Author Topic: FDIC for BitUSD  (Read 26223 times)

0 Members and 1 Guest are viewing this topic.

Offline luckybit

  • Hero Member
  • *****
  • Posts: 2921
    • View Profile
  • BitShares: Luckybit
Well, perhaps I'll end up being one of the lone dissenting voices on this, but it seems to me that it might be worthwhile to think twice about exclusive reliance on 3).

I really worry that 3) could be a significant drag on the value of BTSX and on buy-in into the network. It all goes back to the question of how BTSX holders and investors would feel about potentially open-ended debasement. In a real, live network with 3), how often would short positions get blown out, and how much dilution would need to occur to withstand multiple disruptions over time?

I have the same concerns. We probably should not call it debasement. We should also avoid words like dilution because it causes massive communication hurdles. It functions more like a stock buyback vs dilution. You cannot debase a stock but you can dilute it. It probably will not affect us very much in the long term though if handled right.
What if there is not just one BitAsset, but several correlated BitAssets, e.g., BitBTC, BitGold, BitCNY, etc., which could potentially increase rapidly in value and cause blown shorts? Would all of these need to be supported by BTSX holders under something like 3), and would that compound the potential problems?

I suspect not many people here have a good idea about how often and how large the triggering disruptions would be--I certainly don't. There seem to be significant unknowns here. But I feel pretty sure about one thing: in general, significant uncertainty + potentially large, open-ended future dilution = anathema to buy-and-hold investors.

The way to do think about and frame it is that the DAC provides a loan to itself in order to solve short term disruptions to it's operations. This is very reasonable and a better way to think about it.

But that loan must be paid back and there is a limit to how much of a loan it can be. The cap of 2 billion XTS should never be changed. The XTS burned on the other hand can be recreated again up to 2 billion XTS in the case of an emergency but there should never ever be more XTS than when we all bought into it which means the earlier you buy into it the safer you are.

The other point is that if you get in significantly later you're probably safe too because disruptions will not be as severe. For example if in a year we have 1 billion XTS left then it would take an extreme disruption to cause the creation of 50% more XTS once the market cap is in a high enough amount.

Suppose for sake of argument we do have a creation of 100 million of XTS in an event. As long as the burn rate, transaction fees and other mechanisms adjust to pay off the loan then the best time to buy would be immediately after these events because the burn rate will be at it's highest as the network seeks to rebalance and pay off the debt. The debt would be considered paid off once the amount of XTS reaches a lower place than it was at prior to the loan due to burning / buyback functions.

BM et al. deserve props for thinking proactively about this issue, and I hate to complain, but I really do hope a better, more investor-friendly solution comes along. What about some mixture of (1), (2), and/or (3)? Are these three really the only alternatives?

BM is offering us a solution which could work. But we should definitely refine this solution in any way we possibly can and even think up a better solution.

I recognize that the XTS supply needs to be dynamic. I also recognize there has to be a deflationary incentive for people to want XTS in the first place.

So we started at 2 billion and now we've burned through a few hundred thousand XTS.
As long as the burn rate is faster than the rate of creating new XTS we all win.

So we can create new XTS just as long as we increase the burn rate in response to eventually balance things out. The whole thing should seek equilibrium and balance so that overall the burn rate is always greater in the long term than the creation rate.

So basically the company burns and this functions like the company buying it's own stock. All of our stock appreciates in value when this happens and it's generally good for us in the long term.

But when a short term problem is faced the algorithm must seek to solve the short term problem in the most efficient manner. Efficiency should include maintaining the burn rate so that whatever increase we experience in XTS is temporary. Transaction fees could increase in response or the burn rate could increase temporarily as a way to try to balance the concerns.

So really instead of thinking of it as taking one for the team, or "debasement" which is a word we should never use in this context. We should think about it as the network providing a loan to itself and the way to pay that loan back is with an increase in fees. Think of fees as taxes which are used to pay back the loan in the form of an accelerated buy back. If this doesn't happen very often then it's not going to be a big deal just as long as we frame it as a loan, a buy back, but avoid toxic words like "debasement", "dilution", "inflation", because no one likes these concepts.

If someone cannot pay their debt we can bail them out but that has to be accounted for.
« Last Edit: August 12, 2014, 06:57:56 am by luckybit »
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

Offline luckybit

  • Hero Member
  • *****
  • Posts: 2921
    • View Profile
  • BitShares: Luckybit
One of the nagging issues we have been facing is what happens when the collateral behind a short position is not enough.  Our options have been:

1) Freeze all markets until someone who holds BitUSD is willing to sell at a loss.
2) Debase BitUSD by allowing unbacked BitUSD to circulate
3) Debase XTS by issuing new XTS to cover the loss.

Test net 12 implemented Option #1... but this option depends upon someone being willing to "take one for the team" to restart the markets.  This is a principle I don't like.
Debasing BitUSD everytime this happens would gradually break the peg and BitUSD would become an asset that is correlated to USD by some constant factor that changes slightly every time a short is blown out.   This makes the system appear to be broken.

Option 3 was considered too risky because someone could manipulate the market to print up unbounded XTS.  Fortunately we have introduced price feed / rate-of-movement restrictions that prevent this particular attack from yielding unlimited XTS.    Without being able to attack in this manner Option 3 returns as the best overall approach. 

What this means is that all XTS holders stand behind the "market peg" and that those who hold USD have a stronger peg.

I prefer option 2. BitUSD is really just a polymorphic asset which doesn't require a fixed supply.

Option 3 is acceptable if there is no better option. Generally it's true that XTS exists primarily as a unit of measure and not as a unit of account. BitUSD on the other hand is meant to be the unit of account. The BitUSD gets created from XTS and XTS has to back the BitUSD for that to happen.

I'd prefer if BitUSD get created out of thin air rather than XTS becuase XTS is supposed to not act in that way but I realize it can solve the problem with option 3 and the problem will only be a temporary problem. It's a problem mainly because we don't have a critical mass of people using Bitshares.

If we had a critical mass of people then the risk of this occurrence diminishes. XTS generally can be created or destroyed but we want it to always trend toward destroyed because that is the source of the dividend action.

So really it's not a dividend it's more like a buyback when the XTS gets destroyed. When XTS needs to be created it can be created. But there should be some kind of cap on how much XTS can ever be created.

So perhaps we could end up going back up to 2 billion XTS because that cap can be set in stone forever but the amount which gets destroyed is what can be pulled from to create.
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

Offline amatoB

  • Full Member
  • ***
  • Posts: 62
    • View Profile
Well, perhaps I'll end up being one of the lone dissenting voices on this, but it seems to me that it might be worthwhile to think twice about exclusive reliance on 3).

I really worry that 3) could be a significant drag on the value of BTSX and on buy-in into the network. It all goes back to the question of how BTSX holders and investors would feel about potentially open-ended debasement. In a real, live network with 3), how often would short positions get blown out, and how much dilution would need to occur to withstand multiple disruptions over time?

What if there is not just one BitAsset, but several correlated BitAssets, e.g., BitBTC, BitGold, BitCNY, etc., which could potentially increase rapidly in value and cause blown shorts? Would all of these need to be supported by BTSX holders under something like 3), and would that compound the potential problems?

I suspect not many people here have a good idea about how often and how large the triggering disruptions would be--I certainly don't. There seem to be significant unknowns here. But I feel pretty sure about one thing: in general, significant uncertainty + potentially large, open-ended future dilution = anathema to buy-and-hold investors.

BM et al. deserve props for thinking proactively about this issue, and I hate to complain, but I really do hope a better, more investor-friendly solution comes along. What about some mixture of (1), (2), and/or (3)? Are these three really the only alternatives?

merockstar

  • Guest
This happens because of xts volatility, right?

so in theory, in a more mature system (higher market cap) its going to be much less likely to happen than it is in testing because a higher market cap = a more stable price.

so as more people adopt bitUSD and btsx, the need to print extra btsx to cover shorts becomes less and less likely.

training wheels. right?

if I understand this correctly, then I agree with everyone else. do what needs to be done.

Offline Shentist

  • Board Moderator
  • Hero Member
  • *****
  • Posts: 1601
    • View Profile
    • metaexchange
  • BitShares: shentist
I'm not sure people understand how monumental 3) is.

So, here is the design I had in mind, when contemplating an alternative to BitShares X.

Let's say the unit is U for simplicity.

BTC -> U cross-chain conversion is allowed in the chain itself through smart contracts. Not like XCP or MSC, but delegates/smart oracles would ensure proper conversion between BTC and U.

Then from U you have ability to convert to v.USD (which is your BitUSD). I'll just use BitUSD since it's more familiar with the audience.

Huge difference here is that, instead of expecting a market depth to get created, market feeds are used to let the DAC itself issue BitUSD by destroying U.

When user wants to liquidate BitUSD, it gets converted back to U, but that's done by the DAC itself, which provides LIQUIDITY!

It does it with 2% around the median market feed, so it still allows for people to bet within the spread, but if some heavy buy or sell comes in, it can not spook the market.

Value of unit U: Why hold it at all? Well it receives dividents from all the trades. (similar to BTSX)


Major difference between my design and 3) is that the DAC itself provides liquidity. While BTSX expects the market to provide that.

UPDATE: although 3) is a step in that direction (DAC providing liquidity)


in short the DAC is the marketmaker!

maybe a needed function

Offline bitmeat

  • Hero Member
  • *****
  • Posts: 1116
    • View Profile
I'm not sure people understand how monumental 3) is.

So, here is the design I had in mind, when contemplating an alternative to BitShares X.

Let's say the unit is U for simplicity.

BTC -> U cross-chain conversion is allowed in the chain itself through smart contracts. Not like XCP or MSC, but delegates/smart oracles would ensure proper conversion between BTC and U.

Then from U you have ability to convert to v.USD (which is your BitUSD). I'll just use BitUSD since it's more familiar with the audience.

Huge difference here is that, instead of expecting a market depth to get created, market feeds are used to let the DAC itself issue BitUSD by destroying U.

When user wants to liquidate BitUSD, it gets converted back to U, but that's done by the DAC itself, which provides LIQUIDITY!

It does it with 2% around the median market feed, so it still allows for people to bet within the spread, but if some heavy buy or sell comes in, it can not spook the market.

Value of unit U: Why hold it at all? Well it receives dividents from all the trades. (similar to BTSX)


Major difference between my design and 3) is that the DAC itself provides liquidity. While BTSX expects the market to provide that.

UPDATE: although 3) is a step in that direction (DAC providing liquidity)
« Last Edit: August 12, 2014, 01:24:29 am by happypatty »

Offline spartako

  • Sr. Member
  • ****
  • Posts: 401
    • View Profile
wallet_account_set_approval spartako

Offline bytemaster

Quote

The analogy is almost perfect... the US prints new dollars to keep depositors whole when banks run out of collateral backing their loans.   

US Bank creates USD backed by collateral of a house.  Housing market collapses and loan defaults so there is insufficient collateral to "buy back the USD" to take it out of circulation.    With FDIC they print new dollars to cover the loss... a bail out of the shareholders paid for by everyone.  This would be like allowing unbacked BitUSD to circulate.

What should happen is that the banks owners (shareholders) should make good on its loans (those that lent the bank USD, ie: depositors).  The only way for a real bank to do this is to sell shares in itself to raise the capital.  If it is unable to raise enough capital by selling shares then it should give the depositors the shares.   We are effectively building this bailout through share issuance into the system.

The shareholders are borrowing USD into circulation with a promise to pay 1 USD worth of shares in the future.   When we borrow them into existence, we lend them to the short which then sells them.  We lend to the short because the short has provided enough collateral that the shareholders consider it "low risk".   When viewed from this perspective it makes everything perfectly clear and natural.

Where the bank issues new shares in itself to cover the loss.   

In other words... there is no limit to how much can be printed because XTS holders implicitly back all BitUSD issuance.  If issuance starts to get too high (diluting the shareholders beyond 2B XTS) then that means delegates need to burn more and charge higher transaction fees to cover the risk.  The system can still earn a profit, the shareholders just have to choose to "save in advance" or "dilute as necessary". 
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline tonyk

  • Hero Member
  • *****
  • Posts: 3308
    • View Profile
I agree with your thoughts here:
https://bitsharestalk.org/index.php?topic=6793.msg91116#msg91116

re: 3)

The rules on that, is something the community  can chime in ,if you have thought about them already.

Or the system just gonna buy them right away anytime the collateral is not enough?

The system will buy right away at any price up to 33% below the median price feed.

Are you sure we can not find potentially better solutions?
1. As in giving some chance for the market to recover first?
2. Possibly other improvements?
Lack of arbitrage is the problem, isn't it. And this 'should' solves it.

Offline Riverhead

re: 3)

The rules on that, is something the community  can chime in ,if you have thought about them already.

Or the system just gonna buy them right away anytime the collateral is not enough?

The system will buy right away at any price up to 33% below the median price feed.


And the pool of funds it draws from is increasing BTSX supply? So the delegates burn it (along with lost private wallets) and the market mints it? Will there be a hard cap on 2 billion supply or is it mathematically highly unlikely the mint rate will catch the burn rate?  Or is that something the delegates will need to control with their income? If the latter then we need to reconsider the policy of not raising fees once the "crisis" passes.
« Last Edit: August 11, 2014, 08:31:25 pm by Riverhead »

Offline bytemaster

re: 3)

The rules on that, is something the community  can chime in ,if you have thought about them already.

Or the system just gonna buy them right away anytime the collateral is not enough?

The system will buy right away at any price up to 33% below the median price feed.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline tonyk

  • Hero Member
  • *****
  • Posts: 3308
    • View Profile
re: 3)

The rules on that, is something the community  can chime in ,if you have thought about them already.

Or the system just gonna buy them right away anytime the collateral is not enough?
Lack of arbitrage is the problem, isn't it. And this 'should' solves it.

Offline MrJeans

  • Hero Member
  • *****
  • Posts: 599
    • View Profile
  • BitShares: mrjeans
Option 3 makes the most sense to me.

Offline bytemaster

Would this mean price feeds would be permanent? I'm OK with that, just curious.

No... eventually the price feeds will just become the 24 hour moving average.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline Riverhead

Would this mean price feeds would be permanent? I'm OK with that, just curious.