The economic incentives of staking in Serenity


#1

For an initial reference, I put together a twitter thread on this subject a few weeks back but I’d like to formalize the discussion here.

I used the latest spec to calculate the sliding scale of staking payouts versus total network at stake and it looks as follows:
DnabodbUcAAAmyV

I’d like to open up the discussion around these numbers. There are many complexities involved that I think should be considered:

  1. What return is the average user looking for considering 32eth of capital?
  2. What returns will “competitors” built on Ethereum be able to offer for returns on ETH investment. This will impact the attractiveness to current ETH holders.
  3. What returns will “competitors” not built on Ethereum be able to offer for returns on capital. This will impact the attractiveness to people who may have entered ETH to stake otherwise.
  4. What will the cost of running a single validator be, cutting into this return?
  5. How many holders will stake no matter what the return, just like running a node at a loss today?
  6. How much do we want staked in the network to feel minimally secure?
  7. Staking rewards are taxable, how will this cut into incentives?

Here are some of my initial thoughts on the above:

  1. Somewhere around 4% return.
  2. This will fluctuate wildly on DeFi apps. Relatively risk free interest will be around 0.25%-1% and riskier loans with counter-party risk will be around 5-10%.
  3. There are 2% savings accounts today but one must also consider inflation. Over a period of say 15-20 years it’s not all that difficulty to average 4-5% returns post-inflation in the US stock market.
  4. Gathered some VERY preliminary data here and seems a single validator will require: 1-5gb storage and 256kbit/s of internet connection.
  5. There are 13,000 nodes today but it’s hard to answer this one.
  6. If I’m reading the spec right then we need 128 validators for a minimum committee size. So 1024*128=137,072 validators, or ~4.2mn total staked ETH.
  7. Would be somewhat similar to other investment vehicles.

What does this mean? Well, I’m not really sure yet and that’s why I wanted to start this discussion. I personally don’t really see a scenario where the network gains many validators under the 3% return rate (~15mn staked). If that is acceptable for the network to run efficiently, the perhaps it’s fine, but if not should we consider increasing these initial numbers with the downside of slightly higher inflation?


#2

I wouldn’t spend too much effort on tweaking parameters right now and rather say let’s just try it out. When the first phase of Serenity will launch, the value to be secured will be quite low, so as long as we don’t start with something completely off we’re probably vastly overpaying for security anyways. Then we have some hard numbers on what validators will do and we can adjust accordingly in the next fork.

Of course I don’t want to discourage any discussion, the more we know now the better. But I have my doubts that those kind of predictions can be very accurate (I’m not an economist though).


#3

I’m not too sure I agree with that. As Ethereum ages, it becomes harder and harder to coordinate hard forks, which would be required to tweak parameters. It’s best to attempt to get it right up front and not rely on the need for constant tweaking. IMO, the need for constant tweaking of the block reward has been less than ideal.

Also, what do you mean by “the value to be secured will be quite low”? There is immense value secured by the network today.

Finally, my research above would say we are potentially underpaying, not overpaying.


#4

As Eth2.0 is supposed to be rolled out in several phases each of which requires a hard fork anyway, this shouldn’t be too much of an issue I hope (note that the forks mostly do not affect the PoW chain, only the beacon chain and the shards). Look here for a preliminary roadmap.

That’s true, but that value will stay on the PoW chain in the beginning, it will not move immediately to the shards (because in phase 0 they don’t exist, and because in phase 1 they only store data, there’s no state execution on which most contracts depend on right now).


#5

@econoar thanks for the post. :slight_smile:

IMHO we should agree on a fixed global ETH inflation (e.g. 1%) and distribute it to the current validator set, no matter how big/small the set is. I have high hopes for such a model when it comes to onboarding validators, because it provides really high rewards for the validators in the beginning (we’re rewarding early entrants for the high risk they’re taking, see bellow), and later on (when Eth 2.0 matures) it reaches some equilibrium. One nice thing here is that we don’t need to care where the equilibrium is, as long as we have the critical number of validators (137,072).

From this it’s obvious that we will need way more validators than we have miners now (quick note: this current number of nodes is not equal to the number of actual miners, but it can certainly serve as a reference).

On the other hand, future validators will face a few disincentives compared to current miners:

a) they risk their stake/capital by applying for this risky new job (PoW is very well known/proven and PoW miners don’t risk losing their capital),
b) the rewards are generally lower than in PoW (AFAIK, payback period in PoW mining is usually around 6 months and after that you usually make some decent profit for at least 6 months after that, then you buy new GPUs and start all over again - my rough estimate is that you can make at least 30-50% per year, but you have to continue investing to stay at that level),
c) they’re making a more serious commitment than in PoW mining (you have to be online 24/7, in PoW you can plug/unplug anytime you want),
d) higher entry-level investment (you can get a small PoW rig for ~$2,500, and PoS validator will have to invest ~$6,000, plus I believe this price is unrealistically low and it will go up in future).

There are more disincentives, but IMHO these would be the main ones (there are also a few advantages, of course).

This being said, I think we should have very high rewards in the beginning to attract as many validators as possible (to reach that threshold), and just let the system reach equilibrium later.

One last note: I’m not a big fan of a super-low/close-to-zero ETH inflation. What’s the point of it, do we really need ETH to moon? I mean, who cares expect the speculators? :stuck_out_tongue_closed_eyes:


#6

Can you elaborate on this, please? To be specific, how will we be paying validators in Phases 0 and 1, is that decided upon? Thanks! :slight_smile:


#7

Thanks for the response!

This is definitely a possibility. A fixed inflation rate takes out the complexity of the sliding scale and could be an easier way to find the proper target.

This I don’t agree with. It’s well beyond just “ETH to the moon”. We’re talking about a platform that hopefully some day secures Trillions in value (already is securing Billions). It’s a necessity that ETH’s value continues to appreciate as it keep the network a lot more secure, incentivizes staking through valuable returns, and makes it more costly to attack the network.

No one is going to trust valuable assets on the network if we’re killing investment value through rampant inflation.


#8

This is true, you’re right, I completely lost that perspective. :blush:


#9

Well, it’s not decided until the community starts running the chain :slight_smile:. I’m not sure how final the numbers in the spec are, I wouldn’t be surprised if they end up being changed according to the outcome of this discussion.

But according to @econoar’s calculations above, for the smallest staked amount of 1M ETH we get a yearly interest rate of 12% or in absolute numbers 120k ETH or $25M.


#10

Thanks for the answer. :slight_smile:

I guess this would definitely be a nice incentive to onboard early, while at the same time global inflation remains almost the same. :ok_hand: This has nothing to do with economics, but I would also be extremely careful with slashing in the beginning (I spoke with a lot of people who are considering becoming validators and they’re all so scared of it :slight_smile:).


#11

Hmmm, who will set the rate? Do we need a “DeCentral Bank”? :wink:
Although it does add complexity, it seems like we might want to be able to set / change inflation rate so that we can manage value of the total ETH asset pool vs other cryptos / fiats. Since OP is comparing investment in ETH to investment in Fiat (“savings accounts”), it does seem like we might want to consider building some form of this lever.