If you would like to view the full recording of the live Crowdcast event, you can find it on our Youtube channel. You can also read the full Quarterly Report by Dustin Teander and Sami Kassab here or Ryan Watkins’ Q3 Quarterly Report here.
Dustin Teander (00:04): All right, I think we’re live! GM, GM. We’re here for the Compound Q4 analyst call. If you’re not familiar, we’ve done one, and this is our second iteration actually. We did one back in Q3 with Watkins and some other people. This is our second iteration. And, my name is Dustin, we have Sami here as well. We’re both from Messari and we kind of co-authored the report together and we have Getty Hill here from GFX Labs, a really prominent member inside of the Compound community. But you know what we’re going to be covering today is very similar to kind of if you’ve ever been familiar with equities calls and kind of Q4 results, that’s what we’re looking into. So it’s the health of Compound, the metrics that matter and truly trying to peel back, just what went over Q4. Before we do all that, Getty, why don’t you walk us through your background, GFX Labs and the role that you guys play within Compound?
Getty Hill (01:10): Yeah, sure thing. Hi everyone and thanks for having me here. A huge fan of Compound and have been dabbling in the protocol for a long time. My background comes from three years of trading and running delta neutral strategies at Grapefruit Trading and then myself and my buddy on the trading desk decided to leave and start GFX Labs and we’ve been involved in Compound for a very long time. Loved it when the protocol launched COMP tokens and gave this opportunity for individuals to come into the protocol and interact and improve the protocol and do fun things. So we’ve done things like update the price oracle, add new assets to the protocol, all sorts of fun things and just real big fans of money markets and Compound in general.
Dustin Teander (01:53): Awesome man, all right let’s just jump into it. Sammy, you want to give us the 30,000 foot level and give us the macro results?
Sami Kassab (02:08): Yeah sure thing, let me just go ahead and share my screen here. Is everything all good, can you see it on your end?
Dustin Teander (02:17): Yep.
Sami Kassab (02:18): All right, so we’re gonna go ahead, get things kicked off, and start with a little macro overview. So we know that Compound is a financial service business and this matters because they’re influenced to an extent by the broader crypto market and the macro environment. There’s a lot of action that went on in Q4: we saw two breakouts in bitcoin leading to new all-time highs and we also saw a breakout in total crypto market cap with a new all-time high there as well. Due to this, we saw Q4 ending with quarter ending highs in several key categories. So we’ll step through this macro chart that’s on the screen, walk through the key performance indicators as well as the financials, but one thing I want to preface with is just looking at this Q3 data. So at the end of Q3 there was a COMP distribution bug that was introduced into the protocol. We’ll go into this deeper later on but why this matters is that… So this bug was introduced on the last day of Q3, so September 30th, and the bug caused some panic, caused some fear, there’s some uncertainty, and so about $3 billion of closed loans and withdrawn deposits affected the Compound market. So the Q3 metrics came in basically a lot lower than what they would have been for example if that bug was introduced the next day. I just want everybody to keep that in mind while we’re walking through this because when we look at the percent growth of Q4 it can seem high for what it really is. But if that’s a little bit confusing, it’s a lot easier to visualize once we go on to the next slides where you can see the liquidity and the capital moving throughout the Compound markets. So starting out with the outstanding loans for Q4 we’re at about $6 billion, that’s about 7.5% growth compared to Q3. Outstanding deposits were at $14.3 billion, pretty much stayed flat compared to Q3, very low growth there at 1%. Quarterly originations, also known as borrowing volume, came in at $11.4 billion. This was down compared to the last quarter. And then deposit volume was up quite significantly to $63 billion. Liquidations were at $34 million, this was a slight increase from the last quarter, and then we had aggregate utilization that is sitting at 41.7%. Pretty much for the past year we’ve just been ranging between the high 30s and and low 40s. Moving on to financials, total interest income for the protocol was $90 million. When we look at the interest, when we take out the interest paid to depositors which was $80 million, we are left with a net interest income of $10 million. That’s a pretty decent growth from last quarter where you’re looking at 11.1%. There were no grants paid out this quarter. The grant program ended I believe in September of last year but the Compound grants team is currently working on the second iteration which should be out in 2022. Currently no date has been set yet. But that led to the net income of $10 million still, close to 20% growth, we’re running a net margin of 11% on the protocol, and when you take out the token incentives paid which was about $55 million, the adjusted net income for Q4 was a negative $45 million. That was a significant improvement from Q3: about 41% growth. Now before we move on, Getty, here’s a question for you: what do you find to be one of the most important metrics or KPIs that you like to keep your eye on?
Getty Hill (06:43): Yeah great question, I think as these protocols have transitioned, especially compound out of its total infancy and have been around now for a bit over a year, really one of the more interesting things to begin to think about is that net interest income line there and the growth and the steady progression of the protocol over the last four quarters coming out to about a little over $40 million on the year is really impressive considering the amount of resources that are being deployed to earn that income. It really demonstrates that these protocols are being able to run themselves and generate some nice income in doing so.
Sami Kassab (07:22): Great, yeah, I agree. Dustin, do you want to take this one?
Dustin Teander (07:35): Yeah. So what we’re looking at here is our outstanding loans for Compound. It is really going to illustrate a lot of what Sami is talking about particularly around the COMP distribution bug. So if you look at the right quarter of that screen, you can see the almost rectangular trough, that’s that period of the COMP distribution bug. From that point on, we quickly resumed loan levels pretty much, obviously a little bit lower than where we were at the end of Q3, but stayed pretty flat until December. So what’s happening there? You have to remember that Compound is driven by the borrow or the demand for leverage. During that time we had bitcoin hit essentially two new all-time highs at the end of October and then end November as well. So that’s kind of fueling a lot of that growth there or at least during that period. We’ve mentioned the COMP distribution bug a couple of times. Getty, do you want to just explain real quick what is it, what does it mean when we say that, and walk us through a little bit of that timeline in that early October period?
Getty Hill (08:47): Sure thing, so for everyone who might not be super familiar with Compound, Compound when it was started came up with the COMP token as part of incentivizing participants to deposit liquidity in the protocol and distribute the COMP tokens in a decentralized fashion. The Labs team decided that it would be an interesting idea to distribute the COMP tokens for market participation. For example if you were to deposit some USDC in Compound, you would earn your pro rata share of the amount of USDC deposited and some COMP tokens that were being distributed to that pool. For the longest time since the inception of the protocol and that feature, it was distributed in a 50/50 split where the same amount of COMP, let’s just say it was like 20 COMP per block, just making up a number right here. 20 COMP per block, obviously it’s not that number, is getting distributed half to all the suppliers and half to all the borrowers and the thought was: hey, it would be really neat if we could adjust those parameters individually. So if we want to give 15 to the suppliers and 5 to the borrowers, we should be able to do that. So Proposal 62, originally, the premise was to introduce this new piece of code that would allow the parameters to be adjusted individually. Unfortunately, there was a small error in the code that was literally resulting in a greater than or equal to sign that was missing. The greater than sign was missing in the code and it ultimately introduced a bug into the code in calculating some of the supported assets and how they’re essentially indexing these rewards and how the math was being done. What happened essentially was a bunch of people who had assets in the protocol that were earning yields all of a sudden earned substantially more than they should have otherwise been able to and were able to claim a significant amount of COMP. In some cases it was more COMP than was even in the existence of the protocol so those were automatically failing transactions. In other cases users were able to claim 10,000 or 50,000 COMP from the protocol’s incentive pool. Luckily, it didn’t hurt any of the users’ participation in the protocol. Really what it was here is it was an unfortunate event for the protocol itself and all the COMP holders as it entered a significant amount of COMP into circulation that wouldn’t have otherwise all been immediately entered into circulation, but everyone’s loans and positions in the protocol were entirely safe and no one had to be concerned about their coins being at risk or being stolen or any of that. It was really just an unfortunate event for the protocol distributing a lot of money to individuals in an unfortunate fashion. Luckily, there was a lot of great effort amongst the community to figure out exactly what had occurred and went wrong and to react to it to the best we can. Now this is where it gets interesting, in the way of how governance functions and feel free to cut me off here because I can probably talk about this forever but what ended up occurring here, because compound is in fact entirely run by governance, the only way to change any code ever is to introduce a governance proposal and the governance proposal timeline is about seven days. So once the change was introduced to the protocol, the protocol’s hands were tied. Governance can’t actually rush to its aid, there is no multi-sig, Robert Leshner doesn’t have a special key somewhere that allows him to run in there and fix everything up. It is truly decentralized here, in a painful fashion in this instance where the protocol was bleeding funds. So long story short we managed to get the fix in place and once that governance cycle had gone through everything was back to normal, and then it became ‘how do we prevent this bug in the future’, which I think we’ll get into more later in this call. But that should be a quick synopsis on what occurred there on the back of that event.
Dustin Teander (12:54): No that was great. Yeah, very good detail. So once that was resolved, that was like mid-October by that time, the proposal resolved it and got enacted. From then on it was almost more or less business as usual. But a couple of key things that we can look at in this chart that I thought was interesting is that: one, it is pretty clear that the light blue and the dark blues are both stablecoins so stablecoin borrowing obviously dominates because people want their cash denominated in stable assets. So the mix between these two is actually kind of interesting. This is the first quarter that we’ve seen DAI almost outpace USDC so DAI came in initially to the quarter as not the most borrowed asset (that was USDC) and then over time, it’s a little bit harder to see in the chart but you see USDC with a slight decline while DAI is basically increasing. I think DAI increased outstanding loans roughly like 34% while USDC was slowly down 7%. I think that’s going to be an interesting narrative that we talk about. I don’t know what your opinions are Getty, about the growing importance of DAI in the protocol relative to USDC.
Getty Hill (14:10): Yeah, what’s interesting is that throughout DeFi summer it was seemingly evident that a lot of funds in crypto were very fluid and that money was moving at a rapid pace and what we’ve seen now on the back after a year plus time is that there’s actually a substantial amount of funds that are actually quite sticky and that have a preference to certain protocols and certain assets when it comes to what they want to borrow with, what they want to leverage themselves on, and where they want to leverage themselves. I think a lot of the data throughout this report and a lot of data in general supports that there’s a significant amount of funds that are interested in simply supplying ETH in a protocol they’re very familiar with and borrowing a coin that they’re very familiar with in the ecosystem, DAI. Similarly, there’s a lot of DAI interest on the other side supporting rates, stabilizing them and keeping them low to make it more attractive in many cases than USDC borrowing thanks to the quite colossal size that DAI is in the Ethereum ecosystem.
Dustin Teander (15:14): Yeah, no I agree. I think it is interesting that we are starting to see that it is almost more adopted as time goes on. You brought up the interest rates which is a good point. So the interest rates typically follow the market correlation. So as bitcoin starts breaking out there in November, you see in the right half of the chart that the interest rates start to spike for the protocol. Back to it, USDC and DAI are the two most borrowed stablecoins. USDC is really driving that average borrowing interest rate spike. If you look at the DAI chart, it’s almost flat to your point, and that’s really driving a lot more demand on the borrowing side.
Getty Hill (16:05): Yeah I think it provides certainty for people when all the DAI that’s in existence is all essentially these dynamic fixed rate loans as MakerDAO governance picks a rate for individuals to borrow DAI at and it provides a steady hand of like ‘hey, this is what the cost of DAI is at its source’, and that kind of sets the pace for everyone else where the dollar rates, dollars are king ultimately and are fluctuating wildly and typically are driven by the the cost of capital at some of the big lenders and then particularly in the leverage markets we look at perpetual markets and whatnot. That’s perhaps where the big dollar driver is in crypto.
Dustin Teander (16:49): Yeah, it’s a great point. I’m glad it’s almost becoming more prevalent and more recognized. Because like that’s what we really saw this quarter is that DAI became more adopted within Compound and being a decentralized stablecoin, it gets more into the ethos of DeFi in general. Another side of this is the supply and Sami you’re going to get into this a little bit. I think in December, Coinbase announced their DAI DeFi yield program, which if you look at the day they announced, which was I think December 8th or 9th, at that point on, 75% of all the DAI that was deposited during the quarter was in that time period. So you’re getting a lot more DAI influx into the protocol during this period and in the future we’re probably going to see a little bit more. Getty I know you’ve done some work on the Maker side to get direct DAI injection, so there’s just a lot happening on the supply side as well.
Getty Hill (17:56): Yeah, absolutely. The direct deposit module that we’ve been working on with MakerDAO is currently functional on Aave, and essentially the idea is that ‘Hey, the protocol has a lot of capital and they are sufficiently overcollateralized, why not print some new DAI, put it into circulation, the protocol gets to earn a nice yield and they also can keep a steady hand on interest rates throughout the world’. So they want to make it such that DAI really becomes a prevalent asset and while Circle might not necessarily be out there injecting hundreds of millions of dollars into protocols directly, it certainly shares some influence through some of the big trading desks and lending partners that they have. But MakerDAO really going out of their way to set up direct collaborations with protocols is a really neat way of DeFi supporting DeFi and providing liquidity in a win-win scenario where Compound gets a lot of liquidity and these protocols that they are going to plug into, in this case it could be hundreds of millions of dollars worth of DAI at a non-existent cost of the protocol. MakerDAO also gets to put more of their currency in circulation and keep rates low to make them the preferred borrower asset for individuals who are seeking leverage and it’s just a win-win for everyone. So we’re really excited about bringing that direct deposit module to Compound. That’s something we’ve been working hard on and I think it will provide a significant amount of value.
Dustin Teander (19:31): Yeah I totally agree, I think that’s huge. All right Sami, why don’t you jump in and since we’ve been talking about deposits here, could you give us an overview of the quarterly deposits?
Sami Kassab (19:44): Yeah, so I was saying that deposits is largely a similar story to what we saw in outstanding loans. You see that the effect of the Compound distribution bug in October like Dustin was saying, you’re really starting to notice DAI take over as one of the most, well it is actually the most, borrowed asset. It’s also becoming a more popular asset to deposit and one of the main reasons that Dustin already touched on was due to this Coinbase DeFi yield product integration. So they allowed their users to basically submit their DAI and then gain yield on it. I think this really goes back to what Getty was saying about getting that sticky liquidity into the protocol and helping out both the Compound protocol as well as helping out Coinbase and their users. Because of this integration, you saw DAI flip well… So let’s start with wrapped ether. So wrapped ether has always stayed as the largest deposited asset on Compound and historically the second most popular deposited asset was USDC. What we saw in Q4 was DAI taking over as that second most important or most popular deposited asset. One thing to notice is that wrapped ether makes up such a large percentage of the overall outstanding deposits but in the previous quarters it began to flatten a little bit and there’s probably a lot of reasons for this. One of them could be the whole Ethereum 2.0 staking option. I don’t know if you have any thoughts on this Getty, but another thing to note is the growth in outstanding deposits from Q3 to Q4 was relatively low. It was about 1%. But this chart is in US dollars and so if you were to remove the asset price increase from bitcoin and ether from Q3 to Q4, you’d actually get a negative growth rate for outstanding deposits. So you’re starting to see deposits slow down into the Compound ecosystem but again it could just be because the general macro markets got a little bit heated towards the end of the last year and then you can see there’s a little bit of a sell-off that happened in the end of December and you see deposits start to taper off as well.
Getty Hill (22:43): Yeah, I think in the Q4 period and especially after the Compound error… You know one of the claims to fame for Compound was that it’s the safest protocol and the safest money market that exists: longest running, never had an error, and while 62 was an error to users, either way it’s something that scares people the fact that such a thing could have occurred is frightening without a doubt. While it didn’t hurt anyone, it’s hard to change that perspective. I think that certainly didn’t do the protocol any favors and that’s why we’ve done stuff like having OpenZeppelin come on and be the full-time auditor for the protocol and audit every single governance change that’s made from here on out. In addition, some of the larger participants and voters in the protocol are taking their position as large governance token holders a bit more seriously on the back of that. I think those two things will make a big difference in light of the error that occurred on that COMP distribution bug. As far as other things go, it’s definitely a competitive environment these days when it comes to money markets. Aave definitely has done some clever things to really put themselves as a dominant participant along with MakerDAO and so in addition to Compound competing then with MakerDAO and Aave, at the end of the day Compound’s bread and butter is margin leverage. Users are posting up coins like ETH and go long with cash stables and we’ve seen the likes of FTX and Binance really surge up in their leveraged product offerings, particularly their perpetual products, throughout the world. At the end of the day we are competing for the same market share: users who are interested in getting levered long in their assets. With the domination that’s occurring at FTX and others, those are all competing products that we have to take into account as we continue to build out Compound.
Sami Kassab (24:47): Yeah, I think that’s a really good point that you added. The market is competitive and you’ve really got to rely on adding new features and continuing to build that trust. I think you touched on the OpenZeppelin partnership which is a really important tool especially because Compound is, I would say, correct me if I’m wrong, one of the only protocols or one of the first protocols specifically to go down the route of decentralized governance and decentralized development as well. So they’re pretty much paving the way and setting an example for other protocols, and of course the first mover is going to encounter a few difficulties here and there but it’s all about the community learning from that and adapting and continuing to make improvements.
Getty Hill (25:37): Yeah exactly, that hits the nail right there on the head. We’re the ones who are experiencing a lot of the growing pains being the father of DAO governance and the governor bravo contract and alpha contract now being so widely used in this space. I think MakerDAO is a great example of a protocol that had a lot of its own challenges being the first and foremost DAO that exists in the space. We’re similarly going through a lot of those growing pains when it comes to ‘how do you incentivize participation, improvements in the protocol, security and risk management in the protocol?’ So the likes of OpenZeppelin getting contracted and getting paid millions of dollars to come in here and do great work and have full-time individuals dedicated to the protocol, I think is going to be immensely helpful for participants and for the protocol’s progression in general. And then also the likes of innovating on risk parameters and trying to do risk management. These are all super new interesting things that are certainly not perfect today, but everyone is doing their best to figure it out and ultimately that’s the beauty of DeFi: solving these weird problems that you won’t really find elsewhere. Luckily there’s a lot of incentives in many cases to drive that, and everyone’s willing to have the hard conversations to drive that progress.
Sami Kassab (26:58): Yeah, and I think we can touch on the kind of that incentivisation process as well later on but moving on to depositor interest rates, Q3 saw all-time lows as far as depositor interest rates go and the Q4 interest rates were relatively in line with that as well. So if you take an average of the annualized deposit rate for Q3 it was around 1.8% and that average for Q4 was about 2%, so a slight increase there, but I think these rate stabilizations indicate the product maturity at Compound and there’s a consistent demand for deposits and that demand is constantly met with the borrowed demand as well. I believe that’s really best highlighted with this utilization chart. So one thing to notice is that this utilization ratio has been trending down since around the end of 2020 and that highlights that you have depositor growth continuing to outpace loan growth which is bringing down borrow rates for borrowers, especially with the rapidly maturing stablecoin market. We’ve seen the stablecoin market go exponential in 2021. It has just skyrocketed. So the utilization rate is going to decrease especially as supply starts to meet demand.
Dustin Teander (28:50): Awesome, alright so what we’re looking at here is protocol income and Getty, like going back, you picked my favorite metric as well on the macro chart. To me, this is the culmination of what we’re going after. You can have a bunch of liquidity in the protocol, you can have a bunch of borrowers, but it depends on the take rate or the reserve factor that we’re charging here. This is specifically an interesting thing. So that’s the name of the game for Compound. What matters at the end of the day in my opinion. So I want to spend a decent amount of time on this chart picking apart not just what happened and the results but also looking forward and how we think about the revenue mix in general and directing certain levers we have to pull in a strategic manner in order to optimize revenue growth. First, just to give a quick narrative of what we’re looking at here, we see from Q3 to Q4 about 10% growth but factoring in grants costs, it’s roughly 20% growth in net interest earnings. Going back to my little DAI vs. USDC talk, we are again seeing DAI increase relative to USDC in terms of protocol importance. So it’s driving, in this image, roughly 50% of Compound’s income in Q4 and that is due to a couple of reasons. For example, its reserve factor is higher than USDC so on a per loan basis it’s more income for the protocol. So Getty, let’s talk just a touch about what you think about when you are looking at this kind of data and what levers do you think we should be pulling in order to incentivize different revenue mixes and stuff like that. How do you approach the revenue mixer of Compound and how do you think about it?
Getty Hill (30:50): Yeah, so there’s two big components to revenue today. That is one, the primary driving factor and especially the one in the long run that drives income for the protocol. That is assets to borrow from the protocol and the protocol takes its cut of that reserve factor in the form of these assets. So at the end of the day, we need people to not only supply assets to the protocol but most importantly borrow assets. And that’s what we see here on this protocol income slide, usually seeing DAI denominated things people at the end of the day are most interested in leveraging long in their assets. The market, I think, thankfully for my positions, is very interested in borrowing tons of ETH and going the other direction. So what we see here is a lot of the protocol revenues are going to be denominated in stablecoins. So what that means as far as growth and how we increase these numbers over time, I mean supporting more assets that individuals are interested in using as collateral to borrow then you know more dollars are going to seek that margin long. So whether it be supporting Eth2 assets like Lido’s staked ETH product or other similar staked ETH products or supporting LP tokens, as we look towards the future, those are the things that will be the most pressing developments to drive protocol income up and general TVL up in a very natural fashion. Now, the other half of that coin, which is the not so pretty part, is the COMP incentives that we distribute to the protocol right now. The portion up here, while we made $10 million in Q4, I don’t know the exact number of top my head, but if I had to ballpark it from the research I’ve done, we probably spent 10, 50, 70, or 80 million COMP incentives over that period at the same time and so a lot of the income that we’re earning is in many cases the symptom of a lot of those incentives as we’re incentivizing people to deposit assets in the protocol, we’re also similarly incentivizing them to borrow assets from the protocol and we’re earning a small percentage of the money that we’re paying them to borrow in the form of stablecoins. So the other side of the coin is not so pretty but then again I think we might talk a bit more later in the call about how we use these incentives in an efficient fashion to reduce the protocols annualized spend of COMP incentives, which is now a little bit north of $100 million a year.
Dustin Teander (33:28): Yeah, no I mean it’s definitely a discussion I want to have because I find it pretty interesting as well because it obviously makes sense when you’re a newer protocol and you’re starting up liquidity and stuff like that. But it almost loses its effect because… It’s a big statement, but if the borrowed assets are stablecoins, then we’re putting all this essentially sell pressure on the COMP token and you’re keeping COMP flat, or the COMP emissions, roughly, then you’re losing a per dollar impact for the supply. So the reward rate that I get, let’s say it was 4% flat, no incentives for depositing, as time goes on we’re seeing that juice get to be four and a half as opposed to previously when the token is higher, it’s a higher margin there. If you could see it, we got a slide on COMP emissions there.
Sami Kassab (34:29): COMP emissions. Here.
Dustin Teander (34:33): Yeah, okay I believe it is that. So across this period Q3 to Q4, that’s 30%-ish less in US dollar terms that we’re emitting in COMP tokens but it’s only 5% less or so in actual COMP tokens, so it means we’re getting less bang for the buck more or less.
Getty Hill (34:53): Yeah exactly, the protocol is distributing largely a fixed amount of COMP. The amount of COMP that we’re distributing has really not changed significantly over the period. It occasionally has as we’ve introduced assets or taken assets out of that mix, but it really has remained the same. But at the end of the day a lot of the protocol’s participants, whether they be individuals or trading firms or funds, whatever it may be, a lot of them are denominated in dollars. They report back their P&L in terms of dollars and so they look and see ‘okay well while the COMP tokens are remaining a fixed number of tokens, ultimately what we care about is what’s my bang for my buck’. So if COMP is trading at 800, the protocol is distributing notionally significantly more dollar value than if it’s trading at a hundred dollars. So what’s interesting is we’ve done a lot of research on this and similarly if you were to compare here the amount of participation in the protocol with COMP token incentives paid, it presents a huge decrease from over Q2 to Q4. That is 101 million to 55 million but you’re not seeing the protocol’s participation similarly half in value and that I think is a great representation that ‘hey while there is a lot of participation here that’s interested in these COMP incentives, there’s way more that is actually naturally here’ and that is the stuff that we want to hang on to.
Dustin Teander (36:18): I think that’s a good point. To think about it even more strategically, take it one step further, it’s like ‘all right, we’re kind of paying out COMP rewards across the board on different assets as well, but by going back to our revenue, it was DAI that made up 50%’. We’ve got constant partnerships or integrations at least driving a lot of DAI supplies so it doesn’t make sense to incentivize DAI supply longer or do we direct COMP emissions towards assets that do drive more revenue in order to focus on what matters to the protocol.
Getty Hill (36:53): Yeah absolutely, I think COMP incentives came about from the original idea of the COMP Labs team that kept a good portion of the tokens for themselves and their investors and they said ‘okay, well let’s give the rest of them away to the users of the protocol’. So the original idea was ‘what better way to give them away to the participants in the protocol?’, and this is, mind everyone who is listening who is now familiar with the crazy world of airdrops that exists, this is really before airdrops had really become a thing. Uniswap’s airdrop was something like six or seven months after this occurred in the fall whereas this was back in May. That idea didn’t really exist. It seemed sensible to distribute tokens to participants and so that was really the impetus of this and that’s where it gets a bit more philosophical about ‘do we want to end incentives or do we not?’. It hurts current holders but also the original mandate from the team was ‘hey, these were meant to be distributed to a lot of the users’, and so we’ve seen a lot of interesting debate develop around what’s the best way to allocate COMP tokens most effectively so that it is distributed and decentralized but also such that we don’t needless spending. In my opinion, that is kind of what’s occurring today.
Sami Kassab (38:14): Yeah, I think it’s a good point too because there’s different strategies you could take to… rather than just tapering all COMP rewards to zero you could potentially use that to maybe kickstart new markets, to incentivize people to provide liquidity for that specific market. What are your thoughts? Where do you see this going? Do you see it more strategically being planned and if that’s the case does it just have to be… you guys have to create some type of parameters that have to be set? Because it seems like a really difficult topic to discuss and because it’s very subjective right?
Getty Hill (38:59): Yes, it’s all it’s all quite subjective when I originally went down a rabbit hole being like ‘hey, is there a way now in the back of 62 once we got all the dust settled and we have these new tools, is there a way to materially improve the amount of money that we spend on the protocol and more efficiently spend that cash on the protocol to increase participation and get incentives in a targeted fashion?’, and it really boiled down to being a very subjective process. I think at the end of the day, the most realistic point for what COMP rewards are supposed to be doing is for incentivizing liquidity and markets that do not have liquidity. At the end of the day, the protocol’s mission is to provide a venue for individuals who come and borrow tokens and if we don’t have tokens for these individuals to borrow, then we’re failing our mission there. And that can be quite difficult as we onboard some new markets as the protocol has been doing over the last several months, and turning on COMP incentives for those, starting up some type of regular program where it’s say, ‘hey month one you get this number of rewards or month two you get this number of rewards and it staggers off over a three month scale back down to zero’, and we have a fair amount of evidence now to suggest that a lot of these funds are sticky, and that users will come to the protocol and leave their funds on the protocol. Certainly some will leave but there will be a substantial amount that will provide that liquidity that the protocol needs to function and I think in my mind that’s the direction we’re headed when it comes to COMP incentives. We’re not going to entirely phase them out all together because they certainly play a purpose in strategically improving new markets, but I think at the same time we do need to significantly cut back on the amount of tokens that are entering circulation or at least the style in which they are entering circulation today.
Sami Kassab (40:45): Yeah great point.
Dustin Teander (40:56): What other levers do you think you would be pulling in order to increase revenue? I’m thinking of interest curves. Is there any sort of management of interest curves or anything that you’re looking at there to say like… I know DAI has got a little bit of a different one, but it’s still driving a lot of revenue. Do you look at something like that for USDC or is that even applicable? What do you think?
Getty Hill (41:20): Yeah, it’s a great question. In the grand scheme of the protocol, and since literally back to day one, all the innovation has been underlying the invention of the COMP token, the invention of these interest rate curves, dynamic interest rates, a lot of energy and time and effort has gone into the fundamentals of the protocol and not necessarily the interest rate curves that are chosen, the reserve factors that drive revenue for the protocol, some of the smaller details have fallen to the to the wayside behind the massive improvements in the fundamental protocol that exists. So I think shedding some more time and effort particularly on those issues at the end of the day the protocol is a borrow-lend market and for a borrow-lend market we spend an awful small amount of time improving or researching interest rate curves and setting those up is really important I think to driving the maximum amount of revenue and providing the best rates that we can possibly provide to the market in addition to those reserve factors which ultimately drives revenue to the protocol.
Dustin Teander (42:27): Yeah I know, I totally agree. That’s kind of the next wave of like… ‘Okay, we’ve got something that’s obviously got some market traction, how do we now optimize and fine-tune particularly from a revenue perspective?’. It was definitely one thing that came to my mind as I was putting some of this together. So Sami has kind of prompted me here, so obviously another thing in revenue is liquidations. It has obviously become less of a factor here over Q3 and Q4. You want to give us any insights you got there?
Getty Hill (43:01): Yeah, so I think it was back in Q3, the protocol introduced an improved liquidation system that was brought about I think by the the core team and that was to decrease the liquidation incentive for liquidators and introduce part of that liquidation incentive for the protocol itself so in the case of when a user unfortunately gets liquidated on the protocol, it’s something about an 8% penalty that is put on them, in their portfolio, whatever they are borrowing. About 5% of that goes towards liquidators incentive and essentially that just gives them a 5% improvement on whatever the market weight was at that time to get them to come in and perform that liquidation. I think it’s about 2.8% or so that goes to the protocol, so that’s been an interesting new line of revenue for the protocol. Personally I’m not a huge fan of it, I think it introduces maybe a bit of an interesting incentive for the protocol that doesn’t really need to exist, but at the same time it’s also something that users haven’t really seem to care too much about. So if you want to go the other direction, it’s in theory something you could crank up a fair amount more to drive revenue to the protocol as we’ve seen at MakerDAO liquidation revenues are a substantial amount of income for the protocol over there, especially over the back of what’s occurred over the last month or so, the recent liquidations.
Dustin Teander (44:30): Yep no, it’s definitely a good insight to have. It’s not something that’s long term. You won’t want to be… Like FTX I think has actually done, in terms of the exchange, they have retooled their liquidation engine to attract more traders and for Compound, long term you don’t want to be just liquidating your users because they’re not your users after they get liquidated.
Getty Hill (44:53): Exactly, the protocol is substantially more incentivized to keep the users around and keep them healthy and try to keep them informed of the risk that they’re taking rather than profiting off of their demise.
Dustin Teander (45:08): Right, right. Alright, cool. So that wraps up the financial section we’ve got going on. I want to peel into more of the governance and what’s happening in the governance world as well as looking forward to what the roadmap looks like and things like that. Before we jump in there, I know you gave a little brief overview of what GFX does. Can you paint a picture of who else is in the community, what roles are they playing? We mentioned OpenZeppelin.
Getty Hill (45:41): Yeah, so the Compound community is a fantastic one since it’s one of the older ones in the community so you have both the older crowd who has a ton of experience and seen a lot of the hardships that DeFi has gone through. The folks like Robert and crew and some of the others in the community that go by anon names like Blck and Dennis Bowling and others who have been on discord hanging out long before the COMP token even existed and a lot of those individuals that have seen the protocol go from having tens of millions of dollars to now having six billion dollars in TVL, having liquidations being predominantly performed by individuals and trading firms now being predominantly performed by algorithmic systems that are all entirely on chain. So the landscape has changed significantly which is really cool to see. We’ve got awesome more recent additions to the team and more recently like Tyler Ether in the protocol and we have certainly got some older ones like Arr00 in the protocol who do some great work. So there’s a great collection of individuals who make up the protocol. They are the ones who are day-in-day-out. I’m sure I’m forgetting a few that I should be mentioning, and then there’s also the more recent additions of Don and OpenZeppelin in the mix. They provide a bit more of a strategic steady hand on the protocol and are highly incentivized to participate in a meaningful fashion as well so it’s a nice dichotomy that is constantly developing. It isn’t always super functional and there’s certainly shortfalls regularly, but ultimately it’s a nice and caring community for the protocol and it’s improving that protocol over a long period of time.
Sami Kassab (47:27): Yep, so touching on that, I know that there’s this DAI direct deposit module that is basically a partnership between MakerDAO and Compound and then when you also think about the Coinbase DeFi yield product integration. How do you incentivize people to go out and get partnerships like that? Because those are such important partnerships for the protocol but it seems hard when you don’t have some type of centralized business development team that’s working on forming these partnerships and connections. So how do you really push the community into constantly looking for new opportunities to grow the protocol?
Getty Hill (48:20): The nicest way you can put it is that it’s very challenging. There’s not a lot of incentives so a lot of the participation that does come about in Compound and lots of DeFi is just from people’s genuine interest in the protocols and perhaps from the notoriety that comes with participating in these protocols. Similarly there’s very few things in the world where someone with a keyboard and some smart ideas can affect billions of dollars in capital and that allure I think is quite seductive to a lot of individuals and that’s one of the reasons I certainly got involved in finance and trading, and then in protocol governance. Ultimately, is the fact that anyone can do anything really in this space as long as you just write a good logical argument about it, you can get pretty far just repeatedly performing that strategy. I think in the long run though, unfortunately that doesn’t really work out forever. That’s a nice way to get jump started and lean on the community in the early days but in the long run I think we’re going to see DeFi not morph into what exists in the corporate world but certainly lean on some of the progress that they’ve made. There’s certainly a tried and tested model in corporate governance and just in general how companies function. I think we’ll begin to see some more of that. GFX Labs has our own proposal up right now that were that we are in talks with the community about becoming more of a core contributor from the outside of the protocol, introducing some interesting ideas like KPIs and what not to incentivize our participation to make sure it’s worth our time and similarly the protocol gets what they want from the relationship and everyone can dedicate meaningful resources to improve these massive protocols. But ultimately it’s a lot of interesting challenging conversations that are being had that are slowly driving the whole DAO space forward.
Sami Kassab (50:13): Yeah, thanks for sharing your views on that. Now, we’re kind of running out of time, so maybe we can wrap up with talking about gateway for a bit so I know gateway was announced in March 2021 and it was supposed to be this independent blockchain that would serve as the infrastructure for cross-chain interest rate markets. So you’d have various start ports on different blockchains and they would all report back to the shared liquidity, and there are talks about using a cash stablecoin and I believe Compound Labs came out with an update a few weeks ago. As we all know in crypto time moves 10 times faster and so a lot has developed in the cross chain ecosystem and I’m assuming the team has woken up possibly to the new challenges, especially the the cross chain problems that we’ve been witnessing lately, when you think about like the Thor chain hacks and the wormhole hacks that were going on, but can you give us a little bit of a summary on the new approach that Compound Labs is
planning for? The v3 of Compound?
Getty Hill (51:29): Yeah, absolutely. So they took a lot of the feedback from participants over the period. While gateway was a really credible idea, it was also a very big and hard to execute idea. As to your point we’ve seen with some of these protocols trying to do cross chain things. It is a ton of work, extremely risky, and it spreads out the code base substantially. So I think the team decided to take a more iterative approach to developing gateway and saying ‘okay well, what are the things that we need most right now, what are the good ideas from the protocol?’, because there’s a lot going on at gateway that was fantastic apart from the cross chain implementation, there were a lot of really great features regarding capital efficiency and assets in participation and leverage. One of the really great things I think particularly, retail is going to enjoy the new protocol. It’s substantially simpler. It’s a one-way market in the sense that there is a list of assets that are supported as collateral assets and then there is a single defined asset that users can borrow. So unlike today where you can borrow and supply anything and it’s very much a generalized money market, the protocol decided… Seemingly Labs decided to make a bet and I’m not going to say words for them so this is my impression of the decision making here was ‘if the vast majority of the capital in DeFi and these money markets is interested in going levered long, let’s build a protocol that is optimal for levering long’. So let’s set it up in such a way where ‘hey, there’s a single asset here that you were borrowing and that is dollars or it can be any stablecoin and then governance can add assets on the collateral side and say these are your supporting’. Let’s remove all the logic from borrowing any asset, let’s simplify it down to only supplying assets as collateral, they can’t be rehypothecated, and it makes the code base substantially simpler, substantially more gas efficient, which I think retail is going to appreciate because to use compound today is not inexpensive. In the long run it’s going to make it also that, back to the gateway point, substantially easier to deploy and process and run on other chains. So it’s very much an iterative approach and definitely certainly a tad different from what gateway was proposed as back in last March, but it’s one that Labs feels confident they can deliver and that everyone is very excited to see and use.
Sami Kassab (53:52): Yeah, I found the idea of just one borrow asset pretty interesting, because when you look at the current Compound protocol, most of the borrow demand is for stablecoins. So if you just had one stablecoin that was worth borrowing, it would just significantly reduce your code base and complexity required for going cross-chain. So overall I thought it was a very interesting idea and approach.
Getty Hill (54:20): Yeah, it should be good fun. I’m really looking forward to having it and the numerous tools in it regarding capital efficiency are certainly improved as well. So I think, who wouldn’t want to use essentially cross margin MakerDAO? A big reason why I personally don’t use MakerDAO is the fact that while I own multiple things, to have separate margin accounts for every single asset that I have to maintain is a bit of a pain. I’d much rather use a system that allows me to put that all underneath one roof, in my mind, mitigate some risk that way when it comes to liquidation, and that’s exactly what they’re delivering.
Sami Kassab (55:00): Yeah DeFi is already complex enough, so reducing that complexity and keeping it as simple as possible for retail is just as important I think, like you highlighted.
Getty Hill (55:14): Yeah, and the nice part too is there will be a good opportunity here for v2, the current protocol that exists, to run in tandem for a long time as v3 exists. There’s a lot of room for a generalized money market that exists as we’ve seen and there’s also a lot of room for just a specifically margin-long version of Compound. So I think these things will operate well in tandem. I certainly think v3 will end up being the most popular one because people ultimately love being levered long like myself, but it should be good fun.
Sami Kassab (55:47): Yeah, I’m excited. Dustin, any last questions you can think of before we lose Getty?
Dustin Teander (55:53): You know what? I think we should probably ask the chat or the community if they have any questions. Feel free to drop it in there. We’ll stay on for the next minute or two. Get anything answered.
Getty Hill (56:07): Well it’s been a lot of fun talking about Compound. I could honestly go on endlessly about this protocol and a lot of the other protocols that exist. I think it’s a really interesting time to be in DeFi, in the money market. While there is this kind of quiet bear market occurring in DeFi, there’s a lot of really bright people that are putting their heads down and developing some novel protocols that are going to continue to progress at the forefront of DeFi.
Sami Kassab (56:35): Yeah, I think the Compound community owes you and GFX Labs a big thank you for all the effort that you guys are putting in. I got a chance to read through your proposal, your GFX Labs proposal, as well and I really like the topics that you guys are wanting to focus on for the new changes of Compound. I’m looking forward to seeing GFX Labs play a bigger role as a contributor in the Compound ecosystem.
Getty Hill (57:07): Appreciate that. Thank you very much.
Sami Kassab (57:11): Yeah, well it looks like no questions really. So we’ll go ahead
and just wrap up. Thank you so much again for being here, it was a pleasure speaking with you.
Dustin Teander (57:24): Yep, appreciate it Getty.
Getty Hill (57:27): Cheers, everyone have a nice rest of their Thursday.
