just USDC regulatory risk, Maker smart contract risk, Maker volatile collateral risk, and a looser peg. Actually the USDC regulatory risk is even amplified because while Circle can blacklist individual USDC holders for compliance, they could only blacklist the whole PSM, that would result in a haircut to all Dai holders.

And we can rationalize that realistically the risk of a USDC blacklist is miniscule for any number of reasons, but it doesn’t matter what we believe, if Dai holders are worried about USDC, then it damages our product-market fit. Decentralization and censorship-resistance is the only advantage we have over centralized stablecoins.

I think its key to realize that the value of Dai on the market at any given moment is equal to its redeemable value at ES ± some premium or discount based on supply/demand, market dynamics, etc. We don’t control the second term, except for the supply aspect, and that’s slow and imprecise. And it conflates monetary policy and peg management with our business development.

Isn’t it much simpler to just adjust the redeemable value? I believe that Dai holders will be much happier with a coin that holds its peg and is redeemable for $0.99 at ES vs a $1 ES guarantee + USDC tail risk when ES is never supposed to happen anyway. We trade a tail risk for a parameter we can raise and lower as required.

1 Like

I believe one of the main reason is the risk premium.
Holding usdc seems smart because it is not supposed to change in value. but that is not 100% true, it carries a risk the default risk. This risk is actually relatively easy to calculate because USDC borrow usd at 4% to match its accountancy so that means the risk is probably around 4%.

On the other side our risk team can calculate the risk on a eth vault. it is 0.4% if I recall.
USDC profit at 0% is -4%
eth-c at 0.5% is 0.1%

On top of the 4%, the PSM carry also a risk which is not included in the 4%.

1 Like

I largely agree with this positioning but the Growth team will have (and likely already has had) a tough time garnering integrations with a loosely pegged Dai. Without the PSM, our stability melts away and with it more market share to centralized stablecoins. We are somewhat between a rock and a hard place here.


We don’t have to get rid of the PSM entirely, but it should be a scalpel when we are using it as a hammer.


I don’t think this is true. I am probably one of the most vocal critics of how quickly the PSM is growing. The reason I joined the forum was to raise this issue in my first thread.

I absolutely believe cutting rates was the right move. This is ultimately the right way to boost supply. I think we should be using the PSM as the thing that fine-tunes DAI to get it pegged. Specifically, if the PSM disappeared, DAI’s average value over any extended period of time such as a week should be $1. If we’ve achieved that, then I consider the peg requirements satisfied.

DAI being fully backed is a non-negotiable for me - if DAI is unbacked, it will probably get bank run at some point and the mere fact that it can get bank run will shake confidence. Decentralization, as Paper pointed out, means different things to different people but on a protocol level, we are decentralized but the assets we use as collateral are clearly not (and I think the community is okay with that).

I don’t think that there is a trilemma here and I would question the assumption that we need to rank them in some order and can only pick two out of three. What we need is possibly centralized but always diversified collateral. The PSM is backing 60% of DAI is not diversified and that’s the problem for me. Cutting rates and moving faster with RWA are solutions.

Perhaps we should ask ourselves what we’re doing on this side of the rock. We have other ways to do large-scale supply management with the PSM reserved for the week-to-week fluctuations. I think we can and will solve that problem, but the DAO is not built to turn on a dime, unfortunately.

The good news is that now lots of attention is turned to the phenomenon of the PSM and money supply management in general. We already see lots of different ideas about how to deal with it – either as a problem or as an opportunity – but this is one area where the DAO is certainly not standing still.