A Brief History of DeFi — Part 3: The Failed Use Case

Hermetica
7 min readFeb 28, 2024

--

2021 was a high point for DeFi innovation. The following years would see a significant decline in on-chain activity as multiple attempts to innovate new use cases failed, sometimes spectacularly.

Algorithmic stablecoins, notably Terra (UST), set the industry into a tailspin that it is only now recovering from.

Algorithmic Stablecoins

Algorithmic stablecoins, or undercollateralized stablecoins, are crypto tokens pegged to the value of a fiat currency, usually the US dollar.

This is done via a stability mechanism. These stability mechanisms increase the supply of the pegged asset when demand pushes the price above the peg and decrease the supply when market selling pushes the token below the peg price.

In our first part of this series, we discussed CDPs, or collateralized debt positions, wherein a user mints a stablecoin by depositing an unstable asset, like Ether, and “borrows” a certain amount of its value as a newly minted coin. It is important to note that most of these protocols are overcollateralized, as the value of the tokens they mint must be significantly less than the value of the unstable assets deposited. CDPs allow users to borrow 75% of their collateral value, with a liquidation threshold of 85%.

Algorithmic stablecoins use a similar overcollateralized model as CDPs. However, they only partially back their currency with this method and back the rest of the stablecoin’s market capitalization using a “stability asset.”

A stability asset is an unstable token created by the protocol and can be used in protocol governance similar to stock equity. This stability token is minted or deleted based on the inverse demand for the stablecoin. When the stablecoin trades above the peg, more stablecoins are minted and sold into the market. The proceeds from this sale are then used to buy back and delete the stability asset / governance token. This reduces its supply similar to a share buyback in equity markets.

Conversely, when the stablecoin is trading below its peg the governance token is minted and sold into the market. The proceeds are used to buy back the stablecoin thus reducing its supply and raising its price back to its peg.

This process of governance token issuance to reduce supply is like a bank issuing equity to cover losses sustained during lending. Silicon Valley Bank (SVB), First Republic Bank, and others all three attempted share issuances during runs on their banks in February 2023 which was more or less successful. Algorithmic stablecoins systematize this process by betting on sufficient demand for their governance token to balance small fluctuations in their stablecoin price.

The problem arises, just as with fractional reserve banks, when faith in the protocol’s ability to restore its stablecoin peg falls too low. This makes it impossible for investors to rationally buy the stablecoin or the governance token.

Terra-Luna

Terra ($UST) was a pure algorithmic stablecoin only backed by its governance token, Luna, without CDP backing. Luna was also a proof-of-stake (POS) network that required users to stake their Luna to run the network. Unlike other POS networks transaction fees on the network were paid in UST instead of Luna, which is the case for most POS networks. As a result, Luna paid its stakers in UST rewards.

The Luna Foundation used this feature to incentivize a lending market called Anchor. Anchor allowed Luna holders to borrow UST against their Luna below the interest rate paid to UST lenders. Anchor paid lenders a fixed rate of 20% in UST and charged Luna borrowers a dynamic rate of less than 20%. The differential was partially offset by Anchor harvesting UST rewards from borrowers’ Luna collateral. However, this was still insufficient, so the Terra Foundation created more UST in order to pay out UST depositors.

This created a “flywheel” structure where market participants were encouraged to buy Luna, borrow UST, and then lend that UST back to anchor. This captured a spread between the borrowing rate and lending rate. The process of borrowing an asset at one rate and lending at a higher rate is called a carry yield.

This incentive system caused UST’s market capitalization to grow by 900% from $2 billion to over $18 billion in about one year. Defilama, a data provider that tracks DeFi metrics, estimated that at its peak the Terra-Luna ecosystem was at one point approximately 15% of all DeFi total value locked (TVL).

The Terra-Luna Catastrophe

Terra-Luna saw an increase in capital in Q4 2021 and Q1 2022, as the rest of the industry came down from the euphoria of 2021. Yields across DeFi began their long decline in October 2021 from thousands of APR to hundreds and eventual sub-10% yields.

Most yields in DeFi came from token incentives and not the customary borrower APR, which is how yield is generated through sustainable lending. Rewards from token incentives could only last as long as there was enough speculative demand for newly released cryptocurrencies. This speculative demand collapsed in Q4 2021 as the US Federal Reserve began its belated battle with inflation and raised short-term interest rates. This caused inflows to stop and eventually reverse resulting in lower prices and yields across the crypto space — all except Terra-Luna and Anchor.

As previously mentioned, Anchor’s deposit yield was fixed at 20% regardless of how much UST was lent through the protocol. As yields fell across crypto many market participants were forced to use Anchor to maintain the high yields they had become accustomed to not understanding the potential for catastrophic loss in the event of a UST depeg event.

On May 8th, 2022 a large amount of UST was sold through Curve’s UST-Curve3 pool triggering a UST depeg which eventually spiraled UST to 2c on the dollar. The protocol responded to this depeg by minting billions of Luna tokens and selling those into the market to buy back UST and stabilize the price. Realizing that Luna would be diluted Luna token holders lost faith in the project’s ability to make them a positive return net dilution and so they began selling their holdings, further accelerating Luna’s decline until it became worthless.

The fallout from the Terra-Luna collapse wiped out approximately 15% of DeFi directly, and another 20% from the indirect fallout to AMM/lending pools in other ecosystems and to the value of cryptocurrencies at large.

DeFi TVL has not recovered from the UST depeg and sits at levels below Q1 2021.

Why did Algorithmic Stablecoins Fail DeFi?

DeFi brought financial functionality to the blockchain: AMM DEXs, money markets, and overcollateralized stablecoins allowed users to trade and borrow against their cryptocurrency assets; Liquid staking tokens, Perp DEXs, and DOVs gave users access to additional on-chain liquidity and sustainable yield-generating opportunities. But any further attempts to increase capital efficiency failed.

Why? Because Terra-Luna copied the design of the banking system, but was missing a critical piece.

Banks work on a fractional reserve model where banks borrow from depositors on a short-term basis and lend these deposits out as long-term loans for interest. At any one time banks hold a small fraction of total deposits in reserve to cover depositors who seek to withdraw. Banks historically held less than 10% of assets as cash reserves, making their leverage on deposits 10x or 1000% on average. This high leverage ratio allows banks to optimize their reserves in lending operations bringing in a sizable carry return on every dollar deposited.

Sounds familiar? This is the same structure as Terra-Luna. UST can be analogized to bank deposits, overcollateralized Anchor borrowers to bank borrowers, and Luna to bank equity. In the same analogy, the UST depeg is functionally the same as a “bank run.”

However, unlike Terra-Luna, banks have an ace up their sleeves that allows them to lend at eye-watering leverage and have no fear of failure — the government.

If the bank in question is deemed important enough by the federal government that bank will receive a large amount of freshly created cash often called a “bailout.” Banks that receive bailouts might be forced to change top management or have some equity investors whipped out, but cash depositors are never whipped out. Thus, depositing US dollars at a domestically domiciled bank is risk-free.

Terra-Luna was a bank that held a synthetic US dollar but without the Federal Reserve’s backing. Without the Federal Reserve’s support, Terra-Luna was doomed. All banks run into problems eventually, it is in the nature of being highly leveraged that one day you will not be able to cover your debt to depositors.

Therefore, when a turndown in the crypto industry occurred (ironically, but not surprisingly when the Federal Reserve started raising short-term interest rates) UST collapsed as outflows trumped inflows and the fractional reserve pyramid scheme crashed down.

--

--

Hermetica

The first Bitcoin-backed, yield-bearing synthetic dollar protocol.