The Death of Terra
The Rise and Fall of a Great Mission
There are so many interesting projects in the crypto space - this must be what it felt like in the late ‘90s/early 2000s when the internet was new and had a ton of excitement. Some projects are loved, others are hated, others are hated for being loved. Terra was hated for being loved.
The most interesting heckle towards Terra happened around March 14th this year when a $22 million bet on LUNA went down. Algotrading tweeted that he basically had reservations against LUNA and wanted to bet someone $1 million that the price of LUNA would be lower in a year. Do Kwon, the founder of LUNA, responded and the bet was on. They both worked out the terms in real time on twitter and each sent $1 million to a wallet address owned by Cobie, who would serve as the middleman. The final bet was that if on March 14th 2023 the price of LUNA was above $88, Do Kwon would win. If its under, Algod wins.
Around the same time Gigantic Rebirth enters the scene and bets $10 million. Do Kwon accepts, and the bet was on.
The great thing about public blockchains is that its easy to see whether these guys really put the money in or not. Seen on Etherscan:
Wallet address: 0x4Cbe68d825d21cB4978F56815613eeD06Cf30152
Algod $1 million transaction can be seen here
Do Kwon $1 million transaction can be seen here
GCR $10 million transaction can be seen here
$10 million match from Do Kwon can be seen here
Terra overview
Terra is “a family of programmable digital currencies.” Terra set out on a great mission; to create stablecoins for various regions in the world. It seems as though the vision was to become a stablecoin generator machine for the world. Terra had a US dollar pegged stablecoin (UST), a Korean Won pegged stablecoin (KRT), MNT pegged to Mongolian Tugrik, and SDT pegged to IMF Special Drawing Rights. Terra’s flagship stablecoin was UST.
Why are stablecoins good?
Stablecoins allow you to spend fiat money in crypto form. For example you would use UST over USD because;
UST is easier to spend - can be spent anywhere in the world, outside of traditional banking rails
Lower fees vs traditional banking rails
Faster settlement vs traditional banking rails
Blockchain recorded transactions allowing
ease of audit
automation of transactions with smart contracts
Store your money in a digital wallet and have access to other crypto applications including DeFi.
In the case of Terra, you had access to the Terra network’s savings protocol called Anchor, which at a stage allowed you to earn 20% yield on your savings.
How Terra works
Terra is an algorithmic stablecoin that is not backed by a fixed deposit of fiat, crypto, or other assets. Instead the value of Terra’s stablecoins is maintained by an arbitrage incentive made possible through a token called LUNA. LUNA is a reserve asset backing UST and acts as a stabiliser for the currency peg, absorbing the volatility of UST. LUNA also serves as a governance token for the Terra ecosystem.
For any transaction using any of the Terra ecosystem’s stablecoins, a small transaction fee is paid out to LUNA stakeholders. The more people that use UST, the more fees generated and accrued to holders of LUNA, making LUNA more valuable.
How it works is that users burn LUNA to mint Terra, and burn Terra to mint LUNA. All made possible via the Terra protocol’s algorithm. Remember stablecoins need to have a 1:1 peg with the underlying currency. Here are 2 scenarios explaining UST and LUNA;
In an expansion: When the price of UST is too high, relative to its peg, supply of UST is low and demand is too high. Because Terra seems attractive, users will be incentivised to burn LUNA and mint UST as more people want UST. Minting UST makes the pool of UST larger, until supply matches demand.
Users mint more UST with burned LUNA until UST goes back to its target price if $1.
Because LUNA is burned, the LUNA pool gets smaller, increasing the price of LUNA.
In a contraction: When the price of UST is too low relative to its peg, supply is too large and demand is low. UST is then burned, which decreases its supply, and LUNA is minted from burned UST. The decrease in UST’s supply causes scarcity and the price of UST increases. The LUNA pool increases and LUNA’s price goes down.
Arbitrage opportunity
Because Terra’s protocol enabled users to always trade 1 USD worth of LUNA for 1 UST, and vice versa, this incentivised users to maintain the price of UST, creating arbitrage opportunities as market forces played out. 2 examples:
If 1 UST is trading at $1.01, you can burn LUNA for UST to the value of $1 per LUNA token to mint 1 UST, then sell that UST for $1.01 - profiting $0.01. When you sell UST for USD you’re effectively “exiting” the Terra ecosystem and therefore the UST goes back into the UST pool. This increase in supply, reduces scarcity, and reduces the price of UST. This will happen until all of the arbitrage opportunity is gone and 1UST = $1 again.
Alternatively if 1 UST is trading at $0.99, you can buy 1 UST for $0.99, sell your UST for LUNA to the value of $1 per LUNA. You can then sell your LUNA and profit $0.01. Users can keep doing this until enough UST is burned such that the arbitrage opportunity disappears, and UST returns to its peg of 1 UST = $1.
The Terra ecosystem couldn’t successfully function as a currency however without having an underlying asset in reserves. For Terra this was the Luna Foundation Guard (LFG).
LUNA Foundation Guard
LFG is a non-profit organisation based in Singapore, and acted as a reserve organisation, backing the Terra ecosystem. LFG was mandated to purchase assets over time (mainly bitcoin and ETH) to build up reserves for the Terra ecosystem.
Terra Transformation
In March 2022 there was approximately $4 billion in bitcoin, ether and LUNA sitting in the Luna Foundation reserves.
The Luna Foundation Guard (LFG) decided to start buying more bitcoin to back Terra and embarked on a strategy to back UST entirely with bitcoin, and gradually relying less on the algorithmic functioning of UST. Terra started implementing a system where instead of burning 100% of LUNA when people seek UST, Terra burns 60% of LUNA and the 40% is used to purchase bitcoin.
Terra became an aggressive buyer of bitcoin using its LUNA reserves, buying bitcoin on price dips
The idea for buying bitcoin to try have the Terra ecosystem eventually entirely backed by bitcoin was based on the fact that currencies generally need reserves to be deemed valuable.
Why does a currency need reserves?
Currencies (money) are a medium of exchange with an underlying value. If I give Kim a $100 note - the note itself doesn’t mean anything. The underlying value of the note (since the advent of the fiat system) is that Kim can redeem the $100 (by law according to the United States government) by either
buying something else with the note, or
creating more money with the note (save it and earn yield, or invest it and earn a return)
The ability to redeem a currency for something else has over time solely become a function of whether there is an open market for the currency. If a currency is instead backed by an underlying currency/asset, then someone can redeem the currency for the underlying asset from the issuing party (the government). This method of currency governance is extremely risky however because you would need to ensure that if all people wanted to redeem all of the currency in circulation, you can guarantee that you’ll make them all whole by providing each of them with the underlying asset at the same price, in order to avoid panic and rush selling.
In this light it is easy to recognise why to date no one has been able to successfully implement a currency peg. The closest success story is that of the PBOC and managed pegs. The PBOC sets a daily target exchange rate and allows the currency to fluctuate within a few percent of that rate. Anything more or less then the Bank steps in to stabilise the price. Here are some examples of famous currency pegs that broke over the years
Gold Standard: US dollar peg to gold was abolished in 1971 to prevent selling pressure on USD from nations looking to redeem dollars for gold.
Thai Baht: broke from USD peg in 1997 as Thai government ran out of foreign currency reserves
Chinese Yuan: China revalued its currency higher against the USD in 2005, amid mounting market pressure to more accurately reflect the sate of its economy.
Swiss Franc: In 2015 Switzerland broke its peg of 1.2 to the Euro due to a weakening euro against the dollar. A weaker euro against the dollar meant a weaker franc against the dollar.
Essentially when you peg a currency to another currency the bet is that the two completely different systems will work together and as the underlying currency (currency to which another currency is pegged) gets stronger so will the pegged currency.
This is exactly what stablecoins are betting on. In the case of Terra, deciding to add more bitcoin to reserves was an attempt to strengthen Terra’s defence capabilities, and to add a little more creditability to the Terra ecosystem; given the belief in bitcoin and the highly performant track record that bitcoin has been able to build to date.
Problem with bitcoin reserves
Bitcoin volatility:
Being a relatively nascent asset, bitcoin is still extremely volatile. This puts your reserves at risk and can land you in a position where you are unable to defend the peg in an attack because if the value of your reserves falls, so does the belief that you can make people whole if they want to redeem their UST.
Arbitrage:
Adding more bitcoin arguably made Terra more vulnerable. We all know that whales (real ballers) are generally against bitcoin so seeing a newcomer like Terra, with a “tricky” product, add a volatile asset to the product functioning was bound to entice whales to speculate using LUNA.
Here’s a summary of how UST was killed.
The attack
Around May 10th, 2022:
UST got hit with an attack to try and break its 1:1 peg with the US dollar
Some say the attacker was Citadel and Ken Griffin - the guy who had to bail out Melvin Capital after they were short-squeezed on Game Stop (GME) by a group of retail traders
The attack worked and the UST peg broke from $1 to $0.64, before bouncing back to $0.90.
This was really a canary-in-the-coal-mine because stablecoins can’t have their pegs break to such low marks (not even Tether, with all the criticism against it, has ever broken to such a low mark).
The Luna Foundation Guard had to sell reserves to try support the peg. This hurt the rest of the crypto ecosystem because large sell orders put downward pressure on the price of BTC and ETH.
Playbook
Sourced from The Milk Road and Route2FI
The attacker borrowed 100 000 BTC from Gemini, which they shorted
The attacker also borrowed about $1 billion in UST from OTC trading desks
The attacker then cleared out liquidity pools on Curve Finance. Approx. $350 million worth,
Now there was no more liquidity on Curve and the UST:USD peg starts falling to around 0.97-0.98. A little more fear sets in and people start to withdraw their reserves in Anchor.
The price of LUNA starts tanking, but the attacker still has about $650 million in UST which they start selling on Binance. This causes a massive de-peg, and the LFG steps in by selling BTC to buy UST and hopefully restore the peg.
At first LFG and another UST backer called Jump started selling more crypto assets (mainly ETH and BTC), again to buy UST and try bring the peg back. They succeeded in bringing it back to $0.91
The attackers then started dumping UST, creating a death spiral and LFG realises that they can no longer defend the peg, so they let it bleed
In 48 hrs, 50% of all funds were withdrawn on Anchor
Result
The peg went all the way down to $0.31
Price of LUNA dropped from $90 to $24
What I think was really the objective was shorting bitcoin. So the attacker likely levered up to profit highly from that short, which was made possible because
the attacker cleared liquidity pools causing panic.
The attacker sold UST, increasing supply and reducing the price, causing fear, and causing
the LFG to sell their bitcoin in large orders to defend the UST peg, meaning
the price of bitcoin falls dramatically, and
the bitcoin short pays out
Terra introduced some safety measures:
$1.5B loan from their reserve - $750M of BTC was sent to over-the-counter trading firms to help protect the UST peg, $750M worth of UST to accumulate more bitcoin.
Terra moved another $1.4B from their reserve wallet to protect and bring back the peg
The peg came back up to around $0.80. Still obviously not good enough for users.
The price of LUNA then went back up to $31.
But by that time all trust in the network was lost and at the time of writing, UST is effectively dead - with the peg at around $0.09.
In the weeks leading up to this event the LFG reserves were about $4 billion, made up of different crypto assets. After this debacle, the reserve had no bitcoin left, and only about $195M in other crypto left.
The defence wasn’t enough as people realised the vulnerable state in which this whole episode left Terra. The drastic fall in reserves meant that Terra would be unable to defend another attack. In the space of about a week or so from that defence attempt we saw:
LUNA hyperinflation: to defend the UST selloff the Terra protocol minted so much LUNA that the supply of LUNA went from around 725 million on May 5th, to about 7 trillion by May 13th
The death of Terra
Coinmarketcap: UST peg completely broken
Coinmarketcap: LUNA went from $116 a coin to $0.
Success of Terra
At its peak:
Many Korean retailers used the Terra network through a payment application called Chai to accept payments in KRT
The top 14 Korean banks, the number 1 convenience store, the largest gaming publisher, and the biggest book store in Korea were all transacting in Won through the Terra blockchain
Over 2 million users spent $1 billion a year in KRT for transactions
In Mongolia, over 40 000 people a month were using Mongolian Tugrik on the Terra blockchain to buy goods using a payment app called Mimi Pay.
At its peak, UST was the fourth largest stablecoin in the market behind Tether, USDC, and BinanceUSD.
At the height of its success, there was approximately $100 million to $200 million worth of demand for UST per day.
Lessons learned
With regards to negative comments about Do Kwon
I don’t know him personally so I’ll keep emotions out of it
It takes a lot to build something of value and even more to build a multi billion dollar project. Most hating on Do Kwon likely don’t have a leg to stand on.
Not cool to kick someone when they’re down
Cocky gent, but quite skilled and deserves credit for building Terra
His overconfidence may have lead him to overlook an important feature (no reserves!) as LUNA became too big, too quickly
This humbling event might help him do greater things which I hope he does. Everyone loves a redemption story!
He should look to fork Terra and re-build (maybe a new name) - because the vision was great
Final thought: Thomas Edison probably blew up some stuff along the way as well. He once said that he didn’t fail, he just found 10 000 ways that don’t work.
With regards to stablecoins
A currency needs enough reserves to protect the value of the currency
Stablecoins need enough reserves (must actually over compensate) to protect the currency peg and to protect against current inherent crypto volatility
Perhaps an over-collateralised stablecoin might work. At its peak the Luna Foundation Guard only had $4 billion in reserves while LUNA’s market cap peaked at $41 billion.
There’s weight in the previous statement because DAI is an over-collateralised crypto-backed stablecoin that has stood the test of time
If you go fiat collateralised: (USDC, Tether, etc) the problem with these stablecoins is that they are vulnerable to seizure by governments (because the fiat is controlled by the governments)
If you go crypto collateralised: The problem with these stablecoins is that crypto is nascent, therefore volatile and attracts speculators
If you go algorithmic: the algorithm needs to work perfectly or risk ending up like Terra. No one has been able to figure these out yet.
Final thought: it seems everything just seems to take us back to bitcoin and ethereum. These assets have stood the test of time, and I think the answer is a mix of;
Decentralisation: no single point of failure
Large “reserves”: in this case meaning that because there is no underlying asset that derives the value of bitcoin or Ethereum, they have become markers of success. The success of bitcoin and Ethereum (due to their independent nature), has allowed people to trust those networks over time, allowing people to incrementally build value on top of those networks, in turn validating the Bitcoin and Ethereum networks. Independence and time-in-the-market has appeared to make the Bitcoin and Ethereum systems impenetrable to date as networks, and most valuable as assets.