Teetering on the Edge: How Black Thursday Exposed the Flaws in the Crypto Markets - Ep.164

Transcript

hi everyone welcome to Unchained your no-hype resource for all things crypto I'm your host Lauren if you enjoy Unchained or unconfirmed my other podcast which also features a weekly news recap after every interview please give us a top rating or review in Apple podcasts or wherever you listen to the show Kraken is the best exchange in the world for buying and selling digital assets it has the tightest security deep liquidity and a great fee structure with no minimum or hidden fees whether you're looking for a simple Fiat on-ramp or futures trading Kraken is the place for you cipher trace cutting-edge crypto currency intelligence powers anti-money

laundering blockchain analytics and threat Intel leading exchanges virtual currency businesses banks and regulators themselves use cipher trace to comply with regulation and to monitor compliance what's the best way to spend crypto the MCO Visa card lets you spend it anywhere visa is accepted including your coffee shop or the Apple Store all with up to 5% back download the crypto comma and reserve yours now today's guest is Kyle Somani managing partner at multi coin capital welcome Kyle hey Laura get through back on the show you wrote a couple of really interesting blog posts last week that

refer to some of the events on March 12th which we'll call Black Thursday for listeners who because we're gonna be referencing events throughout that day throughout the show and on that day due to panic over the coronavirus and its effect on the economy and amid a much wider sell-off the crypto markets crashed in two legs as he wrote about in one of your blog posts and you say in the first blog post say in the second leg the market structure of the crypto markets broke so why don't you just kind of broadly describe the events of that day in crypto so on just for clarity so if you look at kind of the like coin market cap you'll see a

lot of the so reported in the UTC time and actually the first leg down happened the twelfth and the second leg thirteenth and the u.s. it was basically morning and evening and so I'm just going to refer to the two legs as morning and evening just for simplicity so the first leg down happened at about 5:00 a.m. Central time 6:00 a.m. Eastern Time on the 12th we don't know with a hundred uncertainty what caused it but definitely someone who held a lot of Bitcoin started selling to Bitcoin the two best theories are one that someone was risking because they were kind of concerned about global macro and as a result of the coronavirus and the second theory is that the plus token scammers from kind of last summer

sold off Bitcoin and presumably they wanted to do that in the guise that was kind of being aligned with the crota virus they had a chance to see the u.s. futures markets you know limit down the night before and so they kind of assumes and everyone would rightfully just assume that the broader markets were the cause so those are the two best theories it feels like a high probability that one of those is correct there is a possible third theory or there are other possible theories but none of them that I think make a ton of sense and so it's very likely that one of those two things cost the first leg down on the 12th that moves Bitcoin from about 7,500 to about 5,800 or so over the course of about an hour so it's a very large and very fast move and then kind of prices stabilize

throughout the day I can tell you from our side know we were with a lot of other market participants exchanges market makers desks and stuff and obviously the day for all of those guys was really hectic and a number of people got margin called meaning that they had posted collateral to borrow some assets the value of the collateral the value of the assets they borrowed decreased such that the value of the collateral was now higher than the value of the asset that were borrowed and so assuming the value of the assets bar was higher than the day of the collateral and so those lenders were you know trying to work with with the borrowers to figure out how to either get more collateral or repay the loans and so that kind of back-and-forth went in between a lot of

lenders over the course of the day and what appears to be the case is that some number of borrowers were unable to either post more collateral or repay the loans so the lenders in the evening started liquidating collateral and that's what caused the second leg down unfortunately during the second leg down then we kind of started hitting some mechanical problems in the crypto markets and that's kind of why I said the market structure broke yeah and you have a really great dissection of how the crypto markets have these unique characteristics that kind of played into what happened there so I don't know if you could remember all nine of them but you laid out nine different ways that they you know are kind of distinct from traditional markets or or have their own

unique characteristics so can you at least lay out the major ones if not all sure so the most unique thing about crypto is that there are a lot of trading venues where people trade they may all trade the same assets you know Bitcoin etherium right Ripple etc in traditional markets by and large you can more or less think of like let the equities as like there being a single exchange where equities are traded in FX markets there are a handful of trading venues like the primarily major a few major banks and dealers around the world like three or four kind of major venues where people trade FX but in crypto there's about 10 or so major venues where people trade and although the underlying you know assets are core the same Bitcoin aetherium

those things the way they're traded is very different so you've got spot markets you've got futures markets you've got perpetual swaps which are a form of the futures market and you have options those being the major ones and then you know different exchanges have different they stray different products so not all the exchanges trade all the products for example coinbase only traded spot they don't have any derivatives if you look at bit neck Smith makes only trades futures and perps DARAB it really just dominates options trading and so you've got all these different venues and people trading different products and of course is the underlying price of Bitcoin changes the the price of these different derivatives products is also changing in real time and so what's really and then on top of that you've

got different kinds of traders who cannot access certain venues right so if you just say like I want to sell or buy 10 million dollars a Bitcoin let's just say like the generally speaking that the not sensible thing to do would be to go to a single exchange and like just eat through the order book on that exchange the the right thing to do and there's a lot of kind of people who formerly know how to do best execution on this stuff is you know you look at the order books across all of these exchanges and you'd write pick off the orders off of those off of all of those exchanges you'd space it out over time and do those kinds of things but that's assuming you can access all of the order books and crypto a substantial majority of traders in the space actually cannot access all

of the order books or even some of the order books some people can only access one order book well for example a lot of Chinese retail traders basically only trade on movie a lot of people who don't want a kyc only trade on bit max and so you've got people who structurally do not want to spread their trades across lots of venues and so as a result of kind of all of these you know unique dynamics of like different products being traded across different venues and the fact that people are not you know doing quote unquote best at best execution and kind of taking liquidity across all that in venues you end up with pretty meaningful price deviations between those venues and 99 percent of the time those price deviations are close to the zero or like maybe they get a little bit beyond the make or taker

fees but but not much and arbitrage errors will kind of just arbitrage those and like that's fine but what happened on the 12th was that those prices started getting so out of whack arbitrage er started having to shuttle money between the exchanges because like the puts a the price was high I want to change on the on the other but let's say that they're the the money they had on each exchange was was backwards so they had to basically flip the positioning of like the Bitcoin let's say the USD T that they had across those exchanges well in order to do that they have to use the block chains and you know withdraw from what exchange and deposit to another exchange and that takes time and so on the 12th like that started happening in such free high frequency that basically the Bitcoin blockchain in

the etherion blockchain became entirely congested so people were using up all of the block space on both chains and so gas fee started to go up the network sorted the congested transactions started to sort of not going through and so that was like a big problem just making the whole system more difficult to use on top of that because the price went down a lot of miners sort of turning off their mining machines because you know their revenue was decreasing and the cost of electricity is fixed and so it actually became unprofitable for them the on mining and so as when they do that that means the rate at which new blocks are coming out is slowing down which is obviously the decreases the overall throughput of each blockchain and so those two things kind of happening at once basically made it so that people

could not move you know arbitrage errs could not move money between exchanges quickly enough and so the prices started to really meaningfully deviate between those exchanges on top of that then things got even pause there like that was already a lot yeah well I mean what you're describing just sounds like one of those terrible feedbacks feedback loops like it just was essentially a downward spiral you know the problem was being exacerbated at every step of the way but one thing that I wanted to ask about because it was a little bit confused when I read this in your blog post you know one BTC is one BTC and whether or not you choose to sell it at four thousand or eight thousand or

twelve thousand or whatever it is is presumably under the miners control so I kind of don't necessarily understand why it is that they would just turn off their mining equipment immediately at that second because yeah I maybe I'm you know I just don't understand how they run their business but I feel like yeah I could see an argument for being like oh well you know they could just hang on to it until the price goes back up right so they can and certainly miners historically have very much speculated on the price of the coin as the derivatives markets have matured specifically the futures markets and the options markets miners have become a lot of miners have gotten blown out like in

2018 as the market crashed and so they become a lot more a lot less willing to take on price risk of Bitcoin and so they're fortunately now they can hedge this to a large degree where they can say hey my cost of mining Bitcoin is five thousand I can get it guaranteed lock in price of seven thousand and so I'm gonna do that and I'm not going to take on the fact that maybe the price falls to six thousand or five thousand or four thousand so a lot of miners are becoming a lot more rational about that and but but if you're so that has started to happen however there are a lot of miners who still are not hedged and so for them right if you're a miner let's say your cost of mining Bitcoin is six thousand and as the price of Bitcoin

it's 6,000 maybe you keep on mining but like maybe you say hey it's 5,500 right am I gonna mines for 10% loss like if it it's 5,000 am I gonna mine for you know almost 20 percent loss 17 or 18 what you know whatever that math comes out to be and a lot of people start to say hey this just isn't worth it like I'll just wait for the price to come back out before I start mining why would I mind now and have a guaranteed loss and so that that started happening on the 12th and then one other thing I wanted to ask about was you did talk about how there were obviously increased transactions plus higher gas fees and transaction fees and so why was it that those higher fees were not enough to offset that the you know the decline and the value of

the block reward yeah so I haven't looked at the ratio recently but right if you just say a minor makes 100 cent of revenue I think right now is something like 90 to 95 percent of my no revenue kind of generically speaking comes from block rewards and only five to ten percent comes from transaction fees I don't remember the exact math breakdown but it's it's something to that effect and so like fees going up like three or four X like doesn't it doesn't offset and often like where they move the needle right okay all right so um then you know in the beginning you talked about how part of the reason for the first sell off was perhaps you know coronavirus or or maybe pin the plus

token but then for the second one that was what was the cause of that again was that the miners returning office yes so the cause of the selling was that there were lenders who had lent out assets and the borrower's became insolvent or were close to becoming insolvent and so the lender started liquidating the collateral of the borrowers what's interesting is there's two different types of ways to get leverage in crypto there's you can use do it on exchange using bit max or finance or any of the derivatives exchanges all that you take on leverage and if you follow any of the twitter accounts like well then I think one's called whale calls I think another one is called bit next rect there's probably other ones as for the other exchanges and basically right there if you follow

those accounts they'll start tweeting out you know this this this number of contracts liquidated in this price that's a they're just like BOTS that tweet that out so those people are taking on leverage on the exchange and those exchanges have rules and say hey look if you're levered long either short or either you're never long or levered short if the price moves against you there are some threshold at which the exchanges will just they will take your collateral and they will just certain markets selling your collateral right so to make sure that you stay solvent and so that happens all programmatically in the exchanges handle that the other way to take on leverage is to basically call someone in this case a lender so think people like Genesis or block Phi or matrix port or those kinds of companies and those kinds of companies say hey you know if you

want to get a yield on your Bitcoin deposit as a Bitcoin with me and I'll give you a yields like a bank and if you want to borrow some Bitcoin you know call me and I'll lend you something and so those guys started to you know the people who borrow Bitcoin and eath and whatever else they had borrowed and probably dollars in this case those people started to become insolvent and so in those cases the lenders don't typically programmatically market sell on the underlying exchanges they always prefer not to liquidate people's collateral and so they were quite literally calling up you know the people who borrowed and saying hey Bob hey Joe hey Jim you know you're close to being insolvent you know you need to deposit more collateral or we're gonna look at your clock or or liquidate your collateral and so yeah it sounds like

they had about 12 hours or so of you know back-and-forth and dialogue and obviously things did not get resolved in that 12 hour period and so the lenders you know contractually had to start liquidating that collateral and so that's what caused the second leg down okay and so like on this chart that you have on your blog post there's what does this look like maybe a period of about 12 hours or so where there does not appear to be very many liquidations at all it's sort of flat after the first big spike in the morning and then then it continues to go up is that what kind of what happened they're like what why why was there this period where it doesn't look like there were any liquidations correct so that's what happened during that for that 12

hour period was the lenders were reaching out to the borrower saying hey guys you know here's your balances here's your you know solvency ratios and whatever you know you need to either deposit more collateral or you need to return the loans and you know the tracks between the lenders and borrowers you know they have language saying hey we're gonna give you this much notice if you don't respond in six hours or four hours then we have the right to liquidate your collateral once you respond you know there's like a whole bunch of provisions in those contracts about these kinds of situations that very formally defined right like how quickly things need to get resolved in this case it seems like it was about 12 hours it's kind of about the amount of time in the whatever contracts were written and obviously things did not get

resolved and so then at that point the lender is legally you know had had the right to claim the collateral and then in order to ensure that their depositor state solvent they had to liquidate that collateral Wow okay so there was another thing here and I was trying to there was a lot of speculation about this online but basically so there was one point where so in your blog posts you said that there there were only 20 million dollars worth of bids left on the entire bit mix order book and over 200 million dollars in long positions that needed to be liquidated and you say that that could have crashed the price of BTC 2-0 briefly but that would happen was that bit mechs ended up going down for

maintenance and I did see some people speculating that bit mechs had kind of gone offline to save the price of Bitcoin like what do you think happened there and how also how does this play into what you were saying before about the Bitcoin blockchain being backlogged yeah so there's a few kind of compounding variables here so all of this happened during the second leg down which was in the evening in the US so once the lender started liquidating collateral the obviously the price is right a Bitcoin started going down more at that point there were people who had leverage on bit max I mean people had levers on all of the dirt as exchanges but on that specifically there was more people who were who were levered long and so those those

liquidations started to cascade and the problem was that over the course of the day people had already done a lot of arbitrage errs because of the the the price movements in the first half of the day and so a lot of market makers have already moved collateral off of bit necks to take advantage of arbitrage you know on other venues and so the amount of collateral available on bit was lower than normal on top of that that so the thought of that you know the blockchain was slowing down because you know the miners are hurting their machines off and because there were just so many transactions that were shoveling around and so that was and so is the the price of sort of cascading down on bit max what happened was the market makers basically stopped showing up this is actually common that like in general

when volatility is increased market makers widen their spreads in general like market makers their business mandate is every single day don't lose money and like if there's a day where market maker lose money like that's a very bad day for them and and there's all those businesses are run with a very strict risk controls on making sure that they're not taking on too much risk and not quoting spreads too tight so that they lose money and so on days like you know on March 12th right like those are days that the volatile is really high and they don't to lose money and a lot of market makers may have gotten may have ended up losing money in the first half of the day and so they were already just very risk off and scared and a lot of market makers just said hey this volatility picking up this is getting

crazy we don't know what's going on and so a lot of market makers just stopped providing liquidity entirely and so as the price of Bitcoin started cascading down in the evening market makers just didn't show up to provide liquidity on bit necks and people keep getting liquidated there's a few more and it wasn't not only that there were obviously not all market makers disappear but you know some percentage did and so even the ones that wanted to provide liquidity at that point they just didn't have collateral on the exchange and so they were trying to shuttle collateral from you know coin based finance where we'll be wherever else they had collateral they were trying to send it to MIT max but the blockchain was slow and so they couldn't get collateral to MIT mix and so they couldn't actually provide liquidity and

so as a result of all of that there ended up being a point where there were two hundred dollars of outstanding liquidations that needed to occur but there was only 20 million dollars of orders on the bid side of the book and so I mean had the engine you know had they not turn the system off or or how did I come down I mean the prize a Bitcoin would have gone to zero yeah and it sue if that would have happened on bit max like what would have happened to the price of Bitcoin where oddly across the market I mean nothing like that would have only survived for like 1 or 2 minutes 3 minutes like people were obviously trying to send money to Bitcoin and there is a price right like if sorry yeah right like let's say Bitcoin is trading for 500

ours on bit max in $4,000 on coinbase like there are people who are gonna say right like when the price discrepancy is that large like I don't care if I I lose everything a bit max blows up in the next five minutes I'm willing to risk that to try and make you know 8x in 5 minutes right and like there are people who have that risk tolerance and have this sophistication do all that stuff so as the price of discrepancy gets wider it obviously encourages more people to show up and so at some point they would have shown up but like the the spike on bit max you know it could have printed 0 and that's just like a very bad luck obviously Wow yeah oh my gosh just so much craziness alright so in a moment we're gonna discuss how all of this played out in defy and also look at what

different on chain technologies could be solutions or preventative measures to these issues but first we're gonna get a quick word from the sponsors who make this show possible why should you get an MCO Visa card from crypto comm first it's a beautiful metal card you can top up the card with crypto and spend anywhere visa is accepted you also get up to 5% back every time you spend on all spending including your morning coffee gas or even a new phone you know they'll pay for your Spotify and Netflix - you'll love the unlimited airport lounge access and interbank exchange rates if you travel a lot there are so many cool perks loaded in one card download the crypto dot-com and reserve yours now will the world follow France

and advocate banning privacy coins will government backed stable coins become the new Fiat are distributed and peer to peer exchanges just a flash in the pan the answer is maybe virtual currencies can flourish and create a new private and more versatile economy but grand vision can't happen without keeping Krypto clean and that requires support of governments and accountability for bad actors privacy enhanced compliance using cryptographic controls has the potential to preserve anonymity without compromising legitimate investigations cipher trace is working on this vision of the future sign up to stay up to date on the

privacy enhanced compliance initiative and receive authoritative crypto AML reports quarterly WWF root race comm / keep crypto clean today's episode is brought to you by Kraken Kraken is the best exchange in the world for buying and selling digital assets with all the recent exchange hacks and other troubles you want to trade on an exchange you can trust Krakens focus on security is utterly amazing their liquidity is deep and their fee structure is great with no minimum or hidden fees they even reward you for trading so you can make more trades for less if you're a beginner you will find an easy on-ramp from five fiat

currencies and if you're an advanced trader you'll love their 5x margin and Futures Trading to learn more please go to Kraken com that's KRA ke n comm back to my conversation with Kyle Somani so let's now talk about how these market failures on Black Friday Black Thursday played out and defy you wrote in your first blog post that defy quote outright failed which is interesting phrasing to use so what happened there and let's start the story with maker Dow yeah so of the major defy protocol is the one that definitely had the most problems

and on March 12th was maker so for those who don't know makers a way to get leverage like everything else in crypto getting leverage is generally like the top thing people like to do they like to trade with leverage in the case of maker you post aetherium or eath is collateral and you withdraw dye as a form of debt and the people who do that are generally doing so for one reason which is they want to go large and long or left with long 'if and so they're taking that die and selling it for youth alright so they're getting levered exposure to the price of aetherium so on the 12th during the first leg down price leg down bitcoin in the theory on both crashed pretty hard as a result of that a lot of those make

your doubt collaterals became under collateralized and so people started liquidating those positions on traditional centralized exchanges like bid max they exchanged automatically runs as liquidations and they just will market sell or market pie and they'll just you know eat orders on the order book right to execute those transactions in the case of maker now there is no order book right and that's not a centralized exchange and so in those systems anyone who this term is kind of generically used as a keeper keepers run some software now software was written by the maker dive team and that software basically it just looks at you know at the price of each looks at you know the collateralization of all of the outstanding loans and it says hey you know if anyone is under collateralized you can go liquidate them and you get

some reward for doing so i was like a few percentage points or something and so there's a rational motivation for people to do that so that that system broke down in kind of a couple different ways two to three different ways on the cert on Thursday the 12th so the first is that the if your network became fully congested and so you know blocks like people gas fees were going up and blocks were full the keeper software was not written to actually dynamically adjust gas prices it was basically using some setting to just say hey set the gas fee it I'm just gonna say one penny there's implicitly but like the market wasn't clearing transactions with anything below let's say 10 cents and so like the people who were running the

keeper software that out of the box keeper software we're trying to liquidate trend you know the liquid collateral and we're just not doing so because the software wasn't configured to do this some very smart person realized this realize what was happening and so they went into the system and they just changed the parameters and they said hey just change the gas price and increase the gas price so that transaction will get including you other miners will include the transaction well that person realized is that the way that the liquidation is working maker is that those are actually there are collateral auctions and so the idea is like there's a hundred dollars for sale right like you can bid up that's a $99 and then you okay $200 and then do you make $1 profit right so the what some guy realized hey no one else's

transactions are going through so I can bid $0 and I can get $100 and so some very smart person did that and collect it I think six million dollars or eight million dollars of eath over the course of yeah I think the total was 8 million but at least one of them was maybe like 4.5 but anyway yeah so yeah so someone with millions of dollars and paid effectively nothing for it and so that was that was the first failure that happened now that's fixable I mean if the keeper software was written to dynamically adjust gas prices that wouldn't have happened so that's a pretty fixable problem the second problem was that liquidity in order to run those liquidations you need to be able to you have to compete in those auctions with die we need to have die and it turns out that there weren't

enough keepers who had enough liquidity like had enough die on hand to actually engage in those auctions and so the price of die I went up to I think a dollar and ten cents because people realized that like this was happening we're spying too high at any price to try and run this trade so that's obviously a bad thing you want the price to die to stay about a dollar so that was the second problem the solution to that I'll come back to the solution and then the third problem was the Oracles so right the maker system needs to know the price of eath USD in order to make sure the system is staying solvent and if the price of 80 USD falls too low then you know the keepers can come and do the liquidations well again the etherion blockchain became 100 percent congested and so the makers either were

not configured to adjust the gas prices or they just chose not to submit transactions outright because the price of eath kept crashing and so during the second leg down when touched I think $88 on coinbase when that happened you know the the the last price that maker reported that was formally incorporated in the system was about adding 101 dollars or something like that and it turns out there were a lot of volt source EVPs in the maker system that we're going to get liquidated at $100 and so like the conspiracy theory is that the or the people running the Oracles realized what's happening they probably realized that hey if this huge amount of collateral gets liquidated

the people taking that eath collateral may choose to sell it for dollars and cause the price of eat the Cask a further remember a couple minutes ago I was saying that you know in times of crazy volatility market makers just stop showing up because they don't know what's going on and so you know there's like what the beef is down at $88 or when Bitcoin was down at 3800 or 4000 at those prices there's very very little quiddity and so if someone's trying to market sells five million ten million dollars of eath that's gonna crash the price and of course there were you know other maker vaults with even lower you know like lower liquidation thresholds and so it's very possible that eath could have just cascaded down from 88 that like it have the maker Oracle's you know been reporting the

price correctly it's very possible that needs could have cascaded from 88 down to 50 or 40 or 30 dollars because market makers just weren't showing up and so those were kind of the three big problems but defy that that you know even though I read your blog post before my jaw is still like on the floor because I mean the thing in fact the Oracles is it like if the Oracle really had just dipped one dollar lower to 102 I mean think about it like and it's so I actually did not see much chatter about that but were there any conspiracy theories about what happened there because it almost feels it like in both

instances like things were manipulated I mean I look at this other people who run the maker Oracle's the way it's designed is there are some people who are publicly disclosed and some people who are not publicly disclosed and that's like a safety mechanism I think it's actually very smart designed to maker team came up with and I generally agree with it some of them I remember all the participants off without my head the ones that are publicly disclosed but the maker team has them written down somewhere but like I think the 0x team has you know one of the keys the maker team is one of the keys I forget also the keys obviously even the people who are anonymous who we don't know who they are obviously those people are vested in the interest of maker in the long run right like they're not bothering to do

this and because like for fun and because they think maker are stupid they're obviously doing this because they care and presumably a lot of them own a lot of MKR a lot of them also probably own a lot of eath too right like I think the probability of that is extremely high and so I mean it was very rational that they realized what was happening and said oh my god just stop we can't prove it I'm not accusing them of anything but on the surface that is a very very positive explanation' right right listeners that we do not know the facts we're speculating here but but in terms of the other explanation would that just be the congestion on the etherion blockchain correct so that's the other theory is

that it was just congested and the in the software was not configured correctly in the trend you know they weren't including the gas fees you know to actually get included in the blocks Wow okay all right so let's let's now just talk about kind of you know let's just play through like a little hypothetical here so if those liquidations had occurred at $100 as they were supposed to you what do you think would have happened next particularly in terms of the whole maker system and then also the knock-on effects throughout defy I mean I think had had the Oracles you know reported the price down the $88 I think each

would have traded down to probably 30 or 40 and that would have caused more liquidations and it's very possible that a hundred percent of the collateral and maker could have been liquidated I don't think I don't think maker would have become insolvent in the sense that like the the system would like break like the contracts that are broken I just think that you know ninety or a hundred percent of the users would have got liquidated and that would have you know outright shattered confidence in the system yeah well except for the congestion on the floor I've seen but anyway yeah and then so what about the rest of defy so I mean yeah there's interesting second-order impacts on the rest of defy so die is collateral for a lot of other protocols and so for example in compounds right you get some

interesting effects if 8a II is collateral and a ton of other protocols so compound lend for me you know B Z X dy/dx all these things and so had the price of eath crashed you know really low that would have caused cascading liquidations on other you know t5 protocols which would have probably further cascaded the price of the AP even lower which would have been dad the price of dye also like like as if if maker started to become insolvent the price of die either could have skyrocketed because like die would have been contracting that's unclear or the price of die I could have crashed because people realized oh my god the system is insolvent I don't know how those dynamics would have played out on you know I mean a five-minute basis I

don't know I don't think it's possible to know but certainly there's a case where die could have become you know started trading well below it all over you could have traded at 70 cents or 50 cents or something and if that happened then a lot of the loans that are collateralized by die you know also what a probably become insolvent which would have caused even more liquidations and so this would you know this already cascaded to some degree into other protocols for example compound you know compound had a lot of liquidations firing primarily because it youth collateral being low but it could have caused you know just like like 100 million dollars or more liquidations could have occurred all right and actually so one of at one point earlier when you were talking about how if the

die price had gone off its peg you said you would talk about the solution to that later is that the US DC thing that maker doe adopted later on yes so today so so one of the core problems was that the price of died I went up to a dollar 10 and it was very hard for people to run these collateral auctions to actually buy the Biggie's collateral in the system and so one of the problems was that people couldn't get died and so today in order to do that you have to you know deposit eath and you have to effectively take on risk exposure like today if you want a mint new die you can do so at any time but in order to do that you need to be able to be able to hold eath right and take on take on risk

at least a price of eath and a lot of market makers don't want to necessarily take on the price risk of holding eath in order to create more die and so the idea here that there's a pretty clever is the maker team said hey look we're going to allow anyone to create die using us DC collateral of the USD see the price is very stable at $1 I think it's never gone more than a penny up or down and so we're gonna let anyone create died using USD C and that's basically the solution then that you can deposit USD C create die and then you know run any of these auctions and so you can effectively guarantee the price of die won't go up to a dollar ten again this obviously is controversial because it introduces you know trusted collateral into maker which has been a

pretty controversial debate in that community for a while I think the way that the maker team has has proposed this and I think I think was voted in last week or you know recently to to be used is that they said hey if you're gonna borrow deposit US DC and withdraw dye the interest rate it's gonna be really high I think the interest rate is set at 20% or 25% and so it's designed to deter people from leaving lard like creating large amounts of dye in the system and leaving it there for a long time because the 20% interest rate is really expensive but of course if you're just gonna borrow it and repay the loan in five or ten minutes then you don't really care if the interest rates 20% and so they designed it very clever where they said hey look we can solve a

liquidity problem by allowing for this other form of collateral but we just expect no one's gonna leave these loans outstanding for more than the you know a few minutes oh that's interesting okay yeah I did see some people kind of upset about that but I feel like what you just explained is is kind of a good way to take that criticism into account okay so now let's go to the second blog post that you wrote which covered all these different solutions for potentially dealing with some of the market failures that we saw in the crypto markets so obviously you had talked about how one of the issues here is the fragmented markets of crypto but you actually say

in the second piece that you expect even more fragmentation so why is that and then what are some other potential ways to reduce the volatility without more consolidated trading yeah so I mean if you look over the history of crypto from I'd say up until the second half of 2017 they were generally speaking no more than three to major liquidity venues all those liquidity venues basically only traded spot those exchanges for the most part didn't have any form of kyc a coin base I think did and crack and they've like none of the other ones did and so there was just a pretty limited number of venues obviously a number of more participants is much smaller the sophistication of mark participants was much lower you didn't have you know firms like jump and vrw and Susquehanna

encrypt up at all at that time okay our W is there but but basically no one else and so there was the market was not very sophisticated is pretty limited what happened in the second half of seventeen is you got a lot more sophisticated exchanges show up so by Nancy my name it showed up it makes got really big will be at okay and all those guys sort of sorted or the growing and sort of adding a lot more sophisticated products meanwhile you gotta got Chinese government cracking down on you know like exchanges in China and as a result of that basically will be okay and by an answer the kind of sort of the only three exchanges that are allowed to function in China and although I don't know this for a fact I think with a pretty high probability that if you try

and launch a new exchange in China the government will basically not let you get off the ground I'm some of those three you know seem to have some sort of kind of just they were already here and so there are lots to keep living but they're no one's allowed to launch new exchanges so those guys have kind of a kind of inherent grandfather basically Claus from the government's so it's unlikely you're gonna see new people show up there Korea basically the same thing kind of happened and so and then we've also heard see the institutionalization of the US market with firms like CMA and backs getting into space and so for all these reasons right the market structure has really started to diverge where you have just more trading venues and people have meaningful network effects around each those trading venues there's always a

lot of traders who value kyc don't want kyc you know there are chinese retail there in american institutions and so as the markets which are it's become very clear that there's different types of market participants who want different things and that really supports structurally having lots of different exchanges is it clear the answer is six versus eight versus 12 versus 18 I don't know that's a pretty wide range but but it's clear that the market is going to support quite a few different exchanges structurally at least until some of these types of parameters change and so I don't expect you know anytime soon that like that Marco structure dynamic is good it's got to change very much and then another option you looked at was potentially disallowing extreme leverage such as the 100x leverage

that's allowed on bit Mex or the 125 X leverage allowed on Finance how would that how how would you know disallowing that help and why do you think that this is likely not something that they'll actually implement yeah so I mean they offer that much leverage because some people want that right they're not offering the product and like like no one's using it and so people are trading with that much leverage a lot of people are not trading with that much leverage but some people are and and so and you know these exchanges are unregulated and so these businesses rightfully so want to serve their customers and if a customer say I want to trade with how much leverage

then the exchanges you know want to let them do that so I think it's unlikely those exchanges will disallow it basically because of their and their capitalist and they want to serve their customers the problem with that unfortunately is that if you make it too easy to get to a leverage you can very quickly get cascading liquidations and that's what we've seen happen time and time again they the second leg down on bit max you know when there ended up being two hundred million dollars liquidations on you know only 20 million dollars in the order book that right if you simply disallow a lever let's say the neck scaffold average of 10x I'm fairly certain that that situation would never come up right and no but why do you think oh so you're saying that the the exchanges won't I just allow it just because since there is consumer doing a

customer demand they they'll just meet that demand correct I mean like that crypto like one of the rules a crypto centers your persistence if your customers want and you know nonsense of them right like that's what they want they know the risk anyone who trades about much leverage knows what they're doing if they understand that's very risky and so they know what they're doing fair enough let's hoping when other option you looked at with circuit breakers and I saw that your partner Tushar Jane tweeted a god as a suggestion and his tweet got you know a fair bit of pushback eventually actually in recent days Kobe did introduce a circuit breaker on its derivatives platform but do you it seemed like you believed that

the exchanges are not likely to agree on a global circuit breaker provision so can you talk a little bit about how circuit breakers would help and why you think they're not probably gonna be a big solution yeah so if you look in traditional if you look at like the New York Stock Exchange for example they have circuit breakers I think the rule is if the price moves seven percent or more in one day then they are in some period of time then the starter breakers kick in and they just freeze all the markets that works really effectively when the underlying assets only trade on the single market in this case it's a US equities but if you think about it like if you Institute let's just say that circuit breaker is 15 is price of Bitcoin goes on 7% in 50 minutes well those those term because there's

multiple menus at the table those terms become very challenging there is no like single price at the coin at any moment in time the price between all of the exchanges varies a little bit you know it's usually very small it's usually a few basis points but like in times of volatility in times when the circuit breaker would be going off like that's definitely not the case right like the price of the exchanges is different and so you end up just creating right like even if all the exchanges shake hands and say yes well all institute a 30-minute circuit breaker if the price movement is you know X and no some of them are of minutes it's still really hard because the circuit breakers are gonna go on and off at different moments in time because each exchange has their own local price and so what you end up

doing is you end up just screwing a lot of market participants because of just like the market market microstructure both as certain exchanges shut down right and then others hopefully hopefully follow and then also on the reverse side of when those exchanges you know light back up right and turn back on and start taking trades again because they would end up like turning on at different times and so you have a lot of like very very difficult problems with market microstructure and doing that in a fair way and as a result like someone's getting it hurt really someone's gonna make money from that obviously but someone's gonna get hurt in what I would consider an unfair capacity because they just can't compete at that micro structural level so logistically it's very very hard to do time I mean even just agreeing on

notions of time the terms of price in all those kinds of things and obviously volatilities gonna be very high in these moments and so and then also there's just there's a strong incentive to to break grains let's say all the extremes got together and shake hands and say hey you know we're not gonna do this and then let's say like say there's 10 people and nine of them do it and one of them doesn't because that worse and says great I'm the only one open for business so all of the trading has to happen on my exchange right now I'm gonna collect all the trading fees and so you have a real tragedy to Commons where it's just it's super difficult to get people on board and even if you could get them on board the actual micro and structure of the implementation it would be very very

very hard yeah yeah that makes sense so in the second essay you actually also reviewed a whole bunch of different tech solutions for increasing throughput in lowering latency on the blockchain networks themselves so what you started with some of the different technologies on a Bitcoin network why don't you talk about some of the different ones there and you know what your assessment is for their potential to help in these situations yeah so two major kind of proposals for improving the aggregate throughput and decreasing latency of Bitcoin transactions are Lightning Network and side chains and I think there's kind of structural problems with

with both in the case of lightning right like in order you have to pre collateralize all the channels and each channel is a bilateral channel between two parties so today they're about ten major venues and so if you wanted to you know have channels between all the exchanges that's about a hundred different channels between these folks and there's a hundred channels and that means each you know each exchange has to collateralize you know one side of that and and so these exchanges would have to be terms of balance sheet management it's pretty risky for them to say hey I'm gonna freak lateral eyes all these channels and like I don't know if people are gonna use this I don't know which of these exchanges my customers are going to want to send money to and that really limits the flexibility because every Bitcoin that's in one channel cannot be

using another channel not only is it limit their flexibility in that way it also limits their flexibility in terms of things like let's say loans so it's a lot of the expanders in starting to loan out assets and so you know like how they think about risk of their own balance sheets as they have all these can interests it's really difficult to reason about this running a lightning note is expensive it's risky because it's up you know it's a hot wall as always online and it's very capital e and efficient because they don't know you know where customers are gonna want to send the money and so they have a large amount of overhead for maintaining this they know for a fact that customers are not going to use it the vast majority of time and and then even if their customers did start using it right

like people will because the liquidity is divvied up across a hundred different channels you know there's gonna be like like the arbitrage errs are gonna very quickly exhaust that channel and then that channel is gonna become sitting there useless right like one okay the first few people to arbitrage will get instant transactions but then as people trying to follow along that same corridor then that channel gets exhausted and so you end up having to you know Rica lateral eyes the base kind of anyways so I'm pretty skeptical the Lightning is like a core solution to this problem it just creates a lot of mechanical complexity and you end up just breaking up basically liquidity between these different channels so I'm skeptical there the other chains yeah so the other major proposal for

this side chains so I think the block stream team proposed side chains I think as early as either 2013 it's not thirteen then definitely four twenty yeah and so it's been talking about for a long time today the usage of side chains on it Bitcoin is effectively zero the only real side chain out there is the liquid side chain made by the block stream team liquid has been around for more than twelve months now and still as effectively no usage and a handful of the major changes are signed up to use it but they're not using it and there's also quite a few missing people I know by Nance is missing coinbase is missing crack this by saying I think a few others are missing too and so you know you don't have everyone on board side chain theoretically should solve the problem but the exchanges are just

choosing not to use it and so the kind of natural question is well why extreme doesn't have some reason to use side chains and that it would very clearly help them save on gas fees on the blockchain you can bet that all of the exchanges paid a lot of gas fees on you know on March 12th because the you know transactional costs were going up on the block chains as networks became congested and so they're actually incentivized to use something like liquid cuz it'll actually help them save money I don't know for like I haven't called all of the exchanges and like had conversations with them and asked them and this visit a question but as I think about it the the most obvious reason I think they don't use it is that the examiners don't trust each other they are competitive you know brutal

competitors they generally don't like each other and there's a lot of ways we're like liquid could backfire so the way liquid works is it's an 11 of 15 multi-sig and so if any eleven of the exchanges clue they can screw you know they can really hurt other exchanges and so you know if I'm left at coinbase is a good example here or CME or backed right if I know a lot of the other people on liquid or you know Asian exchanges that are unregulated if those guys you know call each other and say hey we're just not gonna send these coins to point base you know they're you know the things we're supposed to that we can they can just choose not to and they're gonna take over themselves and given the fact there's like almost no legal recourse right like this would be totally

uncharted territory exactly it's really game through it'll be dangerous for for coinbase and so I think that's a big reason why these exchanges haven't opted in to use it is because they just didn't trust each other and that wasn't 1115 configuration that's it's very easy to see how you get you get hurt not for a million dollars I mean you could get you know hurt out of fifty million dollars I mean it could be a huge huge loss well so do you see any way to solve these kinds of congestion issues you know during times of high volatility like that I'm a Bitcoin I mean yeah I'm Bitcoin I'm not really optimistic I I don't see how to do it on on Bitcoin I think with the the smart contract platforms that have you know newer just approaches to scaling as well

as they have so it's kind of full scripting languages that allow for more advanced more advanced smart contract logic I'm more optimistic but on Bitcoin based on my kind of current outlook I'm not I don't see how it's gonna get done all right well let's talk about a theory and quickly what you know are some of the different solutions there and what's your assessment of those yeah so on aetherium there's a handful solutions roll-ups are kind of a common one trying to think what are the other ones was another you mentioned in your blog post yeah that's right starting is the other so I'm starting first so you know eath 2.0 is working on charting a lot of the new layer ones coming out or also working on charting as well and they have that other one flavors of charting

and they have different nuance as to how they do what they do but the really big problem I have this charting is that it's not clear that it actually helps in this specific situation where you have a lot of money needing the shuttle around and kind of much of what I'll call like random ways meaning you just have rank new random deposits and random withdrawal addresses and so the reason that I'm concerned that it doesn't solve the problem is that in a charted system if you send the transaction from sharda The Shard be like you send money you know and you need to go do a cross shard transaction you end up doing a transaction on sharda and on short d and so when you do that if you think about that and then if 100% of the

transactions are cross shard then what you end up didn't affect you end up getting is that the aggregate throughput of the system ends up approaching the the throughput of a single shard because every transaction takes up a second you know it takes up space on the second charred and so that's really concerning because the because then you end up like right if people are struggling money around to run these arbitrage errs and all these transactions in the being cross shard then the whole you kind of defeating the whole purpose of charting and so I'm very concerned about that kind of happening obviously in any of these cases you would not end up with a hundred percent of transactions being cross shard but certainly a meaningful percentage would be and like and that's

gonna be again during a time of volatility a time when things are moving around and so I the idea that sharding is gonna increase throughput by 10x right 50 extra 100x I'm very skeptic a s-- 90% of the time but like the during the time when the systems are breaking and when everything is you know volatile and going crazy that's when you need it to really you know maintain that throughput advantage and during those moments of time is when I think that's gonna really get tested really it's terrible but when you were talking I realized that it reminded me of the hospital bed or ventilator issue with the corona fire I was like oh it's the same problem umm right anyway he's time everything is good like building systems to run during peacetime

is easy like building systems to work during wartime is really hard yeah well and then but what about optimistic roll-ups yeah so optimistic Wallops have a similar problem where they couple problems so one is the at least of vanilla implementation of optimistic roll-ups doesn't sort of the latent doesn't solve a latency issue because you still rely on the underlying layer one blockchain for for clearing transactions so you increase throughput but you don't increase or decrease latency there's a problem basically is that if these exchanges end up using their own optimistic roll-up chains then you haven't really you're basically recreating the same thing as sharding where you have to end up having to do the same transaction on multiple chains or multiple optimistic world chains in order to produce the

kind of same effects and so it's like I look at these exchanges these guys don't coordinate on much of anything and saying that hey they're all going to agree to use the same optimistic roll-up chain it just seems to me to be very unlikely like they haven't done that in the past you know technology's still new these guys all seem to want to kind of like have control of their own systems and so I just find it unlikely that all of these your players agree to use the same same optimistic world of chain and if they don't then you end up basically and kind of the same kind of sharding logic that I just walked through Wow all right well let's see you kind of started to go into like how you think we could solve this issue and I know this is maybe more speculative or you know off

the future but why don't you go into that a little bit yeah so I mean my favorite solutions to solve this problem I think there's one interesting market structure solution which is the introduction of prime brokers I think prime brokers really helped a lot because they basically allow you to have a single centralized entity off chain basically manage all the collateral across all the venues and allowed the different funds and traders to do all the trading and then just net all that stuff out on some frequency I kind of after the fact prime brokered you know very big business in traditional markets people like prime brokers the challenge is that you a you need everyone to trust the same prime broker or prime brokers which encrypted today there are no like single

entity that everyone trusts and indeed that entity needs to have a lot of cap because you need to be able to have enough capital to basically face all of your clients or so if you're a prime and let's say all of your clients have you know five billion dollars collectively you need to have at least five billion dollars and probably a lot more in order to be effective fry broker and so that's been that's been kind of a problem is that there aren't prime brokers I expect we will eventually have prime brokers so to know me for example today is trying to do that they're not there yet but they're certainly moving in that direction I have a lot of players are trying to do something similar the good thing about prime brokers is that they don't really change the market they don't require a change in the market

structure all right you don't need to get the exchanges to agree on anything you don't need to introduce some new standards this can be very much like you know the the exchanges want to do business with the trader would in this case the prime and the traders want to you know have prime services and so it's a very kind of organic evolution of the current market structure so I'm I'm optimistic that that will help mitigate this problem over time the other solution I'm really interested in is the introduction of new layer one assets you know hearing people have been talking about prove mistake and shorting and all these things for years you know for the last couple years you've seen a lot of really smart people say hey I have different ideas and different ideas and how to scale these systems and a lot of

really smart people are working on scaling new they're ones of those the one I'm kind of most interested in is called Solana and they're they're really the team focused on optimizing just the single shard of the kind of maximize throughput in a single system so I have most throughput and lowest latency and one shard and the idea is is that if you can actually do that then you know all of these problems around throughput you know around moving Loney around around low latency all those things get just magically solved you don't have to worry about new trust models new security models cross chart of transactions like there's all these other things are just really untested and just introduced lots of new variables and we just don't know how those things are going to actually work in the real world but we all know today

with very high conviction that like you have a chain and it has assets on it and you shuttle shuttle them around and exchanges you know take the pauses of withdrawals and like that market structure is very well understood by everyone in the market and so I like that that that that kind of approach plugs into the current market structure very well the downside of that approach is you need people to agree to use which is a non-trivial thing to do definitely non-trivial but it is still early days in crypto so we'll see what happens all right well thank you so much for explaining all this craziness it has been a great having you on the show thanks for coming on Unchained hey Laura thank you so much for having me it was a pleasure to be on thanks for tuning in

to learn more about Kyle and multi coin be sure to check out the links in the show notes of your podcast player some of you have heard the weekly crypto news recaps I read on unconfirmed and have asked me for the links to the stories I mentioned there you can also get them delivered right to your inbox with my weekly newsletter which comes out Fridays go to unchain podcast calm right now to sign up Unchained is produced by me Laura Shin with help from her first recording Anthony you NAT Daniel knows Josh Durham and the team at CLK transcription thanks for listening [Music] you [Music]