APRs and APYs
The whole of DeFi uses the incentivized token reward model to incentivize liquidity. For example, a new AMM DEX protocol wishing to attract liquidity for a certain new pool of X and Y token, may decide to distribute their native token to Liquidity providers of this pool. Generally speaking, the real return or yield from a liquidity pool are in the trading fees that are accumulated from the trading volume of that pool. Trading fees are subject to market activity and not something that is fixed. Usually, a protocol will take the fees generated from trading volume over the past 24 hours to compute an estimated yearly annual return. This is not the best measure of what a user can expect to receive.
The protocol will also emit token rewards for this pool following some emission schedule. The per second(s) APR for these rewards is governed by the following relationship.
As we can see from this relationship, the APR is dependent on the USD value of the token emission for that day and the amount of liquidity providers that have staked their LP tokens into the reward contract. Therefore, both the trading APR and the reward APR are both variable. The estimate given by the protocol is literally just that and can change every day based on these factors.
There is another issue, if the reward APR is 10% and the token price of the reward token suddenly drops due to some FUD or market conditions, then the APR is also subject to volatility. It’s a double-edged sword with regards to reward tokens. If a certain amount of accumulated token rewards have been earned over a period of time, if you don’t realize these rewards as soon as they are earned your actual APR is subject to the volatility in the reward token. You could end up with a much higher yield if the token appreciates in value or a much lower yield if the token drops in value. Hence why V1 vaults aim to realize any yield from rewards as soon as they are earned or over a period of 24 hours by selling them for the vault’s asset type (sold for stable coins and redeposit them back into the strategy) to achieve a generic compound effect. In version 2 we may take a different more hands on approach to harvesting/compounding as our level of fund management will be greater in this version.
If the underlying strategy/pool is displaying a certain APR, then this is essentially made up of trading fees and incentivized rewards. The Incentivized rewards are what can be compounded and not the trading fees (as they accrue to the LP position).
If the underlying pool/strategy has an incentivized reward APR of 7%, this would translate to a 7.25% if we were to compound every 24 hours (CP = 365) for a year (Y = 1). This may not look like much but if we were to make Y = 10 years, the APY would be nearly 100%. Assuming you put 100 USDT into this pool without compounding you could assume (all things being constant) that 7% annually for 10 years would be equivalent to a 70% appreciation on your principal 100 USDT (170 USDT after 10 years). The difference in compounding is almost 30%. This difference increases exponentially over time as Y increases. The higher the TVL in Marhaba Emplifai vaults the more compounding becomes a logical undertaking as there are more rewards to compound and costs become cheaper and more distributed.
Say a user deploys the same vault strategy themselves. Without selling the reward tokens for some other stable asset you will be subject to the volatility in the reward token over time and attempting to compound any rewards may be irrational as gas costs will be more expensive for a single user with less funds. Rewards may also not be significant enough to compound. During bull markets the reward token APR may be higher as token prices usually soar in value. However, most users forget to take profits/realize gains in bull markets. These factors coupled with the time and navigation required to manually execute the underlying strategy makes the act of depositing into a vault a logical undertaking.
The vaults will have a raw APY (prior to fees being deducted) that is governed by the following:
Last updated