Establish Parameters for Yield Reserve

Establishing Parameters for the Yield Reserve account should be a top priority of the community as it directly effects staking rewards and can increase them in a major way. As of right now, 90% of the incoming yield from deposited bLuna and protocol fees are going to the Yield reserve wallet, which has approx $560,000 UST and can be seen here: terra1tmnqgvg567ypvsvk6rwsga3srp7e3lg6u0elp8 and 10% are used to purchase ANC and distribute as staking rewards.

As a community, we need to agree on a threshold and metric for the Yield Reserve account and what is to be done with the surplus funds once the threshold is met. The purpose of this post is to start that discussion.

I propose we set the Yield Reserve threshold at 30 days of yield payout. In simplest terms possible, this means that the Reserve wallet could cover the Yield payments on all deposits for 30 days if the inflow was zero. So, currently there is $168m UST on deposit, at 20.22% means $33,969,600 per year in payouts, which comes out to $93,067 daily and about $2,800,000 for 30 days.

Once the Yield Reserve wallet meets this threshold, so it has a balance $2,800,000, then 100% of the surplus is used to buyback ANC and distribute as staking rewards. By my math, at current deposit and borrow levels we should reach this threshold in a little less than 30 days. After the threshold is reached, instead of 10% of inflows getting paid out as staking rewards, we are talking about 100%, so it is a very simple and handsome 10x of staking rewards. Not to mention it will generate an additional $100,000 worth of buying pressure on ANC daily.

As deposits increase, so does the Yield reserve threshold balance, as it will require more UST for 30 days of payout.

Thoughts? Feedback? Ideas?


That’s an interesting idea - and I think it’s an interesting metric to define the reserve amount. it would surely incentivize many to stake ANC and increase ANC buying pressure drastically.

However, I think the main purpose of ANC in the terra ecosystem is to function as a savings account, in both bull and bear markets - hence 30 days is unlikely to be sufficient.


Fully agreed with the idea in principle following our discussion on Telegram.

Worth noting that Anchor reserves should be ready for sudden deposit increases say if a major fintech launches an Anchor integration. It is possible that deposits increase very quickly by say $0.5bn, meaning that what has so far been a month’s reserve now becomes only good for a week.

What deposit levels can be expected in say 3 to 6 months? And how fast can borrowing interest catch up to ease the potential sudden pressure on reserves?

One final idea, if we can determine Anchor deposits’ growth rate, the reserves could simply grow accordingly. So say if last week depositors’ interest payable grew by 10% → the reserve needs to expand to match that, and any excess yield can be distributed to ANC stakers.


Would be great if we have an analytics dashboard!

1 Like

Agreed. A dashboard where we could see the realtime data would make it much easier to monitor and decide if adjustments are needed.

30 days might prove to be insufficient, it is just a starting point and adjustments can be made as needed. However I do think it is important to hit the threshold quickly so as to ramp up the staking rewards and buying pressure on ANC as soon as possible.

This is an important point. If there is a huge windfall of ust deposits on ANC we need to be sure that the Yield Reserve can handle it. Keep in mind that if there is a large deposit and the Reserve drops below the 30 day threshold, the inflow distribution will revert back to 90% yield reserve 10% staking rewards until the reserve wallet is replenished back to 30 days.

Monitoring the growth rate I think would be very helpful in determining if the Reserve amount would need to be adjusted, say from 30 to 60 or even 90 days reserve. This is why I think a dashboard would be helpful as someone else suggested.

1 Like

Another thing that must be taken into account is the supply of Luna and the capacity it has to support UST demand. Out of the 390m circulating supply 20m is in bluna 20m18 = 360m which can be borrowed max 50% LTV, 360/2 = 180m UST deposits would be the “max” Anchor could support right now. So it would be highly unlikely that demand could jump 500m in a short amount of time. There would first need to be a large collateral build up of bLuna to support this. Even 100m bLuna at current price/staking reward could only support 100m18 = 1.8b * .5ltv = 900m in deposits. This should be taken into account when we think what is an appropriate amount of time to build the Yield reserve for. If we add other liquid POS assets then this changes this question significantly but I think its safe to say that bLuna being created and thus allowing more UST deposits can be ballparked. We also must consider if the 20% yield is sustainable which would effect our planning for the yield reserve.


What if we put it on a 180 day curve? As the reserve gets closer to 180 days of coverage, the % changes. For example, say we have 0 reserve on day 1, 100% goes to the reserve. day 2, 5% of the 180 coverage is met, so 95% of the excess is directed to the fund.

That way it sorta balances out. 180 days might not be the right figure.


@Freshcakes - good points, but from what I understand Anchor is already engineered to incentivise borrowing when deposits go up to maintain sustainability. With lots of sudden deposits, the borrowing APY would get ridiculously low up until equilibrium is restored. Now, how quickly can that happen - I’m not too sure so I agree with you that reserves need to be built up sufficiently to sustain Anchor deposit interest payments meanwhile.

@SuperPhly - I like your thinking. Depending on the levels of reserves, we could be more or less desperate to re-fill them, so a dynamic rate sounds sensible.

Yeah hard to guess what the right target # days is for Anchor reserves. I suppose the following needs to be taken into account to determine that target:

  • What is the max sudden increase in deposits that we can reasonably expect?
  • How fast can locked bAssets be adequately increased?

I can’t really tell with a good degree of confidence even what orders of magnitude we are dealing with for either of those, does anyone have a better informed POV?

1 Like

A curve is an interesting idea. I’m just thinking in terms of maximizing staking rewards and buying pressure on ANC as quickly as possible while still establishing a safe reserve… I’m not sure this is the best model for that, but I’m open if its what the community wants.

True, but remember a surge in deposits means a surge in demand for UST, which means LUNA price goes higher, in which case all the existing collateral (currently 18 mil Luna) is worth more, so that mechanism helps the issue you are raising. But it is a valid point you make which needs to be considered as we make our decisions.

Along with that if Luna price increases APR will go down (Assuming all else equal).

@Freshcakes @SuperPhly I think a sudden increase in deposits is a very legitimate concern that we need to account for. I also think 180 days reserve is quite high, and building up that much of a reserve organically will take too long and delay our staking rewards and buying pressure benefits… even on a curve. Here is an outside the box suggestion that might solve both problems: what if we established a 30 day reserve organically, as well as transferred 2.5-5 mil ANC tokens from the Community fund to the Reserve wallet, kept in ANC and held in case of an emergency windfall of deposits. In such a situation, the ANC would be sold for UST and cover the reserve until the system rebalances at equilibrium. After all, we are sitting on almost $500 million in community funds… I don’t think this would be a bad use case for some of them. Thoughts?


Makes total sense to me, with current deposits a $500m fund can cover interest fully for 5+ years. For the fund to be worth only one month’s reserve, based on my napkin math we’d need over $31b suddenly deposited in Anchor (200x the current deposits). A very safe margin.

Taking the idea further, do we even need to transfer millions of ANC to sit in the reserve wallet? What if we just bake in a rule enabling the use of the Community fund to cover reserve shortages ICE?

So summarising the draft proposal with the idea above:

  • The target size of reserves would be 30 days’ worth of deposit interest payments based on current deposit levels
  • If the reserves size is less than that, a significant part of excess yield goes to fill the reserve (say 50%?). If it’s 30 days or more, all excess yield is distributed to stakers.
  • In the unlikely event that reserves prove insufficient to handle a sudden spike in deposits, we’d allow the use of the community fund to cover deposit interest payments. When the borrow/deposit equilibrium recovers, the amount borrowed from the community fund gets repaid from the excess yield.

All good thoughts, it feels like we are getting somewhere. I would make some minor tweaks to your summary of the proposal:

  1. Target size of Reserve wallet is 30 days interest payments of current deposits, held in UST. Once target is met, 100% of inflow goes to buyback ANC and paid out as staking rewards
  2. If reserve wallet falls below 30 days, 50% of inflow is diverted from staking rewards to refill reserve wallet back to target size of 30 days payout.
  3. As a safeguard from a windfall of deposits, 2.5 million ANC tokens will be transferred to the reserve wallet from the Community fund (making the total value of the reserve wallet approx 180 days payout reserves). If the UST in the reserve wallet is depleted from a windfall of deposits, the ANC will be swapped for UST and paid out as yield until the fund rebalances.
  4. In the event that the ANC in the reserve wallet runs out, we have a rule in place authorizing whatever is needed to payout the necessary yield to be drawn down from the community fund, this will be paid back when the protocol rebalances. (this would be a code red emergency if everyone in crypto decided to deposit all at once)

Am I missing anything?


This seems like a very good summation and I think the idea of using community funds (especially given how much $ they are worth now) to backup the reserve wallet makes a lot of sense!


Agreed. Would love to get feedback from a few more community members or even a team member before putting up the actual proposal.

1 Like

as a community member Ill just say this: I would definitely vote for this as it’s laid out here, I’m not smart enough to know if the parameters laid out are the right ones, but its a damn good start.

1 Like

All sounds good. Let’s get input from some relevant Terra dev / insider regarding feasibility and making sure we’re not missing anything. Who’s the right person to poke?