Beyond DeFi 2.0 – What Is Driving DeFi’s Current Innovation?
After a prolonged stagnation, DeFi has once again resurfaced among crypto’s leading narratives. The resurgence is primarily being led by the contentious term “DeFi 2.0”, coined by the pseudonymous developer of Alchemix Finance, Scoopy Trooples. In a recent Twitter thread, Scoopy highlighted a number of second generation protocols building upon the 0 to 1 innovations created by the first generation of DeFi protocols such as MakerDAO, Uniswap, Compound, and Yearn. This classification sparked infighting over the categorization of DeFi protocols and pulled the attention away from the actual shifts occurring under the hood. Twitter timelines have been left in constant confusion with everyone asking the same question about their favorite OlympusDAO fork:
Rather than debate the differences between arbitrary naming conventions (DeFi 1.0 vs DeFi 2.0), let us take a step back and examine the macro themes that are driving the latest wave of DeFi innovation. A year ago, such a report would have covered trends like vampire attacks and launch strategies. Today, we will explore the birth of Liquidity-as-a-Service (LaaS) protocols in addition to the rise of protocols that automate, enhance, or extend existing DeFi economic models.
Liquidity mining, the heart and soul of DeFi Summer 2020, has recently fallen out of favor as protocols struggle to cope with the after effects of mercenary capital providers draining value from their systems. The liquidity mining model provides short term incentives for liquidity providers and creates a perpetual expense on protocols’ balance sheets. As a result, projects are realizing they need better systems to ensure sustainable liquidity while aligning long term incentives for investors. As the problem becomes more well known, projects are beginning to specialize in LaaS. Using a LaaS provider, protocols may buy their liquidity outright from the market or rent it from protocols that are designed to offer the cheapest, yet highest quality liquidity.
Olympus DAO was the first project to create a viable alternative to the liquidity mining model using its novel bonding mechanism. By issuing its native token OHM at a discount, Olympus is able to purchase LP positions from the market to create “protocol owned liquidity” (POL). The recent launch of its Olympus Pro service marked DeFi’s first LaaS offering by introducing Olympus’ bond model to the wider DeFi ecosystem. Olympus Pro provides projects with a custom implementation of the Olympus bonding mechanism while also introducing a new demand channel for their native token. Projects that purchase their own liquidity will accrue any revenue generated from trading fees but will also bear the impermanent loss (IL) associated with price changes. This approach will likely be best suited for larger projects with less volatile token prices in order to minimize IL.
Unlike Olympus, Tokemak is a protocol designed solely for liquidity provisioning. At a high level, Tokemak will function as a decentralized market maker. Tokemak’s native token, TOKE, represents tokenized liquidity and is used to influence the direction of liquidity across DeFi. Protocols are rewarded in TOKE for seeding liquidity into the Tokemak ecosystem and can influence liquidity direction by staking their TOKE. Since the liquidity is ultimately controlled by Tokemak, Tokemak retains the trading fees associated with the controlled assets but also bears any impermanent loss. From a protocol perspective, this approach is the most similar to liquidity mining; the protocol still “rents” liquidity, does not receive trading revenue, and is not exposed to IL. However, rather than incurring perpetual costs to sustain liquidity, protocols are rewarded for their liquidity participation in native TOKE rewards.
Fei Protocol x Ondo Finance Partnership
In addition to a revamp of its core stablecoin protocol, Fei Protocol recently announced plans for a partnership with Ondo Finance to provide an inexpensive, short term LaaS option for DeFi. Protocols will be able to deposit their native token into Ondo liquidity vaults for a specific amount of time and have their deposits matched by newly minted FEI. This token pair is then sent to an AMM for liquidity provisioning. The Fei x Ondo design is attractive for projects that want to generate on-demand liquidity with no upfront cost to acquire the other end of the liquidity pair. In exchange for seeding the other half of the liquidity position, Fei charges a small fixed fee at the maturation of the vault. Since the projects themselves are functioning as the liquidity providers, they are entitled to trading fees but are also exposed to potential impermanent loss. At the end of the period, Ondo returns the token liquidity provided net of trading fees (positive) and IL (negative). This strategy gives protocols a new way to strategically provide liquidity over short time periods at very low cost.
Altogether, LaaS providers are introducing methods to create sustainable liquidity at a lower cost to protocols. While every option will have its pros and cons, it is becoming increasingly clear that LaaS is in the early stages of weaning the current system off of its addiction to liquidity mining.
Second Order Protocols
The second category of projects reshaping DeFi is what I will refer to as “second order” protocols. Leveraging DeFi’s composable nature, these projects are built on top of existing DeFi infrastructure to either automate, enhance, or extend existing DeFi economic models and processes. Since they sit on top of existing “money legos”, their utility comes at the cost of compounded risk.
Yearn’s creation of a yield-as-a-service platform gave projects an automation strategy to adapt to other specialized functions across DeFi: find a process that is time or gas inefficient and package it up for a small fee. In return, the ecosystem becomes more accessible and efficient. There are now several projects being built that automate specific micro-processes within DeFi. Popsicle Finance takes assets to be used for LP positions and manages the liquidity positions across multiple exchanges and layer 1s. Convex Finance recycles $CRV and Curve LP tokens for boosted rewards, trading fees, and governance features. Although slightly older (in crypto terms), Pickle Finance provides auto compounding services for yield aggregator vaults that lack auto compounding as a native feature. The emergence of projects that provide automation-as-a-service are proof that if a specific problem is not fully solved by a dominant protocol, new players will step in to finish the job.
The first iteration of DeFi protocols created several 0 to 1 innovations that now serve as the bedrock of our programmable financial system. A few of the most prominent examples include Maker (collateralized debt positions or CDPs), Compound and AAVE (decentralized interest rate markets), Uniswap (automated market makers), and Yearn (yield aggregators). Unlike these first generation DeFi protocols, model enhancers do not actually introduce any new operating models for DeFi. Rather, they recycle the outputs from existing protocols to provide a more optimized model for the end user.
One of the fastest growing projects of the last month, Abracadabra.money, serves as a great example of a project using the model enhancer strategy by creating CDPs from yield-bearing assets. The CDP model was originally introduced by MakerDAO as a way to create a permissionless credit system. Users mint the protocol’s native stablecoin Dai using a variety of overcollateralized vaults. The common knock on Maker is that its CDPs are capital inefficient since its assets remain locked in vaults and do not earn any interest. To improve the efficiency of the CDP model, Abracadabra uses existing yield-bearing assets as the collateral to mint its native stablecoin MIM. Not only does this put a user’s idle assets to work, it provides the borrower with a deeper liquidation backstop. Since collateral in Abracadabra is continually accruing interest and increasing in nominal value, the probability for liquidation decreases as a function of time, but never to zero. Although model enhancers like Abracadabra do not introduce any novel primitives to DeFi, this category of projects has found product-market fit through increased capital efficiency.
Model extenders are protocols that introduce a similar 0 to 1 innovation that first order protocols created, creating new overall value for the system as a whole. Unlike their base layer counterparts, these primitives are only possible through the use of stacked DeFi protocols.
Alchemix Finance is one example of a model extender. At first glance, Alchemix’s use of CDPs built on top of yield-bearing assets looks very similar to Abracadabra. However, subtle differences allow Alchemix to create a capital efficient, overcollateralized borrowing model that is not subject to liquidations. Liquidations will always be possible in the Maker / Abracadabra model because the value of their collaterals are denominated in a different base asset than the issued stablecoins. Unlike Maker and Abracadabra’s multicollateral stablecoins, Alchemix’s loans are synthetic versions of the underlying collateral making them denominated in the same base asset. For example, ETH loans are issued in the form of alETH, a synthetic asset pegged to the price of ETH. This allows Alchemix to remove price risk from its model and ensure no possibility of liquidation. As a result, the full amount of collateral can be put to use across DeFi to provide an ongoing source of loan repayment. While many will cite the “self-repaying loan” as Alchemix’s primitive, the real extension is the ability to absorb the opportunity cost associated with the decision between consumption and saving.
Another group of model extenders take DeFi derivatives and deconstruct them into separate pieces. Examples of these types of projects include Pendle and Tranche Finance. This optionality provides users with more customization and hedging strategies. Given their high level of innovation, the number of model extender groups is scarce compared to model automaters and model enhancers. However, the unique properties of these projects could serve as the seeds that create DeFi’s “killer apps” that bring the masses into DeFi.
The rehypothecation that model automaters, enhancers, and extenders make use of does not come without trade-offs. The composable nature of these second order protocols result in compounded risk that is no longer only controlled within the protocol. If the protocols that generate their inputs are to be exploited, the second order protocols will fail as well. This creates a house of cards structure where the failure of any foundational pillar will topple the entire tower. Ceteris paribus, or all else equal, the value added by second order protocols is offset by the risk introduced from composability. Despite the carefree marketing that many of these projects use, it is important to understand that risk is more present than ever within second order protocols.
Outliers and Concluding Remarks
The grouping of these themes is not meant to be all-encompassing. That is, a protocol may be pioneering innovation but does not fit nicely within these broad categories. One example is Rari Capital’s Fuse product. Fuse takes Compound and AAVE’s interest pool model and applies it at a granular level allowing users to create an interest pool out of any combination of assets with reliable price oracles. It is still providing meaningful innovation by introducing a permissionless interest pool structure despite not functioning as a LaaS provider or being built directly on top of other protocols.
LaaS providers and second order protocols are not defining characteristics that should be used to call something “DeFi 2.0” either. Trying to apply a catch-all label is not something that should be done proactively while the space is undergoing such rapid evolution. We can save that job for future historians. If we are lucky, we will be able to make it another full year before we address the inevitable “DeFi 3.0”
(Visited 23 times, 1 visits today)