The Voltaire SPO poll: more is needed to increase decentralisation

The Cardano Foundation is about to run a poll on mainnet asking Stake Pool Operators their opinion about changing some of the network’s parameters. Delegators also have a voice in this experiment as they can move their stake depending on the vote of their SPO. The poll will go live this Monday (May 15th 2023). The goal of changing those parameters is to achieve maximum decentralisation, but can this be realistically done by only changing two parameters?
The problem is that big stake pools which are run by a single entity could potentially become too powerful and pose a threat to the security of the Cardano blockchain. We need a way to distribute the available stake to as many pools as possible. The current situation where a few entities hold a very large portion of all stake is not desired.
The poll is proposing changes to the k parameter and the minPoolCost. Currently k is 500 which means the network strives to have 500 fully saturated pools. When k would be increased to 1000 the network aims for 1000 fully saturated pools and therefore the saturation point of all pools will be automatically lowered from 70 million to 35 million ADA. This means that entities running multiple fully saturated pools need to split their pools to accommodate all the ADA they hold. This will increase their operating cost due to the fact they need more equipment to run those extra pools.
The argument for increasing k to promote decentralisation is often that delegators who stake at a big pool will move to a smaller pool. Half of them will surely move to another pool, but if they also move to another operator remains to be seen, especially in the case of exchanges. The staked ADA remains with that single entity and the desired increase of decentralisation did not occur. Not a win.
The next option is the minPoolCost. When the network distributes rewards to the pool’s delegators, the minPoolCost goes to the pool operator first and then the delegators get the rest of the amount. Proposed poll options are either to keep the 340 ADA, or reduce it to 170 ADA.
Running a sustainable operation will get more difficult and sometimes impossible for small operators when their income is goes from 340 to 170 per epoch. Many smaller pools especially in developing countries might give up because of this. The poor will again get poorer and the rich again get richer. Not fair and again not a win for decentralisation.
Small pools often want to be able to set the minPoolCost to 0, because then their delegators receive higher rewards. By promising higher rewards they hope to attract more delegation. Covering their operating cost can be done by setting a variable fee of for example 1%, but in the case of small pools it’s often 0%. Great for delegators, but the operator is not making any money and therefore this kind of operation is not sustainable in the long run.
In my opinion increasing k to 1000 and keeping the minPoolCost at 340 ADA is the ‘best’ of the options. It raises the cost for the big Multi Pool Operators that need to split their pools. Also it keeps a sustainable income for operators of small pools. Therefore my vote will go to “Increase k to 1000 and keep minPoolCost at 340 ada”.
In the meanwhile we should think about a way to better promote decentralisation. We want to avoid having only few entities who control the majority of the staked ADA. This becomes even more relevant as Voltaire is apporaching, because probably a higher amount of ADA staked in a pool will have a bigger weight when voting.
My proposal for improvement would be making the minPoolCost a function of pool size and pledge. The minPoolCost would stay a fixed amount, but when the pool grows, a percentage is added to it. When the pool shrinks the added percentage shrinks as well.
The pledge shows how invested an operator is. Should an operator with a low pledge earn the same as an operator with a high pledge?
Currently the ROA of a pool is determined only by pool size. The proposed function would be an additional way to govern the ROA of a pool. Large pools could be forced to pay a higher minPoolCost and would therefore have less ADA available to distribute to their delegators, making the pool less attractive for delegators and making them move to another operator. Should the pool have a high pledge compared to the stake, the function could ‘soften’ the minPoolCost increase.
This won’t be a solution in all cases and for all MPO’s. Those exchanges who have a 100% variable fee on their pools distribute the rewards to the delegators themselves and thereby avoid the networks’ reward distibution scheme. Usually those kind of pools also have a very low pledge and maybe the network should just distribute less rewards in general to a (near) saturated pool with a very low pledge. If the operator wants to earn more they should at least increase the pool’s pledge. This is however a completely different problem.
Adding a function that takes pool size and pool pledge into account and adds a pecentage to the minPoolCost is in my opinion a good start to promote decentralisation and should be investigated.