Ep. 31 Apr 18 - 2019¶
Video | Audio | Discussion
| References | Person | Text |
|---|---|---|
| # / 00:00:03 | Richard Brown | Hello everyone. Welcome to the April 18th edition of the scientific governance and risk meeting. This is our thirty first call, so congratulations to everyone for surviving this long, it's a bit of a milestone, it's an achievement in my books. It's not easy to keep momentum up on these things so I'm very encouraged to see the number of people that are continuing to join these calls and engage. |
| # / 00:00:27 | Richard Brown | It's ... What's happening here is interesting, it's unique and we're all on the forefront of this very new and unusual wave of online or [inaudible 00:00:38] governance and so thanks again for joining. Keep the preamble light, we are desperately interested in getting as much engagement as we possibly can from this audience today, so if you have a question, please type it in the chat on the side. We'll ask it if you don't have access to a microphone, if you do have access to a microphone please jump in at any time and ask your questions. |
| # / 00:01:03 | Richard Brown | Today's meeting agenda looks like this: we'll have a brief thought of the week from Steven, I'm going to talk about some of the issues that arise from the way that we have polling and governance, or voting set up and then we'll hear from two analysts which Cyrus will introduce us to at the risk part of the call. I think that's it from me Steven, I'm gonna hand it off to you. |
| # / 00:01:28 | Steven Becker | No, thanks man. Well, to reiterate what you're saying and like ... firstly, hello everybody and thank you for actually participating on these calls. I must admit that last week, we may have had a record turnout, and I'm certainly hoping we can take that momentum forward because the contributions that are coming in, the work that's getting done, the ability to facilitate it's just improving from week to week, so and you know, just on that basis, you know, thank you very much. But, you know, one thing I like to do is just reiterate every single week exactly what the themes are that we cover. I think it's crucial to know these themes because it defines very clearly the spectrum of the risk that is considered in the system. |
| # / 00:02:10 | Steven Becker | I mean, the first theme, the one that we've been, you know, really centering in these discussions is about the demand and supply and balance or, another way of looking at it is the demand and supply equilibrium. The second, clearly covers the collateral types that we will start presenting in the future, the collateral portfolio as a whole, the parameters to those collateral types and parameters ... and portfolio and the adjustments to those as well. I know that was a bit of a mouth full, but really it's all about collateral types. And then, the last theme that we'd like to cover in these calls are the exogenous risks to the system. |
| # / 00:02:45 | Steven Becker | So, again, as you know, Rich pointed out, we are probably going to be tackling the demand and supply imbalance and we're probably touch lightly on the governance process, highlighting a few of the issues, we need to keep in mind when voting and I'm kind of looking forward to hearing from the two community members that will contribute ... that have been contributing to the risk working group. Really great examples of how contributions in the background are starting to, you know, show real merit and gives us a good sense of how to keep the momentum going. But, again, you know, keep in mind this is decentralized scientific governance and ultimately, what is considered here is being facilitated by the foundation in this very short hour, with the objective of helping MKR governance to make decisions and needs. And, again, and I will always keep reminding everyone that this is a continuous function, governance is a continuous action, so please be involved as much as possible. On that note, turnover to Rich again if you don't mind. |
| # / 00:03:54 | Richard Brown | Nope, don't mind at all, it's my job. So, thank you Steven that's a great introduction and also it provides me a useful segway because a constant theme the last four or five calls has been this idea that governance needs to iterate over time and, we're constantly identifying additional challenges. Maybe the theme or the name of this segment should be 'who would have guessed that governance is hard' because that's sort of the thing that we're figuring out, there's edge cases that need to be dealt with. |
| # / 00:04:23 | Richard Brown | And, what I wanna talk about today is the interaction between polling and voting. We've seen some interesting behaviors arise from those mechanisms, we've iterated a bit over the last couple of weeks, where it became apparent that we really needed to tighten up the cadence of how we initiate a poll, how we discuss those results and how we initiate a vote and most importantly, how quickly we can encapsulate that entire process. The idea is that we wrap that up in a seven day period, that allows the community to have enough agility to respond to market pressures and adjust to the peg as they see fit. It also allows the organization enough breathing room to handle the mechanics in the backhand and the intendant steps that it takes to actually get that ... those proposals showing up in the portal, which are ... I won't bore you with it, but are not inconsequential. |
| # / 00:05:29 | Richard Brown | Here's where that leaves us though, so we have this cadence on paper it looks really good, when put into practice, like anything in the world of software, you start identifying edge case and one of those edge cases and one of the those edge cases is what I want to talk about today because it, internally has spawned a fantastic number of conversations and also in the rocket chat and in the sub Reddit of potential solutions and so we no end of solutions for this problem but we don't know which one is the best one. So, that was a mysterious little preamble. I'm going to provide some detailed context and, or order to do that I am literally just going to read what I wrote, what I posted in the sub Reddit, so if I sound stilted and or mechanical for a while it's because I'm literally just reading things out loud. I'm posting the link as well for people that want to read along with me. |
| # / 00:06:20 | Richard Brown | So, here's the context, I just kind of briefly went over this but we have a governance cadence, Monday we present the poll, it has a default set of stability fee options, the poll collects signals for three days. On Thursday, we get together and talk about how the poll went and discuss other issues. On Friday, we throw up an executive vote, which presents a new picture, a new state of the system that can be chosen by voters that contains the winning poll option. So, basically in the mall that we've been playing with so far it's whatever number had the most signal from the poll, it was the number shows up for the stability fee and the executive vote and then we all determine whether we want to support this new picture of the system. |
| # / 00:07:05 | Richard Brown | This is a fairly clean way of doing things and ... but it carries some assumptions with at and we might be spoiled with the way that this process has worked in the past because what has happened up until just very recently is there's been a tremendous amount of voter engagements, a great deal of interest, we present this executive votes ... |
| # / 00:07:26 | Richard Brown | Actually, back up a second, we have the poll and up until this week there's been a very clear signal about which one of these poll options were most enticing to the community, which, you know, makes us all have a great deal of confidence in the way the system is behaving and then we take that clearly winning poll option or we put it into an executive vote. And, in the past, we've also had the satisfaction of seeing that executive vote be ratified within ten to twelve hours, same day basically, tremendous amount of MKR moves from the old state of the system into the new state. That creates a great deal of clarity and then that free us up to have a bit of a weekend before we begin this process over again following Monday. |
| # / 00:08:08 | Richard Brown | The problem that we have discovered though, is that we're not always ... we can't live in a world where we expect that type of clarity. This week is a perfect example of the poll where I pulled out some rough numbers but for the leave things alone option we have nineteen percent of the turnout, by "raising it by one" is zero percent of the turnout basically. "Two percent" is twenty three percent of the turnout. "Raise it by three percent" is thirty eight. "Raise it by four" is eighteen. So, that's number a number soup but there's no clear, clear winner in the poll that we have this week. We have a range of options. |
| # / 00:08:48 | Richard Brown | The expectation was of the hope or the thesis I suppose was that when we have a poll with a wide range of options with no clear winner as the clock ticks down people will begin to capitulate towards their next best choices. Alright, and then this model we saw that activity, but very little of it. Twenty one hundred MKR move from three percent to four percent since last night, so there is some juggling, but not much. So we're in the situation right now where we have sort of an ambiguous poll. |
| # / 00:09:20 | Cyrus Younessi | Can you clarify something, when you said people ... you expect people to capitulate to their next best option, you mean that if somebody sees that their vote has no chance of winning, they'll redirect ... they'll remove their vote from that option and put it towards another one? Is that correct? |
| # / 00:09:38 | Richard Brown | Yes, precisely. |
| # / 00:09:39 | Cyrus Younessi | Okay. |
| # / 00:09:41 | Richard Brown | Yeah, so, I'm theorizing how the system works I think that that's been a fairly common assumption that people will go “okay, well I'm not gonna get my two percent, but it looks like three percent of my next best choice and that one already has a winning vote, so I'm just gonna dump in there or something. But, that's a largely uncoordinated and organic series of capitulations and negotiations that need to happen offline, so I can't really speculate too much on what that might look like. But, to get in ... or to pick up the thread though, so we have a situation we have sort of an ambiguous poll and that creates some issues where we need to decide, well, if we have four options in poll with five options and only one of them is in the lead by whatever, ten percent, is that really a situation where we need to through the rigmarole of having an executive vote, so I'm gonna leave that question open for now and I'm gonna go back to another problem that we saw in the system over the weekend. |
| # / 00:10:36 | Richard Brown | And, that's the problem where we had ... last week we had the poll, it was fairly clear which one of the options the community wanted the most. That was placed into an executive vote and we also sort of, I don't know who we all are, I sort of sat around on Friday and waited for the end of the day and to see who won, and it was very clear that nobody won in that picture. And, on Saturday, nobody ... there was no clear winner as well and on Sunday, there was no clear winner as well and that was probably because I needed to tee up the next poll for the system on Sunday and that made me ask a very uncomfortable question. |
| # / 00:11:14 | Richard Brown | So, if there's a poll coming on Sunday ... or Monday that is asking for which of these range of options do we want to see. It presupposes that there's an executive, that there's a clear state of the system, with which we're basing these increases on and if there is no perception of a clear state of the system, we run into some situations. We don't know what to poll on on Friday and I wanna bring some clarity around that because this is kind of a hard one to wrap your head around. |
| # / 00:11:47 | Richard Brown | We have an executive, a continuous approval voting system, we have a situation where we have an existing state of the system has been ratified. We add a new executive vote and then people can, MKR token holders can decide whether they prefer the new system or the old system and I think that we've fallen into a trap where we assume that the new picture of the system is better because it's new and it's more recent and it has up to date thinking behind it. |
| # / 00:12:16 | Richard Brown | And, so it's easy to fall into the expectation that once this new polling ... voting option goes into the system, it's just a matter of time before people come to their senses, move their MKR over and then we move forward again. That is not a safe assumption to make. |
| # / 00:12:30 | Richard Brown | Now, we actually have three states in the system that I can sort of identify where we have a single vote in the system that has the vast majority of the staked MKR and that's the way the system has been configured right now, we have an incoming vote where there's this expectation that this vote is an improved state of the system and we're all waiting for people to move from old to new, that's state number two. And then we have state number three which is there is actually a disagreement where there's a significant number of voters who have no intention of moving into the new picture of the system, they're happy where things were and they're not going to move their MKR. That creates ambiguity because that means that we don't know which state of the system we're polling for. |
| # / 00:13:20 | Richard Brown | We could set out a poll and then people could change their minds and then start moving into a new state of the system and then ... mid poll. So, having a poll that is dependent on a known quantity for the system is very difficult. Actually, I'm gonna stop there cause I talked for like ten minutes about stuff that's might not be immediately familiar to a lot of people in the chat so, if there are questions or comments or immediate suggestions ... maybe everybody understands it and I'm just being dead horse here. Cyrus, you were chatting in the side, was there something that we need to address? |
| # / 00:13:57 | Cyrus Younessi | Someone suggested that they were planning on capitulating for the executive vote. So, if their option didn't win, then the governance will ... they would just vote for their next best preferred on the executive, so - |
| # / 00:14:09 | Richard Brown | Yeah, that raises another problem though, where we might run into a situation where we have ... an executive vote goes up and it doesn't get the majority of the MKR staked into it, so we have the current state of the system then we have a new sort of weak executive poll that didn't reach ... didn't get ratified and then a week afterwards, we might end up with another sort of weak executive poll which no one will ratify and so we might have this long list of executives that never actually get executed. And that creates even more ambiguity in the system, which is a puzzling sort of world to live in. |
| # / 00:14:53 | Matthew Rabinowitz | I guess this is a question for the way the system is architected and designed. Are we voting for a specific stability fee or are we voting for a change in the stability fee? |
| # / 00:15:04 | Richard Brown | Yeah, well that's a good question. So, that puts us into a position where we can start talking about are we voting for relative increases or absolute increases? So ... but that also leaves us in sort of this gray area because if it's absolute, there's a poll for raise the stability fee to eleven percent or something, fifteen percent and then at some point during the course of the week a new executive was ratified that is ... sets the stability fee to fifteen percent, then that poll is now invalid. Or, if the ... well, I guess that's enough of an example there, because of the sands are constantly shifting underneath, the polling system, even relative and absolute values have a potential for being signaled for but no longer appropriate and then we'd have to be in this position where we are presenting executive votes for a system that no longer exists or we're in a situation where we're vetoing polls because they no longer make sense and that's a risky situation to be in, I think. |
| # / 00:16:17 | Richard Brown | Does that make sense Matthew, or have I failed to understand your suggestion? |
| # / 00:16:21 | Matthew Rabinowitz | No, I mean that's it. I mean, I guess, I'm trying to ... like if you fast forward what you were just saying to the part of if you just had a vote every ... and I'm just making it simple right now ... even minus two, minus one, the delta is the question, the real number is the absolute. So, if we're at eleven and a half now, the vote maybe should have minus two, minus one, zero, one and two. And that's just the executive vote, and it should be a continuous role, but you're really not voting for the change. We really, in my view, should really be voting for what we think the actual end stability fee would be. We logically should know what the state was beforehand and so it's very similar to these telegram polls that they run, where you can retract your vote and cast a different changing vote and it's just a live continually voting system. |
| # / 00:17:15 | Richard Brown | Yeah, so that's good because that turns over a rock that I've been considering as well because we have this polling system, and I think that ... and I could be completely wrong and if somebody has deeper insight, please stop me before it's too late, but the polling system was initially designed to gather signals around things that don't actually result in a state change. So, the MakerDAO principles, or we should do X, Y and Z in the real world or something. The executive votes primarily are to execute an new state of the system and so one of the questions I've been asking myself is is it not the correct place to be polling for a range of options, when instead, possibly, there should be a range of the options that are simply placed into the executive voting system and that's simply it. |
| # / 00:18:05 | Chris Padovano | Wait, I have a technical question. I may be under the mistaken belief, but I thought that each individual MKR could vote on up to five different slates at the same time, is that correct? |
| # / 00:18:21 | Richard Brown | I have no idea. [crosstalk 00:18:23] |
| # / 00:18:26 | Chris Padovano | Cause I think you get multiple votes per MKR on different slates, or the way we're currently doing voting. So ... is anyone technical on the call that can answer that question? |
| # / 00:18:42 | Richard Brown | I'm not recognizing any names. |
| # / 00:18:45 | David Utrobin | I could ask in one of our internal channels. |
| # / 00:18:48 | Richard Brown | Yeah. |
| # / 00:18:49 | Chris Padovano | Yeah, cause for some reason I thought that each individual MKR could vote up to five slates, so you wouldn't have to ... presumably, if you're in favor of a three percent stability fee for one week and then in the executive vote you could still vote on the next slate while still locking in your three percent from week one. |
| # / 00:19:16 | Richard Brown | Yeah, so we'll need to dig into that. So, I've done, hopefully I've done an adequate job in explaining the conundrum that we're facing here. I don't expect us to spend the rest of the call digging into the minutia of this, I just wanted to put it on people's radars and encourage them to come into that thread that I posted in the chat and give us some suggestions. But, I'm gonna leave some ... I'm gonna make the water even murkier by adding some questions that are of standing, so ... We also ran into another issue where there potentially could be a disconnect between the demographic that's engaging with the polls versus the demographic that's engaging with executive. |
| # / 00:19:58 | Richard Brown | I think that it's worth considering that polling seems to be a friendlier option for smaller maker holders to create a signal, but executives fall prey to the commons issue of why should I bother if things aren't going my way, I can't move the needle, I'm not a whale, all the rest of those issues that are common to every voting system that's ever been created. But, perhaps people that are willing to vote in a poll are not willing to vote in an executive. So, we might be seeing a disconnect between signals and what large holders actually are hoping to see. It could be vice versa, large holders have to deal with custodians, they have a lot of multi sigs and they have to dig things out of safes in order to access to the ledgers. So, maybe they're willing to go through those hoops in order for an executive, but they're not willing to go through that hoop just to signal some information on a poll. It's hard to tell. |
| # / 00:20:54 | Richard Brown | The other major question we need to address is, does the polling system have a hard dependency on our recent state change, having successfully executed in the system? One of the things that I wanna get some clarity on is are we allowing enough time to pass between an executive and a poll, and this is something that we didn't get into today, but it's something we need think about. So, we have an executive on Friday and then the following Monday we kick off another poll, so the intervening weekend, is that enough time for the market to react, is that enough time for people to understand the new state of the system, is there any breathing room here, you know. And is it just a fantastically inconvenient time for people to be moving their tokens out of cold storage, something we need to think about. |
| # / 00:21:38 | Richard Brown | And this is one of the puzzlers for me I think is, how do we know whether a new picture of the system, a new vote, has been placed into the portal has failed or not, or has a lack of support? We have an existing picture of the system, we've put in an executive and after two days, three days, five days, it doesn't reach a majority, does that mean we can all assume that this is a weak proposal and is there any utility in even having it in there. So, there's been a lot of options thrown around where we can potentially start thinking about thresholding, this amount of MKR needs to vote in order to have an executive. This executive needs to have enough MKR in it to be valid and if after x number of days there's not ... there's like five percent of MKR in the new executive after five days, maybe it just gets cleaned out and goes back into the old executive picture and we clean up the system that way. I'm not sure, that might actually break some of the fundamental assumptions of how we've designed this thing. |
| # / 00:22:38 | David Utrobin | So, Mariano Conti just replied and he said that yeah, that's technically true so any MKR can vote on up to five slates at a time. |
| # / 00:22:48 | Richard Brown | Alright, that's interesting. So, Chris, why were you asking that question? Is there a solution here that I'm missing? |
| # / 00:22:54 | Chris Padovano | Yeah, cause we're talking about capitulating votes and moving votes around it seems that you don't need to - |
| # / 00:23:00 | David Utrobin | Hold on, actually I'm wrong. He said not five slates, one slate with up to five addresses. |
| # / 00:23:06 | Chris Padovano | Got it. |
| # / 00:23:06 | David Utrobin | Okay on different addresses. |
| # / 00:23:08 | Chris Padovano | So, the question ... yeah, then we're right back at square one. It doesn't ... |
| # / 00:23:15 | Richard Brown | Alright, so I'm not gonna bog us down with game theoretical conundrums today, that's now officially on everybody's radar and I've posted the link. So, let's continue the discussion. There's been some interesting discussion, there's been some interesting solutions presented in the chats already. They need to be fleshed out a bit, so, I'm going to leave that there unless somebody has a link to how to fix governance that we can all enjoy, where they have a magical solution for us, if not ... Cyrus did you just BRB us? Are you back? |
| # / 00:23:49 | Cyrus Younessi | I'm back. |
| # / 00:23:50 | Richard Brown | Oh, okay. I'm going hand it off to you cause you have something special for us today, so what's going on in the world of risk? |
| # / 00:24:00 | Cyrus Younessi | Sure. So, today we're gonna hear from as mentioned two members of our risk working group. Vishesh has spoken in the past and he will be kicking off the state of the peg today. We're also going to have Primoz with us to report on CDP statistics and I've realized we've never really had them give proper introductions, so Vishesh and Primoz if you want to give a quick bio please feel free. And, tell everyone else, as you all know, we're pushing hard on this decentralized risk governance approach. If you're interested in getting involved, please reach out to me and with that I'll hand it off the Vishesh. |
| # / 00:24:47 | Vishesh Choudhry | Thanks Cyrus. Yeah, so just as far as quick introductions I'm not sure if I've ever introduced myself properly on the call but I am computer science and biology by education, worked in, you know, the academic space in a lot of hard science research fields and then ... since then, got involved with a few tech startups, particularly a health insurance startup that I worked at for a couple of years where we did a lot of kind of analytics and risk-modeling and financial modeling. Since then, moved onto a crypto focused research firm last year and then just in the past month or so, I've kind of broken off and kind of started doing consulting on my own and publishing, you know, graphs and tweets and posts a little bit more publicly. Just kind of trying to get a little insight and analysis out there and into the space and into community calls like this, so we can help inform what's going on. |
| # / 00:25:52 | Vishesh Choudhry | Now, on the data side, let me go ahead and share my screen here. Okay. You can see these graphs. So, here was the price peg, so essentially the previous stability fee increase was around March 22nd and the most recent one was April 14th, so what we saw was basically the price peg had been kind of declining for a while and I think that rate of decline had kind of slowed or come to a head with a bit of a valley just before the most recent increase. But, that was following a huge spike in transaction volume. What we saw in the last few days since the increase was a drop in that transaction value but the ... again we should discount it because it's a lighter volume but there price peg, starting to recover a little bit from that sort of ninety five ish, ninety six ish low to about ninety seven to most recently and then I think Lev was showing similar numbers, on stablecoin.science. |
| # / 00:27:20 | Vishesh Choudhry | So, this was just sort of the same thing in the context of what's been going on with ETH price, so it is interesting to kind of compare. With the previous increase, we were dealing with an ecosystem were ETH price was rising. With the most recent increase we have sort of seen a few points dip in the ETH price. Now, I think the overall view is sort of lukewarm in the sense that it looks like we're starting to get a little bit more healthy stability back into the system, but again, with lower trading volumes and with ETH price having come down in the last few days, we should probably discount like how much we can attribute that purely to the stability fee increase versus other independent variables outside of the Maker ecosystem. |
| # / 00:28:10 | Vishesh Choudhry | So, and just like ... I think it's a continued trend that's been observed, a bit of an inverse relationship in terms of what's going in with ETH and what happens with DAI. On the supply side, so, I think positive again, indicator, but wanna keep things tempered is since the few days before the most recent stability fee increase, there was a bit of a ramp up in the outstanding supply. But, I think that tends to happen just before an increase and then after that increase that supply has come down a bit, mainly I think due to the large volume of MKR get .. of DAI getting wiped on April 15th, which was the day after the increase. I'd also ... a separate thing that is not shown here that I kind of tweeted out, just following a few transactions, there were a few big transactions where people were borrowing DAI on secondary platforms like Compound and Dharma and then going ahead and using that to pay down their CDPs, which I think was kind of the behavior that we try to incentivize with increasing the stability fee, is to just kind of make it a little bit more expensive so that people consider wiping down their debt. And they are in, what I'm seeing, taking advantage of some of those secondary platforms as a lever, or an alternative platform to obtaining DAI. |
| # / 00:29:40 | Richard Brown | Vishesh, can I ask you a quick question? Or, should I recommend? |
| # / 00:29:44 | Vishesh Choudhry | Yes. |
| # / 00:29:45 | Richard Brown | I'm wondering at what point would you feel comfortable saying that we have historical data at this point. Like can you ... how many more times do we need to do this before you can say that the last two to three stability fees had these effects and we just saw it affect again. |
| # / 00:30:01 | Vishesh Choudhry | Yeah, and there have been a few offline conversations on that topic, in terms of just how we're looking as timing and frequency as a variable in a lot of the research that we're doing here. |
| # / 00:30:13 | Vishesh Choudhry | I do think it's a little bit premature for us to start to say that we have trends and historical data. I think we're sort of at that stage where we're seeing some motifs, some patterns, that, you know, we can maybe start to make some very light predictions on and then kind of test the quality of those trends by how well they hold up to what we see in the next couple of iterations I would say. But, I think we certainly probably need at least a few more weeks before we can start to say anything for sure, because I think the real trick is we want to see a variety of independent environmental conditions, right. |
| # / 00:30:52 | Vishesh Choudhry | So we wanna see stable ETH price, we wanna see volatile ETH price, rising, falling, a lot of different types of conditions, both when we keep the stability fee steady, when we increase it, and we essentially wanna be able to control for as many different variables as possible. And, I think there's just been too much noise and too many independent variables changing and a lot of changes that we've been making, to really draw like historical trends or conclusions right now, it's just too [crosstalk 00:31:25] I feel. |
| # / 00:31:26 | Richard Brown | What ... I don't know what tool you're using on these graphs, but is it possible to get like a big vertical red line for every time that we've increased the stability fee? |
| # / 00:31:35 | Vishesh Choudhry | Yeah, I can put the stability fee ... I can demarcate it a little bit more clearly. These compound graphs have stability fee graphed as well, but I can drops some pins on the other ones for when those increases were. |
| # / 00:31:53 | Richard Brown | That'll be awesome, thanks man. |
| # / 00:31:54 | Vishesh Choudhry | Yep. So, just as mentioned before, transaction volume, big spikes just post the previous increase and then this was actually something that we had mentioned last week, was that transaction volume was healthy. It has been dropping since then, but I think that is a little bit natural in terms of what's been going on in terms of the outstanding supply. I think it's primarily been people wiping down debt which is the healthy behavior that we're looking for. |
| # / 00:32:25 | Vishesh Choudhry | So, what was interesting with the secondary lending platform rates was in the past, it had kind of been cheaper to get DAI from Maker. More recently, it's been cheaper to get ... Oh, sorry, it had been cheaper to get DAI from Maker, but now that we've made that a little more expensive, I think what we've seen is fewer people taking their DAI, getting it from Maker and supplying it on these secondary lending platforms. But, I think a healthy, semi-sustained level of people when they're in a crunch, going to those platforms and getting DAI and using it to wipe down their debts. |
| # / 00:33:09 | Vishesh Choudhry | So, I think the role of these secondary lending platforms has shifted is what I'm saying and I think it's shifted a little bit more to what it's supposed to be used for, which is kind of a last resort for getting ... and I know there's been a discussion about whether Maker should be the lender of last resort or not, but I think people have kind of been going to theses secondary lending platforms recently only when they really need to. Whereas, in the past, there's been a ton of activity of people just drawing out DAI willy nilly and then supply it on these platforms to try and get some kind of yield because it was really cheap to do so. |
| # / 00:33:47 | Vishesh Choudhry | So, I think this is a healthier shift, but there's probably a lot of debate about that. So, in terms of the volume, as I had mentioned last week there had been this kind of continual uptick of the supply and borrow volume on these platforms, slight dip just post the stability fee increase in terms of the amount being supplied. That has recovered a little bit, I think that was a lot of people essentially trying to reclaim access to some of their DAI to try and wipe down debts. |
| # / 00:34:23 | Vishesh Choudhry | One new addition, which ... apologies, it does seem a little bit messy, but when I zoom in it'll be a little bit cleaner. This is something that was discussed a little bit last week that I wanted to just show, to answer a couple of questions. The questions that we've gotten were like when people are drawing out DAI, are ... is it new CDPs that are drawing out DAI, or is it existing CDPs that are kinda levering up their amount of debt? |
| # / 00:34:52 | Vishesh Choudhry | And so, what I've just kind of done was looked at, okay, at a daily level how much of the DAI being drawn is being drawn with CDPs that were created, let's say, within twenty four hours, or CDPs that had already been existing, you know, more than twenty four hours ago. And so, essentially, what we see is ... and then I did just the nine day simple moving average and then graphed the ratio of the two as well. So, what we've seen which is kind of interesting is since just before the last stability fee increase, more draws on new CDPs and then that kind of falls off. And, then, just leading up to the next increase more draws on new CDPs. |
| # / 00:35:39 | Vishesh Choudhry | So, and this kind of is I think an indirect indicator of something else that I wanted to show, but will likely have to show next week, which is kind of the overall age of this debt and I think what this speaks to me is a little bit more circulation in debt being drawn and wiped down, which is, to me, healthier behavior in terms of people not sitting with so much stale, outstanding DAI, paying down some of those debts and then having some new people come in and draw out DAI when they need a loan. |
| # / 00:36:13 | Vishesh Choudhry | So, I think this is a potentially healthy indicator, but also, given what's been going on with ETH price, I always wanna kind of qualify, because when ETH price is going up, we tend to see people want to lever hard on DAI and then ... lever hard on ETH and use DAI as a pathway for doing so. And then when that ETH price starts to fall a little bit, we do tend to see ... we did see some liquidations as well, a little light compared to in the past obviously, because it was a very small drop in price. But, in the past month or so we have not really seen very many liquidations. So, again, that drop in ETH price does tend to be a healthy moving the sludge through the pipelines kind of factor for Maker, in my opinion. Sorry, i don't wanna take up too much time. So, I think that was the gist of what I wanted to show today. |
| # / 00:37:07 | Alex Evans | Actually, it's up to you. Do you wanna pull up the compound, I just have one really quick question on that. |
| # / 00:37:12 | Vishesh Choudhry | Sure. |
| # / 00:37:14 | Alex Evans | If that's okay, I don't wanna take up time. |
| # / 00:37:20 | Vishesh Choudhry | The volume or the rates? |
| # / 00:37:22 | Alex Evans | The rates, the rates. So, this is saying it's still ... so, what's the difference there between the borrow rate and the Marker stability fee? |
| # / 00:37:35 | Vishesh Choudhry | Yeah, it's been about thirteen, it's dropped recently to like twelve six, versus the eleven point five stability. |
| # / 00:37:42 | Alex Evans | We're getting close though. The one thing that I'm trying to understand as well is ... so, my understanding is when you borrow from Compound, the collateral that you post, you're also supplier for. So, you're earning interest on your collateral. Is that correct? |
| # / 00:37:55 | Vishesh Choudhry | So, my understanding was you have the ability to do so. I was not sure if it's like automatic function, but broadly, yes, that's I think how it works. |
| # / 00:38:04 | Alex Evans | So, technically, potentially you net the interest rate that you're earning on your collateral. It may be the case that the borrow rate is very, very close to the current stability fee. |
| # / 00:38:16 | Vishesh Choudhry | Yeah, I think that's - [crosstalk 00:38:19] |
| # / 00:38:18 | Alex Evans | So, let's say you're earning a percent on it right, then they're about equal. |
| # / 00:38:23 | Vishesh Choudhry | Yeah, so I think you always have to discount the borrow rate. And so, if you discount from twelve six versus eleven five, it's probably very close, it think that's accurate. |
| # / 00:38:35 | Alex Evans | Right. No, it is, you're paying ... your debt is lower than your collateral by value so one of them is ... maybe you don't need a lower interest rate, but if there's more of it and so potentially it could even be higher, depending on what the rates are are, I don't know what the supply rate is for ETH or whatever equivalent on Compound right now. Like, yeah, I would also be curious right now that we've seen a lot of Dharma and potentially looking at those rates here more seriously as a thing. Though, in the past, we have discussed the issue of subsidy and looking at other secondary lending platforms that are now getting some volume and looking at data from those. But, anyway, that's a good thing, so potentially we could be coming to a point where it's ... where we're getting close between Maker and Compound. |
| # / 00:39:26 | Chris Padovano | I mean Dharma right now is giving you eleven percent like Joe was saying, but, you know, definitely some loss leading going on right there. I'd encourage everyone to just go use the interface to see that, you know, there definitely subsidizing those rates up to a million bucks. And, I wounder if Compound is doing something similar. |
| # / 00:39:54 | Cyrus Younessi | Alright, well thanks for that Vishesh. I guess before we jump into kind of the conversation, let's have our second presentation happen as well and then we can discuss after. |
| # / 00:40:08 | Primoz Kordez | Yeah. Hi guys, can you hear me? |
| # / 00:40:11 | Cyrus Younessi | Yes, we can. |
| # / 00:40:12 | Primoz Kordez | Okay, cool. So, just really short introduction of myself and my team because I have a lot to show you today and we don't have much time. So my background is in financial risk management. I currently run company called Block Analytica which focuses on crypto fundamentals, research and investing. We're a team of three and we're based in Slovenia in Europe. |
| # / 00:40:35 | Primoz Kordez | Before, I start sharing my screen, let me just say we intend to develop tools to measure this kind of measures that I'm going to show you today. I made this in Google Sheets and it's not automised yet, so hopefully you will soon such things online, and in more user friendly environment. Further bureaus are working on addressing DAI on chain transaction and white listing certain addresses to get familiar with DAI demand because I believe currently DAI demand is mostly unknown and the most important ... yeah, you could call it variable, when addressing supply and demand imbalance. |
| # / 00:41:13 | Primoz Kordez | Okay, so let me start sharing my screen. Okay. Hope you can see this. So, yeah, basically I know that CDP debt is pretty concentrated, around few hundred CDPs, although there is more than, I think, two thousand active CDPs out there. My goal was to see what amount of CDPs had some sort of impact when it comes to debt issuance and the repayment on daily basis. |
| # / 00:41:43 | Cyrus Younessi | Sorry, can I cut you off for a sec? Do you mind zooming in just a little bit? I don't know about everybody else, but it's a little bit hard to see. |
| # / 00:41:51 | Primoz Kordez | Better? |
| # / 00:41:53 | Chris Padovano | Should be [crosstalk 00:41:54] |
| # / 00:41:54 | Cyrus Younessi | A little bit more. |
| # / 00:42:00 | Cyrus Younessi | Yeah, that's fine. Thank you. |
| # / 00:42:02 | Primoz Kordez | Okay. So, we can see here that twenty ... two hundred fifty largest CDPs, they already represent ninety percent of whole debt. Nevertheless, for the research we used a thousand CDPs, the whole history of them which corresponds to ninety eight percent of whole debt. Here we can see historical data, but obviously I'm not going to go through the variables here. |
| # / 00:42:29 | Primoz Kordez | So, this is one the first interesting things. We focus on last 10 days because there was a lot of wipe activity in this period. This graph is pretty much similar what we use to look at mkr.tools, but I converted everything to DAI so it's easier to compare. As expected, there was a lot of lock and draw activity in the first half of the period, when ETH price was rising and there was a lot of free and wipe activity in the second half of the period when ETH was falling and also stability fee was announced to get increased at that point. |
| # / 00:43:10 | Primoz Kordez | What we wanted to do next ... Sorry. Yeah, what we wanted to do next is to see how much of this activity, so this, our total DAI numbers, values, for each type of event, so lock, draw and wipe and we wanted to see how much of this activity is presented by top single events. So, to see just like the maximum wipe on particular day, how much of total wipe on that day this will present that. And, interestingly, it represented about third of the activity and it's only one event. |
| # / 00:43:51 | Primoz Kordez | So, we also looked at top five events for each day, again, for every event state and averages are pretty high. You could say like five top event represent three quarters, at least three quarters of all activity on that day. And for specifically, for wipes which were pretty high on those days, on the second period, yeah, these top five events represented like three quarters. But, these five events are not necessarily from five different CDPs, so here specifically, this were only two CDPs. So, I think it's ... this huge concentration of activity per CDP is something that's probably going away with time. |
| # / 00:44:41 | Primoz Kordez | But, on the other hand, it's kind of a good news from research perspective because it's easier to analyze if wipes actually affected the peg because we make more detailed on chain analysis of these CDPs and look if they really behave as it was intended by policy setting. So, this was exactly what we wanted to do next, perform some kind of on chain analysis or to see if these CDPs on this particular day, if they actually freed ETH from collateral, used it to buy DAI on the market and actually helped to restore the peg because this was the point of increasing the stability fee, right. So, usually ... The usual pattern you can see on chain when people ... when CDPs build leverage, they make these iterations. They draw DAI and then they convert it to ETH and then they lock ETH and they make bunch of iterations and they built leverage. |
| # / 00:45:39 | Primoz Kordez | So, we actually wanted to see if, when there were wipes if we could see similar pattern occurring and if people actually bought DAI on the market and if this actually helped restore the peg and we specifically focused on just two days and this were 14th and 15th of April because you could see a lot of wipes here. And, also, on this two days we could see that DAI price increased, so I was hoping that I would be able to connect wipes and people actually buying DAI back to repay the debt with this increase of DAI price and higher volume. And it wasn't necessarily the case, so, let me show you this. So, these are two CDPs on those two days. Just a second. So, on those two days there was four million of DAI being repaid, there was four million of wipes and four million point four of ETH in dollar value being created. And, eighty percent, up to eighty six were just from these two CDPs. So, you just look at on chain of these CDPs and see how they behave. |
| # / 00:47:01 | Primoz Kordez | So, the first CDP we found out that it was actually free from the collateral and it was used to buy DAI, but this CDP used HitBTC to do that, which was interesting to me. So, he actually deposited like in many chance ETH to HitBTC converted it to DAI and repaid his debt. Now, why did you do that? Obviously because of higher stability fee and also because ETH price was falling at that point, I wouldn't say his collateralization ratio was low, that he needed to do this. |
| # / 00:47:46 | Primoz Kordez | So, I guess that is the good news. He actually did it, but perhaps not so good thing is that he actually used HitBTC, so, that's why we can't fully see this straight here. If we were including this data, HitBTC, obviously on this two days, price of DAI should be a bit higher so it would look much better to state in other words. And, also, one interesting fact, I was actually amazed HitBTC has ten million of DAI deposited there. So, obviously some people are actually using HitBTC to trade DAI it's ... I am not sure, there are rumors about wash-trading but just one interesting thing about it. |
| # / 00:48:32 | Primoz Kordez | The other thing is the second CDP, so we believe wipes of this CDP were associated with low collateralization ratio because it almost fell to one hundred sixty and I guess the owner needed to lower his leverage and he actually made this iterations that we usually see when people build up leverage and followed it again on chain. Why did he ... how did he perform the ... how did he bought DAI and we couldn't actually see any of the major venues, be it the decentralized exchanges or centralized exchanges. We could see where the funds came from, but we're not sure, was it ... It could be that there was OTC desk in the middle, it could be that were some spare DAI somewhere although I don't believe it. Or, it could be that it was borrowed on some secondary lending platforms and I'm pretty ... really not sure, it's hard to say. But, we are really certain that none of these wipes on these two days where we thought, okay, this is helping with the peg. It didn't actually, this trades that we see here, I guess they were mostly associated with some guys going, hedging their positions and buying DAI because ETH was falling at that point. I guess it's related with that. |
| # / 00:50:02 | Primoz Kordez | So, that I would say is the main conclusion about it. Just the last thing and then I've finished. This is something that we played a bit recently. It doesn't have to do with what I just presented. It's just we tried to estimate the leverage of CDPs, I think that's one metric that I'm kind of missing is basically something you can adjust from collateralization ratio, but what I was curious about was how much people actually lock their own money inside versus how much of the collateral they have, right, and collateral is of course made by borrowing. So, that's the point of leverage long. And this number is aggregated number for thousand CDPs and I was ... I thought it was going to be higher but it's not really, it's one point four six and I think it's a good metric to have because if you can monitor it through time, you can assess some kind of risk position of CDP holders, how they behave two time. |
| # / 00:51:09 | Primoz Kordez | And that's basically all from my end. I guess I already spent a lot of time so if you have any feedbacks, you can also find me on Rocket Chat, or on Twitter. |
| # / 00:51:20 | Cyrus Younessi | Cool, thanks, that was awesome Primoz. Thanks for taking us through that. Couple questions from the chat, you may want to address some other time, but a lot of people asking if you can share the spreadsheet, maybe in just read only mode. |
| # / 00:51:35 | Primoz Kordez | Yeah, I was actually thinking to do that, I just need to decide whether we should publish a blog with it and then I'll need a day or two or ... definitely, I'll publish it. |
| # / 00:51:49 | Cyrus Younessi | Cool. So, my first takeaway, or my immediate takeaway from this was that, you know, earlier this week when we saw like three, four million DAI reduced, you know wasn't ... it was just from literally just a couple people they were just going to HitBTC and to me that doesn't really seem like the kind of response we're hoping for. Does anybody else agree? I mean, do you guys get what I'm trying to say, I mean we're not triggering any kind of mass closures or big supply reduction in any meaningful percentage of the community, it just seems like just the fringe, kind of like the fringe market makers who are just still playing with idle capital, you know, just kind of ... |
| # / 00:52:41 | Richard Brown | Well, is that the assumption we should make? Cause I've been trying to wrestle with this myself and do we have a distributed demographic of ... at the top end, we have people with a high or very risky profile and/or potentially smart money behaving in a vastly different manner than the majority of the other CDP holders and if that's the case, is it possible for us as a community to try and find actions that would move them without moving everybody else at the same time, if that makes sense. I'm [crosstalk 00:53:14] |
| # / 00:53:13 | Cyrus Younessi | I don't think that's possible. |
| # / 00:53:15 | Richard Brown | [crosstalk 00:53:15] ... that cannot be gamed as though, obviously, you look at like here's large positions, what if we have, some kind of magical stability fee that just applies to whales or something. Obviously, stupid idea, but all these kind of fixes are gameable in the extreme so it's a tricky one, right. |
| # / 00:53:45 | Richard Brown | Michael, what did you ... You just posted a link there, what's the genesis of [inaudible 00:53:48]? |
| # / 00:53:50 | Michael Dunworth | Oh, hey guys. Thank you very much for all the data points, it's super interesting. I think a lot of the data points are sort of these very noisy kind of, you know, it feels like they're very noisy. |
| # / 00:54:05 | Michael Dunworth | When you break it down, everything comes back to trying to determine how risky the system relative the underlying asset and how is that underlying asset performed ... performing int the market and so, basically, the biggest indicator that I can see that gives you the most amount of, you know, data from that which is gonna be wipes, draws, dies like all that kind of stuff. It all comes from the collateral ratio and the collateral ratio, I posted about it and I've thought about it. |
| # / 00:54:41 | Michael Dunworth | I put a post in the MKRgov sub Reddit and then I've talked to a lot of people on Rocket Chat about it, but I think all these things considered, you need to really, really emphasize that being the indicator of how you determine the stability fee. You know, the stability fee's gone really ... I mean, sorry, the price and the peg has gone really wonky over the past, let's call it, six weeks and in that six weeks, you know, looking at that spreadsheet that I just put. The first time that we've actually had any kind of recovery is when we've actually been within those upper and lower bounds in our ... like when the stability fees actually fit those upper and lower bounds, I mean it's not ... I haven't been super diligent on it, just cause I don't have time, but I would really, really, really, really encourage people, if they can, to sort of crack open that rabbit hole. And, I think what they'll find is this is the best guiding compass for us determining what the stability fee should be. So - |
| # / 00:55:49 | Richard Brown | Michael, those are fascinating topics of discussion. With the top of the hour though, I don't want to silence you, so if people want to hang out and talk about whether the collateralization ratio is the magical number that we haven't been paying attention to. Let's continue the call afterwards, but, it's late in a lot of places in the world and people have meetings. So, Steven did you still want to give us a thought, before we shut things down? |
| # / 00:56:14 | Steven Becker | Well, yeah, by all means. Firstly to Vishesh and Primoz, that was brilliant, thank you very much for, you know, participating and obviously Michael I do not wish to cut you off ever, so I'm gonna keep this short and obviously if folks want to, you know, keep on, so be it. And I'll also leave it up to Rich, but again, thank you very much for you participation folks, really appreciate it. |
| # / 00:56:39 | Richard Brown | Yeah, that's great. Thanks everyone for joining. I'm not sure what we do in this situation because we've had a couple of really interesting conversations post call. I've been in the habit of stopping the recording and ... but potentially people might be missing some interesting things, so maybe today is an experiment and we'll continue to record if people have questions they wanna ask and see how that shakes out. So, Michael like I'd be interested in hearing a bit more because in my mind the collateralization ratio is sort of this following indicator, and I'm not sure if I'm using that term correctly, of overall health of the system, the recent changes in the price of ETH and its sort of knock on effect on how healthy or unhealthy positions are. Using that as an indicator that shows an immediate ... |
| # / 00:57:29 | Cyrus Younessi | [crosstalk 00:57:29] Well, I think ... |
| # / 00:57:29 | Richard Brown | ... system is what freaks me out about. Sorry, go ahead. |
| # / 00:57:32 | Cyrus Younessi | I mean, the collateralization ratio is a function of the ETH price, right, as the ETH price rallies people have basically higher risk preference and it just, it really ramps up the DAI creation. But if they're, you know, this is obviously the problem we keep coming back to and, you know, if we don't have an accompanying increase in the demand for that DAI that's being created, then we're gonna see DAI price suffer, so the solution [crosstalk 00:58:03] |
| # / 00:58:04 | Richard Brown | [crosstalk 00:58:04] ... talking about this CR as being like a symptom, two steps down. Why don't we just focus on ETH price being like the thing that's really - |
| # / 00:58:10 | Michael Dunworth | Well, the ETH price is going to determine ... If you look at the PETH pool of how big the PETH pool is, right, which is the underlying count of assets, so if you go, you know, two million ETH or let's call it PETH whatever, that's been relatively constant and the thing that swells and, you know, expands and contracts is gonna be the collateralization ratio because you look at the consistency of the PETH pool is about, you know, within one or two percent and that basically says people aren't putting more in. Or, the system's not getting more or, you know, having more withdrawn overall like within one or two percent, yet, for some reason, it's swell up to four hundred and twenty percent collateralization ratio, down to three hundred and fifty percent and that's ... It becomes this very organic indicator that is the core foundation of everything which is, are people growing the system and putting more PETH in or are people getting out of the system? |
| # / 00:59:11 | Michael Dunworth | So, all these CDP activity, they're all a function of how this sort of like big, you know, system indicator is going to react to it. |
| # / 00:59:29 | Richard Brown | Okay, yeah. That's interesting. I'm hoping people that understand the system or the interplay between the market and MakerDAO can chime in with some deeper insight. It just still feels like this is ... the collateralization ratio is a blending of a few others, risk appetite, recent prices in ETH - |
| # / 00:59:47 | Michael Dunworth | But what determines the risk appetite? |
| # / 00:59:51 | Cyrus Younessi | So, Michael, can you distinguish between, do you think people are borrowing because they have a lot of room to work with, like they have a lot in assets and they can borrow against it, or because they think ETH is going to keep going up? So, let's say ETH was worth ten times as much as it is today, so the demand for levering up on ETH specifically is gonna be much lower, but you'll have a lot higher collateralization ratio. So would that necessarily lead to a weak DAI price? |
| # / 01:00:29 | Michael Dunworth | It wouldn't lead to a weak DAI price if there is not any ... if there's no ... like if there's a better on ... if there's a better discovery of the pricing mechanisms, so basically liquidity being sourced, you know, cause that's going to determine ... You could have a five thousand percent collateral ratio, and if it's a one to one peg between the environment in which people trade and exist, which is going to be the on chain markets, versus the off chain oracles, that delta, if there's no delta between those, you can have a zillion percent collateral ratio and it's not going to affect the stability fee. Like, it's still going to be one to one and it's gonna be smooth sailing. Cause basically ... Well, the premise that I'm talking about is oracles are one thing, the collateral ratio is a very organic indicator of the environment itself and if we look at on chain trading where we're saying, DAI to ETH should be the same as USD to ETH. |
| # / 01:01:33 | Michael Dunworth | It's sort of like these markets that ... every market is different, so government restrictions in China means China might be trading at a premium, KRW was trading at a premium in 2017 because of these capital outflows, it doesn't mean basically, there is different ... every single country or region has different, whether its through put restrictions, regulatory restrictions, so when you try and proxy in the, you know, this is what the price should be, that's like trying to say KRW in Korea, the price should be this. And, it's like look, whatever dude, that's America. In KRW, or South Korea we have regulatory restrictions and capital outflow restrictions so we're gonna pay a thirty percent premium. |
| # / 01:02:21 | Michael Dunworth | And so, when you look at on chain discovery, price activity, you need to be very conscious of that and that's what this model basically takes that into account because otherwise you're fighting gravity basically. It's, you know, you can't compete against the environment and whether that's through put constraints, regulatory accessibility and it ... Yeah, I think that's largely why there is a lot of ... that's factored in in this model, yeah. I don't know if that makes any sense to anyone, but ... |
| # / 01:02:55 | Richard Brown | Yeah, thanks for that. This would be back testable if it was true though, right? |
| # / 01:03:05 | Michael Dunworth | Yeah, I mean I've only put the numbers there when I've like had a chance to like drop them in but yeah, you guys could back test it. |
| # / 01:03:15 | Cyrus Younessi | I mean, it's clear that, at least to me, I think that, you know, the solution is still the same, which is still having to make borrowing more expensive but so I mean, do you have thoughts on how we could use this? Like, let's assume we were at a dollar right now, and then we were to see a, you know, a sharp ETH move, you know, how can we ... I mean we would have to create a model that would, you know, input ... so, I mean but this ... isn't this just going back to just inputting the ETH prices as a factor into determining what the stability fee and everything should be. I mean if the collateral ratio is just derivative of ETH price. |
| # / 01:03:57 | Matthew Rabinowitz | Yeah, but collateralization also implies that you've got a constant, basically, more or less a constant like you were mentioning earlier Michael, about having the number of ETH that's locked in, that PETH, right. I mean right now it's kind of somewhat tapered out around two point one million and it hasn't really launched higher. So, extract what you want from that. Maybe there are ninety eight million ETH holders that don't feel comfortable locking their ETH in a DAO, but if you end up having significantly more collateral accessible but just not locked in, doesn't that screw with the numbers? |
| # / 01:04:30 | Vishesh Choudhry | So, one ... I think that that's broadly the right instinct Matthew. One thing that I would say is volatility versus just absolute level is an important thing. So, what I think is a big driver of collateralization ratio is yes, the absolute ETH price, but also the volatility of ETH and so essentially, when we see ETH being very volatile, we tend to see a corresponding relative increase in collateralization ratio. When we see ETH prices are high, just based on simple math, we see an absolute increase in collateralization ratio and one of the different [crosstalk 01:05:13] - |
| # / 01:05:13 | Michael Dunworth | Yeah, and that's because the underlying asset hasn't changed as aggressively so if you've got a one or two percent, you know, deposit or withdraw into the PETH pool, that's the system effectively and if that's not changing and that's constant as like a count of assets in the pool, then ETH can run up a thousand percent and the pool is gonna keep blowing up exactly in line with the price. But, if the deposits of ETH are running parallel, so the actual count, if it goes up a thousand percent or ... yeah, it goes up a thousand percent in price, then, if you have a thousand percent increase in deposits into the PETH pool then they should be almost one to one. Just thinking on the top of my head. |
| # / 01:06:12 | Cyrus Younessi | Well, I mean, I'll have to think about his for a little bit. |
| # / 01:06:15 | Michael Dunworth | I also, I also .. yeah, I know it's a bit wacky or whatever but it's going to be very, very challenging when there's not sort of consistent, defined parameters of what qualifies as a trading venue that is acceptable for an oracle price, or a price feed in general and what does not qualify for it and the reason why is because, you know, if the oracle feeds are coming in from all these places, some places have wash trading, some places have price restrictions, withdraw restriction and all these kind of bells and whistles so if there is any kind of parameters that are available to the community, I think that would be really helpful as well, cause I think the aim of the game here is, before everything else, it's very hard to test or, you know, use data if we can't find any kind of statistical significance and it's really hard to find statistical significance when we don't have anything, you know ... a lot of consistency or continuity between what's being measured. Yeah. |
| # / 01:07:27 | Michael Dunworth | But I mean, look this is all like a really ... not by the way, please don't take this is criticism. This is exciting as hell and I'm thrilled to just like throw, you know, different thoughts out there. I'm not saying I'm right or anything, but just fostering more ideas to help us move forward as a team. |
| # / 01:07:46 | Cyrus Younessi | Yeah, absolutely. I don't think anybody is construing this as criticism. No, yeah, it's an interesting idea, for sure. |
| # / 01:08:05 | Richard Brown | So, I'm not sure what we do at this point, cause we're in uncharted territory so, should we commit to like chatting for the next twenty minutes? Or, do you wanna wait til things just wind down? Should we stir up the pot with some more hard questions? |
| # / 01:08:15 | Matthew Rabinowitz | But also ... let's maybe rewind for, you know, basically an hour beforehand when you mentioned where things ended up with the poll, kind of left it into an uncertain area. So, what do we do? Do we have an executive vote tomorrow? Do we do another poll? |
| # / 01:08:32 | Richard Brown | That's a great question, so ... and this is something that we've been tossing back and forth and I think that there's an enormous amount of risk if we just kind of go, okay, we have this governance system, everyone's kind of agreed to it, we've committed to it, we've talked about it endlessly but this time we just don't like the way it's gonna turn out so we're not gonna do it. I think that's gonna be ... the blow back would be fantastic, so, we're committed to this system. |
| # / 01:08:55 | Richard Brown | The community has designed it, the community has voted for it and so the next steps are we go ahead with the executive tomorrow, with whatever happens to have shaken up in the last hour, which I haven't looked at yet, but I think we're gonna do a three percent increase executive vote tomorrow at 9am PST and then we see how it goes and the next steps, I'm guessing are is if we still run into this situation where there's ambiguity, we have weak executive votes, we have polls that don't quite show any clear direction. Then we start ramping up some of the options that I wanna keep on exploring in the MKRgov link, do we have thresholding? So, for the poll, that is twenty percent of all maker need to show up before that poll is valid or does one of the poll options need to be twenty five percent higher than the second best option? I don't know. |
| # / 01:09:51 | Richard Brown | Jaikut presented a very interesting idea in the chat as well, which I haven't had time to completely digest but seems like it might be workable, where he talks about first past the post voting, which like everything else in the governance world everyday is a new euphemism that I need to learn. It seems like that could potentially be an option as well, but I think that the course is set now and changing that would sort of create a crisis of confidence in the way that we have governance at MakerDAO, to not do it because we don't like it. Does that make sense Matthew? |
| # / 01:10:34 | Matthew Rabinowitz | Yeah, most definitely. So, I guess one of the questions would be what if it doesn't after ... let's imagine, you know, just you know, very theoretical that it doesn't hit that enough votes of the three percent stability fee increase, what point does it actually activate is kind of my general question and then if it doesn't hit that, but we yet hit another poll a week later, is that what you're saying would be the window where we'd kind of stand down? Or? |
| # / 01:10:58 | Richard Brown | Well, that's where all the ambiguity comes in and I don't wanna make those decisions arbitrarily, so people would need to ... there would need to be a significant amount of engagement with the thread in MKRgov that just presents a clear solution to us, I think, because these are the risks I was discussing at the top of the hour where ... or the last hour, where we could potentially have four or five, well maybe I'm exaggerating, but we could have a lot of very weak proposals in the executive poll and then the executive part of the portal and nobody has a very clear understanding of what the picture of the system is. |
| # / 01:11:31 | Richard Brown | We're no longer ... the community is not longer providing forward guidance, like this is the way things are going because there's just indecision at that point, so, we need to figure out, do we expire executives that get no engagements? Do we need to organize more campaigning in the community where like get out to vote stickers or something where people actually signal. I'm not sure. But there's a lot of things we need to decide. |
| # / 01:11:56 | Richard Brown | I've posted ... well, it's a good segway while there's some more uncomfortable silence is I've posted a link to the discussion again in the side chat and I've also posted a link to the first post suggestion by Jaikut. My expectation is we spend the next week, just aggregating opinions and then, in the next call, we have a clearer picture of which way the community is leaning as far as solving this problem and whether it really is a problem. Look, there's all kinds of things on the table, the other one that we just kind of came up with as the call was proceeding is like is the polling architecture a bad match when it comes to determining what a risk parameter should look like. Maybe that's just ... maybe polling is for intention or business changes or foundation changes and executive votes are simply ... I mean, we go straightly into an executive vote for a parameter change, I'm not entirely sure. |
| # / 01:13:19 | Richard Brown | Yeah, I think that Christopher, you've signaled some support farther up. Unfortunately I lost track of the questions in the side bar, so I'm wondering if things got missed there. But, I think that also leads back into this idea that I'm very, very, very hesitant to allow these meetings to be a decision making forum, I don't want ... we've already seen this misunderstanding in the community where people have referenced these calls as the governance committee and that's kind of antithetical to the entire thing that we're trying to achieve here. This isn't where important people gather to make decisions for the rest of the ecosystem, it's where people with deep insight get together, ask some really cool questions and then return that back to the wider group for decision making. |
| # / 01:14:19 | Richard Brown | So, what do people think? Should we open up a can of worms? Or should we consider the conversation sort of organically having wrapped up and then leave it where it's at? |
| # / 01:14:30 | Richard Brown | Cyrus, do you have any other major topics that you wanted to get into? |
| # / 01:14:34 | Cyrus Younessi | Not really. |
| # / 01:14:35 | Richard Brown | Alright, I think that it's been a long week, people have things to do, so let's stick a fork in it there. Alright, thanks everybody. I'll say it five times. Continue the discussion in the MKRgov, I'll be posting a summary there, I'll be posting a bunch of links, I'll be posting links to Primoz's spreadsheets and Vishesh's graphs and video, audio and transcripts as well, so, let's continue the discussion. Thanks everybody. |
| # / 01:15:12 | David Utrobin | Thank you for hosting the call Rich. |
| # / 01:15:15 | Richard Brown | No problem, it's my job. |