Skip to content

Governance and Risk Meeting: Ep. 42 (July 4 - 2019)

Video | Audio | Discussion

References Person Text
# / 00:00:00 Richard Brown There we go. Hey everybody, welcome to the July 4th edition of the Scientific Governance and Risk meeting with MakerDAO. My name is Richard Brown. I am the head of community development. I will be moderating today's call. We're not going to have a huge turnout. Apparently there's some kind of holiday going on somewhere, so we're going to plow ahead with what we have on our schedule, talk about the peg. We're going to hear from Matthew about his weekly narrative. I'm going to talk a bit about a secret forum that we have. That was my big teaser to try and generate some traffic today. Did not work, but we'll talk about it anyways.
# / 00:00:42 Richard Brown As always, excuse me, we're intensely interested in hearing from our community, so if you have question, type it in the chat. If you have a microphone, jump right in and let us know what you're thinking. Don't be afraid to interrupt.
# / 00:00:59 Richard Brown To kick things off, we should talk about governance, because that's half the reason we're here today. Something interesting happened in the poll, and I want to call this out, because this is a recurring theme that we discussed over and over in these meetings. That is voter turnout and what does that actually mean, what's the impact of it, what is continuous approval voting, how do we put the signals in the system and why are those important?
# / 00:01:28 Richard Brown Excuse me, I have a cough today.
# / 00:01:30 Richard Brown It's important for numerous reasons, but today we saw an example that's happened a couple of times in the past, which points out one of these edge cases, and one of these edge cases is that if you don't believe that your vote matters, then somebody else's vote does. It's a hard lesson to learn and it's one that people need to internalize, especially in this ecosystem.
# / 00:01:51 Richard Brown But up until early this morning, there was a fairly clear signal that the voters had been interested in raising the stability fee by 1% to 17.5%. At some point in the last few hours, someone sneakily moved their MKR into 18.5.
# / 00:02:14 Richard Brown So, we have this 11th hour, excuse me, shifting of direction, and if only one or two large holders are showing up for these polls, then one or two large holders are going to be setting the stability fees for us. So it's yet another reminder that people should do their best to keep their hands on the wheel when it comes to these polls.
# / 00:02:40 Richard Brown I don't want to dig into it too much. I don't think it's that alarming, because based... I'm not smart enough to know. But I'm assuming that when Vishesh gets into his numbers, we will see that perhaps this is not too outlandish a decision to be made.
# / 00:02:55 Richard Brown Second thing I want to talk about today is we have a secret forum and I'm calling it secret to try and create some buzz. It's no secret at all, it's just unadvertised. What has happened... what has happened? What's happening is that we have lots and lots to go over in these calls. We have very, in the governance ecosystem in general, there's a lot to talk about.
# / 00:03:17 Richard Brown And we've been talking or repeating this message that these calls have a very limited time window associated with them, they're North American centric, there's all kinds of reasons why this is a very, very small demographic, a Venn diagram of our constituents are going to show up in these calls. What we need are venues where we can continue longer term, deeper debates, and really dig into these issues that have more of a window, a time window that allows for these discussions to occur.
# / 00:03:54 Richard Brown Up until very recently, that solution for us was Reddit. Depending on your experiences on Reddit, in my opinion, Reddit is not a good solution for anything, but it's probably the best one we have. But primarily, it's still very time restrictive. You have, depending on the activity in that subreddit, you have a day or maybe a week, in order to interact with a post before it begins to disappear off the front page. That's not great for us.
# / 00:04:25 Richard Brown So what we've been talking about for the longest time, but now actually have, is a forum dedicated to issues around MakerDAO. Now, we haven't talked about this to the wider world yet, we haven't advertised it. But this is a special group of super smart people. So we're going to invite everyone here to participate with us, in the hopes that we can begin to pre-seed some content there, get some interesting conversations happening. And when we do bring this to the community, there's something for them to bite onto.
# / 00:05:02 Richard Brown And so if you want to have a click on that link, you'll see that this is Discourse. It's the crypto's favorite forum solution, we did some research, and this is the best one out there by far. So that's what we're going with. We have right now, three major categories, more categories are coming. These sections are spread across major verticals of MakerDAO. So we have governance, we have risk, and we have community development, that are represented right now. More departments from Maker are coming on board soon. So we'll be able to cross pollinate some of our communities.
# / 00:05:39 Richard Brown Please take some time to dig through these things, because there's already some interesting stuff. Vishesh has been in there, Matthew has been in there, we have some interesting discussions about collateral types already, which surfaced some opinions that we had not anticipated, particularly around REP. So there's some good conversations to be had there.
# / 00:06:02 Richard Brown How this relates specifically to governance though, is this. We have a problem in governance, where we have these chicken and egg situations, where the existing portal, we understand the limitations around that, there's one vote at a time, one poll at the time. If you decide to signal on a poll, you can't signal on a vote and vice versa. That's all going to be solved very, very soon with the new version of the portal that is being released.
# / 00:06:28 Richard Brown We also have this problem where our polling system is how we surface signals in this ecosystem. So we want the community to talk about something and signal their interest in a possibility. We use a poll to gather that information and then we move that into an executive and then in the executive vote, make a decision about how to change the system based on those signals.
# / 00:06:53 Richard Brown The chicken and egg situation comes down, where we have to ask ourselves how do we get signals into our signaling system? And that's a tricky one. So does somebody just decide, do we come up with rough consensus? Do we measure by upvotes on Reddit to figure out what goes into a poll? These are all very, very hard questions.
# / 00:07:12 Richard Brown Happily, I think the forum may solve that problem for us. And the way that it solves that problem for us is, it creates, like I said, this venue for deep, engaged, long-term debate from highly engaged stakeholders to really hammer out an issue, but it also has a light reputation system built into it. It also has some fairly decent voting and polling mechanisms built into it.
# / 00:07:40 Richard Brown So someone can come up with an idea, if we need to change an X around the polling system, they put up a well reasoned argument, people debate, debate, debate, and then there's a poll and people can signal their support or lack of, for that issue. I think that's probably a great way to start surfacing signals, that we can gather consensus on, and then that's what leads into the polling system.
# / 00:08:11 Richard Brown It immediately raises some questions though, because right now we have skin in the game, a governance system where if you want to alter the way that the system works, it's easy. Just go buy some Maker and then go to the portal and then start voting. And then there you go. It's a model that the world understands and there's countless examples of it in action, in the traditional business world.
# / 00:08:37 Richard Brown But that's skin in the game. So that's Maker token holders voting on signals. In the forum, we have no way to determine whether the people that are voting on these polls are Maker token holders. I don't think that's a problem, because there's different ways we can look at the community. There's 1000 ways we can look at the community, but from our perspectives, there's this.
# / 00:08:57 Richard Brown We have the general community of MakerDAO enthusiasts that are interested in the ecosystem for a myriad of reasons. Then, we have these overlapping Venn diagrams and one of those Venn diagrams or Venn circles, I'm not sure what they're called, you know what I'm talking about. There's the people that are interested in governance. And that's a much smaller circle and there's also in that overlap, there's people that are interested in governance, MakerDAO and hold Maker, which is an even smaller circle.
# / 00:09:26 Richard Brown But that isn't the only source of signals that we need to take into account when we're talking about governance. One of those major sources of signals are the community at large, we need to know what the community wants and as Maker token holders, take that into account when we're voting.
# / 00:09:45 Richard Brown So the fact that the forum is available for anybody who's interested and engaged with the system to start generating some signals for consideration, I think is a good thing. So there's actually a post in the forum that discusses these very ideas, and so I like the concept that in the forum, anybody can come in, like we don't need civil protection. We don't need to verify whether the people that are expressing their opinions own Maker, because that doesn't really matter. We just want to get a general sense of what people are thinking and whether the ideas are good, and whether there's data to back it up.
# / 00:10:25 Richard Brown That, I believe is enough to achieve enough of a consensus to a lot of things into a polling system, where we're just sort of refining that opinion. There's, like I said, there's a thread in the forum. If you disagree with me, if you have some additional insights, please join us there and let us know what you think.
# / 00:10:50 Richard Brown I think that might wrap it up for me today. I don't want to dig too far into it. If you do have suggestions about the forum or you're used to Discourse or if you've used these things in other crypto projects in the past, we also have a site feedback category. So if you want to offer some suggestions about how to improve the forum, please let us know as well.
# / 00:11:17 Richard Brown All right, at that point, I am going to stop. 15 minutes after, so that worked out well. And I'm going to hand it over to you Vishesh, to give us an update on the state of the peg, if you're ready.
# / 00:11:31 Vishesh Choudhry Okay, all right. I will share my screen here. Can you see everything all right?
# / 00:11:43 Richard Brown Yeah, looks good.
# / 00:11:47 Vishesh Choudhry Okay. So I think I shared a little sort of preview in the forum, and I think maybe that's actually a good segue to start this time. I shared a little preview in the forum. I just, my two cents Rich, I like the forum. I think it's a good UI, easy to use, pretty clear. So might be something that I try to use moving forward.
# / 00:12:15 Vishesh Choudhry But yeah, I mean, the basic takeaways were for the past week, the peg is low enough where we should pay attention to it, but I don't think it's a crisis, it is still at levels that could totally recover on their own. The debt ceiling is definitely within like a distance from the supply that we should start to talk about raising that debt ceiling. And I know some of those discussions have already started on the forum and the chats as well and we briefly touched on it in last week's call. So I think that's something to circle back to.
# / 00:12:53 Vishesh Choudhry And then the reason for this is, by and large, ETH and what's been going on with ETH price movements. We've seen a lot of trading volume, we've seen a little bit of the peg suffering, we've seen some increase in supply. So a lot of these metrics tend to indicate leverage and levering up behavior. So that kind of makes perfect sense. ETH price dips a bit, people lever up, they mint DAI in the process.
# / 00:13:25 Vishesh Choudhry Stability fee is a little bit lower right now than it had been before. So early, it had been, when things were kind of, quote-unquote going well. So it's I think, very facile to like just say the stability fee causes X, Y and Z, and I think a lot of people just want to get to that quick answer, but I think it's a little bit more complicated, where the stability fee changes the reactivity of the system to stressors like ETH. That's kind of like the gist of what I'm going to say today.
# / 00:14:02 Vishesh Choudhry So to just touch on, I think it's maybe good to mention this every now and then, I try not to show it too often because it's kind of a boring graph, but the cumulative loans that have ever been issued is roughly 360 million currently. The last time we checked on this sometime in May or early June, it was like right at 300. So that's like a 20% increase in basically a month.
# / 00:14:34 Vishesh Choudhry That's, I think, pretty significant growth. And you can sort of see this kind of curving upward. So there's been, I think, a fair acceleration in the amount of debt that's ever been issued and the amount of debt that has ever been paid back or liquidated has kind of tracked in that rate, although I think that margin between the two has widened, which is good. I think that shows growth. And we just want to make sure that that growth gets realized in terms of revenues, because it does also come with some risks. So that's just kind of like a high level thing to touch on every now and then.
# / 00:15:13 Vishesh Choudhry So specifically, the peg has kind of, since 17.5%, increased, drifted down a bit, had a little bit more variability than it did. And with the 16.5% decrease as well, had a little bit more variability than it had at 19.5%. And that was something that I had touched on as kind of an indicator that it's a little bit slippery. And so a stressor could push it a little bit further than it could have at 19.5%. And that stressor happened and so it did pressure it a bit.
# / 00:15:47 Vishesh Choudhry Not the end of the world, I think totally recoverable. Even on its own and I think this is a discussion point that has happened on the forums and can happen in this discussion is, to what extent do you need to, as a community, step in and actively push that peg around? Versus allow the kid to drive their bike to school, and so on and so forth, and let it roam free a little bit in the hopes that it will come back to $1.
# / 00:16:18 Vishesh Choudhry And I think it's reasonable to allow that degree of freedom, because there is always a cost with tightening that up. And that cost is just how easy and profitable is it to use Maker through CDPs, which is the stability fee, that stability fee, like 100% determines profitability. And so it's the trade-off between kind of tightening the variability in the system, versus allowing growth to happen a little bit easier.
# / 00:16:50 Vishesh Choudhry But I think what we'll see is that growth has been happening despite 19.5% stability fee, 17.5, et cetera. Like even at these levels, we've seen the system grow. And so I don't think you necessarily need to lower the stability fee, I don't think you necessarily need to increase it. I do think if you increase it, it will reduce some of this variability. I don't think it will necessarily 100% contract the supply.
# / 00:17:22 Vishesh Choudhry Although basic economics says that if everyone is operating rationally, they would refinance some of those over to secondary lending platforms, where there has been an increasing supply. And I think that increasing supply is a really interesting metric to touch on, which we will, a little bit there.
# / 00:17:39 Vishesh Choudhry So the ETH price, I'm sorry, when I reload these, all these become unselected. The ETH price has been running up for most of the month and then we had a of a sharp drop just about in the last week, then a little bit of a run up and then another sharp drop.
# / 00:18:03 Vishesh Choudhry So what's interesting is ETH and DAI had kind of tracked in the broad movements, although ETH has been sort of drifting up overall in the longer run, and then DAI has kind of been steady and then lately drifting down a little bit with ETH. And it's a little bit tough to do correlations for this, because it's not a one to one relationship. And so these correlations change over time with other variables.
# / 00:18:37 Vishesh Choudhry But at the end of the day, what we're saying is, there's sort of a blue skies fairweather relationship between ETH and DAI, which is a little bit distinct from the downturn relationship between ETH and DAI. And I think ultimately, it's like when people are highly levered on CDPs, then DAI has one relationship to ETH. And when people are not highly levered, then DAI has a different relationship to ETH.
# / 00:19:06 Vishesh Choudhry And that's the underlying mechanic that I think is going on. And so what I think is currently happening is people are not heavily levered up on ETH, or were not heavily levered up on ETH. ETH price sort of dipped down severely and then people started seeking that leverage again, especially at a slightly more attractive rate.
# / 00:19:27 Vishesh Choudhry So the supply here, supply has been really interesting. So as we had this series of stability fee decreases, the supply had been coming down. It's too easy to say that this was as a result of the stability fee, because there was also a lot going on in terms of secondary lending platforms through this April-May period.
# / 00:19:56 Vishesh Choudhry What is pretty interesting is, as the supply was kind of steady or relatively steady between 19.5% and 16.5%, and ETH was making price movements during that time, pretty significant price movements, it is a data point that somewhere between this 19.5% and 17.5% stability fee, was a relatively, quote-unquote steady state variable for Maker at that time.
# / 00:20:32 Vishesh Choudhry Now, it's not just the stability fee in isolation, there's a lot that goes on with secondary lending rates and secondary lending volumes, that's kind of this buffer layer that is also moving constantly and moving much more wildly than the DAI supply. But just in terms of the DAI supply, things were fairly quiet in that time period. And I think that was an indicator, that at least for the purposes of minting new DAI, that range was a bit of a steady state.
# / 00:21:06 Vishesh Choudhry As the supply as the stability fee was reduced to 16.5%, similarly ETH was making price trend reversals, I believe during that time. So you see, just around first week of June, the price had been coming down and down. And so I think somewhere around this time period, people decided to start to seek more leverage again. And so that definitely does contribute to the increase in supply.
# / 00:21:42 Vishesh Choudhry And so there's, I keep seeing this debate and this is why I kind of harp on this, between the stability fee, what its actual meaning is, what its actual effects are, and how much all of these things would be happening if the stability fee had never changed, that was a concept that was raised. And my point is just it's not that simple. The stability fee is kind of like the volume. But ETH is the one that's changing the channels.
# / 00:22:10 Vishesh Choudhry And so at the end of the day, the signals are coming from ETH, but the sensitivity of the system to those signals is hugely affected by the stability fee. So I think just when you have those things happening together, that's when it's very clear that this supply will increase. And so that is exactly what happened, the supply has increased pretty significantly from around 80 million at the beginning of June to 90.1, 1 million or 2 million, currently. This graph actually cut off the 24 hour delay.
# / 00:22:52 Vishesh Choudhry So the circulation of debt, I think just kind of echoes this a little bit. So what we'd seen was, with that 19.5% increase, you start to see just over time, a decrease in the age of debt. So more and more of that old debt was getting paid back, more and more debt that was being... more and more of the outstanding debt was new. And so this was kind of continuing and then dropped to 17.5, 16.5, that started to reverse. And then 16.5 holding steady, the average age started to drop again for open debt, but for closed debt was holding relatively steady. So what that just means is more new debt was being drawn out.
# / 00:23:41 Vishesh Choudhry Part of this sort of intermediate period where the new debt was not being drawn out that aggressively, was because a lot more of that new debt was really coming from secondary lending. So the demand for DAI was either people purchasing DAI at a discount, or borrowing it from secondary lending platforms during that time period.
# / 00:24:08 Vishesh Choudhry So yeah, we just see kind of this long term... And it's interesting, it's got this sort of periodicity to it that I mentioned, long-term decrease in the amount of new debt that's being drawn out of new CDPs versus old CDPs.
# / 00:24:25 Vishesh Choudhry Although what's really cool is in the past week or two, that has started to actually increase, so a few more new CDPs being used for loans. And I think this was actually, some people who had refinanced secondary lending platforms, coming back to Maker and so some of those folks may have closed out CDPs and had to reopen new ones.
# / 00:24:48 Vishesh Choudhry So in terms of what has been going on with ETH prices. I'll just double check to make sure... We're still streaming, okay. Usually, Cyrus will interrupt for a question or two. So in his absence, things are too peaceful.
# / 00:25:08 Vishesh Choudhry So right around May, we'd started to see the collateralization come up. And that was pretty much nominally with price. But then we saw this sort of commensurate decrease again with price, but more aggressive than the ETH price had been decreasing. And so then when we saw sort of a run up again in collateralization, this was in some extent due just to the nominal value of ETH.
# / 00:25:36 Vishesh Choudhry And I've had some feedback, which I think is really good, is to show this sort of in dollar terms as well, where sort of ETH price adjusted is what I mean. With that collateralization ratio, had started to run up partially due to just price and basic math, but it was a little bit more muted than the price would justify.
# / 00:26:00 Vishesh Choudhry But when the price started to drop, the drop in collateralization was a little bit more aggressive. And so what that means is sort of a reinforcement of how we had been thinking about the collateralization ratio, which is obviously it moves when ETH price moves, and that's the short time scale effect. But when the collateralization ratio moves down, because ETH price decreased, there are also liquidations, people start to see more risk in that. And so people will tend top up, sort of as a counter effect to what the price is doing.
# / 00:26:46 Vishesh Choudhry And then similarly, when the collateralization ratio runs up, because ETH price is increased, people are not, they're not too quick to pull out collateral from those CDPs if they think that that ETH price has run up too quickly, and might have a chance of dropping again. Because once burned, twice shy, they don't want to get liquidated, but to some extent, other people, especially on the higher range of collateralizations I think tend to do pull out some of that collateral, so that they can realize some of their profits, pay back other loans, whatever they're going to do with that capital. I think they don't necessarily want to pay the cost of having more dollars sit in the system. So you see this kind of counter balancing effect on both ends is what I'm saying.
# / 00:27:42 Vishesh Choudhry So just to touch on the price, like I said, has been suffering a little bit, but nothing too extreme. So I think when the ETH price potentially runs up a little bit, you will see some cashing out behavior. If it happens, obviously. And then that supply should theoretically decrease a bit. And then some of this leveraging will unwind, and you'll see more buying pressure on DAI. And so that could just totally even out on its own, and you may not have to do anything at all.
# / 00:28:19 Vishesh Choudhry If you do increase the stability fee, it's not necessarily going to be a bad thing, because it will simply help to tighten up and increase more pressure, more incentive for people to pay back those loans, which I will talk about in a second, because I want to talk about demand.
# / 00:28:38 Vishesh Choudhry So like I said, these were two new metrics, actually new, I mean, you all may have seen some previews of it on these calls, of the lending rates on variable term platforms. So that's dYdX and Compound right now. Outstanding borrow and supply weighted interest rates.
# / 00:29:02 Vishesh Choudhry And so basically, what this is getting at is, on a weighted average basis, how much are people paying to borrow DAI? And how much are people receiving for supplying DAI? And so what this means is, if somebody was to draw out DAI from a CDP, at least for the supply rate, if somebody was to draw out DAI from the CPD, how much is the cost of doing so mitigated by the ability to just temporarily lend it out on a secondary lending platform, if those people are not immediately selling it or using it for something else? If they're just sitting with it, how much can those platforms mitigate that cost?
# / 00:29:46 David Utrobin Vishesh, your voice cut out a bit.
# / 00:29:55 Vishesh Choudhry And the weight of the borrow rate, so that's basically the more important one, I think, which is the alternative... I'm sorry. So I understand the weight of the borrow rate is basically this alternative pathway for obtaining DAI.
# / 00:30:06 Vishesh Choudhry So that's the more important of the two metrics here. I think it tells you basically, how cheap or expensive is it for people to obtain DAI from other similar platforms, than directly from a CDP. And then the supply rate is basically just how much is the cost of just holding DAI mitigated by the ability to lend it out on secondary lending platforms.
# / 00:30:33 Vishesh Choudhry And so what's interesting is, you see, there was this valley sort of in the past week or so of those rates, they've kind of come down a bit, and then started to run up again. And so this buffer between the stability fee and these rates is actually tightening up a bit, I feel. So it does like drift a little bit at times, but that's not a huge margin, it's just sort of been a comfortable... and I think this tends to be a pattern, a comfortable 3% or 4%, occasionally 5%, occasionally 1%. And so on average, I think it tends to be around 3%or 4%.
# / 00:31:17 Vishesh Choudhry The supply or the outstanding amount of borrow and supply is pretty interesting. So as I had mentioned in May, you've kind of seen this explosion in the amount of DAI that was being borrowed on secondary lending platforms. And that share basically had come right out of the amount of DAI that would have theoretically been minted from CDPs.
# / 00:31:45 Vishesh Choudhry And so I haven't shown it here, but there's a write up I'm working on where you'll be able to see that, which is essentially the total outstanding loan volume, taking into account CDPs, taking into account secondary lending platforms, kind of just has been consistently increasing. But the supply of DAI had come down, even though that had been increasing in sort of a perfect complement to the increase in borrow from secondary lending platforms.
# / 00:32:18 Vishesh Choudhry So basically, that borrow had come right out of new DAI that would have been minted, which I think is actually a good thing. Because what it allows is sort of the narrowing of the gap between supply and demand. And that reduces our oversupply problem that had been experienced in the past, which makes the system more robust, reduces some of the downward price pressure on DAI, and I think is overall a net positive. It's effectively value being thrown down the drain, is being recovered, in the sense that risk was being taken on unnecessarily has been shed for a premium.
# / 00:33:01 Vishesh Choudhry And I think that's good is, when you have risk that you don't want, and you're able to offload it so that someone else can get compensated for taking on that risk, that is a good thing. And so it's just an improvement.
# / 00:33:18 Vishesh Choudhry And so what we'd seen is, what I think is a little bit more curious, the borrow is explicable, I think, pretty well. What's really curious is this supply on those secondary lending platforms has run up as well, and sort of jumped in the past week or so. And so as long as this supply exists, there is that buffer that I've been talking about. When this supply is increasing, that makes me sort of think that the increase in supply of DAI might be an increase in oversupply, rather than an increase in demand.
# / 00:34:02 Vishesh Choudhry And so this has been a big debate lately. And I know we always want to believe that the demand is increasing. But it is entirely possible that that has, by and large, or at least to some significant extent, been an increase in oversupply. Now, I don't think it's quite as simple as saying, okay, well, the supply on these comparable platforms was around 14-15 million, and had run up to around 19.5 million. So there's 4.5 million that's attributable to oversupply, and the remaining 5.5 is attributable to demand. I don't think it's quite that simple.
# / 00:34:48 Vishesh Choudhry But I do think that there is some relationship that is close to that, which is maybe in a more percentage weighted term basis, the amount that this supply increases versus stays stagnant, relative to the increasing DAI supply, gives you a hint as to how much of that increase in DAI supply was due to a legitimate increase in organic demand.
# / 00:35:16 Vishesh Choudhry Now, organic demand for DAI is actually a very tenuous concept as well, because there's a lot of question of whether you're going to start counting leverage as an organic demand use case, versus not. I think, right now, you kind of have to, because leverage is sort of that big driving use case for DAI. There's also sort of that volatility shelter, there's legitimate transaction volume, there's the need for flash capital and just legitimately permissionless loans, et cetera.
# / 00:35:51 Vishesh Choudhry Unfortunately, I think trading volume relative to transaction volume has... I don't have that graph here in front of us. But trading volume relative to transaction volume has been high. And just historically for DAI. And so I do, unfortunately, think that means that lending derivatives, leverage, et cetera, are right now big drivers of the usage of DAI.
# / 00:36:17 Vishesh Choudhry So just a little bit of a preview here is stability fees have started to map out a bit. And so those stability fees have significantly, the paid stability fees have significantly increased as a lot of people know, between sort of March and July. But the unpaid stability fees, it's my sense, have been increasing at a faster rate than the paid stability fees.
# / 00:36:45 Vishesh Choudhry And so this is something that I've kind of encouraged us all to think about in the past, which is to what extent the age of the underlying debt gives us a clue as to what's going on with circulation and by extension, fees. Because if that circulation is low, then you can see a lot of unrealized revenue. And unrealized revenue is simultaneously future potential revenue that may be underpriced by the market. And so that has a lot of implications for people looking at revenue and valuation.
# / 00:37:20 Vishesh Choudhry But then it also speaks to a level of risk in the sense that some of this revenue, you're paying a cost for, in the sense that it's capital that's not realized today, and so the time value of money, that's capital that you're losing. And then on top of that, in the event that some people sort of take this odd exit option and decide to never pay back loans, and have abandoned debt, which is something that we have really no visibility on, we can start to look at sort of stagnant CDPs, and see if this is something that is prevalent at all.
# / 00:38:01 Vishesh Choudhry But it's very hard to get visibility on that level of risk. I think it's been a lot of sort of theoretical discussion about whether that is an actual risk or not. And I know that that was something that like Rune chimed in on in the past, where there's the intent for there to be a sort of exit tax in the switch over to MTD, where if people are not paying back their CDPs, they can't just sort of switch over and get liquidated at a profit. But there may be the intention for some people to just pay the liquidation penalty and sort of not pay their fees.
# / 00:38:44 Vishesh Choudhry But again, there may be an element of market rationality, there may be an element of data that we just have a hard time getting eyes on. And this is a really, I think, hard and unpredictable problem to model out.
# / 00:39:00 Vishesh Choudhry Okay, I know that was a lot, I'll kind of pause, allow people to ask questions. Do we know what proportion of DAI is supplied to secondary lending platforms? So, yeah, we do. So at least for dYdX and Compound, which are a big portion of the amount, it's about 19.5 out of 90. And then LoanScan has Dharma's as well, which I haven't checked on recently. But I think if you broadly add those three together, that makes up a big portion of the amount.
# / 00:39:45 Richard Brown I think the question there though, and Steven, correct me if I'm wrong, what portion of that is actually being lent out? And I think isn't the assumption that with these variable lending platforms that they are actually lending out almost immediately like the DAI is being put to work?
# / 00:40:01 Vishesh Choudhry The utilization rate?
# / 00:40:03 Richard Brown Yeah.
# / 00:40:05 Vishesh Choudhry Yeah. So I haven't graphed the utilization rate here. But I think these platforms, especially like Compound, for example, set the rates so that they maintain a relatively steady utilization rate, the interest rates, I mean. But I have not graphed that over time, it should just be the ratio of these two, which if you look at, currently, Compound, 13.8 supplied and 8.5 borrowed. So that gives you a sense. And then similarly for dYdX.
# / 00:40:45 Richard Brown I'm going to jump the queue and actually ask a question. So you talked about the... Sorry, let me try and formulate this question for you. Okay, before I embarrass myself, I'm going to give us some thoughts. So let's focus on the questions and decide then.
# / 00:41:09 Vishesh Choudhry Okay, can you help me a bit, [inaudible 00:41:10].
# / 00:41:11 Vishesh Choudhry Yes, so revenue is a misleading term that maybe I shouldn't even use with regards to Maker. But stability fees are, quote, are paid, and cause the deflation of Maker, and so its overall appreciation for everyone. And so Maker, like stability fees are paid literally in Maker. And so that amount of Maker that is used to pay those stability fees goes into a burner wallet where it's never seen again. And that burner wallet currently is not burning MKR, as I understand, but it will be eventually.
# / 00:42:00 Richard Brown Okay. I think I figured out my questions. I'm going to jump in there, sorry for this Vishesh. So we're talking about the, you said that people draw DAI out of a CDP. And if they're not actively putting this into a use, or they need to store it for any certain amount of time, they can stick that DAI into a secondary lending platform and sort of shave off some of the hit they're taking with the stability fee.
# / 00:42:24 Richard Brown But is there some way for us to figure out which... the DAI that's in secondary lending platforms has this already been through the ecosystem once its capital has been deployed by people who've taken it out of CDP, done whatever it is they need to do? It's now in the ecosystem and after it hits the ecosystem, it ends up... another demographic takes that floating DAI and sticks it into the secondary lending platforms?
# / 00:42:48 Richard Brown Or is the expectation that a significant amount of DAI in the secondary lending platforms has gone directly out of it, or come directly out of a CDP? Is it possible for us to understand that?
# / 00:43:00 Vishesh Choudhry It is possible, I believe. Have I done it? No. I think it involves a fair amount of detective work in the sense of tracking transfers between addresses, which is, I think, a pretty heavy chain analysis question of, to what extent can you track the movement of the money straight from the CDPs, all the way to being supplied to those secondary lending platform contracts?
# / 00:43:31 Vishesh Choudhry Alitheo had done some work on this with regards to the overlap between Compound and Maker. And I think what they had determined was there was a pretty heavy overlap. I don't think they looked into, to what extent did, like what's the proportion of that relative to the DAI that was not overlapping effectively, and had gone through second or tertiary hands before getting to those secondary learning platforms?
# / 00:44:06 Vishesh Choudhry I am very curious about this. But no, I have not computed that answer at this point.
# / 00:44:15 Richard Brown Thanks.
# / 00:44:17 Akiva Dubrofsky Yeah, I have a quick question. I'm wondering about this whole idea of letting the DAI return to the peg on its own. I'm wondering if it's potentially an issue, because I think we want to move DAI back and forth above and below $1 as often as possible, so that we inspire confidence in the eyes of market makers and arbitrageurs. The longer DAI says below $1, the less market makers and arbitrageurs there will because they won't be able to get as good returns if it's staying below $1 for a significant period of time.
# / 00:44:49 Akiva Dubrofsky I think we need to take an aggressive approach and move the stability fee to get back it on par. That way, in the future, the deviations can be smaller, because there will more of these arbitrageurs and market makers.
# / 00:45:02 Vishesh Choudhry So I think the, and this is not a new question by any means, this is sort of an ongoing discussion, but I think the first part of that is 100% true, which is when and by definition, so if DAI perfect case, if DAI was literally always $1 and never traded above or below, market makers would have 100% confidence that in the event that it ever... Actually, this is probably a fallacy. If DAI had always perfectly been $1, then if it drifts below $1, would market makers immediately think that, oh, I have strong confidence that this will return to peg?
# / 00:45:47 Vishesh Choudhry I don't think the answer is necessarily yes. I think it is helpful for market makers to see DAI oscillates above and below $1 pretty predicatively. And I have seen it drift from peg before many times and still return to the peg. I actually think in that case, there is a greater confidence that in the event it dips in the future, that they will arbitrage it in the hopes that it will return to peg.
# / 00:46:18 Vishesh Choudhry I do definitely think that the further it drifts from the peg for the longer period of time, 100% that reduces the confidence of market makers. And I don't know if there's any market makers currently on the call, but I love hearing from them.
# / 00:46:32 Vishesh Choudhry So if there is, feel free to interrupt me, any of you, but the one thing is, I don't think the second part of what you were saying necessarily follows from the first part, where just because that does necessarily increase the confidence of market makers, doesn't necessarily mean that we should then use the stability fee to respond to times when it drifts from the peg. I think there are really important questions of, how much has it drifted from the peg and for how long, before we start to... There should be kind of a noise pass filter before you start to use that stability fee as a lever to be changed.
# / 00:47:15 Vishesh Choudhry So then there starts to become important questions of, is 0.984 for a week, long enough to say we have a problem that needs to be, where we need to turn the dial to solve it, versus is 0.99 for two days enough for us to say that we should change the stability fee? And so having that kind of cut-off really well thought through is important, both in terms of magnitude and in terms of timescale.
# / 00:47:45 Vishesh Choudhry So what I was saying before is, I think, I'm not saying you should never use the stability fee to fix the peg, what I'm saying is 0.984 might be within the acceptable threshold. And the amount of time that it has been drifted from peg, might not be long enough for us to be seriously concerned. Because if we do change the stability fee, in perfect, sort of one to negative one response of movements above and below the peg, we may actually just double the amount of error and noise in the system.
# / 00:48:20 Vishesh Choudhry Because the peg will naturally move a little bit. The question is just how much is a little bit, versus a lot? And I do think that 0.984 is within kind of the gray area. But I think if you did increase it by a point, half a point, at this point, that's not wholly unwarranted. If you increase it by a point or two points, I think that's maybe more than... Sorry, more than 1% right now. I think that may be a little bit aggressive. And it's very hard to get to a mathematical equation of what is aggressive, versus not, or too aggressive versus not, in terms of exact percentage terms.
# / 00:49:05 Vishesh Choudhry But I think based on experiences of how much has been changed before and how much the price moved in those times, maybe, and something a little bit less, 1% or less at this point is justified. I think 2% is more than is probably justified, based on the amount that it has drifted and how long it has drifted for. Sorry, long answer.
# / 00:49:32 Richard Brown We have another question in the chat, Vishesh, from Stephen.
# / 00:49:36 David Utrobin We lost you Vishesh.
# / 00:49:39 Richard Brown Can you actually read out the questions, for the [crosstalk 00:49:41]?
# / 00:49:41 Vishesh Choudhry Yeah, sure. So the velocity of arbitrage for DAI is a little slower than fiat backed coins. Do you think this would speed up in the future? Or do you think speed is important at all?
# / 00:49:51 Vishesh Choudhry Speed of arbitrage in the sense that how quickly DAI drifts and returns from peg is important, for sure, I think, and I don't want to just sort of become the justificar on what is or isn't important. But I think what you can say is when that arbitrage cycle is slower, there is a little bit more risk, both for the system and from the market maker perspective. But again, I don't want to speak for market makers. I would much prefer they speak for themselves, in terms of their incentives.
# / 00:50:37 Vishesh Choudhry Is it a problem? Like is it a significant problem if that arbitrage cycle is slower than the ability for a reserve backed coin to mint or burn? I don't think that's necessarily a problem. I think it's always going to be like that, there's always going to be that delta. I think arbitrage mechanics and liquidity constraints are always going to cause sort of that true economic arbitrage from a decentralized market perspective to be a little bit slower.
# / 00:51:13 Vishesh Choudhry I think as DAI becomes more liquid and more heavily transacted on, that cycle is going to speed up. And of course, it'll help, and I think it's good, quote-unquote. But I can't tell you how important that is versus not.
# / 00:51:38 Richard Brown Okay, cool. Thanks for Vishesh, that was amazing as always, we'll have a recap, or we will have a synopsis for us that we will post into the forum and the subreddit, that captures all of the salient points. We're at the top of the hour, almost. But I don't want to skip this next section. So if you feel like hanging out, please do.
# / 00:52:02 Richard Brown Matthew, did you have a weekly recap for us that you wanted to talk about?
# / 00:52:07 Matthew Rabinowitz Yes. I mean, yes. So basically, the main point that I wanted to chat about was, as we migrate from a single collateral DAI world to a multi collateral DAI world, there is not only a question of where we're going, it's also the big picture steps we have to do to get there.
# / 00:52:26 Matthew Rabinowitz And as I understand, that's always a loose point, we're switching, we will be pivoting from what we're doing today, we're almost halfway ignoring our risk policy, and in a single collateral world, and smushing risk policy and monetary policy. And as a function of switching to multi collateral DAI, we are basically bifurcating the two, where we're going to have collateral with a set quantity of parameters, whatever that debt ceiling, liquidation ratio, collateralization ratio is for a given collateral, just pick ETH for a moment.
# / 00:53:06 Matthew Rabinowitz And that will have a given risk premium that's associated to it too, in effect. And this is a risk team question. It's not for me to answer, but should be to the point where it makes the output, any DAI that's minted from that, in effect, riskless. More or less, probably to a 95%. If it's 100%, of course, the risk premium would be infinite.
# / 00:53:30 Matthew Rabinowitz So it goes to a certain degree. And the question is, how do we go from single collateral DAI to multi collateral DAI? In the context right now, by having to mix or basically ignore risk policy, but only using monetary policy, by changing the risk premium, we've ended up with a stability fee for the underlying collateral of Ethereum of whatever we're at now, 16.5%, which by definition of how many times Maker has had to be minted to cover any losses from a liquidation, which thus far, after a 95% drop in Ethereum with zero, we can argue that 16.5% is too high. In fact, we even had it down to 0.5% and it still didn't have any losses.
# / 00:54:18 Matthew Rabinowitz So the question, really, or the community, and there is a mathematical form, it's not just how you feel, but what would be the risk premium, given the same parameters that we have today? I suspect that it's way, way lower than 16.5. The question really is, how do we go correctly, how do we move from single collateral DAI to multi collateral DAI with a new risk premium that's associated to just the risk, not the quantity? Having nothing to do with the price of DAI, just pricing a collateral quote-unquote package, correctly?
# / 00:54:59 Matthew Rabinowitz And this is the, I have ideas of what I would recommend that we do. It's obviously not for me to say what we do, but rather, we need to have rough consensus from the community as to, how do we migrate from one to the other? If we've got a new risk premium, or of the existing collateral pool that we're using, and we're migrating, one would assume that if that new risk premium is materially lower, say 2%, instead of 16.5%, if we just flipped the switch, we'd have a surge of DAI, which would erode the price. So there's a sequence that we have to go through to navigate the waters from single collateral to multi collateral, of which we can go through some of those recommendations now, if you want.
# / 00:55:48 Richard Brown Yeah, sorry. This is an interesting conversation. I think that I followed some of the debate in the forum, but I wasn't able to like follow it closely. So you're saying that, right now, we understand the system as... Sure. I'll say we understand the system right now with ETH as a collateral type. It's been priced in, there's premiums associated with it and all the rest, we know what's going on, basically.
# / 00:56:12 Richard Brown But the minute we add a new collateral type, where we're basically shifting the sand beneath our feet, and we don't understand what impact that new collateral type is going to have on the underlying risk models that has established that stability fee, or the stability that created the DAI that's already in the ecosystem.
# / 00:56:34 Richard Brown So you're saying that, because DAI is sort of fungible, because we don't know which DAI has been backed by which collateral type, it's difficult to predict what the risk that's been baked into the system.
# / 00:56:48 Matthew Rabinowitz Just for right now, the question is, what is the risk? What risk premium would we logically, and this is a risk team question, like what is the risk plus or minus that DAI with 150% collateralization ratio, with Oracle feeds every 30 minutes that come in, with a debt ceiling of whatever number we picked, 150 million, it doesn't matter, what is the risk premium such that, when the risk premium is in effect, an insurance fund, that's actually part of the question, like what is a Maker token?
# / 00:57:20 Matthew Rabinowitz It is, in my understanding of it, it is, in effect, a decentralized credit facility, but it also has the simultaneous effect of being an insurance fund, where we're taking a collateral package, and the fact that we've got a cool liquidation tool, a bankruptcy court in another way of saying it, that allows us to liquidate collateral in the event of a breach of that liquidation ratio. Would that risk premium that we're associating in a multi collateral DAI world, is just what does it take to cover the theoretical losses of a liquidation event? It should have nothing to do with monetary policy, nothing to do with DAI outstanding, nothing to do with the DAI price. Purely to do with the collateral at hand.
# / 00:58:12 Matthew Rabinowitz So if we take 16.5%-
# / 00:58:12 Richard Brown This is the part I'm not-
# / 00:58:18 Matthew Rabinowitz Go ahead.
# / 00:58:18 Richard Brown I just want to say that this is probably, I'm completely unqualified to answer that question. But from an amateur economist perspective, that's being immodest I think, or overly modest, I don't know, sorry, you know what I'm saying, you're saying that the risk in this system has to be looked at holistically. There's no way to compartmentalize this stuff.
# / 00:58:43 Richard Brown And my understanding of how the system is designed and the way that it's supposed to operate is that it is compartmentalized, that there's a liquid risk for token X. And that's why it has a very, has the stability fee associated with this very conservative debt ceiling associated with it. And so the addition of token X, the risk there is essentially capped by how much of token X can be liquidated and of course, 24 hours, if token X goes south rapidly?
# / 00:59:17 Richard Brown And from that perspective, these are individually units of risk. And you're saying that they're not individual units-
# / 00:59:22 Matthew Rabinowitz No, no, no. I'm explicitly saying they are compartmentalized. So when we take our existing system today, just imagine right now we could flip a switch and it was going to exist in a multi collateral world, just assume the software got deployed right now. Does a stability, excuse me, does a risk premium of 16.5% make sense for the Ethereum collateral package that we have with 150% collateralization ratio, does that make sense?
# / 00:59:58 Matthew Rabinowitz My argument is that that is way overdoing the actual risk, because we've embedded monetary policy in that risk premium, not only risk policy.
# / 01:00:11 David Utrobin But isn't that to say that the risk premium is just a component of the overall rate. and that because of the system's design, you have to have sort of a monetary policy like adjustment to the rate at some level? [crosstalk 01:00:26] have to be spread out in some way.
# / 01:00:29 Matthew Rabinowitz Exactly. No, this is, I mean, we're going to watch this video in five years and say, look back and say this was a short-term phenomena where we were using the risk premium to handle monetary policy. We should be dealing, the risk premium for every single collateral should be specific to that one collateral package. And also, the covariance is to see if they are related to other collaterals. It shouldn't have anything to do with monetary policy.
# / 01:00:56 David Utrobin So is the bigger question, do we use in multi collateral DAI the stability fee as a policy tool? Or do we rely solely on the DSR for monetary policy kind of adjustments?
# / 01:01:09 Matthew Rabinowitz That's what we should do. Yeah, I mean, at the end of the day, my respectable view is that we should be looking at each one of the collateral packages in isolation. Again, there's covariances. So if you've got Ethereum with 180% collateralization versus one with 1,000, and you've got them sitting next to each other, you have to include how the two might interact, right? It's not that they're completely isolated.
# / 01:01:36 Matthew Rabinowitz But the risk premium that's associated for, let's pick the one that had a collateralization ratio of 1,000, should be much, much lower than the one that has a higher, excuse me, a lower collateralization ratio. The risk premium for the one that's more risky should be higher. But that's all we're focused on.
# / 01:01:55 Matthew Rabinowitz So the question is, or the point I'm trying to raise is, as we navigate the waters when we launch multi collateral DAI, and again, there's a community question feedback, my understanding of it is when we launch multi collateral DAI, when it is launched, the DSR is not being launched in parallel and should be launched sometime after the fact. The question is, when you move from code set A, single collateral DAI to multi collateral DAI, and then you go from multicultural DAI to a world where you start introducing the DSR, we need to find a phased approach where we migrate without...
# / 01:02:35 Matthew Rabinowitz And the ultimate objective here is to continue doing everything I'm mentioning without degrading the price of DAI during this endeavor. There's an economic sequence to do this, where we in effect can refi out collateral package that has a stability fee that's ridiculously too high and introduce a new stability, excuse me, a new collateral package with identical parameters, but a quote-unquote correctly priced risk premium, and slowly raise that debt ceiling at the same time, we're introducing a DSR to offset the DAI that's minted from it.
# / 01:03:12 Matthew Rabinowitz The point more than anything is, as we start introducing more and more collateral packages, how much we use a given collateral should be totally irrelevant, as long as it doesn't introduce too much risk. Risk should be managed by a risk premium, monetary policy should be managed by the DSR.
# / 01:03:36 David Utrobin But can't you look at oversupply of DAI as a monetary risk? [crosstalk 01:03:43].
# / 01:03:43 Matthew Rabinowitz That would be why you increase, I mean, the idea being is that it's a riskless offset. So if there too much riskless DAI that was minted, you are fully justified in increasing the DSR to, which is a riskless absorbing tool, a mop, to justify it. You're dealing with a riskless point. You can ensure package at that point.
# / 01:04:18 Matthew Rabinowitz The point is, there's a migration path that we have to go through, and we can go through the forums and chat more about that. The question is, how do we do that without disrupting the price of DAI? More importantly, the moment we start actually, what I would say, correctly pricing the risk premium as it relates to giving collateral package, in effect, segregating, bifurcating monetary policy and risk policy, we will end up with a currently model Ether with a risk premium that's materially lower, which of course, raises the question, what happens to secondary lenders that are used to relatively high rates if we start introducing rates that are reasonable, but we now have a secondary tool to absorb the excess DAI, that maybe they don't have?
# / 01:05:09 David Utrobin Yeah, that's fascinating stuff. I'm excited to see how those conversations play out on the forum. That's going to be a really interesting longtail conversation. At least, it already has been.
# / 01:05:23 Matthew Rabinowitz I mean, because if you think about it, if we just segregate for risk and forget REP, we talk about WBTC. I know there are liquidation concerns about how liquid is the secondary market to be able to sell WBTC? And of course, that needs to grow. So what should the risk premium be? Well, it should be pretty high, if there's not a whole lot of liquidity.
# / 01:05:47 Matthew Rabinowitz If there was a lot of liquidity behind it, just imagine somehow that the entire Bitcoin BTC network could be used and utilized, what would be an appropriately priced risk premium with 150% collateralization ratio? It wouldn't be 16.5%, it probably wouldn't be 6%, it'd probably be even lower, because of how the depth of that market to sell it in the event of a liquidation.
# / 01:06:18 Matthew Rabinowitz Now, the secondary question that comes out is, how high do you have to raise the DSR? And this is like another equilibria question. How high do you have to raise the riskless, risk-free rate to encourage enough would be investors, speculators, to buy DAI, to lock it into a DSR, where they're taking zero risk? How high do you have to raise that number? Again, the DSR is uniform in every single one of the... in every single stability in a multi collateral DAI world. How high do you have to raise that, how long do you have to raise that to absorb that excess DAI?
# / 01:07:00 David Utrobin Is there a fear that like secondary lending platforms will totally eat up all of that DAI that should have gone to the DSR? Because it has like I guess higher rates for lending?
# / 01:07:15 Matthew Rabinowitz That's great. Let them. I think that's a wonderful thing. If they're absorbing, if they're buying the DAI, that the excess DAI and finding some kind of use for it, it's no different than you buying it and putting it in your wallet and locking it.
# / 01:07:29 Matthew Rabinowitz So we can kind of dovetail the conversation on that. I'm happy to go on for another half an hour on it if you want it. But the question here is just the sequence. And it's not only what I think, it's how the community kind of views it, and not only where we want to be in five years of having 500 or 600 different types of collateral and a huge portfolio and a stabilized system. But how do we navigate these waters in the next six weeks, 10 weeks, 50 weeks, whenever it is, where we don't disrupt the price of DAI in the process?
# / 01:08:16 Richard Brown It feels like this is a conversation we're going to have to get back to, because it's above my pay grade. But it's also coinciding with July 4th, and almost a complete absence of the risk team. So I'm not sure we're going to get a very furious debate out of that topic today. But let's bring it up again tomorrow. I mean, sorry, next week. And let's dig into this, because it sounds really interesting. I have no answers. So I would like to figure out what those are.
# / 01:08:45 Primoz Kordez I could jump in a bit.
# / 01:08:46 Richard Brown What's up, Primoz? What have you got?
# / 01:08:50 Primoz Kordez I think what Matthew is talking about, so what I believe is that the DSR is probably going to be the main lever to adjust the imbalance. So the thing is that the stability fee will always carry, as I think David mentioned on the chat, is going to carry, he says it's almost like the stability fee has an additional monetary risk premium. I think it actually has an additional monetary risk premium. Because you will always need to finance DSR with stability fee.
# / 01:09:24 Primoz Kordez So with the additional part, upward of the risk premium. This will need to happen. And unfortunately, that's the case. But don't forget, you also have other levers. So you can adjust the collateral ratios and so on. So you have many things at your disposal, where you could adjust the imbalance.
# / 01:09:48 Primoz Kordez But yeah, I totally agree, I mean, I'm not saying there's going to be a problem, but we will need to adjust for this. And there's always a question how we're going to do this. DSR, there's also another tricky thing. So let's imagine there's DSR with, I don't know, 3%, 4%, 5%, whatever. And the demand of DAI outpaces the supply.
# / 01:10:11 Primoz Kordez So in that case, you would need to lower DSR and what are you going to do with stability fee? You are going to lower it as well, I guess, otherwise, you're paying extra yield to the governance. And that's-
# / 01:10:32 Matthew Rabinowitz To me, now we're inserting risk. And that is the danger in this whole equation is, how do you maintain risk policy segregated from monetary policy? Because reducing, if by reducing the stability fee, you mean, reducing the DSR and thereby it reduces, we shouldn't be touching the risk premiums associated to a given collateral for the sake of policy to change the price. Unless the conditions on the ground related to that collateral mean that it's now less risky.
# / 01:11:06 Primoz Kordez Yeah, I totally agree. But what I'm saying is that you might have an additional spread on the risk premium, and you will also deplete it when the DSR is adjusted downward, so you have two-sided effects as well. But yeah, I mean, I agree with you that we will need to compensate for it with stability fee. And how we're going to do it is another story.
# / 01:11:31 Primoz Kordez What I think is also important is, so now let's imagine there's oversupply of DAI from one collateral, which collateral are they going to punish? You talked about that you would prefer to punish the risky ones. What I believe, which one I would punish, if I would state it as such, I would punish the ones who are the most, how do you say it, elastic to the change, right?
# / 01:12:00 Primoz Kordez So imagine if one collateral brings a lot of DAI, and you change the stability fee downward, it is going to be probably the most elastic or the most responsive to the stability fee change. So that's my proposal there, I guess, in regards to that.
# / 01:12:21 Matthew Rabinowitz On that front, yeah, I mean, sorry, just one second, on that in front, this is where like my learning of this, I mean, I'm by no means an absolute expert. I'm just trying to learn as much as I can. I've actually kind of switched back from that mentality of how do you quote-unquote punish, and rather took a lesson from what Vishesh mentioned in the back channel communication of just saying, why don't you just correctly price it on the front end? Where we really shouldn't be punishing any party for minting DAI, if they were priced correctly.
# / 01:12:50 Matthew Rabinowitz If it was taking the parameters of a given collateral package and a risk premium, whatever that risk premium was that made it through the consensus model we're going to have with teams, risk teams, whatever, that meant the DAI that was minted from it was, in effect, derisked or insured or quote-unquote, riskless.
# / 01:13:13 Matthew Rabinowitz So if a risky collateral package, even in an extreme case, like we've talked about with REP and how it has some challenges, even if somehow we come up with a model where we say, listen, the risk premium is going to be 500% for REP, because there's a possibility that it could default. Well, if people then mint DAI based upon some really high risk premium, well, that's fine. That's why it was insured in the first place. That's the point is, when we take a package, and we derisk it to the point where it is no longer risky, it has been insured and quote-unquote riskless, the DSRs used to mop up the monetary policy, mop up the excess or in the scenario you outlined, when you have too much demand, you might have to decrease the DSR.
# / 01:14:04 Matthew Rabinowitz But changing the risk premium meant that it wasn't set correctly in the first place. Or it meant there were situations on the ground that changed. When I started this whole thing was, were the conditions on the ground for Ether as the fundamental collateral base with our current parameters... The parameters haven't changed, the business case, the liquidity market for Ether has probably not changed, if anything, it got better in the last year.
# / 01:14:37 Matthew Rabinowitz So does the risk premium that we're associating with it today of 16.5% make sense in multi collateral world? In two years, definitely not. For the first sequence to how you go from, you have multi collateral DAI launched. What is step number one, two, three, four and five? Step one probably means you have a risk premium that is elevated. But it gets phased out, because it's not correctly priced.
# / 01:15:14 Matthew Rabinowitz David, I cut you off, I apologize.
# / 01:15:17 David Utrobin That's okay, I just had a little bit of a tangent of a thought, where would it make sense in the context of having a lot of different collaterals, a lot of different collateral packages to actually use debt ceilings as a means to dampen DAI supply? Because at that point, debt ceilings would be far less risky on the individual collateral basis. And if you just lower it to below the current amount of DAI, then you could just basically stop DAI creation from like whatever asset you choose. And you could start doing this from like the riskiest assets, or however you want to do it, but I was just going to ask about what you think about that.
# / 01:16:00 Matthew Rabinowitz Yeah, I think it's a very solid policy tool. The sequence that I have in my head that I would entertain or recommend to the community is step number one, you launch with a collateral package that's identical to what we have with Ether today, with the identical stability fee, so it shouldn't change the price. And a debt ceiling of I don't know, pick a number, 120, 130, 150 million, something where it doesn't really rock the apple cart.
# / 01:16:26 Matthew Rabinowitz And at the same time, the next collateral package you launch has identical parameters, but the correctly priced risk premium to be determined. Let's just say it's 2%. But what David just said, with a debt ceiling of, I don't care, $100, $500, something where it'll get used instantly, but it's not enough to change the price of DAI. That would be basically step one and 1.5.
# / 01:16:56 Matthew Rabinowitz Step two would be to launch the DSR and have it be 0.1%, something where it doesn't get a whole lot of interest, but we can test it. Step number two, or excuse me, step three would be to then start to increase that debt ceiling on that second collateral package, and alternate back and forth between that, and increasing the DSR, so that we're not changing the price of DAI, we're not really doing anything that changes the market.
# / 01:17:23 Matthew Rabinowitz And keep doing that, up until the point where folks from the first collateral package start to refi themselves out, if they find themselves lucky enough to get in that spot on the second collateral package. Until we get to the point where you kind of let it loose in the wild. And once it's on its own two feet, where we have correctly priced collateral package B, with a very low risk premium and a DSR that's effectively mopping it up with whatever that equilibria is or would be, then you drop the collateral, excuse me, you drop the debt ceiling on the first collateral package down to zero. So nobody else new can hit something that's incorrectly priced.
# / 01:18:07 Matthew Rabinowitz And yet, folks will naturally over time, refinance themselves out of collateral package A to collateral package B, or whatever else is in the market. We got to find a way to do that without disrupting the price of DAI.
# / 01:18:20 David Utrobin Cool, man. Thank you for the answer.
# / 01:18:28 Primoz Kordez Guys, I just want to switch the debate a bit about the state of the peg and the graphs that Vishesh showed. I think there's one interesting point to it. So I think Rich asked how much of the DAI that's drawn from CDPs has ended in the secondary platforms. And I think, I mean, that wouldn't ultimately make much sense, right? Because why would you pay a higher interest rate for borrowing and then lending it out on secondary platforms? Because the interest rate there is, as much as I know, is now lower.
# / 01:19:06 Primoz Kordez This would make only sense if there are market makers who have spare DAI, and they're waiting the price to increase, so they would sell it in the future. And meanwhile, they want to hold this inventory in lending platforms. But I don't know how much DAI there is in their inventories now.
# / 01:19:25 Primoz Kordez So I think most of the supply of the lending on secondary platforms comes from, I don't know, retail users who just see a huge benefit because of extra yield on DAI. And I think there's becoming, the awareness is increasing on this extra yield. So this on one hand I think helps us with the peg.
# / 01:19:48 Primoz Kordez So I think the situation with the price would be much worse without this. Because we have the demand, the people are buying DAI and making extra yield. And don't forget also, that borrowing from secondary platform also is positive for the peg, because these people who are borrowing there and selling it for Ether or whatever, making leverage, they would otherwise go to Maker. And they would issue DAI there and sell it.
# / 01:20:22 Primoz Kordez So you have this positive effect of the demand and supply on secondary platforms actually helps us with the peg a lot currently. So this should be taken into account about the current price. And we haven't seen such surge for some time, I guess.
# / 01:20:50 Richard Brown Thanks for that Primoz. It's interesting to try and figure out what's going on, at some point, we're not going to be able to avoid having to do some deep chain analysis to find out exactly what's happening here. Because we're going to need that information to mull the system effectively. I appreciate the follow-up.
# / 01:21:07 Richard Brown We're at 10... sorry, not at 10, I'm at 10, everybody is at 24 minutes after the hour. I'd like to wrap it up here, actually, and keep an hour and a half as a nice tight package for these calls. Thanks everybody for sticking around and thanks Matthew for your recap, again. Lots of food for thought. Let's not lose this conversation though, because it feels like, as David said, this feels like a debate that is going to be lasting for weeks or potentially months. So let's keep the momentum going.
# / 01:21:38 Richard Brown And I'll be posting this to our subreddits, I will be posting links to the forums as well. So if you're interested in the summary, keep an eye out for that. All right, thanks, everybody.
# / 01:21:50 David Utrobin Thank you, Rich. Take care, everybody.