Governance and Risk Meeting: Ep. 49 (August 22 - 2019)¶
Video | Audio | Discussion
| References | Person | Text |
|---|---|---|
| # / 00:00:04 | Richard Brown | Hello everyone. Welcome to the August 22nd edition of the Scientific Governance and Risk Meeting. My name is Richard Brown. I am the Head of our Community Development and MakerDAO, we have lots to go over today. I say it every single time, but it's true every single time, so I'll keep on doing it. |
| # / 00:00:20 | Richard Brown | We are enormously interested in hearing from people in this room. So please, I am begging you if you have a question let us know what that question is. Type it in the chat at the side. If you have access to a microphone, please just get on the mic and interrupt us. We don't mind at all. |
| # / 00:00:37 | Richard Brown | The agenda for today is I'm going to talk for ... I'll try and figure out how much of Q&A we need to get through, because I'm the conference facilitator of things. See if there's any interest in getting through that a bit at the top. Bob Wisdom, or LongForWisdom to his friends who will be giving us a brief recap of what's happening in the forums and governance in the last week. |
| # / 00:01:04 | Richard Brown | Cyrus from the risk team will review the risk team mandate, which is another large documents in the same vein as the governance facilitator mandate, that went into the forums recently. He'll walk us through that and we'll hear from Vishesh, and he will tell us what crypto has been up to and what the peg looks like, and actually Matthew, I haven't confirmed with you. But if Matthew has a weekly narrative for us, we'll be talking about that as well. |
| # / 00:01:32 | Richard Brown | So let's jump into things almost immediately. I don't have another speech prepared or a document I want to review, but I do want to just at the top of the hour open the floor to see whether there's questions that people have about the governance facilitator role and whether I can address some of them immediately. Yeah. Well, let's do that right now. If anybody has any questions about the governance facilitator role about the conversation that's been happening in the forums, about any specific item let me know. In the absence of that, and I know it can be a bit intimidating to be the first one to do it. In the absence of that, I might just open up the threads, run through it and see if there are things that we could address. |
| # / 00:02:19 | Richard Brown | I woke up this morning to a significant amount of interaction in there and I haven't had a chance to reading it yet. So perhaps we could attempt to dig into that as well but we don't have a ton of time today. You know what, LongForWisdom, Mr. Wisdom maybe I can put you on the spot here because I know that you are enormously engaged, a very clever person and you do not feel obligated to agree with me. And so, you might be the best person to maybe surface some of the primary complaints or concerns or questions or comments that people might've had. |
| # / 00:02:59 | LongForWisdom | I can do. |
| # / 00:02:59 | Richard Brown | Go ahead. |
| # / 00:03:01 | LongForWisdom | I feel like a lot of them are probably like personally my complaints, so I don't know that you should have potentially say no if I was ever on the side. |
| # / 00:03:08 | Richard Brown | I thought you were a team player, but all right. Okay. It's entirely cool. This is a completely open forum and the hard questions are why we're here. So if you have some questions that you wanted to kick things off with, that'd be great. |
| # / 00:03:25 | LongForWisdom | Yeah. I mean I mentioned this, this in the thread and it's mainly the timeline, right? It seems there isn't enough time to get a good amount of community engagement, I guess in response to the things. |
| # / 00:03:37 | Richard Brown | Yeah, that's something that, yeah, we've obviously been wrestling with this for a long time because that's what these calls are all about, is to interact with the governance community, figure out basically what the landscape looks like. And I bounce these memes in that thread as well fairly aggressively, but defining problem spaces and then figuring out the scaffolding of what the process is going to look like in the future and stuff. I don't have answers to all of the things that we need to have answers for, I'm hoping that the community provides 99.9% of those answers. |
| # / 00:04:16 | Richard Brown | And then the governance facilitator role is in charge of helping to make sure that those things get executed appropriately and on time. |
| # / 00:04:24 | Richard Brown | But the ... yeah, so the amount of time that it takes to properly evaluate these things is going to be the first thing that we need to figure out in the governance space because we're breaking a lot of new ground here. This is the first time that we've formally presented a mandate/proposal for community ratification. We don't know what the level of engagement is so we could have just arbitrarily put in, okay, here's two months to discuss this or two weeks, so we just made a guess. The guess was, it would take two weeks for the community to either express interest or for us to be comfortable with the idea or confident that the lack of interest that we've seen so far is indicative of the future interest, and so we should just try and expedite this process as much as possible. |
| # / 00:05:15 | Richard Brown | That's a long winded way of saying that we just arbitrarily picked two weeks, and let's see how things are going to happen. The expectation I think is that if there was no community interaction with this thing, there's no debates then we were just moving into a forum poll and they're moving into the governance portal. If there was a fantastic number of community interaction, significant amount of debate, and then we would just the way things are like, nothing is written in stone here, so we could usually determine that there's just too many open questions in the forums. Right. And that's we'd push this thing off until all those questions have been answered. |
| # / 00:05:57 | LongForWisdom | Yeah, I think it's difficult because the actual content of that proposal and the rest of proposal are both they both seem really strong. I don't have much feedback towards the actual proposal itself, it's really good, but it's about the process side which is what I have an issue with currently. |
| # / 00:06:13 | Richard Brown | Yeah. And I think that the important thing to remember is that these are broad strokes and everything that we're doing right now is an iterative process. This is something that we've had, you would not believe the number of conversations that have been happening in the foundation over the last weeks and months to carve out what governance potentially could look like. And we've identified three different schools of thoughts here. And this might be getting too far into the weeds, but there's the canonical source of truth camp. There's the mechanical manual camp and then there's the iterative camp. And so let me clarify what those might mean. |
| # / 00:06:55 | Richard Brown | The conical source of truth camp believes that or is suggesting that we create a framework with sufficient level of resolution and present to the community as a mass package. So here's the binder of the rules or the principles and the commandments and the processes that will be adhered to say yes or no. If you say no, then we go back to the drawing board and we come up with something new but that's one camp. |
| # / 00:07:29 | Richard Brown | There's another ... I shouldn't say camps, because I mean, camp applies division. That's these are just ideas that have been floated. And then another idea was that we ... this idea of a manual almost. And so here's the 15 rules of order that apply to establishing a forum poll and then here's the levels of forums and thresholds and written rates and algorithms that we use to determine what a consensus actually looks like and then here's, et cetera, et cetera. And let's discuss that process where it goes on and on and on. That's the creating a manual of what the future will look like. |
| # / 00:08:09 | Richard Brown | I have concerns about both of those paths because we're in a very, very, very nascent space. In crypto and in general we're in a double Mason space when it comes to open governance and DAOs and having that layered on top of crypto, I have very little confidence in our capacity to predict the future three months out let alone six to 12 to 18 months out. I think that at that level we're just engaging in futurism, which has a very limited historical level of accuracy when you get into that thing. The model that I'm hoping that the community will adopt or is at least comfortable with is this idea that we iterates. We come up with really good plans. |
| # / 00:08:56 | Richard Brown | We have a very clear definition of what the future might hold and what we want that future to look like. And then we pivot and iterate as necessary as we approach that goal line that we've all agreed upon. And so that's a very, very long winded way of saying that we come with a really good idea. We were hoping that this governance facilitator mandate is a really good idea and the next step is that we see whether people think or agree with this, that it's a good idea, if they have any ideas about how to make it a better idea and then we do it and then revisit in two or three months and say, "Okay, well, let's fix this part, let's change these numbers, let's do this slightly different," and then we can just ratify it again. And then we iterate on this thing. |
| # / 00:09:40 | Richard Brown | I have a tendency to look at these things as if there were a software and I don't think that that's a bad way of doing things. We can release version one of the good term, of the governance facilitator mandate. And then when it's time to release version two, we'll re ratify and then version three and then six months from now. So we might have something completely different or we might have this with all the edge cases ironed out. I'm not entirely sure, but that's the way I hope things progress. |
| # / 00:10:08 | LongForWisdom | Yeah. I think I share that like iterative ... it seems like a good way to go. |
| # / 00:10:12 | Richard Brown | Yeah. So, that was a fancy way of saying that we have no idea what the future's going to look like. We can make our best efforts and we are following the same plan of release, early release often. And I think that as long as we're conservative about the powers and the scope, and the impact of these changes, we should be in a good place because we're in a very ... we're always cognizant of the foundation. I think we should always be cognizant in these community calls that we're not just playing around with arbitrary dials. |
| # / 00:10:44 | Richard Brown | We're not ... what we're doing here is serious business. This ecosystem is small, but we are the linchpin, and it's hard to say this without sounding immodest, but MakerDAO is the linchpin of this emerging DeFi ecosystem. There's third party vendors that rely on our stability. There's people that rely on our stability. And so governance in my mind is not a place to be having a radical experiments, but it is a good place to take baby steps and plan things effectively and iterate and fix as we go along. I'm hoping that that's the way that we evolve this thing. |
| # / 00:11:29 | Richard Brown | Did you have a ... well, I know that you did, so you want to move on to concern number two LongForWisdom? Did you have more that you wanted to go through today? |
| # / 00:11:44 | LongForWisdom | Yeah, I mean there's more I can talk about. I don't want to use up too much time though, so it's up you. |
| # / 00:11:48 | Richard Brown | Yeah, that's a good point actually, because I always start these things off saying I'm going to keep it quick and then bang, 30 minutes have gone. So if we can, this is one of the things that we need to figure out as a group too. And I've been pushing this idea and we'll see how whether everybody else agrees and we all align on this, but these governance calls are not a decision making forum. We cannot expect that the course of MakerDAO or the ecosystem is shifted in the hour a week that we have allotted to this North American friendly time zone where we're all speaking English. That's not very global. It's not very banking the unbankedy of us. In these calls we have chats, we surface issues, we set tone, we communicate things, we help people who want to come along and advocate for anyone for secure position or role or idea. |
| # / 00:12:41 | Richard Brown | This is a great place to do it, but it's largely a communication medium but governance happens in the forums. And so if people have questions we don't address in these calls, please, please, please hit up the forum thread and let's continue the debate there. I'm going to post the link one more time in case people are not familiar with what it is we've been talking about for the last 15 minutes and I have my own homework to do. There was four or five really substantial comments that were left while I was asleep. And so I'll check those out after these calls are over. But I think that unless there are additional questions about the governance facilitator's specific concerns we'll probably leave it there and then potentially pick it up again. |
| # / 00:13:28 | Richard Brown | After the top of the hour we are all going to stick around, or anybody who wants to is welcome to stick around for three minutes to just have a generic Q&A. So maybe once things have been warmed up a bit, people can ping me for some questions. But I think that, yeah, you're right. Yeah, LongForWisdom must keep track of the time and I'm going to hand it off to you actually for governance of the clients to come and give us a recap, that'd be great. |
| # / 00:13:55 | LongForWisdom | Yeah, absolutely. |
| # / 00:13:57 | Richard Brown | Go ahead. |
| # / 00:13:59 | LongForWisdom | So the major two threads I guess, well, I wanted was the risk team mandate, which was ... [inaudible 00:14:05] a couple days ago. So that's had some feedback on already, some from me, some from others. I would suggest everyone looks at that. Rich posted the results of the collateral asset priority poll, so there's a whole breakdown in there of the results and what that means for the prioritization of ... with the initial set of client collateral assets. So that's interesting. |
| # / 00:14:37 | LongForWisdom | Again, as Rich has been saying that we've been talking about the facilitator thread is still fairly active. We had to, yeah, there was Michael [inaudible 00:14:48] re-posting a post he posted previously, which is talking about goals setting for the foundation ... sorry, for MakerDAO, which I thought was interesting. Let's just read and have a look at that, it seems so, I think that, okay. |
| # / 00:15:07 | LongForWisdom | Other than that, I haven't really changed much in the other sections. One thing I would say is the trying to seek consensus around the exponential rate stepping for stability peoples it's still going. We are at 66% consensus on it. One of the options now I think there's 15 people who have taken part in that. So we have 10 people who have 15, that want the same thing, so no impressive numbers yet really. But anyone else who hasn't signaled in, and they would like to, that'd be awesome and that's pretty much the summary of some of the stuff going on, but it's a limited space. |
| # / 00:15:46 | Richard Brown | All right, thanks for that. That's good that you've raised those, that last point to because this is something we have their lingering concerns and questions. I keep on saying concerns, I should, there's these lingering tasks that we need to complete now that we have a more robust polling solution available to us. Two of those decisions are ... does the weekly cadence for the stability fee make sense? Do we need to think about two weeks instead? There's a pile of pros and cons in that thread. I'm going to copy that, I'm just going to click ... yeah, [inaudible 00:16:25] forum thread mix in check. We always have equipped, the way that we do these rates, stepping numbers into these polls like zero, one, two, three, four, five, six, seven, eight is not enormously intuitive. It gives us a linear range, but it doesn't give us granularity and it doesn't allow us for larger leaps should an emergency arise. |
| # / 00:16:50 | Richard Brown | So two questions outstanding, should we consider exponential rate stepping? So 0.5, one, two, four, eight in these polls in general or stability fees in specific? And should we be voting every two weeks on the instability fee? Many don't want to see to get lost, I know that we have this new polling system, we should consider wrapping these discussions up and then getting them into the governance portal proper at some point in the next couple of weeks. So please have a look at those two threads, make your voices heard in those polls and then we will start figuring out exactly what that process is to move things from the forum and into the governance portal, which is part of this emergent iterative process that I keep talking about, that we should ... it's time to bite that one off. So we should try and do that soon. |
| # / 00:17:41 | Richard Brown | Right. I think this is probably where we should hand it off to you Cyrus because there's a lot to go over with your risk mandate, I think. |
| # / 00:17:52 | Cyrus Younessi | Agree. I want to start off by apologizing in advance. I'm traveling overseas and my internet is not great. I actually got cut out of the call earlier for about five minutes. Well, I was trying to troubleshoot it, so I hope that doesn't happen again. Okay. I've added the link to the chat for the risk team mandate. There was, I guess I'm going to start off by talking about a little bit about the history of this document and how it came about, how we got to creating this mandate in the first place. |
| # / 00:18:41 | Cyrus Younessi | I think from the early days of MakerDAO there had always been this notion of having decentralized risk teams. And then when Steven came along I think he introduced the notion of this decentralized risk function, which is basically this philosophical mandate behind how risk for MakerDAO might operate. |
| # / 00:19:07 | Cyrus Younessi | And he posted a three part series on risk, I guess just about a year ago actually and I posted a link for that in the chat as well. I highly recommend everybody read this three part series. It's phenomenal. But I think as we have been getting closer to MCD launch, there are a number of operational aspects to how risk teams might operate. Some very specific nitty gritty questions and as I started to field more and more of these questions, offline made sense that we would create some sort of almost an FAQ style document for risk teams. Actually that's how this risk document was originally structured. |
| # / 00:20:04 | Cyrus Younessi | I think as we started to explore that style though, we realized that this would be ... there aren't just operational questions for candidate risk teams, but also for MKR holders themselves and realize that it would be nice if we could get everybody on the same page for how risk teams might operate in the community to minimize any potential chaos. While the notion of a purely decentralized governance framework is definitely a very short term goal, as of today, I think it's difficult for example, the risk team to come along and go through any negotiation tactics with some arbitrary DAO. Can you guys still hear me by the way? |
| # / 00:21:09 | Richard Brown | Yes. Loud and clear. |
| # / 00:21:11 | Cyrus Younessi | Okay, cool. Right, so this document also transformed into something for the benefit of MKR holders as well as prospective risk teams. And by ratifying this mandate through an official governance poll, it helps the other governance aspects such as Richard's role as the interim governance facilitator to help streamline the process, make it operate as smoothly as possible with minimal friction. One of the concerns that we had on the purely decentralized version was the amount of overhead that would go into basically onboarding the risk team and the risk models and the risk parameters, and all this stuff, just a dizzying amount of edge cases and potential votes and so forth. |
| # / 00:22:12 | Cyrus Younessi | So I think by creating this mandate, we're essentially putting forth a first pass at how we might do a practical implementation of risk governance. And so with that, I'll jump into the doc a little bit. I'm not going to run through the entire doc, but just run some of the more interesting points especially the first few sections I think are more or less familiar with the existing MKR community, nothing groundbreaking here. I think it's a nice introduction for perspective risk teams coming along who may not have a perfect overview on MakerDAO. Right. |
| # / 00:23:03 | Cyrus Younessi | So, really the meat of it in terms of new information is two things. One is what are these risk models going to look like and how do we actually onboard new risk teams? For the first question regarding risk models, we call them risk constructs, which is a more, a broader term that incorporates not just any specific model, but also almost a more philosophical version of the risk models. So it includes these principles of scientific governance, how we expect these risk packages to look like, what are the standards to which the MKR community should hold these risk models? We certainly don't want risk teams to come along, right? |
| # / 00:24:05 | Cyrus Younessi | Okay, so here's a hypothetical worst case scenario. Risk team comes along, they say they understand everything perfectly within two days. They submit a risk model and they demand that the risk model gets approved the following week. Without providing sufficient time to vet this information, without upholding risk teams to a certain standard of what these risk models might look like, things might get chaotic and even worse, potentially subjective. Right? You might have ... I think it would open up the discussions to capture from various interests, and I think that's a lot of the purpose of this doc is to minimize the ability for biased ecosystem actors to come in and capture the risk process for their own benefit. |
| # / 00:25:10 | Cyrus Younessi | With that being said, the very basic risk construct could comprise of three parts. One is obviously the dataset and the data being used. I think without supplying a good data set, it's difficult for the community, for MKR holders and everybody to verify the legitimacy of pretty much everything else, right? You need to start with good transparent data and any transformations made to the data such as filtering for wash trading or any other analysis should obviously be specified so the community has a good grasp of what they're looking at. |
| # / 00:25:58 | Cyrus Younessi | Following the data set, this is where things get potentially a little confusing. So we have a general model and an applied model. And the applied model is what you might expect from a typical risk model, which is risk team comes along, they say, "I evaluated ETH, here's all my fundamental analysis, here's all the financial models I'm using and these are the risk parameters that it spit up." While that is at a first glance seems okay, we're asking prospective risk teams to start off with initial general model that is basically describing the methodologies that they might use and the logic behind them. |
| # / 00:26:59 | Cyrus Younessi | The reason being because we're not entirely sure to what extent MKR holders or the community at large might be able to interpret or process finalized risk models without sufficient background on the subject. So for example, if a risk team were to come along and say that some utility token is completely worthless, why exactly? And so potentially a general model would be a place for risk teams to spell out their methodology behind how they might evaluate particular token classes. |
| # / 00:27:49 | Cyrus Younessi | Now, of course this information makes it into the applied model as well, but I think part of the benefit here is that the applied model will almost be a deterministic output based on the dataset and the general methodologies. |
| # / 00:28:09 | Cyrus Younessi | So potentially one issue we might face is there's going to be a lot of, all right, chaos and drama and bickering over the final proposed parameters, right? Some communities favor token gets very unfavorable risk parameters and they come in and they say, "Well, I don't think token X, Y, Z should have such a high stability fee. That doesn't make sense. It should be 2% or whatever." Right? And instead of debating the outputs of the risk models, we would strongly suggest that we debate instead the general model and the methodologies put forth. |
| # / 00:28:56 | Cyrus Younessi | Again, it maybe somewhat restrictive in some senses, but I think the goal in the early days is to minimize chance of capture, and I think this is also naturally how we were planning to do it as part of the interim risk team. So over the next month we're going to release basically some general frameworks on how one might evaluate the risk of crypto assets or basically all the different token structures, how we might look at the data, how we might put that into financial models and so forth. |
| # / 00:29:40 | Cyrus Younessi | The general model is really the meat of the risk construct, afterwards the applied model is basically taking that formula and applying it to a specific token and you would say, "Okay, based on our methodologies as described in the general model, we would take this approach to evaluating a certain token." |
| # / 00:30:10 | Cyrus Younessi | Oh, and one of the big yet subtle implications from this whole process is the difference between voting in risk teams versus voting in risk models. There was a lot of discussion on what it would mean for a risk team to be voted in, in the first place. If a risk team were to be voted in, would it make sense for them to just be able to go and submit any risk parameters from there on forward? Obviously not. And really the power of a risk team is limited to their ability to continue to output these risk constructs and essentially upon the approval from the community of that first risk construct that will essentially designate that team as an official Maker risk team. |
| # / 00:31:22 | Cyrus Younessi | Outside of that, they don't particularly have any special powers because going forward anything they want to do, they would still have to justify their work and continue to create new risk constructs. In a roundabout way, you could say that we're somewhat abstracting the concept of a standalone risk team being voted in and having any arbitrary power. They certainly won't have the privilege to come in and say, "Things have changed and we're going to drop the stability fee," just through a simple post, right? We would want them to somewhat go through the same process again. |
| # / 00:32:13 | Cyrus Younessi | Now for as long as they're submitting risk constructs that are attached to the original general model, that's more or less okay but I think as soon as a risk team would want to branch off into different assets or different classes of risk in the MakerDAO ecosystem, then they would essentially need to recreate and resubmit a general model for approval. That's again, another reason why risk teams by themselves will not have arbitrary power, right? Because for example, let's say we were to onboard a specialist for just for crypto assets such as ETH and then two months later they say they want to start evaluating stocks or other traditional assets. Right? We would obviously want to avoid those kinds of situations. I'm going to stop here for questions on everything, including the risk models, of course. |
| # / 00:33:38 | Cyrus Younessi | Yeah, it's a lot to take in for sure and a lot of these decisions are very nuanced for a number of reasons and it's not entirely clear why certain decisions were made the way they are. So yeah, I obviously encourage a lot of discussion in the forum and on the calls and I'll actually get to some of those questions later today. |
| # / 00:34:16 | Cyrus Younessi | "Is it always assumed that the distribution of price paths is normal?" No, it's not, but maybe that's also a little bit too specific a question for this risk mandate. |
| # / 00:34:45 | Cyrus Younessi | Okay, let's move on. Let's finish the doc and then I'll circle back with some other questions and the subtle implications of the mandate. |
| # / 00:35:00 | Cyrus Younessi | Okay. We've defined that we propose that a risk team would submit risk constructs in this certain manner, which is dataset general model and applied model, which is calculation of the risk parameters. Now the second big question is, how or through which forum do they actually go and submit these risk constructs? And so, I mean, at the extreme of decentralization, it could happen anywhere, right? Risk teams could come along and say and just post risk models anywhere and they get traction and whatnot, then that's totally fair game. |
| # / 00:35:53 | Cyrus Younessi | I think similar to the notion of Rich and his interim governance facilitator role, we're also proposing a particular methodology through which we think it would make sense for the process to play out. And this process is described here under onboarding risk teams. Basically first step, risk team comes along, they go to the forums and they post some credentials about themselves and they say, "This is who we are, this is what we've done in the past and we'd like to evaluate this particular class of risk or whatever." Right. And from here, this is where we're identifying or envisioning two separate tracks. Okay, so one is this long term vision where there would then be some negotiation with the DAO where MKR holders through governance polls may approve or reject a particular risk team's proposal for what it is they want to do. |
| # / 00:37:18 | Cyrus Younessi | Well, that sounds nice in theory. We're also trying to support a less friction process through the Maker Foundation where prospective risk teams might come to us with their proposals, with almost like a contractors or are the grants program, and we can take it from there and then we can start assigning risk teams to various aspects, and then the voting process would still go through the standard governance framework. It would just be the ... maybe some of the overhead would be reduced, because as I said, one of our concerns is that risk teams might come along and they might look up this entire project up and down and just say that this is too much work in the early days. |
| # / 00:38:31 | Matthew Rabinowitz | Cyrus, one quick question that is related to that. Eval and there's the general process but how do we more or less "enforce" that with actors that don't follow this? |
| # / 00:38:45 | Cyrus Younessi | Well, I mean, that's part of the reason why we're putting this mandate up for a vote and to be officially ratified so actors who or not ... so the implication of this mandate passing is that actors who circumvent this proposal should not be legitimized and presumably this would be one of the roles of the governance facilitator. Does that makes sense? |
| # / 00:39:24 | Matthew Rabinowitz | Yes. |
| # / 00:39:26 | Richard Brown | Well, yeah, I'm not sure. I think there's going to be some more nuance to that. There needs to be a pipeline with which work comes into the system and there's going to be. And Cyrus, feel free to cut me off if I'm way off base here, but the risk teams in the future, we'll all be working together and bouncing ideas off each other and vetting each other's work and potentially we'll have meta risk teams that are just in charge of making sure that other risk teams have clean data sets. And we hear there's also been discussions about weighting mechanisms and all the rest. So I think that there's going to be a significant number of checks and balances in place to ensure that there's a certain level of rigor being adhered to. And I think that if the fact that some teams have just gone off the rails and aren't adhering to the basic principles outlined in the mandate, they'll probably get identified fairly early in the forums and hopefully there'll be some social selection in place. |
| # / 00:40:24 | Cyrus Younessi | Yeah. In general, I'm not envisioning that to be a big issue early on. I just don't see dozens of risk teams coming along and trying to do their own thing. |
| # / 00:40:40 | Cyrus Younessi | I think that the process that we've outlined is actually we try to strike a balance between specifics and keeping certain aspects sufficiently vague and as Rich has mentioned many times, we can't know how the governance process will look six months from now or a year from now and so forth. It's almost entirely certain that this entire framework will need to be updated and then put up for vote again with everything that we've learned. This document was designed to set a certain threshold of safety and responsibility for how risk teams might operate, but also leave some room for evolution and iteration. |
| # / 00:41:40 | Cyrus Younessi | So for example, in the section on the different types of risk teams, currently we're only identifying really just two types of risk teams, which are, one is just risk teams that submit risk parameters, which are obviously the most fundamental one and then teams related to oracles. But I could probably come up with a dozen different types of risk teams that don't actually submit risk parameters and these are all roles that we will identify as they come up and then I think we can officially formalize these roles. So when new risk teams come along, they say I want to be this particular type of risk team and there's a standard to go by. |
| # / 00:42:39 | Cyrus Younessi | Oh, another aspect that I missed in terms of onboarding risk teams with the Foundation bootstrapping initial risk teams is the notion of compensation. One of the biggest questions is always how these teams get compensated. Will there be stability fee funding? And those they're obviously great questions and I think we do envision stability fee funding from the DAO. Well, one reason is just, I think we're a bit time-crunched to be able to develop a full framework for that right now. I also think it's somewhat unnecessary in the first few months but that is a topic that you could expect to see fleshed out over the coming months and that's obviously a great example of something that could then get inserted into this document and then put up for a re approval. |
| # / 00:43:51 | Cyrus Younessi | Okay. Let me run through some of the questions in the forum and then and I'll just do a quick recap and then we can end the discussion there. Okay. So LongForWisdom brings up several concerns. One is on the timeline of when we're expected to vote in the stock. I think you guys spoke about this earlier on the call, is that right? I think that's when I got cut out. |
| # / 00:44:25 | LongForWisdom | Yeah, I've been talking about this for ages. |
| # / 00:44:29 | Cyrus Younessi | Okay. That's fine. There was a ... okay. What were some of your other questions? |
| # / 00:44:41 | LongForWisdom | There was one about general models and ones, what if we decide we don't like the general model. |
| # / 00:44:50 | Cyrus Younessi | Right. So changes and deviations from the general model. The idea is that once a general model is voted in, then if you wanted to build an applied model based off of that, you can't just go and deviate from a methodology that's already been voted in. If you wanted to do that, you could resubmit general model with changes and say, "Hey, I've updated my thinking on certain aspects or we've been researching this new class of financial models that we'd like to use." and go from there. But if you were to basically present a document that says. "This is how we're going to evaluate risk," and then turn around and do something else entirely, that's going to cause huge problems in terms of being able to keep the risk function legitimate. So, that's all I was referring to there, is that if you wanted to make changes, it really should go through another vote. |
| # / 00:46:04 | LongForWisdom | Yeah. So any specific changes in the general model is not changes for a specific collateral or a specific value. |
| # / 00:46:09 | Cyrus Younessi | Exactly, right. |
| # / 00:46:12 | Cyrus Younessi | Yeah, and I mentioned earlier, but one of the key things we want to avoid there is we want to avoid debates on the outputs of these parameters. We don't want to get into a flame war over this stability fee being too high or too low, but rather encouraged discussion on the methodologies and the models that went into determining those parameters in the first place. |
| # / 00:46:42 | LongForWisdom | Yeah, it makes sense. |
| # / 00:46:55 | Cyrus Younessi | Were there any other questions from the forum that I maybe didn't get to today? |
| # / 00:47:03 | LongForWisdom | I mean, my other one was about responsibility. That's where you said responsibility remain solely on the Maker token holders other than the risk teams. |
| # / 00:47:13 | Cyrus Younessi | Right. What I was going for over there was that ultimately MKR holders vote in all these risk parameters, right? And the risk teams cannot by themselves set parameters. All they can do is go through the models, go look at the data and output suggestions. I mean, the notion of responsibility, I'm literally just speaking from a community standpoint, nothing heavier than that. What I'm trying to get at is that it should be on the community to manage and oversee these risk teams and the risk models and the suggested risk parameters that they output, right? I mean, presumably risk teams could put on models that get rejected, right? We don't want to fall into this trap where we just always assume somebody else is doing the work and that everything's okay and I'm just trying to get at that there should be some scrutiny at all times and I think that can be achieved by defining where the responsibility for MakerDAO lies in the first place. |
| # / 00:48:37 | Cyrus Younessi | And ultimately that is on MKR holders. Risk teams are for now just almost contractors. Right? And we can talk about the incentives structure when that comes about and we of course would like to align those incentives so that risk teams do well when MakerDAO does well and vice versa. And in fact that's almost what I'm referring to here. Is that MKR holders should be cognizant that they need to make sure the incentives are aligned so that they don't fall into any traps. |
| # / 00:49:16 | LongForWisdom | Okay. Yeah, that makes sense. |
| # / 00:49:18 | Cyrus Younessi | Right. Akiva asks about a political risk question on the forums. |
| # / 00:49:26 | LongForWisdom | This is a separate thread, I think. |
| # / 00:49:28 | Cyrus Younessi | Oh, okay. |
| # / 00:49:31 | Richard Brown | Akiva, do you want to read other questions for us? |
| # / 00:49:34 | Akiva Dubrofsky | Yeah, my question is basically Rune has said on Reddit that if you get opinions for collateral onboarding from law firms then you can do it in such a way there's no legal risk but there's always those remains, political risk. I'm wondering what the foundation has done to help the DAO diversify its political risk especially away from US interests. So I'm wondering if you guys have pursued any collateral onboarding efforts in either Iran, Russia or China. |
| # / 00:50:16 | Cyrus Younessi | I would probably redirect that question to Rune himself. I haven't particularly spent too much time the more political aspects of onboarding collateral versus the more low level evaluations themselves. That's a really interesting question though and I haven't actually ... I haven't heard it that way before. Yeah, I don't know right now. I know there's a lot of efforts going on from various teams in the organization to help minimize any potential frictions there from the legal team, from the Biz Dev team and so forth, so yeah. Unfortunately, I don't have a good answer for you today, but I guess I will pass the forum thread onto other people at the foundation to chime in. |
| # / 00:51:24 | Cyrus Younessi | Okay. So let me wrap this up with a quick TLDR, just a one minute overview. The things to focus on here are risk constructs and our expectation for how risk models will look like and how they will be presented to the community. And then secondly is this notion of risk teams being approved through approval of the risk construct and how we'd like that process to play out in the forums, and that the Maker Foundation is going to be doing some initial bootstrapping of risk teams until we can get some more fully featured, decentralized frameworks in place. And that's basically it. |
| # / 00:52:28 | Cyrus Younessi | All right, let's move on and then if there are any questions later, we can get to them or please post them in the forums. Next on the agenda is monetary policy by Vishesh. It looks like we have one question. |
| # / 00:52:55 | Cyrus Younessi | Guillermo if you'd like to post it or get on the mic. All right. Maybe you can ask afterwards then. Vishesh, are you on the call? |
| # / 00:53:20 | Vishesh Choudhry | Yeah. |
| # / 00:53:21 | Cyrus Younessi | Okay, cool. I'm going to hand it over to you then. |
| # / 00:53:23 | Vishesh Choudhry | Yes, since we're near the tail end, I'll go quick. So big picture, I mean, ETH price has been down a bit lately. It took a sharp jump back down around the 15th. So what's happened is in collateral terms, so obviously in the total dollar value of collateral has gone down over the course of the past two, three weeks, somewhat significantly, but in nominal terms we can look at that same amount of collateral. It's been a little bit more flat. So just to touch on that a little bit, essentially with that drop in ETH price, there was also more collateral that was added into the system, which is as we've seen many times now a fairly predictable reaction. People get scared when price drops and liquidations occur, and so they add in a bit more collateral as a hedge against the risk of further liquidation. |
| # / 00:54:36 | Vishesh Choudhry | In DAI terms, I mean, the supply has been relatively steady around in between 76 and 78-ish million. And so we saw essentially a few draws and wipes, but essentially more or less balancing out over the last couple of weeks so that overall DAI supply has actually been relatively steady. Those two have a good graph on that but yeah, the DAI's price has been relatively steady during that time frame. |
| # / 00:55:14 | Vishesh Choudhry | What has been going on with DAI? So over the last seven days, essentially the trading volume has been relatively low, so averaging up to again around 2 million per day. In the last 24 hours, it was about 1.6 so these are relatively low levels of trading activity. The price has been sitting slightly above a dollar over the last seven days and same is true over the last 24 hours, so it was a little bit higher a few days ago. And then in the last 24, 48 hours has come down slightly, but again, still slightly above a dollar. So, all in all pretty stable even despite ETH making significant moves, so we have not seen a significant amount of re-leveraging with that drop in ETH price though in terms of, as shown earlier in terms of collateralization ratio, the risk has gone up a bit because that collateralization ratio has come down though the amount of debt has not changed significantly in that time frame. |
| # / 00:56:32 | Vishesh Choudhry | So yeah, all in all not a ton of activity with DAI and Maker despite a fair amount of activity with ETH, which is fairly predictable given that the amount of leveraging behavior has been comparatively low lately, so it's unclear whether some of those highly risky CDP owners have halted that behavior or if they're waiting for the price to come down even more before getting back in a leveraged level. So it will be interesting to watch and see what happens, but again, at the current levels the stability fee, it's been a point of discussion I think a bit on the chats. |
| # / 00:57:28 | Vishesh Choudhry | So just to touch on that briefly I mean, essentially the drawing and wiping activity has not been significantly impacted by the price movements in ETH. So, as has been previously discussed, that stability fee is effectively a modifying variable that impacts the relationship between what happens with DAI supply and what's going on with ETH price as an input. And so, given that that amount of sensitivity has been comparatively low lately, effectively it means the stability fee is if it were to be lowered slightly, I don't imagine it would have a hugely significant impact on the DAI price. |
| # / 00:58:17 | Vishesh Choudhry | There is a case to be made that given the lower levels of demands that it may be reasonable to lower that stability fee slightly. One other data point was so ... and I would encourage ignoring this spikes, so essentially the rates on dYdX did shoot up for a day and that was somewhat interesting, but to explain that in a minute. Essentially the rates the weighted average were and supply rates have been more or less steady. They've come down a little bit since some of that hype at the end of July with InstaDapp, with the changes in how Compound was working but has nestled back into that like, two to three point spread between the weighted average rate and their stability fee at least for the borrow rate, the supply rate has come down a little bit to know around 11, 12%. |
| # / 00:59:22 | Vishesh Choudhry | So this is only showing since beginning of July here, but that's basically a steady state supply rate when not much is going on in terms of borrow volumes for Compounds and even dYdX. The dYdX these rates are a little bit less predictable. So that borrow volume ticked up slightly over the past few days, but more or less has been pretty steady. The supply volume has continued to increase a bit although that increase has been a little bit slower on Compound than dYdX over the past week or so. In fact, you could say Compound has been more or less flat, whereas dYdX has picked up a bit. |
| # / 01:00:10 | Vishesh Choudhry | So what does that mean in terms of rates and excess supply, et cetera? So for dYdX, that excess supply had been continuing to grow but took a pretty decent dip. So that's where a lot of those, the rate jumps came from, is that dip in the excess supply, so essentially probably just a large refinancing or large movement or large play by a few players on dYdX, because those overall volumes are pretty low. So a few large power users can move those numbers around fairly significantly. |
| # / 01:00:49 | Vishesh Choudhry | For Compound there was also commensurate dip in the excess supply. But if you actually think about it in percentage terms, it's not that huge. Right? So we were at 9.6 is the peak there and it dipped down to about 7.6 whereas dYdX is dip down to about 2.3 ... oh sorry. Yeah, dip down to about 1.9 in excess supply around that same time frame. Again, this is just shown it a little bit clearer fashion, so the utilization rate on Compound is coming down quite a bit since that spike that we discussed to about 75, 78%, which is a pretty steady state for Compound I feel. And then for dYdX it had been holding steady in the 70s with one random-ish spike up to about 85%, and then again nestling back into relatively steady levels. |
| # / 01:02:00 | Vishesh Choudhry | So, what does this mean? It does mean that the buffer that exists on the secondary lending platforms is still just sitting there. But given what's been going on with the DAI supply, I think it's actually a more reasonable interpretation that through liquidations and through some people abandoning their positions, there was a decent amount of DAI that made its way onto the market that was actually probably picked up by traders, arbitrators, people that are looking to lend that DAI out for you in a relatively risk neutral fashion. I think that's where a lot of this excess supply is actually coming from, so it's curious what would cause this DAI to recirculate and be paid back into the loans that it originated from because if you draw the links out it has to come from somewhere, and so there are CDPs that are sitting out there open that are recurring interest on the negative area under this curve. |
| # / 01:03:16 | Vishesh Choudhry | It's going to be interesting to see at what point this actually circulates because clearly a 20.5% stability fee is not enough to prompt that and so that's what leads me to believe that these are potentially just completely capitulated loan positions where the CDP owner has written them off and has little to no intent to pay it back and may even just want to get liquidated to get out of the position. It's hard to really speculate as to what people are thinking with regards to CDP management. The point of that is this does not necessarily mean that the stability fee has to stay at 20.5%, and this is where sometimes we get into trouble by thinking a little bit too simply. |
| # / 01:04:09 | Vishesh Choudhry | But I know in the past we've really discussed a lot of this idea of, okay, well, if there's excess supply then it can reasonably be interpreted that that excess supply is oversupply and so if there is oversupply, then you should not lower the stability fee because there is DAI sitting out there that's not being consumed, and so demand levels are being outpaced by supply levels. That is potentially still true but I do think we ought to think a little bit differently about cases where that DAI is sourced from CDPs, that a CDP owner may have potentially given up on, and this DAI is now actually held by traders who have just picked it up on the market and in which case this may not actually necessarily speak to the "over supply level" or by extension indirectly the demand level. So we may have to get a little bit more refined our thinking here to really understand whether it makes sense to lower the stability fee. But my inclination is that it would not be out of the realm of reason to lower it by half a percent or maybe a full percent. |
| # / 01:05:30 | Vishesh Choudhry | Okay. So that was a quick hit. I will stop and pause for questions. |
| # / 01:05:44 | LongForWisdom | That's another question. I think just a point the executive for 18.5%, I think it's nearly passed. So it might still be going down by 2%. |
| # / 01:06:07 | LongForWisdom | Yeah. Or like 73,000 of 75,000 needed. |
| # / 01:06:14 | Richard Brown | That's interesting because there's been this an extreme reluctance to move that thing for the longest amount of time and I actually stopped watching. So thanks for the heads up. |
| # / 01:06:25 | Richard Brown | I've also have a bad habit of not waiting long enough through these awkward silences to figure out if there's any more questions but let's reframe those. So there's lots of things happening in the governance forums. Cyrus had to jump off, so unfortunately there's not going to be an opportunity to dig into the risk mandates, but I'm sorry, and this is what I always do as well, and I feel horrible about it afterwards. |
| # / 01:06:50 | Richard Brown | Matthew, do you have a weekly recap or a narrative that you wanted to go through or any other topics? |
| # / 01:06:55 | Matthew Rabinowitz | Not so much a narrative, but it's really a general question for the community/for the folks. I wasn't near at the very, very beginning when DAI first launched as a whole, but in what tools were used to allow for ... it's an open ended question of what was used to allow DAI to float at one when it first first first launched, where there were really immature market forces? Well, I assume the foundation took an active role in ensuring they then targeted peg of one and that's an open ended question that lead into the question related to multi collateral DAI because now the market forces are much larger out there to arbitrage and take advantage of that. When multi collateral DAI launches, there are a number of questions outstanding about how we migrate from A to B. Leave those aside for a moment. That's it, I guess, the foundation question and the community question. What was done in the past and how much of a role will the foundation be taking, I guess for the proprietary trading desk just to help ensure that it sits at one and doesn't severely deviate at all? |
| # / 01:08:04 | Richard Brown | Yeah, those are all good questions. I don't have answers to any of them unfortunately. So we can make note to follow up in the next call. But it's my understanding that it's not really the foundation's position or historically, I don't believe that management of the peg, it needs to happen at the market level. So if there are issues then we need to tweak the system as opposed to tweaking the market. I think, can you give me some more insight into why you think that things would materially change with the release of MCD? Why that additional management would be required? |
| # / 01:08:42 | Matthew Rabinowitz | Well, I mean, definitely just generically you go from system A to system B. System B has no market and you increase the supply in that market. Well, until you have matching demand for it. The price will be, there'll be downward pressure. Then the system A side as you're closing DAI positions, somebody has to buy it back from what is currently a stable system, so there's inherently upward pressure, the challenges in managing. System A is not so much of an issue as much as it is maintaining the peg on system B, while we're migrating over ... how do you call it? Such that we don't overload the multi collateral DAI sides in system B with too much downward pressure. |
| # / 01:09:27 | Richard Brown | Vishesh you- |
| # / 01:09:27 | Matthew Rabinowitz | That was the genesis of my question, was what was done at the very, very beginning in single collateral DAI because I wasn't there, so I don't know to help ensure it to be around one because whatever was done then needs to be done a 100 times more now. |
| # / 01:09:48 | Richard Brown | ... Yeah. I think that we'll probably somebody from the risk team or Vishesh, pretty much might have deeper insight into how to manage this, the initial launch of MCD. In my mind, I hadn't created a clear distinction between these two separate ecosystems as we release. I think that the [inaudible 01:10:09] seek some more clarity about what those actual mechanics look like, and how transparency will the movement from single collateral or DAI V1 to DAI V2, how quick and easy will it be? What tooling will be released to the community? What the level of appetite is for that migration? It's a brave new world that we live in so we've seen some examples of that in the past and I think the Compound has gone through the same thing recently. |
| # / 01:10:39 | Richard Brown | But from what I've noticed, I haven't done a the study of it is that the transition phase was fairly rapid and the migration will be too for them, it didn't seem like it had that many stumbling blocks. But I'm not sure from looking from the outside in, and I'm hoping that we'll see the same thing. It's really hard to predict. Attempting to turn off a legacy system or to reduce interaction with the legacy system is a tough one especially since there's more connections, more tendrils leading into the MakerDAO ecosystem than potentially into the Compound ecosystem. So it's hard to predict, this is something we've been talking about more and more as questions arise, as MCD becomes a real thing and people are starting to ask questions about these actual implementation issues. |
| # / 01:11:32 | Richard Brown | And we've seen in the past that at least with the Maker token, asking people to redeem and or migrate really surfaces the amounts of [inaudible 01:11:45] crypto that you have in the ecosystem. There's a lot of positions that I think that Vishesh just touched on. There's a lot of ... it's impossible to figure out exactly what the motivations for CDP holders are, and why some of these things have gotten dark and what their intentions are. It's going to be very hard to predict what the migration paths or the interest in the migration path is going to look like. |
| # / 01:12:07 | Matthew Rabinowitz | From an [inaudible 01:12:07]- |
| # / 01:12:07 | Richard Brown | So there's another- |
| # / 01:12:07 | Matthew Rabinowitz | ... yeah, to follow on that thinking, I mean, that goes to I guess more to the general advocacy on my end at least of not necessarily controlled, but at least a gated and throttled model so that we don't open up the flood gates to allow all of the DAI to be migrated at whim from one system to the other to try to at least stem the flow. I mean, just for nothing else. I mean, we've seeing makerburn.com and you can see how much Maker would be burned right now if single collateral DAI, if everybody paid back at something like 7,000 Maker would be acquired or burned off. Sounds cool except when you go look at unit swap, there's only 4,000 of it available of? Maker, which would cause some crazy price spike in MKR, which is not the objective. The migration, just to what extent it can be throttled and controlled, it's recommended. |
| # / 01:13:02 | Lasse Birk-Olesen | I think, can I give a comment here? |
| # / 01:13:02 | Richard Brown | Yeah. |
| # / 01:13:03 | Lasse Birk-Olesen | I think one mitigating factor is that the migration contract will also allow converting multi collateral DAI back to single collateral DAI, so if there's a large price difference, there will be an incentive to arbitrage between the two DAIs. |
| # / 01:13:22 | Matthew Rabinowitz | I did not know that, that is excellent. |
| # / 01:13:26 | Richard Brown | I didn't know that either. That's quite interesting. |
| # / 01:13:30 | LongForWisdom | Can I ask how that actually works because DAI has to be backed by CDP. So how does that contract work, does it create CDPs on one side to the other and close one and open another one? Yeah. |
| # / 01:13:42 | Lasse Birk-Olesen | Yeah. The migration contract creates one CDP where all single collateral DAI can be deposited as collateral and based on this collateral in MCD, the Multi Collateral DAI can be issued from there and vice versa. If you want to move back your multi collateral DAI into a single collateral DAI, the migration contract can release the single collateral DAI from the CDP. |
| # / 01:14:14 | LongForWisdom | That's awesome. Thanks. |
| # / 01:14:15 | Matthew Rabinowitz | So the single collateral DAI is a defined collateral of multiple collateral DAI. |
| # / 01:14:25 | Lasse Birk-Olesen | Yes. That's the proposal. |
| # / 01:14:35 | Matthew Rabinowitz | Okay. |
| # / 01:14:40 | Richard Brown | All right. I think that's ... yeah, those are great questions. I think that this, like I said, as MCD becomes more and more of a reality trying to hammer down exactly what these mechanical steps are going to look like becomes more of a pressing issue. And so we're going to be hearing from the integrations team and Lasse's team a lot more in the future and the product team as well once we start really seeing information about what the actual UI layer tools or the smart contract layer tools ... excuse me, are going to look like to help people get into the new system and the redeeming tools to convert single to multi, single collateral DAI or DAI V1 to DAI V2. As that tooling gets released to the community, we'll have a greater sense of what those paths are going to look like. And I know essentially the appetite for people to move from one to the other, which is going to be interesting to see. |
| # / 01:15:39 | Richard Brown | All right. Do we have any other questions? We're at 20 minutes past the hour. I think I am going to take advantage of this awkward silence and maybe put an end to the call. If people have questions, please continue the conversation in the forum thread. Actually, I didn't post the link to that yet. Actually, let me do it now. |
| # / 01:16:02 | Richard Brown | Limited utility possibly. Well, people should be familiar enough with it. So head at forum.makerdao.com And at the very top there you'll find a link to the discussion thread specific to this call and in the next hour or two we'll be posting links to the audio and the video. Summaries will be coming soon, and that's the place to continue the discussion and ask questions that weren't answered. |
| # / 01:16:29 | Richard Brown | I think that is it for today. Thanks for LongForWisdom for the questions, Vishesh for your presentation, Cyrus for your risk team mandates and Matthew for the hard questions as always. |
| # / 01:16:40 | Richard Brown | All right. Thanks everybody. See you next week. |
| # / 01:16:48 | LongForWisdom | Thank you. |