[Signal request] Set PSM fees to 0%

Motivation

I’ve published the idea here. Most (all?) financial systems started with coins that weren’t swappable at par but fixed this problem as it is much better to have the same unit of account. It’s inspired by the Suffolk System which solved this problem in 1824.

It was also suggested in The case for Clean Money, which is a good TL;DR of this Signal :

Bringing the ecosystem to have a mindset of 1 DAI = $1 will take a bit of time so it’s important to start early (payment rails should be upgraded, broker-dealer so take it into account, other stablecoins protocols should join, …).

USDC is eating the on-chain volume market share and we shouldn’t stay passive.

The second step will be when our balance sheet will be fully invested (RWA, bonds, institutional vaults, plenty in the works). Then we will be able to increase the DSR. DAI will be $1 but yielding a better rate than the fiat-backed stablecoin (that can’t really compete).

Reducing the USDC is another kind of problem and making it difficult to deposit USDC in the PSM is more adding friction to DAI than anything else (people want to own DAI but we tell them to use USDC instead or to pay a premium, paying a premium to lend us capital at 0% …).

Pros

Stop making DAI holders pay

I estimate that at least $6M are spend on stablecoin swap with DAI (only for LP fees). While this is good for liquidity providers (like me), this is clearly making our customer pay for using DAI. It’s a tax on DAI adoption.

This equates to 0.1% on bank transfers. While uncommon now in TradFi (I still have one bank charging 0.15%), this shouldn’t exist at all in DeFi.

Simplify the perception of DAI

Currently, DAI is a stablecoin which value is around $1. This would anchor the perception that 1 DAI = $1. DAI would simply be a dollar with a Maker logo on it. You could swap it for free for fiat-backed stablecoin and get $1 there.

Facilitate Real-World Asset lending

DAI not being equals to $1 is quite an issue when dealing with RWA. This would significantly improve the value proposition.

Some are going to go through Circle then pay the 20bps to DAI (PSM) to repay their loans. For 90 days invoice financing, that adds up quickly (almost 1% yearly).

Fuel the future DeFi repo market

The SocGen deal is about creating a repo market in DeFi. The repo market is a short-term funding solution to provide liquidity to actors that have a temporary deficit in it.

This works best if all the stablecoin are at the same price so the liquidity is merged (think of Aave when you lend/borrow any kind of “tier 1” stablecoin at the same rate). Even a few basis points is an issue as there is a lot of transactions.

Simplify future payment rails

When I buy a $100 product on an e-commerce website, it cost me $100 no matter the bank I’m using. A merchant will likely take USDC or USDP at face value but not DAI (with the exception of our shop).

PS: Actually, I wasn’t able to confirm what Coinbase Commerce is doing but I assume USDC is at face value while DAI is not.

Increase DAI presence in DeFi stablecoins

Having a strong peg will force other stablecoins to fix their peg as well. They will be able to leverage DAI for that, hence lending capital to Maker at DAI interest rate (at max DSR so 1bps). Read more here.

It would be better for them to use DAI than USDC directly as it feels less centralized, less management is needed, and we provide an MKR backdrop.

Cons

Less revenues

So far the PSM generated 4M of profit for the protocol.

Most of those revenues are generated when we expand DAI supply (or when we introduced PSM-USDP). In a stable environment, the fees from the PSM are below 5% of revenues.

Not great for Curve

Curve is in the business of providing stablecoin swaps. We would provide a better alternative for DAI <-> USDC <-> USDP (assuming efficient markets). This might be an issue for their business model.

Increased exposure to USDC

Such a move will be arbitraged by market makers with Curve. This might lead to increased exposure of 200M in PSM-USDC (thanks @Primoz). using 1inch to get more liquidity providers, the increase might be 250M. This should be solved by institutional vaults, RWA and, more precisely, by short-terms ETFs.

Should MakerDAO set tin/tout of PSM to 0%?

  • Yes, tin/tout to 0% for all PSM
  • No, leave tin/tout as it is
  • Abstain

0 voters

Next Steps

This signal request poll will run for the next 2 weeks, until Monday, October 25th. If a majority of forum voters (excluding abstain votes) are in favor, the proposed changes will be submitted for an on-chain governance poll the following week.

17 Likes

Strongly support this, obviously. I believe this is the biggest way we are currently shooting ourselves in the foot on a daily basis when it comes to momentum and growth. 4 million in revenue, or whatever, is nothing compared to allowing Dai to be a viable alternative to USDC and other centralized stablecoins.

There’s also the context that a lot of large scale collateral solutions are finally seeing the light of day, like staked ETH, the direct deposit module for aave, and real world assets / backbone collateral such as corporate bonds, so the short term growth in USDC exposure should be seen as a positive rather than a negative, since it means more capital available at 0% cost that we can allocate to other assets and earn profits from.

IMO we cannot afford to drop the ball on this so I will follow this up with a MIP that would allow MKR holders to have the final say on this, should the forum process block it. Again, this is the biggest way we are shooting ourselves in the foot on a daily basis - seeing Dai left out of solutions that include all other stablecoins, and also letting people think of Dai differently as simply a “not-quite-stable” coin rather than a real 1:1 USD stablecoin. It is really a quantum leap in terms of Dais role in the market and we shouldn’t let it be delayed longer than the quickest our governance processes will allow.

as an example I learned a while back that most OTC desks have special markets they call 1:1 markets for stablecoins like USDC and other things where there is a 1:1 arbitrage available somewhere. Since Dai doesn’t have the 1:1 arb, it doesn’t get this kind of treatment and most OTC desks instead consider it a volatile cryptocurrency, using different, more complicated and more expensive processes to price it and make it available to their users. This has forced me to use USDC over Dai in multiple transactions in the past because it creates a significant friction and inefficiency, especially at larger scale.

12 Likes

You are right, FTX has a deposit section called USD; where you can deposit USDC, TSUD, PAX or USDP, HUSD, BUSD and DAI.

DAI is listed as a volatile currency, and it is surprisingly very expensive to withdraw DAI through that exchange.

Like that, I am sure there are many more cases.

4 Million feels good and comes in handy, but we should always look to think ahead to what we want for DAI.

We are looking for DAI to be the strongest and most solid alternative to centralized currencies, and also to be without a doubt the benchmark in terms of DEFI.

I will vote YES, worth $4 million, when it is in our interest to position DAI as the most solid and robust digital dollar 1-2 years from now.

2 Likes

I’m in favor of lowering the fees for PSM, but would prefer an intermediate option instead of going all the way to 0%. Maybe starting with 0.05% fee in, 0% fee out, and then reassessing for further reductions (potentially all the way to 0%) after a month.

Another option we could consider is “centering” the peg around 1% by reducing fee in but increasing fee out slightly above 0%. For example we could charge 0.025% for both fee in and fee out, which would result in a 0.05% spread / round trip fee from USD to DAI and back, but would create market conditions where the peg would trade right around 1:1 most of the time and never more than 2.5 bps away from peg (versus up to 10 bps off peg currently). This preserves our ability to earn some revenues from stablecoin swaps (maybe even growing revenues if we see a lot more volume), and also ensures we don’t negatively impact Curve pools which serve as an important driver of DAI demand.

5 Likes

An issue I have with this is that the fee_in and fee_out can be valuable tools to encourage rebalancing between PSMs. For example, if we feel we have too much in the USDC PSM and would like to switch some of the reserves to USDP PSM, we can do stuff like:
USDC fee_out at 0
USDC fee_in at 0.20% (essentially temporarily stop USDC intake)
USDP fee_in at 0, or even -0.01%
USDP fee_out at 0.2% (essentially temporarily stop USDP outflow)

I’m trying to figure out if there’s way we can do this without fee_in and fee_out but I haven’t found anything obvious, if anyone can chime in would appreciate

5 Likes

When the guy who thought all of this stuff up is using USDC that’s an issue in and of itself that needs fixin’. Like seeing Ronald McDonald eating a Whopper, the King of Denmark waving the Swedish flag or your father kissing some woman who ain’t your momma. Heresy!

NB: Changed vote per McMichael’s note.

1 Like

As written they don’t currently support negative fees. Just as an FYI :slight_smile:

3 Likes

The fees in the PSM serve several purposes that keep the system from going off of the rails.

The PSM module was deployed to help take on as only as many 3rd-party stablecoins as necessary to get Dai’s peg back down to an “acceptable” level after a market downturn. It’s stated purpose was to be able to take the hit from a market crash and get enough Dai on the market at a reasonable price to let vaultholders cover their positions. Moving the TIN to 0% means that we will take on to our books all of our competitors stablecoins until that market is completely saturated. According to the market, Dai is objectively more valuable than the other stables. I like to think it’s subjectively because we’re censorship resistant, but Dai’s value will only equal other stablecoins when we’ve cut enough corners on issuance to counteract it’s real benefits vs. other stables.

PE has been working for the last year and a half on various solutions to get USDC and other regulated stablecoins off of the books, and this does the exact opposite of that. Maker governors should be prepared to take on many multiples of the current portfolio regulated stablecoins if this is adopted. Having the PSM’s full or at even greater percentages of Dai backing than what we’re currently seeing is going to limit the appetite to take on more stables in a market crisis (when we need it the most).

Without any spread, there’s no limit to how much stable risk the protocol takes on, and zero friction for anyone looking to pass that risk to us. We would be taking on blacklist, technical, and insider risks for 0 premium in an environment where we absolutely can and should expect to be rugged on these tokens. This is a short-term fix that exacerbates a long-term problem.

There are also potential legal considerations around this change, which is why the Foundation did not want to touch the PSM. I’d strongly recommend getting a legal core unit to weigh in on the implications here.

15 Likes

I find @brianmcmichael’s arguments compelling, and am abstaining for the moment.

In particular, our reduced capacity to accept more of these in a time of crisis is one aspect I had not fully thought about. As it is, we hold nearly half the USDP float, for example.

I could see and support this policy if we had some method at hand to dispose of incoming PSM stables — @SebVentures’ RWF CU is exploring liquid, low-bandwidth bond funds. That is certainly what was in my mind at first: offload all but a billion into actual USD proxies to get safe yield and lower counterparty risk.

But we certainly don’t want to just grab them and sit on them, and don’t think anyone is suggesting that. I would feel more comfortable seeing a clear method to sterilize the counter party risk larger amounts bring to our portfolio first (likely bonds or other fixed income that scales safely).

Perhaps we should table this until we already have in place a way to put the stables to work without further exposure to credit/regulatory risk. Horse first, cart second, so to speak.

4 Likes

I think :thinking: the Curve markets will still have a position here as the pools will just get more efficient IMO

I’d like to second this. A few months ago, it seemed that we were talking about the fact that Circle could blacklist USDC and our exposure to it was unacceptable high. Even right now, it is an existential risk to us. This led to do two things

  1. Lower rates on other collaterals to boost DAI supply from there

  2. The introduction of the USDP PSM to reduce our USDC exposure

It appears that both solutions are working. DAI is now more backed by non-stablecoin assets than before this move and within the stablecoin backing, we are seeing USDP take over from USDC. This move to remove the tin completely torpedoes those efforts. So I am struggling to see why we would do this. If I’ve missed something (very possible), please do point it out.

Can I ask how this 200M-250M figure came about? And do you mean this amount in total or do you mean a rate of increase in the PSM? It seems unlikely that the cost of this move will simply be adding 200-250M USDC to the PSM.

Finally, regarding market share, yes, we are losing out to USDC but in a twisted way, we are losing out for good reasons! As Brian points out, the premium on DAI is not an accident - it clearly shows that users prefer it precisely because it is not backed by a centralized entity. Throwing away those fundamentals to increase short-term market share doesn’t seem like the best move to me.

4 Likes

I believe the 250M figure is based on the DAI currently available to sell on DEX platforms for over $1 (of USDC or USDP). This amount would probably be filled immediately because it represents risk free arbitrage. But we may also see greater demand for DAI if it trades at a 1:1 price beyond this - in some ways this is the point of lowering PSM spread to 0%, it increases DAI’s utility for RWA, OTC trading, and other use cases.

1 Like

Regarding USDC exposure

We have had this problem for more than a year now. One year ago the stablecoin ratio was 53%, it’s 56% now. We kept market share.

Regarding the tail risk of USDC, no fees can compensate for that. Taking 20bps or 0bps is quite the same when you lose your principal.

The last published solution to solve USDC exposure was gelato USDC <-> DAI where we have all the protocol risk (105% LR is not helping). We can still send those PSM stablecoins in DeFi protocols and achieve similar results.

DAI being more valuable

DAI > $1 can be seen as people wanting DAI because it’s better. Yet, we have only 5-6% of market share and not growing above-market growth.

Most new stablecoins are above peg mainly for farming (directly or indirectly). There is an inherent link between the yield you can achieve and the stablecoin buy pressure. I wrote about it last year and while the theory was not fully fleshed, it was kind of correct. The interest rates are now quite in line between USDC and DAI on Aave. This means that people don’t care much about DAI or USDC, they care about the yield. More precisely yield + stablecoin risk.

It’s difficult to analyze in detail as some players might be delta hedged (debt in DAI ~ asset in DAI so insensitive to DAI price).

Impact on stablecoin exposure

The $250M figure is coming from 1inch, how much DAI you could get with a risk-free arbitrage USDC <-> DAI (the PSM). Could there be second-order effects? Sure.

We have close to $600M DAI on Curve. Let’s assume it no longer make sense to swap there, and they all look for yield in another stablecoin, the net impact might be up to -$350M USDC in the PSM.

We might be able to do the same thing with D3M (depressing DAI yield on Aave to make people move to USDC lending), but the farming incentive makes things less simple.

I don’t see the problem, if it’s working, we just have to double down.

1 Like

This is important to elaborate on. The Curve pool is not a bucket that we can reach into and swap $250mm in liquidity out of to get the price to peg. The price on Curve is more akin to a tidal watermark. $250mm represents a single swap which will get Dai price parity for a block or two, but which will quickly be refilled from arbitrage trades coming from every other exchange on the market. We’re not looking at bailing out the Curve pool to get to peg, we’re looking at bailing out the entire DeFi market.

So far we’ve taken on 3 billion USDC to get the Curve balance from about an $800mm difference last year to the $250mm difference now. As we get closer to the peg the liquidity gets deeper and it requires us to take on increasingly more USDC to reduce that spread. Our 3 billion USDC represents the long tail pricing that gets us to the price that we’re at now, but that’s small change compared to what it’s going to cost to hit true market parity.

Unrelated, I’ll add that from a vaultholder perspective, generating Dai in the traditional (revenue generating) way is very expensive in gas fees, even for large amounts. The handful of extra basis points on the market can somewhat help in the recovery of those costs and reduce the cost of vault ownership. By further dampening the benefit of generating Dai in favor of these risky and unprofitable solutions we continue to discourage vault ownership, which is still the bread and butter of this protocol.

4 Likes

Maybe the fact that DeFi exploded again in spring is of somewhat importance here.

We introduced the PSM on December 28th with a 1% tin decreasing to 0.1% over 7 days (a 90% cut!). The PSM was filled quickly and was free to increase from January 27th. There was no significant impact over the next month (which I would say is even a bit odd and would require more research).

In fact, the main driver of stablecoins was the crash end of May. Our lending activity is in line with ETH price. There is an imbalance between DAI demand and DAI supply as ETH price didn’t increase as much as stablecoin DeFi activities. Should we penalize DAI holders?

By having DAI above a $1 you win when you borrow but lose when you repay so … ok you win the SF on the 0.1% of the loan. We should maybe ask what our biggest borrowers think.

The $250M number is from 1inch so all DeFi on Eth of which $200M is Curve. Sure there is a bit more, but hard to think a lot.

1 Like

Sorry guys, slightly emotional rant incoming. I really don’t like the argument that because massive USDC exposure is bad then we should shoot ourselves in the foot and deliberately damage Dais position in the market as a stablecoin. If that’s the way we wanna go, why not just shut down the PSM entirely and turn into rai? It’s not like the 4 million in income per year does anything to protect us against the hypothetical blacklisting risk of USDC, and similarly its not like having 3 billion vs 6 billion USDC makes a difference. In both cases it’s a fatal exposure.

Why do we want to hurt ourselves if it doesn’t actually help us in a meaningful way? This is really Makers biggest achilles heel: getting stuck in a spiral of self-flagellation where we are shooting ourselves in the foot on obvious decisions that anyone else would be able to get right, but we faceplant on them so we can pat ourselves on the back and feel like we are being “risk averse” even though we aren’t because we are refusing to move forward so we can deal with the problem at its root.

Choosing growth and accelerating forward actually provides us with a real, proper, long term solution to the problem of overexposure to stablecoins rather than half measures like shitty UX-destroying fees. We can solve the problem of USDC exposure without also punishing our users. Most of all, having a huge amount of USDC accumulated in the short run will act as a big, flashing beacon to all the potential RWA providers out there, proving to them that Maker has a lot of very scalable business to offer and this will help them muster the massive motivation they will need to jump through all the hoops and barriers to set up proper scalable and secure RWA facilities. The larger Maker is, the more Dai there is in circulation, and the better Dai is as a product, the more likely we are to attract the kind of high powered partners we need to place 100s of billions of future collateral in globally diversified bonds and sustainable and resilient flagship projects outside the US.

I am working really hard trying to get these kind of deals set up, but they are very complicated and it takes a lot of luck and effort to get it done. It sucks to put in all that effort if the DAO is also meanwhile making things harder by choosing to have Dai be a worse product for crypto insider virtue signalling reasons (when it doesnt actually solve the problem!). It just makes it more difficult to explain to potential partners of the future why we are a rational and reliable partner they can count on to help their business thrive and survive.

In the grand scheme of things, in terms of exposure and in terms of where the project will be 5 years from now, 3 billion USDC or 6 billion means nothing, and there’s no difference between the two, so we should just choose growth and focus on real, endgame solutions rather than half measures.

The other massive possibility having a 1:1 PSM opens up for are direct deposit modules that have a simple PSM integration allowing them to deposit large amounts of Dai that gets converted into USDC at a 1:1 rate into aave and compound. We could potentially deposit all 6 billion USDC or however much we will end up with in this way, within a very short amount of time.

10 Likes

I am relatively new to this space but here are my thoughts.

It seems we should keep the Maker mission in mind as we decide these key strategic issues.

“The mission of the MakerDAO products and the Maker foundation is to create an unbiased currency for the world, which of course means for everyone all across the world.”

I believe this means we want to develop a GLOBAL unbiased currency, where Maker becomes the dominant (or at least a key) stable coin worldwide and can be easily used by anyone. So we need to design a stable coin that has broad appeal across the world. Thus what key features does THE stable coin need to have? I strongly believe that two key features are mandatory:

  1. Folks want frictionless transactions: thus no transaction fees.

  2. Folks want value stability: thus minimal ‘bank run’ risk

The Maker folks have built something wonderful here and created alot of value. But there is still a long way to go before making Dai a global unbiased currency. This audacious goal cannot be achieved without taking risks. A low risk strategy at this point would result in Maker being a niche success, at best. Being true to the mission requires taking risks to build a better, more appealing stable coin and in turn significantly increasing Dai usage.

I just dont see how maintaining transaction fees helps us achieve the audacious mission…

Lets stay big and bold in our vision and execution

ADDED LATER:

The market structure and economics of stable coins is iimportant to issues such as this. Do stable coins have INCREASING RETURNS TO SCALE? Basically this means that for a given stable coin its marginal costs of production decrease as their volume increases for all volume levels. For Maker that would imply that creating/servicing the 100th Dai would be cheaper than the 99th, and the 1,000 Dai would be cheaper than the 999th and so on.

Search engine is a typical example of a business with increasing returns to scale. Once Google builds their search engine, the marginal cost of servicing additional customers is very low & and the development costs per customer fall as they are split across a larger group of customers.

If stable coins have increasing returns to scale, this implies the industry should become a natural monopoly or at least an oligopoly (a few huge players) at scale, tending towards a winner take all industry. I suggest that stable coins ( and digital currency) do have increasing returns to scale. I also suggest that stable coins have network effects which means that the stable coin becomes more valuable to a customer as other customers adopt it. Products with network effects such as Facebook tend towards monopolies as well.

Here is my main point: assuming that stable coins have economies of scale and/or network effects, the stable coin market at scale will have only one or a small group of huge competitors. Thus there is a race to scale to become one of the few winners. Slow careful growth will be quite risky as it allows as other stable coins scale more quickly than Maker.

QUESTION: do folks believe that stable coins have economies of scale and/or economic network effects? If so, Maker needs to prioritize rapid growth. (I didnt say reckless growth - one might suggest that Tether is growing recklessly)

4 Likes

I’ve been very consistent in my remarks around setting PSM fees to 0. I contend that the likelihood of getting USDC blacklisted is significantly lower than the scaremongering of crypto-libertarianism and paranoia makes it out to be. Given that the current USDC on the balance sheet is already fatal, there’s no difference between a little bit fatal and a lot fatal. We should maximize the benefit we get from having USDC on the balance sheet. By setting fees to 0, we can suck up USDC into the protocol like a Dyson. This serves multiple advantages.

  • A hard peg at $1, which has multiple tangible benefits around integrations, RWA, and marketing.
  • The more USDC we have, the more drypowder we have for re-allocating that USDC into whatever configuration we want our balance sheet to look like. This includes ESG bonds and green works projects as proposed in @rune 's clean money proposal.
  • Scaling the Dai supply which is critical for the Maker Protocol to sustain it’s leadership in the market.
11 Likes

I can add that more USDC backing means more flexibility when investing in longer term products because it lowers duration mismatch risks. It is much easier to offer fixed rate vaults, liquidity mining, invest in ETF bonds/RWAs and so on if you know that you won’t need to immediately start thinking how to decrease DAI supply if for some reason stablecoin backing decreases to lower levels.

It seems part of the community is still confused about “too much USDC” narrative and I do understand it, because at the same time we are making efforts to decrease or hedge USDC exposure: lower rates, G-UNI LPs for blacklisting reasons, diversification with different PSMs, generous fixed rate vaults and many other attempts to boost supply from non-USDC collateral. But in reality we are mostly switching 0 yield USDC collateral to something more yielding and this is actually a good problem to have, unless you are very worried about holding USDC right now. We are probably talking about <10% increase in PSM which won’t change things much risk wise, but yes it might have negative PR for those who still look at the USDC backing narrative differently.

13 Likes

As someone who a few years ago would have gone with some kind of mantra about keeping USDC exposure low at all costs, I’ve turned around and think scale&ammunitions dominate here. Thinking long-term and in terms of phases actually helps here. Fees at 0 give us a stronger path to the next phase at a cost of 0 new existential risks.

3 Likes