Terra: An $18 billion run

Terra (UST), a cryptocurrency that’s supposed to be pegged to the US dollar, de-pegged this week. Terra’s market cap of $18 billion was wiped within days. I’m going to cover how this happened, and dive into what this means for the future of stablecoins. If you’re familiar with how the depeg happened, fast forward to the section called “Why stablecoins?”.


How does Terra work?

Terra is an algorithmic stablecoin. If that sounds like gibberish to you, bear with me.

Terra’s objective is to be pegged to the US dollar, i.e. 1 unit of Terra is meant to be equal to $1. Always.

This is achieved using two coins: Terra and Luna. You can always exchange $1 worth of Luna for 1$ worth of Terra. For example, on 1 April 2022, the price of Luna was $101.54. This means that, on that day, you could have exchanged 1 Luna for 101.54 Terra.


Whenever Terra is created, Luna is destroyed and vice versa. Let’s look at how the peg is maintained with two examples:

  1. If the price of Terra is $0.90, people will buy Terra and trade it for Luna. Because 1 unit of Terra can always be traded for $1 worth of Luna, traders will make $0.10. The trading for Luna reduces the supply of Terra, and drives the price of Terra up.
  2. If the price of Terra is $1.10, people will buy Luna and trade it for Terra. As above, they will make $0.10. Because Terra is being created (and Luna is removed), the price of Terra falls.

Terra is an ecosystem and operated other stablecoins. For simplicity, we’re only going to cover the one pegged to to the US dollar. The logic in this deep-dive applies to any other currency in equal measure.

So why did the depeg happen?

The issue with Terra/Luna is that it requires the market to have inherent trust. As a holder, you need to believe that Luna is actually valuable, and the trades mentioned above hold.

Imagine the price of Terra falls to $0.90. One of two things can happen:

  • People start buying Terra and trade it for Luna. They hold the Luna because they believe it’s a pure arbitration opportunity. or,
  • People trade their Terra for Luna, and sell the Luna immediately. This reduces the price of Luna and makes other people want to sell Terra. It’s a classic downward spiral.

In order to avoid the second, 2 mechanisms were put in place.

Mechanism 1: Offer a 20% return

Offered a 20% return for anyone who held deposits of Terra. This was done on the Anchor protocol.


Most crypto traders prefer to keep “uninvested” money in a stablecoin to avoid volatility. Earning a 20% return while you do this is pretty appealing. When the depeg happened, people withdrew their deposits immediately adding further pressure to the price of Terra.

Mechanism 2: Build up reserves of other crypto assets like Bitcoin

The Luna Foundation Guard was established as a non-profit organisation by Do Kwon (founder of the Terra ecosystem) and the Terra team. They built up reserves of Bitcoin. If a depeg were to happen, the Bitcoin would be used to buy Terra and bring it back to $1.

As per the Do Kwon’s tweet below, this is exactly what they tried to do when Terra depegged.

How the depeg was forced

The depeg happened as follows.

Someone bought a tonne of Terra (~1 $Billion). They also borrowed $3 billion of Bitcoin and sold it. As a result, the price of BTC fell and caused panic in the markets.

Then they sold Terra. Because Terra was being sold and the market was in turmoil, the price of Terra started falling. LFG responded by selling BTC.

This was what the attacker wanted all along because they had shorted BTC (betting the price would go down). They used these proceeds to pay back the loan on BTC and make a profit.

By the time LFG realised this, people had lost faith in the currency and it was on its way down.

Incidentally, George Soros made $1 billion by shorting the British pound in similar fashion back in 1992 (commonly referred to as Black Wednesday).


This is the worst collapse in crypto by a mile. People parked their money in Terra for its stability. And it vanished within days. It’s one thing to lose your investment in a risky asset, but losing your money in an asset who’s main objective is stability is scary.

LFG released stats on their remaining reserves. There’s approximately $87 million left in LFG’s reserves. The Terra and Luna coins are worthless.

Why stablecoins?

Most of the goods and services we pay for today are denominated in fiat. Stablecoins are a tool that provides the benefits of crypto (easy, fast transfers) without having to deal with the volatility of Ethereum or Bitcoin.

There are two types of stablecoins:

  1. Asset-backed stablecoins where users can exchange their coin for $1 (or a different currency) using assets held in a reserve. These assets can be fiat, crypto or a combination.
  2. Algorithmic stablecoins like Terra where the protocol maintains no assets or some assets but far less than the market cap of the stablecoin. The more collateral the protocol can avoid the better off it is financially.

Asset backed stablecoins are safer but require you to the trust the other party. How do you know there are reserves? How do you know for sure that these reserves will be used to make the promise good?

Tether is an asset-backed stablecoin. It uses a mix of commercial paper, fiduciary deposits, cash, reserve repo notes, and treasury bills as collateral for its stablecoin. Due to the uncertainty in the markets, it also momentarily unpegged and continues to trade at a slight discount.


In theory, this is fine because you can always redeem your coin for $1.

But traders have questionned whether Tether actually has the reserves to back up the stablecoin. They are famously opaque about their reserves.

My key takeaway:

Stablecoins are as safe as their issuers.

At a minimum, people should hope for decentralised control where a single entity cannot take unilateral actions. After all, this is one of the founding principles of cryptocurrency in the first place.

MakerDAO uses a decentralised, asset-backed approach. Users can get the stablecoin by using other cryptocurrencies as collateral. A single person cannot unilaterally make decisions on the assets. I was looking at their assets though, and was concerned at more than 51% being stablecoins. It seems a bit circular.


Another common criticism of MakerDAO’s DAI coin is the extent of collateral required. Users can must borrow using a ratio of 1.5, i.e. you need to deposit 1.5 ETH to take out the equivalent of 1 ETH in USD. This makes the stablecoin safer (more collateral if something goes wrong) but less efficient.

What does the future of stablecoins?

The ideal stablecoin must achieve all of the following:

  • Stability in price with its currency (e.g. USD).
  • Capital efficient so that users don’t have to lock up assets.
  • Decentralisation to ensure that a single entity can’t control outcomes.

It’s a trilemma. Some believe that an algorithmic stablecoin is simply impossible because you need to trust someone.

But it remains one of the biggest opportunities in crypto and Web3. If you’re building the currency of value for goods and services, your total addressable market is everything over a long enough time horizon.

UXD protocol is trying to build an algorithmic stablecoin that solves all of the above. They claim to be decentralised, stable and capital efficient. They claim to achieve price stability using futures contracts. Their docs have a more thorough explanation of how they hope to achieve price stability.

For now, I’m operating with the following principles for any crypto investments:

  1. It’s a risky asset, invest what you can afford to lose.
  2. If it sounds too good to be true, it probably is. A return of 10%+ and above in the market should be scrutinised.
  3. If I can’t understand it, I’m not touching it.

To close

My heart goes out to all those impacted by the fall in Terra and Luna. If you lost a significant portion of your wealth because of this, please know that you will recover.

Looking ahead, stablecoins are a necessary innovation for crypto to go mainstream. Asset backed stablecoins with sufficient decentralisation are the best option for now. An algorithmic stable coin has a lot of potential but is challenging to maintain. It’s hard to say never, only time will tell if someone can figure this out.