Hacker News Reader

Interesting, this reminds me of how perpetual motion machines can be dismissed out of hand by citing the 2nd law of thermodynamics. As with plausible-seeming perpetual motion machines, the key to a popular stablecoin seems to be to make it complicated enough that it's difficult to analyze from first principles.

I think you've hit the nail on the head. Stablecoins exist in the non-crypto world [1] and they all are slaves to the Trilemma - there are trillions at stake if someone could figure out a way around this, but, much like the 2nd law of thermodynamics, you can't make the math work. Crypto doesn't offer a technology solution to this issue since it's not a matter of tech, but rather financial math.

If you're a fan of crypto and still reading, here's a hint at the next financial innovation after pegged currencies that you can try to replicate in the cryptoverse: XDR [2]

[1] https://en.wikipedia.org/wiki/List_of_circulating_fixed_exch...

[2] https://en.wikipedia.org/wiki/Special_drawing_rights

Special Drawing Rights was one of the stablecoins on the (recently defunct) Terra platform.

You present yourself as deeply knowledgeable about these topics, I’m surprised you missed it:

https://www.investopedia.com/terra-5209502

I don't read parent's comment as advocating the creation of a stablecoin pegged to XDR, but rather repurposing the principle of targeting (flexibly) a basket of underlying reserve assets as a unit of account in cryptoland.

But choosing the same basket as XDR would still satisfy that goal, and picking different currencies would be a trivial difference. And if that's the extent of the idea, the parent could have just said so, rather than taking on this coy mystique of "oh, I'd reveal my deep insight, but, novices like you have to do the research and figure it out yourself, else you won't be ready for it".

DAI sacrifices 3 [+] so it should be able to meet 1 and 2, though, right? Much the same as GUSD maintains one by obligating itself to honor redemptions.

[+] Dai's monetary policy is not autonomous; new units can only be issued for sufficient collateral and must be destroyed when value falls too low, a constraint that would make a central bank not be regarded as autonomous.

Hard to see how the GP missed this obvious fact...

In terms of free capital flows with no fixed exchange rate, this is the "it works until it doesn't" part of the problem. DAI only survived the recent run because they had a rainy-day fund in their "PSMs" or Peg Stability Modules.

The PSM takes your USDC and mints new DAI and gives it to you. When DAI is >$1 you can make an instant profit doing this. (take 1 USDC, trade it for 1 DAI that you can sell for $1.001.)

The PSM takes your DAI and gives you USDC (while stocks last.) When DAI is >$1 you can make an instant profit doing this. (buy 1 DAI trading at $0.999, and swap it for 1 USDC worth $1.)

Notice the issue? The PSM within DAI is contingent on having enough USDC (proxy for USD) to keep the peg. When that reserve runs out, the PSM mechanism breaks and the peg fails.

To suggest otherwise is financial alchemy.

If the auctioning functions correctly, there will always be more reserves as compared to Dai. If the "reserves run out", then there would be no Dai to buy.

But DAI doesn't control USDC, which means that they cannot control the reserve piece of the equation. This is the independent monetary policy + free capital flow piece of the trilemma.

The auction function doesn't work if USDC suddenly isn't worth 1 USD, because then you're facing down a reserve that isn't worth "more" than DAI, even though the algo still thinks it's in balance.

Basically you've tied DAI to USDC, which in turn is tied to Coinbase, which in the event of a bank run/bankruptcy event will freeze assets and cause massive capital flight causing people to sell their USDC for less than 1 USD as they flee for safety. This causes DAI to become unbalanced because their reserve of USDC isn't actually worth what it should be in USD terms, causing DAI to also lose the peg in terms of real US dollars.

I'm telling you, there isn't a way to make this stuff work. It's financial alchemy.

That is a legitimate concern, but irrelevant to the trilemma you've mentioned.

If USDC loses value, those vaults will be liquidated and the collateral will shift over to alternates - ie. eth. The scenario you are mentioning only plays out if USDC falls so much (and quickly enough that the auction bots don't respond) that it cancels out the overcollateralization from eth. given that dai is 166% collateralized, it wpuld have to drop a ton, which is difficult to imagine when it is credibly backed by real assets.

But the liquidation of USDC represents the flow of real assets OUT of the ecosystem, thus rapidly inflating the price of ETH and any other cryptocurrency denominated in dollars. Aka if you start selling 2 USDC for 1 USD, then the price of ETH goes from $3000 to $6000 overnight. BUT those dollar signs in front of the price aren't US dollar signs, they are USDC dollar signs. Which means that I can engage in massive arbitrage between exchanges.

Exchanges respond by sealing themselves off to protect their $3000 ETH from the $6000 fake ETH trade prices, thus creating a massive freeze in the market, thus wrecking the redemption system that underpins DAI.

It's very, VERY important to realize that all of these things are interwoven. There's no such thing as "USDC collapsing" without it also taking all the crypto traded on Coinbase with it.

... no. The price of Ethereum in USDC is irrelevant. There is no arbitrage to be had. If the price of USDC falls, you have to spend more and more USDC to mint one Dai because Dai is tied to the price of the dollar, not USDC.

Explain to me how you arbitrage USDC falling out of Dai?

I really encourage you to read up more on this. There are real risks that come with Dai as a stablecoin (including eth or USDC collapsing very quickly) but they are not the risks that you are identifying.

My background is in international finance, and therefore when I promise you that DAI cannot defend a virtual peg without holding large amounts of actual underlying US treasuries or physical USD, I'm only making that promise because it's been true for every currency that has ever attempted to maintain a peg with the US dollar. Maybe it's "different this time" but I sincerely doubt it.

I likewise encourage you to think about this from first principles and read up on the mechanics of a currency peg. You MUST be able to quickly sell T-bills on the market swiftly to defend your currency peg, otherwise you will see your currency break from the pegged range. DAI is shifting the reserves from directly holding USD or T-bills to holding USDC instead, which isn't the same thing as USD and WILL cause problems for DAI.

empirical mathematic principles

I wish this was a joke. How about you try to collect the $1m Millennium price with your proof of the Riemann hypothesis based on "empirical mathematic principles"? To be 100% clear about this: there's no such thing as empirical mathematical facts, there has never been and there will never be, not in this universe or any other. No matter how many (finitely many) roots of the zeta function you collect, they will never constitute a proof.

To your point: you do realize that stablecoins like Dai have given up monetary autonomy, right? The amount of Dai in circulation is governed by users/the market, not the protocol or a central bank. You seem to be acting in bad faith, because you deliberately only discuss two points, but according to this theory, a stable exchange rate and capital mobility are indeed compatible if you're willing to let the market dictate your monetary policy.

What? All of these coins obviously sacrifice number 3....

And what is an "empirical mathematical principle"?