Interest on BitUSD - A Proposal for Review

[tonyk]

If you talk about the interest proposal here, the (rather raw) calculations are not taking into an account this 'shorting pressure discount' so we must be safe in that regard.

I felt like most of the money that was expected to be paid in "interest" was due to taking advantage of this spread.  That's why I feel it was relevant.  As far as paying out the collected bitUSD as "interest"...  I hadn't considered this but it seems overly complicated and does not appeal to me.

I just realize - isn't that spread producing BTSX and not bitUSD for the system?
How and when are those converted in bitUSD?

[pendragon3]

Yes, simple is a must. If savers cannot transparently see past average interest rates and how much they have earned over the past month, year, and so on, they won't be very interested. After all, it's the savers this would appeal to, not the traders. But isn't the vision for the network that the peg is mainly supported through trading and "arbitrage" activity, not because of a coincidental balance between saver and shorting demand? So, this might not boost BitUSD demand as much in the short term as one might hope. In the longer term, traders will be doing most of the work supporting the peg within tight ranges, so here again the savers are less important in maintaining the peg.

I think if an interest program is implemented, it should be in the main BitUSD, not a separate bitasset. A bitUSD5 would kill off demand for the main BitUSD asset, making it even harder to maintain the peg.

I also worry that even paying interest on the main BitUSD could change people's perception somehow. This changes the entire dynamic. With an interest program, BitUSD would seem less like a twin to the USD and make it harder to eventually maintain a really tight peg.

Finally, consider what would happen if, someday in the future, inflation goes high (which forces nominal interest rates to rise--they will need to be above inflation in equilibrium). Would interest on BitUSD still be competitive if traditional interest rates rise to 10%, 15%, 20%?

Basically what I'm saying is that paying interest seems like only an indirect, blunt way to artificially stimulate demand for BitUSD. It could work in the short term, but it certainly has some drawbacks and could have a weaker-than-expected effect on demand BitUSD since it appeals to savers, not traders. It also could have a number of unintended consequences and weaken the psychological basis for a tight peg.

[Agent86]

Agent86... like always you are a very passionate debater who frequently convinces me of things.   I would like to explain why I don't like your proposal at this point in time.

I have had white board designs for how to reenforce the peg via 2 markets.   1 market sets the "premium" shorts must pay.. ie: the amount they must bid over the ask.  With this 2-market system the short interest would be curtailed by those speculating on the premium market.   The result would be a peg and automatic "interest" rate setting for BitUSD.

BitUSD would always be valued at $1 because the interest rate/risk premium would adjust to compensate for supply/demand/risk assessments of the market.

That said we have a MAJOR bootstrapping problem with all markets.  They require a network effect and high liquidity to function.  It is hard enough to find speculators on the BitUSD vs BTSX market let alone a "meta-market".   A system that requires 2 markets to function is way too difficult to bootstrap at the same time. 

A bitbond market would have similar problems.  A bond market is even more difficult to bootstrap than a pure prediction market, especially if there are different maturity dates, interest rates, that make them non-fungible.   

So while I do not want to dismiss your idea, I just do not think it is viable at this stage in the development of BitMarkets.

Thanks BM.  I don't honestly think the idea is a high immediate priority either.  I just think it is the right direction for the eventual evolution of things.  What are your thoughts that matching shorts below feed with orders above the feed stops shorts from quickly covering when bitUSD is undervalued to maintain the peg.  (matching based on collateral provides better incentives)?
discussed here: https://bitsharestalk.org/index.php?topic=8415.0

[bytemaster]

In the interest of simplification I would like to add the following note:

1) cashing out interest once per day or once per year should yield the same return minus 364 transaction fees.   This will keep all balances "fungible" and reduce the cognitive load on deciding when to handle it.

2) no interest should be paid in amounts less than 1 BTSX.  This will be the "small balance fee".   If your balance is so small that is cannot earn at least 1 BTSX in interest then too bad, so sad.   This will help prevent dusting and situations where the transaction fee to claim the interest is greater than the interest itself. 

[bytemaster]

Agent86... like always you are a very passionate debater who frequently convinces me of things.   I would like to explain why I don't like your proposal at this point in time.

I have had white board designs for how to reenforce the peg via 2 markets.   1 market sets the "premium" shorts must pay.. ie: the amount they must bid over the ask.  With this 2-market system the short interest would be curtailed by those speculating on the premium market.   The result would be a peg and automatic "interest" rate setting for BitUSD.

BitUSD would always be valued at $1 because the interest rate/risk premium would adjust to compensate for supply/demand/risk assessments of the market.

That said we have a MAJOR bootstrapping problem with all markets.  They require a network effect and high liquidity to function.  It is hard enough to find speculators on the BitUSD vs BTSX market let alone a "meta-market".   A system that requires 2 markets to function is way too difficult to bootstrap at the same time. 

A bitbond market would have similar problems.  A bond market is even more difficult to bootstrap than a pure prediction market, especially if there are different maturity dates, interest rates, that make them non-fungible.   

So while I do not want to dismiss your idea, I just do not think it is viable at this stage in the development of BitMarkets.

[Agent86]

I have a concern with the collateral being enough to cover the debt by the future date... there is a responsibility on the bond holders to keep an eye on the price changes of BTSX/BitUSD to make sure their bond has enough collateral value to pay the debt at the future date. In fact, in theory if the collateral isn't large enough to cover a drop in the price of BTSX from when the bond was purchased until when it pays the debt, then the effective interest rate can decrease and even become negative.

The part where I see the biggest challenge is such collateralization working on a blockchain. I can not figure out a way (neither  have I read somebody suggesting working solution) where both conditions are met -repayment of the bond (debt is) guaranteed and at the same time incentive for the issuer of such bond is still present. In other words if the whole amount received through the bond sell is kept as collateral + additional collateral for the interest is needed, where is the incentive for the issuer? Anything less than that collateral leaves the bond buyer exposed to the risk of not receiving his money back.

[edit] It will take a truly independent bond market, where special market participants, will act as bond issuer and will collateralize a series of bonds (with deferent expiration dates), for this to truly work.

A bond that pays 1000 bitUSD on a particular date in the future (face value is $1000) must initially be backed by 2x that value in BTSX by the issuer.  The issuer is responsible to maintain the margin just like a bitUSD short; if the collateral value drops to 1.5x the issuer is also subject to a margin call and bitUSD is bought off the market to cover and bitUSD is deposited into the collateral account.  If the issuer doesn't fund the bond collateral with bitUSD by the maturity date (payout date) they are also subject to a margin call on that date.

I think the chance that the margin call can't find a bitUSD seller at the price needed is VERY small.  In that event, the bond holder gets the BTSX shares instead.  (just like if a company can't pay a bond, then the bond holders can take the company)

If the issuer sells this bond for $950. They can then buy more BTSX to get more exposure to the upside of BTSX.  Or do something else with the money raised while maintaining their exposure to BTSX in their locked up collateral.

[Riverhead]

I think we need to keep this as simple as possible. There is already so much for new investors to get their head around. A complicated rewards program just adds to their confusion.

[bytemaster]

Try 32 bit windows

[okidoki]

What about the original idea of shorters paying interest to the longs - is that lo longer workable? I always thought that was nice and simple because you could charge a fixed amount from shorters relative to how long the short position was open for.

Yeah... I thought this as well... why not fix 5% interest from shorters?
What also could be done is a combination of both.... so the sales pitch would be 5% fix + x% variable...

For now I would just love to be able to install 0.4.12 on windows 8.1 or linux. Windows returns an error that it cannot be run and linux does not compile...

[Empirical1.1]

I'm not totally against daily interest myself.

However I think interest will have a limited incentive effect for traders, speculators and weak hands.

Traders will come when there's a profitable imbalance, speculators are looking to make short term gains and weak hands will dump their BitAssets regardless of interest.

However who would be incentivised by interest are people looking to genuinely store value in BitAssets. Those people are imo are who BTSX is currently lacking (No left side of the order book) They are also the ones who would be most exposed to a BTSX failure (risk) which would not be compensated for.

As I said in another thread if the chance of total BTSX failure is 2% this year. BitAssets are only worth 0.98 to them. Plus without utility genuine BitAsset holders have to take into account the conversion cost to acquire the real world counterpart (Or at least the fiat equivalent) their BitAsset is mirroring.

By translating shorting demand into fees for 'rewards' and also not giving those rewards to speculators, traders and weak hands then you create a very attractive pot for people genuinely looking to hold value in BitAssets.

The result is a correction of the demand imbalance & a very healthy left side of the order book and BitAssets that are worth more than 1-1 to a person genuinely looking to put some of their portfolio into BitAssets.

This increase in demand at or above 1-1 may even eventually nullify the situation of shorts needing to compete much to find buyers. (At which point the interest will be going down & the system will be self regulating.)

Over time as BTSX is viewed as safe, reliable & decentralised. When there are gateways &/or utility plus the many other advantages of BitAssets will make them attractive to genuine holders @ 1-1 or even above without interest.

[liondani]

would it be not fair and motivating that a percentage of taken fees go back as a reward to high volume traders?  "The more you trade the more you win "

[jakub]

I've gone through the whole thread and here are my thoughts:

1. Dividends are a very important enhancement for Bitshares X and definitely worth trying to implement ASAP.

2. The word "reward" has very bad connotations for me in this context: it sounds as if someone is giving me a reward in order to hide some ugly downsides. Interest and dividend are much better and I'd stick with one of them unless there actually occurs a legal challenge - then we can choose to fight it or just change the name.

3. Bonds (agent86) and CD (tonyk) ideas might make sense in the future but I think it's much too early to consider implementing them now. Bitshares X are already extremely complex and we need to give people time to get used to such complexity. Only after that there will be space to think about building more financial tools on top of what we have.

4. The initial BM's idea of only paying interest to long term bitAssets holders seems to be wrong as it encourages hoarding and thus hurts liquidity. I think the interest should be linear and accrued on a daily basis.

5. It looks like a big part of the discussion is a dilemma of how much of the trading fees can be transferred to bitAssets holders. Surely there must be some sweet spot: we'd like to tax the trading a bit so that we can make the bitAssets more attractive but we don't want to tax them too much because that will kill liquidity. The sweet spot needs to be flexible over time so this idea seems to be addressing it well:

I think it would be great if the shareholders had ways to define the amounts to be paid to the delegates, amounts to be paid to other specified accounts (workers), the amount of BitAssets to store as insurance for each BitAsset (as a percentage of the total BitAssets issued), and the interest rate to be paid for each BitAsset. Then, in addition to collected transaction fees, the DAC would automatically adjust what percentage of fees were collected as BTSX vs BitAsset from the bid-ask overlap to collect enough BTSX and BitAssets as fees so that the DAC could afford to do following (in order of priority): pay delegates their income; pay workers their mandatory income; grow insurance fund of BitAsset if necessary; pay rewards at prescribed interest rate to BitAsset holders; pay workers their discretionary income if any; pay leftover as dividends to BTSX holders.

7. Also this looks like a good concept to me:

A short position has collateral from two sources: one is the BTSX paid by the BitAsset buyer (I call this BTSX the reserve); the other is the BTSX paid by the short seller (I call this BTSX the margin). The collateral = reserve + margin. The current rule is that the margin needs to be at least equal to the reserve initially (200% backed initially). At some point the short seller covers or the DAC does a margin call. If the price of BTSX goes down with respect to the BitAsset, then some of the margin is needed to cover the BitAsset debt when covering. Whether the short seller covers at a loss or the DAC does a margin call, the mechanism would work the same way it currently does. However, if the short seller covers at a profit, things would work a little bit differently compared to today. If the short position is covered at a profit, that means that only a portion of the reserve is needed to pay the debt and the remainder is profit. This remainder should be split according to some percentage specified when the short position was created and one part is paid as profit to the short seller and the other part is collected by the DAC as a fee (in reality the cover would require excess BitAssets to not only cover the debt but also the value of the fee according to the current price so that the DAC essentially trades the BTSX fee for the BitAsset fee). These fees are collected to pay variable interest to the BitAsset holders in a way similar to what I described earlier.

Since shorts would be limited to the peg price, short sellers need some other way of competing with each other to get their shorts matched. The market rule would be that in the case of a tie in price, the short position with the highest percentage of profit burned as fees wins. The percentage of profit burned should adjust according to market dynamics to balance the BitAsset long and short demand since as the percentage of profit burned goes up the short sellers lose their incentive to short but BitAsset buyers have an increased incentive to hold the BitAsset due to the higher expected interest rates. If the price of the BitAsset in terms of BTSX drops even a little from the peg that it was shorted at, the short seller still has an incentive in the short-term to cover to get their percentage of the small profit. Agent86, do you see any liquidity problems with this method? Anyone see any issues with this proposal?

Indeed it sounds like Capital Gains Tax but it shouldn't be discarded just because nobody likes taxes. Shorts who've made a profit are more in the position to share it with bitAssets holders. This concept looks attractive to me because I feel it might hurt liquidity less than the current bid-ask overlap fee.

[xeroc]

A quick question:

delegates are payed by the transactions fees ... currently paid in btsx.
once we can pay transaction fees in bitUSD or any other bitAsset are those fees still going to the delegates? or are they going to the fund?
will the delegates then earn the fees as bitUSD or will they be exchange automatically for btsx?
can delegates expect to see a list of different fee 'currencies'/'shares' collected?

[santaclause102]

I see the need for interest on BitUSD but why offer more interest if BitUSD is held longer without selling (not only totally but proportionally)?

Apparently the non-linear interest is already history. BM mentioned that somewhere in the mumble.
https://bitsharestalk.org/index.php?topic=8423.0;topicseen
https://soundcloud.com/beyond-bitcoin-hangouts/bitshares-community-and-developer-hangout-9-5-2014
Sorry, I forgot the timestamp

Ok
What is the status on: How is it decided who's trade is executed if there is a demand/supply imbalance between bitUSD longs / shorts? -> Amount of collateral if there is more demand to short bitUSD?
In case more ppl want to go long bitUSD / short btsx then there would be no limit upwards regarding at which prices bitUSD can be shorted?

[Markus]

I see the need for interest on BitUSD but why offer more interest if BitUSD is held longer without selling (not only totally but proportionally)?

Apparently the non-linear interest is already history. BM mentioned that somewhere in the mumble.
https://bitsharestalk.org/index.php?topic=8423.0;topicseen
https://soundcloud.com/beyond-bitcoin-hangouts/bitshares-community-and-developer-hangout-9-5-2014
Sorry, I forgot the timestamp