Balancer Labs — Set AMA Series
This is a transcript of a recent AMA that was held between the Co-Founder and CEO of Balancer, Fernando, and the Set community in our Discord channel. The AMA was held on the 16th of June at 1:00PM PDT, 2020.
You can check out Balancer here.
If you want to be involved in future AMA’s, be sure to join our Discord channel.
Inje: Hello everyone! Let’s get started!
Let’s give a warm welcome to Fernando from Balancer Labs! Fernando is the Co-Founder & CEO of Balancer — a non-custodial portfolio manager, liquidity provider, and price sensor. Fernando, it’s great to have you.
As a reminder for everyone participating — please keep the discussion respectful at all times.
Hey Fernando — feel free to give a quick intro about yourself and Balancer, and then we can dive into the questions!
Fernando: Hey guys! I’m Fernando, CEO & Co-founder of Balancer Labs, the company that’s pushing the development of Balancer protocol. The idea is for Balancer Labs to be dissolved in the future as we transition into a totally decentralized governance model.
We have recently started distributing BAL tokens (which still have not been deployed but will be soon) to liquidity providers on Balancer. Every week we distribute 145,000 BALs, i.e. ~7.5M per year.
In my opnion, the coolest thing about Balancer pools is their flexibility. You can have up to 8 tokens, flexible weights (from 2% to 98%) and a swap fee!
As you saw mentioned here in the chat, there are already pools made up of Sets! We are very excited with the idea of Sets being on Balancer as this enables them to be traded on another venue which = possibly more inbound liquidity and also gives BAL tokens to the LPs!
I can talk more about the types of pools if you like:
1) Private pools: only the creator can add liquidity but they are fully flexible (new tokens can be added, weights and swap fee can be changed etc)
2) Shared pools: anyone can provide liquidity and get pool tokens (BPT) in return. These pools are immutable to protect the liquidity providers from an attack by the pool creator.
Another cool concept is that of smart pools: you can have a smart contract create and control a private pool. Then anyone can send tokens to that controller which funnels them down to the private pool they control. That controller issues its own ERC20 shares that represent ownership of the private pool it controls.
This enables you to have the best of both worlds: anyone can provide liquidity to that private pool and the smart contract controlling it can still change things according to some well known logic. Example is surge fee pools: in times of liquidity crunch (think Black Thursday) these pools can increase the fees thus attracting more LPs and offering better trades. Everyone gets a better deal (think of UBER when it’s raining).
Reuptake: What happens when weights change in a private pool?
Fernando: If you change weights and balances proportionally then the prices don’t change at all. If you only change the weight and not the balance, then you are effectively changing the prices. This can create a dangerous arbitrage opportunity.
Jeremy Guzman: How are the Set-based pools weighed? Or can this be created as a private pool as well?
Rabmarut: The Set-based pools can be created by anyone with any weights. Right now I’ve created most of them. We have a 50/50 pool with ETHBTCRSI and ETHRSIAPY2, and a 40/40/20 pool with those two Sets plus 20% ETH.
Fernando: Yes, any pool can contain any ERC20 provided it’s fully compatible with the ERC20 standard. Now there are a few different combinations (for example with WETH and without WETH).
Rabmarut: There were a couple of different strategies I devised for pools of Sets. I mainly focused on the ETH/BTC RSI Set and the ETH/USD RSI APY Yield (now II) Set because I already held those and believed that one would perform very well in a bull market and the other very well in a bear market (if the end goal is to accumulate ETH).
On one hand, I believe a 50/50 pool with a high trading fee (~10%) can grow over time without any external influence beyond the minimum arbitrage required to maintain price parity with the underlying assets. I belive arbitrageurs would utilize this pool to make profits and as a result help liquidity providers to “swing trade” the two Sets and collect fees over time.
On the other hand, adding ETH to this pool (such as in the 40/40/20 format) and using a small trading fee allows a potential secondary market for Sets to form on Balancer. This pool, too, can perform very well over time but relies heavily on daily trading volume in order to grow.
I’ve constructed a model here: https://gitlab.com/rabmarut/balsim
And have performed a couple of backtests using the Sets backtest spreadsheets.
Inje: What are the unique characteristics of Balancer that differentiate it from Uniswap?
Fernando: Basically the flexibility. If you create a 50/50 Balancer pool with 0.3% fee the pool will be identical to a uniswap exchange. So you can create a pool with a controlled exposure. This 75%/25% MKR/WETH gives you a lot more upside if MKR goes up, compared to a 50/50 pool
Ganesh: Fernando, can you please walk us through a numerical example where the initial COMP listing on Uniswap left money on the table?
Fernando: Sure! This is the paper I’ll need to run you through the maths: https://balancer.finance/wp-content/uploads/2020/06/2020-06-11-Balancer_pool_value_calculation.pdf
You don’t need to look at it in depth now, but the relevant formula is this one (this formula shows the variation of the pool value in USD terms (or any other price reference)).
Let’s use the example of COMP listed on Uniswap. The exchange was created with 25,000 COMP and 2,000 ETH. That’s an initial price of $18.4 if we consider ETH at $230. Imagine COMP at $90 (it’s less than that now, but it went above $100 at some point) — this is a 5x increase!
So if you use the formula above, you can calculate how much the pool value changed from the time COMP was $18.4 compared to when $90.
A 50/50 pool like on Uniswap would have:
- Initial value = 2000*230 + 25000 * 18.4 = $920k
- Final value = Initial value * 5^0.5 = ~$2.057M
A 80/20 pool on Balancer would have:
- Same initial value = $920k
- Final value = initial value * 5^0.8 = ~$3.334M
So the LP would end up having about $1.3M more using Balancer.
In other words, Balancer allows you to get the upside of trading fees at the same time as you keep the exposure on the tokens you are more bullish (which Uniswap does not allow by forcing you to go 50/50).
Rabmarut: It sounds like one of the big features coming to Balancer soon is the distribution of BAL, and subsequently a governance mechanism for the protocol. Besides this, is the team working on anything else right now for the next iteration of Balancer?
Fernando: We are already working on Balancer V2 which will have many improvements in gas costs and new exciting features! Also, for all of you programmers out there we are hiring.
Inje: Care to share what those new features are?
Fernando: I can’t share all of them but the costs to deploy a pool and make a trade will be reduced by a lot. We’ll enable pools to set circuit breakers (which is something a lot of teams asked for) so that the pool does not buy more of a token if its price reached a floor.
One problem that AMMs have is that if one of the tokens in the pool has a infinite supply bug they can drain all the pool value. With our V2 circuit breakers your pool will just stop absorbing that token at some point, protecting the value of the pool.
Rabmarut: Do you see Balancer as more of an index fund factory or a market making utility? Or is it truly both? And can you explain some of the possible use cases of smart pools?
Fernando: I think it’s both. This post shows how high fees can be used to make Balancer feel more like portfolio manager. But my guess is that there will be a lot more liquidity around pools with low fees (seeking retail trading fees and BAL tokens) than high-fee pools.
Rabmarut: Do you think there is a place for both high and low-fee pools on Balancer, or do you think over time all pools might trend toward some equilibrium fee (such as Uniswap’s 0.3%)?
Fernando: I do think they optimize for different trade-offs, so I expect both to exist in the future. Maybe private pools could be used more as a wealth/treasury management tool so could be using on average higher fees than shared pools.
Ganesh: In Uniswap, the ratio of the token balances is the exchange rate and some teams inadvertently use that as an oracle. They have since fixed that in V2 by checkpointing the balances so that you don’t drastically change the balances/hack the oracle with a flash loan. What can you do with the token balances on Balancer, say for a 4-token pool? Or conceptually what does that mean?
Fernando: Sure, that’s something we are considering for V2. You could have a Uniswap-like price accumulator for each pair. Balancer V1 is like Uniswap V1, very dangerous to be used as a price oracle because of sandwich attacks.
mjc716: Are you making a concerted effort to be a venue for initial DEX offerings? Or are the COMP and UMA examples you’ve mentioned more just a way to demonstrate Balancer use case/value prop?
Fernando: We want to be seen as a building block, a primitive that other projects can build on. We really believe in that value prop of starting liquidity for new projects, see this post by our CTO Mike. But we are not forcing the market to use this concept. If it’s good and finds product market fit teams will eventually naturally start using it.
So the answer is: we’ll let the market decide if that’s a cool value prop. We expect Balancer to be used in cool/crazy ways we hadn’t anticipated
mjc716: I agree mostly, but there’s still sort of a Schelling Point dynamic with Uniswap that could be broken with some features built specifically for this value prop in my opinion.
Fernando: Great point! We are actively building a smart pool factory that will serve as a template for teams to deploy LPBs (liquidity bootstrapping pools) very easily. I think one of the major hurdles at the moment is the daunting task of programming a pool that flips weights. But we are doing that so any team can just deploy it out-of-the-box.
Rabmarut: Do you think it may become much more commonplace over time for crypto investors to hold baskets of assets (BPTs) instead of just individual assets? In traditional (retail) finance, it is much more common to invest in a fund than in an individual stock, but right now in crypto people are holding individual tokens. Will protocols like Balancer enable a shift in the investing mindset of crypto communities?
Fernando: Great point! I do think that we’ll see many of the trends that happened in the conventional finance world (like index funds getting more widespread) happening also in DeFi. We are seeing synthetics (e.g. UMA and Synthetix) growing a lot in adoption. Expect the same with options and derivatives.
I think the magic of DeFi (which can’t be done in traditional finance) is the trustless composability of protocols. Insuring things in a decentralized fashion with Opyn or Nexus, opening long/short positions, buying stocks synthetics etc all in atomic transactions. This is going to revolutionize finance as we know it.
Inje: Do you guys plan on adding protocol level fee? Or are you leaving that up to BAL token holders to decide?
Fernando: Right — this will be definitely a very important decision that the governance will make. If BAL holders decide to introduce a protocol fee it has to be done at the right moment (the protocol must have proved itself already and be adding real value to all stakeholders, be they LPs or traders) and also at the right amount (high fees can cause a fork).
Rabmarut: Piggybacking on that, what do you expect the governance levers to be for BAL holders? Besides a fee collection, what else might holders want to vote on?
Fernando: Deploying and maintaining Balancer on other layer 1s (AVALabs, Near, Polkadot etc), deciding to go to a layer 2 on Ethereum (which one?), adding new features, giving out sizable grants etc.
Inje: Is there a plan on who will implement those upgrades once Balancer is dissolved and governance is up to the community?
Fernando: This is a great question that many projects are working on @Inje. We have a lot to learn from other projects like MakerDAO, Compound and other decentralized governance ones. So the answer is “we’ll see =)”
