Rich Get Richer

Published by Mario Oettler on

In PoS, block producers and validators receive a reward for their service. Typically, those rewards are created from thin air or paid through transaction fees. Running a validator node is not very expensive. Compared to PoW, both electricity and hardware costs are neglectable. This means that validators don’t need to spend the earned ETH. Instead of spending them on electricity, they would be well advised on staking them again. This increases their share of the total staked coins (if other validators don’t re-stake their block rewards).

The issue is that each newly created coin goes directly to the validators and stays there. This fosters their position in the staking process, making it impossible to get rid of them. Other users have to deal with the remaining coins outside the staking pools. But since also transaction fees flow to the validators (or get burned like in EIP1559), this reduces the number of outstanding coins, making the validators absolutely and relatively richer.

This is a serious concern regarding PoS because it cements the position of validators infinitely. If you are rich at the beginning of the process, you stay rich and get richer with every new block.

Numerical Example

Suppose we have four validators who contribute to the total stake in the following way. Per month there is a total block reward of 100. It is distributed among the validators according to their share on the total staking amount.

Staked coins (staking share)45 (45%)25 (25%)20 (20%)10 (10%)
Coins from block reward45252010
Coins after block reward90504020
Staking share after block reward45%25%20%10%

In this scenario, the relative shares remain constant, although the number of coins controlled by the validators increased.

If we introduce some fixed cost for running the client, the situation changes.

Staked coins (staking share)45252010
Coins from block reward45252010
Coins after block reward85453515
Staking share after block reward47.22%25%19.44%8.33%

We can see that the validator with the highest initial stake increased its influence from 45% to 47.22%, while validators C and D lost influence.

The consequence is that PoS systems could end up with a clique of a few powerful validators or even a monopoly that influences the consensus mechanism.