This post is an attempt to identify the key developments of DeFi in the past few years, the innovations that have been spurred as a result, and all of the subsequent implications on where DeFi is headed in 2022. Advanced readers will understand the concepts behind AMMs, CRV, and Olympus, but we introduce these concepts in order to establish a foundation from which we can make our judgements. Ultimately, a lot has happened in the DeFi space since 2019, and we attempt to articulate those developments here. Following along on Twitter or Discord quickly becomes a herculean effort, so we attempt to distill all of the information here. This is a long post, but well worth it to gather a well-rounded understanding of where we started in DeFi and where we are going.
The overarching theme that ties all of these separate discussions: “He who controls the spice controls the liquidity”. If I’ve already lost you or you're currently rummaging around amidst paprika and cumin, just hang in there - I promise it will make sense.
The primary function in the traditional finance world that enables the purchase and sale of securities is the central limit order book. Let’s say I want to buy Amazon stock: I can go to a broker dealer, who serves as the middleman to enable my purchase. They will provide me with my stock in exchange for my dollars. This model works well and enables brokerages to sit in the middle of an enormous pool of economic activity, buying and selling securities to “make the market” on an asset for the purpose of price discovery. AMZN might be trading at $101, and I want to buy it for $99. The brokerage will facilitate this transaction by buying AMZN at $101 and selling it to me for $99. In this scenario, the broker dealer takes a $2 loss on the trade itself, but takes a fee from me and the market participant on the other side of the trade. At scale, this is an enormously profitable business for broker dealers that control the central order book.
In 2019, Hayden Adams launched Uniswap on the Ethereum mainnet, with the purpose of eliminating the need for a central order book. Uniswap implemented a simple yet elegant mathematical formula to enable price discovery on an asset in a pool of liquidity. The formula is:
K = y * x
where K is the total supply of assets in the pool, y is the reserve of asset y (price * supply), and x is the reserve of asset x (price * supply). A more technical explanation of this formula can be explored here, or if you like short videos, here.
Let’s say I want to swap my ETH for another ERC20 token, like USDC, similar to the previous example where I traded my dollars for a security. Uniswap’s ETH/USDC liquidity pool has a reserve of both assets, kept stable by reaching equilibria at K. When I swap my ETH for USDC, there is now more ETH in the liquidity pool than USDC, thus the reserve of asset Y (ETH) is now greater than the reserve of asset X (USDC). In order to reach equilibria at K, one of two things must now occur. Either the pool adds more USDC to balance the equation, or, the price of USDC in the pool increases. With many market participants, this model works effectively to enable price discovery on an asset, as arbitrageurs will subsequently enter the pool make the market on USDC back to peg. So voila - I can frictionlessly trade my ETH with an unknown market participant at fair value without the need for a central body to facilitate the transaction (i.e. in a trustless fashion). In order to incentivize liquidity for this pool so that market participants can swap within it, liquidity providers are paid a small fee on every transaction with the pool.
Barter and exchange is the basis for all economic activity. For centuries, banks and central authorities with writs and mandates have enabled economies of scale to rise, but over time they become rent-seeking entities that extract enormous sums of value from the market participants. The automated market maker (AMM) was a revolutionary piece of financial infrastructure, because it programmatically eliminated the requirement for a centralized entity by pooling incentives across a decentralized network of participants to enable the barter and exchange of assets.
Curve Finance is another take on an AMM, but it prioritizes price discovery for stablecoins. Launched in 2020, it attempted to create a low-slippage market maker for stablecoin pools leveraging more efficient algorithms. The intention was to provide something of a fiat savings account for liquidity providers, earning high interest rates from lending protocols without the exposure to more volatile assets. Because both assets in a Curve pool are stablecoins (i.e. DAI and USDC), trading between them causes minimal volatility compared to other pools like ETH/USDC. Rather than focusing on “balancing” a pool like Uniswap, Curve concentrates liquidity near the ideal price for two stablecoins and subsequently injects liquidity where it is needed the most. In doing so, Curve achieves much higher liquidity utilization on its pooled assets, meaning better execution and prices for the user. Curve has evolved to introduce non-stablecoin pairings, but for the purposes of this article we’ll focus on the original interpretation of the Curve methodology. We’ll explore the mechanics behind Curve’s governance token CRV in more detail later, as they are key to understanding the evolution of the ecosystem.
While these various AMM implementations have given rise to a vibrant ecosystem of decentralized financial products, they have drawbacks. This model suffers from the problem of mercenary liquidity, or the idea that liquidity providers will provide assets to the pools that derive them the greatest yields. Yields are high when liquidity is low, because you are taking greater risk by providing your liquidity to such a pool. As a liquidity pool grows its assets over time and gains price discovery stability, yield rewards for LPs shrink. This in turn causes liquidity to flee from lower yielding pools to higher yielding pools, and a quasi “run on the bank” scenario can take place. This scenario is not exclusive to liquidity pools, it can be applied to entire protocols as a result.
While the NFT boom of 2021 saw a huge influx of Ethereum users enter the space to buy jpegs, the financial savants of DeFi were hard at work solving the issue of mercenary liquidity, also known as “rented liquidity” for the reasons mentioned above. Enter OlympusDAO, a protocol attempting to solve this issue.
The core goal behind Olympus is the mission to create an algorithmic reserve asset that is not pegged to the dollar. Any stablecoin, decentralized or not, pegged to the dollar inherits its gradual devaluation over time. As inflation rises in a low interest-rate environment, and more dollars are printed by the Federal Reserve, the purchasing power of your dollar goes down.
Olympus has a few core functions that enable price discovery on its decentralized reserve asset, the OHM token. At the center sits the treasury. The treasury is a vault (smart contract) holding funds collected by the protocol. OHM is backed (not pegged, as they’ll be quick to correct you) by assets collected in the treasury, and has a guaranteed collateralization ratio, meaning the price of OHM should not fall below the value of its backed assets.
The Olympus mechanism that generates profit for the treasury is bonding. The treasury sells bonds to protocols looking to earn yield on their assets. This flips the issue of mercenary capital on its head: rather than renting liquidity like an AMM, Olympus owns its liquidity and sells bonds to generate profit. This ensures greater stability for the treasury, eliminating a “run on the bank” scenario. The liquidity comes from market participants buying and staking their OHM in the treasury, which is where the game theoretic (3, 3) concept comes into play that you have undoubtedly come across on Twitter. By staking your OHM, you are entitled to the rebase rewards the protocol generates at every epoch. Bond sales generate profit, which the treasury uses to mint new OHM and reward to stakers.
This is also where the memeworthy yields arise. Similar to how a smaller liquidity pool will generate a higher fee for LPs early on, the Olympus treasury attempts to deliver insane yields until a certain level of liquidity is achieved. Over time, the Olympus DAO, governed by the OHM token, can choose to reduce the rebase rewards as protocol owned liquidity increases. As the protocol controls the funds in its treasury, OHM can be minted or burned by the protocol to influence the price of OHM. Again, the idea is to create an algorithmic reserve currency, but something that mimics a store of value more akin to gold than anything else. We think this is a bit of a misguided strategy, as the tokenomics surrounding that discussion are not quite sound. However, Olympus has proven its model for bootstrapping liquidity for new projects, which is indeed a promising innovation compared to what we’ve experienced in the space previously.
OHM remains a dramatically volatile asset, down over 60% in the last 30 days. However, the theory goes that over time as the protocol is able to amass greater protocol owned liquidity, OHM will gradually achieve greater stability, since the treasury can be more reflexive in controlling the outstanding supply of OHM. Long term stakers, aka those participating in (3, 3), aren’t worried about the price at any given moment, since the game theory suggests the thesis is to acquire as much OHM as possible over time through rebase rewards. Over a long enough time horizon, the price of OHM becomes negligible relative to the sheer volume of rebase rewards stakers receive.
There has been a ton of noise around OHM of late, as the token continues to crater. The profit-generating mechanism of bonding is not strong enough to support the weight of an ecosystem with average yields still above 4,000%. The token’s slide in price serves as evidence to this understanding. But this correction is healthy long term as the (9,9)ers - individuals taking loans for leveraged exposure to OHM - get liquidated out of existence. The attention Olympus has garnered this year has been staggering, but distracting. If Zeus and the Olympus team are able to continue shipping products that can further contribute to that profit-generating machine, we think the Olympus ecosystem is now in a good position to enable more organic (i.e., less leveraged) growth. At the end of the day, Olympus has amassed an enormous pile of cash and has various ways to make that pile of money grow. As investors, that is the making of a business we appreciate in any market.
The hype around Olympus obviously comes from the ridiculous rebase yields it boasts, but underneath that, there is a novel attempt at governance tokenomics. Tokenomics, for the most part in DeFi, is rather shitty. Governance and protocol utility tokens are rarely able to truly capture a protocol’s value for holders, save for a percentage of the fee a service reaps in addition to some voting rights perhaps. But if a token is to serve as an ownership vehicle for a protocol akin to equity in a corporation, it needs to be more than that. So we give Olympus credit for the attempt, even if the rebase yields over-promised and, for now, have under-delivered. At the very least, Olympus has utilized the OHM token to demonstrate a novel attempt at rapidly acquiring liquidity in its treasury, which is table stakes for a project to survive early on in a violent market.
OK - back to Curve Finance. In order to fully grasp the innovation behind Curve, it’s important to understand the mechanics of its governance token, CRV. The CRV tokens is the primary tool used to direct liquidity injections across the Curve ecosystem. Investors can lock their CRV tokens on a 4 year vesting schedule, receiving veCRV (voted escrow CRV) in return. veCRV receives fees from Curve liquidity pools as well as voting rights in the Curve DAO. The latter of those incentives is deeply important, and is what has given rise to the CRV Wars.
The power of the Curve system is the ability for veCRV holders to vote on how CRV inflation is distributed by assigning weights to each pool's "gauge". Curve Finance emits new CRV tokens on a predictable schedule. Curve’s liquidity providers will receive these newly minted CRV tokens, but how much they receive is determined by the gauge weights as voted on by veCRV holders. This creates a flywheel effect, explained well by the good folks at Average Joe’s Crypto. It works like this: an investor provides liquidity to a particular CRV pool, and buys CRV tokens to lock them and receive veCRV in return. Using their veCRV tokens, the user can vote to increase CRV emissions for the pool he/she is providing liquidity to. As the pool receives CRV emissions, the user will lock CRV rewards again to receive more veCRV, further increasing future emissions.
As the Curve ecosystem has grown and developed, protocols have become the largest buyers of CRV tokens.
Convex Finance has become the largest holder of CRV tokens by enabling a secondary market for locked CRV tokens. Convex was built for exactly that purpose: to acquire as much CRV and veCRV as possible. In order to receive the largest possible share of rewards, veCRV holders must lock their CRV for a maximum of four years, a horribly long time horizon for the typical DeFi investor with the attention span of a mouse. That’s where Convex comes into play, by enabling a user to burn their CRV (irrevocably) into cvxCRV tokens, thus tokenizing your veCRV. This has two important features: 1) it enables a liquid secondary market for previously illiquid veCRV tokens, and 2) since CRV tokens are permanently locked in return for cvxCRV, Convex’s veCRV holdings increase indefinitely.
The Convex model has proven to be an enormously effective method for acquiring an outsized market share of CRV tokens. This is important, because recall that owning these tokens enables the holder to determine which liquidity pools will receive the greatest share of rewards.
[Redacted] and its native $BTRFLY token are attempting to leverage Olympus’ rebase mechanism to position itself as a power player in the CRV war. It purchases CVX, CRV and OHM through bonds, and prints an insane APY for its users. It, too, has become a dominant force in the CRV wars, attempting to bootstrap liquidity in pursuit of voting power. [Redacted] is a fascinating example of how DeFi primitives can be mutated to achieve a different set of objectives. A great take on it from Knower can be found here.
Now that we’ve laid out all the groundwork for what gave rise to DeFi and some of the new approaches to tokenomics, we can begin to piece it all together and attempt to make judgements on where it’s headed. The stage is being set for the CRV wars now, as protocols are only just beginning to catch on to the power associated with acquiring a significant amount of CRV tokens. Separately, the rebase mechanisms behind Olympus, and the myriad forks and iterations of it, will see their greatest test in 2022, as the trees have been shaken violently. The FUD is in, and there are many left questioning whether Olympus will be able to achieve its vision of an algorithmic reserve currency, and whether the bonding mechanism is a viable tokenomic construction to enable such a product. It’s effect on the CRV wars is even further yet to be determined, as it figures out its own systemic risks.
So where is all this headed? Enter Dani and Andre, who really need no introductions, and if you’ve read this far we assume that you’re familiar with their backgrounds. The Godfathers of DeFi are collaborating (bullish in and of itself) on a new project called Solidly that attempts to combine the rebase mechanism behind Olympus with the power of owning voted escrow tokens, like veCRV.
Essentially, Solidly works just like CRV, but with a twist on the way you are able to receive LP rewards. Rather than being provided a voted escrow token in exchange for providing liquidity, like veCRV, you receive an NFT, which can then be staked to receive rewards. The key difference between the CRV model and the Solidly model has to do with emissions schedule. While CRV emits new tokens to veCRV holders on a predictable schedule, there is no max supply. CRV can, in theory, mint new tokens ad infinitum, which introduces inflationary ambiguity to the value of CRV. Solidly attempts to solve for this by emitting rewards for those staking the NFT on a predictable schedule, but with a max supply. The emission of new tokens from this supply leverages a rebase mechanism, hence the term ve(3, 3). Here is a good video on how the model might work.
Does this solve the issue of DeFi tokenomics? That remains to be seen, especially considering so much of the DeFi space is driven by hype and FOMO, and the fact that many degens will ape into anything that Andre touches. We’ll be observing this space closely for 2022, as these new liquidity models have been tested rigorously in recent months. Whether or not they will survive and enable DeFi to move toward a new paradigm for governance tokenomics is up for debate, but it most certainly will be one of the most interesting innovations in DeFi to come this year.