Ep. 36 May 30 - 2019¶
Video | Audio | Discussion
| References | Person | Text |
|---|---|---|
| # / 00:00:00 | Richard Brown | Hello everyone. Welcome to the May 23rd edition of this scientific governance and risk meeting. I'm Richard Brown. I never introduce myself on these things. I'm the head of community developments at MakerDAOour We have, an interesting call today. We have a guest from a product team and Chris, correct me if I've gotten your department wrong, I'm not entirely sure where it is but Chris [inaudible 00:00:26]. Is going to join us and talk about something that's near and dear to all of our hearts. And that is how the governance portal works now, but more so how it's going to work in the future and solicit some ideas. So that's going to be interesting for us. Steven is going to have a chat about how governance is an ecosystem and then we're going to hand it off the risk because, we've been sort of short changing risks in the last few calls and there's a lot to go over there. |
| # / 00:00:53 | Richard Brown | So I'm not going to monopolize or speechify today, which I'm sure lots of people would be happy to hear no more, no less so than the people in charge of the transcripts. And do I ever talk a lot? one thing I'd like to remind people is, we're very, very interested in community feedback. The reason we have these calls primarily is to hear from the people that join these calls. So if you have questions, you got comments, if you have suggestions, please do not be hesitant about sharing those with us. [inaudible 00:01:25] them in the chat at the site if you don't have access to a microphone, if you do have access to a microphone and please feel free to jump in and interrupt us at any point. I'm going to paste in a link in the chat to our sub Reddits. |
| # / 00:01:40 | Richard Brown | This is where we continue the discussions that we start in these calls. So in these calls, we don't make decisions like this is not a governance committee. This is where we debate things we talk about important current events, memes, misconceptions, new trends in the ecosystem. We debate those a bit and then we continue the discussion in the sub Reddit for the next week and then we start making decisions around those things. I think that is going to be my preamble. And I'd like to cut myself off there. So Stephen, I'm going to hand this off to you. |
| # / 00:02:15 | Steven Becker | Thanks Rich. First. Hello everybody and welcome. Thank you for joining us today. Before I get started, the one to thank everybody that approached me during the New York blockchain week. it was really, really quite cool to meet all you guys. And there was one common question I was getting at the NY blockchain weekend. I think it's very important and that is, what the hell is that painting in the background? So that I thought for a second I'll just get out of the pictures. There we go. |
| # / 00:02:48 | Richard Brown | I wanted to ask that for six months. |
| # / 00:02:51 | Steven Becker | Everyone can see it I think [inaudible 00:02:53] and [inaudible 00:02:55] specifically and I'm very happy with the little pause. But anyway, we're here to discuss something in a very interesting in repect to governance. There's an ecosystem that's developing around each stakeholder to the MakerDAO protocol. And obviously with respect to CDP users, with respect to DAI, DAI is effectively the onchain economy, liquidity providers but more so governance. It's I think it's one of those elements where you have a look at the scope of governance and asked about plain and pointed questions about the how do you organize yourself, how do you go about addressing certain issues and addressing certain items that you wish to consider as a group as opposed to as an individual? Now, to that point, what I would like to do is encourage people to self-organise and develop an ecosystem around governance. |
| # / 00:03:48 | Steven Becker | I think it's very important for this interaction to be consistently there as I think it actually underlines the continuous approval system that is in place. So I think it does go hand in hand. So I would really welcome the community to explore everything from that organization through the tooling and considering dashboarding, that can actually pertain to governance. Also keep in mind that any improvements to the tooling or dashboards or processes that you think could benefit to the MakerDAO protocol. Please talk out and please approach Rich because we do have a grants program that can help with that. But the one thing I do want to reiterate is that no matter how you self-organize ecosystem and how you wish to change your doing or provide tooling or dashboarding for the process, the one aspect could keep in mind is that the online scientific governance stays in place. |
| # / 00:04:56 | Steven Becker | The exposure to a plutocratic system is inherent to any kind of governance. But scientific governance with respect to MakerDAO helps at least to the extent possible minimize the risk of whales influencing the system. But given that I need to go through the normal and typical disclaimers. The first one is that in this meeting, we still in this meeting, we still have three things that we consider the demand and supply imbalance, collateral types, risk parameters and portfolio. And obviously the third one, being exogenous risks. This is decentralized scientific governance that we are trying to altogether facilitate and actually work on it and provide momentum behind. So, that's what this hour is dedicated towards. And also finally, this is the constant reminder that governance is a continuous function. So please be involved as much as possible. So with that, I'm going to hand back over to Rich. |
| # / 00:05:56 | Richard Brown | All right. Thank you Stephen. That's, I think that there's an interesting dovetail here too that we can think about it because, governance from my perspective, and this is simply my perspective, it's not make your perspective, but from my perspective, it's a com... It's an extremely hard problem. It's a capital H, hard problem. And, basically every discipline you can think of including, the crypto space. And one of the things that I personally have been wrestling with for a long time is, how do we improve the mechanisms with which we use to conduct governance and MakerDAO and how do we rely on the ecosystem to fill in some of those blanks? |
| # / 00:06:38 | Richard Brown | And, I don't want to get into too much a philosophy here, but there's an opportunity for us to continue with the same model that MakerDAO sort of used with most of his tooling is that the foundation creates crypto primitives and those primitives are designed to be picked up by, the rest of the ecosystem. And the DeFi space puts those things together and various combinations. And then we have an economy and it's, that's an amazing thing to watch. There's a potential here to look at governance from the same perspective. So we have a layer one version of governance and that, oh, it's called layer two, which is a portal where people can vote. And then there's all kinds of challenges that we see with, stakeholders and we've talked about those challenges in the past, but, apathy and turnouts and people not knowing whether MKR makes a difference or not knowing whether a and sort of casting about for is, can I just get my MKR to somebody else who does know how to vote and isolate delegation mechanism. And as Stephen pointed out, there's an opportunity for the ecosystem to step up and start answering some of those questions for us. And we're really interested to see that happening. |
| # / 00:07:49 | Richard Brown | Which also provides me a really great segway introducing Chris Bradbury, who is in charge of figuring out the answers to a lot of these questions and providing us with, the perfect solution, which I'm sure is coming very, very soon. No pressure Chris. So, why don't you introduce yourself and then we can talk a bit about what we can expect in the portal in the future and how you guys are approaching them. |
| # / 00:08:14 | Chris Bradbury | Yeah. Cool. Thank you Rich. As you said, yeah, I'm Chris, part of product team here, ultimately responsible for governance portal and the CDP portal too mainly, yeah, I might even, I said, well, like probably more than aware that governance portal has been around eight months. I think since we had our first foundation principles and we haven't really touched on it since we've released it really that day. It was simply designed for that governance proposal, the foundation proposal, which was a yes or no vote. And over this time it's, it's kind governance as evolved a lot and we haven't really kept up with that on their dashboard, mainly through like MCD being a priority on this. We were going to start looking at it now. And really what this, what this is about really is actually just to request feedback, from all of you who are using the governance portal every week, voting. |
| # / 00:09:07 | Chris Bradbury | And ultimately it's what do you want to see? I don't want to go into details now, or necessarily hear all the feedback on this call. but what I'm going to do is I'm going to drop a post on Reddit, just after I finished, basically just outlining our plans, kind of some of the background and what we want to do moving forward. But just to give you a general idea of what we're looking at here is you've probably seen over the last few weeks that we've all seen it expanded polling from a couple of votes or yes, no vote in the very first vote to a couple of different options. From a percentage increase point of view to more recently when I was becoming more stable to 10 or 11 options. And the dashboard is ultimately completely broken. |
| # / 00:09:47 | Chris Bradbury | We've got a long list of options and it's difficult to actually find or see all the options in one place. So the first step we're actually going to do is we're going to disconnect polling from executive voting from kind of behind the scenes in the sense that we are at the moment polling uses DS chief and it uses the same smart contracts under the hood basically. So we're going to take that out. And with that we're going to redesign the entire UI. We're obviously super like, have a lot of knowledge of how governance works under the hood and I don't think we're best placed to decide how that new UI should look, how it should work and so on. So, yeah, I want to hear from all of you really. What do you like in the current one? What don't you like in the current one? What would you really like to see? The new version of polling we have quite a few of the requirements for and we have a general idea of what we want to do with it, which involves ultimately just emitting events as opposed to actually voting on a particular contract. |
| # / 00:10:46 | Chris Bradbury | And with that, that expands the possibility of bringing more than just MKR and so on to what kind of sentiment is out there. Ultimately, the goal though is to bring more people into governance and the governance process. We have, I think it's 8 or 9,000 unique MKR holders out there. and recently we had 194 unique addresses partaking in governance. So we want to grow that exponentially, kind of within the year really and, and keep that growing more and more. And the more MKR we can get voting or more unique addresses really we can get voting is the kind of measure of success for us really to get someone to drop a summary in the mkrgov sub Reddit. |
| # / 00:11:29 | Chris Bradbury | Please feel free to jump in the discussion there. I can't guarantee that all of the kind of feedback will be implemented into the new design. But what we're going to look forward really is patterns in what people are asking for and what people want and where those patterns, we'll put together some prototypes and then hopefully in like four to six weeks, come back to this call with some prototypes and we'll kind of do the same sort of process. We'll talk about them. We'll see what you think and yeah, hopefully get as many of you are involved as possible. |
| # / 00:12:00 | Richard Brown | That's sounds interesting. |
| # / 00:12:01 | Chris Bradbury | [crosstalk 00:12:01]. |
| # / 00:12:02 | Richard Brown | Well, yeah, I probably, I'm going to be the person that jumped in there because I, I could potentially ask you about 27 different questions about governance and the future of the portal looks like, and maybe that might be too much for this call. But I want to touch on something that you did mention. You said moving from away from the chief to events. And one of the things that, one of the questions I get from the community frequently is where do I see the data for a previous governance activity? And my response is that's where we have a few different tools. The community is coming up with more, dashboarding is coming, but ultimately that's the third party ecosystem thing that we want to foster. And obviously the next question is, okay, well where do I find that data? So in this new model that you guys are planning or architecting, will people still be able to find all the relevant activity on chain or are they going to have to go to an API to get that? |
| # / 00:12:56 | Chris Bradbury | Ah, nice. So that's one of the crucial, requirements. We have been looking at this. Everything will still being on chain. So you'll still be able to analyze all the events and see all the events. And I'll see what is being voted for. And yeah, and hopefully we're going to bring a lot more of that history into the new version as well. That's certainly one of our plans. Certainly a more detailed breakdown of what people are voted on. Yeah, everything's staying on chain. It's still going to be fully auditable and [inaudible 00:13:22]. So, yeah. |
| # / 00:13:27 | Richard Brown | And what are the major highlights of the new features that are coming out. So obviously we have some restrictions in the existing model and we've actually run into those, in the last week we knew that they were there and this is the first time we kind of like hit them head on. And one of those is that, we, the executives have a very continuous approval voting means that we can have a potentially a week executive vote sitting in the system that never gets ratified or it doesn't get ratified before the next poll starts, which is a problem. We're running into a new problem this week where the new executive is not going to get ratified until prior to the newest new executive coming into the system. So I'm actually not entirely clear on what the mechanism there. I'm assuming we just swap out the old unsuccessful executive with the new unsuccessful executive. Is that true? |
| # / 00:14:18 | Chris Bradbury | Well actually I think it's probably a discussion point for this call actually say kind of on the specific of the executive vote. I think it's the executive vote is the same as the winning poll this week. So because it's continuous approval voting. There is no dates or anything attached. So ultimately we can just leave the executive vote carry on basically at 17.5%. So really suggestion there is just the need that as it is onto your point of view, yeah. Multiple votes happening at one time. That's, that's the primary driver for changing it right now at the moment. Yeah. We can only have one vote going on realistically at any time, whether that's polling or executive with a new system, we can have as many polling votes happening at any time and you'll be able to vote on five polling votes at, at once if you'd like to. |
| # / 00:15:02 | Chris Bradbury | You're still going to be able to vote for one option within each poll, but you will be voting yes. You could have, for instance, like a cadence polled at the same time as having a stability poll. At the same time, there's maybe proposing new collateral types to add a to the system and it won't matter if they're overlapping at all. You'll be able to go in and you'll be able to read them. You'll be able to vote. Likewise, if there's something happening in the core system on the executive side, you'll still get to partake in that with the same knowledge that there's no MRK one to the other like we saw in the last week. |
| # / 00:15:36 | Richard Brown | Well that's okay. I assume a lot of the complexity comes into things because right now it's, it's dead simple because the MKR gets locked and we know where that MKR is. We know who owns that MKR. What mechanism you guys are going to be using to help people voting on like five votes at a time plus an executive? Is there some kind of an IOU token involved or snapshotting or? |
| # / 00:15:57 | Chris Bradbury | Ultimately, snapshotting. So we'll probably take it. It's not been fully decided, but I suspect we'll take the, the kind of final block at the end of the poll as a snapshot of everybody's balances. And when you're in a venue alternate, just admitting your address, but you're voting with in the option you want to vote for. So at that block where the poll ends will then take the balances if of all the available, like known contracts that we have plus your wallet balance. So if you've got MKR locked into a vote proxy and locked into chief and locked in and just held in your wallet, we'll actually sum all of those balances up and add that to the running total of the MKR, on that poll. One of the crucial things for us as well though is to actually enable the UI for people to be able to see kind of what that tally might look like, unless of course there's these options where actually it might be better to have hidden results until the end where you don't actually know what's, what's going on until the end. And I'm hoping a discussion forms around that point. But yeah, we'll be using some of them and archive made of some sort to to keep tally, for the UI if we need to. |
| # / 00:17:02 | Richard Brown | Well then that leads into another question I was going to ask. So are you guys primarily looking at things from a user experience perspective and additional feature says perspective or are you also contemplating things like putting requirements or refining how governance works or you're talking about thresholding and upper and lower bounds on particular types of votes or medians between different polling options. Are you looking at any kind of deep mechanisms like that or is it just UI UX. |
| # / 00:17:30 | Chris Bradbury | At moment is, it's just more of a UI, UX thing from a goal really of just increasing the amount of people coming in. I think one thing we learned a lot from the last vote was that a, or the very first vote, the foundation appraisal is we don't really know what's going to happen in the next 12 months. I think some of the requirements we had that it would only ever be "yes" or "no" votes in polls as one of the first things we talks about. And that that quickly changed. Unfortunately, the system we used enabled us to be able to do that. But like if you ever look at chief now there's just a ton of a blank spells basically in chief that just say "signal your support for x". They just getting used to that three day period and then never, never used again. |
| # / 00:18:11 | Richard Brown | Like one of those additional requirements that was specifically around yes and nos. It's become apparent that one of the major use cases for the portal was picking a range of numbers in an easy way. Is that something you guys were playing around with like paradigms for? |
| # / 00:18:26 | Chris Bradbury | Yeah, I mean at the moment that kind of very short term solution to get away from this issue with the dashboard, the kind of long dashboards, it's actually, we're just going to have the kind of headline title on the dashboards. and hopefully this change will happen in the next four to six weeks as well, but kind of ultimately the headline title on the Dashboard and you'll click into it to go into the actual proposal and instead of having 11 proposals, they will simply notice be one proposal and something like a dropdown to choose which way you want to choose. And then- |
| # / 00:18:54 | Richard Brown | Yeah. That'd be nice. |
| # / 00:18:59 | Chris Bradbury | And slightly longer so in [inaudible 00:18:59] redesign that we're looking at, we're open to all sorts of options. I kind of one of the things I think about is even when we create the poll is actually able to choose the input option type to wheter that's a dropdown or- |
| # / 00:19:12 | Richard Brown | A slider |
| # / 00:19:12 | Chris Bradbury | ... Or whatever, whatever you, whatever you think like, that's, that's where the feedback's going to be so important because I think we've got some ideas Rich, and then you've got a lot of ideas, but we're so deeply ingrained in this. I think, the most important ideas are from people that just use it, but ultimately twice a week, like Monday to Wednesday, then on a Friday, ideally. So, yeah, it's just a cool way to get involved in much possible in the next two weeks. I think there's been a couple of questions about wireframes, and so on in the thing and realtime board, and so on. we're actually using a [inaudible 00:19:48] at the moment which we can share publicly. So as we move on we can drop some things [inaudible 00:19:53] and then, yeah, I can add things into Reddit or drop them to Rich to show the community. |
| # / 00:20:00 | Richard Brown | All right, that sounds like a good place to wrap things up. So let's make sure that we clarify what people can do to get involved. So you're going to be posting something to Reddit soliciting feedback and is there going to be a form or something where people can interact or is there going to be adding stuff to that thread or how do you want to solicit those suggestions? |
| # / 00:20:19 | Chris Bradbury | But this, for this one, it's, it's more, I'm looking just for discussions within the threads in Reddit. well I'd quite like to see is it's kind of background in evidence, not evidence such for a reason why you might want something. So just to kind of two words of saying like bigger buttons, it's probably not that helpful to us but yeah, kind of a reason why and then ideally where a discussion can take place which will help us understand the actual demand and reasons for actually doing it- |
| # / 00:20:45 | Richard Brown | That sounds good. So we can iterate on this thing a bit and then maybe get you back in a couple of weeks to talk about next steps or what the progress looks like. |
| # / 00:20:54 | Chris Bradbury | 100%. Yeah. And then if there are any individuals that would be up to do like a one to one session after these next two weeks, then that would also be great as well that we can take individuals through the portal. Actually let them use it, they'll use a prototype and kind of point out in individual bits that they maybe find confusing or they can't actually see what to do. |
| # / 00:21:15 | Richard Brown | All right. Sounds good. All right. Anything else you wanted to leave us with or are we good? |
| # / 00:21:19 | Chris Bradbury | No, that's something, that'd great. |
| # / 00:21:22 | Richard Brown | That's awesome. It sounds like you have some work to do back there. I'm going to let you go. Thanks Chris. |
| # / 00:21:24 | Chris Bradbury | Yeah, cheers, thank you. |
| # / 00:21:28 | Richard Brown | All right Cyrus, I'm going to let you pick it up because we ate up another 25 minutes and we need to get into risk. So off to you. |
| # / 00:21:36 | Cyrus Younessi | Sure. Vishesh are you on the line to go through the peg? |
| # / 00:21:43 | Vishesh Choudhry | Yeah, I can do that. Okay. so just as far as what's been going on lately, let me just start with DAI actually. So what's been interesting because this is a new- |
| # / 00:22:11 | Cyrus Younessi | Are you sharing your screen? |
| # / 00:22:12 | Vishesh Choudhry | I am. Are you not able to see it? |
| # / 00:22:16 | David Utrobin | We do not see it. |
| # / 00:22:44 | Vishesh Choudhry | Okay. Are you able to see that? |
| # / 00:22:50 | Richard Brown | No, you might have to kill one of your monitors. |
| # / 00:22:53 | Vishesh Choudhry | Damn, that's the saddest thing I ever heard. |
| # / 00:22:58 | Richard Brown | I'm imagining you with like a 25 monitors set up or something. |
| # / 00:23:01 | Vishesh Choudhry | But it's like not wrong. I'll just kill the Safari entirely. It's not just multiple monitors, it's multiple browsers. |
| # / 00:23:17 | David Utrobin | Do you have the option to like share window when you click down there? |
| # / 00:23:22 | Vishesh Choudhry | Yeah, I was, was doing that. Sorry for the technical delay. Okay. How about now? |
| # / 00:23:40 | David Utrobin | Negative. |
| # / 00:23:43 | Vishesh Choudhry | Highly dissapointing. |
| # / 00:23:50 | David Utrobin | I noticed your graphs got a new URL. |
| # / 00:23:53 | Vishesh Choudhry | They did. |
| # / 00:23:56 | David Utrobin | [inaudible 00:23:56]. |
| # / 00:23:58 | Vishesh Choudhry | Yes, I updated the URL so it's less... a little bit more official. Okay. I'm just going to unhook the monitors here, back to doing it old School. Apologies for that. I don't usually have an issue. |
| # / 00:24:21 | Richard Brown | I was kind of sorry to see that we lost the visa Vishesh branding. |
| # / 00:24:30 | Vishesh Choudhry | He still got me on Twitter. All right. How are we doing? Can you see it? |
| # / 00:24:41 | Richard Brown | No. It might be thoroughly confused. Maybe if you like leave and then- |
| # / 00:24:44 | Vishesh Choudhry | Reload. |
| # / 00:24:48 | Richard Brown | Did he just drop off? We could do other things while we wait though. Matthew, did you want to have a chat today? |
| # / 00:24:56 | David Utrobin | [inaudible 00:24:56]. |
| # / 00:24:59 | Richard Brown | All right, sure. |
| # / 00:25:02 | Matthew Rabinowitz | Yeah, so basically, this is just kind of a continuation from the call I guess from last week where I had some people ask me, let me just go ahead now I'll try and present, let me know if that works. Anybody? |
| # / 00:25:21 | David Utrobin | Nope. |
| # / 00:25:22 | Matthew Rabinowitz | All right. There's an issue with Google then. |
| # / 00:25:25 | David Utrobin | Yeah. Likely. |
| # / 00:25:26 | Matthew Rabinowitz | Yeah. So this was more related to, this was more related to the iterations between the stability fee in the DSR. Some people kind of reached out to me and asked basically, kind of fast forward the video where we are today before a DSR and where we'll be in two, three, five, seven years, kind of what we're going to look at it. It's manually we're going to be having a from a governance level, we're going to be going back and forth on what happens when you lower the stability fee and trying to identify the signal of lowering the stability fee than increasing the DSR, and iterating over and over and over. Basically a human version of a PID. But without the I and without the D just the P, the proportional. Basically if DAI is above one, what do we do? And building out that matrix of decisions that we all basically as a group agree is the fundamental decision making you should do to help stabilize DAI back down to a dollar. |
| # / 00:26:32 | Matthew Rabinowitz | And the interesting thing becomes, as we get through that point, there will be, whatever that turnover point where the DSR kind of ask about asymptotes hits a derivative of zero where it starts to arc over where would we find the, not necessarily the sweet spot but the point where we have a meaningful direct change, a controlling change in the amount of DAI that's issued. Once the DSR we can start to lower it and also also at the same time also bring down the stability fee at the same time. Once we get to that window, we're going to find, well this iteration process we'll continue basically bouncing back and forth. This again is not a week of work. It's not a month and it's not a year. It's a couple of years to get that. The challenge again becomes this is still just the P. |
| # / 00:27:24 | Matthew Rabinowitz | All of that is just human inputs and like if we can imagine, we can all probably draw it on a sheet of paper there, five, nine, 10 scenarios that we can all fundamentally agree with. What is the logic right now? Question is right now is how do we identify the difference between going between 19 and a half percent to 17 and a half, but why wasn't 17.37 picked as an example? Right, and it's just right now we're just using, yeah, logical guessing, reasonable logical guessing to determine that and a lot of those decisions are going to be outside the scope of what we as people can identify. |
| # / 00:28:03 | Matthew Rabinowitz | We have something directly unique that central banks don't have any access to, which is we have constant immediate access to the amount of supply that is issued and what happens to it on a minute by minute basis. Same thing. Whenever we have a DSR, we'll have the reverse of that of knowing how much is locked up and what that does to the price and all that data is, kind of mentioned this on the last call, we're going to want at some point implemented PID. The question is really where and some of the challenges in that kind of chatted with Vishesh about this offline also would be ... |
| # / 00:28:40 | Matthew Rabinowitz | Some of this is very similar to the way, like we mentioned in autopilot, auto driving car or truck just generates an amazing amount of data for 25 hour drive in or 24 hour drive, like five terabytes of data and they have to, we need to ultimately find a way to rip through that data to do some type of deep learning to identify what would be the best ways. Over time, this is not a small engagement to identify. Imagine we had a scenario, let's call it, one or call them at our hundred scenarios or more where as an example if DAI is above one, but it's trending up with a slope of X. But the supply is coming down with a slope of Y. What historically have we done, what historically do we recommend? And then how do we get that information into a blockchain. So blockchain isn't going to process that. |
| # / 00:29:36 | Matthew Rabinowitz | Or more specifically it is, it could process it, but it's way too slow and burning gas to do that would be way, way, way too expensive. So that kind of deep learning we're going to need to be able to iterate on an optimized level is how we're going to ultimately pull this target sweet spot, the overall stability fee down from ... It'll go below 10 and we have the DSR, but getting it to the plant and to stick below four, five, we're going to need algorithms to help us get there. The challenge again will then be how do we set it up such that can't, the system can't be gamed, can be gamed with how do we set up so they can't be attacked and maybe the answer is ... and these are just mean speculating, again, I'm not stating what we will do or what we must do, this is ignorance is bliss. |
| # / 00:30:28 | Matthew Rabinowitz | The concept of just external computation and how we bring that ID component on chain probably through some type of oracle's distributed Oracle's same way we do with pricing. Anyway, that that was the main takeaway from that, some people asked about was how many iterations would it take to get there. And will that be as simple, smooth wine and absolutely it won't. And we will absolutely joyfully hit the black swan that nobody expects. And what scenarios, how much of that will manipulate our data. |
| # / 00:31:06 | Richard Brown | Well that's a good point. Thanks for that Matthew. That was great and about how willing we are to trust algorithms and what that transition looks like. And I think that as a group perhaps and Cyrus, anybody can correct me if I'm wrong, but we're moving towards this picture where, manually we had to do things manually until we understood basically the broad strokes of how managing risk primers work. It's inevitable that we need to have an algorithm, a somewhere that's crunching a great deal of data and providing some signals for us. And that will transition into this stage where we vote on parameters based on what the models are telling us in any one moment. And then potentially there's some point in the future we'll be able to take our hands off the wheel a bit and let those models or sandbox those models. But let them basically drive some of the decision making process. Is that accurate Cyrus or is that still kind of hard work? |
| # / 00:32:02 | Cyrus Younessi | Maybe in the super long term. I mean I think using models is fine, but I'm very, very hesitant to have any sort of on chain algorithm set our parameters for us. |
| # / 00:32:13 | Richard Brown | Even if they were bounded like very tight one? |
| # / 00:32:16 | Cyrus Younessi | Yeah. I mean I think we need that human gut check for number of years to come in so many ways it can go wrong. |
| # / 00:32:26 | Richard Brown | Yeah. That kind of ties into this larger, I guess a philosophy and as oppose that ultimately this needs to be governed by humans. That needs to be the final arbitrator. And so we're balancing engagement with the system and potential apathy and all of the concerns about plutocracy and all the rest of those things that go along with governance. On one side that's one of the major cons and the other major con, the side of machines don't care about humans and they'll do whatever machines were told to do. And so having an algorithm go haywire is ... |
| # / 00:32:56 | Cyrus Younessi | I think there's also manipulation aspect to it, right? If you know exactly how the algorithm that's setting your parameters works I'm pretty sure you can back out the cost to manipulate it up to a certain number. |
| # / 00:33:12 | Richard Brown | That's an interesting ... there's precedents here too, and it happens a lot with different organizations I speak with, is that people come up with an algorithm or they come up with a bot or a keeper or some kind of mechanism, and then they're successful for about three months until somebody, until a ninja notices that there's chances/opportunity for either exploit and/or profit and then they rapidly get pushed out of that market that they've created for themselves. And I think that same model applies here, that even if we did have a complex algorithm that we liked and was capable of managing the system, it's impossible to predict that black swan genius that might come in and start playing with our inputs. |
| # / 00:33:53 | Cyrus Younessi | Yeah. Or maybe something as simple as a human veto as Josh said in the chat could be good enough. |
| # / 00:34:01 | Vishesh Choudhry | I might suggest maybe slightly the opposite of a human veto, where it's just the human pushing the button and then there's sort of a, maybe a recommendation is a good spot to work towards, right? Where we have an algorithm that's recommending a certain change or a certain amount and that's just like an input that we take into account in pulling the trigger at least for like the medium term, I would say. Maybe longterm we can really get to kind of this point. But to me the whole PID model is more of like an ideal to work towards as like a broader process. So like with the PID thing, I think there's, there's two like major problems that we have to solve. One is what are the right inputs and then two is how do we properly respond to those inputs? |
| # / 00:34:44 | Vishesh Choudhry | And a PID algorithm is like mainly about how do you respond to those inputs? I think we are squarely in the stage of like discovering what those inputs should be. Even from the process of like me manually running 20 models on the weekend. Like it's, it's really hard to figure out what the right variables are and you kind of have to explore and investigate. it's a really, really human process. |
| # / 00:35:07 | Cyrus Younessi | Actually. I don't want to put you on the spot Vishesh but do you want to talk a little bit about that, about your research into the right inputs and how difficult that's been and maybe some of the challenges. |
| # / 00:35:20 | Vishesh Choudhry | Yeah. I haven't published more about this because it's been a more difficult and I haven't gotten to like really good concrete answers and I want to like make sure I get to those before I start sharing stuff. But the basic ideas I got asked a lot like what is driving DAI price? What's driving, what's the impact of the stability fee? Really like seemingly simple questions. and the answer was often "it depends". Right? And the, the latest like bull run is a or a crypto spring as my poll is indicating, is really kind of challenging a lot of those assumptions and relationships. Like most of the correlations that I found running up through, end of April, completely broke in May. And that's something that we really need to take into account is and what extent are the environmental conditions that we're envisioning, like in terms of what's going on with crypto in general, what's going on with Tether, what's going on with ETH price, just how impactful they are for like what happens to the system. |
| # / 00:36:24 | Vishesh Choudhry | We can't simply look at the correlation between stability fee and DAI supply. you can get to an okay correlation if you like, filter out a lot of those external factors, but that's just not realistic. So yeah, it's extremely challenging to find good, regressions with solid, our squared values that like seem to have random residual clots because there are underlying patterns there that are just not being taken into account because there's variables not taking into account. And so like for example, I had initially, like my first, first stupid rough pass was to try to run a lot of regressions on ETH price and DAI price and like I was getting okay correlations, but the point was I was really missing this concept of like DAI price is not a monolith and I was missing this concept of ETH price is relative, right? To people's perception. |
| # / 00:37:22 | Vishesh Choudhry | So just because ETH price is low, it doesn't tell you anything, is ETH price low relative to where people expect it to be long term? That's a really different variable that has a stronger correlation. And then, it's still an iterative process. Right? So I've gotten to better correlations now than I did two weeks ago, but there's still not good enough in my opinion. |
| # / 00:37:41 | Cyrus Younessi | Yeah. I think we've talked about this a couple of weeks ago. Right? Trying to forecast the demand for leverage at various price points. |
| # / 00:37:48 | Vishesh Choudhry | Yeah. And not even forecast, but just explain current levels. and then forecasting is like a level up. |
| # / 00:37:54 | Richard Brown | We have about 15 minutes left Vishesh, do you want to try a screen share? If we are all manually just click on this screen share, it works, but it's not going to take over for us. |
| # / 00:38:04 | Vishesh Choudhry | Yeah, I think, and I don't know if Google updated something, but there's another screen that pops up which is me as well, but my presentation. And so if you click on that, you should be able to see the screen. If anyone has any problems with that, let me know. |
| # / 00:38:20 | Richard Brown | [crosstalk 00:38:20]. |
| # / 00:38:21 | Vishesh Choudhry | Okay. So I'll try to go through this a little bit quicker. So basically this is a new view that, for most people won't have seen, in terms of a breakdown of what's been going on with DAI Price, in the last 24 hours. And I've kind of been checking in monitoring this, over the past week or so. I think broadly DAI has been hovering just above and just below a dollar, right? Which is to me, proper functionality for DAI. There's always, there should always be, I think this kind of bell curve distribution above and below. |
| # / 00:38:56 | Vishesh Choudhry | I think you almost always see, kind of skew where there's more trades below a dollar than above a dollar, even if it's like potentially centered on a dollar. And I think that makes a lot of sense, particularly in terms of what we were talking about, this being a fairly asymmetric system where when DAI starts trading above a dollar, it's very quick and easy for the system to arb it down a little bit. And so there's always going to be kind of this more dramatic cutoff above a dollar, but broadly, you can see the volume weighted average price in the last 24 hours is like a dollar o two. So, that to me is broadly reasonable. Some people have kind of commented the most recent trades sometimes is below, but that tends to balance out over time. |
| # / 00:39:46 | Vishesh Choudhry | Like you have to filter out for some of the noise. So this is the challenge with looking at more and more realtime data, especially with something like DAI with really low liquidity is you start to introduce a lot of noise to the system, which is why in the past I've preferred much more so to look at like a longer running volume weighted average. So just to hop over and so volume has come up a little bit, but it's still like relatively low levels. we saw, a little over a week ago, two huge, huge volumes. And so let's jump over to that. |
| # / 00:40:18 | Richard Brown | Before you move on. |
| # / 00:40:19 | Vishesh Choudhry | Yes. |
| # / 00:40:19 | Richard Brown | Can I ask, do you think this is because inventory levels had been cleaned out or is there some other mechanism at work? |
| # / 00:40:26 | Vishesh Choudhry | This being the volumes or? |
| # / 00:40:29 | Richard Brown | Yeah. |
| # / 00:40:30 | Vishesh Choudhry | No, I think it was just like longterm this is I think a fairly reasonable volume for DAI, like DAI volume has always been kind of low, but, it was just a crazy spike, and in a very atypical time period that was happening over a week ago. And so I will actually look at that on the longer running chart. So as you can see here, like you see that inane spike in volume as opposed to the more long term. So what's interesting is as the stability fee has come up, the DAI price in like a volume weighted average basis has come up a bit as well. I don't want people to just necessarily think it's as simple as that. But looking in the last couple of weeks, you saw again this crazy spike that I have been talking about. This was a lot of people, as we actually got to touch on, last week, this was a lot of people clearing out debt, a lot of people kind of coming full circle from leveraging up on ETH, to cashing out on some of those profits, buying back a lot of DAI. |
| # / 00:41:41 | Vishesh Choudhry | So this kind of spike in volume was to me a lot of people purchasing DAI. Now, there was a big open ended question, last week, which is what are people going to do with this DAI? And everybody was thinking, Oh, DAI is trading 2, 3 cents, above the dollar. Everything is perfect now and we should lower the stability fee. I had suggested that that was maybe a little bit premature and we shouldn't act on kind of this really short timescale trends, which appears to be the case because this is kind of leveled out a bit. But it has been hovering above and below a dollar. |
| # / 00:42:15 | Vishesh Choudhry | So that is to me saying that with the current levels of DAI held and being traded with the current stability fee with the current ascending supply and with youth price holding fairly steady, we're actually at this kind of equilibrium point where DAI is operating normally. So this is really good. I think this is a good data point for us to have in the bank because now we have a frame of reference for understanding what DAI looks like in a "normal scenario". |
| # / 00:42:45 | Richard Brown | Sorry, can I interrupt you again Vishesh? |
| # / 00:42:46 | Vishesh Choudhry | Yes. |
| # / 00:42:47 | Richard Brown | Is this something that we should probably call out regularly in these calls and these meetings is to try and figure out what the proper cadence is and what kind of like a tale of data is required before we start making decisions? Because this kind of requires like with... if we don't wait a week, we're not really going to know with any kind of confidence what happened. Right? |
| # / 00:43:10 | Vishesh Choudhry | Yeah. So, and I know you've asked me that question before. It kind leads me into a trap and give you a concrete timeframe. But what I will say- |
| # / 00:43:18 | Richard Brown | Create a picture for me. |
| # / 00:43:20 | Vishesh Choudhry | I will never, I will never purport to do that, but what I will say is a week is definitely too short. Like we cannot make significant monetary policy decisions on what happens to the price within like a three or four day period. That to me seems a little bit ridiculous. |
| # / 00:43:36 | Cyrus Younessi | I mean on the flip side though, we were making them, when the DAI price was weak, we are raising him every week. |
| # / 00:43:43 | Vishesh Choudhry | Yeah. But I don't think that was a week to week response. Right? That's different. That's not us saying that something happened within the course of a week and now we're going to make a change because of that. That was us looking at a longterm trend and saying we know by a wide margin that were below where we need to be. And so let's continue to make increases on a relatively weakly timescale, which is two different things in my opinion. |
| # / 00:44:06 | Richard Brown | Yeah. That's one of the things I've been concerned about from ever since activity in the governance system ramped up is that we might get lulled into thinking that this is the new normal when it might not be in that we should continuously evaluate what cadences makes sense. |
| # / 00:44:20 | Vishesh Choudhry | Yeah. So I think that's the constant balance that we're managing here is more time is our friend and it helps us understand how trends kind of level out. But at the same time if we know we're not in the place where we need to be, inaction can be bad for very obvious reasons. So it's definitely a balance we have to continue to manage. What I would say is if we're looking at longer term data trends and making actions on that, that makes sense. If we are responding to intraweek events, and making monetary policy changes on that, that does not make sense. And I think like at the current amount of information that we have, that's probably about as much as I'm like willing to conclude, but I'm sure we'll learn more as we go along. |
| # / 00:45:08 | Richard Brown | Thanks for that. |
| # / 00:45:10 | Vishesh Choudhry | So, kind of look at ETH rice and DAI price. So, and there's a number of different lines here that people can play around with. The gist was, what's been happening to kind of a slightly smoothed out, longer weighted average of DAI and ETH price. So what's really interesting here is in the past there had been a stronger correlation as I mentioned, inverse to what's been going on with DAI and ETH price. Recently, that has become much more correlated, the course of this month. So that's kind of broken the model a bit, which to me says that there's a deficiency in the model. And so what my suspicion is I need to get a lot more refined on how we think about ETH price short versus long term. And like maybe there's two models here. One like what does DAI do in a bull market situation and what does DAI do in a bear market situation? |
| # / 00:46:13 | Vishesh Choudhry | Because those clearly seemed to be different things. And I think it makes a lot of sense, like intuitively if you think about it because in a market where ETH is trading cheap relative to where people think it should be, they're going to lever up and DAI price is going to suffer. And so there's going to be kind of an inverse relationship, in a market where people have already done that and now they're just cashing out on that ETH they're trying to run away with a 40% profit or a 30% profit. And so they don't really care about taking a one to 2% haircut if DAI is trading above a dollar. And so in that scenario, you can actually maintain this kind of positive correlation. And so what we need to do is expand our view so that we have a more comprehensive model that can explain both of those situations in which we haven't really been doing up until now because we haven't really seen this kind of positive relationship very much. And so, I think that's what I'm saying where we get to steal a lot of different types of environmental conditions. We learn more and we can refine our models and get better at describing what the hell's going on. |
| # / 00:47:25 | Richard Brown | Sorry Vishesh I keep on interrupting you, but that there's something interesting. I just came up there. You're saying that in bull, we don't have to be super responsive with lowering the fee but in a bear we should be fairly responsive and raising a fee. Is that the takeaway? |
| # / 00:47:42 | Vishesh Choudhry | Yeah, I mean it's, I think that's fair. And what I will qualify is, it's like the bull through the China shop. You know what, in the bull market there is such a strong signal that is kind of wiping out any sort of cost. Like, so there's this calculator that I shared a while back in terms of what's the profitability of leveraging ETH. Like that depressive effect on DAI price of maybe some people are levering up, and they are selling DAI is so outweighed by just the pure transaction value of people cashing out into DAI, that you'll lose all of their signals. It'll just be overwritten and you'll never know what the hell is going on on the other side of the equation because all you see is the strongest signal. |
| # / 00:48:25 | Richard Brown | So euphoria in a bull blurs the signals and fear in a bear kind of clarifies them? |
| # / 00:48:33 | Vishesh Choudhry | Yeah. All that adrenaline, we don't know what we're thinking. |
| # / 00:48:36 | Richard Brown | Interesting. All right, thank you. |
| # / 00:48:36 | David Utrobin | Do you think that there's, there's like a built in bias in the market to like just make it really hard to lower the stability fee because I think that, like do you think it's possible to basically have an artificially high stability while still seeing DAI behaving like it's at an equilibrium point when in reality the equilibrium point might be like a broader range of stability fee? |
| # / 00:49:02 | Vishesh Choudhry | Well, so there's definitely like a concept of a, like a local equilibrium vs a global one. Right? So you can have local versus global maximum. So the idea that, and I think this is a really good example for what's going on right now. It's possible that what we are dealing with is a local equilibrium and not a global one. Because as I mentioned, there were a lot of people that cashed out on kind of this full circle leveraged ETH position. And so those people in what I wanted to get to was what's going on with burns and wipes, sorry, burns and draws. Because those people might be sitting with DAI, which means whatever they do next in the longterm might move the equilibrium. So even though right now we might be stable, it's possible for their next action, which is actually just a reaction in a longer timescale to move that equilibrium and potentially break us off of it. |
| # / 00:50:04 | Vishesh Choudhry | And so then it's really a semantic discussion of like, do you consider that to actually have been at equilibrium or not? I don't know. I can't answer that for you, but what I can say is it's entirely possible that right now we are stable, but just that kind of full circle swing of what had been going on with leverage has yet to conclude that we might yet move off that equilibrium. |
| # / 00:50:27 | Matthew Rabinowitz | Yeah, or we might be in the eye of the storm. |
| # / 00:50:31 | Vishesh Choudhry | Yeah, that's it. That's a fair. |
| # / 00:50:33 | Cyrus Younessi | I'm hesitant to draw too many conclusions from the bear market of last year because we were at the same time seeing this, this massive expansion of DAI suppliers MakerDAO was just kind of hitting the limelight and getting popular. I mean, I wouldn't expect another exponential growth in the DAI supply if another ETH bear market hits like we did last year. So everything around the stability fee mechanics, it is almost useless to me, I think from the data perspective. |
| # / 00:51:04 | Vishesh Choudhry | Yeah, I mean, so I think that's totally fair and I don't think I'm necessarily making a lot of conclusions based on what happened, over a year ago. But, I'm mainly talking about like what are kind of the trends in the last, I don't know, let's say six months. |
| # / 00:51:22 | Cyrus Younessi | Got it. |
| # / 00:51:24 | Vishesh Choudhry | I'm so sorry. I just want to make sure we get through a lot of the important part. So supply. Supply has been coming down in the past few weeks, fairly consistently. I had all these prepped, but I reloaded the page. So yeah, you can see here, there were a few like major burns. and I kind of tweeted about this. May 16th was a oddly eventful day. So what we saw actually was shortly after the 19.5% raise, supply had been coming down. There was this weird little run up and then immediately precipitously drop back down. So what I can say is the flywheel is moving and circulation of debt is very strong, so this is a really interesting metric to see. I think it's one of the most important, most underrated metrics about Maker is just how quickly, is debt circulating through the system. |
| # / 00:52:29 | Vishesh Choudhry | So this was just a tremendous amount of churn. What we saw I can't necessarily explain why there is this run up in supply. But what does make a lot more sense, is that supply then significantly dropped as a lot of people were paying back that with newly acquired DAI, a newly purchased DAI. |
| # / 00:52:51 | Richard Brown | Was that a single large holder or was this some kind of group? |
| # / 00:52:55 | Vishesh Choudhry | I think this was ... And I'm forgetting now, but I think this was a cluster of large transactions. |
| # / 00:53:06 | Cyrus Younessi | That's interesting. |
| # / 00:53:06 | Vishesh Choudhry | But- |
| # / 00:53:06 | Matthew Rabinowitz | Yeah, that was also CDP #5199, they burned like a $2 million in one day. 2 million DAI. |
| # / 00:53:12 | Richard Brown | But something to keep in mind too is that some of the times we're looking at trends and these trends represent a single person doing a single thing, right? |
| # / 00:53:20 | David Utrobin | #3228 also scaled out, he has like a 50,000 DAI debt now. |
| # / 00:53:26 | Vishesh Choudhry | Yeah. And anecdotally as well, there've been a lot of reports of sort of refinancings from Maker to other platforms. So that naturally has an associated dip in supply. So just to actually touch on that real quick, because we only have a couple minutes here, so this is what I was talking about in terms of the circulation of debt. The age of debt has just dropped precipitously in the last few weeks. Again, this is sort of around that, May 16th period has just been coming down, which is I think really good. So if there's more debt circulating through the system, there's more realized fees, there's less, accrued outstanding fees and that's just, a healthier state because there's less of a disparity between what's on the surface of what's underlying in the system. And just to touch on the secondary lending rates. |
| # / 00:54:23 | Vishesh Choudhry | So there's kind of this healthy buffer, I think that's been maintained around 3 percent, between Maker and secondary lending platforms. Dharma is a bigger spread, but those are subsidized. It's really hard to say what the actual, like rates would be without subsidies. And I know that's a conversation. It's been discussed before. it's been suggested that we look at things like Nuo as well. But, there've been a lot of issues with Nuo lately, so I'm fairly comfortable just looking at Compound. It's like the most reliable signal at the moment. and so similarly around that May 16th period, what we saw was kind of a drop, in the rates because there is a corresponding drop in borrow volume. So as fewer [inaudible 00:55:11] which is really interesting because it kind of flies in the face of the narrative that people were refinancing until something like Compound. |
| # / 00:55:18 | Vishesh Choudhry | So, Dharma also didn't really experience a huge change in volume. So I'm not really sure where those people were potentially refinancing to or if they just completely exited there a long positions. It's really hard to say what people are doing on other platforms or outside of Maker. I'm going to skip over this now. Collateralization ratios. So just in that same time period, a ton of ETH is removed, which makes sense because a ton of DAI was removed. So why would people just keep eating sitting? But the collateralization ratio has been running up, and been holding fairly steady, which I think again is a metric to watch because it speaks to the idea that people want to maintain this buffer, which means there is still potentially worried about a drop in ETH price. So I think people are fairly neutral on ETH right now. They're not really levering up, they're not really, clearing out their collateral. They're just kind of waiting to see what happens next. Okay. I think that's time. |
| # / 00:56:23 | Richard Brown | I want to jump in super fast, we're at the top of the hour. I think that there's more cool conversations to be had, so I'm willing to hang out for another half hour if everybody else is. I think, yeah, I think that we could probably just leave it there. So if we want to have a free form Q&A please hang out back to you Vishesh and/or Cyrus. Vishesh or Cyrus? |
| # / 00:56:57 | Vishesh Choudhry | I'm happy to answer questions if people have them |
| # / 00:57:00 | Richard Brown | Yeah. Was there anything on the side chats that didn't get addressed? |
| # / 00:57:06 | David Utrobin | Yeah, I think, there was a question about one of Vishesh's graphs about the correlation coefficient. |
| # / 00:57:17 | Vishesh Choudhry | [inaudible 00:57:17] which graph were you referring to? I'm assuming it's ETH and DAI price. |
| # / 00:57:24 | David Utrobin | Yeah, I think it was the ETH and DAI price graph when that question was asked. |
| # / 00:57:28 | Vishesh Choudhry | Yeah. So, that's one of the things that I've mentioned prior on one of these calls is ETH and DAI price correlation is not so great. And I forget and I think it's around 1.4 or something, but DAI and ETH ratio of short term price to longterm price is stronger. I think that's above 0.5. But with the caveat of in the last few weeks that correlation has really come down. So that's why I really have not felt comfortable sharing that more publicly is just because I don't feel it's a robust enough model yet to explain what's going on. |
| # / 00:58:15 | David Utrobin | Fair enough. Cool. |
| # / 00:58:20 | Richard Brown | All right, well they're not going to jump in there with a random question then. I'm not sure where I got this idea from. It was somewhere I saw something on Twitter, I was Reddit. But the question was asked if there is a logical lower bound on what the stability fee can be, in light of all these alternate lending platforms out there, is it possible that, it just, it's never going to make sense to go below 10% or 8%, once the industry sort of determines what the most common lending rate is? |
| # / 00:58:51 | Vishesh Choudhry | So this question was posed ... It's been posed on these calls, it's been posted on Twitter, like how do we look at the relationship between the stability fee and secondary lending rates? I want to caution people from using the secondary lending rates as a target for stability fee. Even with like a built in buffer, it is a good input to understanding what's going on, as a response to the existing stability fee. But you just get into a very screwy feedback loop if you start to use it as a target. And I know we've talked about that before. As far as like what happens if we start lowering the stability fee and is there like a realistic floor there and can we discern that for, from other platforms? Sure. If you start to look at what's on those other platforms as a floor, that does make sense because if you've learned the stability fee down to at or below what's on some of those other platforms, you get it back. |
| # / 00:59:56 | Vishesh Choudhry | You get us back into some earlier economics that we were in, which was kind of screwy where people are highly incentivized to just draw out day and lend it, or to refinance loans from other platforms back onto Maker. Neither of those things are, good for the system I feel. But that's when we're talking about the floor. There's a lot of bad things that can happen between here and the floor I think is the point. And so sure we can think about what the floor is, but I don't even think that's the right discussion to have. I think what we're doing now in terms of what's the right level to be at as a target is a much healthier discussion. And I think naturally Maker should be the most expensive. Like it should be the lender of last resort because it's the easiest place to get a loan from. Maker will always lend you out DAI no matter what. |
| # / 01:00:50 | Richard Brown | Well, unless the debt ceiling is hit. |
| # / 01:00:52 | Vishesh Choudhry | Yeah, that's fair. And then we get into, some even more complex economics, but I think we're in a bad spot if we hit the debt ceiling in my opinion. |
| # / 01:01:01 | Richard Brown | So interesting how those conversations around debt ceiling sort of dried up. |
| # / 01:01:07 | Josh Wisniewski | What about if Maker's always lender of last resort, do, do these secondary markets, that doesn't quite make sense because the secondary markets are always going to have to be more expensive, on fees then Maker or otherwise the lenders aren't making any money in. It has to be subsidized somehow else. |
| # / 01:01:27 | David Utrobin | Yeah but that's assuming that the lenders on those markets are sourcing their DAI from CDPs on Maker. They might as well... they could just be buying it on the spot market with their own capital and deploying it completely independently of Maker. |
| # / 01:01:42 | Vishesh Choudhry | And I think you would want that happening to right? So if those people are incentivized to buy up DAI, that helps to support the price floor on DAI. |
| # / 01:01:55 | Matthew Rabinowitz | Yeah. And also could be just a short term phenomenon altogether. I mean at the end of the day Maker, which mimics a central bank, it basically is a central bank that has a retail book of clients, thus circumventing, mitigating, removing retail banks. Then the question is why would a person put money in a retail bank to earn what they could earn directly from a central bank? |
| # / 01:02:20 | Vishesh Choudhry | Are you talking about the DAI savings rate now too as well? |
| # / 01:02:23 | Matthew Rabinowitz | No, not even that. I mean it's more of just the stability as a whole. I mean, I don't want it to date poorly, but we may find that the whole secondary market is just a short term phenomenon. |
| # / 01:02:37 | Vishesh Choudhry | I mean I think that's totally possible. And I think there are definitely parts of it that are not going to stay the way they are. Like if you look at, as I mentioned, like Dharma 14.5%, I think it was ... I'm curious to see what it moves to in the absence of subsidies. |
| # / 01:02:55 | Matthew Rabinowitz | I think those subsidies just ended by the way. |
| # / 01:02:56 | Vishesh Choudhry | Did they? |
| # / 01:02:57 | Cyrus Younessi | Yep. |
| # / 01:02:58 | David Utrobin | On Dharma? |
| # / 01:02:59 | Vishesh Choudhry | Interesting. |
| # / 01:02:59 | Cyrus Younessi | Yep. |
| # / 01:03:01 | Vishesh Choudhry | [crosstalk 01:03:01]. And then it's a fixed time period loan so I'm sure, it'll take some time to propagate as well. |
| # / 01:03:12 | Cyrus Younessi | One interesting thing actually to talk about is a tracking when big chunks of Dharma loans will mature and potentially bringing a lot of DAI back into circulation. |
| # / 01:03:29 | Vishesh Choudhry | Yeah. Especially if the subsidies just ended, I'm sure you'll have a wave coming up. I guess, what are they, three months? |
| # / 01:03:38 | David Utrobin | Pretty much 28-day loans and there were also 90-day loans. |
| # / 01:03:42 | Richard Brown | Does anybody know of any analytics tools for what's going on with Lever right now? |
| # / 01:03:48 | Vishesh Choudhry | I mean Loanscan, does have some stuff, but- |
| # / 01:03:53 | Richard Brown | [crosstalk 01:03:53] for Lever as well? |
| # / 01:03:55 | Vishesh Choudhry | It's just for Dharma. |
| # / 01:04:01 | Richard Brown | So, I know that there was some significant, some large capital moving into Lever and I'm presumably that's all going to come out at the same time. So it'd be nice to know when that's going to happen. |
| # / 01:04:11 | Vishesh Choudhry | I agree. This is all really great stuff to watch out for. I'm sorry, just to answer the question real quick, in terms of the question was about the secondary lending platform rates and sort of is that a natural or healthy buffer in terms of Maker being more expensive? I mean if you look at Compound right, it's a pretty organic situation. There's no subsidies going on there. It is similarly, a non-fixed time period loan. So I think it's a pretty good data point to look at. At least one of the best ones out there right now. And I think what we've seen is there is this kind of natural buffer that tends to form after the Maker stability fee had come up above a certain point. There was, and enough time was given to propagate. |
| # / 01:05:11 | Vishesh Choudhry | There was this buffer that started to form between the Maker stability fee and the Compound rate. And there's no artificial short term pressure that I can think of that would cause that to happen. And then later reverse, it seems to be a pretty natural trend. And I'm curious if there's disagreement on that. |
| # / 01:05:34 | David Utrobin | Well, I don't know if I'm framing this correctly or not, but what is the pressure be like the risk of the rate changing. So like that spread, that 3% buffer is really just like people trying to judge like the length at which they're willing to deploy capital on either side or payback capital or hold capital? |
| # / 01:05:53 | Vishesh Choudhry | But there's no, there's no fixed time period on Compound. So it would just, those rates would move accordingly. So I don't think that would be a factor for people is the short answer. What I do think is interesting is if you look at supply versus borrow volume, on some of these like Compound in particular, there is always been like a very healthy difference between, sorry, not healthy is actually the wrong word. There's just been a significant difference, between the amount that's being supplied on there and the amount that's being borrowed, which is really interesting because, presumably people purchase that DAI are eating the cost of capital and are just keeping it sitting on there. And there's, right... In the last month or so, there's been roughly four, three to 4 million DAI that people have just kept parked on Compound that has never been taken up into alone that they're not earning any money on. |
| # / 01:06:52 | David Utrobin | Isn't that a function of like the Compound, like the formula that they use, like they want to insure a float there because once it gets smaller the borrow rates get higher and then subsequently like [crosstalk 01:07:05] rates going higher? |
| # / 01:07:07 | Cyrus Younessi | Aren't they getting the supply rate? Vishesh what do you mean they're not getting any? |
| # / 01:07:12 | Vishesh Choudhry | So you only received this unless I'm misunderstanding how Compound works. You only received the supply rate. If you loans are taken up |
| # / 01:07:20 | David Utrobin | No- |
| # / 01:07:21 | Cyrus Younessi | Isn't it just a giant bucket? |
| # / 01:07:23 | David Utrobin | Yeah, it's a bucket, you get the supply rate no matter what. And like the borrow rate is always higher than the supply rate. Like right now the borrower, it's like 18% or something or 16%. |
| # / 01:07:33 | Cyrus Younessi | The borrower rate is going to track closely to the stability fee because there's that arb there, right, where you can just borrow from MakerDAO and lend it on ... Wait. |
| # / 01:07:50 | David Utrobin | Well- |
| # / 01:07:51 | Cyrus Younessi | THe buyer rate will try it closely because of the demand from basically speculators for what they're willing to pay to borrow. But the supply rate is obviously factored in it. |
| # / 01:08:03 | David Utrobin | Yeah. That's the whole thing is the supply bucket takes all of the revenue from the borrow bucket and splits it amongst itself. So it doesn't matter if your DAI is actually borrowed, there's going to be a float of DAI sitting in Compound that's never borrowed, that's meant to be there. And that float is like literally just correlated with that spread on the borrow supply interest rates. |
| # / 01:08:26 | Vishesh Choudhry | But even in that scenario, it just doesn't make sense to me why, I'm like, there would be this additional risk factor that creates three or 4% buffer between the Maker stability rate and the borrower rate on Compound. |
| # / 01:08:46 | Josh Wisniewski | Sorry, just to jump in, I'm thinking about what you just said, David, that you basically, if you're going to, you're going to buy the DAI off the open market and then put that in Compound instead of taking it directly from a CDP. And what I started thinking was, well, if you've got, $1 million that you want to deploy to put that in a CDP, you need 1.5 million to get a million DAI out. But if you buy a million on the open market, then you can go straight to Compound and lend that at the rate. |
| # / 01:09:18 | David Utrobin | Yeah. |
| # / 01:09:18 | Josh Wisniewski | So you have that 50% difference on how much capital do you actually need to get the rate that you want to get the spot rate? |
| # / 01:09:28 | David Utrobin | Yeah, exactly. This is, I mean suppliers and secondary lending markets, like it doesn't make sense for them to source DAI out of CDPs if they're projecting that their costs on borrowing is going to be higher than their revenue on supplying or the profit from supplying. So it really makes the most sense for savers and for people that just want that passive, sort of low risk revenue stream. Yeah. It's just a different segment of the market. It's just a different group of people, different group of capital. |
| # / 01:09:58 | Josh Wisniewski | Well, you're kind of like, if you're, if you've got, I can't do the math in my head right now, but basically if you have to pay 17 or you're getting whatever, 17% on, 1.5 million, that's a quite a bit different than 17% on a million of capital. |
| # / 01:10:21 | David Utrobin | Yeah. But I'm a little confused what you're saying because, in like the case where you're borrowing the DAI, you're, you're paying that 17% well 19.5% and a half percent. Right? And like, whatever the stability fee is. Like there's nobody who's actually lending that diets just, just, the stability fee, it gets burned and MKR. But on Compound, if you have 1 million bucks and like the supply rate is 8% or whatever, what we're seeing right now, listen, that's 80 grand a year on your million. And if you have a million and a half, guess what? That's 80 grand plus for you, it's 120. |
| # / 01:10:56 | Josh Wisniewski | Right. Yeah. Okay. That makes, yeah, that makes more sense. |
| # / 01:11:00 | David Utrobin | Yeah. If you have 1.5 million in DAI, you're making money on Compound and dYdX is also a pretty, yeah. Mateo has been mentioning in the chat that dYdX is seeing a lot of volume as well. And noticed that the supply rates have been consistently higher than Compound on dYdX. And Vishesh I think that it's worth tracking dYdX supply and borrow rates. I think the supply rate is more consistent than the borrow, but I know, I don't know, |
| # / 01:11:30 | Richard Brown | Yeah, they've been exploding recently. |
| # / 01:11:32 | David Utrobin | So like, yeah, I saw like 30% like several times throughout the week. I was like, damn, wish I had 1.5 million. |
| # / 01:11:43 | Cyrus Younessi | What keeps the borrow and supply rate on compound from converging closer to each other? |
| # / 01:11:51 | Chris Bradbury | So I think it's the way that they coded it because they're always ... the less the spread, the higher the borrow rate is pushed up. So like if there's a lot of people borrow, like if there is a smaller spread between the borrow supply, sorry, the borrowing volume and the DAI supplied, then the borrow rate jumps up and so the supply rate jumps up and more people are incentivize to jump in and supply DAI on that platform. So there's just like, I think it's just the way that it's coded. |
| # / 01:12:30 | Vishesh Choudhry | Yeah. If you, if you watch those two rates there's a clear like mathematical relationship between the two. |
| # / 01:12:37 | Cyrus Younessi | Right. Because that ... But my question is like how is that derived? |
| # / 01:12:42 | Vishesh Choudhry | I don't know. They're I don't know Compound's like internal algorithms for setting the rate. |
| # / 01:12:47 | David Utrobin | They've posted it publicly. It's in somewhere. |
| # / 01:12:51 | Richard Brown | It's interesting though because there was a [inaudible 01:12:53] right there for the stability fee in the DSR though, right? They have a cleaner, less complex system, but there's the same general principle at work. |
| # / 01:13:03 | Vishesh Choudhry | Yeah, definitely some lessons to be learned from what some of the other platforms are doing. I mean I think that buffers is decreased as the Maker stability fee has increased this one point, which is interesting. |
| # / 01:13:22 | Josh Wisniewski | Once we get DSR, does that start to confuse what stability fee should be looking like? Like the correlation between DSR and stability fee because we've got two variables to tweak at that point? |
| # / 01:13:37 | Cyrus Younessi | Sorry, can you repeat the question? |
| # / 01:13:40 | Josh Wisniewski | So when we go to Multi Collateral DAI, we're going to, we're going to have stability fee for each collateral and DSR, which I believe is basically risk team only going to be setting stability fees and we kind of forget about those as general governance and then DSR is what we float around a holding hand? |
| # / 01:14:04 | Cyrus Younessi | No, I think there's definitely going to be governance on both, but they're going to be more tailored to what the fee represents. So I think DSR will be kind of more of a monetary policy type of tool. And the risk premium, I mean the stability fee, the collateral parameter I think will be rebranded as the risk premium rate and that will entail or collateral specific discussions. But ultimately there will be governance on both. |
| # / 01:14:39 | Josh Wisniewski | Okay. I'm just preempting sort of the idea that it's going to get real confusing real quick once we hit Multi Collateral. |
| # / 01:14:48 | Cyrus Younessi | Yeah, it is. That's actually why I was asking about, that like inherent bias in being in having like a hard time actually lowering the fee. Like it's really easy to, figure out when to heighten it, but when do you know that DAI is stable but supplies dropping. I guess like just start lowering it when you really see a consistent premium. Yeah. But like that might take a long time. I don't know. Yeah, I'm just thinking out loud. |
| # / 01:15:16 | Vishesh Choudhry | No, that's a really, really good point to keep in mind though, is, and I'm going to stress this multiple times. It's an asymmetric system, so it's not the same to raise something versus lower it. And like, especially when you start thinking about some of those discussions that were had around a debt ceiling and all of that is like the perfect example. Just because you wind something up, doesn't mean you can undo the same steps, to unwind it and just, I think that's particularly true with something like DAI, where it's got sort of inherently different economics above and below a dollar. |
| # / 01:15:53 | David Utrobin | Yeah. And if, and if, if you guys watched last week's call, for those of you that are in here, Matthew, he talked a bit about, like what it would look like for us to do, like the ideal balance between the stability fee and the DSR. And that initially it will be high and like we should work to iterate it downwards. But again it's theoretic and it's just thinking out loud. |
| # / 01:16:22 | Cyrus Younessi | All I know is when MCD launches, Richard's not going to sleep for months due to governance. |
| # / 01:16:29 | David Utrobin | No, we're going to have to get them bigger canteen for his coffee. |
| # / 01:16:32 | Cyrus Younessi | Coffee's he's going to need something stronger [crosstalk 01:16:34]- |
| # / 01:16:35 | Vishesh Choudhry | Diminishing return on coffee. |
| # / 01:16:37 | Richard Brown | Yeah. I've already hit the diminishing returns. |
| # / 01:16:42 | Vishesh Choudhry | But I do want to touch on a serious point real quick, which is what do we manage towards, right? So what do we look at? What's the guiding, like North star, especially as the system becomes way more complicated? Because we can envision a lot of great complicated systems about how we manage this. But in my experience, the simplest like, most constant guiding principle is going to get, 80, 90% of the way there. And right now what I'm thinking is you want to manage up the DAI supply with DAI operating at or around a dollar in the long term, right? So if you have both of those things happening, if days growing and it is stable, that is good. And I think that's- |
| # / 01:17:28 | Richard Brown | It's a interesting point though, because right now we're hyper focused on stability for you because we have this hard cap of whatever, 100 million or a sweet spot of 80 circulating. But we also need to take into account competition and adoption in the MCD world. So that's a great point. For sure. |
| # / 01:17:48 | David Utrobin | I'm going to say exactly that. Like market share for stable coins is something that we should be cognizant of. |
| # / 01:17:53 | Cyrus Younessi | No, I don't think so. I don't buy that at all to be honest. I think the real- |
| # / 01:17:59 | Richard Brown | Contrarian. |
| # / 01:17:59 | Cyrus Younessi | I think the real, the real debate is going to be on actually what Vishesh just kind of, he casually threw together two conflicting points about keeping DAI stable and growing the DAI supply at the same time when those two notions may not work in tandem very well at all. I mean, it's easy to grow the DAI supply, right? Just lower the stability fee and- |
| # / 01:18:26 | Richard Brown | [crosstalk 01:18:26] simplifying things that are like, the DAI supply is a result of an ecosystem full of actors and like we're just one tiny part of it, right? Like business development, a wider DeFi space. Crypto in general builds up the supply. We need to make sure that we're not inhibiting that. Right? |
| # / 01:18:44 | Cyrus Younessi | Right. But I think you could just increase the debt ceiling on ETH and put the stability fee zero and then- |
| # / 01:18:51 | Richard Brown | Got you. [crosstalk 01:18:51]. |
| # / 01:18:54 | Vishesh Choudhry | I agree with Cyrus. Like I just want to be clear. I agree with Cyrus. I don't agree that Cyrus disagrees with me. So I agree that DAI supply and DAI price are often like in direct conflict with each other. But that's kind of my point is there exists theoretically this proper balance of growing the supply slowly and steadily while keeping DAI at a dollar in the like medium to longterm timescale. And when both of those things are happening in tandem, I think that is excellent. When DAI supply is relatively steady and DAI price is relatively steady at a dollar, I think that's pretty good. Right? |
| # / 01:19:34 | Richard Brown | We maintain tension between DAI supply and the stability fee in our ecosystem. And we balanced that tension against the wider market than that though. Right? |
| # / 01:19:42 | Cyrus Younessi | But even that if you naively assume that the total stability fee is this, some of the DSR and the risk premium rate, if you just set all the risk premiums to zero and then just set the DSR such that DAI still equal to $1, then you will exponentially grow the DAI supply because we're essentially taking on risk for no reward. |
| # / 01:20:09 | Vishesh Choudhry | And then I'm not sure that scenario would even be possible. I'm not sure that's a realistic scenario to construct. |
| # / 01:20:14 | Cyrus Younessi | Sure, it is why not? |
| # / 01:20:17 | Vishesh Choudhry | Because in that situation, like there's a question like what is the actual supply of DAI there as well? With respect to the DSR. But in that scenario, either one of two things like realistically the supply of DAI is actually not what you think it is. Or like DAI is actually not going to be able to stay at $1. |
| # / 01:20:40 | Cyrus Younessi | [inaudible 01:20:40] the demand though. Right? Because let's say a global DSR of 5%, keeps DAI stable no matter what collateral assets are in the portfolio, 5%, we'll keep it stable. And then you have two assets, a risky one another and high risk and a low risk one. If they both have the same overall stability fee, then people will just keep borrowing off of the riskier one. |
| # / 01:21:13 | Vishesh Choudhry | And locking that DAI up in the DSR you're saying, or a good portion of it? |
| # / 01:21:19 | Cyrus Younessi | Not necessarily locking it up in the DSR, but they're basically, taking advantage of lower rates through MakerDAO than they could get elsewhere for that same asset. |
| # / 01:21:31 | Vishesh Choudhry | So I guess there's, there's two, there's three questions here. And the third I'm just going to jump to which is okay, if not that, then what is your definition of success, which I'm really curious about. But the first two is if there are five DAI and those five DAI are equal to a dollar, is that success? And then the second question is if all the DAI that's out there is actually just locked up in the DSR and DAI is trading in a dollar, is that success? |
| # / 01:22:02 | Cyrus Younessi | Right? So I think success is keeping DAI stable while maximizing the risk reward frontier. |
| # / 01:22:14 | Vishesh Choudhry | Can you clarify what you mean by "maximizing the risk reward frontier"? |
| # / 01:22:23 | David Utrobin | So, you have to factor in the financial incentives for the MKR holders, right? So they have this incentive to a keep DAI stables but maximize the buy and burn. |
| # / 01:22:37 | Vishesh Choudhry | Right. But that's my point, maximizing buy and burn, it's an equation and one of the variables in that equation is the total amount of DAI being drawn and wiped. |
| # / 01:22:49 | David Utrobin | Yeah. And I think not only that, it's also the scaling of the total DAI supply in this and the scaling of demand for DAI. |
| # / 01:22:57 | Cyrus Younessi | Right. But I'm saying you can optimize for increasing, draw and wipe at the cost of having extra latent risks in your collateral that you're not factoring in. That's, that's exactly my point is you can take some junk grade bonds and assign it a really low risk premium and then you'll have enormous borrowing against it because it's so cheap to borrow against what's typically a risky asset. |
| # / 01:23:27 | Vishesh Choudhry | So I don't think that's unreasonable, right? If the simple answer is simple answer is there's no simple answer. But not to be cute. Like if the answer is that we should think about the overall growth and scale of Maker as a system and the amount of DAI is being transacted, the amount of DAI is being drawn and wiped and do we think about DAI being stable and we think about systemic like health risks right in the system like that is, if that's the answer is just add a third kind of category we're not thinking about, which is what is the systemic risk. I think that's totally reasonable. I'm just trying to get to like what is a good definition of success? |
| # / 01:24:03 | Cyrus Younessi | Right. So yeah, I'm saying is that definition of success is not necessarily DAI at a dollar and there's a trillion DAI in circulation. It's actually DAI is at a dollar and there's a very healthy collateral pool. Now the goal is to maximize the DAI supply given a healthy collateral pool. |
| # / 01:24:23 | Vishesh Choudhry | Yeah, that's great. Sounds great. Yeah. |
| # / 01:24:27 | Richard Brown | All right, well that's, everything got spicy, which is nice. And Sarah's got full of life, which is also good to see as well. But we're at the 10:30 mark. I need to jump into another meeting, so I have to stop the recording. So if anybody wants to hang out [inaudible 01:24:40] but- |
| # / 01:24:42 | Cyrus Younessi | [inaudible 01:24:42] to summarize. |
| # / 01:24:44 | Richard Brown | So going off the record at this point. |