The problem with Anchor rewards

Anchor currently pays out three types of rewards:

• Reward for borrowing

• Reward for staking the ANC token

• Reward for providing liquidity in the ANC-UST liquidity pool

There is one major issue with the on-chain data on these rewards, however. Currently, only the rewards for providing liquidity are labelled. This means that the transactions where a user claims rewards from borrowing and transactions where a user claims rewards from staking the ANC token cannot be differentiated. Therefore, we have to come up with a clever solution to be able to differentiate these rewards in another way.

Introduction

Anchor is a decentralized savings protocol on the Terra blockchain. It allows users to deposit UST to earn a stable 20% APY in the EARN section. Anchor also allows users to borrow UST. Users can provide bLUNA and bETH as collateral and borrow up to 80% of their collateral value in UST. Naturally, users have to pay a borrowing interest in order to be able to borrow UST. However, to incentivise borrowing, Anchor currently pays out a Distribution APY in the ANC token. This can be seen as a reward for using the platform by borrowing. This dashboard aims to analyse what users do with these borrowing rewards. Specifically, we will look at the percentage of borrowing rewards that are sold for UST within 7 days of receiving them. This percentage can give insight into the demand for the ANC token.

Approach

In this dashboard, the following approach was chosen to differentiate the borrowing rewards from the staking rewards. We only looked at users that use Anchor for borrowing, so they do not stake ANC tokens and do not provide liquidity. This means that the rewards this group of users receives can only come from borrowing.

The next step is to find out the percentage of these rewards that are sold within 7 days. This was done by gathering all the ANC to UST swaps by the "borrowing only" addresses. Then we posed three conditions: the swap had to be from ANC to UST, the swap had to be at most one week after the claim of borrowing rewards, and finally, the swapped amount of ANC had to be equal to the amount that was received by the claim of borrowing rewards. We will consider swaps on Astroport and Anchor.

Of course, this approach is not ideal. For starters, all users that borrow as well as stake and/or provide liquidity, are not considered in this dashboard. Secondly, we only 'count' a swap if the exact amount of ANC that is claimed from borrowing is sold for UST. This means that if only a part of the rewards is sold for UST, the swap is not considered. Finally, we only consider swaps on Anchor and Astroport. This means that if users were to sell their rewards on, for instance LOOP Markets, the swap is not 'counted' as well.

Results

The approach above resulted in the following numbers:

Conclusion

This dashboard aimed to analyse the percentage of borrow rewards that are sold within 7 days. In order to answer this question, a few constraints had to be defined. In order to isolate borrow rewards, only users that borrow on Anchor were considered. Also, we only considered the reward as sold if the exact reward amount was swapped for UST. Finally, we only considered selling ANC on Anchor and Astroport. With these constraints, we were able to get to the following key findings:

• Over the last 6 months, about 42% of borrowing rewards were sold within one week.

• The percentage of rewards that were being sold, dropped from about 60% in October 2021 to about 30% in January 2022.

• There has been a slight increase in the selling of rewards from January to now.

• In order to reduce reward selling, more financial benefits to holding ANC are necessary.

First, we look at the number of claims of borrow rewards over the last 6 months. All the claims of users that only borrow on Anchor, that is.

The number above represents the number of claims where the reward was sold for UST within 7 days on Astroport or Anchor over the last 6 months.

This means that over the last 6 months, about 42% of all claimed borrow rewards were sold for UST within 7 days. I think this is a surprisingly high percentage given the conditions we defined in the approach section. We only look at swaps where users swapped the exact amount they claimed in borrowing rewards. We are not even considering users that sold only a part of their rewards!

Next, we will look at this percentage per day over the last 6 months.

As can be seen, the percentage of borrow rewards that are swapped to UST was over 50% during October 2021. From November 2021 up to mid-January 2022, the percentage dropped to about 30%. This drop meant that more users were not immediately selling their borrowing rewards. Since mid-January, the percentage seems to even have increased back to about 40%.

How to reduce reward selling

Rewards are usually quickly sold off if there is little utility for the token. I think is this the case for the ANC token as well. Aside from the usual staking and liquidity pooling, the only other incentive to hold ANC is to participate in governance proposals. I think that participating in governance is not enough of an incentive to keep users from selling their ANC tokens. I think more users would not sell their rewards if there were more financial benefits to holding the ANC token. Examples:

• Increased APY in EARN

• Reduced cost of borrowing

• Increased LUNA/bLUNA rates, so receiving more LUNA when burning bLUNA.

Of course, no research was done if any of these suggestions are actually possible. But I think these kinds of benefits to holding the ANC token are necessary in order to keep users from selling their rewards.