Not Legal Advice 11/9/19 – Dai hits $100 million; Crime doesn’t pay; FBI Director Wray speaks to Congress

Welcome back to this week’s edition of Not Legal Advice!

Once again, I’ve been remiss in typing up my weekly newsletter on a weekly basis due to travel – one of the downsides of solo practice is that one has no minions to dispatch to the far sides of the world – and this time, to San Francisco, where I did a panel with the inimitable Josh Stein of next-gen digital securities firm Harbor, among others, at SF Blockchain Week. Well done to the organizers for putting on a great conference.

This week’s a short one, as there really hasn’t been a whole lot in the last 14 days that I’ve found particularly interesting:

  • Dai hits $100 million in outstanding CDP contracts; crypto bros still don’t understand risk
  • Crime doesn’t pay: an update on the Brian Haney arrest
  • FBI Director Christopher Wray talks crypto to Congress.

1) Dai hits $100 million in outstanding CDP contracts; crypto bros still don’t understand risk

The Block reports:

The number of outstanding Dai has reached the protocol’s built-in “debt ceiling” of 100 million— an all-time-high for the nearly two-year-old stablecoin project. CDP 15336 minted the Dai that boosted the outstanding supply to its limit.

MakerDAO, the issuance platform behind Dai, had an original Dai debt ceiling of 50 million, which was raised to 100 million in July 2018. The MakerDAO team and community members plan to execute a governance vote this Friday to raise the debt ceiling by an additional 10-20 million.

Yes, a “decentralized stablecoin protocol” has “governance votes.” I’m not sure either.

The Block continues:

Early last week, the Maker Foundation announced that it will be rebranding its Collateralized Debt Position (CDP) in preparation for its November 2019 Multi-Collateral Dai (MCD) release. The new user interface of the Maker Protocol after the release of MCD will label CDPs as “Vault.”

What is “Dai,” I hear you ask? Dai is a so-called “stablecoin,” a cryptographic token which is designed to always hold a peg to a fixed, external unit of account – in Dai’s case, the U.S. dollar.

Dai accomplishes this, we are told, through a series of smart contracts on the Ethereum blockchain which issue the Dai coins and lock up an amount of Ether in excess of the Dai as collateral to back the “loan” which has been issued. This was known as a “collateralized debt position” but, perhaps because the organizers of the scheme have some dim awareness of the regulatory consequences of issuing securities which are backed by collateral pools and making them available for public sale, the Dai people are now changing the terminology of these smart contracts to “vaults.”

CDPs/Vaults expire in one of two ways. First, someone can pay back the Dai debt plus interest, which the scheme promoters misleadingly refer to as a “stability fee,” at which point the CDP dies and the locked Ether in collateral is returned. “Stability fees” can only be paid in MKR, another shitcoin which was issued by the original scheme organizers. In the alternative, if e.g. the value of the collateral pool is impaired, the CDP may be liquidated and the collateral used to repurchase Dai from the marketplace to ensure all Dai are backed by a quantity of Ether with a dollar value that is greater than or equal to the dollar value of all Dai in circulation.

How this works is a little complicated, but the team over at Reserve summarizes it well:

The process by which this happens is somewhat complicated. It involves two different on-chain auctions that try to raise enough capital to make the CDP debt free. To fully understand the process, you may have to spend some time thinking it through after reading it. If you don’t fully get it, don’t sweat it: full understanding is not necessary for following the rest of the analysis.

Here is how it works: first, a “debt auction” tries to repay the CDP’s debt through MKR dilution. The debt auction buys Dai, paying with newly minted MKR. The Dai is burned, to cancel the CDP’s outstanding Dai debt. The purpose of the debt auction is to ensure that the debt is repaid even if there is insufficient collateral in the CDP to repay the debt.

Simultaneously, a “collateral auction” buys MKR with the CDP’s collateral. The collateral auction sells enough collateral to cover the debt, accumulated interest (called the “stability fee”) and a liquidation fee. In Single-Collateral Dai, the liquidation fee is 13% of the collateral in the CDP — that is, they take 13% of the user’s locked up collateral capital when a user’s CDP gets auto-liquidated. The smart contract finally returns the remaining collateral to the CDP holder and burns all purchased MKR.

This is all, ultimately, just a complicated and extremely long winded procedure to repackage exposure to Ether in such a way as to drive demand for the MKR token. It is really only useful if you either (a) have a bunch of Ether and want to lever up and go long on more Ether or (b) you want to use a smart contract to obfuscate the source of your funds, which is something you really should not do.

The entire system is vulnerable to adverse movements in both ecosystems. As Dai is now expanding with “multi-collateral Dai” which is backed by many different kinds of coins, soon it will be vulnerable to adverse movements among a range of different cryptocurrencies.

The risk has not gone away. It has merely changed form. DeFi Bros have difficulty understanding this. e.g.

Current mood:

Long have I had suspicions about whether Dai is for real. My skepticism about the scheme before it launched was reinforced when the loss of one bot on one sketchy overseas exchange operated by an unnamed “third party market maker” resulted in the Dai dollar peg not just breaking, but shattering, until the bot was restored. Put another way, the brilliance of the Dai stablecoin system – at least back then – wasn’t the reason Dai held its dollar peg. A bot was.

And this isn’t me saying this. It’s the founder of MakerDAO, Rune Christensen.


Put another way: back in 2018, the volume on busiest market in Dai by far, on a $1 million trading day, dropped to $300 when a single bot went down.

DeFi Bros struggle to understand why this is also problematic.

“In reality Dai remained stable on all other exchanges” is a worthless argument in that context. The context being that we just discovered that a huge chunk of the market was not bona fide trading. If most of the volume of the coin can be traded by one bot, were we wrong to trust the numbers before the bot was discovered? What reason do I have to trust the numbers now? What reason is there to trust the rest of that volume on other exchanges? How do I know they’re legit?

I’m not saying here that the MakerDAO team knows anything about these bot operations. Far from it. Indeed, Rune refers to a “third party market making bot.” A third party with whom I should greatly like to speak who, apparently, never decided to reveal him or herself to the world.

I don’t know who operated the bot. I also don’t know how the bot operator communicated this information about bots on Bibox (the exchange) to the wider world. I don’t know why they were spending all that time wash trading on Bibox or what they stood to gain from it. I don’t know why the wider crypto community and stablecoin bros alike were not the least bit distressed by this event. All I know is that it happened, and I have never seen an explanation for why the scheme should have worked when that bot was up yet it broke catastrophically when that bot was down, as occurred in January of 2018. In the last two years, journalists haven’t followed up.

What I do know is that there’s no magic or innovation in wash trading around a fixed price point to make a market look real, on the off chance that is indeed what’s going on.

Charts of derivatives that are repackaged exposures to Ether should look like they are repackaged exposure to Ether. Dai does not. In the eyes of a dispassionate observer this should raise questions about market integrity. When Dai first broke its peg in early 2018, daily trading volume was around $1 million and the total market cap was around $3 million. Now, daily trading volume has reached highs of up to $50 million. All of which is to say that to the extent that bot training wheels first put in place back in the day are still in place, those training wheels are being asked to hold up an increasingly large rider and will be placed under greater degree of stress.

I stand by my prediction, first made in 2017, that Dai will eventually implode. But for the bots, after it fell on its face in 2018 it would have stayed down, just like previous collateralized stablecoin schemes such as BitUSD and NuBits, both of which failed (in BitUSD’s case, it failed after five days). The bigger the scheme becomes, the more difficult it will be for Dai’s training wheels providers – mysterious figures in the shadows, operating bots that generate volume for fun and profit – to hold back adverse market movements.

If we learned anything about risk-obfuscating schemes from the global financial crisis, we know this: the bigger they get, the harder they fall.

2) Crime doesn’t pay: Silk Road trafficker pleads guilty

Breaking the law is bad and dumb. Breaking the law with cryptocurrency is exceedingly dumb. Hugh Brian Haney was arrested in July of 2019 in relation to Silk Road activity dating back to 2012; this week he pleaded guilty to two charges and now faces a maximum of 30 years in prison.

3) FBI Director Christopher Wray talks crypto to Congress.

Which brings us to our next news item. An interesting fusion of the crypto-means-cryptography universe and the crypto-means-cryptocurrency universe happened in Congress this week. As reported in CoinDesk:

Wray noted encryption is touching every aspect of emerging tech such as instant communications:

“Whether its cryptocurrency, whether it’s default encryption on devices and messaging platforms; we are moving as a country and world in a direction where if we don’t get our act together money, people, communication, evidence, facts, all the bread and butter for all of us to do our work will be essentially walled off from the men and women we represent.”

First, to clear something up: most cryptocurrencies DO NOT encrypt communications. Bitcoin is chief among these crypto-critters-that-don’t-encrypt-transactional-data. Bitcoin really only shields one bit of data – the private keys of the users – from government surveillance. But it doesn’t stop the government from tracking what different keyholders do and how funds on the Bitcoin blockchain move around.

Some privacy coins, such as Monero or ZCash, do encrypt transactional information. Opinions as to which method of encryption is superior and e.g. the merits of ZCash doing a weird international math druid ritual to generate the coin’s SNARK public parameters are legion and do not bear repeating here. What does bear repeating here is that it would be very foolish to presume that these encryption methods will be secure forever.

Second, we should be cautious before we throw encryption out the window. Crypto that can be defeated by the FBI can be defeated by anyone (which isn’t a dig at the FBI, it’s just reality – Fort Knox wouldn’t be safe if it had a secret, unguarded, publicly-accessible back-door, and neither is code under the same circumstances).

I have yet to watch the entire hearing (and will likely do so tomorrow) but from this little, brief tidbit, what’s interesting from my point of view is how cryptocurrency and cryptography are starting to crop up in the same breath. And, unlike the 2010s where the interesting tech was about sharing cat pictures, virtually all of the interesting tech I can think of operates in this weird zone of enabling dissenters, since platforms like Twitter and Facebook are essentially tools of the hard-left anti-Trump #resistance establishment now.

As my friend and Israeli secret agent Maya Zehavi observed:

And I added:

What a time to be alive.

Here’s a picture of some marmots, licensed under the Pixabay license.