RE: Algo Stablecoins Won't Save You

[sorin.cristescu]

By "The main use case is" I didn't mean that's what they are used for now. I meant "that's what they could be most useful for" ... provided they could get regulated and mass adopted. I agree that now, they are used as you say for "greasing the wheels" of the crypto financial machinery.

About algorithms, you are wrong: youcan program them to check liquidity and stop redemptions if liquidity is out of band. I don't understand your reasoning when you write:

"the rate of redemptions, which NO Algorithm can force, either way, you can spoof it but there is no hard rule to stop it."

This is code, for Pete's sake, you can force whatever you can think of ! Why couldn't you program a rule to simply prevent redemptions, as long as redemptions are done by the algorithm ? You send your troubled algo stable to the smart contract to get back whatever the backing is. The algo checks a formula. Formula out of band ? Well, you get back the answer "sorry, the bank teller is closed, withdrawals are temporarily unavailable, check back later, don't call us, we'll call you !"

Comments

[chekohler]

How am I wrong? UST proved that algorithmic stablising both with native and 3rd party assets is very much valuable.

LOL that's not how code works mate, because you're dealing with an oracle problem, if your coin is traded on a market and it has to reflect the market price or it will be clipped until your treasury is depleted.

Do you know these things work? because I do, a algo stablecoins market price is NOT set by the smart contract, it's set by the market, the market price feeds come from external service providers like exchanges. The more lisitngs you have the more exposure and the less control you have over managing the peg.

Its NOT a self contained system, so when someone shorts your stablecoin or collateral backing it, on one of the data providers, arb traders come in and clip the premium on exchanges yet to catch up and then the on-chain has to correct flooding those markets and being the floor, diluting its collateral and breaking the peg.

[sorin.cristescu]

Can you walk me through this step by step? So say my algo A is traded on markets N1, N2, ... against other cryptos, including other stablecoins, say collaterlized like USDT or USDC.

Next an attacker borrows my algo A in a large quantity. However, the quantity he's able to borrow depends on M (the monetary mass), id est how much there is available. If you want to sell A (trying to break the peg), there needs to be someone out there to lend it to you. If the total monetary mass of A is reasonably controlled, how much can the attacker borrow ?

So now say the attacker has borrowed as much as he could find and sells enough A on the market N1 that the whole order book there is cleaned and the N1 market lists A as $0.01 because there are no more buyers on that market. Some people may call "break the peg" but that doesn't qualify, in my opinion. The smart contract might or might not be programmed to defend the peg on each and every market. Why would it ? That would be stupid and open A to manipulation. At any rate, that is a CHOICE of the programmers. I can have oracles, I can look at markets. So what ? What forces me to engage in defending the peg on individual markets ? The peg is guaranteed on the blockchain, nowhere else

"A" fails and the peg is broken only when there permanently is no way to redeem it at par from the issuing smart contract on the blockchain. That is the central authority which has ultimate responsibility over the peg, not an individual market, whichever that my be, not ANY number of individual markets.

The fact the UST failed proved that it was not programmed right, NOT that no algo can be programmed right and all are doomed to fail.

So what happens next, can you please indulge me with an exploration of the mechanics ?

[chekohler]

I literally explained it already, check above, if you still don't understand here is an in-depth deep dive into the fundamental issues of algo stablecoins, please read it

[sorin.cristescu]

Your explanation is based on an underlying assumption I do not share. Your assumption is that

[...], a algo stablecoins market price is NOT set by the smart contract, it's set by the market, the market price feeds come from external service providers like exchanges. The more lisitngs you have the more exposure and the less control you have over managing the peg.

An algo stablecoin "price" is indeed set by the markets. Just like the dollar price expressed in euros is set in the Forex markets. But if the goal of your stablecoin is to facilitate payments between individuals and businesses (rather than merely enabling meaningless financial shenanigans), then the price (expressed in, say, dollars) set by a market at any given moment is mostly irrelevant to what should be the main mission of a stablecoin, facilitating peer-to-peer transactions in the "flesh" economy. And that, at least as long as there is a credible theoretical path of returning to parity.

You need to see the evolution in time of the relationship between the stablcoin and its peg as a continuum. At one extreme, you have nearly microsecond-level parity, and then you need to manage the peg as you say, which can be very tenuous. At the other extreme, you have a "permanently unstable stablecoin" such as the SBD which lacks appropriate stabilization mechanisms and therefore sits most of the time far away from its peg, to the point where it completely lost credibility as a stablecoin.

In between, you have stablecoins such as HBD which are not that stable but still reasonably so.

In terms of stability, I measure the quality of a stablecoin by its ability to stake a serious claim at being "close to the peg most of the time". But one can be more or less strict in that definition and the stricter one is, the more difficult managing the peg becomes.

The article you quoted is intellectually shallow and mostly useless.

[chekohler]

You're just spitting out word salad trying to obfuscate my points, let me cut this debate short because you're talking yourself in a circle.

Here's a simple issue the algo stablecoin backed by a native asset, relies on the market cap of the asset backing it, the market cap is set by the amount of buying pressure of the asset, if there isn't enough buying pressure for the underlying the market cap quickly evaporates as in the case of Luna, Bean and many others prior.

once the underlying breaks, there is nothing to support the debt issued in stablecoins and it collapses.

You might think you're safe because the so-called market cap shows backing, but the volume to maintain the market cap is razor-thin. The issue with algo versus real fiat is that central banks have unlimited liquidity and a bag holder in their citizens to keep shoring it up.

If you're going to keep talking around the point, then there's no point continuing the conversation and you can happily stay in your delusions.

[sorin.cristescu]

Despite my "word salad" you seem to finally get my point, or perhaps you just fumbled upon it by chance.

You rightly moved the focus from "maintaining the peg" to the comparison with fiat currencies and the conditions of "collapse" (especially in the underlying asset).

That's precisely what I've tried to point out:

1. The main goal of a stablecoin is NOT (or should not be) to "maintain the peg", just like the goal of a currency board is not to "maintain the peg" but rather to heal the underlying economy. "The peg" is thus merely a means to the real end, which is to facilitate transactions, to stimulate an "economy" (crypto economy), to allow for a smooth integration between the "crypto economy" of the underlying asset and the outside fiat economies. Saying that "the purpose of a stablecoin is to maintain the peg" is akin to saying that "the purpose of the dollar is to maintain parity with the euro" (or the other way around). A credible algo stablecoin creates or strengthens the economic bridges needed for the cryptoeconomy of the underlying to "trade" with the much larger fiat economies. This is the yarstick one should use when looking at how good an algo is
2. Yes, an algo can collapse and most present-day algos are ridiculously easy to make collapse ...because their underlying cryptoeconomies are lopsided and tiny. But even fiat currencies collapse regularly, despite being backed by central banks which have "unlimited liquidity" and their citizens to hold the bags. They can "collapse" both internally (in my definition, of an economic collapse) AND "externally" (in your previous definition of "breaking the peg with whatever currency they had a peg with"). Think at the recent collapse of the Sri Lankan economy, at the "external" collapse of the Turkish lira, at the regular collapses in the Argentine peso, at the Greek and Cypriot crises back in 2011 - 2013, etc. Further back, think of the 1991 "collapse" of the British pound when it broke the peg to the ECU that the British government and the mighty Bank of England, the famous "Old Lady" were committed to defend! Economies collapse, fiat currencies collapse despite central banks with unlimited pockets, despite having millions of users. So yes, algos are going to collapse too, and more often than fiat currencies, but NOT because they are inherently worse than fiat currencies, but because their underlying economies are much, much smaller and a lot more imbalanced.

In short, there is nothing inherently bad about algorithmic stablecoins. They shouldn't be expected to be more stable than what the size and health of the underlying cryptoeconomy warrants. A "bad" algo is one that ignores the M * V = P * T equation of central banking and has no "circuit breakers" built in. But good algos can be designed. Even they are not going to be insulated from a crash in the underlying cryptoeconomy though, so ... simply don't focus too much on stability, that's mostly a marketing decoy. What you should focus on is whether the algo is spurring investment and growth in the underlying cryptoeconomy.

By that yardstick an algo should be judged, not by how good it is at maintaining the peg. UST was poorly designed with no "circuit breakers", there was no underlying LUNA economy to speak of (aside from financial engineering), so yeah. But that doesn't mean "there can be no good algos"

[heretickitten]

Wow, both of your arguments had a lot of really good points. I really think that both of you together make a powerful team in breaking down the fundamental obstacles of stablecoins.

I'm surprised that "word salad" came up! Ahaha @sorin.cristescu, you are talking about really complex stuff, but I want to really stress to @chekohler that what @sorin.cristescu is saying is NOT word salad. It is just really dense, and hard for humans to talk about.

So much math...!

I think @sorin.cristescu truly understands what's going on here, but the funny part is that I think @chekohler does too. It's just an issue of collaboration, and fitting both of your ideas together to build knowledge as to how to engineer a more powerful stablecoin.