Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Nadav Hollander: Dharma – A protocol for tokenized debt

Episode Date: December 25, 2018

Lending is one core pillar of the economy, enabling one person or company to be entrepreneurial with someone else’s capital. However, in the traditional banking system processes in lending are often opaque and the barrier to entry into this market is high. The emergence of easy to use Decentralized Finance is one of the hallmarks of 2018: To date, DeFi has brought us crypto-collateralized stable coins, decentralized exchanges, tokenized credit default swaps, trustless derivatives, and decentralized margin lending. We’re joined Nadav Hollander, co-founder and president of Dharma. Dharma is a decentralized protocol for credit products which connects debtors with creditors through a transparent mechanism. The protocol itself is agnostic towards the collateral and terms used, however, the team recently introduced a crypto-collateralized margin lending application running on top of the Dharma protocol, Dharma Lever. Topics covered in this episode: Nadav’s background and how he became interested in both blockchain technology and debt The vision behind distributed lending The mechanics of the Dharma protocol The role of underwriters in the Dharma protocol Dharma lever, an application for margin lending on the Dharma protocol Dharma’s business model The future of decentralized finance Episode links: Dharma Dharma Lever Introducing Dharma Lever (article) Request For Blockchain Lending Startups (article) Current and Future Approaches to Unsecured Lending in Dharma Protocol (article) Dharma protocol: Debt & Liquidity for ETH (video) Thank you to our sponsors for their support: Deploy enterprise-ready consortium blockchain networks that scale in just a few clicks. More at aka.ms/epicenter. This episode is hosted by Brian Fabian Crain and Friederike Ernst. Show notes and listening options: epicenter.tv/267

This is Epicenter, Episode 267 with Gaston and Darth, Hollander. This episode of Epicenter is brought you by Microsoft Azure. Configure and deploy a consortium blockchain network in just a few clicks with pre-built configurations and enterprise-grade infrastructure. Spend less time on blockchain scaffolding and more time building your application. To learn more, visit aka.m.m.s. Hi and welcome to Epicenter. My name is Brian Fabian Frienne.

And my name is Friedrich Ernst. We're here today with Nadav Hollander, so we're going to go into the conversation in just a bit. He's the founder of the Dharma Protocol, which is exciting new lending marketplace and lending protocol on Ethereum. Yes, and they've just started building a first application on that protocol, namely Dharma Lever for margin lending. And we'll talk about that and the business model and where they're going to go. Okay, and with that, let's go to the conversation. Nadav. We're here today with Nadav, Hollander and Nadav is the founder of Dharma Protocol.

Dharma is a protocol for lending, loans, and all kinds of debt instruments on Ethereum or on blockchain. So, interesting project. I've actually met Nadav together with Meher in February. We visited him in his apartment, and there were just two people at the time, just starting out. And in the meantime, they've made really incredible. progress and as I was checking out a little bit what has been built on it and what people are doing with it. They've come an enormously long way in a year. So it's definitely one of those

interesting new decentralized finance projects and I'm really excited that you're joining us today, Nadav. Thank you so much, Brian. Appreciate that. So tell us how did you first become interested in, I guess there's blockchain and then there's debt? Like which one of those came first. Yeah, no, definitely blockchain first. Let's talk about that. So I was a student at Stanford studying computer science in 2015. And I had heard about Bitcoin and had like many in the space sort of dismissed it initially as just being kind of like a bit too fantastical or I didn't really see what was particularly like interesting about it or what was particularly useful. about it. And I never really dove that deepened it. But in 2015, I took a class at Stanford

called Bitcoin and Cryptocurrencies that Dan Bonet and Joe Beno taught. And I just completely fell in love with the space. And in particular, learned a lot about Ethereum, which was, you know, at the time, a very new project. And got really, really excited about that. Actually ended up meeting Fred Erism from Coinbase in that class because he came in and gave a talk, Fred Erisom being the co-founder of Coinbase. And eventually, like, that led to me working at Coinbase as an engineer. I actually started off as an intern there. And it's been down the rabbit hole ever since, pretty much.

Cool. So, and then you became interested in debt after that? Yeah, so I think like where where I got into where I started kind of reading a lot into the history of debt and kind of got interested in the topic in general was as a result of starting to think about like the fact that Coinbase as a company sits on a tremendous amount of cryptocurrency that's just like undeployed. It just sits in its coffers, not accruing interest for their users in any way. And I just thought that the idea of having, like, wouldn't it be nice if your Bitcoin when it was sitting in Coinbase was earning you some sort of interest rate like it would in a bank? Like, that's a really powerful concept to have a bank that is effectively kind of globally accessible, that is stateless that anybody within an internet connection can access. And so I started to get like really jazzed about the idea of, you know, interest accruing debt instruments in the context of cryptocurrency. And I was surprised to see that there's really a lack of formal credit markets or credit market infrastructure in the cryptocurrency space at the time.

And so really that kind of led to me both trying to learn a lot about how credit infrastructure has gotten built in the legacy financial system over the years, but also kind of understanding, you know, like what are sort of the early green shoots of. of a formalized credit market in the cryptocurrency world and where it likely will start to go and eventually end up in the future. And so how did you go about learning about the traditional credit market? So was there any exposure you had to this beforehand? In a formal setting, no. So like definitely, you know, I didn't work as like a loan officer at a bank or anything like that,

if that's what you're asking. But I picked up a lot of it from osmosis just in terms of like, basically like going around to a lot of the smartest people that I knew that either worked in the financial industry or worked in some sort of like ancillary financial services and kind of just like picking their brains and kind of shopping around the idea of what would eventually become Dharma to them to kind of see what they thought of it. And, you know, in conversations like that, you learn things like, oh, you know, know, like, you know, subpoint A of this thing that you're coming up with doesn't make sense,

but actually sub point B, like, you know, solves like X, Y, Z problem for us. And just kind of doing a lot of conversations like that until I kind of picked apart more and more of what are the real gaps in the existing sort of credit market infrastructure of the world and what are the problems to which blockchain tech is actually applicable. And something that was very influential on me from, you know, from a standpoint of getting educated on how, you know, like on the history of the world's debt markets was a book called Debt the First 5,000 Years. This is a very, very popular book in the cryptocurrency industry, often for actually different

reasons than my reasons. It's not necessarily like people are excited about it because it affirms certain narratives about how money came to be. But yeah, I think it's like a fantastic extensive treatise on the history of debt and how it's evolved in the modern world. Cool. So this led you to found Dharma. Before we actually deep dive into what the protocol actually does, what does Dharma mean?

I mean the name. Yeah, yeah. So Dharma, at least in the Hindu concept, is essentially, and you know, I always worry that I'm doing a bit of a butchering job of translating it. But as far as I've understood it, is effectively like the concept of like obligatory. or things that you like ought to do. And I thought that in constructing a kind of universal canonical system of debts or obligations, that it was sort of a fitting concept in many senses.

So let's dive into the Dharma Protocol. I guess first, can you, like a very high level describe how does it work and what are the parties that make up the, you know, the main players in the protocol. Right. So I think the best way, the best way to think about what Dharma protocol is is, I like to think of it as like a shared settlement infrastructure for peer to peer lending.

So what does that mean, right? Let's take an analogous example. I'm sure many of your listeners are familiar with the zero X protocol, which has kind of gained a lot of prominence over the past couple years. The zero X protocol can be felt. of as a shared sort of settlement infrastructure that is like a public set of smart contracts on the Ethereum blockchain that takes kind of like a standard order schema that defines like, you know, like party A wants to sell token X to party B takes that like message and then

execute some sort of action on behalf of those two parties. In the case of zero X, the kind of action that it's facilitating is, you know, a trade between two parties. So it takes a, you know, magic signed string of text that says, I, you know, token holder A, want to sell token X to token holder B. And then it like swaps those tokens out from their two accounts once it validates the message is correct, aka it settles the actual trade. Dharma is a similar concept, but for loans rather than trades, right? So it's basically a shared settlement infrastructure for executing lending transactions, where instead of that order essentially saying, I, you know, token holder A want to sell token holder X to token holder B, and instead says,

I, token holder A, want to lend token X to token holder B. And when it executes that transaction, it both pulls the tokens out of token holder A's account and sends them over to tokenholder B, but also kind of like initiates the loan agreement and kind of creates the contracts that can be used to administer the loan agreement over time. And we'll dive a little bit more into that in a second. So that's at a high level kind of the way to think about

what the whole system is, you know, what the goal of the system is, what its mandate is. Now, there are several actors in the Dharma Protocol that are worth defining. So first of all, we have what are known as relayers. So what relayers do in Dharma is extremely similar to what relators do in the zero X protocol, i.e., they essentially host centralized order books that borrowers and lenders can post debt orders and offers onto in order to find other counterparties.

So, you know, I gave the example earlier of like token holder a, wanting to lend like X tokens to token holder B and them kind of coming up with some sort of like magic string that if you give it to the Darmusmark contracts, it would settle the actual loan transaction. That's not particularly useful if token holder A and token holder B don't already know each other, right? They need some sort of way of finding the counterparty. And so that's where relays come in. And, you know, a very naive model, if I am trying to take out a loan, I can go onto a relayer, I can, you know, craft one of these signed messages and basically post it onto the relator's order book so that effectively other cryptocurrency users can go and kind of browse

through that order book, look at different debt orders, and choose which one they want to fill. And the reason why these relators do this is because they earn a fee every time these loans are actually, these loan orders are actually filled. So again, very akin to the concept of relays in the zero X protocol. Now, there's another class of actors in Dharma, which are called underwriters. So what do underwriters do? Underwriters essentially earn a fee for, for playing a few roles. One, they originate borrowers. So they actually like, you know, in some way, shape or form, get the borrower to the doorstep of the relayer, be that programmatically or be that quite literally,

you know, like linking them over there. And then B, they underwrite the risk of the loan, meaning that they make some sort of prediction as to the likelihood of the borrower repaying the given loan, and they cryptographically commit to that likelihood in a message. So they basically say, like, I believe this borrower, has a 0.9% chance of defaulting on this loan and not repaying.

And crucially, the reason why they are held accountable to this prediction is because if that loan is ever actually filled, so if somebody eventually does come around and take the other side of the order, then their prediction gets sort of like immutably recorded into the blockchain so that you could, you know, empirically evaluate over time. time, you know, the degree to which an underwriter's predictions have been accurate. Now, there's some caveats around this, and we'll probably discuss this a little bit later, but that's the basic model. And so essentially, the end-to-end flow, when you have both an underwriter and a relayer, goes kind of like this. So imagine I go on to, you know, like www. Loans.com. And that, like,

the operator of that website is, you know, in the back end acting as an underwriter. So effectively, I just went there. I don't know anything about Darmor Protocol. I don't know anything about crypto assets. I just want to take up some sort of loan. I, you know, like fill out some sort of form there that says, like, I want to borrow, you know, $100,000. What Loans.com is then going to do is they're going to, you know, run their sort of proprietary

algorithms that are going to determine, like, you know, what my creditworthiness is. They are going to then sort of like display some sort of like, you know, a sample sort of like term sheet, right? It's going to say like, okay, we can give you $100,000 at a 3.14 APR interest rate at like X, Y, Z terms. And if I am on board of this, then I'm going to like consent to that. And then they in the back end are going to like, they're going to sign a message that. that essentially is one of these debt orders that I described earlier, but they're going to attach their sort of prediction of my creditworthiness onto there. So you can see that they think that I have a 0.9% chance or whatever of defaulting. Then they are going to go and broadcast

that order onto different relayers order books so that, therefore, a lender can come in and look through those order books and essentially choose whether or not they want to invest in my loan. And so effectively, all in all, you have what is basically a decentralized distributed credit market with different actors who are kind of facilitating the pricing and the counterparty matching of different loan issuances. This is decentralized very much on the side of the creditor and on the relayer. But the underwriter and the debtor, you have to know something about that. right? This can't be a purely on-chain reputation system that you're building up. Yeah, yeah. So I would definitely say, I view kind of this first version of the system that we've built as being like heavy on mechanics, light on reputation, right? And so we basically built like the very kind of like the very foundational mechanics of how this sort of decentralized credit market can work.

But it has a very loose and weak notion of reputation. And what I mean by that is that like, The only sort of reputation metrics that you have on a underwriter for how good they are at predicting defaults is kind of their historical performance. Unfortunately, for a lot of reasons, that historical performance can be fairly easily gained, right? So, for instance, as an example, attack, like I as an underwriter could go and generate a bunch of, like, fake dummy, you know, borrower and lender accounts. and then basically simulate a bunch of loans that are all just me lending to myself and, you know, perfectly predict their accuracy every single time and kind of build a false sense of reputation that were somebody to, you know, kind of naively trust this reputation system, I could use to then, like, defraud them of a certain amount of money. So, so really the the sort of reputation system that

exists in Dharma today is is a weak reputation system. It's really, just kind of a some sort of empirical signal that helps you evaluate whether to trust an underwriter. But, you know, in all of our documentations, we always kind of heavily communicate that, like, we communicate two things. A, underwriters need to be kind of trusted entities. Like, they're not, this is not a protocol in which you should be willing to trust an underwriter on the basis of simply their, you know, their public key and their history. You should also want to know, kind of what sort of, like social capital they have. Like what is their actual like brand, their reputation? Are they like a trusted company? Are they in a regulated sort of jurisdiction? Things like that. And the second thing

that we emphasize is that for a lot of reasons, unsecured lending in Dharma is is kind of an experimental feature set right now. So we really, we really, really strongly discourage people from from doing unsecured loans and meaningful volume right now because kind of as you pointed out, the sort of reputation system that we built into Dharma is very naive at the moment. The other thing that I would emphasize, or the last point that I'd add with respect to underwriters being trusted entities is like the way that I like to think about it is if you were to invest into something like a token sale, you would never just invest into like a blind address, right? You would never just go onto a website that just had like nothing but an

address and, you know, said, like, these are the terms at which we're raising at. You're always going to evaluate that investment on the basis of a lot of social cues and signals and, you know, like, the quality of the white paper and the quality of the team and all this other good stuff. And I view underwriters as being very similar in this regard. Like, they are effectively facilitating some sort of investment in which you are going to need to kind of, like, judge on a basis of a lot of social cues how trustworthy they are. And so, I view them as basically being the fact that it is not a trustless system doesn't mean that it's not valuable in terms of the sort of distribution that it begets.

If you've listened to previous episodes with Marley Gray and Matt Kerner, you know that Microsoft is committed to providing enterprise grade tools and infrastructure for blockchain developers. Well, the Azure blockchain workbench is perfect for organizations building consortium networks. Take the Ethereum proof of authority template, for example. It's ideal for permission networks where consensus participants are known and reputable. Ethereum on Azure has on-chain network governance that leverages Parody's extensible proof-of-authority client. Each consortium member has the power to govern the network or delegate their consensus participants to a trusted operator. And Parody's WebAssembly support allows developers to write smart contracts in familiar languages like C, C++, and Rust.

Azure blockchain workbench was created on the same principles that drive all production services. in Azure, so you know you're relying on secure, redundant infrastructure that can scale. And with built-in services like authenticated APIs, off-chain databases, and secure key management services, you can scaffold your infrastructure in just a few hours. To learn more about Azure blockchain workbench and how Microsoft is advancing blockchain usability and enterprise, check out AKA.m.s slash Epicenter and start building today. We'd like to thank Microsoft Azure for their support of Epicenter. So you mentioned before, right, that the domain focus today,

to build these like mechanics of these decentralized credit system. Now, I would love to understand how does that compare with, you know, the mechanics of the normal credit system? So, you know, in particular, right, you mentioned underwriter. Of course, underwriters do exist, right? In the traditional world, is there an analogous thing to relayers? So, so like on the one hand, you know, are there particular things that you add that are needed a decentralized world that they're not needed in traditional world? or maybe things that you're able to remove and entire parties that you can cut out.

Yeah. So I think it's important to evaluate this from two perspectives. One is the perspective of the initial issuance. And then the second is kind of the context of like secondary markets and how these instruments get traded around after they've been issued. So if you think about it, like the number of hops between, you know, like when you put your dollar into a bank account and, when it eventually gets utilized by some sort of borrower is fairly immense. Like there's kind of like an insane amount of different like intermediaries that kind of like stand in between your dollar in and like the dollar out that gets lent out to somebody.

And the problem is that a lot of these intermediaries are not necessarily operating in like a highly kind of programmatic sort of efficient manner. are they're, you know, very old world financial types. And so like, you know, to give like a super simplified crude example here, let's just say your money is kind of accruing interest in some sort of pension fund. And that pension fund then goes and buys like a CDO or collateralized debt obligation, which contains like, you know, thousands and thousands of different mortgages packaged up into like a fancy instrument. And the administrator of that in of the of the of the collateralized debt obligation is taking some sort of cut on that. And then one layer down from there,

you have like some sort of like some sort of like bank that is actually issuing those mortgages. And they're the ones who sold the mortgages to the CDO administrator. That bank is similarly in some way, shape or form going to be taking a cut from the dollar. that went in there. And then, like, a layer down, you have some sort of, like, originator that actually went and found the, um, the borrower and advertised to them and got them to the doorstep. Um, and that originator similarly taking some sort of pound of flesh and, you know, so on and so forth. And so the point is, is like, um, you know, there, there are, there are intermediaries in the Dharma credit market and there are intermediaries in the traditional financial

system. But our kind of hope and our goal with building Dharma is that at least in in Dharma, those intermediaries will all be kind of entirely programmatically accessible from anywhere in the world. And B, they will be like highly minimized in what they can do. As in like, you know, like the sort of trust that you need to place in them will be will be minimized or the kind of attack surface of what the bad things that they can do can be minimized. And then three, they'll be like almost entirely transparent, as in like any sort of, you know, anything that they may do that is fraudulent or anything that they may do that is good will be kind of auditable on chain. And so, so that's kind of the way that I like to think about the, what the Dharma credit market does for initial issuance in comparison to the existing financial system. Now, there's a whole other topic conversation about how Dharma factors into secondary trading of different debt instruments.

But that's a little bit of an esoteric rabbit hole that I think maybe we can save for another time. That's super interesting. So in principle, the Dharma Protocol is pretty bare bones as it is. So in principle, any loan can be structured as a debtor and underwriter together see fit, as long as the creditor actually buys. it. So in principle, I don't actually need to collateralize my loan at whole or maybe with my reputation. So basically if I

borrow money from Brian, Brian will know that we work in the same office and we will see each other many days and that it will be really awkward if I don't pay back. But what

collaterals do you actually expect to see first? And this is kind of moving into the into the Dharma. So you pioneer this Dharma lever that actually works on top of Dharma. Right, right, right. So yeah, I mean, the way that we see Dharma evolving is that we think that the first use cases for decentralized lending that are really going to take off are those that have like the least sort of external dependencies on the real world. And those at the moment happen to be kind of collateralized on chain by other crypto assets. So they don't really rely on

any notions of off-chain reputation or identity. And the reason why we think that these are the most relevant or useful in the short term is that, as we discussed earlier, like the reputation system for underwriters is fairly undeveloped. It's fairly primitive in many senses. So though, yes, in theory, today, you could go and take out an unsecured loan via Dharma. It's not an extremely scalable sort of system for that right now. Like, you know, it'd be really, it'd be really hard for there to be like, you know, thousands of different underwriters that were all doing unsecured loans. And it would be hard for a creditor then to be able to evaluate which ones to actually trust.

So with that being said, our focus in the short term is instead to facilitate. So essentially, like, evangelized Dharma's usage as the sort of canonical credit market for, were collateralized loans just to like get the actual distribution mechanism kind of embedded into the cryptocurrency ecosystem and then over time to start to layer on kind of more sophisticated notions of reputation that enable things like unsecured loans. And so you ask kind of what sort of things would be used as collateral in the context of these collateralized loans. And basically far and beyond the biggest category here is just other crypto tokens. Right. So like really, you know, a classic example would be like, I hold a bunch of ether and I don't want to sell my ether.

I want to like keep holding onto it or I don't want to incur some sort of tax liability or what have you, but I want some sort of stable liquidity. So what I can do is I can basically post an order onto the Dharma credit market saying I am willing to put up, you know, X amount of ether as collateral and get, you know, like some amount of U.S.D. coin or die or what have you lent to me against that. And the value of that USD coin or die would be less than the total notional value of the collateral that I've put up. And so there's really like the lending use cases that are enabled by this primarily fall into the category of like speculative like borrowing. Like it's basically like, uh,

taking on margin positions. Like if you want to like lever up your position in a certain crypto asset or you want to short a certain crypto asset, like this is this is a great mechanism for doing that. And this is kind of what we see as being like the primary beachhead for decentralized lending today. And this is this is what we are primarily focused on in the moment. And if you like, I'm more than happy to kind of dive into lever in our efforts there. Yeah, no, it would be great. And of course it's pretty on some of that. I mean, I totally agree with you. It makes perfect sense as a use case.

On some level, it is quite funny that, you know, in this hyper volatile space, right? Now you can, like, level up more. Yeah, I mean, I think so. I think that's a wrong way to think about it, though. And the reason why I'd say that is that, so first of all, like, yes, like, you know, like, it's kind of wild that people take on leverage in the cryptocurrency space. and they do it a lot. And, you know, I can't necessarily say that's like, you know, sound investing strategy

to do that in how an extra volatile market as is, but, you know, to each their own. But I think it's also important to realize that, like, margin loans can be used to short assets as well, right? So to basically, like, bet on them going down in value. And that's really, really important for kind of bringing maturity to a market. because if you kind of like if you don't have kind of like short pressure on an asset, then it is more likely to be like extremely volatile. Whereas like if traders have a means of actually doing things like shorting, you know, denticoin or what have you, they will do so.

And that's going to pressure the price down. And that's going to kind of add another price signal into the market that's going to help make things more rational. And so I think it's a bit more nuanced than saying like, you know, This is like providing like, you know, hot dice to a like, like, uh, gambling addicts. It's really, it's more, it's more about like building kind of like the, the basic fixtures of a credit market or a basic fixtures of a sophisticated trading market.

So that prices can become more rational. So, so talking about lever, right? So you can, I, I, will it be possible to kind of borrow any Ethereum asset and put up any Ethereum asset as a collateral or how does it work? Yeah, so really quickly, what is Lever? So Lever is the first underwriter in the Dharma ecosystem that we are building. And so to be clear, anybody can build underwriters in Dharma. We don't gate that in any way, but we in particular, for reasons that we can discuss

perhaps later when we're talking about things like business model, have chosen to build this first underwriter. And the market that this underwriter is focused on is basically like high volume margin loans for anybody in the world that has an internet connection. And so the way I like to think about what lever is as a product is kind of like a shape shift for loans almost, where you can kind of show up on a website. You say, I want to borrow ETH collateralized by, you know, die. We'll give that example again.

Lever will then go out and kind of scan through the Dharma credit market and find like the best sort of offer that fits your parameters. It will then sort of display that offer to you. You will then be able to go and send your crypto assets to some sort of address and then instantly receive your principal kind of sent to you. and all the sort of complexity of actually, you know, like filling the loan on the Darmus smart contracts and interacting with Ethereum nodes and all that sort of stuff is abstracted away from the end user. So you have this like very simple Web 2.0 style product that wraps around the entire experience. And so that's what lever is. So just on lever, let's say I want to, I'm going to put in some dye as collateral and borrow some ether. And then this die gets held in a smart contract?

Yeah, that's correct. Yeah, yeah, yeah. So the collateral is like sort of trustlessly escrowed in a smart contract. The only conditions on which it's like released is either if the user defaults, in which case like the collateral kind of automatically gets liquidated for the principal. And kind of like the remaining principle is sent to the. to the lender and the remaining collateral is sent to the borrower, or in the case of a margin call. So if the price of the underlying collateral drops to a certain point where the loan is no

longer over-collateralized, then there is a sort of liquidation mechanism for making sure that the lender doesn't lose their capital. So all in all, like the system is still kind of, the crypto assets are collateral by smart contracts and not kind of custodied by some sort of arbitrary trusted third party. Okay, so you're talking about two ways of closing the loan. Presumably, as a borrower, I can also just return the loan, right? Right, right, right. So, yeah, the third way in which you can access your collateral is just making kind of repayment in full, in which case your entire kind of collateral deposit is returned to that makes complete sense. So let me go into the liquidation process. So presumably if the value of the

asset I have put up as collateral drops in value or I fail to make interest payments, my asset gets liquidated. So can you take me through how this liquidation actually happens? Yeah, sure. So basically there is a price speed that is kind of, uh, kind of periodically informing the chain of what the price of the two assets is with respect to each other. And effectively, what the smart contract is doing is it is using that price feed to keep track of what the loan to value ratio is of that given loan. The loan to value ratio being essentially the ratio of the value of the principle that's been lent out to the value of the collateral that's underlying it. And once that like the LTV or loan to value ratio crosses a certain threshold, the loan becomes eligible for liquidation. So what happens then is that anybody can come to that smart contract with a amount of the principle that is sufficient to repay the lender and basically purchase the collateral in the smart contract with that principle at the current price.

So essentially, like, you know, to give a quick example, if I, if we're, if we have a zero interest rate loan that is for $100 USD and is collateralized by $150 worth of ether. And for some reason, we have now kind of liquidated this loan. then I as a liquidator can come in and I have like $100 and I got $100 in USD and I can purchase $100 worth of ether from the smart contract in USD or USD coin or whatever. And then the contract is going to take that USD coin, return it to the lender. It's going to take the remaining collateral and send it back to the borrower. And the reason why I as a liquidator would want to do this is kind of twofold. Like either A, I am the underwriter. So, you know, in the case of like lever, you know, like we are initially going to be doing a lot of this because, you know, we are underwriting these loans.

We have an interest in seeing them be kind of service correctly. We have an interest in making sure that lenders aren't going to lose money, et cetera. And we're also earning a fee as an underwriter. So we're being compensated for this. Or there's some sort of optional liquidation. discount on the actual underlying collateral, in which case, like, there is a sort of like arbitrage opportunity here where we can go and, you know, use our $100 to purchase, you know, like the $100 worth of ether at some sort of discount and then immediately sell it at the real

market rate so that we make some sort of delta there. And so that's kind of the basic liquidation mechanism. Because that's how it works at Baker, you know, that you have a, a discount or basically a penalty that gets paid by whoever gets liquidated. So, but here that's not. Yeah, I'd say that we, yeah, the system maker's system is, it has a lot of kind of parameters to it that are sort of optimized for creating a stable coin and like really disincenting defaults to a really big degree because it kind of like chips away the stability of the stable coin every time that happens. And so, so really like, you know, like I think Maker has like a default

penalty of like 13% or something like that. And so like in the case of Dharma, we don't have these same sorts of constraints. Like we don't need to, we don't need to impose these sorts of like very, very sort of draconian rules. And so you don't see those same sorts of mechanisms used in the context of Dharma. So basically the discount that, that you're given for purchasing this loan, Does this depend on the kind of collateral that I put up? Because basically there are many kinds of tokens that if you actually purchase them, you're influencing price quite heavily. So basically they just don't have a good market depth, right?

So basically if it's a large position and you go on to any decentralized exchange or often even centralized exchanges, you move the price a lot just by actually purchasing that amount of token at market. price. Is that factored into that somehow? So yeah, that's an excellent, excellent question. So basically the liquidation discount or the fee that the underwriter earns, which again, they're kind of interchangeable for how you want to compensate a liquidator for coming in, is absolutely like, you know, if, so it's parameterizable, right? It's like part of the order that that gets like broadcasted onto different relays. So you can, it can go from zero to infinity.

right? And or at least in the case of the liquidation discount zero to 100%. And so, and so yeah, what you described is exactly accurate, right? Like, it's like if a big factor that goes into deciding what that discount should be or what that fee should be is the liquidity of the underlying collateral, because in particular, if it's, if the system is based on the liquidation discount, in order for you to like execute that arbitrage, you have to, you have, to cross the spread twice and basically like you know like purchase the collateral and then immediately sell it thereafter and in particularly liquid assets that spread can be very very wide and you can be like losing you know like two four percent or something like that just by crossing

the spread and so for that reason like the yeah you are 100 percent correct in saying that like the liquidity is a big reason why that fee is parameterizable This is super interesting that leads me to my next question. So basically, where do you get your price feed? Because basically, in principle, liquidation then can be an enormously profitable endeavor. So basically, if you get a price feed and you can somehow manipulate the price feed as a liquidator, that gives you a way to game the entire system, no? Yeah, absolutely.

Yeah. And so initially we are going to be like running our own price feeds as a sort of trusted operator. And then, you know, in particular, like I mentioned earlier, we are the underwriter of these loans. And so, you know, I actually think this is an excellent example of where you have sort of like, earlier I spoke about how like underwriters are meant to be kind of like trusted, but like trust minimized in a sense. Like how you, how you like, you want to minimize the amount of trust assumptions you have to make about them. And so I think Dharma lever is an excellent example in this sense because lever is the underwriter of these loans. that it's originating. And for the most aspects of the loan process, you don't need to trust

lever the underwriter, right? Like, you don't need to trust us with respect to making sure that the loan is actually collateralized. You don't need to trust us with respect to making sure that the loan has like some sort of liquidation mechanism that's going to happen to that. All of these things are managed and administered by smart contracts. The only thing you need to trust lever the underwriter with respect to is operating this price feed correctly and accurately. And so, and so yes, there is somewhat of a trust assumption. Yes, there is a way in which lever could defraud you, but you have some sort of auditable track record where you can see, well, okay, the price feed that they've been using for the past, you know, like two years has like

always been accurate within like some sort of confidence interval. And so, and so I view this as like an excellent example of what I think the first underwriters in the Darmine network are going to look like, i.e., like trust minimized actors. They're really, um, they, they have some trust assumptions baked into them, um, but they're, they're like, you know, either trust assumptions that can easily be verified, um, you know, ex post facto or, um, like, uh, yeah, that's, that's basically the gist of what I'm trying to say. So you mentioned, right, that basically, Dharma is a trusted party. to some extent in this context.

And so what does a regulatory profile look like for Dharma levers? Is this going to be accessible to anyone or is it going to be restricted to, you know, accredited investors? Do you guys need some sort of license to do this? Yeah. So I think the, with respect to accessibility, at the moment we are planning on having the product be accessible to both retail and accredited investors, with that being said, like we're not going to go kind of like full, you know, anonymous deck style origination, right?

Like we are going to be KYC customers that come in because we're going to be accepting fees. And if, you know, if you are domiciled in the United States and you're accepting fee revenue from pretty much anybody, you like need to make sure you're KYCing them. And so the system will have at least a little bit of gating in that capacity. Now, you asked an interesting question about, you know, what sort of licenses we may need to, you know, get in order to operate this business in a compliant manner. And the, what's interesting is that though Lever is, like, effectively facilitating the liquidations and is facilitating the origination of these loans, Lever is a non-custodial product, right? Like, we are never, like, actually, like, touching people's principle and, you know, executing actions on their behalf in any way. And so it's not necessarily, like, akin to, say, like, a lending club as a peer-to-peer lending provider where lending club actually, like, stores your dollars and could, you know, in theory, if you, if lending club went out of business, you could not have access to your dollars anymore.

We are more so effectively like a kind of like interface for the underlying Dharma credit market from which you are finding your credit liquidity. And so what that means, and, you know, I can't necessarily dive into this in a super sort of like deep manner on this podcast right now. But it really creates a very nuanced regulatory analysis of like what this product's role is in from a regulatory standpoint. because it's not very accurate to describe it as a money transmitter. It's not very accurate to describe it as a money service business because there's not an actual sort of component of it that involves us touching people's money so much as us just acting as like a nice interface

that bundles up the operations of the underlying Dharma credit market in a way that's easy for users to interact with. That's interesting. So there are regulations that actually touch these types. type of marketplaces, but I think maybe this is not the place and time to go into that. There are a couple of projects that do very similar things that have sprung up in the recent months, such as Compound, D-Y-D-X and Box. How do you see yourself in relation to them?

Where do you think your strengths are? Where do you see your position in the ecosystem? I think Compound and D-Y-D-X are the one or two that. I'll dive into right now because they're they're the ones that I'm most familiar with. So they both are very different in their own respective ways. So if you look at compound, for instance, like compound is a money market and not a pure to pure lending market. And so what that means is that like any lender that puts their money into the like compound

market for a given asset is guaranteed the same rate that all of the other lenders have. and that rate floats over time on the basis of like supply and demand. And that what's great about that is that, you know, with compound like in one click, you can like start earning your interest right away. And that's really, really cool. But the problem with that is that if you have an imbalance in the market where there are, you know, a lot more lenders than there are borrowers, which is exactly what the crypto market looks like right now, then there's this really uncomfortable situation where in order to make

sure that all the lenders are getting the same interest rates, the lenders get like a really, really low interest rate, like sub 1%, and then individual borrowers have to get charged really high interest rates in order to compensate, like, all of those lenders. And so again, it's really just like you have a sort of tradeoff here in the design of compound as a money market where, on one hand, both the lend and borrower side have instant access to what they're trying to do, which is awesome. But on the other hand, that lends itself to have. having often like less attractive interest rates. So Dharma on the other hand is like an order based protocol. So instead of there being like, you know, you as a lender put your like money up

and immediately start earning interest for it, instead you as a lender sign a debt offer and broadcast that offer. You basically say, I'm willing to lend this much at this rate. Kind of like you would post an order onto an exchange's order book. And then if somebody else eventually comes around and says, I'll take you up on that. That rate sounds great to me. Then your assets start earning interest for you. And so, you know, not all lenders in the Darmus system are guaranteed to earn some sort of interest rate. It's only those whose orders get filled. And so what that means is that, you know, again, going back to this sort of like tradeoff analysis, you don't necessarily have like, at least from like the lend side,

an instant access to earning interest rates. But because there's like there isn't a sort of guaranteed interest rate for all parties that are in the system, that means that lenders can get higher interest rates on the Dharma credit market. And similarly, borrowers can get lower interest rates on the Dharma credit market. So it's really kind of like a tradeoff space between those two, those two different projects. Now, with respect to DYDX, DYDX is much more similar. similar to Dharma in this regard and that it's based on like an order protocol. And many different elements of the system are very, very similar to how we designed, how we designed lever.

But I'd say like the biggest difference is that the underlying smart contracts of like the, of Dharma lever are just like the general Dharma credit market, which is built to be like a generic credit market. that can be used for things other than just margin loans. And so we view that as like a sort of strategic differentiator where we can do margin loans and also, you know, like crypto kiddies backed loans and things like that. And use the same sort of underlying infrastructure to have that be served by a unified credit market.

But with that being said, you know, at this point in time where everybody is using decentralized lending, the only thing people are using decentralized lending for is for, you know, speculative loans, I'd say that the projects do look very similar. Interesting. So as the space matures, I assume you'll move into, you'll move away or you'll move to other projects than Dharma level, such as applications that are looking, that look into more illiquid assets. So for instance, say I want to take out a loan on my company or a mortgage on my house that

are not as easily gaugeable as the tokens that I have to put out, have to put out as collateral on a Dharma lever. As far as I know, you guys haven't actually done an ICO. What's your business model going to look like? And how is it going to be different for those very straightforward Dharma lever type applications and for other applications building on top of the Dharma protocol. Yeah, so we definitely, we took an approach where we decided not to do an ICO during the kind of 2017 phrase.

And in particular right now, we're quite happy with that decision because frankly, we have enough sort of regulatory issues and analysis that we need to worry about in our day-to-day operations. I think adding, you know, like the whole world of securities law and, you know, unregistered securities assurances to that is just another reason not to sleep at night. And so the way in which we plan on making money is kind of building out ancillary services and businesses on top of the Dharma credit market, the first one being lever, right? And so like lever is an underwriter in the Dharma credit market. We think that it's going to bring a lot more volume to the Dharma credit market.

We think that's going to incentivize a lot more relays to join and that's going to incentivize a lot of other underwriters to join, and that's going to be great, and that's going to, you know, continue to bring more liquidity to Dharma lever and make us more money and, et cetera. And so, so really, like, the short-term business model, like, in this immediate, like, year is very much focused around, like, lever and making sure, like, you know, we are earning fee revenue through lever. And that fee, mind you, is, like, the underwriter fee that we're taking, which is baked into the protocol.

But in the future, I think that we want to kind of start, like, spinning up, uh, taking the lessons that we learned from building lever and spinning up like other types of underwriters in other sort of related industries. And so you can imagine that looking like us spinning up an underwriter for lending to minors, for instance, or spinning up an underwriter for lending to, you know, various like crypto protocols that need credit liquidity, like for instance, like layered. two scalability protocols. And that's a whole subtopic of its own that I can that can cover. And so there's a lot of really interesting kind of businesses that can be spun up around it. But, but at the moment, our focus is on, you know, like we've built and deployed the underlying

Dharma credit market. We are now building the first underwriter in that credit market. It is an open market. There can be many other underwriters, but we, you know, like our goal is to kind of earn our pound of flesh by earning fees as that underwriter. Do you think at some point there's going to be a role or a necessity to have some kind of token in the Dharma protocol? Yeah, so that's an excellent question. I think that, you know, we spend a ton of time thinking about like token models and which ones do and don't make sense. I think that the two categories. So, first of all, to give the quick answer, the quick answer is like maybe.

I don't know. Like, you know, there's always, there's definitely a possibility. that at some point in the future, it will become clear that, you know, there's a gap in the protocol in X way and that a token would help solve that gap. And if we truly believe that, like, you know, there's a good model that would add value to our ecosystem and also, like, capture value to some degree, then we probably would pursue something like a token. So what are the models that we think potentially make sense in this regard? And I will heavily caveat by saying, like this is highly, highly speculative.

Like we don't have any plans of doing an ICO right now. We don't have any plans of issuing a token or anything like that. Like this is just totally just kind of hand-wavy speculation. So those two models are either like governance tokens or what I would kind of define as like almost like an insurance token. And I'll talk a little bit about what I mean by that in a second. So the first is, you know, like kind of the same. When we're talking about governance tokens in the context of. of a sort of public settlement infrastructure like Dharma, then the best sort of analog to think of

is zero X, right? So, so CRX is the governance token of the like the zero X settlement infrastructure and it can be used or, you know, in theory will be used in the future to govern upgrades to the zero X protocol. Now, the problem with this model is that it's unclear the degree to which it is valuable to govern upgrades to that settlement infrastructure, even if it is entirely broadly adopted by like everyone in the industry, because it's not clear that the switching costs between using, say, zero X protocol or using, you know, a forked version of zero X protocol are that high. If the switching costs were super high, then there would be a very, like, compelling case to be made for why, you know, we need to have some sort of governance mechanism in place.

But if it's just a question of like, you know, like setting my token permissions to like some other set of smart contracts, then it's a it's not entirely clear whether it's necessary for there to be kind of community governance of the actual like settlement infrastructure. But again, what I'd also emphasize is like, you know, it's very possible that I'm wrong here. It's very possible that there is value to these sorts of governance tokens, et cetera. So, so again, all the more reason why we, we have not rushed into building something like this into. our protocol. The second token model that I think is really interesting is what I would define as kind of like an insurance token. And this is a really crude word to describe it. But the best analog that I can think of is kind of MKR in the context of Maker. So MKR is actually a governance token and also an insurance token in this regard. And what I mean by an insurance

token is that like the token acts, the token holders act as the sort of lender of last resort to the to the maker system. So if the maker system, basically in certain cases, the MKR token will be like inflated and sold in order to facilitate some sort of action in the maker system or to like kind of compensate some people if something goes really, really badly wrong. And I view this as being kind of an interesting token model where basically a bunch of

people purchase, like, what is effectively kind of like a share in their belief that the smart contracts of this system are kind of soundly designed and that the economics of the system are soundly designed. And it's almost like they are earning like some sort of premium on an insurance policy in a sense where, you know, in theory, if the system is very, very, like, is functioning very well, the value of the token is going to go up. And if the system actually ends up kind of like defaulting or breaking in some way, then your token is going to get it to get basically inflated to zero. And so I think that these are actually like very interesting token models that will probably start to see a lot more frequently in the wild with Maker

being probably the best proof point for it. So in the context of Dharma, the idea would then be that if I'm an underwriter I can opt in to use some sort of almost premium insurance service by leveraging this token or would this become like a mandatory yeah that's kind of the way I would think about it exactly is like you would have this sort of like opt-in insurance policy

that you could kind of like be looped into again highly speculative I really there's a lot of questions around like how would you, how do you underwrite that risk? How do you like, you know, underwrite the risk of different under, like, how do you permission which underwriters can use that, like, you know, things like that. But I think that those sorts of models are very interesting because, yeah, I just think that it's something that uniquely leverages the abilities of tokens to be like inflated and

deflated programmatically. Cool. Well, we're about coming to the end of our conversation, but it would be just good to hear from you a little bit. What is sort of the long-term vision you have for both Dharma and decentralized finance in general? Like, what impact do you see this having on the world in the next 10 years? So I think what gets me really excited about decentralized finance is that effectively, it really drastically looks. lowers the barrier of entry to building and delivering financial services to people. And so if you think about it like today, if I want to spin up some sort of like financial services

company, let's say I want to spin up a lending company, there's a lot of hoops that I have to jump through, right? I have to not only get all the sort of the regulatory licensure that it requires, that is required to be a lender, but I have to actually go and like raise lending capital. I have to go and get like a credit facility and I have to go like knock on a lot of bankers stores and convince them to lend me a bunch of money so that I can go and then lend that money out myself. And that's like a really bespoke and sort of like wrote process. And if you look at what the internet did to most businesses, that it, what it did to like e-commerce, what it did to kind of like social networks, et cetera, is it really, really drastically lowered the barrier

to entry for building a business and delivering products to people. All you had to do is just go and spin up a website, you know, have a little stripe checkout flow, and all of a sudden I can sell anyone in the world digital products and even deliver it directly to them through the internet. It's a really powerful idea. What I think decentralized finance is going to do is it's going to extend that to the world of financial services, where now all of a sudden it becomes very, very easy to spin up a lending company, or it becomes very easy to spin up like in an arbitrage hedge fund that like arbitrage is

different like markets in the decentralized finance world or it'll become very easy to spin up like a internet based kind of bank where people can earn their own sorts of where people can earn interest rates from anywhere in the world. And the idea here isn't necessarily to say

that like each one of these financial services will be better than their analog counterparts so much as to say that there will just be a lot more competition. like we will have so many more entrance into the financial services market that the sort of quality of them will go up and the prices of them will go down at the same time. And so my vision for what I think decentralized finance is going to be is essentially a world in which the same sort of like hyper-globalization that the Internet did to consumer products and the sort of like, massive increase in quality and decrease in price that you saw come from that when you kind of lowered there's barriers to entry in the context of consumer internet products is going to happen

in the context of financial services as well. And so our hope with Dharma is to kind of build the canonical lending infrastructure that slots into that broader decentralized financial vision and on top of which kind of the financial services of the future are built. cool well that's well articulated and yeah thanks so much for coming on there was i think absolutely amazing project and great how much how quickly you progress so i look forward to seeing what you guys will build in the years to come thanks brian i really appreciate it thanks for having me on and thank you frederica thank you thank you for joining us on this week's episode we release new episodes every week you can find and subscribe to the show on

iTunes, Spotify, YouTube, SoundCloud, or wherever you listen to podcasts. And if you have a Google home or Alexa device, you can tell it to listen to the latest episode of the Epicenter podcast. Go to epicenter.tv slash subscribe for a full list of places where you can watch and listen. And while you're there, be sure to sign up for the newsletter, so you get new episodes in your inbox as they're released. If you want to interact with us, the guest or other podcast listeners, you can follow us on Twitter. And please leave us a review on iTunes. It helps people find the show, and we're always happy to read them. Thanks so much, and we look forward to being back next week.