Governance and Risk Meeting: Ep. 39 (June 13 - 2019)¶
Video | Audio | Discussion
| References | Person | Text |
|---|---|---|
| # / 00:00:01 | Richard Brown | Hello, everyone. Welcome to the June 13th edition of the scientific governance and risk meeting. My name is Richard Brown. I'm the head of community developments at MakerDAO. We'll be hearing from Cyrus Younessi the head of Risk. What is your title Cyrus? How embarrassing. |
| # / 00:00:21 | Cyrus Younessi | I don't even know. I think it's- |
| # / 00:00:25 | Richard Brown | I'm going to crown you the Head of Risk. There you go. Cyrus Younessi, the head of risk, will be walking us through collateral risk today. I'm not going to steal his thunder. I'll let him get into that. I have some things that I'd like to say at the top of the call though. I want to make sure that people are as engaged as we can possibly incite them to be in these meetings. So, in the group chats, I've posted a link to a reddit thread. In that reddit thread, every week, we post a summary of the call. |
| # / 00:00:55 | Richard Brown | All the interesting points, David has been kind enough to actually recap the entire thing for us. He also provides links to timestamps in the YouTube video, so it's easy for people to get catch up, get caught up with what's been going on. In that thread, we continue the discussion for the rest of the weea k and this is meme I keep on hitting over and over and over again. But it's worth doing, so I'm going to keep on doing it. |
| # / 00:01:20 | Richard Brown | We have discussions and we have debates resurface, new issues and new ideas in these calls. But we don't make decisions here, whenever possible, I guess. Then we continue the debate in mediums that have a longer life cycle, or lifespan, than these one hour calls in North American friendly time zones. And currently, that's Reddit. If we surface some important issues, get into reddit, that reddit thread, debate with us. We'll come up with some conclusions, test the temperature of the community and see how things move on from there. |
| # / 00:01:57 | Richard Brown | We'll be introducing new tooling in the future that's a bit more sophisticated than what reddit has to offer at this point, but we haven't quite outgrown it yet. In the interim, continue the debate there. Preamble number two, we are extremely interested in hearing everybody's opinion in these calls. We have a lot of really smart people here. These calls aren't designed to be me, Cyrus, Vishesh talking. They're designed for us to talk and moderate. |
| # / 00:02:26 | Richard Brown | It's our hope that the smart people on this call will be asking questions, at every conceivable opportunity. If they have a microphone, feel free. If you have a microphone, feel free to jump in and ask a question whenever you like. Interrupt us, it's not a problem. If you don't have a microphone, you're supposed to be working, you're in a cubby somewhere, type your question into a chat and one of us will get to that during the course of the call and we'll ask it for you. We also have been having a sort of extended Q&A session at the end of these calls. |
| # / 00:02:58 | Richard Brown | At the 60 minute mark, we transition to general Q&A type plan, and we generally hang out for a half hour or so, just asking whatever might be on people's minds. So if you're interested, or you have some kind of of the wall question, wait till the end and we'll get to it at that point. I think that I'm going to leave it there. It feels like I'm forgetting one of my favorite preambles, but I'm not sure what that is right now, so I'll skip it. Let's talk about governance for a second. |
| # / 00:03:28 | Richard Brown | We're seeing some interesting things beginning to happen in the world of governance. And that interesting thing that I want to talk about is something that makes me happy, I guess and this is cool. We're seeing some rational behavior with the voters, which is nice. Governance is a new thing in this space, and it looks a new thing to the internet in general I guess. But it's even newer for crypto we're all trying to figure out how governance works. |
| # / 00:03:57 | Richard Brown | I like to tell people or whenever I get the opportunity that governance is a capital H, hard problem. Nobody has the solution. Algorithms will not save us. It's ultimately about people and communication and sociology, more than it is about game theory. But I guess there's an overlap; you know what I'm talking about. One of the things that we've seen over the last couple of weeks is some encouragingly rational behavior. |
| # / 00:04:25 | Richard Brown | It's my impression and I'm willing to be convinced otherwise, but it's my impression that the voters are fairly engaged with the data coming out of these calls. And my impression of that is coming from the fact that, during the polling process, the stake weight is being applied in a conservative manner. And it seems like right now the team or the MKR voters are testing the limits of the lower bound of the stability fee that still maintains the peg, which is very encouraging … ze've been hovering around the midpoint of the stability fee. |
| # / 00:05:08 | Richard Brown | So, we've been hovering around the point where we stay where we're at. There's been this mild competition, between voters staking 1% lower than the current stability fee, and a more conservative set of voters seemingly preferring the current stability fee. And in my mind that's an encouraging behavior. If things had gone of the rails, we would see stakeholders with conflicting incentives, attempting to skew the stability fee in ways that might benefit them most, and we're not seeing that. |
| # / 00:05:43 | Richard Brown | So, the priority still seems to be maintenance of the peg, which is what we're all here to ensure. That makes me super happy. I like the … I'm known for my optimism, so I like to start things off with some good news. Second thing I want to talk about is something that's far more murky and far more complex. And we've been wrestling with this in these calls for weeks now. I keep on asking that question without presenting an answer, because they don't have a clear answer in the sense I ask them from the community. |
| # / 00:06:16 | Richard Brown | So here's the issue. We have a governance system that's basic but works. It's basic because we're as a group, as a foundation unwilling to experiment with governance, because this is serious business what we're doing here. We can't really be playing games with this situation or testing out new technologies. We don't understand how that works. We have a very basic polling system. We have very basic governance system, voting system. We poll for signals. We take that signal; we add it to an executive vote. We wait to see if people think that vote is good or not. |
| # / 00:06:55 | Richard Brown | And then when the community moves in that direction, we implement those changes to the system. It works. It's great. Actually this is another thing that I'm going to bring up and this is the dangers of the new normal. And it's moving into a system and then assuming that's the way the world's going to work. And an example of that was we had, the peg had deviated. We were locked at 0.95, 0.96 for a long time. We raised the stability fee. We needed it to be responsive and agile in order to do that. That required a weekly cadence of votes. |
| # / 00:07:31 | Richard Brown | And so every week, we all got together in this forum. We hash things out. Polls were put into the voting mechanism, the community voted. Happily we're in a situation for the last three or four weeks where the peg is healthy. The voters have signals, last two weeks. And this is another great signal is that they prefer not to change the stability fee at all. So we're still looking at 16.5 for this week. So that raises the question, is this new normal, the one that we want to be stuck in? Do we need to vote every week? |
| # / 00:08:03 | Richard Brown | And I think that that's a question that needs to be answered. And we need to figure out how to answer that. So that runs us up against this next conundrum that we have where, in the current voting system that we're presented with, we can only have one poll at a time, even more restrictive, even though we actually have one vote. So, you can vote on a pole. You can stake your Maker into a poll; you can stake it into a vote but not both. We can't have more than one poll at this point. These issues are all going to be solved in the coming weeks, with a new version of the portal that we're affectionately calling 1.5. |
| # / 00:08:34 | Richard Brown | Once that comes out, we'll have some more freedom. But the question, I would like this group to begin debating and thinking about and then finishing that debate in the reddit thread is, we're seeing stability right now because this is the second week in a row that we have not chosen to move the stability fee. Has this given enough breathing room to consider that next week, we don't have a poll for the stability fee at all? Next week we have a poll that is related to cadence. So next week, do we pull for a, should we be voting every week or should we be voting every two weeks? |
| # / 00:09:12 | Richard Brown | And it's my thinking that we should also have another poll … or actually here's where things get complicated. We can only have one poll as I previously discussed, but here's the answers that, the questions that I want answers to. How often should we be voting and what are the steps in that vote? So, do we vote every week? Do we vote every two weeks? Do we vote in increments of 1% for the stability fee? Do we vote in increments of 2%? I think that we can formulate questions that answer, that cover all those options with a single poll. |
| # / 00:09:45 | Richard Brown | That means it's going to be some, for … if I'm thinking about this correctly, four polling options with various combinations of 1%, 2%, one week and two week. And then next week we assume that we're just not going to move the stability fee at all. We're going to devote next week to figuring out what the proper cadence is. The issues, the larger issues at play here are questions around apathy and friction and engagement levels, that the community needs to consider. We have to ensure that the system is secure. |
| # / 00:10:21 | Richard Brown | The system becomes secure by having a significant amount of MKR locked up in the voting system. If, we are voting every week and there's no urgency to that vote, people began to drop off and potentially the security of the system begins to drop off as well. And so these are questions that we all need to answer. We need to hash this stuff out. I don't know how to do that. And let me finish this off with some cautions. I'm a highly skeptical about the notion of rough consensus. |
| # / 00:10:55 | Richard Brown | If somebody wants to go back and listen to 27 hours of risk meetings, you'll find out why. But I could recap it for you. A rough consensus is generally a matter of perception. So, whoever thinks they got rough consensus and whatever forum they happened to have been conversing in and the strength of their opinion generally decides what rough consensus was. You can have 20 different people come out of the same call with different ideas of what a rough consensus is. The reason why we have a polling system is so we don't need to rely on rough consensus. So here's the conundrum. How do we get enough consensus to alter the polling system when the polling system is a thing that we use to get consensus? |
| # / 00:11:37 | Richard Brown | So I'm going to kind of leave that with the community to stew to over a bit. If people have solutions or strong opinions, please fill me in right now. If you want to think about it for a bit, this continues as a discussion in the reddit thread. I'll recap my thoughts. I was a bit over the map, all over the map on this one, but ultimately I want to know whether next week, we swap out the stability fee poll with a cadence poll, and let's figure out those issues. All right. Nobody has jumped into the fray. I'm not seeing anything in the group chat as far as questions go. |
| # / 00:12:18 | Cyrus Younessi | So couldn't potentially vote in people who are approved to conduct polls based on almost like a community delegates, or like how our risk teams would be voted in, could vote in a governance person to run the polls. |
| # / 00:12:39 | Richard Brown | Yeah, definitely. We could … That's a huge step in my mind though. And I think that we'd have to think about that because delegation is … I think when it comes to governance, delegation is the number one request or suggestion or idea that gets floated. I'm not sure that I've seen anything yet there's demand for a delegate system. But people keep on floating that as a concept, and refreshingly some people have come into the chat or it's different forums and said like, “Hey, I don't understand governance. I don't really want to. I just want to know that somebody I trust does and I want to let them vote with my Maker.” |
| # / 00:13:16 | Richard Brown | I think that that's an enormously mature position to take, and I think that fostering that would be great. But here's something else that I've been thinking about recently more and more, we just had this long discussion about it in bracket as well, is that I'm becoming more and more a fan of this idea of where we look to the community to solve these issues. And from that perspective, it's … governance takes on the same model that MakerDAO has taken on. |
| # / 00:13:45 | Richard Brown | It's MakerDAO's was responsibility to promote and foster and grow the core protocol, and then create proof of concept implementation to that protocol and then incentivize the community to build on it. So, we write the contracts, we release those contracts to community and an enormous third party ecosystem builds a lot of amazing tools to adapt and CDPsaver and numerous, numerous other ones. |
| # / 00:14:10 | Richard Brown | Why don't we look at governance from that same model? We've released the chief. We have the, that's the core protocol. We have the supporting portal. That's the reference implementation. And it's my hope actually that the community as a whole rises up and says, “Hey, you know what? We need delegation. We need somebody to be trusted to do this stuff. We need to organize some kind of campaigns around different collateral types or different stability fees or what have you.” |
| # / 00:14:40 | Richard Brown | And then they build out some tooling for us, and then they encourage Maker community to join the experiments with their delegated voting mechanism that interfaces with the chief. I'm concerned that MakerDAO begins to solve a lot of edge cases, and then adding friction to the core voting mechanism that would need to be pulled out of the core voting mechanism, if it turns out that it doesn't work the way that we expect. |
| # / 00:15:11 | Richard Brown | I guess that was a long winded response to your question Cyrus, and I kind of took it in a direction that I find more appealing, but I think that delegation is a great idea. I'm just not sure that we should be the ones that do it. |
| # / 00:15:25 | Cyrus Younessi | I mean, should we offer up some solutions, some ideas? Do we have any potential answers or- |
| # / 00:15:33 | David Utrobin | Or if you consider the idea for delegation, you also, you need to think about it full circle, and make sure that Maker voters have veto power in case like the delegator decides to do something rogue. I guess it's not really like so much of a risk, but it's still a risk. |
| # / 00:15:52 | Richard Brown | I don't think that Maker needs to do that. I think that whoever comes up … So, this is part of it discussion that really got very interesting. I should summarize it. Because we were talking about the … to size into what I said earlier in the call that governance in the crypto space and experiments in general, the governance that MakerDao does is not an experiment governance that we do is utilitarian in the extreme. |
| # / 00:16:15 | Richard Brown | But people talk about future keys. They talk about liquid. They talk about holography. They talk about delegates. They talk about DAO based votes. There's a lot of experiments that can be had in this ecosystem. And I really personally don't think that that's MakerDAO responsibility to experiment with how governance can quickly edge cases of governance. But I think that it is MakerDAO's responsibility to encourage that behavior in the wider community, and do whatever we can to support it and help them out. |
| # / 00:16:46 | Richard Brown | But, if some group of people got together and decide, “Okay, we have a few tricky …” I don't even know if I'm pronouncing it correctly. “We're going to build a DAO around it. We're going to spawn up a Aragon DAO. We're going to allow people to vote, bet on what the best outcome is and we're going to incentivize, Maker voters that use our system with some kind of mechanism, and then we're going to reward people that voted correctly in our future key model.” |
| # / 00:17:14 | Richard Brown | I think that's great. I think that DAO should be spawning up all over the place. There should be … There could potentially be another DAO that's like Maker conservatives or the Maker Hawk party or something. And then they can do whatever they like, start aggregating as much Maker as they want and then, or if it's self-organize themselves into smaller units, parties if they like. But, the amount of overhead sophistication, the complexity, the edge cases that we would have to anticipate as a foundation at the core protocol level to handle each one of these experiments is mind boggling. |
| # / 00:17:50 | Richard Brown | I think that in my opinion, it'd be super nice to see the community come up with some solutions, and then do some experiments and we can see what works and what doesn't work. |
| # / 00:18:01 | Vishesh Choudhry | So Rich, can I ask a question? |
| # / 00:18:01 | Richard Brown | Sure. |
| # / 00:18:04 | Vishesh Choudhry | So I tend to prefer not to give opinions, give analysis instead. But, I'll just ask a question I guess on this topic. It may not be Makers responsibility and I understand it's a huge problem to solve, but like voting delegation is not Makers responsibility, but it is Makers problem. I think in the sense that, it has a big impact on how governance goes down, the level of engagement, which ultimately impacts the types of outcomes that are achieved in this process, which impacts monetary policy and thus impacts the system. |
| # / 00:18:43 | Vishesh Choudhry | And so in that sense, you can't really get away from the problem, even though it is a big one and not necessarily Makers responsibility to solve. |
| # / 00:18:52 | Richard Brown | Yeah. So that's great Vishesh. Let me jump in there because I was not, I usually like to be precise with my language, and I was not being precise. When I say responsibility what I meant was, for us to develop and implement and then release. I'm not sure that that's the case. And I should caveat, this is all my opinion. I don't run MakerDAO. You'd be amazed at the lack of control I have over the greater world. |
| # / 00:19:17 | Richard Brown | But, I think that in my opinion, the community can make this tool, these tools and decide for themselves how they want them to be organized, and then we can help them to do it with like, whether that's with grants, whether that's with additional developer insights, whether that's having conversations in these calls and we can foster that activity. |
| # / 00:19:37 | Richard Brown | I'm just not sure that we should be the ones to implement it is what I mean. |
| # / 00:19:42 | David Utrobin | There's always something else to say about it is that, there's a lot of large Maker holders that are not the foundation, that it is in their interest to fund these kinds of initiatives as well. So, if there's large Maker holders out there that are willing to give somebody a grant, even privately completely apart from the foundation's involvement, that's a route that the community and Maker stakeholders should consider. |
| # / 00:20:08 | Richard Brown | That's a good point David. And it's also part of the models that I think that was part of the initial conception of this system is that, these large holders and these funds and these family practice offices would get together, either get together and/or independently begin to act as these large delegates, right? And I'm not sure that we've seen that yet and I'm not sure that I know why. |
| # / 00:20:36 | Richard Brown | Maybe it's custodial solutions or our roadblock here or its lack of engagement, or whether some of these larger investors are viewing it purely from an investment perspective, or less than a stakeholder perspective. These are questions that I don't think we have answers to, but yeah. |
| # / 00:20:55 | Richard Brown | So, yeah, I guess I'm just adding more questions to the questions, but I think delegation is a great idea and Vishesh if you have ideas, about how the community could potentially organize around that or what those implementations would look like, I think we should continue that discussion, but I don't have those answers. |
| # / 00:21:17 | Vishesh Choudhry | Okay. So perhaps we'll talk another time. |
| # / 00:21:20 | Cyrus Younessi | Well, I think additionally, we're also not voting on a lot of different things, right? We're only voting on the stability fee, which is basically a proxy for the DSR at this point. And we're not really looking at like debt ceilings or liquidation ratios and new collateral types and all that kind of stuff. I imagine engagement will pick up significantly once there's kind of more on the line. |
| # / 00:21:48 | Richard Brown | That's a great point. Like it's early days and we … there's … yeah. So let me organize my thoughts here. So, here's the issues that ultimately we, there's this impression that we're not seeing it as enough Maker getting staked as we would like. We're not seeing enough individual voters turn out as we'd like. But this is a systemic issue with crypto in general. |
| # / 00:22:12 | Richard Brown | If you look at any of the large projects, I think that we're actually coming out ahead as far as voter engagement goes generally, unless there's some outliers that are different, that are doing things very well and we should probably be taking a close look at how they're doing it very well. I hear nothing but good things about Decred. And there's a couple other examples of that, but … |
| # / 00:22:33 | Richard Brown | Like Cyrus said, we're playing with a single lever that moves a single number in a system that is relatively new, that shows a drift from the peg for a couple months over its year long, over a year lifecycle, and now that problem is over. So what is the incentive for all of these people to come out of the woodwork and vote on something that doesn't need to be voted on from their perspective? |
| # / 00:22:58 | Richard Brown | So I think that's, these are early days. If we keep on iterating, as long as the engagement numbers go up over month, over month and we're on the right path, I think. And whether people are apathetic right now, because we lack of delegate solution or whether they're apathetic right now, because they just don't see any point to voting at all because the peg is stable, people are already voting for the thing that they wanted to vote for, so what's the point? |
| # / 00:23:25 | Richard Brown | There's questions we still need to answer. Okay. I'm going to actually got to, it's nice that people actually care and ask questions about governance, but I want to be cognizant of time, because I know that Cyrus has a nuclear bomb of science he's going to drop on us. No pressure, Cyrus. So I'm going to hand that off to you and then we can start talking about risk. |
| # / 00:23:49 | Cyrus Younessi | Okay, cool. Yeah, we're going to start with collateral risk today and then do the monetary policy stuff afterwards. I've got at slide presentation. Just find out how to share this. Okay, good. Okay. Yeah, you guys can see this? |
| # / 00:24:22 | Richard Brown | Yeah, looks good. |
| # / 00:24:23 | Cyrus Younessi | Okay, cool. Okay, so- |
| # / 00:24:27 | Richard Brown | You might want to expand on that. |
| # / 00:24:30 | Cyrus Younessi | Oh, uh, sure. |
| # / 00:24:31 | Richard Brown | How's that? |
| # / 00:24:37 | Cyrus Younessi | Did that work? |
| # / 00:24:39 | David Utrobin | It's like a home theater now. |
| # / 00:24:40 | Richard Brown | Yeah. |
| # / 00:24:47 | Cyrus Younessi | Okay. So we're going to sort of take a step back and go through the collateral risk lifecycle from start to finish. So it should hopefully be a little bit less erratic and less frantic. This is kind of the general process of how this would work. Start off with the onboarding collateral, which is basically how do we ingest new collateral types into the system, followed by a due diligence or fundamental evaluation, from which the results get put into the quantitative model, along with all the trading and market data. |
| # / 00:25:36 | Cyrus Younessi | And how that passes to governance, how we continue an upkeep that governance over time. And then, I think after we go through those, we can start talking about actual … I think by that point we'll be at the stage, where we're going to start talking about specific collateral types and how we're going to be evaluating them. So this is obviously not an outline just for today, because there's a lot of stuff here. I think today we're just going to talk about the due diligence section. |
| # / 00:26:10 | Cyrus Younessi | And the reason I'm skipping onboarding collateral is because, there is a whole application process documentation being released very soon, and I don't want to jump the gun on that. So sometime in the next couple of weeks, I think that should be coming out and then we can talk about it. In terms of schedule, I'm just tentatively guessing. We can probably run through due diligence today, maybe not all the way. Maybe there is a bit of leftover, and then we can talk about the meaty quant stuff for two, three, four weeks. Not really sure how long that's going to take. |
| # / 00:26:50 | Cyrus Younessi | And then we can kind of take it from there and see how the rest of this stuff goes. Final point, this is the one diagram in today's presentation, so soak it up because these things take a long time. Okay. So just really quickly on the onboarding application, I'm not going to spend too much time on this because as I said, it's coming out soon. But essentially what this consists of is a technical audit, which is essentially an evaluation of the contract code. |
| # / 00:27:26 | Cyrus Younessi | Looking at the third party audit reports, looking at what kind of special adaptors, whatever's needed to ingest the collateral into the protocol and so on? There's a risk audit where we kind of basically just try to figure out if this is a complete scam or not, or is something even worth considering? Essentially we can envision a situation where we have hundreds of potential collateral types in the queue, and we need to find a way to filter and prioritize these collateral types. |
| # / 00:28:10 | Cyrus Younessi | So, just kind of a quick rough and dirty risk audit. In some cases we might need to do counterparty audit, just kind of looking especially for security tokens, looking at who the issuer is and that kind of stuff. As I mentioned, this is just kind of a filtration mechanism, where we can prioritize the incoming collaterals so that we can just start tackling them one by one. And there are also some interesting governance implications here, in the sense that how exactly, does potential collateral even get selected for evaluation. |
| # / 00:29:01 | Cyrus Younessi | Somehow, these decentralized coins are going to have to present themselves to the community. I don't know if that's going to happen on reddit or on the calls or whatever. But I think there is a concerted effort to not have everything just flow through the Maker Foundation. And then we just say, “Here, this is kind of … This is what we've been working on in the background.” So yeah, we're working on ways to make that as decentralized as possible. |
| # / 00:29:34 | Cyrus Younessi | So after the onboarding, we can start looking at due diligence. And this is really the meat of today's presentation. Talk a little about the goals, how the analysis works and a scoring framework. Okay. So our main goals for looking at a qualitative assessment for collateral first is ,we want to find risks that are not immediately present in the trading history. So, we want to look for aspects of the collateral that might contribute to a CDP liquidation that we reveal through a deep dive into the organization. And I'll go through some examples of these in a bit. |
| # / 00:30:33 | Cyrus Younessi | And then secondly, after we conduct this assessment, we want to be able to convert that into some sort of numerical score that we can actually input into the risk models. Makes sense. Cool. |
| # / 00:30:53 | David Utrobin | I have a quick question. |
| # / 00:30:56 | Cyrus Younessi | Yeah. |
| # / 00:30:57 | Unknown | Isn't the risk really more around an under collateralization event? Vishesh said on the call a couple of weeks ago that, liquidations aren't necessarily bad things, it's keeps the steady flow of money in and out. |
| # / 00:31:09 | David Utrobin | You still have to determine like the risk parameters though. |
| # / 00:31:16 | Cyrus Younessi | Right. |
| # / 00:31:17 | Vishesh Choudhry | [inaudible 00:31:17]. Liquidations are not necessarily a bad thing. Bad is relative to the specific parameter you're talking about. So, if you're talking about revenues, yes, liquidations produce penalties which produces fees, but also if you're talking about liquidations above a certain size or a certain frequency, then you get into an additional concern when ETH is 3 or 5% of the Maker, like when three or 5% of ETH is locked up in Maker. |
| # / 00:31:47 | Vishesh Choudhry | Then you get into a different level of concern, which is, “Okay, I'm I incurring costs like slippage and things like that and selling that?” So, it's a different level of concern. Sorry. |
| # / 00:32:08 | Cyrus Younessi | Okay. Yeah, and to answer that question, I think, yeah, no. It should be clear that there are certain aspects of collateral type that could lead to very bad situation, for MakerDAO that we'd want to evaluate from qualitative perspective. And, I'll get into examples in a bit. Okay. So kind of mentioned, kind of touched upon this last week, a process for how to actually begin evaluation collateral. My first step is always a classification. Broadly we have bearer assets and registered assets or security tokens. |
| # / 00:32:54 | Cyrus Younessi | bearer assets obviously being much more amenable to our protocol, as you guys all know recourse is not really an issue there. There's essentially no counterparty risk. And big question I get a lot is obviously around wrapped tokens, wrapped Bitcoin and all the others that might come down the line. And essentially wrapped tokens are a form of registered or security tokens, because there is that counterparty risk which has to be evaluated. So essentially, that counterparty risk is an example of the recourse analysis that needs to be done. |
| # / 00:33:45 | Cyrus Younessi | So, what is the process behind actually claiming real Bitcoin from wrapped Bitcoin? Now, it needs to be very carefully studied. You need to study the issuer. In this case it'd be Bitco, but whoever it might be, we're going to have to look at all these in detail. And it should be obvious that the downside here is if for whatever reason, you cannot claim your underlying asset from the wrapped token then well, essentially you won't have anybody bidding on that collateral. |
| # / 00:34:26 | Cyrus Younessi | The collateral price will fall and as it goes to liquidation, again, there'll probably be nobody purchasing that collateral except for opportunistic kind of distressed asset investors, which might look to scoop up a scoop up these assets for pennies on the dollar, and try to try to make money through the legal system. |
| # / 00:34:52 | David Utrobin | So quick question, Cyrus. Are these kind of like the two main classifications, bearer assets versus registered assets and STOs, or are these kinds of like broad strokes? I mean, I know there are types of tokens in within bearer assets and like types of securities within registered. |
| # / 00:35:09 | Cyrus Younessi | Yeah, I don't think there's anything that's … I don't think there's anything that falls outside of these two, these two divisions. And really the importance of this division is mainly that, we don't have to do too much of a recourse analysis. No one in single collateral has ever thought about the liquidation process for ETH, because everybody knows that as soon as you purchased ETH through the auction, it just directly gets transferred to your wallet. In certain situations that may not be the case and we need to be cognizant of that. |
| # / 00:35:44 | Cyrus Younessi | And then now obviously it'd be nice if we could just stick only with bearer assets, right? Everybody wants to just have ETH only or whatever. But, obviously there's an extremely high correlation between lots of the current ERC20 tokens. Okay. So these are some of the, these are some of the bearer assets and security tokens. But, I think we're going to need to go through a careful … we're going to need to develop a careful understanding of these different tokens. |
| # / 00:36:23 | Cyrus Younessi | Maybe not today, but as we, maybe we can kind of do them one by one as we go through actual collateral types, but pretty much these are the main ones as far as I see them, where you have the base layer token such as ETH or Bitcoin, whatever. And these base layer tokens or are typically subdivided into money candidates or non-money candidates. So obviously Bitcoin wants to be this kind of global cash. ETH potentially also wants to do the same thing, but then you also have base layer tokens that are, kind of self-described themselves as not competing for that market. |
| # / 00:37:06 | Cyrus Younessi | And that becomes important as part of the analysis. Utility tokens, I think everybody here generally knows what utility tokens, are a special token that you'd use for particular service. Work tokens, give you the right to do some work and provide some service and you get paid a certain amount of fees from it. Governance tokens such as MKR for one, staking tokens such as OMG and potentially others. I actually don't know of a lot of good Ethereum based staking tokens. |
| # / 00:37:51 | David Utrobin | So really quick Cyrus, so these designations, these classifications, they're separate from the previous slide of bearer assets versus registered assets and securities in that, these can kind of overlap between those two types, right? Because you might have like a, or like a staking token that's off chain that's a security token really? |
| # / 00:38:14 | Cyrus Younessi | Right. Yeah, you can. So, when it comes to the security tokens, really all you're doing is making a counterparty risk adjustment to your risk model. So if you had Ether versus wrapped Ether that was issued by some centralized custodian, then you'd to just factor in the counterparty risk, which is obviously a big question on how to actually do that. We'll get to it soon. So, okay, so after … So at this point you should have a kind of a good understanding of what this token actually looks like, what it does. |
| # / 00:38:54 | Cyrus Younessi | And then you want to probably dig into the actual project itself. You're going to want to look at the team, who they are, what they've done. You're going to want to look at the technology, the competency of the developers, look at the roadmap, what kind of adoption have they come up with. You're going to want to look at their community, the business potential to the total addressable market, who's their competition, what's their business strategy. |
| # / 00:39:25 | Cyrus Younessi | All these aspects need to be evaluated by risk teams, to get a good understanding of what this project actually looks like. We're also going to … I think it's also important to look at the token distribution, and potentially if there was an ICO, what their funding was like, what's their legal structure, what exchanges are they on, potentially who were there early investors, what were the vesting periods for those investors? I think it's fairly clear that the first few steps of looking at a collateral type is, understanding what it is, who's behind it, what does it look like, right? |
| # / 00:40:16 | Cyrus Younessi | Just get a good overall picture. Potentially in some cases, we could also go ahead and build out a valuation model. This would be ideal, although there is a lot of overhead associated with this kind of stuff. But just to give you an example of how important it is, there were legitimate projects at the height of the bull market that were worth over a billion dollars, right? And there was nothing wrong with the team or the product or whatever. It's just insanely overvalued, right. |
| # / 00:40:52 | Cyrus Younessi | And from the MakerDAO perspective, you have to ask yourself, would you really accept in extremely overvalued asset as collateral no matter how legitimate the project is, right? Because you're weighing in your mind, what is the probability that the market snaps back to reality quite quickly? And actually in the most recent bear market, I would say that market participants generally got fairly lucky in the sense that, assets kind of corrected somewhat slowly over a whole year, to go all year for it to hit rock bottom where in the future may not be so lucky. |
| # / 00:41:38 | Cyrus Younessi | Because in some cases it would just take a couple large investors to just completely crush a coin instantly. But if you want to do valuation models, there's some funky ways to do it with these, with the bearer assets. For example, if you've ever looked at what a Bitcoin valuation model looks like, with security tokens it's fairly easy. You can just borrow from the traditional world of how a lot of these assets are evaluated. So at this point, we should have a understanding of what this token is, what it's worth and all that. |
| # / 00:42:22 | Cyrus Younessi | And this is where we start to dig into the risks associated with these assets. So all assets have risks to them, right? And it's our job to reveal what those risks are. Now with traditional assets like equities, bonds, whatever, I think the risks are not fairly understood but extremely well understood. Not really the case with crypto assets. I think a lot of people don't necessarily have the right intuition for what gives these assets value in the first place, and so they are somewhat unaware of the risks associated with them as well. |
| # / 00:43:09 | Cyrus Younessi | So for example, money tokens such as Bitcoin or Ether, the risks lie in its ability to maintain its status as a store of value. And store value is very, is a very mean based value proposition. And as soon as that thing gains momentum one way or another, it really just runs with it. Simple example with Ether in 2017, when ETH was the only money acceptable for ICOs. Everybody needed it to get into these projects. And then when ICOs died, people realized that they didn't need this money any more. |
| # / 00:43:57 | Cyrus Younessi | So that's a risk that we have to be on the lookout for money tokens. For utility tokens, we have what are known as velocity risks, where people just kind of trade, where people don't hold onto the asset for any meaningful amount of time. They just get in and out of them, like they would say arcade tokens. I would say this is one of the categories of tokens that took longest for the general investors base, to really understand the deficiencies of. |
| # / 00:44:32 | Cyrus Younessi | I think as we start to go through specific collateral types that are utility tokens, we can pick apart exactly how serious of an issue it is. Because in some respects utility tokens, to be honest, don't really have much of a value proposition at all. Yet there are still many that persist in the wild, and so we can start to discuss why that is the case. And in particular, branding tends to be a big driver in my mind. A lot of these utility tokens have insanely good branding. |
| # / 00:45:10 | Cyrus Younessi | They're housed by very popular parent organizations where they have a lot of, they have extremely loyal user base who is willing to trade in their bitcoin or Ether for that utility token. But yeah, velocity risk is definitely something we need to discuss. Work tokens are susceptible potentially to being forked out and replaced with a more native based token. Governance tokens suffer from value capture. |
| # / 00:45:42 | Cyrus Younessi | People who own governance rights over certain project in some cases, they have the right to govern it, but they don't have the right towards any of the value that is created by good governance. Staking tokens have a lot of risks in my mind. And then for security tokens you have the risk of the underlying asset, as well as the counterparty risk for the wrapped portion of it. So the big question is how does, how the market treat these going forward? |
| # / 00:46:16 | Cyrus Younessi | Historically, it's been extremely favorable, extremely favorable almost shockingly so. I think you can remember when the market started to turn in last year, there were quite a few people, calling for just the absolute annihilation of a lot of these token categories and it just didn't play out quite as bad. But we have to be aware of how this might occur in the future. |
| # / 00:46:48 | David Utrobin | So, Cyrus, I have a quick question. So is this slide to do with like the estimated shortfall calculation, or does that already happened like during the quantitative side of things, like after all the qualitative stuff? Is this just like the- |
| # / 00:47:03 | Cyrus Younessi | It's a fact risk? |
| # / 00:47:04 | Vishesh Choudhry | Yeah. So these are like the lists of the main risks per like token category, right? |
| # / 00:47:10 | Cyrus Younessi | Right. And then what we're going to do is we're going to essentially take this entire qualitative evaluation, everything from the valuation model to the classification, to all the risks and we're going to basically, input all this information into kind of a model. And that's going to spit out a score that will then go into the more broader quantitative model that calculates aspects like the last distribution. |
| # / 00:47:46 | David Utrobin | Got it. And I noticed in the chart that there's like, there's just talk about how some of these risks overlap, like for example, governance tokens can also like suffer from like fork risks for example. And so just to be clear, like these aren't the singular risks that like each of these … this is just probably like the main one, right? |
| # / 00:48:05 | Cyrus Younessi | Yeah. These are the main ones. |
| # / 00:48:07 | David Utrobin | Each of these ones? |
| # / 00:48:09 | Cyrus Younessi | The first thing is hopefully we would have had some valuation model for these in the first days, which is essentially presuming that these tokens are worth something in the first place. After you have an idea for what it's worth, then you can look at what are the risks to it not being worth what you think it might be. So for example, some of these risks are quite low probability. And for example, I think the fork risk of work tokens is quite low personally. But it will all get factored in. Okay. |
| # / 00:48:59 | Cyrus Younessi | So continuing with the risk analysis, I think that, the broader theme of risk analysis is adversarial mindset. Really anyone, any risk team or anyone in the community, who is really excited to see any collateral be put into the system, has to approach it from the mindset of, what is the worst thing that can happen? How bad can this get? How bad can MakerDAO holders suffer? And this can happen in a lot of different ways. Recourse analysis, very simple example, and I've used this a lot but what is the chance that you don't recover your dollars from tether? |
| # / 00:49:59 | Cyrus Younessi | Exchange, flip these things if the tokens only, there's few exchanges and they start to delist it, that can be extremely bad for the token price. Poor operational management or leadership, we've seen several projects run out of money when they really shouldn't. I liken this to having good trader's instinct. In many cases traders don't really care about what certain signals or certain TA or certain models spit out. They just know that this is a terrible idea through experience and due diligence, and this is really where that kind of stuff comes in. |
| # / 00:50:45 | David Utrobin | How does that fit into like such a kind of rigorous analysis … that seems like a bullet point that's like shooting from the hip, like traders instincts? |
| # / 00:50:54 | Cyrus Younessi | Well, get used to it cause in crypto that's really a lot of … I mean again, that's not like the entirety of the model, but that is just kind of a fact check. |
| # / 00:51:07 | David Utrobin | Is that like a maybe a different way of saying like this is the tail risk of like maybe … I just can't distinguish between like whether it's a feeling or if it's something that's just like, “Okay, I feel like there might be a tail risk that we're all missing here.” |
| # / 00:51:24 | Cyrus Younessi | Right? So first, you have to recognize that something is wrong. Something's a risk and then you can at least go trying to the attempt to quantify that risk and how bad it could get in, and what's the probability of it recurring. But, just even recognizing some of these risks is definitely it's instinctual, right? If you think about before the first exchange hack ever happened, I'm sure people didn't really think that exchange hacks were legitimate possibility. |
| # / 00:51:55 | Cyrus Younessi | Before certain projects blew through $100 million worth of raised capital, I'm sure people didn't think that funding concerns was going to lead to the death of their project. So a lot of these is just kind of built up through experience and due diligence. Okay. So- |
| # / 00:52:15 | Richard Brown | So, we'll [inaudible 00:52:16] the rest of the call. |
| # / 00:52:18 | Cyrus Younessi | We're I'm almost done, I think. Let me try to speed through. Or if we don't actually, we can just continue next week. I don't know. |
| # / 00:52:28 | David Utrobin | You totally lied about that one slide. You were like, “This is the only slide I'm going to show you guys.” |
| # / 00:52:33 | Cyrus Younessi | So it's the only side with it with diagram. |
| # / 00:52:39 | David Utrobin | Yeah, I misheard you. Right, cool. |
| # / 00:52:44 | Cyrus Younessi | So essentially take everything we've just discussed and you're going to want to spit out a risk rating on. This is basically where people call MakerDAO the TCR of crypto or similar to some sort of rating agency, because the grade or the score that we assign to collateral types is what we use to, evaluate the more quantitative risk parameters such as the stability fee. Details of this framework, I guess we can go over another time. |
| # / 00:53:14 | Richard Brown | Are you actually considering like five stars out of 10 kind of rating system for crypto types or is it just going to be, “Here's the package of risk parameters and that's your rating?” |
| # / 00:53:27 | Cyrus Younessi | It doesn't … So the actual risk parameters come after the rating, not before, but essentially the risk rating doesn't really matter, just needs to be something convenient and easy to understand, and needs to be able to just be standardized so they can compare different assets against each other. They don't necessarily map to a direct stability fee, like seven out of 10 is 5% or whatever, although you could in some ways. So let me just do a couple of quick examples. |
| # / 00:54:03 | Cyrus Younessi | These are just like a one-page example, and then we're working on much more in depth evaluations behind the scenes. So for example for Ether, obviously we're going to have to redo Ether's collateral type for MCD, and we might as well kind of apply this new framework and try to come up with the most optimal risk rating and risk parameters. So ETH would come down the pipeline somehow and then we do a technical audit. Ethereum has been audited a couple of times. |
| # / 00:54:37 | Cyrus Younessi | We know that it needs a wrapper, because doesn't work in its native form. A financial audit would probably look at supply, it's inflation, the team, just very high level. And then we'll say, “Okay, Ether just looks good enough to be evaluated more in depth.” So we'll kind of look at, what Ether the token actually looks like. It's base player token, we know what it does. We evaluate it from a fundamental perspective. Look at some of the risks, and then we give it a score, right? |
| # / 00:55:08 | Cyrus Younessi | And that's basically what the end of the qualitative analysis would be. And then the score is what gets transferred to the quantitative model, which we'll talk about next week. Another quick example of utility token, same thing. And then the last thing I want to emphasize is just, is the philosophy here. I think we're going to be seeing a lot of conservatism, common sense. I don't want to rush to do anything we're not comfortable with. Pretty standard stuff, but it's good to give people a reminder that we are, I think the goal should be to take this nice and easy and do it properly. |
| # / 00:56:01 | Cyrus Younessi | Okay. So if you guys from the community want to help out, there's going to be a ton of work to do on this end, verifying data that comes in through the applications, working on diligence reports and all that. Please reach out to me. Right now we're currently working on a Ether risk model, doing both qualitative and a quantitative analysis. And could always use extra help and just reach out. That's it. |
| # / 00:56:35 | David Utrobin | Those was a really solid presentation, man. |
| # / 00:56:38 | Cyrus Younessi | Thanks. Okay. Any questions before we jump [crosstalk 00:56:45]? |
| # / 00:56:51 | Richard | David, you said you were collecting questions from the chat? |
| # / 00:56:54 | David Utrobin | Yeah. So, there was just one question a bit earlier. So I covered the governance token also being at risk of being forked. Then there was the other question of, if there is going to be shitcoin voted in, how do we vote it out? |
| # / 00:57:14 | Cyrus Younessi | Hopefully it doesn't get voted in the first place, but if we need to … so, this'll be part of the maintenance section, which is like a few little bit down the line, but essentially we have to come up with ways of how to delist coins, right? One method would be to just lower the debt ceiling. There's a couple of different ways to do it with varying degrees of urgency. |
| # / 00:57:38 | Richard Brown | Vishesh, do you have a question? |
| # / 00:57:54 | Vishesh Choudhry | Yeah. So, is this kind of like an illustrative question of once the initial modeling is done in the initial parameters are selected, what's kind of the view of the big picture process, for those maintained over time or adjusting them if there are, and how does that fit in with the governance process? |
| # / 00:58:13 | Cyrus Younessi | Yeah, I think every few months it's worth it to do a reevaluation of the collateral assets, see how the profile has changed. For example, if Ethereum has some major scaling breakthrough, that works, everybody loves it, then actually it's lower risk than it was before. So yeah, you can definitely track these and adjust the scoring, adjust the ratings over time. Companies have their risk ratings adjusted over time as well, so it's the same concept. |
| # / 00:59:03 | David Utrobin | How often do you expect that collateral need to be reconsidered and go through the whole motion again? Like, how often should you look at it whenever you see like [crosstalk 00:59:15]? |
| # / 00:59:16 | Cyrus Younessi | Probably preset cycle, maybe every couple months. But then we should allow for emergency situations, where we have to change something on the fly. I think that again goes more into the maintenance and operational aspect of the collaterals. Okay, cool. Let's jump into the monetary policy then. |
| # / 00:59:39 | Vishesh Choudhry | Let me share my screen. |
| # / 00:59:49 | Cyrus Younessi | Yeah, for sure. |
| # / 00:59:54 | Vishesh Choudhry | All right. So can everyone see everything properly? |
| # / 00:59:59 | David Utrobin | Yes. |
| # / 01:00:00 | Vishesh Choudhry | Okay. All right. So, just to jump in, things for the past week plus as has been noted were fairly stable. What's interesting is in the last just 24 hours, ETH jumped up a little bit and people levered up a little bit more on that, which depressed the DAI price a little bit. And so I think this is a topic to consider for the governance votes, which is, you can see over time since the 19.5% stability fee increase, there was this dwindling of peaks and valleys to a little bit more of a stable level. |
| # / 01:00:48 | Vishesh Choudhry | But since the 16.5% increase, I think the valleys have started to pick up again a little bit. And so I think we're starting to see some increased variability in DAI price, particularly the amount of DAI that trades below a dollar. And so I'll dive into that a little bit more in a different graph, it's easier to see. But trading volume has shrank over the past two weeks and then spiked up again last night. So that's just something to keep an eye on. |
| # / 01:01:22 | Vishesh Choudhry | I think those spikes, as we've talked about tend to coincide with one of two things. Either a lot of ETH being sold for DAI when people are cashing out on leverage or far more often people buying, no sorry, selling DAI in order to leverage further up on ETH. And based on what's been going on with some of the secondary lending rates in the past 24 hours or so, I think that was more of the latter where people were levering up in anticipation of a bit of a jump in ETH price, which actually self-fulfilled within a 24 hour timescale. |
| # / 01:02:06 | Vishesh Choudhry | So, yeah, to talk a little bit about the relationship between DAI and ETH price. So as we've noted since mid-April, start of May, the trend in DAI price and ETH price has kind of tracked. I think we're starting to see a little bit of divergence from that, but it's a little bit too soon to say. So in my analysis, the view would be that that tracking was actually a reversal of a previous trend, where people were levering heavily up on DAI and just kind of continually accumulating the amount of ETH leverage that they bought with their DAI. |
| # / 01:02:51 | Vishesh Choudhry | And so they were accruing more and more debt, but at the same time they were levering further up on ETH. Then at the start of May, ETH price had started to run up, some people cashed out and delevered a bit. As ETH price came down into kind of the 230-ish range, we started to see people get a little bit more excited and lever up further. And so that's kind of like the narrative there is, initially there was this huge levering up, then unwinding a bit and now people are starting to pick back up again a little. To talk about what that means for supplies. |
| # / 01:03:27 | Vishesh Choudhry | So, supply had been consistently going down, but in the past couple of days that has ticked back up. Even this kind of cuts off with a 24 hour delay, it's now around 83 million. So, essentially as the stability fee is continually increasing from 11.5 to 19.5%, there was a pretty steady drop in supply. That stayed stagnant all through the 17.5% decrease, but then around the 16.5% decrease, that started to tick back up a little bit. And has now come up a little bit further, but that's probably a short term effect from what's been going on with ETH price. |
| # / 01:04:09 | Vishesh Choudhry | So in general, it's … Again Rich asked me this question all the time, how do we decide what is confidence in terms of a trend versus a short term movement? Like I said, I think time is what generally tells us that. So, as we look at more of these like long run averages I think will get clearer pictures. But, my initial sense is with the 17.5% decrease, there wasn't a huge effect. But with the 16.5% decrease, there has been a little bit of a ticking up in the DIA supply and the amount of leverage that people have been seeking. |
| # / 01:04:51 | Vishesh Choudhry | Now, that leverage that they've been seeking is not just been on Maker. It's also been on these secondary lending platforms, which is a topic that we've kind of touched on. So, I'll circle back to that in a minute. In terms of circulation of debt, this trend has been very clear as opposed to others where it's harder to … Sorry all the data just to reload it as we were talking. So I have to reset afresh. But in terms of the circulation of debt, the trend was fairly clear. |
| # / 01:05:23 | Vishesh Choudhry | There wasn't a huge impact on the amount of old debt being paid back or the amount of new debt being taken up, as the stability fee was increasing all the way up to 19.5%. Once we hit that 19.5% mark, there started to be a legitimate impact where that debt started to come down significantly. And so in that sense, the circulation of debt started to increase and debt was getting younger. What is interesting is that also coincided with this increased traction on secondary lending platforms. So part of it, I think it's good to consider the effects of refinancing. |
| # / 01:06:05 | Vishesh Choudhry | But, at the same time as the stability fee was dropped again to 16.5%, we started to see this trend slow down. And this is despite the fact that in the same time range dYdX and Compound v2 we're sort of getting this renewed traction. So I think that speaks volumes to me in the sense that, 16.5% seems to be a meaningful level, and might be either a balanced point or slightly too low, very hard to say the difference between those two. But, I think there are noticeable effects where it was different than 17.5%. |
| # / 01:06:45 | Vishesh Choudhry | So just to touch on this briefly, I switched up this graph a little bit since the last time we saw it, but there's been kind of this long-term downtrend a little bit, in the amount, the percentage of new debt of all the debt that has drawn out, what percent is drawn out on new CDPs. So, although, and I don't have this graph to show you here today, the amount of unique users that have been interacting with CDPs has continually increased, which is a good metric in terms of usership and traction. |
| # / 01:07:18 | Vishesh Choudhry | The percentage of debt that is being drawn out from new CDPs versus existing ones has actually periodically held steady, and then these peaks have decreased a little bit over time, and the valleys have exaggerated. So what that means to me is that over time, a larger portion of the debt is being drawn out from existing CDPs, even though there may be more a new individual users interacting with the contract. Collateralization, so was … collateralization was one of the more … Sorry, that keeps happening. One of the more interesting metrics in the past couple of days. |
| # / 01:07:56 | Vishesh Choudhry | So, we saw this huge run up in collateralization percentage after April. I think that was a hesitancy for people to make significant changes to their collateral despite the rise in ETH price. So, essentially as ETH popped up, people were scared that it might come back down, and so they didn't make significant changes to their CDPs. As ETH sustained those levels, and it actually started to come down a little bit, people started to clear out some of their debt, and also clear out some of their collateral. |
| # / 01:08:30 | Vishesh Choudhry | And so in that sense, there was overall I think a deleveraging behavior. What's interesting is that again it's ticked up a little bit, in the past 24 hours and it's actually even higher than 470 now. So, again, I think it's just this valley has started to be hit around 16.5%. In terms of the attractiveness of leverage even at like ETH 230 to 250 price range. To just touch on DAI price real quick. So, kind of reflecting what I've been saying, there's been this widened spread, in terms of DAI prices and where DAI is created. |
| # / 01:09:12 | Vishesh Choudhry | You can see that also … you can also see that in the stable coin graph, which shows kind of over time a significant amount of volume since coming up to peg trading below a dollar. Now, I think there was this huge reaction that, “Oh, once DAI if hit that dollar one, dollar two range everything was okay,” which I think was broadly the sense. But then people started to decrease that stability fee. |
| # / 01:09:47 | Vishesh Choudhry | And then over time, around just the end of May, starting of June, we started to see a slight drift back down below the peg. So I think perhaps these are indicators that 16.5% is slightly too low or maybe just right and this is a short-term effect. There's again, no real way to tell the difference between those two things. So this is some of these secondary lending platform trading volume or activity outstanding volume. |
| # / 01:10:20 | Vishesh Choudhry | So what we see is in the past … I'll just show the full scale, since the launch of these platforms, they've gotten a ton more of the outstanding DAI borrow volume on the secondary lending platforms. But in the past couple of weeks, that's actually held fairly steady. And then in the past 24 hours, that actually came down a little bit. But it was both borrow and supply that came down, even though the overall DAI supply ticked up. |
| # / 01:10:50 | Vishesh Choudhry | So that is to me, some people saying, “Hey, it's potentially more attractive to come back to Maker, from some of the refinancing that we had done in the past, because some of the rates on, for example dYdX have shot up rather exorbitantly in the past 24, 48 hours. And dYdX for example, has maintained significantly higher rates, Compound slightly lower. And so, that's something that we'll go into probably more next week or I'll do a little bit of an overview on lending. |
| # / 01:11:22 | Vishesh Choudhry | But in general, I think some of that refinancing has come back to Maker is the answer. And I think that was the last point that I wanted to make on the graphs, open to any questions. |
| # / 01:11:38 | Richard Brown | Have you seen any evidence of people being more willing to buy cheap DAI in light of this empirical evidence that it does return to the peg eventually? |
| # / 01:11:49 | Vishesh Choudhry | I think there is a … it's very difficult to measure people's willingness to buy DAI below a dollar. But I think what we've seen is the same kinds of stressors that in the past had caused DAI price to drift far below the peg have now had a lessened impact on DAI price. And I think that is most likely because people are more willing to arbitrage DAI when it's trading below a dollar. You've seen in the long run more confidence that it will return to the peg and they'll have a chance to make a profit. |
| # / 01:12:21 | Cyrus Younessi | I think it's important to distinguish that … Just looking at this chart, the vast majority of any DAI trading below the dollar is on Oasis, right? And people don't go to Oasis to buy cheap DAI. This is a result of market … This is a result of retail speculators jumping the bid-ask to buy yourself ETH, which results in a cheap DAI price for the market makers who are just spreading the ETH against the Coinbase price. |
| # / 01:12:54 | Cyrus Younessi | The DAI dollar price is not trading at like … is not traded at like 98 cents at all really, if you were to just go straight from the direct exchange rate. |
| # / 01:13:04 | David Utrobin | But isn't Uniswap also a liquidity pool for Kyber for example? |
| # / 01:13:11 | Cyrus Younessi | Yeah, but even Kyber is not very liquid. All the bars below a dollar are 80% oasis. And the other 10% [crosstalk 01:13:21] is AirSwap, which is again just people trading ETH/DAI and just paying the bid-ask spread. And that results in a cheap DAI price. |
| # / 01:13:29 | Vishesh Choudhry | Yeah. So to be clear, the upper tail 100%, if you look at like Coinbase for example, that price, the ETH/DAI price does tend to stay centered closer around a dollar. Actually, this is somewhat uncharacteristic. I think this is an effect of some of the frenzy that happened in the last 24 hours, where you see more of this Oasis long tail towards the lower prices. Usually it's actually Uniswap that tends to dominate the lower tail prices, primarily due to slippage I'm assuming. |
| # / 01:14:04 | Matthew Rabinowitz | But also to some extent because people are going to Uniswap because they need that immediate liquidity and they are willing to take a bit of a haircut. And so, to your point Cyrus, if I understand your point correctly, yes, most of the mainstream trading sources do tend to sit around a dollar. The lower tail of dye prices is usually the one that's in the liquidity crunch and they need to get quick access, primarily because they're trying to refinance or they're trying to lever up and they don't want to miss that jump in the ETH price. That's my sense. |
| # / 01:14:41 | Cyrus Younessi | Really it's just because there is a little bit of an uptick in the market and speculators who hold DAI need to pay a little bit of a premium to get long ETH, given their expectations. |
| # / 01:14:55 | Matthew Rabinowitz | Yes, that time preference that causes that, I think. |
| # / 01:14:58 | Richard Brown | All right, well speaking about time, we're running out of it and Matthew got bumped last week. So can we move on to the weekly narrative and see where we're at after that? Matthew, do you want to take the lead? |
| # / 01:15:12 | Matthew Rabinowitz | Vishesh, are you done? |
| # / 01:15:15 | Vishesh Choudhry | Yes. |
| # / 01:15:18 | Matthew Rabinowitz | So basically I guess last week when I wrote this narrative, and Cyrus of this treads into too much of your turf, just we can punt it also. But it was basically [crosstalk 01:15:30] going through the math of basically what would this stability fee look like in a multiple collateral world, where we use the stability fee in the single collateral world to just purely reference that the part that goes to the buyer in the burn of a Maker token. |
| # / 01:15:34 | Matthew Rabinowitz | But in a multi-collateral world, where, we a stability fee is that culmination, as I understand it, of the DSR, plus whatever the Oracle fees are going to be, plus whatever the risk team will be, plus that call it, risk premium of the collateral package. I don't want to call it a collateral type, but really the package, because it's not only if it's ETH, it's ETH with whatever collateralization ratio, debt ceiling, liquidation ratio of all of those things that you mentioned before about, they can all be different and can change. |
| # / 01:16:18 | Matthew Rabinowitz | It's that risk premium. Those four things equal what would be the would-be stability fee for a given collateral package, right? So the question becomes, if you really … basically let's assume for a moment that the oracle fees and the risk fees are more or less going to be static. I'll have to discount them and let's just ignore them for the time being. That there's a natural tradeoff between the DSR and the risk premium that gets associated with that collateral package, one that's somewhat inverse, where you're taking the DSR, which is a riskless … excluding technology risk. |
| # / 01:16:57 | Matthew Rabinowitz | It's a riskless return, versus a risky return, then you have to insure, right? So, the game … on my a narratives thing, I put together a presentation and forget the fact that it could have been drawn by my third grader, forgive my artwork, but it stresses the point, which I guess I can try and share on here. Well, it's coming up. Right. Can you see my screen? |
| # / 01:17:27 | David Utrobin | Yeah, I see it. |
| # / 01:17:28 | Matthew Rabinowitz | Right. So, as we assume the way they risk free asset, the percentage of a given allocation to a DSR compared to the value of risk component is inverse. The more we see the further you to the right, the less percentage that goes to a DSR, which somewhat leads to the logic that, as assets become more risky, the percentages flip. It's the same graph but in a different interpretation, right? So the logic of this being just then over time, depending on what assets we put together and the DSR being a common event, it ultimately creates … |
| # / 01:18:14 | Matthew Rabinowitz | And I'm just kind of skipping through part of what I wrote here, in the interest of time, right? The credit ratings that you were mentioning before Cyrus, where just in like in the real capital market world, you've got AAA, AA et cetera, all the way to just junk and risky. What we're going to ultimately see is all that stuff in effect become compressed. But that same type of risk characteristics, where you take an asset and below here where, if you were to just imagine that was the United States treasury, that's in the form of a token. |
| # / 01:18:45 | Matthew Rabinowitz | It should be almost exclusively weighted to the DSR with very, very little VaR component, because the probability of having a default with whatever the risk metrics that you outline on the interim risk team, this is all after the fact of everything that was calculated. One of the components that's going to be needed will be the breakdown between what percentage on a collateral package we're going to say gets allocated to the DSR and what percentage gets allocated to the VaR as kind of our launch percentage. |
| # / 01:19:22 | Cyrus Younessi | Wouldn't they be somewhat independent of each other? You would just add them … you calculate the base DSR, and then you just add in with the additional risk premiums. Why would they be at the expense of one another? |
| # / 01:19:39 | Matthew Rabinowitz | Well they equal the ultimate stability fee, right? So even when you're launching it for the very first time, again, pick the US treasury for a moment. They don't have to be at the expense of another, but they're connected, right? Once we know what that percentage are, the question really becomes after the fact. It's like to me there, this is what my point was. There's like a logical step, logical sequence of having the interim risk team determine the collateralization ratio, liquidation rights, debt ceiling for a given collateral package, right? |
| # / 01:20:15 | Matthew Rabinowitz | And also determining the initial DSR and VaR allocation, right? But thereafter, right? What we ultimately need to compute using algorithms PID or not, is what percentage … excuse me, not what percentage, but the sum of those two things. Because when we change the DSR, it changes it for everybody. So we have to know what the percentage breakdown between the DSR and the VaR for a given collateral type will be and it shouldn't change. That's my point. |
| # / 01:20:53 | Matthew Rabinowitz | The percentage should be known and should be set for a collateral package that's floating out there, very similar to a debt security that's traded in the public market now. You know what you're buying and you know what the percentages are. Then the question is, how much do we move it around as we try to find the optimal methods. But the point I'm trying to get at is, we have a scenario where we take one asset that's not risky, a US treasury and an asset that is let's say riskier, Ether. |
| # / 01:21:28 | Matthew Rabinowitz | And the question becomes if we have price degradation, right? We mint a bunch of DAI off of both of them, how do we determine? We have to find a way to determine and distinguish the signal of which thing moves the DSR and which one moves the VaR. And that percentage needs to be known. I don't know how you would do it without knowing it because after the fact, if people mint DAI off of a treasury, you really don't want to change the VaR for it, because the risk characteristics didn't change. |
| # / 01:22:02 | Matthew Rabinowitz | It really should be the DSR that moves to absorb all of that and conversely for the risky asset that moves the DSR incrementally, but the VaR for it is what needs to take the lion's share of the hit. I don't know how you do that without setting them. |
| # / 01:22:19 | Cyrus Younessi | Right. I think I see what you're saying. You're saying that, if some particular asset ends up generating an enormous amount of DAI, and that one specific asset is responsible for essentially the DAI price being weak, then you're suggesting that a penalty should be applied to that specific collateral asset and not to all assets equally through the DSR. Is that right? |
| # / 01:22:50 | Matthew Rabinowitz | Absolutely in mine, correct. Because the challenge is not when you have two, the challenge is when you've got 60. And some of the mint DAI, some of them … and this is again, forgive the drawing. When you have time zero, some of them are minting, some of them are burning and this is only whatever I wrote on there, three or four. But if you had 400 of them, we have to find a way to distinguish what is the blended DSR needing to move up by four basis points or is it down. And it's a gigantic simultaneous equation, that the ID of this out of the PID comes later. |
| # / 01:23:26 | Matthew Rabinowitz | The P of it is ironically very similar to the circuit theory. It's a gigantic simultaneous equation that's all weighted based upon the DAI that's outstanding per given collateral type. That's like, forgive my ignorance, I don't know how I make the assumption that in an MCD world, we're going to be able to see which collateral minted DAI, and which collateral type is burnt that'll be visible on the Blockchain. I really hope so, because we're going to have to find a way to distinguish between collateral A, B, C and D and who was, not necessarily the bad actor … |
| # / 01:24:01 | Matthew Rabinowitz | But again, if you mint a bunch of DAI that's related to a treasury, the DSR is your lion share of the mob, 99.99%. And with that- |
| # / 01:24:14 | David Utrobin | Sorry. But wouldn't that might have manifested just to a low stability fee overall for the less risky asset? |
| # / 01:24:21 | Matthew Rabinowitz | As it should. That's right, it should. And that's kind of the point, why when you go back over to this point, when you have to somewhat not set it and forget it, but you need to set the percentage between the DSR and the VaR. And when you decide to change it in the future, with four our modify it based upon the conditions on the ground, very much like a credit rating. When you say that Ether, because of its new, doesn't matter, scalability features and there's far less risk. |
| # / 01:24:54 | Matthew Rabinowitz | The only way I understand how we can do that is if you leave the percentages the same, but you drop the collateral is it …? Excuse me, you drop the debt ceiling basically down to zero and you launch another collateral package, same underlying collateral, but with different characteristics and a different breakdown between the DSR and the VaR. It's the only way to basically wind down one asset class in exchange for another, of which an efficient market should show up to basically refinance their debt from a more expensive VaR to a less expensive VaR. |
| # / 01:25:29 | David Utrobin | Why can't you just lower the debt ceiling below whatever the current amount of Diatron against a particular asset. That way over time naturally, since people will be paying [inaudible 01:25:40] on those. |
| # / 01:25:42 | Matthew Rabinowitz | [inaudible 01:25:42] that's pretty [inaudible 01:25:43]. |
| # / 01:25:42 | David Utrobin | Maybe I'm understanding then, yeah. |
| # / 01:25:44 | Matthew Rabinowitz | No, no. That's my point. You're lowering the debt ceiling basically to zero for that asset class, that collateral package, so no new DAI can be minted, but the DAI that just needs- |
| # / 01:25:54 | David Utrobin | Right, it winds down. |
| # / 01:25:54 | Matthew Rabinowitz | … to be out there, it just winds down. It's no different than having a collateral, excuse me, a promotional package, when you get a credit card and the introductory rate is 0% and then later it changes and it winds itself down. [crosstalk 01:26:10] go ahead. |
| # / 01:26:11 | David Utrobin | I was going to say so, I'm trying to understand if there is like a conflict here. So you're suggesting that we have to create basically per collateral package, some sort of like risk based ratio between VaR and DSR usage? |
| # / 01:26:28 | Matthew Rabinowitz | Correct. |
| # / 01:26:28 | David Utrobin | Because that doesn't intuitively make sense to me. |
| # / 01:26:33 | Matthew Rabinowitz | No, no. You are taking the exact output … The output from the interim risk team basically or those risk metrics that were outlined. And one of those needs to be, in my opinion, the percentage breakdown between these two. And that's- |
| # / 01:26:49 | David Utrobin | So we're looking at this graph, right? And this graph assumes everybody's on an equal stability fee, right? |
| # / 01:26:58 | Matthew Rabinowitz | Yes. |
| # / 01:26:59 | David Utrobin | But the reality is that, this is like percentage of DSR usage, but wouldn't it be unlike the same type of graph, but instead of DSR usage, actual DSR base rate on all of these different trenches of collateral? Wouldn't it be just a common denominator? Wouldn't it be as simple as that? |
| # / 01:27:20 | Matthew Rabinowitz | Except for the fact that when you have DAI that gets minted, how do you change one VaR without it impacting everybody? Because the DSR is common to everybody, right? It's a common addition. Every single person's ultimate stability. |
| # / 01:27:43 | Cyrus Younessi | So let me try to explain how I think about it and then you can try to tell me just where you disagree. I don't have a defined, rigid viewpoint, but this is just my perspective right now. So anyone who mints DAI, whether it'd be from a super safe treasury, whatever, or from a super risky asset, like Ether, when they create the DAI, they immediately or semi-immediately pass it off to someone else, some third party, some counter party, whomever is their recipient, right? The counterparty of that margin trade or whatever. |
| # / 01:28:31 | Cyrus Younessi | The DSR is essentially a rate that gets paid from the borrower to the recipient that compensates them for wanting to hold that DAI and not in turn, immediately sell it for less than a dollar. However you want to think about it, the DSR is a rate applied that is for the benefit of the counterparty for that leverage trade. Now, it's not entirely clear to me why the recipient of the counterparty of the DAI created from a treasury collateral would require a different compensation rate from the recipient of this DAI trade. |
| # / 01:29:20 | Matthew Rabinowitz | No, they don't. |
| # / 01:29:20 | Cyrus Younessi | Because they both are just holding DAI, right? |
| # / 01:29:22 | Matthew Rabinowitz | Right, that's correct. |
| # / 01:29:22 | Cyrus Younessi | So whoever creates DAI has to pay the same fee to the recipient of that DAI. |
| # / 01:29:32 | Matthew Rabinowitz | Correct. And it didn't matter how risky that asset was. That's correct. |
| # / 01:29:37 | Cyrus Younessi | Right. Now additionally, if the asset is riskier, which is a risk to MakerDAO, there's an additional credit spread, credit premium applied, correct? |
| # / 01:29:48 | Matthew Rabinowitz | Right. |
| # / 01:29:51 | Cyrus Younessi | So what you're suggesting though is that, if there is an enormous amount of DAI being printed off of ETH for example, and it's depressing the DAI price, that you don't think that the DSR should be applied, but rather a collateral system. But my point is though, if the recipients of the ETH-DAI aren't willing to keep the peg stable at that current DSR, then why is the other collateral's DAI counterparty willing [crosstalk 01:30:23]? |
| # / 01:30:24 | Matthew Rabinowitz | Yeah. So, it's not that I'm suggesting the DSR shouldn't be used. I'm saying that it should be used in proportion, and that proportion should be known, because if we follow the model where we're just talking about, what percentage is the question? So like if we ultimately … Let's just take ETH for a second. That's the case example just. Forget we have a treasury all together. We're just going to use MCD, but we're going to forget the M and we're just going to have a DSR and launch it, right? |
| # / 01:30:54 | Matthew Rabinowitz | So our first instinct will be, like I've mentioned before, we need to iterate it back and forth over and over to basically get rid of the excess supply, which means the DSR will find its way to get up to, I don't know, five, six, seven, eight, 9%. And then in both of them, the DSR as well as the VaR for just Ether will find themselves to be compressed and fall back down over time. But while it's sitting at five or six or 7% of the DSR, you absolutely obliterate your ability to ever launch any type of lower risk asset, because the DSR, plus that risk premium is way outside the market because it's way outside the aggregate risks that's tolerable. |
| # / 01:31:40 | Matthew Rabinowitz | Because we put a more risky asset, we're not correctly pricing risk for the asset is my point. |
| # / 01:31:47 | Cyrus Younessi | Right. So there is an intermediate case where if the credit risk of ETH is mispriced, then you will generate excess DAI than what should be, and that could depress the DAI price, which is [crosstalk 01:32:02]. |
| # / 01:32:03 | Matthew Rabinowitz | I'm not really talking about the price, I'm talking … right now, I'm making the base case if the prices just constant and that is temporarily constant. And then you introduce a more risky asset and a far less riskier asset. And both of them mint DAI, what do you do? And how do you determine how can we count …? I'm not making the case that we can't. I'm not saying we have to lock these today. I'm saying that over time, you have got to lock it. |
| # / 01:32:31 | Matthew Rabinowitz | Because how will we be able to determine without sticking our finger in the air and saying which way we think it should go. That the DSR should mop up the excess supply from collateral A, B, or C more. The only way I know how to do it scientifically is to lock a percentage breakdown between the DSR and the VaR for each asset as they're used. And then wind it down as the conditions on the ground change based on the credit, the risks team's analysis of it. |
| # / 01:33:02 | Vishesh Choudhry | How do you get it? |
| # / 01:33:03 | Matthew Rabinowitz | How else could we determine what you were saying? Like the DSR, the DAI buyers that buy DAI and just lock it up. Then they don't care. They're just capturing a yield, right? The challenge is when we increase or decrease that yield, it should be disproportionately in favor of the ones that are lower risks. That's why we have ones on the left that are blue and ones that are on the far right that are purple. |
| # / 01:33:30 | Matthew Rabinowitz | Because if we make it a flat line, then we're basically treating the less riskier asset is getting disproportionately … Is going to have to pay more, whenever a more riskier assets causes in effect problems. |
| # / 01:33:47 | Cyrus Younessi | But I still think it definitely could happen, but I think it's still unproven that the riskier asset will necessarily cause … Will uniquely cause or unilaterally cause a drop in the DAI price. So, there's a chance that that just issue doesn't actually play out. It's not guaranteed. |
| # / 01:34:10 | Matthew Rabinowitz | It'll not only decrease. It's not only a decrease. It can also be, there's excess DAI that got burned from some location and we have to adjust the DSR down, but by what amount? That that's my point. This piece of it's far more mechanical. It's far more closer … It's closer to the circuit theory than anything else, because once we know these percentages, the things that need to change are the risk characteristics, the credit risk characteristics. |
| # / 01:34:39 | Matthew Rabinowitz | And how much we apply the DSR to change the entire system is what makes it interesting. It's that collateral A, we're going to be voting on in effect the DSR which affects absolutely everybody. And we're going to be voting on the VaR for every single collateral type. |
| # / 01:34:58 | Cyrus Younessi | Right. So, can you agree that … yeah, I see what you're saying, I just don't think it's necessarily going to play out that way. Because if a ton of … if some risky asset is generating a disproportionate amount of DAI, but the safer assets and DAI in those markets is trading below a dollar, right? But the safer assets are also generating the small amount of DAI and wherever the DAI goes, there it's trading at a dollar, right? |
| # / 01:35:40 | Cyrus Younessi | For starters, that's no longer like a, “Oh, I need to wait for the …” There should be some arbitrage there. |
| # / 01:35:53 | David Utrobin | There's also like something else to consider, right? Like the fact that … How we're going to be using these as policy tools. So VaR is not really going to be used as a policy tool like the stability fee is not really going to be used as like a pair of policy tool as much as the DSR is, right? |
| # / 01:36:11 | Matthew Rabinowitz | I don't know [crosstalk 01:36:12], that's the point. It's x times DSR, but x is common to everybody and it should be y of collateral A, times the VaR. But the percentage is known, so we're just voting on … No, the x and the y, the VaR is what we're voting on, not the x and the y just, right? |
| # / 01:36:34 | Cyrus Younessi | I think this is the answer to your question and in some respects it becomes not a risk decision, but kind of a MakerDAO managerial decision of sorts. Because if you think about it, if causing a restrictive DSR to some assets allows the most popular asset to flourish and generate the most amount of buy and burn, then potentially that's a decision that whatever MakerDAO want to go with, they can decide [inaudible 01:37:11] straight up. |
| # / 01:37:11 | Cyrus Younessi | But it's not really a risk decision at that point, right? It's more of, which direction does MakerDAO want to proceed in, if that discrepancy even shows up, which I'm skeptical that it would. But it's not entirely clear that MakerDAO would hamstring a particular asset for the benefit of another asset, assuming that it just treats all assets [crosstalk 01:37:35]. |
| # / 01:37:35 | Matthew Rabinowitz | What I'll say is [inaudible 01:37:36]. What I'm trying to get at is that market forces, the ultimate stability key that an asset class comes together. If we ultimately have the underlying DSR too high, because we're not properly allocating the risk, it's not that Maker won't choose to, the market will dictate. For example, just as a base case, if we ever want to put on treasuries onto this system, but the DSR is 5%, there's no use case. |
| # / 01:38:07 | Cyrus Younessi | Right. So then it's a question of like, if allowing treasuries on would require lowering of the DSR to whatever, 1%, I don't know. And then that would in turn cause the other collaterals to go out of whack. Then- |
| # / 01:38:24 | Matthew Rabinowitz | Definitely, this is- |
| # / 01:38:25 | Cyrus Younessi | … you'd have to choose between if you want to treasuries, if you want those other assets. |
| # / 01:38:28 | Matthew Rabinowitz | Right. And this is where I think instead of trying to … this is where tying him down, when you know the percentage between the two, between DSR and the VaR, but you're voting on the allocation of that pain across the board. Then if you … the challenge I thought about just over the weekend was, how else do you do the system that allows it to be inclusive for all collateral types? |
| # / 01:38:53 | Matthew Rabinowitz | Because if we don't do it that way, we'll find ourselves in a scenario where the structure or the priced out of whack for some collateral types, but injecting systemic risks for others when it wasn't really- |
| # / 01:39:08 | Cyrus Younessi | Right. And that's definitely a good discussion to have I think more from like a, almost like a social perspective of the overall philosophy of MakerDAO and its goals and its vision and all that. Because if there is a tradeoff, which they're very … may well could be … It's certainly not for any individual to say what the direction should be, but at the same time it's also not strictly a risk decision. It's more of a operational of between the MKR holders, which we can definitely discuss on this call for sure, but it's not … |
| # / 01:39:50 | Cyrus Younessi | If MakerDAO holders want to allocate those percentages in a certain way at the social consensus level, they're more than welcome to. And then we can try to figure out what the right parameters should be. But I don't think well, we're just speaking from the risk angle, can't say what parameters should or shouldn't be whatever. If the goal is to be inclusive of all collateral types, no matter what, then that's something that certainly needs to be discussed. |
| # / 01:40:22 | Cyrus Younessi | But it's not clear to me what exactly the mandates are probably that we have in the coming months. |
| # / 01:40:28 | David Utrobin | I'm still trying to think of it. And Matthews, I got to probably reread your posts if you go through the math of it, because the way you described that equation started to help me understand, but I still just kind of see the credit premium on all of these various collateral types being exactly that, the premium based on the actual risk of the asset. And then the DSR is really just what goes up and down in response to monetary policy. |
| # / 01:40:59 | David Utrobin | So, you could have consistent VaR based on very kind of well-defined parameters per parameter package or whatever. But then, you use DSR is that adjuster for whenever the DAI prices are out of whack. |
| # / 01:41:14 | Matthew Rabinowitz | Yeah, that makes sense, as long as everything is more or less equally risky. |
| # / 01:41:18 | Vishesh Choudhry | Right. |
| # / 01:41:18 | Matthew Rabinowitz | But if you use the DSR for things that are less risky, especially the substantially less risky, then your percentage-wise blowing that entire market up, because- |
| # / 01:41:30 | David Utrobin | Oh, I see what you're saying. |
| # / 01:41:32 | Matthew Rabinowitz | Right? Because if everything's more or less the same risk category, read and off to the right, using of the DSR is fine. Just imagine like there's a … This is actually really important. Unlike DAI and people who want to use leverage based on ETH, which is an exceptionally elastic desire, where we've seen what people are willing to have stability fees that go from zero all the way up to almost 20, right? In certain other markets, let's say real estate, country debt market, sovereigns, once you reach a certain percentage above where that tolerance is defined by the external market, it's inelastic, there's zero desire, it's blown out of the water. |
| # / 01:42:15 | Matthew Rabinowitz | So if a treasury [crosstalk 01:42:17] sorry about that [inaudible 01:42:18], but if the treasuries equivalent total amount could be three, then that's it. Then you can't have it go higher, otherwise the usage drops to zero instantly. |
| # / 01:42:30 | David Utrobin | So it's a matter of being competitive against other markets and having the lowest rates that make sense for those collaterals that are safer, and not having the safer assets subsidized basically. |
| # / 01:42:42 | Matthew Rabinowitz | Exactly the same with this. |
| # / 01:42:43 | David Utrobin | [crosstalk 01:42:43] the rate for … Yeah, okay. |
| # / 01:42:45 | Matthew Rabinowitz | The more risky assets can't overwhelm or disproportionately take the … What we call the mapping tool that'll subsidize that effect by the less risky tool. |
| # / 01:42:57 | Cyrus Younessi | Can't or shouldn't, right? |
| # / 01:43:00 | Matthew Rabinowitz | Well, this is the challenge. If you don't do it, if we don't walk it more or less [crosstalk 01:43:05] and so on, it'll just go. |
| # / 01:43:06 | Cyrus Younessi | But I'm saying there could be a rational reason to not lock into … for example, if MakerDAO … I really don't think this will be the case, but for example, if there's some phenomenal asset XYZ that is generating insanely high stability fee at low risk, then to hell with everything else, right? Potentially that's one … again, that's not even the philosophy of MakerDAO as far as I'm concerned. |
| # / 01:43:38 | Cyrus Younessi | I'm just saying, it definitely needs to be carefully spelled out and come to consensus on how we want to drive this protocol forward. |
| # / 01:43:49 | David Utrobin | There's also one more consideration in the fact that the DSR base rate is charged on all outstanding DAI, whereas the DSR actually collected as a result of blocking DAI up in the contract is not that full amount. So you still have a surplus of stability fees entering into the Maker coffers so to speak. |
| # / 01:44:13 | Cyrus Younessi | Right. |
| # / 01:44:15 | David Utrobin | And that could produce a structurally lower DSR rate generally, maybe. Maybe I'm not thinking about that right, but I'm a newb and this is really interesting, but I think I got to go in a few minutes. |
| # / 01:44:29 | Matthew Rabinowitz | Right. Well, it's something for us all to continue with the discussion. The point I'm trying to bring up is not that this concept should be applied to any aspect of the risk in my respectful opinion, needs to make those decisions about how risky a given collateral package is, with all of its characteristics and define it, plot it and set it. And if we decide later we don't like it, then we need to drop the debt ceiling or let it organically die off and replace it in another structure that's more or less risky based on the conditions on the ground. |
| # / 01:45:04 | Matthew Rabinowitz | After we have that, if the DSR, VaR is locked or given collateral package, again, the rest of it's math to determine what is the proper non-subsidized efficient allocation of the DSR across all of the risks back. |
| # / 01:45:21 | Cyrus Younessi | Cool. Let's strike up the conversation offline. It's definitely interesting. |
| # / 01:45:28 | David Utrobin | Yeah. This was a wonderfully interesting conversation guys. Thank you. |
| # / 01:45:35 | Cyrus Younessi | Cool. Vishesh, to answer your question about parameters really quickly, who can change them or how often? Who is obviously MKR holders change them. How often, cadent to something that we have to figure out eventually. What parameters can I think? Any parameter. The subject to be changed … and what do you mean by, does that model converged to MCD? |
| # / 01:46:04 | Vishesh Choudhry | Well, that's what you were talking about in terms of if you start to disproportionately penalize riskier assets, you run into the potential where it's only like determined by market forces and it's determined by logical evaluation of risk. It's a more stable system. Could theoretically be that the answer to the most stable system in that scenario is just one single collateral type that is like relatively well trusted. |
| # / 01:46:33 | Vishesh Choudhry | Which although would make sense in terms of the individual value at risk and DSR allocations, might overall from some sort of holistic bird's eye view perspective, not be optimal, because we are trying to maintain this relative lack of correlation between assets. |
| # / 01:46:52 | Cyrus Younessi | Right. |
| # / 01:46:53 | Vishesh Choudhry | You need to keep in mind. |
| # / 01:46:54 | Cyrus Younessi | Is that why you said shouldn't to my answer earlier …? To my question earlier? |
| # / 01:46:58 | Vishesh Choudhry | That's definitely a factor, but I think that's just objectively … It's what we think we should or shouldn't do. It's not that it's the only structure that works. |
| # / 01:47:10 | Cyrus Younessi | Right. |
| # / 01:47:13 | David Utrobin | Anyway, I'm stopping the recording here, you guys, so thank you for awesome call as usual. We'll talk to you guys throughout the week and excited for next week's call. |
| # / 01:47:21 | Cyrus Younessi | Okay, cool. Sounds good. |
| # / 01:47:24 | Vishesh Choudhry | Cool. |