The Blockchain Debate Podcast

Motion: Algo and fractional stablecoins are flawed (Bennett Tomlin vs. Sam Kazemian)

• Richard Yan, Sam Kazemian, Bennett Tomlin • Episode 29

Share episode

Guests:

Bennett Tomlin (twitter.com/bennetttomlin)
Sam Kazemian (twitter.com/samkazemian)

Host:

Richard Yan (twitter.com/gentso09)


Today’s motion is “Algo and fraction stablecoins are flawed.”

A good stablecoin can sustainably hold its peg, and recover quickly from a premium or discount. This is a basic requirement for stablecoins. 

An obvious design is the bank coin model, where coins are backed 1-to-1 by fiat. But this creates a single point of failure and incurs compliance overhead. 

Hence MakerDAO, which made a smart contract driven stablecoin, and is de-coupled from the banking system. But it requires over-collateralization. 

So new designs popped up and tried to make the next capital-efficient stablecoin to allow under-collateralization, with innovative collateral adjustment mechanisms. We call these algorithmic and fractional stablecoins. 

Historically, most of these coins failed to hold their pegs. Is there a fundamental problem? Or can these challenges be overcome?

The two debaters today include the founder of an algo/fractional stablecoin that has been holding its peg relatively well since launch, that is about half a year, as well as a well-known critic of various stablecoins.

If you’re into crypto and like to hear two sides of the story, be sure to also check out our previous episodes. We’ve featured some of the best known thinkers in the crypto space.

If you would like to debate or want to nominate someone, please DM me at @blockdebate on Twitter.

Please note that nothing in our podcast should be construed as financial advice.

Source of select items discussed in the debate (and supplemental material):



Guest bios:

Bennett Tomlin regularly publishes articles about fraud in the crypto space via his blog. His dayjob is data scientist and fraud investigator in the pharmacy benefits area.

Sam Kazemian is cofounder and CEO of Frax Finance, a stablecoin project that brands itself as the world's first "fractional-algorithmic" stablecoin. Sam also started Everipedia, the first decentralized online encyclopedia on the blockchain.


Algo and Fractional Stablecoins Debate

INTRO

[00:00:00]  Welcome to another episode of The Blockchain Debate Podcast, where consensus is optional, but proof of thought is required. I'm your host, Richard Yan. Today's motion is algo and fraction stablecoins are flawed. 

[00:00:19] Richard: [00:00:19] A good stablecoin can sustainably hold its peg and recover quickly from a premium or discount. This is a basic requirement for stablecoins. An obvious design that achieves this goal is the bank coin model, where coins are backed one-to-one by fiat, but this creates a single point of failure and incurs compliance overhead. Hence MakerDAO, which made a smart contract driven stablecoin and is decoupled from the banking system, but it requires overcollateralization.

[00:00:48] So new designs popped up and try to make the next capital-efficient stablecoin to allow under-collateralization with innovative collateral adjustment mechanisms. We call these algorithmic and fractional stablecoins. Historically, most of these coins failed to hold their pegs. Is there a fundamental problem, or can these challenges be overcome?

[00:01:08] Let's look at the two debaters today. One of them is the founder of an algo fractional stablecoin that has been holding its peg relatively well since launch. That is about half a year. And the other guest is a well-known critic of various stablecoins.

[00:01:23] If you're into crypto and like to hear two sides of the story, be sure to also check out our previous episodes. We feature some of the best thinkers in the crypto space. If you'd like to debate or want to nominate someone, please DM me at @blockdebate on Twitter. Please note that nothing in our podcast should be construed as financial advice. I hope you enjoy listening to this debate. Let's dive right in!

[00:01:42]DEBATE

[00:01:42]   Welcome to the debate. Consensus optional, proof of thought required. I'm your host, Richard Yan. Today's motion, Algo and fraction stablecoins are flawed. To my metaphorical left is Bennett Tomlin arguing for the motion. He agrees that algo and fraction stablecoins are flawed. To my metaphorical right is Sam Kazemian arguing against the motion. He disagrees that algo and fraction stablecoins are flawed. In fact, he thinks they work very well and will continue to do so. Let's quickly go through the bios of our guests. 

[00:02:12] Bennett Tomlin regularly publishes articles about fraud in the crypto space via his blog. His day job is data scientist and fraud investigator in the pharmacy benefits area. 

[00:02:22] Sam Kazemian is co-founder and CEO of FRAX Finance, a stablecoin project that brands itself as the world's first fractional algorithmic stablecoin. Sam also started Everipedia, the first decentralized online encyclopedia on the blockchain. 

[00:02:37] Welcome to the show, Sam. And welcome back, Bennett.

[00:02:41] Sam: [00:02:41] Hey, thanks. It's great to be here.

[00:02:42]Bennett: [00:02:42] Glad to be here. 

[00:02:43]Richard: [00:02:43] So we normally have three rounds: opening statements, host questions, and audience questions. Currently our Twitter poll shows that roughly 58% agree with the motion, and 25% disagree with the motion. After the release of this recording, we'll also have a post-debate poll. Between the two polls, the debater with a bigger change in percentage votes in their favor wins the debate.

[00:03:05] Okay, so let's get started. Bennett, you are in the for position. Go ahead and explain to us why you think algo and fraction stablecoins are flawed. And maybe just to start out, define for us what algo and fraction stablecoins are per your definition. 

[00:03:22] Bennett: [00:03:22] Okay. So for me, there's a lot of overlap in those two categories. The distinction is away from stablecoins that claim to be fully collateralized or overcollateralized. So stablecoins like DAI would fall under the overcollateralized category and stablecoins like USDC and Tether both claim to be in the fully collateralized category. Algorithmic stablecoins exist on kind of a spectrum of collateralization from ones that have no direct collateralizations to ones that can have up to a fuller, nearly full collateralization and they rely on various mechanisms to basically make up the difference - relying on incentives, game theory, and other things to try to keep the value of this thing at $1, whatever, despite having less than $1 of other value that supports it.

[00:04:15] The reason I am so skeptical of fractional and algorithmic based stablecoins is that they represent this alluring piece of alchemy that basically by combining this set of assets together and creating the right set of incentives, you can effectively get something for nothing. You can take 60 cents of value and get a dollar out of it. And I think that that can work for a period as long as you maintain sufficient interest and your incentives are appropriately well aligned, but there will eventually come a point with all of these that the incentives end up misaligned or there's alternative incentives that motivate someone enough to attack it and break the peg.

[00:04:56]Richard: [00:04:56] Okay. Great. So Sam, I'm sure it's not the first time you've heard skepticism towards undercollateralized and algorithmic stablecoins. So go ahead and explain to us why you are against the motion and why you think that these types of stablecoins will actually work well. 

[00:05:13] Sam: [00:05:13] Yeah, definitely. First of all, I'm actually pretty surprised about the poll pre-debate. I was coming in here expecting 90% plus to be against algorithmic stablecoins. Cause they're very popular thing to be against because to be honest, most of them have failed the way that Bennett has actually said. And it's pretty obvious that there is some flaw in those algorithmic stablecoins' design. So first I want to actually just define what I personally...

[00:05:46] Richard: [00:05:46] yeah. So 58% actually agree with the motion. And 25% disagree. So essentially 58% are dunking on algo and fractional stablecoins. 

[00:05:55] Sam: [00:05:55] Yeah. I was actually expecting even more to be honest, so that's promising to be honest. First of all, to be clear during these debates, my definition of algorithmic stablecoin is a stablecoin that has a price peg or target, and it has the same trust and security guarantees as Bitcoin or Ethereum or fully decentralized digital assets. That's the definition that I use for an algorithmic stablecoin. That's why I think it is so important and it captures people's imagination because it is the holy grail answer to all of the "Bitcoin is too volatile" or "Ethereum is too volatile to buy things with and it's a good investment but it's too volatile." 

[00:06:47] I the theme of algorithmic stablecoins embodies the answer to here is this class of digital assets and it should, the ideal one, should have the same security and trust guarantees as Bitcoin, namely, it should be permissionless. It should be decentralized. It should be censorship resistant and thus, in order to afford to actually get those properties, you need to have a very, very robust system of digital assets, either partially or fully backing it and not having wrapped fiat coins or securities tokens or, in the case of the normal fiat coins, Tether and USDC and stuff, they don't obviously qualify as algorithmic stablecoins because it's just bank settlement using blockchain technology. So that's the first thing I just want to clarify and define what I mean by algo stablecoins. 

[00:07:47] And that brings us to the next situation which is what Bennett was talking about, is that actually possible? Is it actually possible to create, a stable unit token, primarily pegged to the dollar because that's basically everyone's unit of account currently. 79% of global trade is denominated in US dollars for lack of a better term. It's the SI units of economic value currently. We can debate if it's going to be for a long time or imminently, it's going to change or something, but for now it is. 

[00:08:26] And so the next question that I think we're going to be exploring during this debate is, is it possible to create something that has security guarantees of a Bitcoin, of Ethereum, fully on-chain with decentralized digital assets and decentralized algorithms and bonding curves and all of these things while following a price peg of probably the U S dollar, right? And the US dollar being a product essentially of the Federal Reserve of the US government, is it actually possible to do something like that? So that's, I just wanted to first define my own stance. And obviously I think that not only is it possible, but I think it's one of, if not the most exciting areas in DeFi and blockchain, that's why I founded FRAX, which, we call ourselves as the first fractional stablecoin because we introduced this stability mechanism to the crypto scene.

[00:09:25]So my own view is that it's possible. It's very difficult. And for reasons that I think we're going to explore here, it is inevitable that there will be a algorithmic stablecoin that is big enough and has enough Lindy effect. And that is sound enough in monetary theory that people will use it to the tune of at least multiple billions of dollars within the next one to three years, which in crypto is a long time, right? And, but it's very imminent and in the, in terms of financial timelines. 

[00:09:58]Richard: [00:09:58] Okay. Great. So let's get to round two. So I'll be directing questions to you both. I'd like to open with a question to Sam actually. So your stance is that algorithmic and under collateralized stablecoins work. But we have seen in the marketplace that quite a few stablecoins have tried to do the fractional algorithmic route and have not been successful.

[00:10:24] And failure mode here is very much in the sense that they we're unable to hold the peg. So FRAX so far seems to be holding up reasonably well. Can you maybe talk a little bit about various approaches that have been tried in the algo under collateralization space and why they have failed to hold up and why FRAX has been successful? 

[00:10:48] Sam: [00:10:48] Yeah. So, first of all, I want to pretty much set a distinction here is that there's probably going to be many, forks of FRAX or of the idea of a fractional algorithmic, stablecoin famously iron recently and everything else. But I just want to make sure that there is a clear distinction here in the sense that even though a lot of these protocols might share the same code or even literally part forks of code that I literally wrote myself, it doesn't mean that the execution is the same, right? 

[00:11:26] For example, we don't say that Ethereum classic being double spent and attacked, is a Testament to Ethereum, the ETH networks security, right? Even though they share almost identical code, the network effects the execution and the community of users is not in any way comparable. So one of the first things I want to say is that it's easy to dunk on failed algorithmic stablecoin projects. It's much more difficult to actually analyze why they failed. It's much easier to say, Hey, look, here's another one. It failed, but it's as persuasive as saying Ethereum classic got double spent so ETH is insecure, which is to say it's not very persuasive. In terms of the recent ones, do you have any particular projects that you'd like me to kind of, go over?

[00:12:20] Richard: [00:12:20] Yeah, for example, you can talk about FEI. Yeah. 

[00:12:24] Sam: [00:12:24] Yeah. Yeah. So FEI is actually one of the most recent and largest attempts at an algorithmic stablecoin. I actually, when they launched, I actually immediately said that their direct incentives, Was very flawed in the sense that it was essentially protocol control of capital controls, right? It would prevent you from exiting and selling your faith depending on if the secondary market price outside of the protocol's controlled value of ETH was under a dollar. This just seems exactly backwards, right? If the point of blockchain technology is to create open global markets with the right incentives, then literally creating capital controlled sub-markets is the exact opposite of aligning incentives, right? And so 

[00:13:15] Richard: [00:13:15] Can you quickly describe for us how FEI works without going into too much detail? Maybe just in broad strokes, how it works. 

[00:13:23] Sam: [00:13:23] Yeah, so, so to FEI's credit, one of the main interesting things about FEI was that they only have ETH as collateral, but they came up with the concept of liquidity collateralization. Essentially what they said is that it doesn't matter so much. If there is a lot of collateral that backs a stablecoin if there's difficult  ways of redeeming it or it's slow. What actually really matters is if there's a lot of collateral in liquidity pools particularly, for example, a Uniswap pool against the stablecoin FEI that can be exchanged at Peck. And so what they designed is they designed a system where people place ETH into a Uniswap liquidity pool against FEI, and there you can always go there to purchase FEI for $1 worth of ETH and then you can also sell FEI for a dollar worth of ETH at any time. And obviously the price of ETH is volatile. So the algorithm or their protocol would balance and reweight the liquidity pool so that the price is always at a dollar. So that, that was the primary idea, right? And then the secondary idea, which is where all the problems started was that.

[00:14:47] Essentially if the peg for FEI broke the way that they would incentivize people not to sell at that time into the liquidity pool would be to say, if you try to sell one FEI, which is worth a dollar, you'll actually be penalized some factor based on how far away the prices from the peg. So for example, if the price is 2 cents under the peg, you would be penalized 4% or 4 cents.

[00:15:18] If you try to actually sell it, you would only get 96 cents on the dollar worth of ETH. Correct me if I'm wrong, if that's a good or accurate representation of it, Bennett, but I think that does it fair justice, right?

[00:15:32]Bennett: [00:15:32] Yeah. FEI had the, their protocol controlled value is basically them using the reweights in the pool to try to force the price, to be at $1 for each FEI by pulling out their liquidity and putting it back in until the price is at a dollar. This was done at regular periods of four hours per reweight, I think.

[00:15:55] And then the direct incentives are what you described where if the price was- it because it was exponential. It got really weird, really quick when the peg actually did break. So like when we saw FEI get down to like below 75 cents, the penalty for that ends up being like 450% or whatever. And so the incentives would have gotten ridiculous if they'd actually been left on, but FEI of course turned the direct incentives off the day it launched.

[00:16:24] Sam: [00:16:24] Yeah. And so actually I have a thread about that, which is one of my first predictions is that direct incentives don't work and they'll turn them off very quickly, which is exactly what they did. And so you can, you can ask me how come I'm very pro algorithmic stablecoin but I immediately said that this was not going to work well, the important thing is that this was just poor design, right?

[00:16:50] Like the fact that saying you can't sell something at the supposed peg that the protocol is supposed to defend is totally paradoxical, right? In fact, I even said this before. I said that it's essentially what FEI the stablecoin was when they had direct incentives was a game of if you're going to get rugged by the direct incentives penalty, right?

[00:17:14] It's almost like a speculative game where it's like, oh, I'm going to hold this for a dollar. But as soon as it trades outside of a dollar on other secondary markets, then I'm going to be penalized. If I try to sell it for a dollar, which AKA means it's not going to be worth a dollar by the protocol.

[00:17:31] So obviously there's flawed designs and obviously a phase backed by a pretty smart team. And they saw that and they iterated. And direct incentives is not on... They're actually, to their big credit, they're actually holding their peg, backed against ETH liquidity. There's obviously a lot of stuff that they need to figure out, like how to expand the supply now that there's no real incentives other than the fact that it's backed by a volatile asset.

[00:18:02] But one of the things to give them major props about is, I think they're currently the only stablecoin that has zero reliance on USDC or any fiat stablecoin, which is actually a very big deal. So I think it's it's important to shine that light, and congratulate them on that part of what they've achieved.

[00:18:25]Richard: [00:18:25] Okay. Bennett, do you have anything to add in terms of the performance of FEI? I know that you wrote a very long piece analyzing their mechanism. And then we'll go back to Sam and have him talk about why he thinks FRAX is a better option, a better design, or perhaps has gone through better execution and delivering a stablecoin of this genre.

[00:18:47]Bennett: [00:18:47] FEI's major innovation was this idea of trying to basically manipulate the liquidity in the pool to maintain the price. However, every time you have to withdraw your liquidity and make those changes, you are diminishing your liquidity ratio, meaning the amount of collateral they have for each of their coins in circulation.

[00:19:06] And they even discussed this in their white paper, basically saying that without the direct incentives, reweights end up being net negative for protocol controlled value. They turned the direct incentives off because they weren't working and we're probably never going to work. But now what's resulted is them having the only stability mechanism, being three weights, which are continually draining against the protocol controlled value.

[00:19:31] So effectively the position they've put themselves in is that they they're basically making a bet that the dollar denominated value of their ether collateral will increase in terms of dollars, quicker than the cost of them having to reweight to help maintain the peg for FEI. And so well, they have now been able to get back to near peg and maintain that much better since turning back on reweights and turning off direct incentives. I think long-term, they've put themselves now in a position where they are fundamentally unstable that it's fundamentally unmaintainable due to the way their structure is set up. 

[00:20:11]Richard: [00:20:11] Okay. Can you quickly talk about the re weighting mechanism? How is that used to enforce the peg? 

[00:20:16]Bennett: [00:20:16] Yeah. So, whenever FEI was minted, you a corresponding amount was minted directly into the pool, under control of the protocol. So you'd get one FEI and the protocol would get one FEI and it would match with the ether. You put up into the pool to keep the pool at one-to-one. Whenever that would shift, because someone else is depositing FEI or ETH into the pool, making it so FEI is no longer at a dollar when it gets past a certain point and that's the appropriate period, the FEI protocol withdraws all the liquidity and then puts back in the right amount.

[00:20:50] So that by like the calculation on Uniswap, FEI is now back worth a dollar, however doing that has a certain cost. And that cost is what's continually draining the PCV as they do each of those. 

[00:21:02] Richard: [00:21:02] And what does PCV stand for? 

[00:21:04] Bennett: [00:21:04] Protocol controlled value. That's their term for the ether in the FEI that they control inside their Uniswap pool? 

[00:21:11] Richard: [00:21:11] Okay. And you're basically saying that the way FEI can continue to hold this peg is predicated upon an ever increasing Ethereum price? 

[00:21:20] Bennett: [00:21:20] Yeah. Right now, until they come up with some other way to incentivize stability that doesn't involve them basic paying attacks each time they have to do it.

[00:21:32] Sam: [00:21:32] I mean, I would agree with that. I think he's, I think Bennett is totally right about that. So I, it's hard to defend a mechanism that I wouldn't personally build. So I think the real fun will be when we discuss FRAX essentially. 

[00:21:48] Richard: [00:21:48] Right, right. And sorry, just maybe one last comment on FEI and the intuition behind that idea that their peg is predicated upon upon Ethereum's ever increasing price is because if the ETH price were to drop, then the reweighting process would somehow set off a vicious cycle for the FEI price? 

[00:22:11] Bennett: [00:22:11] not necessarily because you're not looking at like cascading liquidations or something like you could in other collateralized stablecoins, it's just predicated on the fact that there is a cost associated with bringing fade back to peg. And so if the value of their collateral is not increasing quicker than they're paying that cost, eventually you run out of collateral with which to pay the cost.

[00:22:32]Richard: [00:22:32] Okay. Awesome. Let's go back to Sam. And Sam, let's have you talk about FRAX. How does it work? Why is it superior?

[00:22:39]Sam: [00:22:39] Yeah. So the thing with FRAX and the notion that we started with was that before we actually came up with it most people are probably familiar with the stablecoin triangle so to speak, it was introduced by Haseeb Qureshi at Dragonfly a few years back where it had this triangle where at the top on one side was fiat coins, like Tether.

[00:23:03] The other side was the over collateralized stuff like DAI. And then the bottom was algorithmic stablecoin back then basis, which never launched their first product and gave their money back from their original raise. But it was one of those projects that tried to be 0% collateralized, right?

[00:23:25] The first ideas of algorithmic stablecoins. We started with FRAX and we said, well, what's in the middle. For on-chain, decentralized stablecoins, there's MakerDAO all the way over at 150% collateral ratio and this kind of leveraged model, and then you have Basis and these seigniorage shares ideas where there's 0% backing.

[00:23:51] And that just seemed to us wrong. In fact, one of the main things that we thought is that the reason they haven't proliferated but, you know, MakerDAO is growing and, and stuff is that there's something fundamentally wrong here. It's almost like it's a bank without a balance sheet. How exactly is something supposed to actually issue currency or liabilities against itself, if it has no balance sheet to do anything with, right? So we asked the question what's in the middle between 150% collateralized and 0% collateralized. And we came up with FRAX and we actually named it FRAX because it stands for fractional algorithmic stablecoin, because the idea behind FRAX is that there's a market set collateral ratio. 

[00:24:38] And what that means is when FRAX first started, it was actually very, very boring. It was a hundred percent backed. And in fact, it was backed by USDC, another stablecoin. So it almost seemed a little bit pointless where you put in one USDC and you get out one track stablecoin. Now, as the FRAX supply expanded, as people wanted to farm FRAX or there were new integrations and people were a little bit more comfortable holding FRAX in their wallet instead of just immediately switching it to other stablecoins the price of FRAX which slightly go above a dollar, right?

[00:25:15] People would pay a slight premium for it. So there'd be this arbitrage opportunity to mint FRAX from the contract and arbitrage it, right? And the protocol used this slightly higher price as a signal to lower the collateral ratio. And so what that meant is it would go to 99% collateral ratio or 98% or 97%.

[00:25:39] And what happens here is that in order to mint one FRAX stablecoin, someone would only have to put in 97% USDC and then burn $3 or sorry, 3 cents of FRAX Shares, which is the governance token. And this is how the governance token captures signage right unbacked value of the stablecoin and the... 

[00:26:05] Richard: [00:26:05] Hold on a second, who is doing the burn? Is the person purchasing FRAX at this moment contributing 97 cents of USDC and then 3 cents worth of FRAX and then the FRAX gets burned by the contract. Is that how it works?

[00:26:18] Sam: [00:26:18] FRAX Shares. Yes. So it's important to note here that there is never a place where value is leaking out of the protocol. The value of signage is always captured by the fracture distribution, right? So if someone wants to mint FRAX, they have to put in a dollar worth of value. If someone wants to redeem FRAX they actually get out a dollar worth of value at all times the ratio of what, to, what collateral to FRAX Shares is based on this market set collateral ratio, which changes based on whether FRAX is expanding in supply or retracting in supply.

[00:26:54] Richard: [00:26:54] So hold on a second. So in that case, someone puts in 97 cents of USDC and then 3 cents worth of FRAX and subsequently... 

[00:27:01]Sam: [00:27:01] FRAX Shares, the governance token.

[00:27:03] Richard: [00:27:03] Sorry, FRAX Shares, and it receives $1 worth of FRAX dollar, right? And then the 3 cents worth of FRAX Shares get burned by the protocol, right? That's the result of this. So now what if that same person wants to redeem- wants to sell back that $1, one FRAX dollar. What does he, or she get? 

[00:27:22] Sam: [00:27:22] They would get 97 cents of USDC the thing they just put back in and then they would get 3 cents worth of FRAX Shares, again, minted from the protocol.

[00:27:33] Richard: [00:27:33] see. Okay. Interesting. Now carry on, please. 

[00:27:38] Sam: [00:27:38] Yeah. And so the main concept behind FRAX is that it uses the market as an indicator of how much there needs to be collateral in the protocol and how much the actual governance token, which is another way of saying the seigniorage capturing token or the value of cruel token of the system can stabilize part of the supply of the system, right? And so that's one of the first things that we introduced to the crypto space is that there is something in the middle between MakerDAO's 150% collateralization ratio and other projects is 0% attempts like basis, like ESD's version one and all of these other fully algorithmic AKA zero collateral designs, which none of them have worked and they've contributed to falsely having people's opinion you know, misinformed about the topic, right? 

[00:28:39] And so one of the things that is important with FRAX is we've never broken our peg as defined by, for example, it being outside of the price band by even 12 hours or six hours, I think even is what our research has shown. I think we've never broken the peg a single time since launching in December, 2020. So we're pretty proud of that.

[00:29:05] Richard: [00:29:05] Okay. And what is the reason why FRAX has been holding up well despite the failures of these predecessors? 

[00:29:13]Sam: [00:29:13] I think maybe Bennett should answer that one as the [inaudible] 

[00:29:17]Richard: [00:29:17] No. What I'm saying is so far, FRAX has been performing quite well compared to its competition, right? But based on the mechanism that you've described, I'm not sure from my end, I can immediately point to a reason as to why it has been so successful. Empirically it has been doing so. Just curious. I guess this could be answered by both sides. What do you think is the reason why it's been holding up so well, and if Bennett wants to go first, that's fine too. 

[00:29:45] Sam: [00:29:45] Sure. Sure. I think the person arguing against, out of respect, should go first.

[00:29:51] Richard: [00:29:51] Oh yeah, sure. Go ahead, Bennett.

[00:29:52]Bennett: [00:29:52] Well, I mean I don't know what the current collateralization ratio of FRAX is, but...

[00:29:58] Sam: [00:29:58] It's 84. 

[00:30:02] Bennett: [00:30:02] OK. For any of these algorithmic stablecoins, especially ones with the seigniorage model like this, and FRAX is a little bit unique, because it did have the partial collateral, is so there is a certain amount of collateral you can confidently capture by redeeming your FRAX dollars. You can get back 84 cents worth of USDC. And so the remaining question then becomes, where does the rest of the value comes from? In this case, so long as FRAX or similar algorithmic stablecoin protocol is attracting interest is growing and the FRAX Shares token is actually values accruing to it.

[00:30:41] Then you can capture some of the rest of that value by effectively collateralizing with something else in this case with the Frack Share, not directly because it's burned and then minted, but effectively the rest of the values coming from the fact share. And so as long as you are able to continue to accrue value to there, the peg holes. And so in FRAX's case it's been growing pretty consistently since December of 2020 and has been largely collateralized for the entire period. So they've been able to build up a certain amount of trust in their protocol overall. And more specifically in the FRAX Shares token as a token of value accrual. So because of that, they've been able to maintain the FRAX dollar at or around a dollar.

[00:31:28] Richard: [00:31:28] okay 

[00:31:29] Sam: [00:31:29] Yeah. And actually I would say that that's quite a fair characterization and what I want to add by the way is like, there is legitimate Integrations and demand for currency, right? And so for example, building this Lindy effect, getting it accepted in lending and borrow markets, getting it accepted in other projects as treasury assets, getting it incentivized exogenously, which means outside of our own community and outside of our own governance token, and being accepted as something of value, right?Something that's, tightly at a dollar. That actually is genuine, exogenous growth, right? It's outside the protocol's demand creating mechanisms, right? It's not circular, right? 

[00:32:19] And so one of the main things that's really important with algorithmic stablecoin design is getting out of this circular vicious cycle where there's this mirror and enticing virtuous cycle that goes exponentially upward. That kind of captures all of people's kind of gambling and degen lifestyles and, and trying to make a quick buck and then everyone leaves on the way down. It's very difficult to actually bootstrap real currency, right? Because exogenous integration and actually having it being valued as money in and of itself is extremely difficult.

[00:32:59] So for example, some FRAX statistics is that we're the third largest curve pool right now with about 300 million TVL. That is obviously not incentivized with fractures. It's exogenously incentivized over 95, 6% of the incentives are exogenous, the second stat that I like to talk about is it is a top Uniswap pair.

[00:33:25] And it has over a hundred million total liquidity on Uniswap, the other things that it actually has. It has integrations and cream and lending markets and iron bank. It has a wire, vaults. It has all of these things that are difficult to build that take time that take trust in a dispersed and wide community.

[00:33:49] This is how money is actually created and actually get social legitimacy. And so the same way, for example, that the us dollar is is used for oil transactions, right? A lot of people might point at, "Oh, well, the US originally started that out of coercion because they won World War II," and everything.

[00:34:10] But the reality is that people use it, right? And so all of these things that the US dollar is used for builds it's network effect, that it allows people to actually reliably settle transactions in it outside of just the recursive nature of just the Fed actually rising and lowering interest rates, but with everything that's being used for the US dollar and the Fed's market operations, you get a very good ecosystem, right?

[00:34:39] And so this is actually the same view, but just a very micro small scale compared to obviously the Fed, we try to, build FRAX out, right? We try to build FRAX in the sense that we need real usage, real demand, actual integrations that don't have to do with the hyperbolic ups and downs of reflexivity.

[00:35:03] And if we can get there the actual collateral ratio, which is market set, right, it looks at the slight demands and retractions of FRAX stablecoin to adjust itself that collateral ratio will adjust accordingly, right? And so now it's at 84% as I as I say these words, right? And hopefully it'll get into the seventies.

[00:35:25] If the market is ready and they actually believe in practice usage, or it might actually increase to 90% or above if there's a lack of confidence or there's a period of, retraction in FRAX supplies. So we take a really, really sustainable long-term and very economically sound approach and one thing I like to say is so far numbers don't lie. We haven't broken the peg, obviously there's a lot of counter examples of projects not holding the peg, but every single day that FRAX actually stays and thrives, I think is another day of data and counter example to the normative narrative of just algorithmic stablecoins are Ponzi or they are a scam. 

[00:36:14]Richard: [00:36:14] Okay. Great. So I think by certain measures, we are still in the bull market, especially for defy. So I think that the continuing faith in defy project and new kinds of stablecoins or the new governance tokens for stablecoins, such as FRAX Shares, definitely contribute to people's confidence. And therefore the ability to, for these coins, the whole peg I'd love to get a perspective from Bennett in terms of what happens when things don't go so well, do things go south? Could there be some kind of vicious cycle that subsequently causes the peg break and, Bennett, feel free to also allude to the black Swan event for a MakerDAO that took place in the first quarter of 2020. This is something I meant to be asking you as well. 

[00:36:59]Bennett: [00:36:59] Sure. So just broadly speaking Often when you really determine whether or not these systems work is when they're first placed under strain and just throughout all of financial history, currency, pegs in general are an extraordinarily difficult thing to maintain. That's how Soros made his fortune, right?

[00:37:18] Was by figuring out how to break a peg and You mentioned specifically the black Thursday event in March of 2020 for MakerDAO. And that's a really interesting example because basically what ended up happening was you had the price of ether itself crashing, which put a whole bunch of MakerDAO vaults in liquidation territory.

[00:37:38] Bunch of people were using automated contracts that didn't take into account the changing gas prices. And so we're unable to get transactions through to either refill or claim their collateral out of their vaults. And so a bunch of the bolts went up for options, went up for auction, sorry. And then because the chain was clogged fees were high.

[00:37:59] The, there was people who broke from like their normal auction contracts and started manually submitting bids for super cheap for some of the ether that was up for auction. And they ended up capturing several million dollars of it. And for MakerDAO, they were able to recover the peg eventually, after that, and I think it also illustrated to them.

[00:38:21] One of the, after that point, they made a pretty big change to their system. And one that corrected some of the issues I'd had with it previously in the initial version of DAI that was first released when there was cases where you could not liquidate all the vaults could not get sufficient value to free collateralize the DAI, your claim on the collateral was decreased.

[00:38:43] And what they did during the black Thursday event was instead, they started auctioning off newly minted Maker tokens to effectively get back some DAI and recollateralize the protocol. This is now also how the multi collateral DAI system works. The asset of last resort is Maker in a case where they need to recollateralize.

[00:39:04] They sell Maker until they're recollateralized incentive wise. That makes more sense because the people you went to punish in effect are the ones setting the risk parameters and stuff who obviously failed to appropriately set the risk parameters. But I think the other lesson that can be learned from the black Thursday event is that in adversarial conditions, in difficult conditions like that assumptions that are made about ease of minting, redeeming refilling vaults, liquidating vaults, claiming certain assets can be challenged due to the difficulties inherent in using the network itself during those periods.

[00:39:43] And the costs embedded in that. And so in certain cases, in things like DAI to calculate like your real risks in those difficult conditions, you need to take into account Not just the collateral you have to put up, but the amount you're willing to spend to preserve that collateral. So like how much are you willing to pay on the transaction fees to get more ether into a vault to avoid getting liquidated or something like that.

[00:40:07] And so the overall problem space gets exponentially larger as you start to factor in those kind of attacks. Again, MakerDAO is a little bit of a unique one because it's over collateralized to start other algorithmic stablecoins. You'll often see some kind of event that triggers a loss in confidence, some kind of execution bug, or just a large entity fighting the peg briefly and even with the first algorithmic stablecoins, like new bits, the Larimer project back in 2015 or 2016 that ever was once you have lost that confidence for the first time and the pegs broken, because someone's trying to break it or whatever market dynamics cause it to break. It becomes increasingly hard to recover the peg because when you are dealing with an asset that is not fully backed by the value it's meant to represent the entire name of the game is effectively maintaining 

[00:41:03] the confidence that there is still some way to get that value out of it, right? And so whenever that's lost, the coins have issues. And this is why I think often the models that have had the most success in maintaining their peg have been those that keep significant collateral and why all of the no collateral ones I can think of have had issues with ever maintaining that peg. The more collateral you have, the easier it is to maintain the confidence that there's value backing it. 

[00:41:34] Richard: [00:41:34] Okay 

[00:41:34] Sam: [00:41:34] Well, the main thing I want to add here is, so I totally agree. The reason we founded FRAX was because you won't hear me defense zero collateral algorithmic stablecoins, because they don't make coherent sense. It's like a bank with no balance sheet, right? And I can't actually defend those designs because we wouldn't have built FRAX that way, right? And in fact, FRAX is supposed to be a fractional stablecoin, right? And the idea behind that is you only should actually have a algorithmic supply of the stablecoin that the market will let you have. Like Bennett was saying, is the market is only confident up until a certain amount, right?

[00:42:22] And they essentially vote literally with their dollars by either selling or trying to redeem the dollars or continuing to hold them, right? And so you can use the price of the stablecoin as essentially a transparent democratic vote, you know, to gauge market confidence. If more people are selling it.

[00:42:44] So the price is slightly below a dollar, so it was 99 cents. Then that should be an indicator of, Hey, increase. The protocol's reserves, you know, 84% is not enough increase it to 85% or 86% or whatever until the price is right back to a dollar. If the price is a dollar and 1 cent in terms of FRAX, then that's the signal that the market is buying FRAX at a slight premium on the open market. So people are going to be minting some FRAX soon, instead of buying it at a premium, that's a signal that people want this stablecoin, even add to this collateral ratio. So perhaps lower it by one, right?

[00:43:31] Maybe take it down to 83% or 82%, as long as the price is continually, slightly above the peg and stop when it's back to a dollar, right? And this kind of market set collateral ratio works for two reasons. One is that, as I said, originally, everyone can come back and get a dollar worth of value for their FRAX, right? So that's clearly something that needs to happen. I actually agree with Bennett about that, right? The difference is how much reserves are there that can be paid proportionally, right? So if someone comes to get a dollar worth for their FRAX right now, the protocol can only give them back 0.8 for USB-C and 16 cents worth of FRAX Shares, right? That equals a dollar of value, but it is not fully in the collateral asset, right? But this system works because it has a market signal from literally people who hold FRAX. And so it's it's pretty important. Yeah. It can be implemented poorly, like iron, you can adjust and change the collateral ratio so that it decreases much faster.

[00:44:43]And which is what they did. And also, it can basically not support the amount of algorithmic supply and during a bad time like Bennett says, during a stress test or something, it can collapse, right? And it should be worth pointing out here that FRAX has withstood many now, and it's six plus months of existence, many negative 30 negative 40 percent declines in both its share token value in the overall crypto market cap value.

[00:45:21] In fact in, in may when the entire global crypto market capitalization shed $1 trillion worth of market cap in a few days, I think everyone remembers that bloody week that three to four day time span where even Terra briefly was trading at 96 cents or something for a day or two, FRAX never went lower than 99 cents or 98.90 cents for any amount of time, more than maybe an hour. And even when FRAX actually was being redeemed for a lot of supply, everyone's still got a dollar worth of value and the peg perfectly returned to a dollar almost immediately. 

[00:46:10] So these kinds of events, I actually agree with Bennett show what what a protocol is made of, right? And that increases Lindy effect, that increases confidence, and it's almost like, we all know that we would like more transparency into Tethers finances and things like that, but it's almost like the Tether effect where it has never significantly collapsed in all of its life. So it becomes almost like a coordination mechanism where it's difficult for people to think that it's going to fail this time, right? And so that's how you build social confidence. So that's the approach that we take and so far, we've only been increasing in terms of confidence and never have broken the peg. 

[00:46:57]Richard: [00:46:57] Okay. So Bennett, you're a skeptic of algo partial stablecoins. So that means you believe there will be a certain scenario with non-trivial probability that these stablecoins will break their peg for an extended period of time. Can you paint a scenario in which this happens?

[00:47:17] What would need to take place in the macro scale? What are some exogenous events that need to happen in order for this situation to play out? And that would be a real threat to holding the pack for these algo partial, stablecoins. And then I would love to hear Sam as to when that sort of thing happens. What would be your rescue mechanism? 

[00:47:38]Bennett: [00:47:38] So I can think of a few different things that could potentially challenge a lot of the algorithmic stablecoins. The first one is for a seigniorage system like FRAX if there's some kind of implementation bug or something else that causes a rapid decrease in the value of the seigniorage share, FRAX Share. Well, you're also seeing people trying to redeem, you can start to feed a cycle, right? Because as people try to redeem the FRAX dollars, a corresponding amount of FRAX Shares are minted to make up the remainder of the value beneath the collateral ratio, right? And so in a situation with a FRAX Share price has already dropping and you're minting more fractures into the market.

[00:48:28] You can start to see an amplification mechanism where that continues to put a downward pressure on the value of the share, making it. So you have to progressively mint more and more as more people try to redeem. Other than that, many...

[00:48:43] Richard: [00:48:43] Sorry, what would be a bug to cause FRAX Shares to drop in value to begin with? My understanding is that these are ERC-20's and the smart contracts are fairly standard. I assume they've been audited and they can't be audited, and so forth. 

[00:48:57]Bennett: [00:48:57] It could be as simple as someone else trying to attack it for whatever reason, right? That's how Soros made his money was in attacking the peg while holding other values out against it, right? And you see this sometimes with the there's more DeFi protocols and stuff now, allowing for like effectively credit, default swaps and insurance against the value of certain stablecoins and stuff.

[00:49:18] So you can effectively take out bets that certain things will crash and then profit from making them crash. And this mitigated somewhat by the way, a lot of those work. in that there's often a governance process for approval of the claims and stuff, or just as you integrate broader and broader into other protocols, you can at times be exposed to bugs in those protocols, right?

[00:49:44] If suddenly a whole bunch of them becoming accessible or. Interactions like that could, or just generally a loss of faith of large whale decides to sell a whole bunch and you see a large decrease... 

[00:49:56] Richard: [00:49:56] I guess their reserve gets hacked or something. stolen. 

[00:49:59] Bennett: [00:49:59] Yeah. A whole bunch are in vaults and there's a bug discovered there and a bunch get pulled and then immediately dumped onto the market. There's a whole bunch of different like implementation risks that exists there. And so then you start feeding the cycle. When you have to start making more of those in order to preserve the value that is there. The other problem is largely related to the bank stablecoins that we were talking about before that many of these used to collateralize them FRAX actually did something I really appreciated when I was reading the white paper for this is that they do not assume that one USDC is going to be worth $1.

[00:50:35] It might not necessarily be. So, Tether has in the past dropped down into the eighties before on exchanges and stuff because of loss of confidence in its reserves and stuff. And so it's theoretically possible that a similar thing could happen to USDC. This could be the result of changing regulatory regimes, new information, whatever, but that's why it's important to make sure that your protocols designed without having the assumption that a certain token is going that certain other token is going to be worth a dollar.

[00:51:09]DAI particularly made a couple of really poor choices in this regard, in that they incentivized to bringing on more and more USDC and built their protocol with the assumption that it would always be worth a dollar like the USDC-A vaults and on MakerDAO cannot be liquidated. And so because of that, USDC now represents like this existential risk to MakerDAO and DAI itself, just because it represents 60% of the collateral and the bolts it's in cannot be liquidated.

[00:51:36] That value cannot be recaptured. In other cases, one of the common mistakes you'll see is protocols will assume that the bank backed stablecoins are fungible, but any USDC token is equal to any other USDC token. This becomes a problem because again, the centralized entity circle, Coinbase, whatever can blacklist and freeze the coins, making them no longer fungible.

[00:51:59] So you believe the tokens in your specific vault or whatever, have a certain value because that's, what's being reported by the Oracle. But actually some of the tokens in there have been frozen or blacklist and their effective value is now zero, but there's often in many of the protocols, no way for the protocol to actually become aware of that fact.

[00:52:17]The other thing you'll often see is just an expectation that certain feedback, stablecoins like USDC or USDT will always be readily available and able to move, this is really the last point, not frozen or blacklisted. And that's relatively likely, but again, we get into the situation where it puts you at, it exposes a large attack surface to regulators who may not appreciate a new form of money being issued.

[00:52:49] That's not under their control, right? And so we've seen previous attempts at legislation of stablecoins in the United States and it is possible to imagine a bill passing through that would require, for example, more thorough KYC, AML of like holders of stablecoins and things like that. That could cause a lack or loss of confidence in a lot of these existing feedback, stablecoins.

[00:53:12] And in those kinds of situations where you see these things that protocols are built around with the assumption that they're going to be worth at or near dollar, you start to see things get difficult when that assumption is no longer true. DAI especially, despite being an overcollateralized stablecoin, is exposed to this and is especially tricky because of its leverage model, right?

[00:53:33] And so as they've suddenly lost that huge portion of value that can't be liquidated and you start to see a decrease in the trust and DAI. You can start to see cascading liquidations as other vaults end up needing to get liquidated causing these assets to be sold onto the market, feeding into these negative feedback loops that can depress asset prices across the ecosystem as a whole. And so fundamentally there's a whole bunch of different- there's a few different mechanisms by which a lot of these coins can end up failing, but often it's rooted in an assumption that a certain thing will always be worth a certain amount or that a certain thing will always be able to act as it has in the past. And I think that that exposes that's what leads to much of the black Swan and catastrophic risk in these kinds of models.

[00:54:24]Richard: [00:54:24] Okay. that's helpful. So, Sam, what I'm hearing from Bennett, I hear at least two things. Number one is that there is a certain scenario where there's a loss of confidence. For whatever reason, it could be an external attack It could be just the price incidentally dropping on FRAX Shares. And these two effects get compounded with each other, subsequently causing more pressure on FRAX's to be minted and then that depresses the prices even further entering some kind of vicious cycle. That's the first problem. And the second problem Bennett depicted is the USDC reliance. I think the reason I suspect the reason why you guys went with USDC instead of USDT is because you think that USDC is more likely to hold the pack longer.

[00:55:13]There's also been a lot of issues with USDC now being circulated on the internet, right? Not sure if it's fraud, I'm not sure if it's real, but this could be an issue. So if let's say these two problems, one of these two both were to take place. What is your plan to rescue the coin to continue to have it hold a peg? 

[00:55:35] Sam: [00:55:35] Yeah. So, first of all, I want to say appreciate Bennett for reading the white paper well because we do indeed you know, make sure to not peg directly to USDC, because like I said, originally, the point of you know, the holy grail algorithmic stablecoin is it should have the same security and trust guarantees as Bitcoin and Ethereum. And so you can't actually build that by assuming USDC or whatever fiat coin is always going to be a dollar. So in fact, in our source code, we even track, the, the value of USDC if necessary, if it does break the it's its own actual peg on chain. Second of all, I want to actually focus more on the second topic cause I think it's much more important because yeah. Anything can break it's peg in a black swan event like, Bennnett even talked about Maker, which is not an algorithmic stablecoin, right? And so the things that apply to everything doesn't seem particularly Persuasive for algorithmic stablecoins.

[00:56:40] I think the main thing that seems extremely persuasive, and it's a very, very big problem that I do want to address is the U S DC reliance, right? In fact, that is one of the biggest problems because we talked about FEI and FEI has no fiat coins, but it currently has a half working protocol. It can't properly expand or retract it only as essentially a tokenized ETH ETF fund, so to speak, right? And, and so, like Bennett was saying is it is leaking value. And basically a bet on ETH's dollar value increasing. The other stablecoins that have full protocols, like FRAX, even non algorithmic ones like Maker, in case readers aren't aware, Maker is about 60% USDC that they have about 2 billion USDC in their possession. One-to-one backing About four point something billion DAI. This I think is a much bigger problem. And FRAX also has a lot of USDC, as I've said in the examples in, in this debate.

[00:57:50] So we're actually doing a very, very fundamental, analysis of how to remove USDC reliance entirely from FRAX in the next two to three months, we're going to have new features and new things that hopefully not only remove FRAX's USDC reliance, but remove the protocols claims on USDC, which is actually really important, right?

[00:58:17] Because one thing I want to stress is that. It's not that the problem is solved. If the protocol, for example, takes USDC and deposits it into wire or deposits into compound, and then gets, cUSDC tokens, right? Cause that's still a claim on USDC, right? And I think this is the main, fundamental thing that everyone should be talking about and even should have debates about, right?

[00:58:46] Because this is the important thing where if a stablecoin on chain can't remove its reliance on something like this, it breaks my own personal definition of an algorithmic stablecoin. It is not actually in my opinion creating value that justifies, it's its own existence, so to speak its unique value proposition.

[00:59:10] And so right now, to be clear, Bennett is right, we do use USDC. We don't rely on its price being exactly at the peg, but we do use it. So we do actually have most of our reserves and USDC claims, some of it's, deposited in different places to earn low risk yield. However, in the next two to three months, FRAX could, and hopefully in my opinion, should become the first algorithmic stablecoin that has never broken its peg and also will not have any claims on USDC and in that situation, I feel pretty confident in saying we've hopefully we'll have reached the holy grail, even more so than, MakerDAO and others that are using a lot of USDC.

[00:59:59]Richard: [00:59:59] Wow. So removal of USDC or claims USDC as a collateral type. Does that mean you're more moving towards a zero USD collateral model? Or does that mean you're replacing USDC with something different? 

[01:00:14] Sam: [01:00:14] Of course, we would still have collateral, but we would remove not just USDC as collateral, but something even more stronger remove all claims on USDC from the protocol. This is really important, right? Because it's easy to weasel your way out of having USDC and the contracts by just depositing in different pool to places like curve wire and compound and saying, look, there's no USDC.

[01:00:40] They can't they can't blacklist us. It's not in our contract. We back it with cUSDC and those are like pooled USDC in compound. That's not exciting, right? That's not the same security and trust guarantees as Bitcoin and Ethereum. And remember our definition of algorithmic stablecoins is very strict.

[01:01:00] It's very clear and concrete and strict. we're going to not only minimize, but hopefully take to zero our claims on USDC and the next couple of months, obviously it depends on our features that we're releasing, working as intended from our research and everything, as well as obviously our peg holding as perfectly as it has that I'm actually extremely confident.

[01:01:27] The peg will always hold because the way we design our stability mechanisms is that it starts off small and it only increases after market signals have shown that the peg is set up the same way our fractional mechanism works. So I'm confident the peg will stay perfect, but let's see how that goes.

[01:01:47] And I think that this is my way of saying that I do accept that it's, critique of relying on fiat coins currently. That is something that is Bearish, so to speak algorithmic, stablecoin designs, but I do think it's possible. And I do think that FRAX, again, will be the first stablecoin that will be able to say it has no claims on USDC and be algorithmically stabilized.

[01:02:14] Bennett: [01:02:14] What will the new collateral be for like minting and redeeming then?

[01:02:18]Sam: [01:02:18] First of all, we haven't really released publicly a lot of the research and mechanisms we're building, but it will be a mixture of assets that we think are extremely important in keeping the exogenous usage of FRAX. So for example, ETH, CRV that is voting for exogenous gauges, other tight integration of the protocol, perhaps we're exploring either other kinds of governance, tokens, like Saddle Finance, which is another StableSwap pool as well as just basically a basket of extremely important utility while it's all digital assets combined with well defined hedges against their volatility. And so the collateral ratio actually adjusts as a soften the basket of their value. That might be a mouthful, but that's that's going to happen in the next few months.

[01:03:19]Richard: [01:03:19] Okay. So replacement with other cryptocurrencies that presumably have better security features or maybe not security features, but can hold them theoretically, hold of value. Stronger than USDC? 

[01:03:34]Sam: [01:03:34] No, I wouldn't say theoretically, hold their value against the dollar stronger than USDC I would say they have they're decentralized digital assets. They don't have a blacklist. They don't have, dollars in a bank that relies on how they work and they don't have a central, custodian that, that requires, them upholding their terms to the FRAX protocol. what it really means.

[01:04:01] Richard: [01:04:01] Okay. Got it. All right. Well, let's move on to audience questions. We got a question from someone named Aquin Littlepage in existing proposals. Is there a way to compute how much capital is required for someone with adversarial interests to destroy it, say like a big hedge fund or George Soros. So if you have someone actively adversarial towards algorithmic stablecoins, and that person is trying to calculate how much capital he or she needs to accumulate in order to execute as an attack, his or her attack. How would that work? 

[01:04:38] Sam: [01:04:38] Yeah. So the first thing I want to say is the, the market set collateral ratio is actually the nascent way, way of saying the exact answer to this, right? Because the collateral ratio is literally the ratio of collateral backing, the protocols money, right? It's liabilities, right? The minted FRAX stablecoins and that ratio is what you need to start calculations of. Okay. If the FRAX Shares as Bennett was saying is not able to back the remaining. 16% of FRAX, right? Which right now is 84% back then you can start running calculations, right? You can say how much capital is needed to redeem this much FRAX given the FRAX Share price is going to go down and this or that. And so FRAX actually does let you on-chain compute that, right? And then as we get even more integrations and more data on chain to analyze people can actually look and they can say something like. FRAX has 300 million TVL and curve, which means that there can be a hundred million FRAX sold into this pool, right? Because of the stable swap in variant, right? Without breaking the peg, then people can start redeeming other FRAX and start minting FXs, which is bad, right? It's dangerous. That's when the dilution of the shares happen, right? And so they can start calculating, okay. After the stable swap pool on curve is exhausted.

[01:06:16] How much more can the redemption process take? How much more can the Uniswap pools take in? One of the important things about being a fully decentralized on chain stablecoin is you can indeed do these kind of. George Soros, adversarial, simulations, and calculations, because everything is so transparent. Everything is so clear. And also that's why it's so important to do a market set collateral ratio because it is the markets input given all of the public data available right on track. So, the answer is yes, definitely. In fact, we do that regularly. We look and analyze how much sell pressure and defending of the peg the FRAX protocol can do given all of its integrations and organic growth. 

[01:07:03]Richard: [01:07:03] How low can the collateralization ratio go for FRAX? So when demand is strong, the collateralization ratio goes down, but how low do you let it go to so the risk is still manageable. 

[01:07:16]Sam: [01:07:16] I mean, Theoretically, it can go to zero, although I don't ever think the market will allow that, right? But theoretically, it could go to zero. And in fact, some people, you know, say that algorithmic stablecoins will never work and perhaps like in this debate. But my main response is okay, great. Then FRAX is collateral ratio will go up to a hundred percent and stay there, right? Cause as soon as it goes down to something like 84, you would expect the next retraction and supply to take the collateral backup to a hundred and perhaps not let it go down as far, right? Stabilize back in a fully collateralized protocol, which has not happened so far. And then the peg has stayed perfectly so it can go down to zero. I don't think it'll ever get there. I think it'll probably get to perhaps 50, 40, maybe even 30, although I would probably think that FRAX would need to become even bigger than Tether or you know, be integrated in so many different places for it to actually ever be able to get to that, but it could go to zero, although I wouldn't hold my breath. 

[01:08:27] Richard: [01:08:27] So right now there's on the order of 150 million or maybe 200 million FRAX outstanding, is that correct? FRAX dollars? 

[01:08:35] Sam: [01:08:35] Actually as of this podcast, there's 229,865,000 FRAX. 

[01:08:44] Richard: [01:08:44] FRAX dollars outstanding, right? So if for whatever reason, a large number of FRAX holders for X dollars holders start selling, then the mechanism, the protocol will respond by raising the collateralization ratio. And at some point the collateralization ratio will be very close to par, right? But how does, how does the FRAX reserve come up with the additional USDC to back up allow them redemption? 

[01:09:14]Sam: [01:09:14] So it mints Frax Shares and allows people to swap it or USDC into the reserve, right? So the way that, just increasing the collateral ratio by changing a number doesn't make, USDC appear out of thin air in the reserves. So, every time the system increases its collateral ratio let's say by 1%, then in order to have the reserve actually have 1% more collateral, it needs about, let's see, it needs about 2.29 million more USDC, right? And so it would meant Frax Shares for 2.29 million USDC and people are able to swap for that and you know, sell their fractures on the market if they want or hold it.

[01:10:01]Richard: [01:10:01] Got it. Okay. Anything to respond there? Otherwise we'll go to concluding remarks. 

[01:10:06]Bennett: [01:10:06] No, I don't know much about like specific calculations for attacking or, and the minimum acceptable ratio is depends on how well you maintain your confidence, right? And so I don't know much to add there. 

[01:10:22] Richard: [01:10:22] Okay. Great. So let's go to concluding remarks and we will have Sam go first. So synthesize your thoughts. Talk a little bit about what you've learned from the other side and maybe any adjustment you've made to your current position, which I doubt there is any. But please go ahead with your final word. 

[01:10:40]Sam: [01:10:40] Sure. So I'll end by saying how I started. I think the definition of algorithmic stablecoin is a price stable on-chain algorithmic protocol that has the same security guarantees as decentralized digital assets like Bitcoin and Ethereum, but is able to keep a fairly stable price target, generally a dollar because of the dollars unique position in global trade, right? So that's the definition I'm going for. Obviously I'm very pro-algorithmic stablecoin. I think that there are designs that work very well with just like FRAX, which empirically we've never broken our peg I do agree with Bennett that the biggest issue, not just for algorithmic stablecoins, but also for other on-chain stablecoins, like MakerDAO is USDC reliance and actually custodial risk, which destroys the whole point of having on chain stablecoins. In fact one of the main things with MakerDAO that's kind of odd is that they're moving more towards custodial assets. And so this is actually exactly the opposite direction we're taking in FRAX.

[01:11:53] So I want people to know the definition of algorithmic stablecoins that I use is something that has the same security guarantees Bitcoin and that's fairly stable. Now, there's been a lot of attempts as we've discussed that haven't worked. And I actually agree that a lot of them have fundamental flaws.

[01:12:11] I don't agree that that means that none of them will ever reach critical mass. And I don't think that being able to come up with very, very corner cases of things that apply to literally any kind of banking behavior, right? Including banking behavior, and traditional finance, like when banks collapsed in the subprime mortgage crisis and other things like that is a unique criticism that algorithmic stablecoins in scope are flawed, right? I disagree with that. And I think that that's an important distinction to make. Those are not criticisms directly at algorithmic stablecoin design. Those are just human banking observations about banking behavior, whether it's on chain, smart contracts running algorithms or traditional financial banking.

[01:13:02] So I think that Bennett actually makes a lot of good points, especially the USDC reliance. Obviously I think that there will be a critical mass. Algorithmic stablecoin. And so far, I'm very confident that FRAX is in the leader in that position. So my case.

[01:13:21]Richard: [01:13:21] Sounds great. Bennett, your final word? 

[01:13:23]Bennett: [01:13:23] I will start with it is impossible for any protocol built on top of either Bitcoin or Ethereum to end up with the same security guarantees, a Bitcoin or Ethereum, because you immediately have additional implementation risk. You take the net, the native protocol risk, and immediately add drone risk of your implementation.

[01:13:42] On top of it, it is fundamentally impossible for something built on top of something else to maintain the same security guarantees of the thing itself. Yes, we're discussing corner cases, edge cases, and unlikely events. But when the goal is to design a censorship resistant resilient system, the most interesting cases are those on the edges, are those in the corners, are those when these things are being tested and really pushed beyond the expected point, because that's when you determine whether or not they've achieved their goal of being decentralized resilient in censorship resistant.

[01:14:22] And so the reason I focus on those kind of cases and that kind of thing is because that is the useful case for these kind of protocols. 

[01:14:29]Richard: [01:14:29] Okay, great. Well, thanks again to Bennett and Sam for coming on the show. How can our audience find your work? Starting with Bennett? 

[01:14:40]Bennett: [01:14:40] I'm on Twitter @BennettTomlin. I am a co-host of the Crypto Critics' Corner Podcast, and I blog at bennettftomlin.com.

[01:14:50]Richard: [01:14:50] Fantastic. How about you, Sam? 

[01:14:54] Sam: [01:14:54] @SamKazemian at Twitter and also on Telegram, it's just my name. @FraxFinance also on Twitter and Telegram. And it was a pleasure being here and some very, very astute and good points made by Bennett. So I appreciate the conversation.

[01:15:12]Bennett: [01:15:12] Likewise. 

[01:15:13]Richard: [01:15:13] Thank you both. So listeners, we would love to hear from you and to have you joined the debate via Twitter. Definitely vote in a post debate poll. Also feel free to join the conversation with your comments on Twitter. We look forward to seeing you in future episodes of the Blockchain Debate Podcast: Consensus optional, proof of thought required. Thank you both. 

[01:15:31]

[01:15:31]OUTRO  Thanks again to Bennett and Sam for coming on the show. As usual, I won't be explicitly taking one side or the other. I will say that these capital efficient stablecoins are fascinating experiments to watch. They're a self-contained and fully transparent study in systemic confidence. Their goalpost is extremely clear and their potential failure is easily quantifiable as the amount of deviation from the peg for a meaningful amount of time.

[01:15:54]What was your takeaway from the debate? Don't forget to vote in our post debate Twitter poll. This will be live for a few days after the release of this episode and feel free to say hi or post feedback for our show on Twitter. If you liked the show, don't hesitate to give us five stars or iTunes or wherever you listen to this. And be sure to check out our other episodes with a variety of debate topics, Bitcoins' store value status, the legitimacy of smart contracts, DeFi, POW versus POS, central bank digital currency, the merits of government bailouts, and so on.

[01:16:22] Thank you for joining the debate today. I'm your host Richard Yan, and my Twitter is @gentso09. Our show's Twitter is @blockdebate. See you at our next debate.