Governance and Risk Meeting: Ep. 47 (August 8 - 2019)¶
Video | Audio | Discussion
| References | Person | Text |
|---|---|---|
| # / 00:00:00 | Richard Brown | Hello everyone. Welcome to the August 8th edition of the Scientific Governance and Risk meeting with MakerDAO. My name is Richard Brown. I'm the head of community developments. In these calls we have a lot of interesting things to talk about. We have a lot of interesting people talking about it and it's very important to us that we collect as much feedback as we possibly can. So please do not be shy about speaking up and making your voice heard. If you have a question, please type it in the chat and David or I will review it periodically to see if there's something we can surface in there. But like I said, please don't be shy about interrupting us. This is a meeting just like any other meeting, otherwise I can all day long [inaudible 00:00:49] like mine. |
| # / 00:00:51 | Richard Brown | So there's no special privileges here so just [inaudible 00:00:55] make your voice heard. Today we have a relatively eclectic agenda. I'm wanting to talk about some governance stuff for a few minutes. Hopefully I'll keep that promise and not talk for more than that. Then we're going to hear from Cyrus from risks and Cyrus is going to tell us a bit about why I asked that priority lists or a thing and why that's important to the organization. Chris Bradbury is on the call from the product team. He's going to fill us in on some of the more sophisticated reasons why the new portal should be interesting to this group. David is going to run us through the governance at a glance that's that we've lifted completely from LongforWisdoms, great efforts in the forum. |
| # / 00:01:38 | Richard Brown | He's going to give us an update on what's been happening there. Vishesh will give us an update on the state of the peg and then we're going to stick around for Q&A session. We've been there's lots to go over, lots of things to discuss. The last months, few months I guess we've been sticking around for 30 minutes post call to just ask random questions and it's been working out really well. So if you have something you want to talk about that is not on the agenda please hang out with us and we will talk about that. I can hear somebody breathing, does that mean you're going to talk or is that mean? I think I just shamed them into muting. All right. Cool. I am going to kick into my agenda. So we, Matthew was just asking a question about it touched on emerging process and how we are going to do things. |
| # / 00:02:37 | Richard Brown | And my answer to that was, I don't know, but it'll be interesting to find out. That doesn't create a lot of consumer confidence in the space. But I think it's should, and one of the reasons why it should is because we have evidence of the system working. We've been iterating on what the governance process looks like for what seems like forever now and happily, the governance process is improving all the time. Which is nice. And so as long as we maintain that velocity, as long as we maintain that agility and as long as the governance ecosystem is better than it was last week or the week before or the week before that, we are doing a great job. And that is the case, in my opinion. Let me segue back into this thing. Two segues ago. |
| # / 00:03:22 | Richard Brown | We have been talking about an asset priority list for a while and we've talked about it in blogs and forums and Twitters about how we need to figure out what to do first that. That I'm going to let Cyrus talk about in a second. But what I want to talk about is how do we assemble that information and how do we organize this group and the wider Maker holding community to align on an idea and to begin to do work and to signal their intentions. And it's been a running theme in these calls that the way that that happens is we start talking about things and then we make decisions based on those talks. Those talks happen in the forums and I'm going to post a link to the chats of where this particular conversation occurred. So we have an asset priority poll. |
| # / 00:04:15 | Richard Brown | A thread in the forum, in the governance section, tagged happily with the signaling text and people know that we're looking for signals. What I did here is explained that we have an asset priority list poll coming up. Offered three different suggestions for what those polls would look like. The reason I did that is because we have vastly improved polling portal, governance portal with a vast improved polling system, which Chris will talk about in a bit. But it's not finished yet. It's a great iteration. It's better than it was before. It's not quite there yet. But so we're in this world where we do not have a mechanism available to us that allows for dragging and dropping a list of things into an order of priority. |
| # / 00:05:04 | Richard Brown | But we do have a mechanism that allows people to signal their interest in a thing or a group of things. So the question becomes how do we use the polling system that we have to determine a priority list. Three options presented where we have one poll for all of the assets. And so there's one poll we're going to prioritize, how would you like to prioritize these seven items. And then we allow people to vote and then we order that list based on stake weight, which is, not basically. |
| # / 00:05:36 | Richard Brown | This is fundamentally how the governance system is designed one Maker, one vote, more Maker, more votes. And so whichever assets have the most stake weighted to them are ones that get bitten off first. The other option is we have a yes or a no poll for each one of these assets. Instead of one poll, with seven options. We have seven polls with two options in each one. One poll for each one of these asset types. This allows voters to express their interest in each one of them. If somebody comes in and votes no on a bunch of asset types, then it pushes us into the negative range. Then whichever asset types have the least no, yeah, the least no votes will keep. |
| # / 00:06:24 | David Utrobin | And currently that's the winning option as of right now. |
| # / 00:06:27 | Richard Brown | Yeah, that's a bit of a spoiler alert because I'm not finished. But the third option is that we have a weighting poll for each one of these assets. And then we'd have seven different polls. And instead of a simple yes, no, people would be able to vote high, medium, or low for each one of those assets. And then we can apply a weighting to that list and then do some math on the back end to figure out that like these three are high, the rest of them were all low. And so what's easy to do some math to figure out what that actually shakes out to be. And then so now that we have this, the proposals in the forum. We have the list of options in the forum. I followed the rules of the forum by referencing the communities guidelines about how polls are meant to be done, which was enormously useful. Anytime the community wants to do work that I don't need to do it. |
| # / 00:07:24 | Richard Brown | That makes me so happy, you would not believe me. And so I followed the rules. I formatted things the way that I was supposed to do in the language that the community said that we should be doing these things. And then we initiated this poll and then some interesting things happened. We set up this poll to last for seven days, I believe that poll ended three days ago. And people signaled their support for those three options that I just discussed. And what happened was option number two, the straight up, yes, no call for each option came out ahead with 43% of the votes. The weighting poll, which I was initially a big fan of, but then I eventually turned on myself and voted for a different poll, came in second at 38%. And then the one poll for all of the assets was a 19% poll. So we had 21 voters came out to vote, 43% chose option two. |
| # / 00:08:23 | Richard Brown | And that's the option that we're going with. And so that's the mechanism that we'll be using in the governance portal when it comes time to initiate a poll. These are all very basic mechanical decisions. But there's philosophy here that's far more important and that is that the foundation in this form of me presented a series of options and solicited feedback. The community came in and signal which parts of that those options they liked and which, and also I've provided some input about how those options should be presented differently, which was very useful. And then there was a clear signal and now we have a path forward. It does raise a lot of interesting questions though, and it's one of these series of questions that I don't think we should answer today or tomorrow or this week, possibly not even this month. |
| # / 00:09:18 | Richard Brown | But you look at this poll and this is, we have governance in action, look at all this sophistication, it's hard to avoid the fact that only 21 people voted though. And that's something that we need to consider as a group. And so this raises a lot of specters. I'm going to cast them as being specters because I think that these are very dangerous things that we need to start thinking about. And one of those is thresholding quorums reputation as possibly something we can consider. Do we want to get into a world where we set limits on whether things count or not? And I think that that's something that we need to have a lot of discussions about. Personally, I think that the world we live in is, the system we designed, is that whoever and this is another favorite meme of mine. Hopefully I can remember this correctly. If you don't believe that your vote matters then it doesn't. |
| # / 00:10:16 | Richard Brown | And if you believe that your vote does matter, then it does. So whoever gets out and actually does something those are the people that are controlling the system. And so rather than sort of begging the community to come out and do something, and then if there are not enough people to do it, we throw in the towel and say, "Okay. Well, we're not going to do anything." I think that's a dangerous place to be. I think that we're going to need to have some teaching moments. And so if 21 people votes out of essentially 1000s of people that are interacting the system, then those 21 people get to make the decision. And that's just the way things are. And so where that leaves us is that there's going to be an asset priority poll that's going into the portal that is loosely scheduled for mid, early to mid next week. And I did it again, I talked too much. |
| # / 00:11:05 | Richard Brown | But I can't stop now because there's really important things we need to get into. And I'm going to try and do this very, very quickly. We're going to be talking a lot about mechanics implementation and process in these calls. But we're also going to be talking about philosophy. And this is where I want to verge into philosophy for a bit because as an organization and as a community and as this governance ecosystem, we have a lot of things happening. There's more moving pieces than we have time in this call to talk about. So that puts us in a position where it's tricky for us to make predictions about what the future looks like. Making predictions in cryptos is a fool's errand, anyways. But it's possible with the two or three months window with some accuracy and we have a lot of things gone to go. |
| # / 00:12:00 | Richard Brown | We have a lot of things we need the community to engage with and we need to communicate that to the community and let them know what to expect. In order to do that, we need to understand what we're up to. We need to understand what our partners are up to. We need to understand what the ecosystem is up to and then we need to understand what governance is up to. The third one is a bit of a wild card because knowing what governance is up to, it's impossible because governance is going to tell us what they're up to. And then we're going to react to that. So we're in a situation where MakerDAO is enormously invested in this concept of radical transparency. If we have information, we want people to know what that information is. The trade off here is that information is subject to change. So we can make our best guesses. |
| # / 00:12:44 | Richard Brown | We can make our best efforts frequently. We are not in control of when these dates arise and so we need to shift those dates. The reason why I'm going through all of this is because in another window I'm looking at all of the secret dates that everybody else in this call probably wishes they could see. And that is not me being cruel. That's because we're about to post that on our website. And this is one of the things I want to caution the community about and how the community give some thought to is this give and take. We can do our very best to outline the 20 things or 30 things that are going to happen between now and MCD launch. But the community needs to understand that those things are going to change and they might get removed, more things might get added. |
| # / 00:13:31 | Richard Brown | Those dates are going to slip. I guarantee it. And I'm frankly not that concerned about it because that's just the way the world works. But we need to create these expectations that we can be radically transparent as much as we possibly can without breaking any contracts. Or we can keep things under our belts and just surprise people with things as they happen. And that's something that community needs to align on before we deal with too much fall out over having missed our dates. Right. I promised to keep it short. I was close to, I'm breaking that promise. So we'll leave it there. Cyrus, do you want to use that as a segue to talking about why the priority list was important and why we went through this polling process? |
| # / 00:14:16 | Cyrus Younessi | Sure. So there are two main reasons for a asset priority list, both of them concerning scalability. One on the risk side and one on the MakerDAO side. So for the risk side I think I just want to stress that there is an enormous amount of work that goes into collateral risk evaluations. Based on the framework we've identified for evaluating these assets, there's just a lot of work to do, right? So there's a qualitative framework and a fundamental analysis, and then there's a quantitative modeling. And what's interesting is that we are trying to do this all in a decentralized setting. Which if you think about it, if we were to do this in a centralized manner, then the risk team could just go and do a bunch of work internally and set some risk parameters. But because we're trying to uphold principles of scientific governance and so forth, we have this goal of making everything transparent. |
| # / 00:15:37 | Cyrus Younessi | All of our models, all of our data, all of our analysis, everything has to be put out to the community and it has to be discussed and it has to be voted on. And there's significant overhead. These evaluations take some time. And I think people may not necessarily realize that there is, well currently there's about seven collateral types in the initial potential set, but in the pipeline there are dozens of collateral assets ready to be added. And until we can scale out risk teams to basically handle more throughput for evaluations, it's just going to take some time to run through these projects one by one. So that being said, how does a risk team go about selecting which assets to evaluate first? First come first serve doesn't necessarily appear to be a good solution. So I think we collectively would think that some sort of prioritization or direction given by the MKR token holders would be more beneficial. |
| # / 00:16:58 | Cyrus Younessi | And we can discuss some criteria for why specific priority list can be chosen. But I think it's just kind of fairly intuitive, right? What has the highest value impact on MakerDAO. What has capabilities of generating a lot of DAI supply adding to diversification and so forth? So right now I think our goal as the interim foundation risk team is to just essentially hit as many of the initial seven as possible until MCD launch. But essentially our process is to go through them one by one to ensure that each one of them are of sufficient quality. I think, for example, we'd rather put out three super high quality reports rather than just a giant batch of mediocre whole-home reports. So that being said, we and other future independent risk teams could certainly use some direction on which assets the MKR community would like us to evaluate. |
| # / 00:18:16 | Cyrus Younessi | So I'm assuming ETH is probably going to be top of the list, but moving beyond that of the next seven collateral assets, it's difficult for us to say what is a priority. In the future, it's not going to be just seven assets on those lists. There could be dozens, there could be hundreds. It's really hard to say, right? So there's some additional benefits to priority list. One is perspective risk teams could come in and they can kind of gauge beforehand the value that MKR holders would derive from certain assets. So if they see an asset just at the bottom of the list, maybe it's not worth their time. But if they see something that's up top and they know nobody is currently working on it, then it could spur some risk teams to get to work. So yes, I mean it's just on the risk team side, it's just purely scalability thing. |
| # / 00:19:18 | Cyrus Younessi | If we had more risk teams then we could get to more assets and the priority list would be less of a concern. If we could just snap our fingers and add every collateral into the system, we would as long as they're not obviously scams or frauds or whatever. Okay. So that's on the risk team side. I think there are some additional considerations on the part of MKR holders and the MKR community. I think it's potentially a bit dangerous to add too many collateral types at once. Let's say during one particular governance cycle we were to batch add 20 collateral types and if something goes wrong, if something happens with the DAI peg, it might be a little bit difficult to concentrate and kind of surgically figure out which asset is responsible for any problems we're dealing with. |
| # / 00:20:21 | Cyrus Younessi | If we could kind of do a slower steady stream of asset additions and basically take our time, look at the data, make sure we're moving prudently, then that would probably be better. So essentially MakerDAO should consider whether they want to throttle the speed of collateral additions to a certain amount. I think it makes sense to a certain extent and right. And so in that sense, a priority list would also be useful. So that basically risk teams, collateral partners, everybody's on the same page about the timeline or calendar of assets to be added into the system. I think that's pretty much it on my side if there are any questions. |
| # / 00:21:20 | Richard Brown | There's no chatting going on in the forum or on the chat room? Is there questions we need to address? |
| # / 00:21:28 | David Utrobin | Well Marson had asked, can you estimate the average time it takes to evaluate one asset for example. And I know that's like a super wide question, but do you have a good answer? |
| # / 00:21:40 | Cyrus Younessi | Yeah, even that's difficult, right? Because it depends on the size of the team. So right now primarily it's just me Vishesh and Primoz and we are collaborating to build to do the ETH risk evaluation. And I can certainly say it takes a bit of time. Additionally, there's also kind of the way that risk models are designed to be submitted. There's also a kind of a generalized model that needs to be attached as well. So in addition to just a specific collateral asset evaluation, there needs to be plenty of documentation surrounding frameworks and ideas and whatever academic research is being used to help support your actual evaluations. And that might just be like a one time fixed costs on our side. |
| # / 00:22:42 | Cyrus Younessi | So for example, if we wanted to evaluate 10 crypto assets, we could use one general framework that we would apply to each of the assets, but that general framework by itself for the first time as considerable overhead to the overall process. So each additional collateral asset would certainly move faster. But then let's say down the line we wanted to shift to something completely different security token, stocks, whatever. Then we'd have to, any risk team would have to then create a new general model that describes their methodologies behind evaluating a particular asset class followed by specific asset evaluations. |
| # / 00:23:32 | Cyrus Younessi | For the first few ones, I mean, it's hard to say because we've been doing a lot of parallel, we're also doing a lot of parallel research that's not directly related to this. There's just a lot of things going on right now at the foundation to kind of isolate specifically how much time the specific asset takes. I would say at a good level of efficiency probably takes several weeks to do one asset. If you think about having to research a collateral asset due diligence, investigate the organization, evaluate its business potential, get the trading data, run some liquidity models, run your general risk models, and then finally actually document everything and write up the report and find a way to submit it to the community. Definitely not a one week turnaround unless it's just sufficiently large risk team. |
| # / 00:24:42 | Richard Brown | That sort of ties back to the predictability of timelines in general is that the complexity, obviously, as Cyrus just pointed out. If these assets are going to vary and the other, like I said, the wild card here is that even once those models or those analysis are finished the community needs to talk about them and figure out whether and what level of debate they're comfortable with. And so that's another big question mark. Even if we did, we're able to predict how long it would take to evaluate an asset. We're not able to predict how long it takes for the community to evaluate that evaluation. We have a lot of complexity here we need to deal with. |
| # / 00:25:21 | David Utrobin | Yeah. There was one more question Frank Cruz just asked, will the different types of collateral assets be located by a risk weighted assets theory and will each collateral asset have like a tier one, tier two, tier three structure for different CDP types? |
| # / 00:25:40 | Cyrus Younessi | So for clarification if he is asking if we're going to kind of model it off of traditional centralized platforms, not exactly. Although there are certainly a lot of similarities. Few differences in our model at least as it pertains to crypto assets, which don't have the same robustness or robustness of trading history for us to kind of accurately bucket what type of risk weight they would be. As we move on to security tokens and more traditional assets, certainly it's an option for risk teams who want to evaluate the risk of these assets using that framework and apply it to MakerDAO, could certainly happen. But no, not for something like [inaudible 00:26:41] or ETH. No, we're not doing it that way. |
| # / 00:26:46 | Richard Brown | Okay. Before we move on to Chris's presentation, there's also a question here, David. Chris Black asked about the necessity of a forum polling mechanisms when we have an official polling mechanism. And this is something that we've been discussing for a while now. And it seems circular in some ways. Why do we need to poll about polls and if we need to poll up the polls, do we need to poll about polls about polls? I think the answer there is no. And so here's the situation that we found ourselves in a few months ago is that having things arbitrarily show up in the official governance portal it is very, very tricky because we can't figure out what the driver for that decision was. And it doesn't give the community an opportunity to debate what those polls are or the utility of those polls or the framing of those polls. |
| # / 00:27:41 | Richard Brown | And so the reason why we have these discussions in the forum is that this is where signals are gathered. And this is where consensus is achieved. And this is where debates happen. And once those things have been hammered out, it's a pre-filter for the existing, for the formal polling system where the making token holders can get in and signal their support for various initiatives. So we need to have a mechanism that allows people to, the community itself to add things into the ecosystem. And that's what the forums are for right now. There's going to be, well we can divide these things up into positive buckets. There's going to be things that the risk teams or the teams that are empowered by the organization, they need to put decisions in front of the holders. |
| # / 00:28:29 | Richard Brown | And they can jump straight into a poll in the governance portal because they've been empowered and ratified by the community to do something, but there's going to be piles of random decisions. So maybe a good example of that is that perhaps we don't need to vote on stability fee every week. This is something we've been talking about for a long time. The foundation could just invent a bunch of ideas and then stick them into a poll and then have the community vote on them. But that's not really what we're trying to do here. We're trying to figure out what the community believes we should be doing. In order for that to happen, we need to solicit back and forth. And so there's a lot of threads in the forum right now about that issues. What is our cadence? What is their stepping rate for these stability fee changes? How should it be phrased? |
| # / 00:29:09 | Richard Brown | It's a long discussion. That discussion has to happen somewhere before it hits the poll, the official poll. And that's why we have the secondary polling mechanism. There's also some interesting twists to this because in the governance portal you vote with the Maker that you hold. In the governance forum, it's the wider community signals their intentions. So we can figure out kind of what the user base wants and then take that into the official arena where we find out what the owners of the system want to figure out whether there's alignment there or not. I hope that answered your question. If not, Chris feel free to ask more questions in the chat or hangout afterwards, after the call. At the top of the hour we'll have some more random Q&A's. Chris, do you want it to take it away for us and give us a demo or talk about the new portal? |
| # / 00:30:13 | Chris Bradbury | Yeah, sure thing. Thank you. Yeah, I just want to just quickly take a short amount of time really just going over the governance dashboard really. Answering the questions we've got or anybody has. Just to show you what it is really and just explain what it's doing compared to what it was doing before. So hopefully many of you have seen it. This is just on our Kovan version at the moment, which is why there's still some polls active. But yeah, just a couple of the changes that we've made to try and make the experience a little bit easier. And to touch on some of the things that we've talked about just previously about having multiple concurrent polls running at any one time and so on. And maybe if you weren't sure how that affects your voting and stuff like that, we'll cover it here. |
| # / 00:30:53 | Chris Bradbury | So the first big thing we've done is moving all of the polls into this, I mean, polling page. So now we don't have a ton of different vote options on the dashboard. They're simply on a single kind of a proposal page, really. descriptions off the pole. And then you vote just by choosing whichever option we've got in there. So at first start this call where we discussed what would be the types of votes, the asset prioritization poll. You'll simply here see yes, no, it's one of those things, in these drop downs. And you'll see a full vote breakdown listed now on a single poll page instead of one each individual poll option. The way polling is done now is it was in DS Chief before using your MKR that was locked up. That is no longer the case. When you vote now, you simply just emit an event from your wallet. |
| # / 00:31:45 | Chris Bradbury | And that wallet has to either hold your MKR or be one of your connected wallets, if you're actually currently voting on the executive voting options as well. So if you've got MKR in the Maker governance system in DS Chief and or you've got MKR in your hot wallet or cold wallet, all three of those MKR balances will be taken into account and voted with. So you don't have to vote through different ways if you've got MKR in a hot wallet, a cold weather and DS Chief. Just either one of the two wallets that you set up the voting proxy with, you can vote with. And you can also vote without saying off of vote proxy as well. So this is to try and encourage and help new users coming into this system now and help them get involved in governance. Not necessarily going the full hog and getting involved in the executive voting by setting up a voting contract and moving their MKR into chief. |
| # / 00:32:36 | Chris Bradbury | If they maybe don't trust it at the first instance or unsure about moving their MKR into chief. They can now just start participating in the polls, keeping their MKR in their wallet and voting on it that way. I'm just going to quickly switch to main ETH as well. And one of the other positives that we had based on some of the feedback that we got before was, we had a lot of people asking for the polling stats ultimately to remain once a poll is closed. We'll see one of the kind of semi annoying things lost on the previous version. As soon as a poll ended, you saw what the winning poll was, but you couldn't see what the vote breakdown was. So this is the poll it just finished now, it tells you what you voted for at the time so you have a history of what you voted for and that will just remain in there now and you have a full vote breakdown as well. So you can see what everybody voted for. |
| # / 00:33:25 | Chris Bradbury | You can see how many unique voters were in this poll and see the number is much, much lower than we want. And the participation is obviously very low and this is based on the total of MKR voting, the total of MKR in circulation when the poll took place. So all these stats will remain now. They won't move around like you might've seen previously and so on. And I guess the big thing here now as well is we've taking a huge step towards decentralization. So all of the content now is hosted ultimately outside of Maker. We've got a, the contract that emits the events is here, and I can share this link if anybody wants to take a look. But this is the contract now that emits events. We don't store the content inside a Maker system anywhere else now. This content is actually stored on GitHub. |
| # / 00:34:19 | Chris Bradbury | And what we do is when you create an event on chain to create the poll, you'd give a URL where the content was first placed and something that as a user or an MKR holder or a voter that you're now able to do is you're actually able to verify that this content is exactly the same as it was when the poll was created. So if you're looking back a day, a week, a month, a year to say, I'm pretty sure that wasn't what we voted on at the time, you're actually able to validate what that hash was by looking on chain that the event that was the hash was created. And you can download the raw metadata that actually comes from GitHub and stored. I don't know if you can see this, probably not. But you can actually download the raw content on there. I think [inaudible 00:35:07] sharing. And you can actually run it through the hashing algorithm. It's explained in our blog about the update. I think we link to it in there. And you can verify that the content matches going forward. I think that covers everything. |
| # / 00:35:25 | Richard Brown | Cool. Before we dig into any questions, Chris, can you go back to the Etherscan page and then click on the events tab for me. Because there's something interesting happening here. This is very cool for the third party ecosystem, developer ecosystem I think. Because now we have this very clear and easy to read audit trail of everything that happened in these specific polls. So it's going to be very, very easy to come up with dashboards and graphs and do some analytics on what people have been up to over time. I'm super excited about this. The foundation is actively trying to incentivize yeah, an ecosystem to arise around governance itself. Right now we have a fairly modest, we have a robust ecosystem around the DAI stable coin about CDP engines, but we're very interested in incentivizing people to build around our governance ecosystem as well. And so I'm hoping that this reduction in friction will be enough to get people to start throwing up some dashboards for us so we can see how things have been progressing over time. |
| # / 00:36:32 | Chris Bradbury | Yeah, exactly. And just one thing I think, which you touched at the very start, this is ultimately just our first iteration of trying to improve governance and improve polling and we've got a number of things on the roadmap that we know we can do to carry on making improvements. But I'd be super keen to hear any feedback on what people think, what people have seen, what they like, what maybe they don't like on this first iteration. So we don't make any mistakes moving forward ultimately. But yeah, I'll leave it at that. If you have any questions, just yeah, let me know. Or just shout. |
| # / 00:37:09 | Richard Brown | All right, thanks Chris. That was great. There's a beat that I, I know you mentioned it, but I want to really hammer this home is that with the polling system now, there are no more hoops to jump through. And so there's no more excuses not to be heard. So if you possess Maker you don't need to lock up your tokens if you want to engage in a poll. So there's literally nothing stopping you. There's no more excuses. So if there's a poll happening, click the link, pick something in a drop down and have your voice heard. It's enormously valuable. |
| # / 00:37:45 | Chris Bradbury | Right. And you can still use a vote proxy without moving the MKR into chief as well. So you can set up that vote proxy just by making a single event from your cold wallet and then keep your MKR on that cold wallet and still vote with your hot wallet for that MKR to come. So yeah, all the barriers have gone in essence now. |
| # / 00:38:02 | Richard Brown | I don't think so- |
| # / 00:38:05 | Matthew Rabinowitz | Forgive the ignorance on this, but it goes back to the question about the voting and having something potentially expire. Subject to community consent or approval. But forgive the technical ignorance on this. I don't know where that fits in with the portal you just mentioned. Is that behind the scenes? Is it in this the UI side? |
| # / 00:38:25 | Chris Bradbury | Did you say how do we manage it when it expires? What's that? |
| # / 00:38:29 | Richard Brown | Let me frame the conversation a bit. So we have this situation and I discovered this to my horror, a couple months ago when I was doing some first passes at gathering analytics on the system, is that there are still people out there that have their Makers staked for a 0.5% increase to the stability fee. So they'd been stuck on us, what we call a spell. So they'd been stuck on a vote. I shouldn't say stuck, they just left their Maker on a vote, it's ancient. And there's piles and piles and piles of Maker that are on expired votes. And so we need to, there's been some discussion in the chats and the forums about how to handle stale Maker. |
| # / 00:39:08 | Richard Brown | If somebody, if a vote is never going to be lifted again, if it can never be put in front or it's unable to ever affect the ecosystem that's essentially wasted Maker at that point. So what do we do? There's been some options forwarded. So either once a vote is no longer applicable that Maker can just go back into a person's wallet automatically. That's risky for a few different reasons because we don't know who owns that wallet. Well, there's all kinds of issues. Or we can come up with the system where after a certain time period that Maker can be poked, which is an interesting idea that we came up with, I think. Where abandoned Maker could be pinged by some sort of a keeper and then that would begin a countdown that would put it back into the person's wallet or put it back into rotation and, or take it off of a legacy voting option. |
| # / 00:40:12 | Chris Bradbury | Right. Yeah. We're actually looking at all these sort of things at the moment actually as an update to executive voting. So this phase of the project was for just the polling, but yeah, that next stage of maybe having a complete, like we look at what continuous approval voting is or looks like or how executive voting might work in the future. One thing we're keen to do is let MCD launch and see how governance kind of changes with MCD being live. And then yeah, use the community and ultimately the smart contract guys to design the best system that's easiest to use. And it's the easiest to put onto an interface basically. Because I think that's one of the hardest things where we're at right now is we have like very complicated system which we've tried to make simpler on the UI. But in reality it's a very simplified version of a very complicated system. And it doesn't help the users coming in at the moment. |
| # / 00:41:06 | Richard Brown | To give you sort of a high level or sort of a slightly ambiguous answer, Matthew, the way that we would answer that question and come up with the solution is the same way that we came up with an answer and a solution for what a priority asset list would look like. So someone would make the case for solution for this abandoned or stale Maker problem, post it into the forums. We would have a debate, there would be a poll. We would gather signals whichever signal got the most poll weights in the forums would be transferred into the governance portal itself. And then we, the community would signal which solution it wants. And then we'd use that to surface it back to the foundation and then the product team could pick it up from there. |
| # / 00:41:53 | Matthew Rabinowitz | I got it. Thank you. |
| # / 00:41:58 | Richard Brown | Okay. Cool. Let me double check the agenda. David, I think that you're up for the what's happening or governance at a glance. |
| # / 00:42:07 | David Utrobin | Yeah. I'll keep it super short so I'm not going to cover everything in thread, as I said last week. I'm just going to cover the new stuff there. So LongforWisdom posted two new threads this week, there was a signaling guidelines thread which actually summarized the old meta governance thread and kind of put some guidelines there. And we're seeking consensus on what a kind of canonical set of guidelines around signaling on the form actually looks like. And so he's inviting anybody to go take a look at that thread and disagree or agree as you wish. |
| # / 00:42:43 | David Utrobin | And I'll post the links up right now, if you guys are interested. And then the other thread that is there is again, about the asset priority poll, which we have already thoroughly covered in this call. So I am not going to go into my notes about it because Richard has talked like extensively about it. So yeah, so those are the only two major new threads in the forum. There have been some kind of scattered comments and continued discussions around older threads, but nothing worth noting. So actually kind of cut my time short to give Vishesh more time and I think Vishesh is up next, right? |
| # / 00:43:25 | Richard Brown | Yeah. Before I hand it off to Vishesh, I want to, there's a discussion in the forum or in the chat as we were talking today, and also I just noticed in the forum again is that there's been some shouting about a dedicated Twitter bot to communicate things that are happening in governance. And this is part of a discussion that we will be having in the next couple of weeks, specifically next week as well. It's on the agenda. But it's the idea of broadcasting and communication. The things that happen in governance and the forums, things that happen in this call, the things that happen in the chats need to be broadcast to the wider world. And we need to come up with mechanisms that allow us to do that. I am looking to the community to provide answers and to help brainstorm with us. One of those answers has already been a suggestion, but I think that who was it again? David yeah, actually you're going to have to help me out, it's embarrassing. What's the name of it? The guy who did the Maker bots. |
| # / 00:44:30 | David Utrobin | MarC. The makerDAI bot. Yeah. |
| # / 00:44:34 | Richard Brown | Sorry, that was a Mark Andre there he is. So we're looking for something that's much like that, but for governance and for risk, that would be great because we MakerDAO we do not expect people to work for us for free. So there's resources available. So if you have an idea about how we can communicate governance issues more effectively, risk issues. If you have experience with Twitter bots. If you think you can pull this off for us and you have some ideas, contact me and we can talk about how we can fit that into possibly a grants program, or to a bounty and we can get something cooking. I think that was a slight segue, but Vishesh. I'm going to hand it off to you and you can tell us what we should do. |
| # / 00:45:26 | Vishesh Choudhry | I won't tell you what to do. I'll tell you what's supported by the data. But yeah, I mean actually on the topic of the governance part, I mean we're also looking into that. So we also, if someone beats us to the punch, but otherwise, hopefully that's something that we'll look to slot into instead of improvements. And actually on that topic, I mean, I think some people will probably have noticed that the stats on one of the sites has been still for the past week or so. That's because on the back end we're doing a significant amount of work to change how all that data is getting pulled in. So hopefully it'll be more live updating, less stillness and less work for us to update it. So it'll be good. All right, so just to touch on here what's been going on with DAI. And I won't take up too much time this time, but basically this is over the last 24 hours pretty much centered on a dollar. |
| # / 00:46:28 | Vishesh Choudhry | Somewhat low volumes around two or 3 million most days in the past few weeks. So light trading, but stable. And so that's what we see as well over the seven day timeframe. Pretty much the exact same volume, weighted average price bout 14 million in trading volume over the past seven days at least on DEX's and Coinbase. Most of that trading happened, like, again, pretty much exactly to dollar or very, very, very slightly below. So again, relatively low trading volumes, like an average of 2 million a day, but at the same time very stable. Because you can see that in the DAI stablecoin graph as well. Some light variation. Very, very light volume, but at the same time centered pretty much smacked up on the dollar. So just to kind of touch on what's been going on outside of just the DAI price. So you can see here ETH price is kind of being consistently, well taking a couple of spikes, but more or less sitting at these lower levels. |
| # / 00:47:45 | Vishesh Choudhry | So I think what we've probably encountered is a bit of fatigue from some of those leverage positions. And you can see in the transactions here the amount of debt that's been wiped over the past couple of weeks has ticked up. But the amount of debt that's been gone out, it's been relatively steady state. So, that means that the supplies come down fairly significantly since about 90 million in mid July to just to beginning the first week of this month down to about 76, 77 million. So that's a pretty significant drop. Part of that can be explained through some of what's been going on with refinancings, but actually not quite that significantly. So this has been a drop in demand, I think we can safely say. But it's more than likely that demand given the way the collateralization has moved. So let me touch on that. |
| # / 00:48:55 | Vishesh Choudhry | And yeah, the transaction volumes have been steady over the past month or so as we talked about. But in the past week or so, they've dropped down quite a bit. So this sort of quiet at the moment. But I do want to test out the sentiment dashboard a little bit, but the collateralization, so what we see is the amount of ETH locked has come down a bit. But the DAI supplies also come down. So these have just kind of moved in lock step. The collateralization has been held relatively steady, so still in this like 360 ish, 360, 70, 80 over the past week or so. Actually two weeks. So not a lot of movement in collateralization, but the supply and collateral have come down. So that's just a reduction of leverage because basically over the past two or three months, the collateralization ratio has not coming down below this like roughly 360% levels. |
| # / 00:50:02 | Vishesh Choudhry | So that's basically at the moment seems like somewhat of a floor for where CDP owners are going to allow the collateralization ratio to go given with the ETH prices currently. And honestly, the volatility that's been going on in ETH, I think has kept a healthy amount of fear in leverage seekers. So I don't think we're likely to see the collateralization dip down much further. Though the DAI supply is something to discuss. So just to talk about what's been going on the secondary lending platforms, what we had seen in mid July was the rates and the utilization shooting up pretty significantly on secondary lending platforms. |
| # / 00:50:49 | Vishesh Choudhry | Part of that is a function of refinancing and part of that is function of InstaDapp and things like that had really opened the flood gates to this back and forth flow of moving CDPs from Maker to platforms like Compound where people had probably not been taking as much advantage of that primarily due to knowledge, user interface just the general experience and the transactional cost involved in moving those things. So naturally that led to a little bit more utilization, but that's actually leveled out a bit. So the only explanation for this drop, and I'll just go to the utilization graph, for this drop in utilization in that same timeframe is supply has gone down. And we discussed this with David a bit on the Maker chat over the past week. Essentially supply has come down, but also utilization has come down. |
| # / 00:51:46 | Vishesh Choudhry | So that unfortunately means there had to have been a reduction in demand, at least a demand for leverage positions. That utilization has come down very slightly or almost none on dYdX. So what's really interesting is when you disproportionately see changes in dYdX versus Compound, I think dYdX has really nestled into a very specific role of when people are seeking high leverage, when they want to go immediately three x leverage and just use the built-in button that dYdX has for doing that. When people are looking for more short term plays where they're willing to pay this higher rate, you can see the dYdX borrow rate is pretty consistently higher than Compound. And occasionally higher than the Maker stability fee, but very unpredictable. So I think that's, these are reasons why dYdX is nestled into more of this short term, somewhat risky or slightly more expensive role. |
| # / 00:52:51 | Vishesh Choudhry | Whereas Compound has been more of the, just think of it as like a cheaper version of Maker lately where people can source excess DAI that has been thrown into lending because someone capitulated their CDP position and someone else picked up that DAI just purchased it, is not paying the stability fee and decides to get a yield. So they'll throw it on the platform like Compound. And that's where some people can get some cheaper DAI. So that's been a little bit more of a stable rate over time. And so it's been really interesting to see that utilization come down from pretty much topping out to closer to dYdX's utilization around 79 80% just in the last week or so. So again, the combination of the fact that supply has come down and utilization on a platform like Compound has come down versus dYdX presumably means that more of those passive leverage seekers essentially not the, like higher risk higher price tolerance lenders and borrowers. |
| # / 00:54:02 | Vishesh Choudhry | So basically people that are, had refinanced CPS to get a bit of a discount because the 20.5% stability fee had discouraged them. Or people who were basically seeking leverage but maybe not all the way to the red line of like three x had presumably kind of given up on those leverage positions. Whereas more of like the dYdX type of users have not changed their behavior as significantly over the past couple of weeks, which does make sense, right? So essentially those seeking more severe leverage presumably have a higher risk tolerance and are willing to hold onto their positions in the hopes that a small rise in ETH price will make them profitable again and allow them to cash it out before they accrue too much in fees. So again, there's a lot to unpack about this. I'm basically consolidating three or four different economic effects into one single statement. |
| # / 00:55:07 | Vishesh Choudhry | So I'm sure I glossed over some nuance there, but you can essentially see the same thing reflected in the cash or excess supply on both these platforms. Which has come up fairly significantly from almost nothing in mid to late July to a solid trouble, eight ish million DAI. And the percentage of this excess supply out of the total DAI supply has gone up fairly significantly because the total denominator has come down. But this numerator's gone up fairly significantly. So again, I think, and this echoes some of what was shared on the chat recently. Demand has dipped a bit though DAI is fairly stable. So I think it's important to consider what can be done to help stimulate and drive this demand because it is stagnating a little bit. And that's definitely something to consider. So borrow volumes fairly flat. Supply volume has gone up, which results in the bottom two graphs. So yeah, just one last thing to touch on is the fees. |
| # / 00:56:28 | Vishesh Choudhry | So essentially the paid fees have had ticked up in that same time frame that I was talking about, mid to late July, and have been fairly flat since then. Unpaid fees have actually dipped during that time period. So that makes sense. Right? You lose some of those unpaid fees because you gained paid fees. But again, since mid to late July, those unpaid fees have now resumed their trend of just accumulating. Whereas the paid fees, it kind of flat lined a little bit. Paid fees tend to go in a bit of a stepwise function. |
| # / 00:57:05 | Vishesh Choudhry | And it'd be really interesting, maybe I can look at next, how the rate of this revenue relates to volatility of ETH price. But essentially that's the, my hypothesis is that as ETH is more volatile you'll see more churn, more circulation in dept. Which is a concept that we've talked about before. Which leads to more realization of paid fees. So ETH being relatively stable at low price levels has led to a little bit of flat behavior in most of the Maker stats and that also applies to paid fees. Okay. So hopefully that was a relatively quick hit. Happy to follow up on things or answer questions. |
| # / 00:57:52 | David Utrobin | Do you have a theory about why more people aren't refinancing from Maker to Compound now that like the rate is even more significantly lower for borrowing? Like I think it's at 16% or 16.2 compared to Makers 20 and a half and there's like 7 million of excess supply. So it's like 7 million of potential borrowing that anybody could refinance. |
| # / 00:58:15 | Vishesh Choudhry | Right. So again, I think part of the initial jump there was some of the buzz excitement and extra visibility around things like InstaDapp. At this point, and I think the market just moves a little slow. It's definitely a factor. What you saw was around July 24th, which is not too long ago, about two weeks. The borrow rates were very, very close to the stability fee. And I think it takes a little bit of, these things are in cycles kind of, it takes a little bit of time I think before people look at okay, there's a solid 2% spread between the Maker stability fee and the Compound rate for example. Now I will go ahead and refinance. I think you should reasonably see more refinancings over the next few days if the compound rates stays around 16, 17%. It doesn't make economic sense for people to not take advantage of that now. Eight to 9 million DAI that's floating on those platforms. But I don't think you'll see a huge change in dYdX because actually the rate there is pretty much smacked out around the stability fee. But for Compound, you do have a solid additional what was it about seven million DAI that should dry up. There's no reason for it not to. |
| # / 01:00:00 | David Utrobin | Cool. Thanks. That's a good answer. Yeah, it seems, yeah. I don't see any questions outstanding in the chart or anything. |
| # / 01:00:21 | Richard Brown | Yeah, so we just crossed the top of the hour. We can transition into the general Q&A mode, I think, if people have questions that they wanted to ask. But for people that are dropping off already, I want to make sure that we get the thank you in. So thank you to Chris Bradbury. Thank you to Cyrus Younessi, Vishesh, David Utrobin and a big thank you to me for my continued administration of these calls. Everybody please hang out and chat. I am going to go dark because I have another meeting, but I'll be listening in another window. Sorry. I forgot my postamble. I did the preamble. Here's the postamble. The conversation continues in the forums. We have a forum thread for every one of these calls. We post summaries of what happened in these calls in those threads and linked videos and audios. Please come to the forum and make your voice heard because that's where governance happens. All right. Thanks everyone. Now we can begin the awkward silence stage I think. |
| # / 01:01:33 | Cyrus Younessi | So you're going to do two meetings? |
| # / 01:01:37 | Richard Brown | Two meetings. I have two chats and I'm composing a tweet as we speak. What did I call you yesterday, a one-jobber? It must be nice. All right. Matthew, create some contention. What's going on in your world? |
| # / 01:02:06 | Matthew Rabinowitz | No contention. |
| # / 01:02:07 | David Utrobin | Any update on like the research around the stability fee predictions using like a PID system? I know that was a big topic a few weeks ago. |
| # / 01:02:22 | Matthew Rabinowitz | Not Exactly. I mean there's, I'm still writing a research paper about that. I started on it, at least drafting it, but it's far from anywhere ready to be published. But using a PID does work, but it's benefits basically drop off the larger the system gets. Because any one, let's just call them, doesn't matter. Let's start imagine we start off with 10 collateral packages, whatever we're going to call them, right? And we start seeing one out of the 10 that starts producing more or less. |
| # / 01:02:54 | Matthew Rabinowitz | And that therefore requires us to change the DSR up or down or risk parameters up or down. But after we have 100 collateral packages or 1000 of them, the impacts from any one given source start to be disproportionately lower than they were initially. So if you use a PID to come up with an algorithmic suggestion, it can help, but it's really, really useful at the very beginning, which is exactly when we won't have it ready by. And the longer the system goes on, the more obvious what the decision tree will be by [inaudible 01:03:30]. So research continuing. |
| # / 01:03:42 | David Utrobin | Nice. Yeah. All right, well I'm going to hop off. So I will see you guys next week. |
| # / 01:03:53 | Cyrus Younessi | I'm going to hop off to, unless there's any questions or anything anybody wants to talk about. |
| # / 01:04:05 | Richard Brown | Yeah, doesn't seem like it. Let's call it. I'll post, I guess the video soon. All right. Thanks everyone. |
| # / 01:04:13 | David Utrobin | I'll have the summary up shortly. |