This is great news, and so necessary for our growth.
For context, did BTER provide this capability previously? Or is this something completely new for BitShares?

What are your concerns about using feed to adjust premium over the peg?

I don't understand how you intend for it to work. My impression, please correct if its wrong, is that as the premium grows, you drop the price feed below its "fair" price (like a fee on settlements) and as the premium contracts, you lift it back toward the "fair" price (lower the "settlement fee"). Is that correct?

If it is, then I have a concern with it. For any static settlement fee, I agree the average trading price would equilibrate at a different level. The possible problem is that when it is dynamic the market may not price it in because it never comes into force, as here:

...I can't see why the market would price this in. If the market is at a premium, then the settlement mechanism is not needed. By the time the settlement mechanism is needed and used, the price must be at a discount where there is no longer a settlement fee. So there is no need for any settlement fee to ever be paid, nor any buyer to price in the risk of a higher fee, no matter how high the premium. So how would this affect supply or demand?

Quote from: lil_jay890 on May 27, 2015, 01:51:56 am

You do realize that most of the benefits will be the same for any other pegged crypto? If they don't charge a premium then it is game over. for us.

What many of you are missing is that it is not economically possible to have a tight peg in a thin market when one of the assets is a volatile crypto currency.

So these hypothetical "other" pegged crypto's with no premium all depend upon a market maker willing to take a financial loss.

We can lower the feed price such that the average trading price is $1.00 and that has always been my stance.  Adjusting the feed will adjust the average trade price and the premium is always relative to the feed. 

It all comes down to a choice on how you want the BitUSD holder to *pay* for their liquidity.   Do they pay up front when they buy it at a premium, or do they pay on the back end when they sell it at a discount? 

There is NO WAY AROUND IT.   Liquidity has a price proportional to market depth and volatility. 

After you factor out the liquidity price all that is left is "speculative demand".   In a bull market shorts "pay interest" by shorting below the cost of providing liquidity resulting in a narrower spread.  In a bear market longs must pay more for "price insurance".

Interest rates can be used to stabilize the supply / demand created by speculation but this stabilization is fleeting because anyone who buys for the interest rate is exposed to interest rate risks.  So whether it is interest rate risk or premium change risks it all works out to be the same thing to traders.   

 

Liquidity constraints in the underlying crypto mean that spreads are always inevitable around the peg, deviating between discounts and premiums. However, there are two key ways to minimise those spreads within that constraint:

i) offering the ability for arbitrage, and
ii) providing incentive for market makers.

Arbitrage is the key to tight pegging. Exchange traded funds provide convertibility (subject to a transaction cost that can reflect liquidity) to their Authorised Parties to maintain close pegging to NAV. CFD providers provide tight spreads because they can arbitrage their book in the external market. I think offering a means of convertibility is essential to get a peg to operate effectively.

Market-makers or spread-traders, trading between bitUSD and USD, could be incentivised to step in where their cost is lower than that of the arbitragers (as they do not bear a spread cost on BTS). They are incentivised because they know there are arbitragers sitting in the wings limiting the variation against them, and possibly because there is a mechanism (e.g. yield) to return value toward parity.

Even with all this, there must always be deviation around the peg. In theory this could be materially reduced by backing a bitUSD alternative with BTC rather than BTS once the tools are in place to minimise counterparty risk.

On a separate point - I have raised concerns about the idea of using the feed price to dynamically adjust the average trading level. I would appreciate learning more about this.

However saying all that I believe the collateral for bitUSD needs to be bitcoin instead of bitshares due to its market depth, liquidity, stability and acceptance as an asset class. This will be possible hopefully with sidechains. This will hurt the bitshares price alot but if it isn't implemented someone else will and that newly created bitcoin backed bitusd will surpersede bitUSD.

Welcome jonnybitcoin.
I agree, that's a big commercial opportunity when we have the tech in place. Some previous discussion raised here:
https://bitsharestalk.org/index.php/topic,15957.msg204725.html#msg204725
It could be a huge boon to the BTS price too.

On whether we should have have a bitUSD centred on parity or at a premium, assuming both are possible:

I do not understand all the subtleties of merchant, payment processor and other third party operations. However, it seems to me that everything that can be done with a bitUSD trading at a premium can also be accomplished with a bitUSD trading at parity, only with different labels on the currency flows. For example the equivalent financial result for all parties can be achieved by having a bitUSD centred on parity, with consumers paying the fee directly to the payment processor. However, maybe bytemaster is right, based on his advice, that the way in which fees are labelled or packaged in the marketplace makes a big difference to actual behaviour. I can't argue with the way other people think.

I do have a concern however that a bitUSD at a premium reduces its utility for a number of other application areas that I touched on in my previous post. A bitUSD trading around parity would serve these applications much better IMO. Regardless of the direction taken in the near term, I would like to continue pursing the method for a bitUSD based around parity, because I personally think it would give bitShares the best long term result.

On whether it is possible to have a bitUSD that trades close to parity:

   I reject the notion that we can establish a peg with 0 premium and 0 volatility in the premium and those seeking this perfection are seeking the impossible.   

This is knocking down a straw man, and perhaps you have misunderstood my goal. What I wish to see is bitUSD that centres around parity rather than a premium. Clearly there will be deviation either side of that with movement in supply and demand, at least to the point that arbitragers are incentivised to jump in and prevent larger deviations. This is the model I am working toward. There is no plan for 0 volatility.

Do you think this notion is impossible? I don't have enough evidence of that yet.

  Lets take a "perfect" and "balanced" system of a simple prediction market that settles once per year at a "perfectly fair" price.    In this prediction market both sides have the exact same liquidity and expectation of profit.  Any deviations from the value of $1.00 is a profit opportunity for both parties.   

  Under such a system what is the "volatility" of the premium?   It is proportional to the holding cost until settlement combined with the demand for leverage on BTS.   I would expect a range around $1.00 that could be off by over 10% at times.    So even with guaranteed settlement at a fair price on a specific date in the future you do not get an asset that is always worth $1.00 with a stable premium.    The best we can do is get an asset that is approximately worth $1.00 and the volatility in the premium is beyond our control. 

This is not the most useful comparison for bitUSD. What you describe here is a futures contract, that expires and settles at a fixed date in the future. The fair price for a futures contract is not equal to the spot value of the underlying asset, because of the cost of carry. That is, the difference between the funding cost related to the currency and that of the BTS in which it is settled. The market will price this in when it compares the cost of establishing a position with similar exposure in the external market. The fair value for the future will also fluctuate as the funding costs in these markets fluctuates. None of this means that the future is trading at wide variations from its fair value, its just that its fair value is not parity. In a different context, we would not argue for example that a discount bond is trading at large deviation from its fair value because it is trading at a large discount to its par value.

In these cases however, large deviations from the "fair value" are contained by the action of arbitragers, and that is really what is important.

A bitUSD is a tracking instrument, more like an exchange-traded fund (ETF), in that it does not expire. The fair value of an instrument like this is its underlying asset value. Instruments like ETFs do trade close to the spot value because of the action of arbitragers in those markets. Authorised Parties are granted the ability to exchange units in the fund for units in the underlying asset, making this arbitrage possible. I think a similar mechanism is possible for a bitUSD.

On whether it is possible to have a stable bitUSD premium:

  If the premium is persistently biased by a certain minimum amount then the market will end up adjusting prices to account for that bias.  This "constant bias" can be relatively stable and can be removed by adjusting the feed. 

Would you mind expanding on this mechanism as I still don't understand it? If the idea is to adjust the feed price downward (equivalent to a higher settlement fee) for larger bitUSD premiums, I've raised my concern about that in the link below. However, I could also be at risk of knocking down a straw man if I misunderstand your intended mechanism.
link: https://bitsharestalk.org/index.php/topic,16143.msg210901.html#msg210901

starspirit, do I understand this correctly: the proposed way to balance supply and demand to create a peg equilibrium is to allow a negative yield? (as well as a positive yield)

Doesn't that mean that if you hold bitUSD while there is negative yield you are essentially getting charged a negative interest rate? Won't this be a hard sell for the general public?

Well, that's a good question monsterer, that I've been debating with myself for a while, and have not reached an absolutely firm view on yet - i.e. is the possibility of negative yield needed at all? Here's some food for thought...

The yield is determined by the free market as the level that equates a pegged currency at parity with its real equivalent for the marginal buyer and seller. So by definition, it must be acceptable to all the current holders of the pegged currency.

Forcing any other yield on the market creates a situation where either supply falls (forcing a higher than natural rate), or demand falls (forcing a lower than natural rate), either of which means that the quantity on issue and in use would be less than if the free market had determined the rate. That's the advantage of flexible yield - in that in maximises and stabilises the quantity on issue as yields can adjust for supply or demand shocks.

Now why would the free market choose a negative yield? Its important to note that this is likely to be very rare. But it is a far more likely situation in the current global environment of suppressed interest rates.

The yield on a pegged currency is normally going to move with external interest rates due to arbitrage. But it would move around that depending on relative supply and demand. For example, it could be higher if issuers/shorts are positively motivated (e.g. crypto bull markets) or lag if they are negatively motivated (e.g. crypto bear markets). In any case, the current environment of zero to negative external rates (after fees) in some currencies like USD or EUR could well lead to a free-market yield in the pegged currency that might calibrate at zero or below.

While bond markets can reflect negative yields very easily by trading above par, negative rates in an at-call market require an actual payment from longs. Cash holders will seek to avoid this if they can. That's exactly why some economists have been recently calling to ban cash, and force people into deposits where they cannot avoid such charges (a proposal I'm against BTW).

Although negative rates are theoretically possible, it's a profound question as to whether negative interest rates on a pegged currency should be allowed or not, even if that's where the free market would calibrate. On the one hand, if external interest rates can go negative after fees, why not for the pegged currency? But its possible for example that even if rates were positive, the mere possibility that they might be allowed to go negative, could be seen as a controversial or negative feature by the market, reducing demand at all times. Like I said, I can see both sides of an argument here.

If we decided we did not want to cross the zero bound, it would not "break" this or any other pegging system. It could still operate. The result of forcing the rate to remain at zero when the natural rate is negative, would be to:

i) reduce the supply of the pegged currency, perhaps dramatically or completely in some extreme situations where shorts demand compensation to participate, and
ii) the peg would break and a large premium develop.

This may not be the end of the world if these situations are rare, but its bound to be an issue of debate also.

Sorry for the long post - but its a complex issue, and not one that is black and white.

I think getting merchants to accept BitUSD is doing it the hard way.

Debit cards will emerge that you can load with BitUSD and spend as ordinary USD.
Merchants won't know the difference.  They get USD and don't care about fiat counter party risk.

So BitUSD will always convert to at least a dollar when it hits your card and any surplus after the card's fees just sits on the card as a small bonus.  You only load the card with what you are going to use in the near future which is a negligible amount of counter party risk.  The rest stays safe in your BitUSD "savings account" until you need to move it to your debit card's "checking account".

Ordinary users will like that they sometimes get a small bonus when loading their card.  It will seem like yield.  Biased to be always a positive number.

Such a system would work everywhere on Day 1.

Except that the ordinary users would need to pay a premium to get the bitUSD in the first place. By this reasoning that will seem like a loss. Biased to always be a negative number. Completely negating the "yield".

I like the sound of the debit card payment concept. However for this discussion, it does not make a bitUSD premium a benefit to anybody. Might as well buy euro above parity to the USD (which it is, and perhaps we could call it premium-USD), then when you load your multi-currency card and get USD, the resulting quantity is guaranteed to be a bigger number. I don't think anybody would believe that is a yield though.

BitShares. A share in the future.
BitShares. A share in freedom.
BitShares. A share in destiny.