Is the Convex Flywheel Sustainable?

Introduction: Understanding the Convex Flywheel

Through the latter half of 2021, Curve and Convex were the talk of the town. After conversations surrounding the “Curve Wars” heated up, many DeFi tokens began to explore veTokenomics, looking to emulate the returns that CRV and CVX experienced.

Convex Finance is a yield optimization platform that has been able to accumulate over 48% of circulating vote-escrowed CRV (veCRV) in its short lifespan. Through their symbiotic relationship, Convex has helped propel Curve to the largest DeFi protocol by total value locked, amassing over $18B in deposits.

To understand how Convex operates, it’s important to know the utility and limitations of veCRV. veCRV aims to align incentives between holders of the CRV token and liquidity providers on the Curve platform. It does this by allowing CRV to be locked as veCRV for up to 4 years; the longer the lock up, the greater the benefits. When users do this they sacrifice the liquidity of their CRV in exchange for three things:

A pro-rated share of fees generated by Curve
Boosted CRV rewards on LP positions (up to 2.5x)
Governance and gauge weight voting power, which directs CRV emissions

Given veCRV’s immense utility but the obvious drawback of locked tokens being illiquid, this created the perfect opportunity for Convex to strike with its liquid staking derivative – cvxCRV. Users can deposit CRV into Convex in exchange for cvxCRV, which yields a very high APR of 50%. The tradeoff here is that cvxCRV is not redeemable and thus only holds it’s peg to CRV via its high APR and incentivized pools. Convex takes users CRV and max locks them as veCRV to be controlled by the protocol forever, boosting its own pools and giving CVX holders the ability to tap into the gauge weight powers and direct emissions.

The above graphic does a good job at summarizing the Convex flywheel and how it has been able to capture so much veCRV through cvxCRV. If you would like to understand the Curve/Convex mechanics and relationship at a deeper level please read our previous post, Convex Continues Compounding CRV from Aug. 2021.

Without further ado lets move forward and discuss everything Convex – including bribes, sustainability, Uniswap vs Curve, and Convex’s endgame.

How Convex Creates Value

At the time of our last report, Convex’s value accrual was largely realized from its staking rewards, receiving 5% of all CRV farmed. Since then, the ability for Convex to monetize its veCRV gauge weight power has been made possible by vote locked CVX (vlCVX). Through platforms like Votium, protocols are able to “bribe” vlCVX holders to direct CRV emissions to their selected Curve pools. Bribing has taken off over the last few months and has generated some serious cashflows for vlCVX holders.

Annualized bribes per vlCVX have grown from $1.76 in the first epoch to over $22.72 at their peak in December representing a 46% APR on vlCVX; over $76M has been collected in bribes since inception. But to understand how and why this works, you need to first know the factors driving bribes and whether this is sustainable.

Firstly, why do protocols bribe?

Simplistically, bribing is an arbitrage on CRV (and thus CVX) emissions. Instead of protocols paying to incentivize their own pool 2, they can pay vlCVX holders to use their gauge weight power and direct CRV + CVX emissions to their specified Curve pool. When Convex first started every $1 spent on bribes could earn ~$5.30 in CRV and CVX emissions. Over time, this opportunity has been milked, and it currently sits at $1 in bribes for ~$1.90 in emissions. So who has been taking advantage of this arbitrage?

Bribes are currently dominated by stablecoin protocols – mainly Frax, Terra and Abracadabra, with smaller participants such as Lido, Reflexer, Alchemix, Tribe and Angle also participating. That said, the bribe market is very concentrated, with the top 3 protocols contributing 2/3rds of total bribes. This makes sense as liquidity is the name of the game for stablecoins and Curve is the deepest, most liquid place to trade them on-chain.

Another interesting development is Curve V2. As Curve forays into their own implementation of concentrated liquidity which optimizes for “lazy liquidity” as compared to Uniswap V3’s design that requires either active management or use of an external vault. This opens up pools on Curve to be created for non-stable assets while also competing for gauge weights. Whether these new non-stablecoin protocols will participate in the bribe wars is uncertain given these assets don’t require the depth of liquidity that stablecoins do.

Many DAOs have been in a heated race to accumulate CVX to either support their own stablecoin / pegged asset with CRV emissions (Frax/Badger/Terra/Alchemix), or to deploy their treasury into a yield-bearing asset (Redacted Cartel/Olympus).

One may notice that the two largest algorithmic stablecoin protocols – Frax and Terra – are both very large CVX holders and Convex bribers. When these protocols bribe, they receive a portion of it back as yield on their vlCVX, thus increasing the attractiveness of the arbitrage. Terra takes this to the next level by deploying their community fund’s UST into the Curve pool they are incentivizing. This allows them to farm the CRV and CVX emissions from their own bribes. These stablecoin/pegged asset protocols like Terra are very sticky holders of CVX as Curve/Convex is an essential piece of infrastructure to ensure their stablecoins have sufficient liquidity on Ethereum.

The end game for these protocols is to accumulate CVX to an extent where they can incentivize their Curve pools in perpetuity. While our hypothesis is the more liquid a pool is the better (we touch on this next section), it remains to be seen if these protocols will continue to accumulate CVX indefinitely. If you’d like to learn more and follow CVX accumulation by DAOs, you can check out the following community resource.

When examining the sustainability of Convex and bribes, we must look at CRV and CVX emissions alongside Convex’s control over them. Firstly, ~633K CRV are being emitted every day to Curve LPs. These emissions decrease by ~15% per year until they are fully expended (20% have already been expended, 50% will be emitted by March 2025 and the remaining 50% over the following ~300 years). It is very difficult to tell what will happen within DeFi over the next few years and although Curve currently has massive emissions, they do last for quite some time, which bodes well for its sustainability. Convex’s emissions, on the other hand, are not a function of time, rather a function of CRV farmed through the Convex platform.

After launch, there were roughly 450 CVX emitted for every CRV farmed. This has decreased down to a ratio of 0.66 CVX per 1 CRV today. So far 33M out of 50M CVX set aside for LPs have been released. As supply increases, the mint ratio between CVX and CRV trends towards zero.

During Jan., an average of 92% of CRV had been farmed via Convex. Since CVX emissions are a function of CRV farmed, if we project out a 90% Convex take rate into the future we can estimate Convex’s emissions. (Note that the 90% take rate is conservative in the sense of how quickly Convex emissions will be depleted; a higher percentage take rate means that CVX emissions happen faster). Over the next 12 months, Convex will emit an additional ~15.5M CVX reaching 47.5M, out of the 50M set for LPs.

Now that we have a solid grasp around CRV and CVX emissions, the remaining piece of the puzzle is to understand what percentage of future emissions Convex controls and can be purchased via bribes. For this, we must turn our attention to cvxCRV – Convex’s liquid staking derivative that receives CRV and permanently locks it within the protocol.

Now, not all veCRV is used to vote on gauges, but Convex controls at least 48% of the current voting power and thus gauge weight influence. How will this change if Convex rewards slow down over the next 12 months?

Well, for that we must pay attention to the percentage of CRV emissions that are being locked in cvxCRV. Not all CRV emissions will be locked as veCRV, so it’s safe to say that as long as this ratio stays equal to or above the percentage of veCRV that Convex owns, then Convex’s share of governance and thus its influence over future emissions will not be diluted over time. (Note this doesn’t account for existing circulating CRV being locked up as veCRV).

The ability for cvxCRV to attract an equal or higher amount of CRV in the future will come down to cvxCRV maintaining its peg to CRV, as well as being able to attract more CRV deposits than regular veCRV. To maintain its peg, Convex has allocated 25% of CVX tokens (distributed over four years) to incentivize the cvxCRV/CRV. Currently, cvxCRV is yielding a 50% APR, which is high enough for CRV holders to choose cvxCRV instead of veCRV. But what factors affect this APR and how sustainable is it?

The cvxCRV APR is generated by three sources:

Firstly, ~ 13% 3CRV APR, which is generated off the admin fees Curve pays to the underlying veCRV.

This is a function of Curve’s success and fee generation as a DEX.
This isn’t super relevant for our analysis vs veCRV, as veCRV generates this same APR.

Secondly, ~16% CRV APR, which is generated by paying 10% of all CRV farmed on Convex to cvxCRV holders.

This is a function of how much CRV Convex can farm, which is dependent on the free boosties it provides users and CVX emissions.
This will likely stay high on a relative basis (even if CVX incentives decrease) due to free boosties and the fact that Convex is “the only game in town”.

Thirdly, ~21% CVX APR, is generated by the CVX emitted from the 10% of CRV farmed. For example, if a pool generates 100 CRV and the CVX/CRV mint ratio is 0.66, that pool will also generate 66 CVX. The 10% fee paid to cvxCRV stakers is then 10 CRV and 6 CVX.

This is a function of CVX emissions (and CVX token price) that will nearly run out within 12 months if Convex continues to farm ~90% of all CRV emissions.

As CVX emissions fade and that portion of cvxCRV’s APR dissipates, the choice between veCRV and cvxCRV comes down to two further things.

First and foremost is yield; in this case, veCRV’s APR that can be generated via its gauge weight power through websites like bribe.crv.finance vs cvxCRV’s APR which is dependent on the CRV APR that Convex can generate by farming CRV through its platform (remember they both receive 3CRV APR).

Second is ownership and liquidity. With veCRV, you lose liquidity but retain ownership of the underlying CRV. However, with cvxCRV, you lose ownership of the underlying CRV in return for a liquid derivative that could potentially lose its peg. In our view, the choice between veCRV and cvxCRV over the longer term is the key question that determines the sustainability of Convex. In order to monitor this, we are watching cvxCRV’s APR relative to that of veCRV along with the percentage of daily CRV emissions locked in cvxCRV. These two metrics act as a leading indicator of Convex’s health.

From a wider view, these assets will ultimately be priced based on the success of Curve as a DEX, so let’s take a look at the macro picture for Curve and how it stacks up against its competition.

DEX Competition and the True Purpose of Curve

Two of the oldest and most popular DEXes within crypto are Curve and Uniswap. Uniswap v1 & v2 were typically used to swap volatile assets, and popularized on-chain trading. Curve focused on “stableswaps,” allowing users to move between stables with low slippage. As a result, Curve created a hub for stablecoin projects to source liquidity. For the longest time, Curve had nothing to worry about – they had their bread and butter, controlling the flow of stables and earning healthy revenues in the process. Then Uni V3 came along, introducing the concept of concentrated liquidity, and drastically changed the status quo.

In May 2021, when V3 was released, Uniswap began to eat into Curve’s market share. This was further exacerbated by the launch of 0.01% fee pools which provided an advantage over Curve. Due to the comparable price execution in V3 alongside a severe undercutting in fees, Uniswap has been able to match Curve’s volumes in stableswaps turning this into a nearly 50/50 market in a little over six months. What is even more impressive is that Uniswap has done this with virtually nil emissions. Compare that to Curve which emits $2M in CRV per day ($750M annualized). So what is the secret sauce that has let Uniswap compete with Curve in its own game?

Uniswap V2 utilized a constant product formula of x * y = k to determine the price of an asset. In order to ensure liquidity across all potential prices, Uniswap spreads an LP’s assets across the entire price curve from 0 to infinite. With V3 came the concept of concentrated liquidity – a more efficient way to provide liquidity and generate returns within a fixed price band. In the case of stablecoins, v3 allows LPs to provide liquidity in a very tight range, let’s say from $.99 cents to $1.01, allowing for significantly enhanced capital efficiency over V2. The introduction of Uniswap V3 alongside 0.01% fees allowed for the creation of hyper efficient pools for stablecoins that operate at extremely high utilization rates.

As we can see above the top 3 stableswap pools by volume on Uniswap destroy Curve’s top pools in the game of capital efficiency. They are all significantly more utilized then Curve’s pools and provide a much higher natural fee APR. So what are people missing in this type of analysis? What do people miss when they say Curve is spending nearly $750M a year to incentivize liquidity yet it’s losing volume market share to Uniswap?

The answer lies in Curve’s importance as a hub for deep liquidity rather than dominating everyday swaps.

This chart makes it clear what Curve has been spending all its emissions on – extremely deep liquidity. This liquidity on Curve is used to reinforce stability of stablecoin pegs rather than facilitate volume. A perfect example to highlight this was the MIM+3CRV pool during the week of Jan 24th, 2021. The MIM token came under pressure after news that 0xSifu, a close affiliate of Abracadabra founder Daniele Sestagalli, was associated with the now defunct Quadriga exchange that lost millions in customer funds.

Let’s take a look at some numbers from this extremely volatile event. Before any news had broken, the MIM+3CRV pool had $1.3B in MIM and $1.3B in 3CRV (3CRV = USDC, USDT, DAI). As news broke and the day progressed, due to a combination of LPs pulling liquidity and regular swaps, the pool dropped down to a ratio as low as $1.1B MIM and $42M 3CRV – a ratio of 96.5/3.5. Quite far off from its 50/50 target. During all this, the pool was able to handle a whopping $1.4B worth of volume. Even though the peg broke down to $0.97, all things considered, the MIM+3CRV pool handled the volume and volatility extremely well. It’s impossible to know exactly what would have happened to MIM if the pool didn’t exist, but its almost certain that the peg would have come under significantly more pressure without Curve.

This type of depth and liquidity is priceless to stablecoin protocols and the proof is in the pudding. If we recall back to the bribe section, Frax, Terra and Abracadabra were the top 3 protocols spending money on bribes to direct emissions to their pools (Frax and Terra are also top CVX holders). In addition to Convex, Abracadabra uses another service to tap into even more gauge weight power, called bribe.crv.finance, which allows them to bribe existing veCRV holders for emissions. The events of last week further strengthen these claims, as Abracadabra’s investment proved useful in enhancing MIM’s stability.

This clearly defines a product-market fit that Curve has carved out for itself, yet it may be hidden beneath volume headlines and gossip over veTokenomics. While Uniswap V3 and Curve are both operating in the same vein, it’s possible they can serve different purposes and can co-exist. Furthermore Curve is looking to expand its reach with Curve V2, extending its support to non-stable assets as it looks to compete with Uniswap on its own turf. Curve V2 is Curve’s version of concentrated liquidity that is optimized for “lazy liquidity”, automatically concentrating for LPs around a moving average of spot price. This provides a comparable result to Uniswap V3, yet with a more passive LP experience and access to Curve emissions as a bonus.

The Future of Convex

Jumping back to Convex, given the protocol’s reliance on Curve emissions, it has started diversifying its operations. This brings us to Frax Finance, which has been a supporter of Convex since the beginning.

Frax is currently the biggest spender on bribes and the single largest DAO holder of CVX. For a bit of background on Frax, they are a hybrid stablecoin that is both collateralized and algorithmic giving them a nice balance of stability and capital efficiency. Modeling after the success of Curve and Convex, Frax pivoted to the veToken model. This works similarly to Curve’s token economics, with users being able to lock up FXS for veFXS in order to gain: farming boosts, earnings from the protocol, and control over Frax gauges (emissions). A key difference with Frax gauges is that they can send emissions to any pool/protocol that integrates with the Frax stablecoin, unlike Curve which can only direct emissions within their own platform.

Out of this shift, cvxFXS was born – Convex’s liquid staking derivative for veFXS. In development is a fully functional staking platform for Frax LPs and a system for vlCVX to control the underlying gauge weight power. This marks the first expansion for Convex out of Curve and into other DeFi protocols.

This acts as a potential signpost for Convex to begin accumulating tokens of other protocols that adopt the veTOKEN model. However, adjustments will need to be made on a protocol level if Convex wishes to expand. Currently, CVX emissions are designed specifically to accumulate more veCRV and support the cvxCRV flywheel.

Therefore, if Convex wishes to expand its reach, it may need to reallocate some token emissions by either deploying CVX from the treasury or redirecting gauge power from their existing veCRV holdings. There is precedent for this with Votium directing 5% of all its delegated veCRV gauge power towards incentivizing the CRV/cvxCRV pool on Curve. With more protocols such as Ribbon, Dopex, and Yearn potentially embracing veTokenomics, we may see Convex leverage its large veCRV powers to incentivize cvxTOKEN/TOKEN pairs in order to accumulate more assets.

Convex has proved to be an extremely interesting project and has had profound effects on the Curve ecosystem and CRV price. We will be paying close attention to cvxCRV to see if it can maintain its attractiveness over veCRV in the coming months to years. As Convex looks beyond the Curve ecosystem and expands to include projects like Frax, we are eager to see how it can replicate its success and utilize the treasury and governance powers to accumulate even more assets. If they are successful in this endeavor, it’s possible Convex can be a “meta-governor index,” giving exposure to governance powers of the most important DeFi protocols.