What's the point of adding stablecoins as collateral?

Since we’ve just received the very first - I believe - collateral onboarding submission by TUSD, I think it would be useful to discuss a question that is often brought up: What’s the point of locking up a stablecoin to generate a stablecoin? Why would anyone ever open a CDP backed by TUSD, USDC etc if they have to pay a stability fee and there’s no possibility of the asset going up in value?

The answer is that not everyone would want to do this, but there are some who really want it: Market makers and keepers.

For market makers, being able to collateralize stablecoins allows for highly flexible and efficient inventory management, and the result is that Dai will become significantly more liquid against any stablecoin that is onboarded as Dai collateral. This is because a market maker will be able to offer large trades at very low spreads, since their capital intensity becomes very low, because stablecoins can have extremely low liquidation ratios. At scale, you can almost think of it as if Dai starts behaving like a fiat backed stablecoin itself, in that it will be so easy to exchange it 1:1 with very low spreads to USD and fiat.

Here’s some specific examples:

Imagine a USD-backed-stablecoin as Dai collateral that has a liquidation ratio of 100.5%, and imagine a market maker with 1 million of the USD-backed-stablecoin in his inventory. A counterparty then approaches the market maker with 100 million USD, and wants to convert that to Dai. It seems like converting 100 million USD into Dai in one go would be very difficult without significant slippage, but with stablecoins as collateral it now becomes trivial. The market maker simply receives the 100 million USD, converts it into the fiat backed stablecoin, and together with his own 1 million in capital he’s able to instantly generate 100 million Dai and close the trade. Because his only costs are the cost of his 1 million in internal capital, and the stability fees on the stablecoin CDP until his trades balance out in the other direction (which in a very liquid market shouldn’t take long), he can charge a razor thin spread and still make a lot of money because the volumes get so high.

Another and more simple use case is for when the price of Dai trades above the peg, like what’s currently the case. To earn some free money, a keeper can take his capital and turn it into a fiat stablecoin 1:1, and then generate Dai, buy more fiat stablecoins (at a rate above 1 USD because Dai is above the peg) and continue the cycle of buying fiat stablecoins with generated Dai. A single keeper would theoretically be able to bring the peg back down by doing this, if the debt ceiling for the fiat stablecoins were high enough. However the strategy does rely on Maker governance continuing to lower DSR and stability fees in this scenario, since otherwise the keeper will be stuck paying the stability fee on non-appreciating fiat stablecoins and could ultimately end up loosing money. However as long as there is a way for keepers to signal that they are this kind of position and will eventually unwind it, it should work well together with governance ensure much greater stability of Dai.

13 Likes

Hi rune, glad to see you in the forum!

What about a third use case? if the Dai adoption was too high so that we lose the peg to more than 1$ with 0% DSR and we almost run out of safe collaterals.
can we use stable coins with a liquidation ratio of 100% and 0% SF to buy us time until we can scale the protocol instead of Global settlement?

1 Like

Yes, that’s similar to the second case I’m pointing to. But it’s important that all collateral types in the system always have an LR of more than 100%, or an SF of more than 0%, since they always introduce some level of risk to the system. (There’s only one exception and that’s for SCD in its specific use as collateral during the SCD -> MCD upgrade)

For centralized USD stablecoins I believe the risk parameters can be extremely low. SF could be less than 1%, and LR could be 100.5% or maybe even less. For centralized fiat stablecoins that are different from USD, the risk parameters would have to be a little higher, and probably would have to be similar to the terms offered when doing forex margin trading (100x leverage).

1 Like

Hi.

Agree that giving MMs and Keepers the best tools possible is somewhat paramount to this project. And hence adding stablecoins as collateral would be a great benefit. However I believe that there are at least some in this community (me included) that are worried about (not necessarily against) adding centralized coins as collateral.

Would it not be possible to create baskets of centralized coins wrapped in a ERC20 and then use this basket as collateral instead? The purpose would be to alleviate some of the centralization risk. For example if five fiat stablecoins were wrapped in a basket (20% each) with a LR at about 130%. Then theoretically one of the stablecoins could basically collapse (assuming the stablecoins are not correlated) and there should still be more than a 100% in collateral hence no MKR printing when liquidating CDP. The way I see this is that some of the centralization risk is moved from MTH to CDP owners. So instead of MTH deciding on introducing centralization risk we can kind of pass on this decision to CDP owners who would have to consider if opening a CDP with this basket of collateral would be worth the risk or not. I realize that a LR of 130% is very different than 100.5% but maybe this would be a more balanced approach considering the pros and cons of centralized stablecoins as collateral?

Assuming I am not missing any details that renders this approach impractical what is peoples opinion on using baskets of centralized coins?

Hi gin,

no basket type of collateral allowed under MCD. Only one type of collateral per CDP. I guess some crypto team could produce a index type stablecoin based on five fiat stablecoins and then use that as collateral. That could work. But doubt it is worth the effort.

Hi. If the basket (created by Maker or not) is an ERC20 and it is voted in as collateral would it not be “allowed” then?

If it allows us to add centralized coins in a decentralized way that would be worth a lot of effort in my opinion :slight_smile:

Using a basket just constrains the flexibility of MakerDAO. Using a separate collateral type for each stable coin and including 5 stable coins would provide the same diversification advantage while allowing the flexibility to set slightly different risk parameters for each stablecoin.

That’s true good point. Didn’t think about that.

Not sure I agree here. Is it not true that if one of the centralized collaterals collapses to zero because of a single point of failure when not wrapped in a basket MKR will be diluted? I.e. not the same diversification advantage as a basket could have prevented MKR dilution?

1 Like

Not sure I agree here. Is it not true that if one of the centralized collaterals collapses to zero because of a single point of failure when not wrapped in a basket MKR will be diluted? I.e. not the same diversification advantage as a basket could have prevented MKR dilution?

As long as the stability fee is set correctly to match the expected loss over time, and the portfolio as a whole is sufficiently diversified, the net effect on MKR will still be deflationary. I know it’s scary to imagine putting your trust in risk models, but being able to do so is the fundamental premise on which not only Maker rests, but society in general due to the role of finance in capitalism.

Now if your concern is that you’re worried about Maker imploding in the short run, before it can feasibly reach strong diversification, the thing to keep in mind is that the alternative to expanding the collateral portfolio as much as possible is to keep completely reliant on ETH, which isn’t really any better. Most people in the ecosystem tend to underestimate the inherent risk of ETH because it seems so old and so permanent when counting in blockchain ecosystem time. But the truth is that ETH has a similar and very real risk of sudden total collapse when compared to centralized stablecoins. While the network itself is decentralized, it has significant reliance on just a few centralized exchanges for its market value, most of which - unlike a centralized stablecoin - aren’t regulated or overseen in any manner whatsoever.

2 Likes

True. Makes sense. I think my counter argument to that would be that isn’t it more difficult to quantify the risk of a centralized coin compared to a more trust less system then? I might be wrong on that though.

But still it seems to me like there is a path to somewhat mitigating the risk of trusting centralized collaterals by using a basket? So still not convinced that this is not a path worth taking at least long term (even though it gives less flexibility in regards to setting risk parameters).

I am not really considering the short run. Long term however I would be worried about MKR taking a significant hit (because of MKR dilution) in a scenario where a centralized collateral goes to zero. If the LR is set to 100.5% and the collateral goes to zero does it not mean that at least in theory close to 100% of this collateral would need to be covered by MKR dilution? Agree that such a loss could be covered by a correctly set SF long term. But short term this would imo likely be a very volatile event which would be very nice to avoid :slight_smile:

Congrats on becoming a dad by the way :champagne:

Is the following attack possible?

  1. Own $x in regulated stablecoin.
  2. Mint Dai using a CDP at 100.5%(!) ratio.
  3. Trade the minted Dai for more of that stablecoin.
  4. Trade that stablecoin around some.
  5. Lock that stablecoin into the same CDP.
  6. Repeat 2-5 as desired (or until you reach a debt ceiling.)
  7. Short MKR with other funds.
  8. Intentionally and publicly announce to the administrator/owner of the stablecoin that you have used it for non-compliant means (in step 4).
  9. The stable-coin issuer is forced to burn the collateral based on these claims due to fear that they might be genuine. The DCS ‘liquidates’ the CDP, forcing a large multiple (due to 100.5% ratio) of MKR to be printed.
  10. Attacker flees the reach of the U.S legal system, or alternatively proves that actually they did nothing illegal with the stablecoin and claim insanity (or something.)
1 Like

This could be a major issue if the entirety of the stablecoin balance would have to be seized in this kind of scenario, without honoring the CDP debt. However it’s also possible that only the net CDP value is seized, meaning that first the CDP is seized and unwinded, and then the leftover surplus is what is seized.

Whether it plays out one way or another would likely vary from jurisdiction to jurisdiction and be a part of the legal risk analysis of the stablecoin.

2 Likes

Got you, I hadn’t considered that it might be fine to unwind first before seizing/burning. Legal risk analysis is going to be critical then.

There always seem to be so many edge-cases and attack-vectors whenever anything slightly non-standard comes up. Something tells me there’s going to be a lot of time spent writing adaptors in the future.

In terms of the arbitrage benefits of regulated stablecoins, I think it’s hard to disagree with your points. It is a much safer and efficient path to Dai stability for market makers compared to minting Dai from Eth.

1 Like

This is the kind of scenario I was thinking about. But maybe the cause would rather be the government of the centralized stablecoin’s jurisdiction shutting down the stablecoin. Or if a majority of an admin multisig were compromised etc.

Something to keep in mind is that both of the scenarios you are mentioning are unlikely to cause the value of the stablecoin for keepers to go to 0.

In both cases the legal claim by legitimate holders of the stablecoin still exists, so as long as the underlying reserves aren’t compromised it should be possible for keepers who can prove that they have legitimately obtained the stablecoin to get legal recourse to their claim on the custodian of the reserves. Even if the reserves are also frozen, eventually the claim should go through if it is legitimate.

Only if rule of law or property rights are compromised (which governance can select for when accepting a centralized stablecoin from a given jurisdiction in the first place), or by some other means the underlying assets backing the stablecoin aren’t there, do we end up in a situation where bidding on the collateral would actually be worthless for keepers.

As an example of what it looks like when a centralized stablecoin is shut down and how innocent users are treated in such a scenario, I think e-gold is a good case to look at https://en.wikipedia.org/wiki/E-gold

I think the creative MKR-shorting attack scenarios are more important to consider with legal analysis (from a risk-of-collateral-going-poof perspective), since they deal with completely unknown territory in terms of case law and legal precedent, so it is harder to predict what happens in those kinds of edge cases.

That improves the situation didn’t think about that.

Hi Rune.

I am a Dai market maker, and I have no interest in borrowing DAI on a CDP funded by USDC or similar tokens because those tokens have extremely high black swan risk. That is, they can be seized because of government action without due process, and then the frozen collateral can destabilize the entire Dai ecosystem.

We know that Dai is being incorporated in some SK-SNARK privacy schemes as a wrapped token. I could easily see a legal order from a financial regulatory body deciding to seize all regulated stablecoin Dai collateral “because of terrorists” and the entire system collapsing due to this.

I believe the path to growth for Dai comes from a higher ETH price, not from accepting dubious, government-compliant collateral bearing “freeze” methods.

I’m much more open to non-stable semi-centralized tokens like wBTC, DGX and the like, because they actually bring additional market types to the table. Given the fungibility of stablecoins and extreme ease of exchanging one for another, trading a slight increase in Dai liquidity for a large dallop of systemic risk seems a bad deal.

Put me down as a hard “no” on this idea.

2 Likes

What is the expected loss of a regulated bank-issued dollar coin like USDC being seized because elements of the US government don’t like non-regulated dollar-denominated (and private inside a ZK-SNARK wrapper) tokens like Dai?

Unless the system is robust against all this collateral going to an effective zero (in which case it should not be allowed as collateral due to losses it imposes on other users of the system) it presents an existential risk to Dai.

Hi IslandHunting and welcome to the discussion,

With regards to growing the DAI ecosystem there have historically largely been two major strategies, with the difference between the two being how the DAI system should interact with the existing financial system. Call them “crypto-only” and “love-KYC” or anything else you want.

With the “crypto-only” strategy collateral is exclusively sourced from other crypto assets, making the system lighter in terms of governance and more resiliant against attack. The downside is less collateral, fewer types of collateral and a more niche profile for DAI. The system interacts only weakly with existing financial systems as all funds first pass through other crypto.

In the “love-KYC” strategy collateral is additionally sourced from fiat stablecoins and any security token we can lay our hands on. The upside is a huge increase in collateral and a much larger role for DAI, at the expense of more complicated governance and more risks connected with all the new collateral. Additionally the new types of collateral will increasingly make for tougher KYC requirements which is hugely unpopular in some crypto circles.

Right now we are more or less headed in the “love-KYC” direction unless external events force a change. It is however very valuable to have participants with opposing views so please continue sharing your doubts.

1 Like

You probably mean low probability catastrophic event, not “black swan” event.
Black swan is something that is hard to imagine in advance. Uncle Sam wanting to dominate the world financial system is quite predictable. Maybe we should worry more about Ethereum black swans.

1 Like