[GIP-11] Token Compensation for Tech & Finance Multisig Members

Gearbox Multisig Comp

This proposal aims to provide a framework for the compensation of multisig participants, on both the technical and financial multisigs that exist for Gearbox currently. I have attempted here to do some research into various existing DAOs and their compensation structures for multisig participants. By no means was this research exhaustive, so if there are other known cases of multisig compensation that are relevant, feel free to either suggest them to me directly or as a part of the discussion of this proposal.

Firstly, I will note that there isn’t a lot of transparency around multisig compensation in various DAOs. This is not because people are hiding the information, but rather it’s simply because this information often isn’t publicly displayed in obvious places (such as the docs or informational pages about governance). This information is often only found in specific threads in governance forums/snapshot proposals. I just thought this was interesting to note, and I guess I’d suggest that when Gearbox does decide on compensation (if any) for multisig participants, that this information be added to the docs so it’s easy for others to find.

Secondly, while I’ve found a number of examples where multisig participants do get compensated, there also seem to be plenty of cases where they are not - for example, I could find no information about Aave Guardians being compensated (I could be totally wrong about this, I looked in their governance forum and snapshot only). Same thing with Compound - as far as I can find, there is no compensation for their community multisig.

OK, that’s enough preamble.

Overview of DAO Multisig Comp (not exhaustive)

Yearn

Yearn’s multisig participants each receive 1 YFI, linearly vested over 3 years as of Feb 2021. This was determined in the below proposal about Yearn’s compensation/retention packages more broadly.

https://gov.yearn.finance/t/yearn-retention-packages/9698

Untitled

As of Feb 2021 the value of 1 YFI ranged between $30k and $45k - lets take $35k as a rough average. As of today, the value of 1 YFI is roughly $20k.

Based on prices at the time of the proposal, and taking into account the 3 year vesting, each multisig participant is being compensated with about $950 per month ($35k / 36 months). Based on current prices, each multisig participant is being compensated about $550 per month ($20k / 36 months)

Yearn also has some interesting data around the activity of their multisig participants. My understanding is that the two least active members of the multisig relinquished their roles voluntarily due to low participation.

I think based on this data, we should anticipate that there may be situations where low participation/slow signing times necessitates multisig participants to step down from their roles. We should also consider setting some minimum expectations for response times and general availability in order to have clear guidelines for participants.

Source: https://gov.yearn.finance/t/yip-62-change-two-multisig-signers/10758

My personal opinion is that we should perhaps let both multisigs run for one quarter, to see how it goes and collect data before coming up with more formal guidelines for response times and availability, but this point is very much open to discussion.

Tornado Cash

Another project that has a multisig where there is compensation is Tornado Cash.

https://snapshot.org/#/torn-community.eth/proposal/0x2b24f61a9429a9e0ebd5e9c716c8b92e9f9edbce4cfde384072aeaf380a43cde (100 tornado per month)

Their multisig participants are compensated with 100 tornado per month. At proposal time this translated to about $4k a month; at current prices, it’s closer to $3k.

Note that outside of signing txs, Tornado Cash multisig participants are also expected to be available to brainstorm and exchange ideas with members.

mStable

mStable multisig participans get a signer stipend of the equivalent of 1000 USD in MTA per month. This is based on USD values, so the number of tokens paid as part of this stipend vary month to month.

Source:

https://forum.mstable.org/t/pdp-37-community-signer-election/696

https://vote.mstable.org/#/proposal/0xc21a6ac9813a3262bc8800341beab3c0731a887657546038c0568b77d82fdbe7

Fei

The Fei/Tribe ecosystem previously had a monthly 1000 TRIBE grant for multisig participants. At the time (Aug 2021), this was worth $650-$800, currently $580. This was then replaced with gas refunds, which was then subsequently replaced by a grant of 21000 TRIBE per month, which works out to 3000 TRIBE per multisig member, which at current TRIBE prices is about $1700 per month.

Sources:

https://tribe.fei.money/t/fip-21-optimistic-approval/3429 (1000 Tribe per month)

https://tribe.fei.money/t/fip-29-oa-multisig-updates/3514 (gas refund)

https://snapshot.org/#/fei.eth/proposal/0xc5e488e497107f96fcd65f1f73f56108c2f2fe992c9ee113eb83bbd76bd9ee7c (21k tribe / 7 members)

Synthetix Spartan Council

The Synthetix Spartan Council gets compensated with 1000 SNX per month. I left this for the end because the associated responsibilities here seem to go beyond just a multisig. These council members are elected by token holders, and the councilmembers have control of certain protocol parameters. However, it seemed (to me) that the Spartan Council bore enough similarities to a multisig to be included here as a data point, as long as we account for the fact that this role comes with additional responsibilities, which in turn means the compensation is higher as well.

1000 SNX is equivalent to about $4000 per month at current prices.

Source: https://blog.synthetix.io/the-spartan-council-election/

Gearbox Multisig Comp Proposal

Here’s the summary of the above multisig compensation findings

The upper bound seems to be bout $4k, and the lower bound is about $500. This range is more or less what I would have guessed prior to doing the research, so I’d say we should stay within this range.

A couple of additional thoughts

  • Multisig holders should be compensated in tokens, not in USD value of tokens. Just like every other role in the DAO, multisig holders should have access to upside. Governance can revisit the amount of tokens paid if we have a situation where the value of the token changes significantly in either direction

  • Technical multisig participants have (in my opinion) a significantly harder job than financial multisig participants. They will be used to deploy new contracts, change params of the contracts, update allowedlists, and in emergency scenarios to pause operations. In other words, they need to be able to read code and understand what proposed TXs will do. Financial multisig just needs to be available and responsive to send tokens, which is significantly less difficult. For this reason, I believe that the technical multisig members should be compensated 1.5x to 2x more than financial multisig members.

  • Gearbox tokens aren’t transferable and have no price right now, so we will need to make some educated guesses about the possible range of token prices. I propose that high bound of GEAR token price estimate * token compensation should definitely not exceed $4k per month, since this is highest comp that I found in my research.

My rough guesses for the range of possible token pricing for gearbox is as follows:

High end estimates:

I’m using FDV because its easiest, even though yes I know FDV is a meme. But it’s bear market so maybe it’s not a meme anymore + it’s the easiest way to do back of the envelope calculations like this.

Aave FDV = 2.1b

Maker FDV = 2.0b

Compound FDV = 1.1b

Synthetix FDV = 1b

Yearn FDV = 700m

To me, 1b is the upper limit of a plausible valuation for Gearbox (for the near future), given that all the above Defi protocols have been around for a while, have significant numbers of users and proven track records, and should fairly be valued more highly than Gearbox given our newness.

That would put the highest end estimate of token price at $0.1 per GEAR (I think this would be dumb, Gearbox should not at current stage be ‘worth more than Yearn example, but I guess its on the edge of plausible, the market is often dumb).

Low end estimates:

I chose some smaller, newer defi projects using the heuristic of “I think people have heard of these” - this is not at all an unbiased sample, but here goes.

Alpha FDV = 312m

Alchemix FDV = 306m

Keep3r FDV = 215m

Reflexer FDV = 120m

BProtocol FDV = 50m

I would put lowest end FDV estimate at 50m, which implies token price of $0.005.

Token Compensation Amount

If the high end token price estimate becomes reality, technical multisig holders should be compensated with no more than $4k per month.

That means at most, technical multisig holders should receive 40k GEAR per month.

Technical multisig members should receive between 1.5x and 2x more than financial multisig members. Let’s use 2x.

40k/2= 20k.

With these numbers, the compensation looks as follows:

High end

Token price $0.1

Technical Multisig: 0.1 * 40000 = $4k

Financial Multisig: 0.1 * 20000 = $2k

Low end

Token price $0.005

Technical Multisig: 0.005 * 40000 = $200

Financial Multisig: 0.005 * 20000 = $100

I think these are acceptable outcomes on both sides.

Given the fact that some multisigs are not compensated at all, the low end estimates seem OK to me, especially if gas is refunded (particularly on the technical multisig where gas costs are likely to be higher).

On the high end, $4k per month seems (on the edge of) reasonable as well for technical multisigs who need to inspect TXs carefully, particularly in this earlier stage when there might be more deployments and changes to the protocol.

Vesting, Payment Periods, Options

As a relatively early stage project, there should be some form of vesting for almost all GEAR payments to contributors, and multisig participation is no exception.

I would suggest a linear vesting period of 6 months to 1 year. 1 year seems easy, so I would just go with that unless there are strong objections.

Ivan suggested in discord that these payments be made quarterly - this seems sensible.

Compensation should be made each quarter for the previous quarter’s efforts, and each quarter there should be an open discussion in the DAO about the performance of the two multisigs, and whether any existing members should step down.

We should also revisit this topic in 3 months to see if it makes sense to set explicit guidelines for participation rates and response times, so we all have clear expectations.

There was also a suggestion of making multisig compensation options on GEAR tokens rather than tokens themselves - I think this would probably overcomplicate the matter, so I propose that we do not do this, at least not for the initial iteration of this compensation plan.

TLDR Summary

  • DAO multisig comp ranges from ~$500 to $4,000, some don’t compensate at all.

  • Technical multisig is more difficult, should be compensated 2x compared to Finance multisig

  • Upper limit of sensible price estimates for GEAR should not compensate multisig members more than upper limit seen in other DAOs, which is $4,000

  • Token compensation should be linearly vested for 1 year.

  • This proposal suggests a comp of 40k GEAR per month for Tech multisig members, 20k GEAR per month for Finance multisig members. Payments will be made quarterly for work from prior quarter, and all tokens will vest linearly over 1 year.

9 Likes

Incredible work!
good to know how other DAOs work.

1 Like

Gm fren. Thank you for putting this together.

Given we have no token price, and i agree it should be based on tokens rather than usd value of tokens, we have three options:

Continue as written but review if gearbox becomes tradable.

Make the token allocations based on the relative work vs those who have already received tokens rather than a hypothetical price.

Leave the multisig unpaid for now to either be retroactively rewarded later or honoured through simply contributing.

Many thanks again for your work on this.

3 Likes

Great work! very interesting to see the data from other DAOs.
I think an interesting way to look into the data is to calculate the percentage of the compensation in comparison to the total supply.

YFI
7 Multisig, 1YFI each, one payment.
per person (till death?): 1/36,666=0.0027%. in total: 7/36,666=0.019%

Tornado Cash
5 Multisig, 100 TORN each, monthly payment.
per person per year: 100x12/10,000,000=0.012%. in total: 0.06% per year.

Fei
7 Multisig, 3000 TRIBE each, monthly payment.
per person per year: 3000x12/1,000,000,000=0.0036%. in total: 0.025% per year.

Synthetix Spartan Council
7 Multisig, 1000 SNX each, monthly payment.
per person per year: 1000x12/244,199,074=0.0049%. in total: 0.034% per year.

Gear Pre GIP-7
7 Financial Multisig 40,000 each, 9 Technical Multisig 20,000 each, monthly payment.
per Financial Multisig per year: 40,000x12/10,000,000,000=0.0048%. in total: 0.034% per year.
per Technical Multisig per year: 20,000x12/10,000,000,000=0.0024%. in total: 0.022% per year.

I am not sure how much work Multisig needs to do, but I’d say this is a good compensation plan. if they work 10 years for us, they get 0.56% of the total supply.

2 Likes

seems reasonably in line with others, considering there are just more people involves and the technical multisig will likely be more laborious than most multisigs from other projects?

Sorry for the late response here! This is a great start, especially since a lot of information is not publicly disclosed. Kudos!

Regarding Yearn’s information, I believe the program you referenced is a retention program, not an annual program, so I don’t believe it is relevant here.

I would push back. Yes, practice based on the examples you shared is to award in number of tokens, but this is generally changing, especially as folks have experienced a few bear markets (everyone hates being undercompensated). Communities also feel cheated when they approve values that are drastically different following a bull run. I do think the protocol needs to provide sufficient upside potential, I just think that should be a separate decision and not inter-mixed with “fixed compensation” approved today.

Agreed - I like the differentiation here, although I don’t like the arbitrary “1.5-2x”. How do we know that is aligned with market?

I appreciate your research here, although I question whether we should use three examples as benchmarks, especially if those three examples did not have a rigorous process. This is an opportunity to set a strong precedent.

Some additional questions:

  • How do we make sure this is an appropriate market rate? C3 can provide benchmark market data for each multi-sig role based on similar roles from traditional technology start-ups or, if more appropriate based on expected workload, compensation for board members of size-appropriate companies. We can also conduct research on other DAOs to ensure we find all possible data points. This will provide guidance on competitive market rates rather than using three examples as the sole reference.

  • Is there appetite to receive stable comp? If there is, you can consider implementing a fixed mix (e.g., 40% cash/60% native). You can also create an elective program which allows receipt of cash in native tokens, perhaps with tokens at a discount if locked for extended periods of time (Balancer recently approved a similar design and it was based off of the Yearn model). This is a method to provide additional upside and ongoing retention, without involving the community regularly. However, that introduces complexity and the proposed method (100% native with 1-year vest) is simple and aligns incentives with the tokenholder, so I am supportive.

  • How long are members expected to serve for? If we want to promote regularity and annual elections, then 1-year vest makes sense. If we expect members to serve for a longer period of time, I question whether a 1-year vest is enough. Perhaps there is an opportunity to provide an initial, one-time grant with a longer vesting period to ensure stability during the first few years (this would also provide upside potential) (e.g., 700,000 tokens with a 3- or 4- year vest).

  • From a governance standpoint, how long does this agreement last and when should we re-visit it ?

C3 is happy to assist in any way we can.

Best,
Cryptowsky/Founder C3

1 Like

Data from other protocols gives a really great way to measure how much multisig signers should be compensated. Great work there!

Ofc, I support this GIP.

I would push back. Yes, practice based on the examples you shared is to award in number of tokens, but this is generally changing, especially as folks have experienced a few bear markets (everyone hates being undercompensated). Communities also feel cheated when they approve values that are drastically different following a bull run. I do think the protocol needs to provide sufficient upside potential, I just think that should be a separate decision and not inter-mixed with “fixed compensation” approved today.

Honestly, I think token rewards here is appropriate because

  1. The DAO is token rich and (relatively) cash poor
  2. The amounts here are pretty small anyways and can’t really be thought of as a “salary” - being a multisig participant is not a full time job or anywhere near it and shouldn’t be thought of as such.
  3. As mentioned, participants probably prefer the upside

If you have examples of multisig members that are paid in USD or equivalent values of USD, feel free to cite them. I honestly don’t have very strong feelings either way.

Agreed - I like the differentiation here, although I don’t like the arbitrary “1.5-2x”. How do we know that is aligned with market?

Yes, its pretty arbitrary - just based on how I feel about the likely work required from the two multisigs. If you have relevant data or a better process to use that can come up with a less arbitrary number, feel free to share.

I appreciate your research here, although I question whether we should use three examples as benchmarks, especially if those three examples did not have a rigorous process. This is an opportunity to set a strong precedent.

Some additional questions:

  • How do we make sure this is an appropriate market rate? C3 can provide benchmark market data for each multi-sig role based on similar roles from traditional technology start-ups or, if more appropriate based on expected workload, compensation for board members of size-appropriate companies. We can also conduct research on other DAOs to ensure we find all possible data points. This will provide guidance on competitive market rates rather than using three examples as the sole reference.
  • Is there appetite to receive stable comp? If there is, you can consider implementing a fixed mix (e.g., 40% cash/60% native). You can also create an elective program which allows receipt of cash in native tokens, perhaps with tokens at a discount if locked for extended periods of time (Balancer recently approved a similar design and it was based off of the Yearn model). This is a method to provide additional upside and ongoing retention, without involving the community regularly. However, that introduces complexity and the proposed method (100% native with 1-year vest) is simple and aligns incentives with the tokenholder, so I am supportive.
  • How long are members expected to serve for? If we want to promote regularity and annual elections, then 1-year vest makes sense. If we expect members to serve for a longer period of time, I question whether a 1-year vest is enough. Perhaps there is an opportunity to provide an initial, one-time grant with a longer vesting period to ensure stability during the first few years (this would also provide upside potential) (e.g., 700,000 tokens with a 3- or 4- year vest).
  • From a governance standpoint, how long does this agreement last and when should we re-visit it ?

Re: all of the above

These are all interesting questions, but I feel they are much more appropriate for the question of how to pay the most active DAO contributors (e.g devs, marketing, etc). I think the multisig comp amounts mentioned are small enough that all of this suggested deep analysis presents a very real risk of bikeshedding.

Of course, if you already have a data on hand (I think perhaps you might since you are compensation consultants), feel free to share it if you want.

The way I would approach this is to pass this proposal (or something with higher/lower numbers based on community consensus), see how it goes, see what the dropout rate/activity rate is, then revisit in 6 months/a year if there is a significant problem with how the multisigs are functioning.

Again, if you have data on hand that would suggest another course of action is better, or if you would like to collect such data and present to the DAO, I’d definitely encourage you to do so.

I guess my opinion on this is broadly that the sums involved are small enough that it makes sense to just try something straightforward and tweak it as needed going forward.

2 Likes

I agree that the awards should be paid in tokens, especially if that is the recipient’s preference. To clarify, my comment is to ensure a constant value at each grant ($4K; value-denominated), rather than a constant # of tokens at each grant (e.g., 40,000 tokens every month; see Ribbon example below). The latter will result in values that will differ from amounts the community initially approved.

We are currently working on creating a robust database of compensation-related items sourced via public forums (and through our client work). As you mentioned, there is not that much information out there. One example we can share off the shelf is Ribbon. They proposed a 3,000 RBN grant per month. This gets at my point above - I caution pre-approving a token-denominated grant because the community is forced to approve compensation today that may be aligned with market, but misaligned in future months. Yes, you can re-evaluate it later, but at that point the train has left the station. You can see that the final Snapshot approval had mixed results.

It will be difficult to find examples that mirror the scope of the multisig at Gearbox (you called this out already with the Synthetix Spartan Council). Another example is the Delegate compensation at Maker. There model is a bit more complex and even includes a performance modifier that was introduced in later months.

I will share any additional relevant examples as we flesh out our database. We have plenty of information on traditional board member compensation. Of course, the scope and time commitments of the multisig members will differ, but it can be used as another data point if appropriate assumptions are made.

This all may seem trivial / bikeshedding, but it’s important to start establishing best practices when we can. This example may be re-used to create multi-sig compensation for other DAOs (just like we did here).

TLDR; If the proposed amount is value-denominated, then I am supportive of your approach of “testing out the waters” :slight_smile: . I do not think 4K is excessive by any means (there is a possibility that it is too light esp for technical multisig, but cannot say for sure without benchmarking and knowing the scope/time commitment of the members). Also supportive of quarterly payments and proposed 1-year vesting.

I replied with a couple of examples and some insights, but the automated spam filter, Akismet, hid my post (not sure why). Someone at Gearbox should be able to unhide, but let us know if you’d like us to re-post.

I think it’s good now? Approved it back. Ye not sure why it flagged.

For pre-launch tokens, value-denomination is not feasible since the price per token is not yet known. Therefore your approach of estimating the # of tokens based on an estimated valuation makes sense and we are overall supportive. There should be a commitment to revisit this after launch.

Here is another example from Balancer. 500 BAL every six months with pro-ration for partial service. The value today is about $8,000 in total ($1,333 per month). The value at ATH is $37,500 ($6,250 per month). The value at ATL is $3,325 ($554). Token-denomination does not make a huge difference unless the price is at the tails. We expect the prevalence of token-denomination to decrease over time (as was the case for traditional companies). It is likely that token-denomination continues in practice because of carry-over of pre-launch practices when estimating value was difficult (as is your case).

We hope this helps!

Good fat article! Thanks for explaining and good to know how other DAOs work.

Gearbox connecting people.
To the moon!
LFG!

Don’t have strong feelings on this tbh, happy to revisit after launch. I guess my sense was “if it becomes a problem because it becomes too much or too little, we can just adjust the number of tokens” Doing it by USD value is fine too.

Actually, I guess the one “bad” situation if we decide to denominte in USD is if the token gets very low, potentially you’re giving out A LOT of tokens to match the USD value. Not sure how to counteract that.

Overall though I feel like either way is fine, let’s see when this actually comes up for a vote (and whether or not there is a price at that point) and we can decide then which way to go based on rough consensus.

2 Likes

Could someone please pull data on the frequency of tech multisig signers’ activity? Aka who signed how many (showing the response time as a result of that). Can suggest numbers then. @ilgiz

Edit: there is more info on Discord risk channel

  1. Financial multisig for last 6 months had 0 activity.
  2. Technical multisig had ~60txs for last 6 months. Here is more info on signer’s activity. So it’s like ~50% less active than YFI governance.

Taking into account p.2, bera market and initial information about other DAOs from @amplice I think numbers should be a little bit lowered. And agree, tech multisig should have multiplier x2.

From current round 1GEAR = 0.015$.
My suggestion is to define budget for both multisigs. Let’s

  • Techincal Multisig (8 persons): 700k GEAR/Quarter
  • Financial Multisig (6 persons): 300k GEAR/Quarter

Distribution of GEARs between signers according their voting activity (example is here).

For tech multisig we can do it retrospectively for Q1-2 22, for financial - we can start from Q3 22 (after DAO raise Financial Multisig will sign a lot…)

Just FYI for everyone, there was also some discussion in the discord about this topic which ilgiz also linked to, if you want to read it you can start at this discord message.

After some brief discussion, I think there is rough consensus to do as ilgiz suggests, which is reasonably close to the original proposal anyways, with the comp brought down just a little bit as adjustment for bear market.

The financial multisig has not had any transactions yet, so there is no need for compensation. The Technical multisig has been active for about 6 months and has done about 60 transactions.

Given that market conditions are very different from when this was initially proposed, I agree with ilgiz that we should lower the compensation a little. He suggests:

I second the above - 700k for tech multisig and 300k for financial multisig per quarter. For the past 6 months, since we have the retroactive data for activity per signer, I also agree that we can just use the suggestion shown in the screenshot below, it seems fairer than any other method.

Just for context, in the original proposal, I suggested 40k GEAR per month for tech multisig participants . With the new numbers, the most active signers are being awarded ~240k, which is basically 6 months x 40k. Basically, the new numbers and the old numbers are very similar if you are actively participating as a signer.

If there are no significant objections, I will make a snapshot vote for all of the above in 2 days. Speak now or forever hold your peace.

TLDR:

  • Going forward, 700k GEAR per quarter for tech multisig and 300k GEAR per quarter for financial multisig
  • Retroactively compensate the technical multisig 1.4m GEAR tokens for 6 months of work so far
  • This 1.4m will be distributed proportionally according to each signer’s multisig activity as shown in image below
  • I will put this up for a snapshot vote in 2 days as long as there is no significant objection

1 Like

As there has been no further discussion here, I will make the snapshot tomorrow (a few days late I know).

1 Like

Seems like it passed, someone from the financial multisig can initiate those now or later. Snapshot

Sharing data with addresses:

signs premia address
Signer 1 49 257,000 0x3C17d03678c9c37947Af86eBD7B0Dfff896cc075
Signer 2 45 236,000 0xb195a8CF4898E6DDc92e435cb118E9842456A45A
Signer 3 41 215,000 0xA0BaCE512E2904ccFDD24AB56feea6621ce14613
Signer 4 36 189,000 0xf5D3dbda5F41A0E26D71B948e29522398e71cFaE
Signer 5 34 178,000 0x0Cec743b8CE4Ef8802cAc0e5df18a180ed8402A7
Signer 6 33 173,000 0x198c4D019BefE68415351E2C0F9b05585cfAAa5b
Signer 7 27 142,000 0x869ad60cDeF31Ae01959288Ff77D51bcB9F09E41
Signer 8 1 5,000 0x1D35DFE2c3B9A0D9f200Ee70a62D73da832606CD

anyone can check the correctness by parsing data for tech multisig transactions.

1 Like