Double treasury taxation of 40% and possible workarounds

During the months that have preceded the Shelley mainnet a discussion about the parameters that are ruling the blockchain was in progress.
Even if the focus was mainly on a0, k and d parameters, IOG researchers were also analyzing the right mix of tau, the parameter that sets the treasury taxation applied to the epoch reward pot, and rho, the expansion rate of the circulating supply.
At the beginning IOG researchers were evaluating to set tau parameter to 5% (0.05) and let unclaimed rewards to flow to the treasury as well.
But then they realized that unclaimed rewards were expected to be a huge amount of ADA due to the possibility, then confirmed, that the average pool pledges would have been quite low compared to the pool saturation point and, due to the effect of a0 parameter, a consistent amount of rewards would not be distributed.
So, in order to mitigate an excessive transfer of ADA from the reserve to the treasury, they decided to update these rules.
Few weeks before the Shelley mainnet launch they chose to set the treasury tax (tau) to 20% and to change the destination of unclaimed rewards from treasury back to the reserve for the benefit of both delegators and pool operators.
This decision however has a side effect that I would like to highlight, not with the scope to criticize that decision but, firstly, to understand if it’s expected or not and then, if unexpected, how to fix it in a way as fair as possible.
Since epoch 211, when the first rewards of epoch 209 have been distributed, I noticed that the epoch reward pot has been roughly split in rewards, treasury tax and unclaimed rewards through following percentages:

  • Treasury Tax 20%
  • Rewards to delegators and pools 30%
  • Unclaimed Rewards 50%

At the moment in ADA we are roughly talking about:

  • Total Reward Pot 39.1M
  • Treasury Tax 7.8M ada
  • Rewards 11.8M ada
  • Unclaimed Rewards 19.5M ada

As you know the treasury tax applies before the reward calculation and distribution, so the whole epoch reward pot is taxed first.
The consequence is that the 50% that goes back to reserve is affected by a double taxation (20% + 20%).
But, except the first round of reward distribution, it also means that all ADAs that compose the reserve will be affected by double taxation sooner or later, because the 50% of unclaimed rewards that is taxed and then sent back to the reserve will be taxed again, at the moment it will be used to pay delegators and operators as you can see in the simple schema below.
The result is that every 50% of the reward pot is taxed by 20% so treasury taxation is actually 40%.

image

My question is: is the treasury tax rate of 40% expected?
Maybe yes, maybe it’s good for the whole ecosystem, but in case it’s a collateral effect in an attempt to distribute more rewards to delegators and SPOs I foresee two possible options to remedy the situation.
The first one is to decrease the treasury tax parameter tau from 20% to 10% in order to reach the expected and desired 20%.
The second one, more complex, consists on applying the treasury tax only to the distributed rewards instead of to the whole reward pot.

I think that connected to this topic other ones could be approached.
For example, if we decide to increase a0 in order to increase the importance of the pledge we need to face the side effect of a decrease of ROI because the reward cut due to a higher a0 will be more severe.
The savings, as result of a decrease of treasury taxation from 40% to 20%, could be used to sustain the cost of an increase of a0 parameter in order to preserve the current ROI percentages, for example.
That’s all, I hope my post will start a constructive discussion.

I take this opportunity to thank Antonio [CSP] and LordWotton [RIOT] for their support on developing the topic and reviewing the draft.

19 Likes

This is a good topic. I will think more about it.

1 Like

I brought this up after the Mainnet parameters was released. In very short, before it was 20% + some variable unpredictable margin. @Lars_Brunjes explained the fixed 40% instead will be slightly less than it was before.
I agree a blog post would help for a transparent understanding.

4 Likes

This issue is an important story for community development. And it’s a real problem.

1 Like

That is correct, unclaimed rewards are no longer going to treasury - they go back to the reserve.

The detailed answer was: (please note that “we” in the following text stands for IOHK)
We changed the way the calculation is done. We had always intended an effective rate of 20%, but with the previous rules the effective rate was dependent on the level of ada participation in the mainnet (since unclaimed rewards went to treasury) - we were returning 5% fixed plus a variable rate (unclaimed rewards). Our calculations showed that with this scheme, we could be returning 50% or more of the rewards to the treasury rather than the intended 20%. What now happens is that unclaimed rewards are returned to the reserves, so they will be distributed to pools and delegators instead.
The treasury now gets a fixed rate of 20% rather than a smaller fixed rate (5%) plus a variable rate that could be as low as 15%, but could be as high as 50%. So pool owners and delegators benefit, and the treasury gets a consistent (but generally lower) fixed return.

3 Likes

Thank you. We hope to discuss this matter with more and more people. If Cardano takes this issue from the community standpoint, Cardano will gain a lot of trust. We have a healthy community. Everything seems to be fine.

Hi, thanks for the official answer, I really appreciate.
I agree that the change applied before the Shelley mainnet launch is an improvement, through that decision IOHK decreased the overall real taxation from rough 55% to 40% and I think everybody can appreciate the effort and the improvement.
If I well understand your (IOHK) statement confirms my assumptions and confirms that this behavior was expected, am I right?
Should I assume, as operator and member of this community, that an overall treasury tax of roughly 40% is what we need to sustain future development of the network?
Because I need clear information to spread to my delegators when somebody asks me for details

1 Like

You may have a different position. But a average ada holder like me supports the comments below.

"consists on applying the treasury tax only to the distributed rewards instead of to the whole reward pot."

Thank you for your post and your suggestions!

I think the phrase “double taxation” is a bit misleading: If all stake was delegated and all pools were in equilibrium (fully saturated, fully pledged and with perfect performance), then only exactly 20% would go to the treasury. If on the other hand few stake is delegated and we have many small pools with low pledge that miss a lot of blocks, much more than 20% or even 40% of reserves would end up in the treasury eventually.

So the tax rate is not really 40%, under extreme conditions it could be anything between 20% and 100%, but in the long run, we expect it to be closer to 20%.

We were aware of this effect when we set the treasury contribution to 20%.

As for your proposals: I like the second idea of only taking a share of the actually paid rewards for the treasury. This would make the mechanism more transparent, I think. We have already had some internal discussions about your idea, and it received some positive feedback.

Would you be willing to put your proposal in a CIP, @feqifei ?

5 Likes

Hi @Lars_Brunjes, thanks for your reply that completes the whole picture.
I’m honored there has been a discussion about my proposal, I will submit a CIP as soon as possible.

6 Likes

Hi Lars and everybody, finally with the support of @Antonio-CSP and @lordwotton I wrote down and forwarded the CIP that proposes a revision of the treasury contribution process flow.
At the moment it’s a simple pull request waiting for review.
In the meantime an easier-to-read format of the proposal can be accessed here.
Antonio-CSP also prepared a paper that compares current and future impact of treasury contribution on the reserves. He will upload it to this thread soon.
Every suggestion from anybody is welcome including those ones that fix my broken English.

6 Likes

Here we go. I have introduced a measure of the share of unallocated rewards that I generally call
inefficiency. I hope there are no errors in the basic assumptions about the current reward distribution mechanism. treasury.pdf (303.8 KB)

5 Likes

great job @feqifei and @Antonio-CSP, I’m very proud to have been able to give my little contribution to this CIP, that if approved I believe will bring more clarity to the Cardano protocol and treasury. I hope that the community will soon start a discussion about this topic!

4 Likes

I have read through the treasury paper, it’s well written and clear, and congratulation really.

Thanks for all of your (@feqifei, @Antonio-CSP and @lordwotton) effort to make it happen.

Hopefully, in the future, only these kind of high-quality paper based CIPs would only be even considered for acceptation.

5 Likes

A great idea and very well presented!
There is a typo at the beginning of the Specification section: “decribe” should be “describe”.
I fully support this CIP.
Nice work!

4 Likes