Incentivized Liquidity and Yield Farming in DeFi
Jul 07 2020 · 13 min read
A panel hosted by D1 Ventures and ChainNews with Synthetix, Balancer Labs, Ren Protocol, and Curve Finance.
The DeFi landscape is evolving into a machine where design experimentation with liquidity incentivizations, governance, and UX are being churned out by a number of projects. D1 and 链闻 ChainNews are lucky to sit down with leaders in the space to cover the growth of DeFi and how their projects are tackling current problems. Below is the transcript from our virtual panel.
Q1: Can you please briefly introduce yourselves and tell us what is the project you’re working on and why you’re so passionate about DeFi?
Kain: I’m Kain Warwick, founder of Synthetix, which is a derivatives liquidity protocol on Ethereum. The Synthetix protocol uses pooled collateral to enable derivatives trading with infinite liquidity and zero slippage. At the moment that trading is mostly done through synthetic assets in the form of individual tokens, but we also launched binary options yesterday and we’ve got synthetic futures launching in the next few months.
DeFi promises permissionless finance for the first time, which is important both for censorship resistance and for ensuring market forces can operate properly. Part of that permissionlessness is composability, or different protocols plugging into each other.
We’ve already seen it enable some serious leaps forward for user experience, like 1inch Exchange which aggregates decentralized liquidity across the ecosystem and allows users to access it from one interface. This is different to the current permissioned Internet, where if you want to interact with Twitter in some way for example, you usually need to interact with their centralized APIs, which are often extremely limited in what you can do and require you to jump through various hoops, rather than just easily building on top of it.
Taiyang: I’m Taiyang Zhang, the CEO of Ren which builds RenVM, a multi-party computation (MPC) protocol. We recently launched on mainnet and the first use case we’ve brought to market is a bridge between BTC, BCH, ZEC and Ethereum. Previously, there was no easy way for DeFi applications built on Ethereum to access these assets. RenVM makes this process simple.
I’m excited for DeFi as it expands the idea that if the most efficient and secure way to transfer value is using a blockchain, the financial infrastructure that this value interacts with should also be built on top of decentralized infrastructure.
Michael: I’m Michael Egorov, Founder of Curve Finance. Curve is a protocol for creating very deep liquidity for similarly priced assets (at least currently) — stablecoins and bitcoin on ethereum right now.
I am an active (daily) user of Crypto since 2014 and DeFi since 2018. Today I can quite imagine life without access to the banking system, but not without crypto and defi. Liquidity depth on chain, and for stablecoins in particular, was quite a painpoint for me which wasn’t adequately solved by centralized exchanges. So I decided to solve it for everyone as well.
It basically allows anyone to create customized liquidity pools that are continuously rebalanced by arbitrageurs. These pools behave like index funds where, instead of the liquidity provider having to pay a fee like what happens in conventional finance, LPs get to charge a fee for traders to use the liquidity in their pools.
Q2: Why is yield farming so popular as a phenomenon and also in the context of memes? And, which marketing approaches are you using and how are the results?
Kain: Yeah so it’s always been my belief that incentives are really important for user behaviour, so at the start of 2019 when Synthetix was in the process of gaining traction, we changed the monetary policy to an inflationary model in March after discussions with the community, which meant that SNX holders who weren’t staking were essentially punished for not participating in the system by being inflated out. This took staking activity from something like 20% to around 80%, and clearly demonstrated that incentives are critical for an emerging network.
As for liquidity mining, it started with an sETH/ETH liquidity pool on Uniswap. We had been talking with the community about getting an on-ramp to Synthetix.Exchange, allowing people to enter the system and start trading easily even if they didn’t hold any Synths. Initially the talk was about an sUSD/ETH liquidity pool on Uniswap, but a community member suggested we go with sETH/ETH, because these two assets should always trade at parity so shifts in value wouldn’t require arbitrage. The incentives came from the SNX inflationary rewards, a small portion of which were diverted from going to SNX stakers to going to liquidity providers for this sETH/ETH pool. Initially it was a pretty clunky, manual process, but over time it was automated, and has been going pretty steadily until just last week when it was finally deprecated, as we now have an incentivised sUSD liquidity pool with other stablecoins on Curve.
One thing we underestimated at the time was this initiative’s power in introducing new users to the Synthetix protocol. Plenty of people have told us about how they heard about this incentive on Crypto Twitter or somewhere else and decided to take part, just as a way of farming some yield on their ETH, and then ended up doing a bit more research about what the project is doing and decided to use those SNX rewards or buy more SNX to actually become an active staker in the system.
Kain: We’ve currently got three liquidity incentives. There is the sUSD Curve pool, which is a pool of stablecoins and allows people to go in and out of sUSD easily as an on/off ramp to Synthetix.Exchange. We pay out a portion of the SNX inflationary supply each week to this pool, much like we did previously with the sETH/ETH pool. This is currently the largest AMM in DeFi.
Then there’s the ‘Bitcoin on Ethereum’ liquidity pool, also on Curve, which pools sBTC, renBTC, and Wrapped BTC. By contributing to this pool, liquidity providers receive BPT liquidity tokens from an SNX and REN pool on Balancer set up by the two teams, which also entitles them to BAL tokens and CRV tokens when they are released.
There’s also an iETH incentive, which is our short ETH token, which inversely tracks the price of ETH. The idea behind that one is to balance out the debt pool, much of which is long either ETH or BTC. We’ve actually spoken recently in the community about starting an iBTC pool to encourage people to hedge out the long BTC bias.
Taiyang: DeFi farming really took off with the launch of COMP, Compound Finance’s governance token. With COMP listing at a high valuation, the notional return for lending on Compound became much higher due to an additional distribution of COMP tokens as an incentive. Borrowers were also incentivized the same way, and this is why we’ve seen such large growth in Compound’s numbers.
In the context of a meme, I first saw a few prominent Twitter users using the phrase “farming yield”. It’s a concept that’s easy to conceptualize: being able to “farm” a new token by depositing existing assets. Structurally, it’s the same as staking assets within a protocol, but instead of being paid out through native token inflation, stakers are rewarded with a different token, or a new “crop” so to speak in the farming context.
Taiyang: As Kain mentioned, we’ve launched an incentivized renBTC, sBTC and WBTC pool on Curve Finance. RenBTC is our protocol’s wrapped Bitcoin on Ethereum. From our perspective, we’ve seen over 800 renBTC deposited into Curve pools to date and as a consequence, over $25mil volume moving through the RenVM bridge.
By incentivizing this pool, we’ve seen much deeper liquidity between renBTC and other pairs, so I’d say it’s been a success.
Michael: Popularity of “DeFi farming” is not a surprise at all. Even way before working on Curve I thought that incentivizing users to help the protocol is the way to use token supply, coming up with WorkLock mechanism for NuCypher, as a much more healthy alternative to an ICO. But certainly, in practice and in DeFi world, Synthetix pioneered those incentive mechanisms.
It was really interesting to see how “farming” became a meme over a course of days and captured all the crypto-Twitter. People are looking for ways to make money, and DeFi farming offers a very healthy, much lower risk alternative to ICOs.
Michael: Currently, as Kain said, we have SNX incentives for SUSD and SNX+REN incentives for SBTC pool. We, however, will very soon launch Curve CRV token with incentive mechanisms. Starting initially with liquidity incentives, we also want to explore incentives for insurance and for trading volume. Trading volume incentives, however, should be approached very carefully to not cause fake volumes like FCoin did (although even in such case traders would have to pay fees to liquidity providers). We will keep architecture modular, so that different things can be incentivized from token inflation.
I think liquidity mining goes much beyond the meme: it’s about decentralization, about getting users of your protocol staked: it’s about inviting them to be owners of the protocol themselves. This is the only way I think a protocol can get a healthy token distribution and become really decentralized. And without a healthy token distribution, onchain governance doesn’t mean anything. It’s just governance theater.
We have started with a very simple proposal for liquidity providers on Balancer to get BAL governance tokens. You get a proportional amount of a fixed 145k BAL per week based on how much of the total liquidity you contributed. We knew there were some intricacies that should be taken into account like how even a pool is but wanted the community itself to step up to suggest improvements to the initial proposal. It worked really well and we now have 3,000 users on our discord, of which many are actively participating in daily discussions.
Q3: What are the risks involved with yield farming and how many different levels of risks are there? Are there any lessons you want to share with our audiences?
Taiyang: The Ren, Synthetix and Curve collaboration is a rather unique case as it makes sense for there to exist an easy on-ramp and off-ramp between sBTC on the Synthetix side and renBTC on our side. These are highly correlated assets and hence why we decided to incentivize the pool. I can’t say I see this becoming more common unless a particular pair of assets being incentivized are related in some way. However, I think a lot of projects are already “collaborating” implicitly through the composability of DeFi. For example, it’s possible to “farm” BAL and COMP at the same time by depositing USDC into Compound, turning the deposit into cUSDC tokens and then depositing them into a Balancer pool.
Taiyang: I’d be wary when saying that it will be different this time for all DeFi projects, because some projects will hit the mark, and others won’t. We’ve seen success with projects like Synthetix and Balancer, and I can see Curve following the same path.
The primary difference between the dynamic of trans-mining and DeFi farming is that in most cases within DeFi farming, users are not paying “directly” to “mine” tokens. Rather, users are supplying assets to a protocol, which does not come at an upfront cost, but rather opportunity cost.
One area to be cautious is when interacting with DeFi protocols that enable one to employ leverage. Many users don’t understand the risks of liquidation within these protocols and as an effect, they may lose collateral they supply. It’s important for any user to do proper due diligence when supplying assets to a DeFi application.
Fernando: I don’t know much about these previous attempts so can’t speak about them. But one thing is for sure: giving away tokens of a protocol to attract users that won’t add real value to the protocol is not long term sustainable. The reason why I think liquidity mining on Balancer is long term sustainable is because the more liquidity there is on Balancer, the better the protocol can serve traders and the more traders there is, the more trading fees (profit) the pools have. It’s a flywheel effect that becomes very powerful and ideally in the long term does not depend on the token distribution anymore.
Fernando: The community was very fast in raising the issue: we quickly reached a consensus that this liquidity was not legitimate in the sense that it was not useful for traders using Balancer. The more controversial discussion was whether we would block this attacker retroactively so they would not get any BAL or if we would avoid at all costs making retroactive changes to the liquidity mining rules. The community ended up deciding not to open up a precedent of changing rules retroactively, which in my opinion was the right thing.
Kain: The exciting thing about DeFi right now is that its genuinely allowing people to do things they haven’t been able to do before. For example, these lending protocols are allowing people to earn a yield on their funds in a totally permissionless way, and with the current state of various global economies this is more important than ever.
The contrast to the examples you listed, like games on EOS and TRON, is that blockchain technology is not even close yet to the point at which it’s offering a better user experience for games than non-blockchain examples. This means that though there might have been an initial bootstrapping method for drawing users, once those rewards or the value of them dried up there was no reason for them to stick around. It’s not definite yet, and a lot can change, but it looks more likely that the value props of DeFi will be sufficient at some point to keep people around, because the user experience is in some way superior to off-chain comparisons, whether it’s for the yield, or the permissionless nature of these systems, or something else.
Kain: Each week, SNX staking rewards are given out to SNX stakers in the form of inflationary SNX tokens locked for one year, and sUSD fees generated by trading on Synthetix.Exchange. This is measured by the start and end of each fee period, once a week, as staking rewards are given to whoever is staking at the end of the period. Some time last year some SNX holders were doing something we called “snapshotting,” which was gaming the system by only staking for, say, an hour over the fee period snapshot, which mean they got rewards without taking on the risk of the debt pool of traders’ profits and losses on Synthetix.Exchange, which is one of the key reasons why stakers receive rewards.
We resolved it with SIP-40. SIPs are Synthetix Improvement Proposals, which are written by either Core Contributors or community members. SIP-40 implemented a 24 hour time lock on unstaking after you’ve staked, but another important element was that we dramatically reduced the risk of stakers’ exposure to the debt pool by preventing ‘frontrunning,’ which was traders taking advantage of delays in pushing price feeds on-chain. This was resolved by SIP-37, which ensured there was a lot less incentive to only stake around the fee period snapshot.
The important thing with any incentive is to monitor it, to ensure it’s serving its purpose, and depending on your governance structure, try to stay agile enough to combat any incentive gaming quickly.
Michael: This is a good point that if what’s being incentivized is not useful, people will come and leave. The point is to either introduce more people to a protocol which is juicy regardless of the temporary incentives, or to have a permanent incentive mechanism with token demand embedded in token economics, or better both.
On Curve, we do have natural (yield) incentives due to effectiveness of our algorithms. Funnily enough, this allowed us to triple our total value locked with the introduction of COMP liquidity mining program: returns for Curve liquidity providers were very much elevated due to stablecoin swap demand coming from COMP farmers.
Q4: What do you think are some unresolved but important problems in the DeFi today? Where do you see new opportunities in the space?
Kain: Obviously with the gas prices as they are, Layer 2 solutions can’t come quickly enough, which is why we prioritised our recent demo with the Optimism team’s OVM solution, to demonstrate the serious improvement in user experience for trading. Whichever L2 solution/s can get a headstart there will see a lot of adoption.
As ever, UX could still be improved significantly, and there are always milestones like Metamask that individually seem to take the ecosystem miles forward in one leap, although of course they’re building on countless iterations before them. But I’m not the kind of person to worry about UI/UX like you see some people doing — if the DeFi value prop is really shown to be there in the next few years, then there’s going to be sufficient incentive for people to jump in to solve its problems.
Taiyang: I think some of the most important problems in DeFi are the experimentation and formalization of governance mechanisms and implementation of a widely-used privacy solution. On the governance side, it’s possible to make changes to protocols which can introduce security vulnerabilities, so it’s important we experiment in a safe manner.
For privacy, I think there’s a huge opportunity to ensure financial transactions between the user and DeFi applications stay private. It’s important for a user to be able to make trades, take out loans and interact with the whole DeFi ecosystem without every action being public.
As a final note, it will be interesting to see how ETH 2.0 changes DeFi with regards to composability, scaling and much more.
Michael: Different farming schemes can have different risks. For example farming COMP may have a problem that building a “farm” for maximum profit may involve being very close to liquidations. Liquidity mining on AMMs should potentially be a relatively safe thing, however one must not forget about risks of underlying protocols which are composed in (e.g. lending protocols being gamed or hacked), price risks of assets being used (which, in case of stablecoins, is a risk of permanent depeg), and, as always, risks of smart contracts having critical bugs.
Writing safe smart contract is no easy, and this is currently a big challenge. We need better tools and languages than what people use today, and we at Curve hope to help making a dent on that front.
Apart from safety, DeFi currently has challenges with gas usage. Every on-chain transaction is really expensive, and Ethereum 2.0 is nowhere near close. Going to different blockchains is not an option, too, because that hurts composability. Of course, L2 can be very helpful, but having each project being on its own L2 would have problems with compatibility, also. Curiously enough, the success of bonding curves can be attributed to expensiveness of on-chain transactions which make order books infeasible (remember EtherDelta?). I think, bonding curves is a superior approach in principle which had a chance to shine due to, essentially, scalability problems of blockchains.
Fernando: We need better usability, scalability and better tools for the average people to be able to use crypto. We are seeing already amazing wallets (like Argent) and services (like Zapper.fi and Zerion) which do a great job at simplifying DeFi for it to be more inclusive, user friendly and reach wider audiences.
Another important aspect is more strict standardization. A lot of important tokens simply don’t follow the ERC20 standard which causes major headaches for protocols which intend to be permissionless and censorship resistant like Balancer. A notable example is USDT that does not return a boolean on ERC20 transfers.