Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - starspirit

Pages: 1 ... 10 11 12 13 14 15 16 [17] 18 19 20 21 22 23 24 ... 64
241
General Discussion / Re: A Generalised Approach to Yield on BitAssets
« on: May 17, 2015, 12:46:07 am »
A middle-ground option: Cash + At-Call Interest Bearing Deposits

For completeness there is a third option which could be worth exploring.

Traditional money markets have 3 options working in unison - cash, deposit accounts, and term investments.

The Yield Option on bitAssets explored above is like a substitute for a deposit account. It could effectively do away with the need for a cash option, which arguably may no longer be required in a digital economy. On the other hand, there may be a good use for a separate cash option (e.g. improved anonymity or other features?) that perhaps has not been fully explored or implemented yet, and perhaps we want to keep this option open.

Having a cash and at-call deposit option working together could still offer at-call yield potential through the deposit market, and so still achieve the same benefits of a yield-based bit-Currency. The main difficulty is dealing with the situation where natural market rates are negative. There would have to be an enforceable way to charge interest on cash to return to shorts (probably method 1 in the OP). An open question is how important negative interest rates really are.

I think it is important to discuss how we see these different components working together.

242
General Discussion / A Generalised Approach to Yield on BitAssets
« on: May 16, 2015, 11:27:04 pm »
[Edit 18 May: After I wrote this OP, I have continued to work on it toward an improved approach. I hope to write that up also in the next few days.]

Here is a yield approach that could be applied to just about any type of bitAsset, so is largely independent of the exact form bitAssets ultimately take. This yield mechanism could also be added at any time, if it were desired to test with a no-yield option first. However despite the added complexity, I would prefer to incorporate yield from the outset to better manage supply and demand.

Basic Approach

- all shorts must specify and maintain a maximum interest rate they are willing to pay to longs (the "max rates")
- at any time, there is a single floating rate paid from all shorts to all longs ("the market rate")
- by construction the market rate is always equal to (or lower - see below) than the lowest of the max rates
- forced short covers other than margin calls, aimed at mitigating bitAsset discounts, are prioritised from lowest max rate to highest max rate
- shorts are able to change their max rates at any time, and new shorts created, subject to not being less than a rate determined by the market rate

The result of this is that bitAsset discounts are always removed through a combination of reduced supply and higher rates. This moderates supply changes compared to relying on the supply adjustment alone.

Managing yield fluctuation

Ultimately, if markets were very liquid, and bond market yields were available to provide a comparison to the (zero-term) market rate, then yield fluctuation would be fairly smooth. In immature markets however, the rate that perfectly calibrates supply and demand is likely to gap up or down and add to yield uncertainty. It may be best initially to find a happy medium even if it takes longer to reach parity, as long as that parity is still assured. These restrictions could be gradually loosened over time as markets get more liquid:

- set a daily lower bound and upper bound on the market rate, being the prior days' rate +/- X% (for example 0.25%, or higher rate to be determined)
- when forced short covers occur, the market rate would effectively rise X% per day until reaching the lowest max rate
- changes to max rates, and max rates on new shorts, cannot be less than the lower bound
- its also possible to lower the market rate X% per day while there is a bitAsset premium (though this may not be necessary if shorts follow down the lower bound)
- the X% limit could be expanded if the bitAsset is a long way from the peg

Dealing with Negative Yields (Payments Longs to Shorts)


In situations where external interest rates in the underlying currency are very low, and supply falls short of demand (both of which we have currently for USD), the natural market rate is likely to be negative. A negative rate would then be necessary to remove bitAsset premiums and guarantee a return to parity. Assuming we wished to incorporate negative yields (an open question), below are a couple of possible ways to deal with this, though developer input would be required on what's feasible:

1. Use a variable transaction fee on bitAsset movements that reflects accrued negative interest paid to the shorts

2. Establish a DEPOSIT token that represents a varying number of bitAsset units as interest accrues in either direction, and create a checking facility denominated in bitAssets that settles as a transfer of DEPOSIT tokens between the parties

Integration with the Bond Market

BitAssets with a yield can be thought of as units being held on deposit. When they are switched to the bond market, they give up the at-call yield in return for the yield offered on the bonds.

243
General Discussion / Re: BitAssets 2.0 (formally 3.0)
« on: May 16, 2015, 10:22:21 pm »
In my view, assuming we can achieve it, the ultimate control mechanism is a yield that works both ways. When demand is too high relative to supply, and bitUSD is priced at a premium, yield falls (also allowed to become negative if necessary), encouraging sellers (and new shorts) until parity is returned. Then you solve all the problems mentioned -

- consumers, merchants, market-makers and shorts can all have confidence that parity is the equilibrium level, enforced by changes in yield either way
- there is no need for scheduled black swans or anything similar

244
General Discussion / Re: Margin calls on shorts - are they full or partial?
« on: May 15, 2015, 10:33:15 pm »
Quote from: pc on May 15, 2015, 04:03:08 pm
See also https://github.com/BitShares/bitshares/issues/1535
That's it! Great to see it on the agenda. I can foresee this being useful if shorts eventually get more flexibility in what they are able to use their bitAsset "loans" to fund...

245
General Discussion / Re: BitAsset 2.0 Requirements & Implied Design
« on: May 15, 2015, 09:14:23 am »
Quote from: merivercap on May 15, 2015, 06:41:16 am
It seems to be better than unlimited forced-settlement, but it seems over-complicated and I don't have as much fear as you in regards to the peg.  Did you check out: https://www.transwiser.com/?  That's the only CNY:bitCNY market I know and it's pretty much 1:1. 
Is there an English version? Unfortunately I can't read Chinese. But it seems to be a gateway rather than an exchange. So its not a free market?

246
General Discussion / Re: BitAsset 2.0 Requirements & Implied Design
« on: May 15, 2015, 05:58:03 am »
Following my last post, I want to consider the limit case where the only bitUSD market is the internal one, as is the case currently. Then BTS is the only path to USD, and the method resolves as follows:

- On any day when the bitUSD:BTS market falls below the price feed, the inferred value of a bitUSD is less than a USD, and this instantly triggers a block of shorts to be selected for calling
- Calling is implemented like a margin cover, by buying bitUSD with the collateral, but with 24 hour notice to the market
- This could continue on consecutive days until the ask returns back to or over the feed price

So in the limit case, the difference to BTA 2.0 is that:
- supply reduction is automatically enforced rather than triggered by settlement requests from individual longs
- the cover price for shorts is a market price, rather than the feed price, with all participants given 24 hour notice to ensure full information and fair pricing

The advantage of this approach might be that a long can no longer manipulate the BTS price to their benefit, because they compete for the BTS made available through the short covers, including with new shorts. Even if somebody owned 100% of the bitUSD, they can only force out blocks of BTS gradually around the price feed, because whenever it goes over that price the short covers stop.

Shorts are protected from unfair pricing by the 24 hour market notice period and by taking account of market conditions in setting block sizes. Block sizes would err on the lower side, and may require several periods to remove the market discount.

As bitUSD found wider external acceptance, the shortest path to USD would be external markets, and preferably bitUSD:USD markets, and its the discounts in those markets that would trigger the short covers. This would demonstrate a commitment to underpin parity in the broader market.

Still just an early idea though.

[Edit: In fact, the short cover procedure is a lot like the way short expiries work, when they are subject to paying no more than the feed price (100%). The main benefit here is that rather than having forced expiries which are often unnecessary when bitUSD is trading at or above parity, we are only forcing covers when there is a discount, and covering from the shorts that are the least keen to hold their position.]

247
General Discussion / Re: BitAsset 2.0 Requirements & Implied Design
« on: May 15, 2015, 04:39:40 am »
Quote from: merivercap on May 15, 2015, 03:56:20 am
Quote from: starspirit on May 15, 2015, 03:08:55 am
merivercap, xiahui135, others interested - I had an initial go at a structure as an alternative to forced settlement in the internal market. You can see it here. https://bitsharestalk.org/index.php/topic,16352.0.html

I read through what I could, but I might need a summary version.  :P

If the idea is that the market can be fixed externally, that sounds good.  We don't currently have any bitUSD:USD markets to reference.  However we do have a bitCNY:CNY market: https://www.transwiser.com/

I'm not clear on the exact mechanisms of the website, but it would be interesting to know.  The peg seems reasonably tight.

Rather than unlimited forced-settlement, I think daily settlement at the price feed for undercollateralized positions should bring enough volume to tie pricing to an external price feed as you want.   We want to have long term shorters to maintain enough supply for bitUSD holders.  Shorters need predictability.



Summary version:
- On any day when the external bitUSD:USD market falls below 1:1, this instantly triggers a block of shorts to be selected for calling
- Calling is implemented like a margin cover, by buying bitUSD with the collateral, but with 24 hour notice to the market
- This could continue on consecutive days until the ask returns back to or over parity

It does not require a bitUSD:USD market necessarily, but if not, it does require at least one external market where both bitUSD and USD trade against a liquid asset like BTC, so that a bitUSD:USD exchange ratio can be inferred (edit: actually any bitUSD market will suffice as long as there is a path to USD).

I think your idea to not force-settle under any price scenario, and just cover under-collateralized positions (which only depends on BTS), would allow the external bitUSD:USD external rate to swing further from the peg, as there is no mechanism to force it back toward parity.


248
General Discussion / Re: BitAsset 2.0 Requirements & Implied Design
« on: May 15, 2015, 03:08:55 am »
merivercap, xiahui135, others interested - I had an initial go at a structure as an alternative to forced settlement in the internal market. You can see it here. https://bitsharestalk.org/index.php/topic,16352.0.html

249
General Discussion / Idea: Settlement forced from the external market rather than the internal market
« on: May 15, 2015, 02:59:45 am »
For consideration, this is a possible alternative to forced settlement in the internal bitUSD:BTS market that could offer the following benefits if it stacks up:

- no internal forced settlement and no opportunity for BTS market manipulation
- supply reduction is triggered directly by discounts in the primary bitUSD:USD market
- forced short covers get a demonstrably fair market price
- it could facilitate a natural buy wall in the bitUSD:USD market (and sell wall, if use yield)
- the voluntary act of self-creation could be the only action that requires a feed from the BTS market

I haven't thought long down this track, so it could fall down somewhere...please attack!

Basic concept

1. No forced settlement of longs, and no constraints on buyers, sellers or shorts in the internal bitUSD:BTS market

2. On any day the ask in the external bitUSD:USD market falls below 1:1, this instantly triggers a block of shorts to be selected for calling (this could be done on yield or collateral). The size of the block could conceivably depend on the level of discount, size of ask orders etc.

3. When such a block of shorts is selected, 24 hour notice is put to the market that a sale of collateral for bitUSD will occur. This gives ample notice for shorts, bargain-hunters, arbitragers etc to prepare for the sale, to get the most effective settlement terms for selected shorts. Shorts in the block could cover before that if desired.

4. At the 24 hour mark, any still uncovered shorts in the block have their collateral sold automatically for bitUSD to cancel their short

5. Each day, as long as the ask in the bitUSD:USD market is below parity, further short blocks will be covered until the ask returns back to or over parity. This guarantees "settlement" and a return to parity as long as there are some willing holders.

How settlement is enabled

I've been a big proponent of the need for longs to be able to force settlement in order to underpin valuation of bitUSD, and thereby minimise discounts to the peg. But maybe this doesn't need to happen in the internal market, where one needs to carefully design some rules so that BTS prices cannot be easily manipulated to force more favourable settlement terms.

The method here effectively achieves the same result, but bitUSD owners can effectively force settlement from the bitUSD:USD market rather than the bitUSD:BTS market, with other parties incentivised to take the bitUSD for USD and offset in the internal market. The way in which this would occur is as follows:

- if there were a lot of demand for BTS, some parties might buy the external bitUSD to sell for BTS in the internal market, especially if they could get a sizeable BTS stake at a lower premium than buying in external markets. Competition among stake-buyers puts an upper bound on the settlement price at which they are willing to buy the BTS from the margined shorts

- if there is weak demand for the BTS, arbitragers will buy the external bitUSD to sell for a premium in the internal market. They need a premium to cover the spread in selling their BTS proceeds in the weak external market. Competition among arbitragers puts a lower bound on the settlement price at which BTS is bought from the called shorts

- market-makers in the bitUSD:USD market may buy the bitUSD at or below parity to sell at a premium later, given ongoing supply reduction is guaranteed to return to bitUSD to at least parity

The settlement price for called shorts, between these bounds, reflects all market and liquidity conditions and is therefore a "fair" free-market price. I don't think it can be easily manipulated, but interested if others can see ways it can.

Some extra ideas

Introducing a yield between longs and shorts

A yield mechanism would allow for situations where the market consensus is that a bitUSD is worth more or less than parity with the USD. As a result, it would allow more moderate swings in supply to balance the demand. It would also limit premiums to the peg. I've discussed this elsewhere, though it should be noted this is not in line with BM's current proposals. I have a process to incorporate yield into the above approach but won't detail it here unless interested.

Creating and cancelling shorts

Given that pegging is enforced in the bitUSD:USD market, especially with a yield mechanism to limit premiums, market-making becomes a more profitable endeavour in that market. BTS owners can easily earn an income from this as follows:
- self-create bitUSD
- use the bitUSD to sell at a premium for USD, and buy at a discount with USD, in the bitUSD:USD market, earning the spread
- when user wishes to stop this activity, buy the bitUSD and self-cancel

It's also possible for shorts to always be self-created and self-cancelled, and never be created in the internal bitUSD:BTS market. This could vastly simplify the internal market. But it would require shorts to self-create with $2 of BTS, then sell their $1 bitUSD for BTS in order to get exposed to $3 of BTS. This is 3:2 leverage, and would not be as much leverage as the 2:1 provided if they can short directly in the internal bitUSD:BTS market.

250
General Discussion / Re: BitAsset 2.0 Requirements & Implied Design
« on: May 15, 2015, 02:33:18 am »
To make a wise choice on the yield vs no-yield option for BTA 2.0, it is important to consider what will give bitShares the best ongoing advantage in the pegged currency space.

The key is to remember that we are always in competition with others. It's not a good enough argument to say yield can be provided in the bond market, so no yield is required in BTA 2.0, if it turns out that a competitor can create a yield-earning equivalent and also have a bond market, a combination that may be superior to our own.

Imagine if a bank, in the face of competition against other banks that were offering interest on their at-call accounts, took the stance that they did not need to offer interest because customers could get interest in their term deposits instead. How do you think that bank might fare?

It is valid to compare the yield and no-yield options, and decide on one as being a better product than the other. Because then we are saying that if a competitor develops the yield option, then we are comfortable we still have a better product. We have to imagine that anything we could build could be built by others, and we need to believe we have the best product available.

251
General Discussion / Re: Idea: Flexible Collateral
« on: May 14, 2015, 11:53:13 pm »
By the way, the first and most useful step toward this would be to allow bitUSD shorts to switch their collateral between bitUSD and BTS. A switch from BTS to bitUSD would allow shorts to then cancel their position from internal collateral rather than having to use other funds held to buy the bitUSD.

252
General Discussion / Re: Margin calls on shorts - are they full or partial?
« on: May 14, 2015, 10:06:50 pm »
Quote from: pc on May 14, 2015, 12:26:20 pm
Quote from: starspirit on May 14, 2015, 10:38:50 am

But I suppose it would be feasible for the call price to be changed to a new level, reflecting the new collateral ratio, right?

I'm only noting that CFD providers and brokers don't usually close all positions unless required. They usually just exercise the discretion to close enough positions so that the collateral buffer is adequate again relative to the remaining exposure.

This just made me wonder what shorts would prefer, given on the surface at least it would not seem a difficult change (effectively just modifying the call amount and the call price).

I know there's bigger fish to fry right now, it was just something I noticed in thinking about my CFD account.

I suppose that since margin calls exist to protect the longs from a black swan, it would make sense to distribute available funds over all margin-called positions in such a way that the principal/collateral ratio is pushed to the same level for all such positions, and to adjust the call price accordingly (to simplify bookkeeping, mostly).

OTOH this is a somewhat complicated optimization that would be useless in a healthy market.
You could try some sort of global optimisation like this, but I think its enough to do it at the individual level. That is, cover 50% of the short, re-establishing 2:1 collateral coverage, and reset the call price. This still provides black swan protection, while minimising the size of margin calls on shorts.

Example:
A $100 short gets set at a price of 100, offering 100 BTS of collateral (100%). The call price is 133, representing a 25% fall in the value of total collateral from 200% ($200) to 150% ($150).
If the call price of 133 is hit, it is sufficient to sell $50 of the BTS collateral to cover $50 of the short obligation. The end result will be that there is $100 of collateral remaining to cover $50 of obligation, giving 2:1 coverage again.
The call price could then be reset to 177 before another call is required on this short.

253
General Discussion / Re: BitShares needs a professional SEO/SEP Delegate... oh, and WHALES.
« on: May 14, 2015, 11:39:09 am »
What BitShares also needs is a crystal clear business plan (product description, market potential, profitability expectation) and strategy (development timetable, marketing plan), even if that means bringing in some good advice. This is required to attract real funds for development and fast-track growth, rather than operating on time charity.

254
General Discussion / Re: Margin calls on shorts - are they full or partial?
« on: May 14, 2015, 10:38:50 am »
Quote from: pc on May 14, 2015, 07:43:45 am
The total position is covered, if possible.

The call price remains unchanged, i. e. if the position is partially covered it will stay in margin called state.

But I suppose it would be feasible for the call price to be changed to a new level, reflecting the new collateral ratio, right?

I'm only noting that CFD providers and brokers don't usually close all positions unless required. They usually just exercise the discretion to close enough positions so that the collateral buffer is adequate again relative to the remaining exposure.

This just made me wonder what shorts would prefer, given on the surface at least it would not seem a difficult change (effectively just modifying the call amount and the call price).

I know there's bigger fish to fry right now, it was just something I noticed in thinking about my CFD account.

255
General Discussion / Re: How do changes to Bitshares design get accepted in the network?
« on: May 14, 2015, 03:42:19 am »
thanks xeroc, monsterer, pc.

From what I understand then, in practice developers decide the version, after community consultation, and promote that version for switching by delegates and users. If there ever were disagreement, delegates (or more specifically, block-producers) could support whichever version they preferred, owners could vote for whichever block-producers will support their network, and users could choose to use whichever network they wish.

Users pay the fees that incentivise everyone. Owners control who is allowed to earn those fees. Block-producers support the network infrastructure. So competitive development teams ultimately need to convince all three of these stakeholder groups that their version is best to ensure a successful transition?

Pages: 1 ... 10 11 12 13 14 15 16 [17] 18 19 20 21 22 23 24 ... 64

Page created in 0.19 seconds with 22 queries.