Exploring validator compounding
Exploring the higher rewards made available with compounding validators in the Ethereum Pectra hard fork.
The Pectra hard fork brings with it the ability to operate validators with balances greater than the current 32Ξ limit, and for consensus rewards to be retained by the validator and compounded rather than returned automatically to the staker. This article discussed the benefits of compounding validators for different levels of staker, exploring benefits and drawbacks for stakers, node operators, and the Ethereum network as a whole.
Compounding validator rewards
The general principle of compounding is well-understood: it is use of static percentage returns on a given amount of capital to increase that capital, resulting in a continuous increase in absolute returns over time.
For the purposes of this article we will use an overall staking yield of 3.5%, with 2.8% of that coming from consensus rewards and 0.7% of that coming from execution rewards. This broadly equates to what is seen on mainnet at time of writing, although there are complications in this that are explored later.
Without compounding, returns of 3.5% over 10 years would look like this:
Whereas with compounding they would look like this:
The difference can be clearly seen when putting both of the above lines on the same graph:
From this graph it seems obvious that stakers should compound validators as soon as possible to gain these higher returns. However, there are other factors that need to be considered before jumping to this conclusion
Realities of validator rewards
The above graph shows the theoretical returns that could be achieved if the validator's balance was compounded daily to increase rewards, however there are a number of features of Ethereum validating that could result in lower achieved rewards.
Effective balance
Validator rewards are based not on their balance at the time that rewards are earned, but on their effective balance. For full information about the difference, please see our earlier article, however for the purposes of this article the important point is that for a validator performing its duties and gaining rewards as expected its effective balance will increase in 1Ξ steps rather than continually in smaller increments. This increase takes a significant amount of time, as can be seen in the table below.
| Balance (Ξ) | Days to increase effective balance |
|---|---|
| 32 | 407 |
| 128 | 102 |
| 512 | 25 |
| 2047 | 6 |
Even at the highest possible balance for a validator it takes 6 days for it to increase its effective balance. This means that in practice compounding is retarded compared to the theoretical curve.
Split between consensus and execution rewards
Not all validator rewards accrue to the validator's balance. Consensus rewards accrue to the validator, however execution rewards accrue to the validator's withdrawal address, which is an execution address. Crucially, these funds cannot be sent to the validator until they reach at least 1Ξ in total, as this is the minimum amount that the Ethereum validator deposit contract will accept. This further reduces the opportunity to increase the validator's effective balance more frequently.
Randomness of rewards
Both consensus and execution rewards have a degree of randomness. Although a validator's attestations are regular, happening exactly once every epoch1, this only accounts for approximately 85% of the validator's consensus rewards. The other 15% come from being a member of a sync committee and proposing a block, which are operations that happen sporadically. A 32Ξ validator would expect to propose a block roughly once every 5 months, and be a member of a sync committee once every 6 years, but given the nature of randomness any individual validator could be unlucky2 (or lucky) and so reduce (or increase) their total income significantly.
Realistic compounding returns
Given that we now understand the theoretical maximum returns, and the factors that could reduce the returns for a validator from that maximum, what are the realistic returns for a compounding validator, and what can we do to improve them?
To help answer this question we define three strategies that can be used to manage validator rewards:
- unmanaged: consensus rewards remain on the validator and assist compounding; execution rewards remain on the execution address and do not assist compounding
- semi-managed: consensus rewards remain on the validator and assist compounding; execution rewards are deposited to the validator whenever they reach 1Ξ and assist compounding
- managed: whenever the sum of consensus rewards and execution rewards exceeds 1.25Ξ the consensus rewards are withdrawn and redeposited to assist compounding
In addition we pick three amounts of Ether as a starting point for our validators:
- 32Ξ the smallest possible validator
- 512Ξ a mid-size validator
- 1,464Ξ the largest validator that continues to compound all of its rewards over the 10-year period (i.e. not exceeding 2,048Ξ balance)
We consider both withdrawals and deposits to be instantaneous. Although in reality this will not be the case, the delays will be relatively small in most situations3 and so we accept this simplification for the purposes of attempting to obtain realistic returns whilst recognizing that these are still being somewhat optimistic.
The chart for our first validator is shown below:
With an initial stake of 32Ξ each additional Ether added to effective balance makes a noticeable difference, which is unsurprising given that an increase of 1Ξ is equivalent to a rewards increase of more than 3%. After 10 years the impact of compounding is clear, with semi-active and active management of the stake providing noticeable increases in growth.
The chart for our second validator is shown below:
Here we can see that the initial stake of 512Ξ results in compounding that is much closer to perfect growth, with the active returns being indistinguishable from perfect growth on the graph.
The chart for our third validator is shown below:
Here we can see that the initial stake of 1,464Ξ results in growth that looks very similar to that of the 512Ξ stake.
Providing a comparison of the different growth can be instructive. The graph below shows the 10-year returns for different initial amounts of stake and management strategies, which is instructive:
Here it can be seen that there is a noticeable difference between the different styles of compounding when a validator's initial balance is low. This is understandable, because the smaller the initial balance the higher impact that a 1Ξ increase in the effective balance will have on rewards. However, as the initial stake increases the differences, especially between semi-managed and managed, shrink, and beyond an initial stake of 512Ξ the differences start to become indistinguishable.
It is important to remember, however, that the difference between the unmanaged and managed approaches depends on the amount of execution rewards that are available to top up existing validators, and these remain beholden to the double random nature of both frequency of proposal and size of reward. So, although the above graph provides a model of how returns would change with different strategies, it should always be kept in mind that the returns for any single validator can and will look different from both the model and other similar validators.
A different way of looking at this is to consider the APY equivalent for each of these validator sizes and methods. This is shown in the table below:
| Configuration | 10-year returns (%) | APY (%) |
|---|---|---|
| No compounding | 35.00 | 3.05 |
| Perfect compounding | 41.91 | 3.56 |
| 32Ξ unmanaged | 39.38 | 3.38 |
| 32Ξ semi-managed | 40.27 | 3.44 |
| 32Ξ managed | 40.85 | 3.48 |
| 512Ξ unmanaged | 40.23 | 3.44 |
| 512Ξ semi-managed | 41.69 | 3.55 |
| 512Ξ managed | 41.73 | 3.55 |
| 1,464Ξ unmanaged | 40.27 | 3.44 |
| 1,464Ξ semi-managed | 41.76 | 3.55 |
| 1,464Ξ managed | 41.77 | 3.55 |
In terms of APY, it is clear that there is very little difference between semi-managed and managed validators. There is somewhat more difference between unmanaged and managed validators, however even then the differences are not hugely significant in percentage terms. To put this graphically, we rebase figure 8 so that its Y axis starts at 0:
This shows more clearly that although there are benefits to different strategies, they are not so large that stakers should feel obliged to compound and consolidate as soon as the Pectra hard fork launches.
Other factors
There are additional factors that may come into play that could result in reduced capital, decreased compounding, and ultimately lower rewards over time.
Transaction fees
The actions of making a validator a compounding validator, consolidating validators, and withdrawing funds from a compounding validator are all Ethereum transactions, and as such cost Ether to carry out. These fees are small during normal network operation, however in times of high congestion the costs to carry out these actions could be considerably higher than the potential returns. It is important to be aware of the current state of the network, and the resultant cost to the staker, when sending such transactions.
Taxation
Depending on the jurisdiction in which the staker resides, it is possible they will have to pay a percentage of their rewards as tax. Taxation may be for all rewards, or only for execution rewards, and may be able to be deferred, however over a period of years it is likely that some tax will have to be paid. When rewards are retained in a compounding validator, it is more complex to calculate and prove tax as compared to the existing periodic withdrawal system, and transactions must be made to withdraw funds to pay tax bills. This is a situation that is likely to change as tax guidance evolves in separate jurisdictions over time.
Costs of staking
Be they costs of a staking service provider, a hosting provider, or ongoing costs such as electricity and internet connection, there are likely to be some amount of on-going costs that will need to be paid.
Network bandwidth and computation
Each validator, regardless of the amount of Ether it holds, generates one attestation each epoch. This attestation needs to be broadcast around the network to be included in the next block, requiring many nodes to receive the attestation, process it, and forward it as appropriate. In addition, the process of attestation aggregation is somewhat costly in terms of resources.
Consolidating validators reduces the number of attestations, whilst maintaining the total amount of Ether that is attesting. This reduces network bandwidth and computation requirements without reducing the security of Ethereum's proof of stake network, which is a good thing overall.
However, the overall impact of any single entity consolidating their validators will be relatively small, even if they hold a significant stake. For a significant reduction in bandwidth to occur, many entities would need to carry out consolidation, which is not without costs as we have described above. As such, we expect network traffic to drop over time in line with the progress of validator consolidation, but not to expect a large decrease soon after the Pectra hard fork.
Overall chain performance
Attestations need to be included on-chain, ideally immediately after they have been produced, to earn rewards. Attestation inclusion is a part of the block proposal process, and relies on other actors in the network to both aggregate the attestations and to provide them to the block proposer. As such, if the chain is not performing well there is a chance of lower rewards due to late (or no) attestation inclusion.
Conclusion
The ability to compound validators is something that brings with it the opportunity of higher returns over time. For stakers who do not need immediate access to their rewards, enabling compounding is the single most impactful process they can undertake to increase their returns. For those with stakes in the hundreds or thousands of Ether, consolidating their validators so that each one has a higher base amount of Ether will provide higher returns although the increase will diminish over time and the level of consolidation should be balanced against consolidation costs and potential slashing risk of a high-value validator. Topping up validators with execution rewards whenever these rewards reach 1Ξ will help to increase returns, although this is an incremental benefit over enabling compounding and is highly dependent on the frequency and amount of execution rewards obtained.
Active stake management, where advanced strategies are applied to validator balances and rewards, may yield higher returns over time, and these remain an area of investigation.