bisq-network / proposals

@bisq-network improvement proposals
https://bisq.wiki/Proposals
44 stars 16 forks source link

Decide on compensation per BSQ for locked up for bonded roles #120

Closed chimp1984 closed 5 years ago

chimp1984 commented 5 years ago

Bonds locked up for bonded roles should be compensated by a kind of interest rate a compensation for each locked up BSQ per cycle to reflect the volatility risk for locking up a bond for long time as well as the risk for confiscation. We should find a consensus what we consider a fair % rate. The rate would be paid until a role gets revoked and would be simply added to the role owners comensation request (no automated payment).

I would suggest 0.5-1 % per cycle (roughly month), but no strong opinions. Please post below your suggestions so that we find a common ground.

UPDATE: Changed the misleading term interest rate to compensation for the amount fo BSQ locked up per time interval. Also added emphasis that this is part of a normal compensation request and no automated process where BSQ earn a interest rate just by locking them up.

This proposal is only for bonded roles not for reputation bonds or future usage of bonds for the off-chain trade protocol.

m52go commented 5 years ago

TLDR; I think a 1% rate is reasonable. I write the following to attempt a reasoning.

This is such an odd scenario that I find it hard to determine an answer based on anything from the real world. Perpetual bonds are probably the closest thing but still very different in nature and purpose.

There is effectively no term for this "loan". The supplier of capital is on the hook for their own performance, unlike a typical loan/bond where the supplier of capital (bank, investor, etc) is trusting the other party to perform (consumer, company, etc). In these cases the rate can be derived from various attributes of that party and the length of the term. None of these conventions are applicable here.

There is currently 130 000 BSQ locked up, and it looks like about 400 000 remains to be locked up (not including arbitrators). So say total bonds for the network end up between 500 000 and 1 000 000 in the long term.

At 1% per month, the network would be paying 10 000 BSQ in interest per month, the approximate cost of one contributor. In return, the network can maintain reasonable confidence that all of its most valued functions and assets are being handled responsibly, as well as the capability for recourse in the event something goes wrong. Remember this assumes a 1 000 000 total, so for now this number will be much lower, probably in the 1 000 to 5 000 range.

This seems reasonable to me. If anything, this is a low rate for the value it provides. But at 12% annualized return for bondholders, I would probably not be in favor of anything higher.

bodymindarts commented 5 years ago

I don't see why there has to be an interest. Locking up allows you to perform a role which you get paid for. This could be viewed as a form of interest already. This is where the value is being added.

Sent from my iPhone

On 12. Sep 2019, at 04:06, Steve Jain notifications@github.com wrote:

TLDR; I think a 1% rate is reasonable. I write the following to attempt a reasoning.

This is such an odd scenario that I find it hard to determine an answer based on anything from the real world. Perpetual bonds are probably the closest thing but still very different in nature and purpose.

There is effectively no term for this "loan". The supplier of capital is on the hook for their own performance, unlike a typical loan/bond where the supplier of capital (bank, investor, etc) is trusting the other party to perform (consumer, company, etc). In these cases the rate can be derived from various attributes of that party and the length of the term. None of these conventions are applicable here.

There is currently 130 000 BSQ locked up, and it looks like about 400 000 remains to be locked up (not including arbitrators). So say total bonds for the network end up between 500 000 and 1 000 000 in the long term.

At 1% per month, the network would be paying 10 000 BSQ in interest per month, the approximate cost of one contributor. In return, the network can maintain reasonable confidence that all of its most valued functions and assets are being handled responsibly, as well as the capability for recourse in the event something goes wrong. Remember this assumes a 1 000 000 total, so for now this number will be much lower, probably in the 1 000 to 5 000 range.

This seems reasonable to me. If anything, this is a low rate for the value it provides. But at 12% annualized return for bondholders, I would probably not be in favor of anything higher.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

clearwater-trust commented 5 years ago

Surety bonds are NOT interest bearing. Creating interest bearing financial instruments should be left to wall street barons and corporate banks. You're thinking of the wrong type of BOND.

meapistol commented 5 years ago

In the BTC community there is a widespread notion that fractional reserve banking cannot occur. This is of course wrong, there will be BTC banks lending out BTC and giving an IOU in return, precisely as in the fiat economy. In a future bitcoin based economy people will be able to lend out BTC to banks and get interest and this is an argument why people will not be interested in locking up bonds without interest. Objects such as BSQ will also be treated similarly to BTC and can be lent out to banks which in turn lend them out to interest-paying customers, subject to standard credit valuation. The future economy will be much more secure, auditable and private, but most of the basic notions of economy will not change, such as interest on locked up assets.

bodymindarts commented 5 years ago

Why should ppl not want to lock up BSQ if its a buy in for taking on a role? As long as the role has a higher expected return than lending to a Bank it shouldn't be a problem.

If we give interest on the bond we are enforcing a 'rich-get-richer' type of economy instead of a meritocracy. Only ppl with money get to take on roles and they get compensated twice over, for the role and the interest on the bond. A meritocracy needs a low barrier for entry and no rent-seeking.

meapistol commented 5 years ago

Agreed, it will be a balance. The compensation requests will reward the effort and the interest together and there is probably no reason to separate out interest. I expect that in a later stage a market will simply develop for these roles. Bisq must of course avoid any type of rent-seeking.

m52go commented 5 years ago

To clarify, I'm 100% dead-against any form of rent-seeking. But I do see work done for a role as separate from the bond for the role.

Consider an extreme case: domain name owner. Bond is 50 000, but there is virtually zero work. Should that role owner forgo their access to capital for nothing in return?

If interest is a slippery slope, maybe it's better to establish better compensation conventions in place of interest for roles that need it.

sqrrm commented 5 years ago

So far I have done most my roles as a free service to Bisq, only charging for costs of infrastructure. That won't work long term considering the bonds required. Since there is a cost of capital it only makes sense to split the work and cost of capital for each role. That's what the market likely would do in the long run anyway and hiding it doesn't fit the spirit of Bisq in my opinion.

chimp1984 commented 5 years ago

@bodymindarts

Locking up allows you to perform a role which you get paid for.

Not all roles are compensated if there is no monthly effort. Domain name owner is such a role. But I agree we could hide the interest rate in the compensation requests, but even then people should have a similar guidance. I think it might be more transparent that way, like contributors should add a BSQ amount to each bigger work package thye have done.

chimp1984 commented 5 years ago

@bodymindarts

If we give interest on the bond we are enforcing a 'rich-get-richer' type of economy instead of a meritocracy.

Bonds are for protecting roles which cannot be meaningfully decentralized from abuse. Roger Ver abusing the Bitcoin brand by running his bitcoin.com page and tricking people into BCH is such a case where it would have been good if he would had set up a 1M USD bond and the BTC community have confiscated him. Bonds are not intended to earn money from interest rate. It just should compensate for locking up funds long term. BSQ can be expected to be more volatile in future as it is now. Not being able to trade in volatile time has a financial cost. This should be reflected by that interest rate. Maybe we should call it different if that creates wrong mental models. Bond lockup compensation or whatever...

clearwater-trust commented 5 years ago

You should not be compensated for locking up a bond. You are obligated to do the right thing because your bond is locked up. Not because you're getting paid. Surety bonds do not pay interest. Perhaps the bond amounts need to be adjusted if the BSQ price is not "volatile" enough.

mpolavieja commented 5 years ago

I think that locking up capital in BSQ is risky. Putting your hard earned savings in such a new instrument is a very generous decission.

In my view, interest should defenitely accrue for non compensated roles. 12% annually sounds a bit high to me. I would say 0,5% per cycle. In any case, the only way I can think of to find a fair interest rate would be through an open auction, which I don´t think is practical at this point.

mpolavieja commented 5 years ago

rent-seeking

Rents have a very important (and deeply missunderstood) social coordination purpose. They are the reward for those who wisely allocate their capital on things that other people find useful, like for example BSQ ;)

wall street barons and corporate banks

Those who have indeed supressed interest rates? I think we should consider what is the purpose of the capital allocation. IMO it is not the same to finance wars (ie soverign Bonds) than to finance a Bisq role, which I assume we all here find it useful.

Surety bonds are NOT interest bearing

A BSQ Bond is equivalent to pledged collateral on a clearing house. That collateral usually accrues interest.

wiz commented 5 years ago

NACK for the following reasons:

1) Is this already implemented in the DAO ? If not, I don't think it's worth the risk of hard-forking the DAO to add such a feature. The bonders would have to color new BSQ each cycle and it seems to add a lot of complexity for a small benefit.

2) The idea of inflating BSQ to pay interest to bonded roles seems to also be controversial, if anything I think it makes more sense to simply keep the status-quo of not paying interest on the bond and instead adjust the amount of compensation for the role using compensation requests. This seems more transparent, more straight-forward, and allows more flexibility per-role.

m52go commented 5 years ago

and instead adjust the amount of compensation for the role using compensation requests

I assumed this is how it would be accomplished anyway.

Surety bonds do not pay interest.

As I understand it, a surety bond's premium is the interest paid to the guarantor. A 10 000 USD surety bond at 1% costs the principal 100 USD to obtain...that price is the interest paid to the guarantor for the power to pay the damaged party 10 000 USD if necessary.

The difference in this case with Bisq is that the guarantor and the principal are the same people. @clearwater-trust let me know if I'm misunderstanding this.

mpolavieja commented 5 years ago

instead adjust the amount of compensation for the role using compensation requests

I think we ware talking about the concept of paying interest, how we call it or the procedure to do it is of minor relevance. I fully agree that if it is decided to be done, the way you suggest to do it would be the best option.

clearwater-trust commented 5 years ago

@m52go I wouldn't call it interest paid. Why are you calling it interest paid? People that have funds locked up in bonds should not get paid for having their funds locked up in a bond. Surety Bond semantics: i suppose the obligee is the DAO, the surety is Bisq network and the principal is (you) or whoever wants to lock up a bond. Even with all of those definitions out of the way I can't figure out where "interest paid" comes into the picture? The principal does NOT get paid by the surety. In fact, it is the other way around.

m52go commented 5 years ago

@clearwater-trust I did some more research and see what you're saying. I didn't understand correctly. A surety doesn't actually provide capital—the principal does. And a principal would pay the surety for that arrangement in order to carry out whatever activity they need the surety bond for. This payment is considered as more of a fee than interest. OK.

If that's the case, then the principal retains control of the capital in question. So they can deploy it to make profits, or...invest it and make returns?

With Bisq role bonds, the capital is being locked up and made unavailable.

Surely there's a difference?

I hope I'm not coming across as difficult. All I'm saying is capital has a cost. I can't think of a single scenario in which access to capital is made available for free.


If it wasn't technically complex, I would suggest doing a dividend instead. Could be adjusted based on the network's success, would be available to all BSQ holders (not just role owners), but would still compensate role owners for locking up big amounts of capital (while also motivating them to do better at the roles they hold).

wiz commented 5 years ago

@chimp1984 maybe it would be easier if you withdraw this proposal and instead propose to increase the compensation rate for certain bonded roles? it would keep things much simpler, and there would be no discussion of "interest paid" which seems to be the major issue.

mpolavieja commented 5 years ago

I would suggest doing a dividend instead

That's very interesting as it rewards only in the case of success, opposed to interest that accrues regardless. It could be a ¿5%? of the result "BSQ burned - BSQ issued" (only if the result is positive). As suggested above, this could also be done manually and through compensation requests

mpolavieja commented 5 years ago

instead propose to increase the compensation rate for certain bonded roles?

@wiz, I agree with that, but that´s what @chimp1984 was already suggesting:

The rate would be paid until a role gets revoked and would be simply added to the role owners comensation request (no automated payment).

As you said, it seems that the problem is confined to the word "interest" (even considering that chimp1984 said "a kind of interest rate per cycle")

clearwater-trust commented 5 years ago

@m52go I know we all have the best "interest" of Bisq in mind... wakawaka.

I understand you are classifying the bond as working capital. Indeed, the funds are working for Bisq, not the bond holder. If there is any interest accrued it should be reflected in the price of BSQ.

I'm no expert but I think there is a reason surety bonds do not pay interest on capital. I suspect it's some kind of conflict of "interest" (see what i did there). Using the Roger Ver example, any interest bearing surety bond would have enriched the Domain pirate, even after the bond is forfeit/burned.

The work being done on a surety bond is by the bond holder (e.g. not being a rogue domain pirate).

m52go commented 5 years ago

Expanding on what wiz and others have suggested...

As I see it, there are 2 ongoing elements of roles:

  1. maintaining a BSQ bond
  2. doing the ongoing work of the role

It's clear that charging the DAO for (1) is contentious, as it's not clear where the value is coming from or what the fairest way of capturing it is.

Maybe let's leave that aside and discuss (2), where there should be more certainty. Role owners have ongoing responsibilities that other contributors don't have that are not really deliverable. These range from relatively minor items like uploading videos for others (YouTube admin) to less minor items like maintaining various nodes (as sqrrm mentioned above). Altogether there's a fair bit of day-to-day work being done by role-owners that's not being compensated.

So perhaps we should think less in terms of "interest on bond" and more in terms of "value of ongoing work". This work doesn't fit the conventions of "merged to master" or "delivered to the end user" but there's no doubt that it's valuable.

Maybe this is a good opportunity to address it.

bodymindarts commented 5 years ago

The extreme example with the DNS entry (high bond / low work load) is valid. I think it is natural to want some compensation for the opportunity cost.

One could argue the return on investment is created by keeping bisq going at all. Without a DNS holder obviously all BSQ would diminish in value, or at least growth would be slower.

I agree that the contention is around the word 'interest'. We should rethink how ongoing roles are compensated. Perhaps there should be a flat fee instead of a payment for work done.

mpolavieja commented 5 years ago
  1. maintaining a BSQ bond

@m52go I think that your suggestion of structuring it as a dividend is fair (let´s say "a kind of dividend"). The fact that locked BSQ accrues compensation and not locked BSQ accrues nothing fully complies with the philosophy that only valuable contributions should be compensated. Although not accurate at all, a fair generic way to know if a contribution that involves no work or very little work is valuable is through profit (ie BSQ burnt > BSQ issued).

m52go commented 5 years ago

The fact that locked BSQ accrues compensation and not locked BSQ accrues nothing fully complies with the philosophy that only valuable contributions should be compensated.

@mpolavieja that's an interesting way of thinking about it.

To me the appeal of a dividend is that it would apply to all BSQ holders, which addresses the concerns of those in this thread against "interest" for a small group of people, while still achieving the goal of this proposal.

While I still like this idea in theory, I dismissed it because I don't see any practical way it could work, as it would require every BSQ holder to make a compensation request. I don't believe the DAO is designed to handle so many requests, and furthermore, I fear such a mechanism would change BSQ's public perception from being a utility token to some kind of investment.

I think the problem at hand is solvable by easier means, such as figuring out a way to compensate for ongoing work. It's a more solid justification for compensation, while indirectly providing a return for locked-up funds.

chimp1984 commented 5 years ago

Burning BSQ is conceptually a divident payment to all BSQ stakeholder. Each unit gets a bit more valuable if the total supply is lower due burnt BSQ. So each single fee payment by burning BSQ is a dividened payment to all BSQ stakeholders.

Beside that locking up BSQ removed available supply as well for the time the bond is locked up. So each bond holder creates a similar effect like burning at lockup time. At unlock time though it behaves more like an issuance (increasing supply my making BSQ available again).

chimp1984 commented 5 years ago

@wiz

Is this already implemented in the DAO ? If not, I don't think it's worth the risk of hard-forking the DAO to add such a feature.

It was not intended to make that automatic, but just to find a consensus what is a fair compensation for bonded BSQ so that all use the same calculation based on their bond value.

I think there is a confusion triggered by the term "interest rate".

All I suggest is to find a conesensus what bond holders should request as compensation for the service to be willing to lock up BSQ (and having the risk to get confiscated as well to be not able to react on volatility with trading the BSQ).

As mentioned, there are some roles where basically no effort is done (domain holder, dns,...) and beside that it would be good to have it transparent how much of a compensation for a normal role is for the bond.

E.g. a seed node operator might spend 80 EUR for hosting costs. He might take 100 EUR for his service to maintain and operate the node. He has to lockup 10 000 BSQ and he wants to get some compenastion for that as well. If we do not agree on a consensus, operator A might charge 10 BSQ and operator B might charge 100 BSQ which would trigger discussions.... Btw. hosting costs should be aligned as well so that there are not huge differences (but some differences are ok) - but that is another topic....

Similar consensus was found for translation work as far I know, but have not followed much that discussion and result. Compensations for work which is highly "fungible" should have a standard payment.

mpolavieja commented 5 years ago

@m52go A dividend does not necessarily need to accrue to all BSQs. Like different classes of stock issuance where not all classes have the same rights. In any case I would not call it a dividend but a compensation for specific role that is calculated in a specific way (i.e. a % of BSQ burnt - BSQ issued per cycle)

chimp1984 commented 5 years ago

I renamed the title to make it more clear and avoid that "interest rate" confusion. Also edited the text to make it more clear.

If still too confusing or too much problems with that, I am fine with closing it as rejected and leave it to others to find a suggestion how to deal with it.

MwithM commented 5 years ago

There's risk involving the tasks you are responsible for, which is different between roles and then there's risk involving cost of oportunity of locked funds and being "forced" to leave funds in a new and possibly volatile asset. This kind of risk is the same for all those who have their funds locked in bonds, even though they might estimate it in a different way because there might be contributors who feel more or less confident about the future of BSQ. I think that removing "interest rate" from the subject of this issue is correct, we just need to find a price for locked bonds. It's normal we don't find a price easily because we need to look to the future to value risks and possible value of this project, and there's not many locked bonds already to compare between different roles. Claiming a compensation request is a good way to pay to those willing to risk their funds, as funds are substracted from the DAO in form of inflation. For the future, if there are ennough trusted contributors interested, maybe we could stablish a periodical auction for bonding price to fulfill the roles needed by the DAO.

bodymindarts commented 5 years ago

2 questions:

chimp1984 commented 5 years ago

@bodymindarts We refer to bonds from bonded roles only, not reputation bonds or those used in the off-chain trade protocol. I will add the "bonded role" part again to the title and the introduction.

It is not necessary and we did not had consensus yet and bonded roles are already used. But it would lead to conflicts if people use different rates as it should be the same for all. Why should one seed node operator charge 100 BSQ and another 50 BSQ for a 10 000 BSQ bond.

Bondes roles are not designed to be picked up the market. They are given to people who have already earned reputation and trust. The confiscation part is just an extra security added to that based on trust and reputation. Also they are not intended to be changed frequently which migth be problematic for some (like nodes where onion address is hard coded).

m52go commented 5 years ago

Another way of thinking of locked-up bond is that it's a commitment to consistently do some quantity of non-deliverable work, as the bond is confiscated if this work isn't done.

Whether this compensation is calculated as a percentage of locked bond, or as a nominal BSQ amount per role based on estimated monthly commitment...that seems to be the main question here.

Upsides/downsides

Percentage of bond Ongoing compensation will not relate to effort. For example, Slack admin and seednode operators both post 20 000 BSQ bonds, but Slack requires very little work (if any) on a regular basis.

Estimated monthly commitment Compensation is more closely tied to performed work, but it's not related to posted bond. Domain name roles lose out here, for example, but I would expect the network to be willing to make reasonable exceptions for such cases.

Obviously the second case seems to be more favored here, but I want to articulate this to show that the first case shouldn't seem so far from the second one.

chimp1984 commented 5 years ago

Obviously the second case seems to be more favored here

Seems so, but not sure if it is caused mainly by misunderstandings (not clear if for bonded roles or for all bonds, etc.).

Maybe we should just close that now and once it becomes a real-life problem (different bond holders will request different % amounts) anyone can give it another try. Does not make sense to burn so much time/energy for such a minor topic.