Author [EN] [ZH] [ES] [PT] [IT] [DE] [FR] [NL] [TR] [SR] [AR] [RU] [EN] [ZH] [ES] [PT] [IT] [DE] [FR] [NL] [TR] [SR] [AR] [RU] [EN] [ZH] [ES] [PT] [IT] [DE] [FR] [NL] [TR] [SR] [AR] [RU] Topic: Ordering shorts by collateral introduces systemic risk: Problem and solution  (Read 392 times)

0 Members and 1 Guest are viewing this topic.

Offline theoretical



This is a small series of ~5 posts.  I invite you to read each one on its own, then stop and think about the post's final thought for a minute before moving on to the next post.

Example from new system:
You use (tie up) $10,000 worth of BTSX to post a large collateral to short 1,000 bitUSD into existance (at par)… 15mins later someone offers for sale 1,000 bitUSD for $970… would you cover?

Sure why not, you just made $30 in 15 mins, and you can get right back to the front of the shorting line by putting up your big collateral again.  If you have a robot doing this you are making $120/ hr.

This comment made me think of a little story.

Alice has $1000 worth of BTSX and she wants to go short $1000 BitUSD at a 2x collateral ratio.  Bob wants to buy BitUSD at the feed price.  But Bob always buys from deep-pocketed Dan instead of Alice, because Dan has more money than Bob and is willing to offer all of it at a 9.99x collateral ratio.

So Alice calls her rich friend Carrie, and upon seeing Alice's business plan (short $1000 BitUSD), Carrie agrees to loan Alice $8000 worth of BTSX.  Combining Alice's $1000 cash with creditor Carrie's $8000 loan, Alice can successfully compete with Dan:  Including Bob the BitUSD buyer's purchase price of $1000 worth of BTSX, the total collateral is $10000 worth of BTSX, Alice now has a 10x ratio that narrowly out-competes Dan's offer.

Giving friends discounts in business deals is against Carrie's religion.  So the interest rate Carrie will charge Alice depends on the risk in Alice's business plan.  But all the credit rating agencies say that, based on Alice's business plan, her possibility of default is very very small.  Especially if the loan has a clause allowing Carrie to margin call the debt and force liquidation if Alice's equity in the deal starts to get dangerously low.

So Carrie is willing to offer the loan to Alice at very good interest rate, and everybody's happy (except for Dan who's just been out-competed, but hey, it's a free market so that's okay).

However, there is one small problem, which I'll explain in the next post...
BTS- theoretical / PTS- PZxpdC8RqWsdU3pVJeobZY7JFKVPfNpy5z / BTC- 1NfGejohzoVGffAD1CnCRgo9vApjCU2viY / the delegate formerly known as drltc / Nothing said on these forums is intended to be legally binding / All opinions are my own unless otherwise noted / Take action due to my posts at your own risk

Offline theoretical

From the story in the previous post, the next natural step is to have a "BTSX money market" to match potential speculators like Alice, with potential creditors like Carrie.

The problem with this situation should be obvious to anyone who was paying attention during the US financial crisis at the end of the last decade.  Alice and other shorters' debts will be very highly rated, so a lot of people like Carrie will put a lot of money into them with an expectation that the principal will be preserved.

If BTSX takes a nose dive and covering Alice's short would dip into Carrie's $8000 principal, then Carrie and all the other creditors will start putting a ton of political pressure on the central bank to print BitUSD or BTSX to make them whole.  Regardless of the central bank's response or non-response, the crisis of confidence will jeopardize the entire system.

You can't stop Alice and Carrie from finding each other and making this business arrangement -- if the blockchain doesn't get built-in support for it, they'll just find a way to do it off-chain.

You can't stop Carrie from mentally turning "this deal is approved by all the credit rating agencies and has a very low probability of default" into "this deal has a zero probability of default."  Humans are notoriously poor when it comes to reasoning about very rare events.

The best way to prevent a financial crisis would be to have Alice require no capital from Carrie at all.  Allow positions like Alice's that want safe, but lower than top-of-book, leverage ratios to offer to pay interest directly to the network.

Should this interest be denominated in BTSX or BitUSD?  Who should get it?  How much interest should be charged?  What about the opening quote in the previous post?  Won't Carrie still want to do something with her capital?  Read more in the next post...
« Last Edit: October 01, 2014, 06:07:11 AM by drltc »
BTS- theoretical / PTS- PZxpdC8RqWsdU3pVJeobZY7JFKVPfNpy5z / BTC- 1NfGejohzoVGffAD1CnCRgo9vApjCU2viY / the delegate formerly known as drltc / Nothing said on these forums is intended to be legally binding / All opinions are my own unless otherwise noted / Take action due to my posts at your own risk

Offline theoretical


The amount of interest to be charged should be some percentage of the amount Alice would need to borrow to put her order at the top of the book, above Dan's, plus some risk premium to compensate the network for the possibility that Alice could default.

A 9x collateral position that needs boosted to 10x will be a lot less likely to default than a 2x collateral position being boosted to 10x.  So the risk premium for the 9x position should be substantially less.

Exactly what the risk premium and interest rate should be, can be determined numerically (and likely even analytically) from a geometric-random-walk model.  I have already written a thread about this:  https://bitsharestalk.org/index.php?topic=9520.msg123853#msg123853

I would add one modification.  Consider a technical trader named Ted, who usually closes his short positions within an hour or two of opening them.  Offering an enormous APR has very little cost to Ted since his position is open for so little time.  But offering an enormous APR does have a great benefit for Ted:  He can jump directly to the top of the book!

To better align Ted's incentives with policy goals, I propose a simple modification which I call the "three-day rule:"  Covering earlier than, say, three days, will be charged the same interest as if the position had been covered after three days.  This is equivalent to charging a percentage fee to open the short, and then offering a credit on the interest invoice equal to the fee paid.  If the position wasn't open long enough for the interest due to exceed the credit, the unused portion of the credit is income for the network, and a loss for people like Ted.

The remaining questions are as follows:

- Who should get the interest?

- Didn't Agent86 and bytemaster just get done explaining that tying up a large amount of wealth in collateral is a desired policy goal?

- Should the interest be BitUSD-denominated or BTSX-denominated?

Read the next post for some answers...
BTS- theoretical / PTS- PZxpdC8RqWsdU3pVJeobZY7JFKVPfNpy5z / BTC- 1NfGejohzoVGffAD1CnCRgo9vApjCU2viY / the delegate formerly known as drltc / Nothing said on these forums is intended to be legally binding / All opinions are my own unless otherwise noted / Take action due to my posts at your own risk

Offline theoretical

I'll say the interest will be in "coins" in this post; I'll reveal whether interest should be BitUSD- or BTSX-denominated in the next post.

The simplest thing to do with the interest is to burn it as fees (if BTSX-denominated) or pay it as yield (if BitUSD-denominated).

We can do something more sophisticated, however:  If we want a large amount of capital tied up for a long time, we can use the interest to incentivize long-term investors to remove their capital from circulation.  We can let a long-term investor like Lenore make an order such as "I'll pay 1000 coins today, in order to receive 1030 coins in a year."  Ludwig might make a competing order that says, "I'll pay 1000 coins today, in order to receive 1025 coins in six months."

The network will convert these competing orders to a common scale -- APY.  If the network's only received, say, 25 coins in interest payments, then it will partially fill Lenore's order (which has a lower APY than Ludwig's order) and leave Ludwig's order unfilled.  To ease network load, maybe filling of partial orders should only happen once an hour or so.

This soaks up Lenore and Ludwig's spare capital without potentially setting up BitShares for its very own version of the late 2000's (decade) financial crisis.  And definitively addresses the issue of how we can still tie up large amounts of capital, even if we allow 2x positions to offer interest to out-compete higher-leverage positions (which is an issue I'm sure Agent86 would raise, based on his quote at the top of this thread).

Should the interest charged to shorts be BitUSD-denominated or BTSX-denominated?  The final question will be answered in the next post...
« Last Edit: October 01, 2014, 06:35:33 AM by drltc »
BTS- theoretical / PTS- PZxpdC8RqWsdU3pVJeobZY7JFKVPfNpy5z / BTC- 1NfGejohzoVGffAD1CnCRgo9vApjCU2viY / the delegate formerly known as drltc / Nothing said on these forums is intended to be legally binding / All opinions are my own unless otherwise noted / Take action due to my posts at your own risk

Offline theoretical

I proposed basically the exact idea in the previous post, which I called "BitBonds," almost a month ago.  I noted then that it was a good way to tie up arbitrarily large amounts of investor capital for long periods of time -- see https://github.com/drltc/bitbond-proposal/blob/master/bitbond.md  But it didn't get much of a response.

The original problem we're trying to solve is that Carrie, who's looking for a long-term low-risk fixed-rate investment, is tempted to become the creditor of her friend Alice.  (Alice needs to borrow money because she wants to short BitUSD at a leverage ratio which is fairly safe, but much lower than the top of book at the short wall.)

We provided a relief mechanism at the speculator's side by giving Alice a viable alternative, paying interest to the network.

We can also simultaneously provide a relief mechanism at the creditor's side:  We can give Carrie an alternative investment that's even safer than investing in Alice.  Instead of becoming Alice's creditor, Carrie can simply invest in BitBonds.

Because Carrie becomes a systemic risk when she partners with Alice, and our strategy for dealing with Carrie is to make BitBonds a better investment than Alice, the solution is obvious:  We want to have BitBonds be denominated in whatever currency would be most attractive to people like Carrie.  Carrie, as a long-term low-risk fixed-rate investor, will be most interested in stable-value currency.  Which means that Carrie will be most interested in a BitUSD-denominated BitBond.  Since the interest on BitBonds is funded from interest charged to shorts, if we want BitUSD-denominated BitBonds, we should charge BitUSD-denominated interest on shorts.
« Last Edit: October 01, 2014, 06:11:12 AM by drltc »
BTS- theoretical / PTS- PZxpdC8RqWsdU3pVJeobZY7JFKVPfNpy5z / BTC- 1NfGejohzoVGffAD1CnCRgo9vApjCU2viY / the delegate formerly known as drltc / Nothing said on these forums is intended to be legally binding / All opinions are my own unless otherwise noted / Take action due to my posts at your own risk

Offline arhag

  • Hero Member
  • *****
  • Posts: 1213
    • View Profile
    • My posts on Steem
  • BTS: arhag
  • GitHub: arhag
If BTSX takes a nose dive and covering Alice's short would dip into Carrie's $8000 principal, then Carrie and all the other creditors will start putting a ton of political pressure on the central bank to print BitUSD or BTSX to make them whole.  Regardless of the central bank's response or non-response, the crisis of confidence will jeopardize the entire system.

Hold on. Creditors like Carrie will be holding a security that has nothing to do with BitShares X. We can't stop them from doing that. If they lose all of their money because of malinvestment, that is too bad for them, but it doesn't mean BitUSD would be undercollateralized. BitShares X only cares about maintaining the peg of its BitAssets. It is not responsible for other investments people want to make. And everyone should be clear on this. There is no reason to take seriously the outside investors whining about BitShares X doing something to make them whole.


Regarding your third post:

Your already complicated proposal is now getting even more complicated (three-day rules?). Seriously, what is wrong with the capital gains tax method I described a while ago. Shorts would only be able to cover through a BitAsset buy order at market rate using the BTSX held in collateral. If BTSX needed to buy the BitAsset for covering was less than the initial reserve (just the "reserve", according to my terminology in that post, not the "margin"), the difference would be taxed at a rate that the short set when creating the short position. This tax rate is used to prioritize short matching at the same price. If desired for the purpose of simplifying the trader's decision making, the tax rate could only apply if the short was matched at the feed price. Extra margin could be kept with an open short order in order to allow it to remain active, within the price limits set by the short, even as the feed price increased (until the margin was no longer enough to meet the minimum 200% collateral requirement at the feed price). Any extra BTSX that was not necessary to meet the 200% collateral requirement at the time the short was matched would not be locked up in the short position; it would instead go back to the trader's balance. Alternatively, the short order could be set up to always use all the margin provided if the short was matched within the price limits by adjusting the amount of BitUSD for sale to satisfy the 200% collateral requirement. If the price of the BitAsset relative to BTSX goes up, the short does not have to pay any additional fees to cover (assuming no margin call). Ted, your fast trader, has no greater incentive to raise the tax rate to jump ahead of the line any more than the medium term traders. If the price of the BitAsset relative to BTSX goes down shortly after opening the short position, the trader has incentive to cover their short early and re-short to compound gains, keeping liquidity high. The funds collected from the capital gains taxes can be used to increase demand for the BitAssets through higher yields. There is still a negative feedback loop, although the propagation of information in this feedback loop might be slower than your interest proposal. Even then, now with the new maximum holding period of a month and with market making bots, the time delays shouldn't be too significant (and if yields saturate at a cap of 5% for BitCurrencies like I hope for, it shouldn't even matter because the demand shouldn't be very sensitive to the tax rates). My proposal can also be modified to allow competing with increased collateral at first up to some limit (like 400%) before then competing on tax rates, although I personally don't see the need for this because I think a 200% reserve is already good enough.


Regarding your fourth and fifth posts (and BitBonds in general):

I have already talked about my thoughts on this issue with you actually. My view is that the money raised from fees on the shorts should be used by the DAC for two purposes: one, used to provide a reasonable yield on BitCurrencies as a marketing strategy to get initial adoption; and two, sent to entities/individuals the shareholders think can grow the value of the DAC best (whether that is regular delegates, business delegates, or workers chosen by proposals ratified by shareholder vote) or alternatively burned as a dividend to shareholders if we are past the growth stage of the DAC. The reasonable yield just needs to be good enough to be better than what people can get holding their fiat in a traditional bank; I think the marginal increase in BitCurrency demand by raising the interest rate higher is not worth the opportunity cost of not using those excess yield funds to grow the DAC in smarter ways. And, as someone else on the forum said, making it too high makes people suspicious and makes them think this whole thing is a scam. I think a yield up to 5% is great. That is up to, meaning it is variable with a cap of 5%. My guess is that in the early stages the yield would be consistently saturated at the 5% cap for a long time (which makes it easier for people to reason about by the way because they can just think they get a consistent 5% per year).

Now from what I remember you want the excess revenue to be given as additional interest via BitBonds. Why? What are the holders risking to get that kind of return by the way. What are we (the BTSX holders) getting out of them holding the BitBonds? Just that they hold their BitAssets for a long period of time? We would already be making BitUSD and the like extremely attractive to hold with the 5% yield compared to dollars in outside banks. I think we can do smarter things with that extra revenue that would be going to BitBonds in your proposal. Besides if they actually believe in the system and they already have more than enough wealth with price stability for the short-term and want higher returns over the medium to long period of time, then in my view they should hold BTSX (at least during the growth phase).
« Last Edit: October 01, 2014, 09:01:08 AM by arhag »

Offline theoretical

Now from what I remember you want the excess revenue to be given as additional interest via BitBonds. Why? What are the holders risking to get that kind of return by the way. What are we (the BTSX holders) getting out of them holding the BitBonds?  Just that they hold their BitAssets for a long period of time?

Yes.  I don't like having shorts compete on collateral.  But having them compete on collateral does help take care of the problem of unbalanced short supply / long-holder demand for BitUSD.

If demand is unbalanced, the price of BitBonds will fall until the yield is competitive with investors' projected return on BTSX.  So it achieves the policy goal of giving large amounts of capital something to do other than short BTSX.

Besides if they actually believe in the system and they already have more than enough wealth with price stability for the short-term and want higher returns over the medium to long period of time, then in my view they should hold BTSX (at least during the growth phase).

The problem is that everybody believes in the system so much that going long BTSX isn't sufficient for them, they want to short BitUSD and have effectively a leveraged bet on long BTSX.  If we offer no-reserve auction on BitUSD, we can balance the two sides.

My view is that the money raised from fees on the shorts should be used by the DAC for two purposes: one, used to provide a reasonable yield on BitCurrencies as a marketing strategy to get initial adoption; and two, sent to entities/individuals the shareholders think can grow the value of the DAC best (whether that is regular delegates, business delegates, or workers chosen by proposals ratified by shareholder vote) or alternatively burned as a dividend to shareholders if we are past the growth stage of the DAC.

We would already be making BitUSD and the like extremely attractive to hold with the 5% yield compared to dollars in outside banks. I think we can do smarter things with that extra revenue that would be going to BitBonds in your proposal.

I would support capping the yield on BitAssets to +5% and using the funds for something else.  I like the business delegates idea.
BTS- theoretical / PTS- PZxpdC8RqWsdU3pVJeobZY7JFKVPfNpy5z / BTC- 1NfGejohzoVGffAD1CnCRgo9vApjCU2viY / the delegate formerly known as drltc / Nothing said on these forums is intended to be legally binding / All opinions are my own unless otherwise noted / Take action due to my posts at your own risk

 

Google+