Aave V2 and the AAVE Token

Disclaimer: Nothing in this article constitutes investment advice.

Now that Aave v2 is live on the mainnet it seems like an auspicious occasion to talk a bit about it and how the AAVE token itself will generate value going forward. The goal of Aave v2 is essentially to bring DeFi one step closer to the masses, by targeting a few key problems roadblocking widespread adoption and asset efficiencies, namely: prohibitively high gas fees for smaller transactions/users, too much convolution/too many steps in users performing operations, and suboptimal asset utilisation rates. Aave v2 seeks to remedy this through multiple improvements to the protocol including gas optimisations and streamlined transaction operations, as well as expanded loan functionality such as scalable native credit delegation, as well as seamless debt and collateral swapping (prospective buyers of AAVE should read this article for full details on v2 functionality and v2 documentation).

What drives the revenue to AAVE holders

Aave v2 certainly sounds like the right step for the protocol, but how do AAVE holders make money from it? To the actual holders of the AAVE token, revenues will be driven by the safety module which holders stake their AAVE to, granting them three sources of revenue: safety incentive rewards in the form of AAVE tokens, a portion of Aave protocol fees, and trading fees as well BAL rewards from the 80/20 AAVE/ETH pool on Balancer which the safety module is built on top of (pool not active yet).

These revenues aren’t quite free for the AAVE holder, as the purpose of the safety module is to de-risk the Aave ecosystem as a whole by acting as a re-collateralisation mechanism in case of “shortfall events”, where unexpected protocol losses are caused by things such as smart contract flaws, oracle failures, collateral asset failures/liquidations, etc. The AAVE token can almost be thought of as bank equity where token holders are subordinate to more “senior” protocol stakeholders such as borrowers and lenders, and AAVE holders guarantee their own assets to act as a buffer for the protection of said stakeholders or the protocol as a whole.

A flowchart of how AAVE is used in a shortfall event

With that in mind let’s look further into the three sources of revenue to AAVE holders. For context, current protocol revenues are around 17mn USD annualised for a P/S ratio of ~75x.

1) Safety Incentives (SI): Safety Incentives exist to ensure AAVE holders are incentivized to stake to the Safety Module. Out of the total supply of 16mn AAVE, 3mn tokens were left as an ecosystem reserve to be used as rewards and incentives to this end, and as of writing only 400 AAVE per day are being distributed (implying a ~20 year emission rate), but this will undoubtedly increase soon — in fact emissions may increase to 1100 AAVE per day soon. This makes revenue and hence value recursive to an extent, as AAVE revenue is partially a function of the price of AAVE itself. We expect these will be mostly exhausted within ~5 years, but will help provide yield to AAVE holders before protocol fee revenues are ready to drive AAVE holder income. One way to view safety incentives is almost as if AAVE stakers are writing put options on the failure of protocol collateral assets and getting paid premiums in AAVE tokens as compensation. Whether AAVE stakers are over or under compensated for this risk may be a matter of personal opinion. For instance there are around 370mn USD in centralised stablecoins deposited as collateral assets on Aave — some view a stablecoin failure event as inevitable and some view it as a completely benign risk.

It is also important to note that incentivization via token rewards has historically worked very well for other protocols — for example the Synthetix protocol performed very well after incentivizing protocol participation with SNX.

How the Safety Module distributes revenues to AAVE stakers

2) Aave protocol fees: With the introduction of v2 the proportion and magnitude of protocol fees received by AAVE holders will be overhauled and decided as a matter of governance. How much may we realistically expect? As it stands, the protocol collects 0.09% on flash loan originations and a paltry 0.00001% on standard loan originations. In the long run, protocol fees are going to be the main driver of income to AAVE holders, as the ecosystem reserve and Balancer related revenues are inflation based and won’t last forever. Saying definitively how AAVE holders will capture protocol fees is not straightforward, with there being many different ways this could happen — there has already been extensive discussion about this on the Aave governance forum — but one thing that is clear is that current protocol fee capture is miniscule and in need of reform if it is going to meaningfully contribute. For example standard loan origination generated 150k USD in fees from January to August 2020, while flash loan fees have racked up around 900k USD this year. Although periods of volatility can generate large fees for the protocol, it still isn’t enough to justify current valuations.

In the long run Aave protocol fees will need to be reformed to make holding AAVE worthwhile

This does appear to be a crossroads of sorts for the protocol, where they can attempt to keep protocol fees where they are (at a near minimum) with the hope of capturing large amounts of new and existing market participants, or they can reform protocol fees with AAVE holders capturing a good portion of this through the safety module, and hope that the superior UX and functionality of the protocol compensates for the increase in fees. We think the far more likely approach is the former, where protocol fees will be reformed at a later date when future iterations of Aave are proven, and the protocol has an established edge and user base.

When fees do get properly reformed, monetisation possibilities include: increased origination fees on standard loans, new flash loan fees, fees from liquidations, fees on interest generated, product switching fees (such as switching from stable to variable interest rates), debt switching fees (e.g. if a user trades their borrowed DAI for USDC once it becomes cheaper), collateral trading/switching fees, margin trading/lending/LP fees (e.g. a small fee attached to a user opening/increasing their margin on an open position), credit delegation origination fees, etc.

How this is done is a very delicate procedure and runs the risk of hurting the user experience if it is felt the fees are too egregious, and on the flipside it runs the risk of creating a disgruntled group of stakeholders who are frustrated at the lack of yield being generated by their AAVE tokens which will perform poorly when the market reprices AAVE to reflect lower revenue expectations.

3) Balancer rewards and trading fees: AAVE stakers will have the choice of just staking AAVE to the safety module, or staking BPT tokens from the 80/20 AAVE/ETH pool on Balancer, with it being possible that BPT stakers and AAVE stakers will share equally in the daily SI rewards. However, in the short run revenues from the Balancer pool may actually drive a much larger than expected portion of revenue to AAVE holders. As it stands the Balancer protocol releases 7.5mn BAL tokens as LP reward incentives annually. The allocation of these rewards is a function of how much “adjusted liquidity” users provide to the protocol. Liquidity is adjusted according to a number of factors which we won’t delve too much into here, but based on current liquidity adjustment rules we expect to see a pool liquidity adjustment factor of ~0.64 for the 80/20 AAVE/ETH safety module pool, which means balancer treats every dollar provided to the pool as 64c (we also assumed a 0.05% swap fee on the pool).

In our view, the Balancer pool will provide a yield of between 3% and 10% APY in BAL rewards and trading fees for AAVE stakers, under a range of plausible scenarios. This pool will likely capture anywhere between 10% to 50% of total yearly BAL rewards (depending on the proportion of safety module stakers who choose to stake BPT tokens), but it is very important to note that the yield and dollar return that this pool generates is mostly driven by the effective price of BAL received by AAVE holders. Like safety incentives this makes the value of AAVE somewhat recursive, because If safety module stakers dump their BAL rewards as soon as they are received this may suppress BAL price and hence expected revenues from Balancer — leading to a lower AAVE price. If on the other hand they hold and reduce selling pressure we may see the opposite.

Competitive Positioning in the Lending Space

Aave mainly competes with Compound, Maker and to a lesser extent protocols like CREAM Finance for liquidity and lending activity. Compared to its competitors Aave will be breaking new ground with some of the v2 functionality such as native credit delegation, debt tokenization and expanded flash loan functionality. Right now Aave is the smallest of the three, making up around 10% of the DeFi loan market in addition to having a lower aggregated utilisation rate than Maker or Compound. It should be pointed out however that Aave as a protocol is also much younger than Maker or Compound, having launched in January 2020 - but it cannot be denied that Aave does have something of an asset utilisation problem.

Aave protocol lags behind Maker and Compound for loan market share and collateral utilisation

With v2 having just launched on the mainnet, an important question is whether it is enough to not only solve some extant problems the protocol has, but establish an edge with it. Ideally Aave will create a sustainable moat by building a superior and unique user experience — which will grow liquidity and the user base — accruing network effects along the way which is sufficient to keep users on the protocol once protocol fees have been reformed.

One example of how this could happen is that native credit delegation (which neither Maker nor Compound has) draws in new liquidity due to heightened capital allocation efficiency (at the moment credit delegation on Aave is handled in a centralised way that cannot scale), owing to the fact that credit delegation will allow underutilised collateral assets of even small users to become available to a new class of market participants who want/need to tap into that borrowing power via uncollateralized loans. Stani Kulechov’s vision for credit delegation is that it “…could become a wholesale debt market. Which means if you are a facility in DeFi, CeFi, traditional finance, you could source part of your liquidity from Aave”. This may work through dedicated liquidity pools users can stake to for the purpose of credit delegation combined with governance to determine who can borrow from those pools.

Successful scalable credit delegation would have the effect of boosting utilisation rates and overall asset yields with the corollary of drawing in more liquidity from yield farmers and hence more TVL. More TVL means that Aave would be more attractive to use for native credit delegators and borrowers, flash loan users and yield farmers themselves when the cycle repeats. This gets even more interesting when we factor in things like Aave’s partnership with RealT to bring tokenized real estate assets into the lending market getting closer to being fully realized. Although this is currently a small market it signals serious intent from Aave to explore the use of tokenized real world assets on the protocol. Native credit delegation for untapped borrowing power on assets such as real estate or commodities could have big potential for a first mover to gain an edge in TVL capture.

In this way it is conceivable that the v2 first mover advantage may allow Aave to gain enough network effects of capital to at the very least provide a short term edge over their competition which will undoubtedly flow through to price. This is all obviously easier said than done however because in time if v2 is a success it will be copied and iterated upon by both Compound and Maker insofar as its possible, but Aave is well positioned to at least recapture a much higher proportion of loan activity in the space next year.

A side note readers should also be aware of is the potential for Aave to rapidly expand their user base and TVL via operations related to the UK electronic money license they secured near the end of August. According to Stani Kulechov the purpose of this is to enable users “to go from Fiat to stablecoins and other assets natively in the Aave Ecosystem and then use these assets in the Aave Protocol”. Being able to set up your protocol as both a gateway from fiat to DeFi as well as the best place to partake in DeFi gives Aave a large potential edge in capturing and retaining new market participants.

AAVE Valuation

At the time of writing AAVE has a fully diluted market cap of around 1.3bn USD (c. $83 per token). Since we know exactly how AAVE generates earnings for holders, we can create valuation models to produce a sanity check on whether 1.3bn USD is reasonable. However, despite its “unicorn” and “blue chip” status it is still early days for the protocol with the market still likely pricing AAVE largely based on its narrative as opposed to hard fundamentals. As the price discovery of the token matures and protocol fees get closer to being meaningfully implemented this will change, but for now it probably isn’t that helpful to build out a detailed financial model for investment purposes.

2020 has been tremendously successful for Aave

For the Aave protocol 2020 has been a great year, with lending activity and TVL increasing by four figure multiples. The Aave team itself has demonstrated that they aren't shy of experimenting or being the first to implement a concept, and this pioneering spirit ensures the protocol has a good chance of benefitting from any future growth in the DeFi space.

With this in mind one way to think about buying AAVE is as if you were buying a long term call option on the DeFi space, as well as specifically on Aave protocol fee capture where your payoff depends on how much fees are able to grow and the proportion of fees that actually get redirected back to you. Of course the price of this option is not equal to the current token price, as we should deduct Balancer related revenues and safety incentives which together could make up anywhere between 200mn USD and 600mn USD in present value under a range of plausible scenarios. This means at current AAVE prices you are paying anywhere between 45$ to 70$ for this option. This is a bargain for those who believe Aave will be a multi-billion dollar protocol within a few years. Others may not be so sure, but hopefully we have helped demystify exactly how holders of the AAVE token can make money going forward.

Student of the Science of Numbers go Up