A quick caveat here is that none of this is financial advice, and to do your own homework. Also, I am learning a lot of this as I go, so am by no means an expert. Teaching (or attempting to teach), I believe is a great way to learn.
If you dive down the crypto rabbit-hole, you will inevitably come across the concept of liquidity mining. My goal below is to explain 1) what liquidity mining (ie yield farming) is, 2) how (and why) you earn rewards for doing this, 3) what the risks are and 4) a way to think about balancing risk and return, especially as it pertains to a concept called 'impermanent loss'.
A brief explanation on what liquidity mining is:
- When you want to exchange one asset for another, you typically will do so by going to a centralized exchange like Coinbase. If you have Bitcoin but want Ethereum, you can go to Coinbase to trade one for the other. This is much more efficient than hoping to find someone that wants to trade with you, so centralized exchanges exist to make this process faster. They have the Ethereum on-hand to give you in exchange for your Bitcoin.
- Primarily due to the advent of smart contracts and the movement towards decentralization brought upon by blockchains, a new crop of exchanges have popped up called decentralized exchanges (DEXs) or automated market makers (AMMs). They are typically "protocols" not "companies" (though often there are "companies" that start them), meaning they exist as code, and are governed primarily by the community (through what's called a DAO).
- Similar to Coinbase, you can go to one of these, like Uniswap, and exchange one asset for another. But, if there is no centralized clearing house, how is there enough of each asset to ensure "liquidity" (ie enough of the asset for efficient exchanges)?
- Liquidity providers offer up their assets to liquidity pools for people to buy from. And they earn fees for doing this as a reward. This is what "liquidity mining" is. You provide your assets to a pool and earn fees for providing this service.
- Those fees are typically in two forms:
- 1) Trading fees when others buy and sell within the pool (often ~0.3%)
- 2) Rewards typically in the form of the protocol's native governance token
So how do those magical rewards work and why do you get them?
- Providing liquidity involves risk (which we will get to later), and so the protocol must incentivize people to provide this liquidity or else they no longer have a liquid exchange because nobody would want to offer up their assets to be bought and sold.
- The solution many of these decentralized exchanges have come to is to offer a native governance token to liquidity providers in a periodic way, proportional to the amount of liquidity being provided. If you provide 10% of the liquidity pool, you get 10% of the daily (or hourly) rewards (again, often in the form of the 'governance token') issued by the protocol in that pool for doing so.
- That governance token has value in the open market (the value of governance tokens conceptually is for another time), and thus, you can convert those rewards into an effective APY based on the amount of rewards you earn relative to the liquidity in the pool you have provided.
What is an example of one of these protocols?
- There is a protocol called Mirror that is seeking to construct a decentralized exchange for synthetic assets. A synthetic asset is just a representation of an asset tied to a real-world price with an oracle. For Mirror, a synthetic asset might be a 'share' of Tesla stock, with the 'ticker' $mTSLA, as an example. This gives someone the ability to buy a 'share' of Tesla in a decentralized manner (ie without a centralized broker like Robinhood), 24/7, globally without any barriers. There is currently a value of over $400 million of these synthetic assets created on the Mirror protocol, with about $50-75 million in daily trading volume. (Note: Mirror is seeing massive adoption in certain countries without access to traditional stock markets, which is an awesome real world use case for Defi!)
- If you want, you can provide liquidity to one of the trading pairs to earn rewards, often in excess of 50% APY (sometimes up closer to 200%+).
- For example, you could provide UST (dollar equivalent stablecoin in the Terra ecosystem) and $mTSLA in equal amounts (buy both and provide to the pool), and then earn $MIR (the governance token of the Mirror protocol) as rewards every day! If you want, you can sell your $MIR for UST to earn dollars and hold no risk on the actual $MIR price fluctuation.
OK, what are the risks of doing that? If I can earn 50+% APRs there must be a catch?
- 1) Protocol and smart contract risk
- Almost anything in crypto holds some protocol risk. By definition, much of this ecosystem operates without the 'normal' backstops of legal protections and formal institutional protections. Nearly all of these ecosystems operate on coded contracts, and are therefore subject to hacks and security breaches.
- 2) Impermanent loss risk
- This one is a bit counterintuitive and can take some time to understand. But here's an attempt to explain:
- When you provide liquidity, you provide an equal value of each asset. If you provide $10,000 in ETH, you need to provide $10,000 in USD, if you are providing liquidity to the ETH/USD liquidity pool.
- Over time, the ETH/USD price will change. In order to maintain equal value of each side of the pair within the liquidity pool, the protocol will naturally change the amount of each side of the pair to do so.
- Typically this means selling the asset that is appreciating in value, and/or buying the asset that is depreciating in value.
- Given the behavior above, holding either of the assets on their own will always be a superior strategy to providing this liquidity, IF you didn't earn rewards. You can work the math out on your own (my spreadsheet is included at the end of this post), but it makes intuitive sense. If you are holding two assets but selling one on the way up and buying on the way down, this strategy will be worse than just holding the assets to begin with. The 'loss' that occurs as a result of the price fluctuation is the "impermanent loss" (the loss of value due to this rebalancing) but is only permanent when you remove your liquidity. If the price ratio is the same when you remove liquidity as it was when you added it, then there is no loss.
Here is a chart that shows how a change in price can drive impermanent loss. For example, if you provide ETH/USD and there is a 200% increase in the price of ETH (against USD), you will incur ~5% impermanent loss, which means you will have 5% less total value than if you had just bought ETH. If ETH goes down 70%, you will additionally accrue about 25% in impermanent loss, amplifying the painful situation of a 70% drop.
But, again, while impermanent loss is always there to an extent (prices never stay exactly the same over time), if you earn enough in fees as a benefit of providing liquidity, you can overcome this risk.
So how should you think about evaluating liquidity mining opportunities and mitigating these risks?
- 1) How long do you plan to keep your assets in the liquidity pool?
- The longer you hold, the more fees you earn over time.
- 2) How volatile is the asset pair you are providing liquidity to?
- The more volatile an asset is, the more likely it is to diverge in price and accrue impermanent loss.
- 3) How much do you believe in the native token that is being issued as a fee?
- If the native asset depreciates compared to USD, your earned rewards will be lower in USD terms, lowering the effective APY.
- 4) How much do you believe in the asset being provided as liquidity (relative to USD)?
- If you are providing liquidity for ETH/USD and ETH drops in price, you will need even higher liquidity mining fees to overcome the drop in price.
Here is a simplification of the above:
- 1) Length of time in pool
- More time → Greater returns (more time, more rewards)
- Less time → Lower returns (less time, less rewards)
- 2) Volatility of asset pair
- Less volatility → Greater returns (less impermanent loss)
- More volatility → Lower returns (more impermanent loss)
- 3) Native token appreciation
- More appreciation → Greater returns (your token rewards worth more in USD)
- Less appreciation → Lower returns (your token rewards worth less in USD)
- 4) Liquidity asset appreciation
- More appreciation → Greater returns
- Less appreciation → Lower returns
So how can you think about approaching any new liquidity mining opportunity?
First, you want to avoid "rug pull" risk and smart contract risk but only working with great projects. This is where you just need to use your common sense. (There are some tools cropping up to help understand some of these risks such as RugDoc). If the project has a large amount of liquidity provided, a strong team, a review/audit of its smart contracts, and a large community around it (Discord, Telegram, etc), you should be fine. But, do your homework and in crypto, nothing is guaranteed.
Second, I would think about the asset you are providing liquidity for. If you don't want to hold the asset, you don't want to provide liquidity for it. You still own the asset, so if you don't want to own it, you don't want to provide it as liquidity.
Third, think about the native token rewards. Do you believe that token will appreciate? Rewards in the form of a token that is worth $0 isn't much good.
Fourth, think about the volatility of the asset compared to the rewards and your expected time horizon. This part I have tried to model out in excel. If you are providing for a highly volatile asset, with low APY rewards, over a short time horizon, you open yourself up to high risk of impermanent loss problems.
Let's say the asset you're providing for goes up 100% in a month. This is great, no doubt! But you would have done ~7% better by just buying the asset itself and not providing liquidity! And the rewards may not have had time to make up for the impermanent loss.
I created a model with a variety of price change scenarios to see what the average expected impermanent loss would be given various changes in prices. Then, looking at the number of days in the liquidity pool, you can calculate the fee returns. The document is below. It's a lot to understand, and I haven't checked everything and made it super pretty, but feel free to reach out if you'd like to go deeper.
Below is what I think is an interesting and unique output of this analysis. It shows the average performance improvement over just holding the asset you are providing liquidity for, given a certain APY, a certain amount of time, and an expected impermanent loss.
So, what is this chart actually saying in practical terms?
It shows how long you should anticipate holding an asset in a liquidity pool to overcome the potential impermanent loss given the APY of rewards. For example, if you are providing liquidity to a pool with 15% APY, all else equal, you will want to hold your assets in the pool at least for about 65 days.
The 'all else equal' caveat is important. Rewards in a valueless token or providing liquidity for an asset that dramatically drops in price will be more problematic for your returns than any short term impermanent loss, but valuation of specific tokens or assets is for a much longer discussion :).
Ultimately, the interplay of these factors is what drives behavior and the relative performance, liquidity and stability of each of these decentralized exchanges. Understanding these dynamics can help you make better decisions about the projects to support and where to provide your capital to earn the highest returns relative to the risk.
Feel free to reach out to me on twitter @jim_shook if you have questions or would like to discuss more! There's a ton for me to learn, so I'm always interested in feedback, additions or suggestions too. Also, leave your email below if you want very periodic updates!