I read the Binance Blog article, and while I tracked with most of what he said, I found the following statements a little troubling:
- The Terra team was slow in using their reserves to restore the peg. The entire incident may have
been avoided if they had used their reserves when the de-peg was at 5%.- The most stupid design flaw is thinking that minting more of an asset will increase its total value
(market cap). Printing money does not create value; it just dilutes existing holders. Exponentially
minting LUNA made the problem a lot worse. Whoever designed this should have their head checked.- The other fundamental flaw was the over-aggressive incentives. Specifically, Anchor’s 20% fixed
APY to push for (in-organic) growth. Let’s strip away all the fluffy stuff and look at fundamentals.
You can use incentives to attract users to your ecosystem. But eventually, you need to generate
“income” to sustain it, i.e., more revenue than the expenses. Otherwise, you will run out of
money and crash.
For the first two point, at least for the mint/burn portion of LUNA and UST, this is just using LUNA as a type of reserve system. In order for it to work well (as in catastrophic pressure), LUNA must always be at or greater in value than UST in regards to market cap. I agree that additional reserves should be there, but you should always make sure that the stable coin has at least 110% value backing it (not just for those on the market, but for all coins that have ever been minted, including those that have been burned for wrapped assets)(or you should have insurance that is not using your product). Lets be honest, even banks don’t meet this level (that is why you can have a bank run in the first place - although they are backed by insurance, something that was missing in this picture). Minting extends a pool of value in exchange for Burning another pool of value. In this case minting was using LUNA to buy UST and burn (which has the side effect of lessening the individual buying power of the individual LUNA, but keeping the market cap the same for LUNA, then using the new minted coins to buy UST, which burns UST then extracting UST from that pool, which makes UST more scares and therefore each UST gains in buying power)(works in the opposite when it needs to go the opposite direction). This is the same monetary policy that every Fiat currency uses. I am not saying it is the best policy, or that it is not crazy, but it is one that soft currency systems use (in my estimation, many soft currency systems are crazy because they uses market manipulation, and in many systems they do not back it with appropriate assets - for instance you should not back a currency with more of the same currency). The issue with this system was that it had an assumption built in to the code that limited how fast it could burn in the exchange, rather than burn at a rate set by volume pressure (this created the run, and it was a run that had fast and deep effects). So at least for this particular mint/burn issue, it is economically sound (or at least as sound as soft currency economics get)(even if it was implemented poorly, and even if there may be more preferable ways of doing it). Basically the mint/burn balancing is really just moving value from one pool (LUNA) to another pool (UST) or vice versa depending on the need of UST (if it is above peg or below peg). The balancing act can be pretty precarious if one of your trusted systems in the code or validation or the network that connects them fails to work as expected. I do agree though that shifting the market cap positively requires revenue into the reserve system in terms of outside positive assets.
- This article has a nice graphic titled “Arbitrage behavior around UST” of the flow of when mint/burn happens to help peg.
On the third point, I am not a fan of debt and loans at all, nor would I encourage them, but the idea behind the Anchor mechanism was that people deposited; the deposits were loaned; and the interest, the collateral used for staking, and the block rewards made up the incentive to loan back to the depositor. I agree that you should not be promising any amount of interest rates, but it appears that they figured out a system that financially could be feasible. There are a lot of formulas in how it figures out each of these (they are in the white paper), and I have not looked at them extensively, but in general the idea would work (and again is fairly similar to what banks do when they loan money). But, again, unlike a bank that has insurance, you would have to have enough so that if everyone pulls out of their accounts at the same moment, the system does not fail (since there is no insurance, and since if they do, it could crash the whole system).
- This article has a nice graphic titled “Wen Anchor?”
For Terra, the reserve was not high enough (both reserve meaning the market cap of LUNA + outside reserves of the TFG), they did not respond quick enough, and even if they had, given the sheer depth of money dropped on the market, the faulty mint/burn mechanism was too slow in burning.
- To complement the linked articles graphic depictions, here is the white paper describing the system
It was a fascinating article (particularly for a central exchange that has its own stable coin).