Interest on BitUSD - A Proposal for Review

[arhag]

Maybe I'm the only one who disagrees with all this, but I really don't like it.

I don't like making money off the spread by matching shorts that are below the feed with bids that are above it.  I think this reduces liquidity in the market and hurts the peg.

I would much rather see shorts prioritized by amount of collateral.

How I would fine tune the market:
I would create a small delay from the time that new feeds are posted and the time they are active, maybe a few minutes.  That way traders can know the exact median feed price a little before it takes effect so all traders are on even ground.  Aggressive shorts can then enter their orders at the exact feed price and compete on who posts the most collateral to be first in line to be matched.  This allows traders to profit by shorting at $1 and quickly buying back at $.99 when they can.

If you do it the way proposed in this thread then you artificially increase transaction costs which hurts liquidity and only attract long term shorts who will be reluctant to cover (they can't turn around and cover at $.99 to make a profit because they may have only got $.95 or less worth of BTSX from their short).

Aside from fact that  this is not the right thread that you disagree with. I also I have my doubts that the actual market will play as described.- i.e. demand for shorting resulting in offering up to 10% discount. I would expect shorts entering very near the 'peg price' and waiting the order to be filled (even if it means days).

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 think I agree with your idea that shorts should be at the peg price and be prioritized by collateral instead. This makes me wonder how much money would actually be generated in BitUSD fees. It still could be a considerable amount from the overlap of bid and asks. The value in that overlap can be extracted as BTSX or the BitAsset being traded (or some combination of each totaling 100% of the value of the overlap). By extracting the overlap as BTSX fees, the DAC provides income for delegates (and, maybe in the future, workers) and dividends to BTSX shareholders. By extracting the overlap as BitAsset fees, the DAC provides an insurance fund for black swan events on that BitAsset and could also provide rewards to the BitAsset holders.

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.

[bytemaster]

What effect (if any) do you guys think the 5% interest will do to the price of Bitassets? Take something like bitUSD which already inflates via the US government by about 2% per year. Then we add 5% on top of that. How will the 7% annual inflation rate affect the price of Bitshares- if it will at all?

I know this is about creating a demand for Bitassets and I think it will, but are there any long term side effects regarding to the bitasset market peg in making bitUSD annual inflation at 7% when the actual inflation rate is closer to 2%???

Bit assets do not inflate. Nor are they debased.  Usd inflation is not relevant. 

[CoinHoarder]

What effect (if any) do you guys think the 5% interest will do to the price of Bitassets? Take something like bitUSD which already inflates via the US government by about 2% per year. Then we add 5% on top of that. How will the 7% annual inflation rate affect the price of Bitshares- if it will at all?

I know this is about creating a demand for Bitassets and I think it will, but are there any long term side effects regarding to the bitasset market peg in making bitUSD annual inflation at 7% when the actual inflation rate is closer to 2%???

[arhag]

Interest bearing bonds require a separate market and implementation from the core BitUSD market.  A BTSX holder can sell a collateralized promise to pay a certain amount of bitUSD at a certain date in the future.  There can then be a "bond market" for these promissory notes.  The present day value of these future promises to pay BitUSD will determine short term and long term interest rates.

Seems like a fine addition to me. Although, I have a concern with the collateral being enough to cover the debt by the future date. A BitUSD holder relies on there being enough traders to sell their BitUSD as the price changes to cover the shorts with dangerously low collateral. Then new shorts can enter the system at the newer prices. Since BitUSD is fungible, the BitUSD holder doesn't have to care about the price changes and can be assured that there will be enough collateral available (ignoring black swan events) when he wants to sell his BitUSD. But these promissory notes aren't fungible are they? Someone has to hold on to a bond that pays a specific amount at a specific future date and has a specific amount of BTSX as collateral backing it. They can always sell it in the bond market, but 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.

I think for many BitUSD holders, they would rather not worry about all of this stuff. They would just like to get some variable interest on their BitUSD balance, just like they would in their bank savings account. Using the BitUSD fees to pay this interest makes things much easier for these users. If they are unsatisfied with the rates they are getting, then they can look into the bond market.

[Agent86]

Please read my proposed bond/CD idea in this thread and through your opinion.
Thanks.

Ok, If I understand that you are talking about using a CD to smooth out the extra bitUSD distributions for people who want something more reliable.  I don't really like distributing the bitUSD fees to bitUSD holders as described so offering CDs on top of that I don't think helps it.  I like using bonds generally though.

Here is an idea. The proposal stays as revenue sharing plan/dividends.

-At the same time everybody can offer true interest bearing CDs. (fix rate/fix time/known penalty for early withdrawal).

-The amount received in bitUSD is kept as collateral, plus additional amount for the interest promised;

 => the person saving the bitUSD is absolutely secured the amount + interest;
and The person offering this deposit benefits is this profit sharing produces more than the promised interest?

CD seems as US term only - here is the link: http://www.investopedia.com/terms/c/certificateofdeposit.asp

Thoughts?

[liondani]

what about replace "interest " with "Bitshares"...  bitshares per year ...  5% value of bitUSD but you get it in the equivalent of  BTSX (?)...

[tonyk]

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.

To tell you the truth, I am in total shock by this statement.
 Do you have time to explain further?

Me too. Agent86, does bytemaster's implementation not appeal to you or the idea of paying interest to BitAssets? I thought the appeal of interest on BitUSD would be obvious; it becomes a super easy way to market BitShares X to the average Joe: you get higher interest rates keeping your BitUSD in BitShares X than keeping your USD in your regular bank.

I'm in favor of interest, I just think the right way to do it is make a bond market…  I proposed the idea here:
https://bitsharestalk.org/index.php?topic=7816.0
I think this has much more potential to be useful and popular and drive people into bitAssets.

Please read my proposed bond/CD idea in this thread and through your opinion.
Thanks.

[x.ebit]

good

[Agent86]

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.

To tell you the truth, I am in total shock by this statement.
 Do you have time to explain further?

Me too. Agent86, does bytemaster's implementation not appeal to you or the idea of paying interest to BitAssets? I thought the appeal of interest on BitUSD would be obvious; it becomes a super easy way to market BitShares X to the average Joe: you get higher interest rates keeping your BitUSD in BitShares X than keeping your USD in your regular bank.

I'm in favor of interest, I just think the right way to do it is make a bond market…  I proposed the idea here:
https://bitsharestalk.org/index.php?topic=7816.0
I think this has much more potential to be useful and popular and drive people into bitAssets.

[Empirical1.1]

Why require a minimum of a month?

I'm not anywhere near the same maths or trading level as you guys but these are my thoughts.

- Instead of selling their BitAssets for BTSX in a bubble or in a panic during a technical problem. BitAsset holders are now incentivised to hold and not create as much volatility.

- Not sure about this one, but doesn't giving people an incentive to hold, take some BitUSD out of general circulation so there will be less available for shorts looking to cover for example thereby increase demand and making BitUSD average closer to the peg than the few % below I hear it averages now?

- Doesn't this proposal reward savers over traders?
Just like keeping your funds for any length of time on a centralised exchange is risky, till BTSX has proved itself so too is holding BitAssets. (BitAsset savers take on a lot of the risks of BTSX with few of the benefits) Maximising their reward attracts genuine saving not just trading demand for BitAssets.

Don't we want more people trading not hording? More trading = closer to peg value over time especially when the price is moving. IF people need to hold onto their bit assets it discourages trading which decreases liquidity. There should be no minimum time. Also more trading = more fees, which is also good.   

Unless I'm missing something?

You may be right. I'm not hot with these things. I'm sure the main guys will be able to clue us in here.

I would think traders would come in whenever there is a profit to be made from an imbalance anyway, they don't need the interest as an incentive.

I also think there is a risk to holding BitAssets long term as well as a conversion cost and lack of utility short term for someone looking to move some of their BTC position over to BitBTC for example so those people are the ones we want to maximise the rewards for.

Also if trading is anything like poker. The traders want as many of the average Joe savers in the game because that's who they get their edge from.

[cygnify]

Why require a minimum of a month?

I'm not anywhere near the same maths or trading level as you guys but these are my thoughts.

- Instead of selling their BitAssets for BTSX in a bubble or in a panic during a technical problem. BitAsset holders are now incentivised to hold and not create as much volatility.

- Not sure about this one, but doesn't giving people an incentive to hold, take some BitUSD out of general circulation so there will be less available for shorts looking to cover for example thereby increase demand and making BitUSD average closer to the peg than the few % below I hear it averages now?

- Doesn't this proposal reward savers over traders?
Just like keeping your funds for any length of time on a centralised exchange is risky, till BTSX has proved itself so too is holding BitAssets. (BitAsset savers take on a lot of the risks of BTSX with few of the benefits) Maximising their reward attracts genuine saving not just trading demand for BitAssets.

Don't we want more people trading not hording? More trading = closer to peg value over time especially when the price is moving. IF people need to hold onto their bit assets it discourages trading which decreases liquidity. There should be no minimum time. Also more trading = more fees, which is also good.   

Unless I'm missing something?

[arhag]

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.

To tell you the truth, I am in total shock by this statement.
 Do you have time to explain further?

Me too. Agent86, does bytemaster's implementation not appeal to you or the idea of paying interest to BitAssets? I thought the appeal of interest on BitUSD would be obvious; it becomes a super easy way to market BitShares X to the average Joe: you get higher interest rates keeping your BitUSD in BitShares X than keeping your USD in your regular bank.

[Empirical1.1]

Why require a minimum of a month?

I'm not anywhere near the same maths or trading level as you guys but these are my thoughts.

- Instead of selling their BitAssets for BTSX in a bubble or in a panic during a technical problem. BitAsset holders are now incentivised to hold and not create as much volatility.

- Not sure about this one, but doesn't giving people an incentive to hold, take some BitUSD out of general circulation so there will be less available for shorts looking to cover for example thereby increase demand and making BitUSD average closer to the peg than the few % below I hear it averages now?

- Doesn't this proposal reward savers over traders?
Just like keeping your funds for any length of time on a centralised exchange is risky, till BTSX has proved itself so too is holding BitAssets. (BitAsset savers take on a lot of the risks of BTSX with few of the benefits) Maximising their reward attracts genuine saving not just trading demand for BitAssets.

[tonyk]

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.

To tell you the truth, I am in total shock by this statement.
 Do you have time to explain further?

[tonyk]

BitUSD is incredibly useful as a trading instrument but the "sales pitch" for the average Joe is something like this:

1) Buy BitUSD today and you will be able to sell it back for 98% of what you paid for it (due to the spread)
2) Once you have BitUSD you get all of the benefits of Bitcoin, instant transfers, privacy, security.
3) As a merchant the "cost" of accepting BitUSD (at parity) is currently equal to the cost of a credit card (a few percent)

Initially demand from traders will to hedge against BTSX volatility will help the market grow, but what we *REALLY WANT* is for people to store their wealth in the system, not just their trading account.  To do this BitUSD needs to offer a return that is explicit.  IE: it shouldn't be done via the value deviating from the peg.

Right now the network earns USD via the market and via transaction fees (yes transaction fees can be paid in USD).  This income is saved for a "rainy day" which we don't actually expect to happen, but could.  The value as an insurance system could be better used to pay interest to USD holders.   If BitUSD were an interest bearing asset, then the interest would compensate for the risk of black swans in the same way that the insurance fund does today.

So far BitUSD has earned $1800 on $485,000 issuance in just 2 weeks of light trading.  This is about .4% yield every 2 weeks or about 10% per year.   I have no idea how this will play out over time, but I suspect that it will result in a significant yield for BitUSD holders.   The more people hold BitUSD the more BTSX is worth.   So it is in the best interest (No pun intended) of BTSX holders to pay interest on BitUSD from the fees collected.

I would like to implement it as follows:
1) If you hold a balance for less than 1 month, you earn nothing.
2) If you hold a balance for one year or more then you earn (YOURUSD / TOTAL_USD) * COLLECTED_FEES_BALANCE
3) If you hold a balance for less than one year then you earn a pro-rated amount.     (YOURUSD / TOTAL_USD) * COLLECTED_FEES_BALANCE * FRACTION_OF_YEAR_HELD^2

The result is that long-term holders of BitUSD earn much higher yields than short-term holders.  The yield should more than cover the cost of the spread and thus your pitch to the average Joe is much stronger:

1) Buy BitUSD and earn more (much more) on your dollars than you do at your bank.
2) Enjoy all the other benefits of crypto.

This feature would have a greater impact on adoption than just about anything else being discussed. 

Discuss.

Why require a minimum of a month and why have a quadratic dependence on FRACTION_OF_YEAR_HELD? What is the purpose in creating incentives for BitUSD holders to hold BitUSD balances for long periods of time? The fact that they have the money as BitUSD in the first place already adds value to BTSX.

And are the rewards claimed at the time of transaction? Do those rewards get immediately deducted from COLLECTED_FEES_BALANCE when that transaction occurs? How do I know when I have officially locked in the specific nominal rewards (in BitUSD)?

I like the general idea, but I worry that your implementation makes things unnecessarily complicated for the user in figuring out the rewards they will receive. It will probably also force extra complications in the client implementations for the purpose of maximizing the interest the user receives from their balance.

For example, what is the best way to maximize my return on some fixed initial sum of BitUSD? Let's define X = (COLLECTED_FEES_BALANCE / TOTAL_USD). Do I want to hold for 11 months and 29 days and then make the transaction right before the inactivity fee deadline, so that FRACTION_OF_YEAR_HELD^2 is as large as possible? In that case my annual reward is approximately P*(1+X), where P is the principal, or a annual return of (X*100)%. Or do I want to break that up into (12/k) k-month intervals to take advantage of compound interest? In that case my annual reward is P*(1 + X*(k/12)^2)^(12/k) assuming X remains constant over the year throughout the entire compounding process, giving an annual return of (((1 + X*(k/12)^2)^(12/k) - 1)*100)%.  Since X has to be less than 1 (if I understand COLLECTED_FEES_BALANCE and TOTAL_USD correctly), this means holding it for a year would always be better than compounding with k-month intervals for any k. But this was assuming X is fixed. In reality, X will fluctuate as COLLECTED_FEES_BALANCE changes (increasing as more fees are collected and decreasing as people claim the the reward). I would hope reward claims would be randomly distributed so that COLLECTED_FEES_BALANCE only has small deviations around a relatively stable value over time. If it wasn't random however, there would be an advantage to the people who claim first. Let's say nearly all BitUSD holders initiated their balance in September and intended to hold for a year to maximize returns. Over the course of a year, the COLLECTED_FEES_BALANCE grew to a large sum S, and in next September, all the BitUSD holders suddenly claimed nearly 100% of COLLECTED_FEES_BALANCE. But the first guy who owns f1 fraction of TOTAL_USD gets f1*S, the guy with fraction f2 gets f2*S*(1-f1), the guy after that gets f3*S*(1-f1-f2), until the last guy gets a tiny fraction of what he could have gotten had he claimed first (again assuming I understand your proposed implementation and calculation of COLLECTED_FEES_BALANCE correctly).

Why create all of this complication for users? The implementation should not have users receive different returns depending on when they move their balances. I don't see why it should withhold rewards for people who held the BitUSD for less than a month either. Why not take a portion of the fees collected in a given block, put it in a fund, and divide the money in that fund by the total BitUSD held by users (the existing BitUSD not collected by the network) at the time of that block to calculate the interest rate for that block. When a BitUSD balance is moved, the sum of the interest rates for each block on which that balance existed (called the ACCUMULATED_INTEREST_RATE) is used to determine the reward on that that balance (ACCUMULATED_INTEREST_RATE*BALANCE). The rewards on these BitUSD balances act as extra BitUSD (effectively coming from the fund) on the input of the transaction that can be claimed by the outputs. This way all of the money in fund can be considered consumed at every block, and the interest payment on balances is continuous over time and cannot be strategically manipulated by timing when to move the BitUSD.

Seems an idea, worth considering in order to iron out the general proposal, imo.
Additionally the first person to point out where my math got screwed up (Thanks.)