Proof of Stake can be cheaper than Proof of Work

There is a famous blog post by Paul Sztorc stating that Nothing is Cheaper than Proof of Work. It is addressed in the Ethereum Proof of Stake FAQ, but while the FAQ dwells extensively on less convincing counter-arguments — it only lightly touches the real break through. The cost of maintaining any cryptocurrency network cannot be less than its block rewards (or any other rewards for the network maintainers) because “marginal cost” = “marginal revenue” which is a basic economic rule for any competitive market. This means that in any POS system resources will be wasted just like in POW. It is quite surprising how this works in a POS system —Stake Grinding is an example of this — it turns the system into a de-facto POW. So you cannot make the system much cheaper by adjusting its internal workings if you don’t change the block reward at the same time. This is the main argument by Sztorc and I agree with it. But

POS can probably spend less on block rewards than POW to reach the same security level

If an attacker wanted to rewrite a part of a POW blockchain he’d need to redo the work done by the miners who created that part in the first place — which according to our MR=MC equality is approximately equal to the block rewards they earned. In the case of a POS blockchain — the rewriting itself would be cheap — but the attacker would have to risk staked coins. I don’t know how to compare these two models — but a model with punishments as well as rewards can have better security characteristic than one that has only rewards.

There is paper published in August Ouroboros: A Provably Secure Proof-of-Stake Blockchain Protocol where the authors claim to create a POS blockchain with security properties for the protocol comparable to those achieved by the bitcoin blockchain protocol.

Bigger market cap makes better currency

In another blog post, that seems to address the Ethereum FAQ linked above, Sztorc admits that smaller rewards would mean less waste, but he argues that if you make smaller rewards in the given coin — then you’ll inevitably move the coin USD price up making the reward constant in USD. Maybe yes, maybe not — but if the price goes up it also makes the currency a better money. With a bigger market cap — you can buy/sell more goods with it, you can store and move more value in it, there is more liquidity in the system. You get more usage for the same input — i.e. the system still gets more efficient.

All of this is a great deal because the costs are enormous. In the end there can be just one digital gold and it wont be the less efficient system, because that one will over time bleed out depleting the value stored in it.