Decoupling transaction fees from delegate pay...

[arhag]

Are you saying the speculators should subsidize the consumers just because they can afford it?  Since they're focused on making money, rather than on convenience in spending, I'd expect fees to be a stronger deterrent to speculators, especially when they can find lower fees in other markets in which they aren't subsidizing others.

Lower fees in what other markets? If people want to short USD (or BitUSD) relative to BTSX because they believe the price of BTSX in terms of USD will go up, and the demand for shorts is much greater than the demand for longs, then the shorts need to give something up to get their orders matched. Whether that is paying a specified interest rate to the bond holders, paying through larger bid-ask overlap fees, paying a capital gains tax, or "paying" through the opportunity cost of lower leverage due to higher margin held in collateral, they need some way to compete with other shorts. The question is what do we do with that value extracted from the shorts: give it to the bond holders / BitUSD yields entirely, or take a partial cut for the DAC expenses and growth. Obviously taking too high of a cut (compared to the added value to BTSX that the cut makes possible) would push the shorts to other markets (for example clones of BTSX that have a smarter policy). Giving some of the extracted value to the longs makes sense since it increases the demand for longs, which doesn't really mean much for the fees collected by the network but does mean that more holders are attracted to keep their wealth stored in the DAC (raising BTSX market cap). But my opinion is there is a limit to how much value we can transfer to the BitAsset holders before the gains from a marginal increase in BitAsset demand is overpowered by the gains possible with more strategic and productive uses of that value (made possible by giving it to the elected delegates/workers of the DAC).

[Troglodactyl]

Conceptually, I think the ideal is that transaction fees should pay for the real costs of transactions, and shorts should pay longs a variable amount depending on the current demand to short.  Is there any disagreement on this?

Yes.
Today the shorts seem  in great abundance.
We will have to change the system to make their life more bearable, long before we reach Visa-like transaction volumes...

I'm not understanding this.  Are you saying we need to make it easier to short?

I am predicting that we will do this before too long. I am not saying 'do it today'. Just that the need for this change will precede the problems you are talking in this thread. So, look for decisions that do not include putting more barriers and fees on the short.

I'm not suggesting more barriers to short. I want it to be easy, but expensive enough to match the aggregate market predicted outcome. If the market overall expects BTSX to rise, fees to short need to reflect that so shorting is betting that it will rise faster than the median prediction thinks it will.

[tonyk]

Conceptually, I think the ideal is that transaction fees should pay for the real costs of transactions, and shorts should pay longs a variable amount depending on the current demand to short.  Is there any disagreement on this?

Yes.
Today the shorts seem  in great abundance.
We will have to change the system to make their life more bearable, long before we reach Visa-like transaction volumes...

I'm not understanding this.  Are you saying we need to make it easier to short?

I am predicting that we will do this before too long. I am not saying 'do it today'. Just that the need for this change will precede the problems you are talking in this thread. So, look for decisions that do not include putting more barriers and fees on the short.

[Troglodactyl]

If BTSX fees are inadequate to pay the delegates' expenses, but BitUSD fees would be, doesn't that create an excellent incentive for delegates to do a hard fork which changes the fee distribution at that time?

Or are we worried that there may be non-updateable clients in the wild at that time, and hard forks will cause those users too much pain?

Getting rid of subsidies is more disruptive the longer they've been paid.  We're better off fixing the problem earlier.

Conceptually, I think the ideal is that transaction fees should pay for the real costs of transactions, and shorts should pay longs a variable amount depending on the current demand to short.  Is there any disagreement on this?

Yes.
Today she shorts seem  in great abundance.
We will have to change the system to make their life more bearable, long before we reach Visa-like transaction volumes...

I'm not understanding this.  Are you saying we need to make it easier to short?

The primary argument here is that in a BitUSD only use case (consumer->merchant, BitUSD<->USD), they pay nothing to the network (only other BitUSD users).  This could yield VISA level transaction volume with no compensation to the delegates.   

The delegates would have to make their money from margin calls and order submission/update on BTSX orders.   Here is the fact of the matter:  market orders will always out number consumer->merchant orders by an order of magnitude.   So assuming there are 4 market order transactions for every consumer transaction then the speculators can easily cover the costs of the consumers. 

Speculators are in the business of making money from their transactions so the fees don't hurt them nearly as much. 

Where do you get that market orders will always outnumber consumer orders by an order of magnitude?  More relevantly, are you implying that this will remain true not only globally, but in our specific market, regardless of the incentive structure we impose on that market?  O.o

Are you saying the speculators should subsidize the consumers just because they can afford it?  Since they're focused on making money, rather than on convenience in spending, I'd expect fees to be a stronger deterrent to speculators, especially when they can find lower fees in other markets in which they aren't subsidizing others.

[Troglodactyl]

Transfers incur a real processing cost. Someone will pay that cost. Separating decision makers from the costs of their actions just introduces perverse incentives, so subsidies are generally a bad idea. Inflation just allows the immediate decision makers (delegates) to offload the risk of their decisions onto the shareholders who elected them, so they can profit even when the system is losing money. More perverse incentives.

[arhag]

A high yield on BitUSD helps drive long-term growth.   A low spread helps as well.  As soon as we can get the spread low enough and the yield high enough so that yield covers to cost in one month, then the peg will be very strong and demand for BitUSD will go to the moon. 

If increasing the BitUSD yield from 5% to 10% barely helps increase user adoption of the DAC compared to say using that value to make the client better, pushing ads explaining the value of the system to new users, helping get more merchants on board, creating third-parties that provide multisig security, or any number of other possible things that can grow BTSX value faster, then it is smarter to give that delta yield to delegates/workers instead of BitAsset holders. Inflating BTSX to pay delegates to fund these useful operations just means that even after going to the moon, some amount of the BTSX value would have been unnecessarily transferred to BitAsset holders.

Let the shareholders decide. Not just regarding how much to inflate (if any) but also how to spend the value generated from all the possible fees that the DAC can capture. The alternative is to let these strategic decisions rest in the hands of the chaos of the exchange overlap fees. I think the benefit from letting shareholders make these strategic decisions outweighs the extra complication involved in the code to allow value to move between BitAsset and BTSX fees.

[bytemaster]

This is why paying delegates via dilution makes a lot of sense.    Rather than having to fund operations from customers, they can fund operations from shareholders and this will ultimately cause a focus on growing shareholder value.

For example:
1) high fees restrict trading volume and bots increasing spread... fees that are higher than necessary to subsidize other areas become a problem.
2) there are many things that delegates could invest in that increase shareholder value by more than the dilution...

A high yield on BitUSD helps drive long-term growth.   A low spread helps as well.  As soon as we can get the spread low enough and the yield high enough so that yield covers to cost in one month, then the peg will be very strong and demand for BitUSD will go to the moon. 

So I contend that giving shareholders the power to control a small amount of dilution via the existing delegate pay system is better than a more complex method of allocating and converting bitUSD fees to convert to BTSX and thus suppressing the bitUSD price. 

 

[tonyk]

Conceptually, I think the ideal is that transaction fees should pay for the real costs of transactions, and shorts should pay longs a variable amount depending on the current demand to short.  Is there any disagreement on this?

Yes.
Today she shorts seem  in great abundance.
We will have to change the system to make their life more bearable, long before we reach Visa-like transaction volumes...

[arhag]

I still think it's a mistake to give asset transaction fees as interest instead of allocating them through the delegates.  The sooner we have solid rules that are obviously sustainable, the sooner people will be comfortable building business on the network.  It's possible that BTSX transactions will have high enough volume to compensate for asset transactions, but if the goal is Visa level volumes in bitUSD, and shorts only pay one time BTSX fees to create bitUSD, it's plausible that relatively stable asset markets with high transaction volume could become millstones around the delegate's necks, and that creates uncertainty and deters serious commitment.  This also runs the risk of inflating BTSX fees to the point where only high volume transactions are cost effective, stratifying a class system for investors.  Even if these concerns are unrealized, subsidizing asset holders from BTSX holders adds unnecessary economic complication and uncertainty, and hard codes a marketing agenda that I think would be better left at the fluid discretion of delegates and investors.

I absolutely agree. Money is fungible. It doesn't matter whether it is BitUSD or BTSX fees. Ultimately we must decide how to distribute this value between BitAsset holders, BTSX shareholders, and delegates. This is why I have been trying to push my proposal to let the shareholders (indirectly) decide how to distribute the value (see here and here).

Conceptually, I think the ideal is that transaction fees should pay for the real costs of transactions, and shorts should pay longs a variable amount depending on the current demand to short.  Is there any disagreement on this?

I think if we want to do shorts paying longs a variable interest, then it should be done in a separate market (the bond market). However, it would take the short pressure away from the BitUSD/BTSX market and so the revenue from bid ask overlap fees would drop, which may not be such a bad thing from the perspective of making the peg stronger. The reduction in bid ask overlap fees means the average yield rates on BitAssets would significantly drop. But that is okay since the BitUSD demand could instead shift to the bond holders who would experience the high returns comparable (or even greater) than the BitAsset yields. The good thing about capturing the bid ask overlap as revenue and distributing it as yields (rather than the bond market) is that the DAC can take a cut for itself. There is a limit to how much we can get away with doing this however since a clone could take a smaller cut and provide more attractive yields for its own BitAssets. But my point is that we should let the shareholders decide what kinds of yields they want to give to each BitAsset to help adoption of their DAC, and how much of a cut they should take for the purposes of helping DAC adoption in other ways (development and marketing).

[theoretical]

If BTSX fees are inadequate to pay the delegates' expenses, but BitUSD fees would be, doesn't that create an excellent incentive for delegates to do a hard fork which changes the fee distribution at that time?

Or are we worried that there may be non-updateable clients in the wild at that time, and hard forks will cause those users too much pain?

[bytemaster]

The primary argument here is that in a BitUSD only use case (consumer->merchant, BitUSD<->USD), they pay nothing to the network (only other BitUSD users).  This could yield VISA level transaction volume with no compensation to the delegates.   

The delegates would have to make their money from margin calls and order submission/update on BTSX orders.   Here is the fact of the matter:  market orders will always out number consumer->merchant orders by an order of magnitude.   So assuming there are 4 market order transactions for every consumer transaction then the speculators can easily cover the costs of the consumers. 

Speculators are in the business of making money from their transactions so the fees don't hurt them nearly as much. 

[xeroc]

I think you see a problem where there is not one. People can still pay the fees in BTSX even if they are trading/transferring Assets. Plus  it is 2 times cheaper to pay in BTSX - enough customers will be aware of that.

I am eager to see how that turns out in the long run ... exciting ..

[tonyk]

almost true ... trading assets involves paying fees in both .. the bitAsset (ie bitUSD) and BTSX... the btsx part goes to the delegates

(first time use in a post)

This doesn't conflict with anything I said. The issue is that BTSX fees are only involved in trade involving multiple assets. If existing bitUSD is transferred between different users as payment, no BTSX fees accumulate. The traders subsidize the payment system users, and if those user bases grow disproportionately (as they should with one subsidizing the other) we have problems.

I think you see a problem where there is not one. People can still pay the fees in BTSX even if they are trading/transferring Assets. Plus  it is 2 times cheaper to pay in BTSX - enough customers will be aware of that.

[Troglodactyl]

almost true ... trading assets involves paying fees in both .. the bitAsset (ie bitUSD) and BTSX... the btsx part goes to the delegates


(first time use in a post)

This doesn't conflict with anything I said. The issue is that BTSX fees are only involved in trade involving multiple assets. If existing bitUSD is transferred between different users as payment, no BTSX fees accumulate. The traders subsidize the payment system users, and if those user bases grow disproportionately (as they should with one subsidizing the other) we have problems.

[xeroc]

almost true ... trading assets involves paying fees in both .. the bitAsset (ie bitUSD) and BTSX... the btsx part goes to the delegates


(first time use in a post)