BitShares Forum

Main => General Discussion => Topic started by: monsterer on April 27, 2015, 08:54:58 am

Title: Ethereum's bitUSD equivalent - has this been discussed and dismissed?
Post by: monsterer on April 27, 2015, 08:54:58 am
The ethereum guys have their own outline for a asset type which retains a stable value (like our pegged assets), called Seigniorage Shares.

https://github.com/rmsams/stablecoins/blob/master/00-main.pdf

The principle idea is that in order to maintain stable value, the blockchain will auction off bitUSD at a discount when the price is too high (increasing supply), and buy it back at a premium (and burn it), when the price gets too low.

How much it buys/sells on the open market is governed by a price feed, essentially.

Has this idea been discussed and dismissed as unworkable? It seems attractively simple compared to our system, or the proposed replacements...

Cheers, Paul.
Title: Re: Ethereum's bitUSD equivalent - has this been discussed and dismissed?
Post by: starspirit on April 27, 2015, 12:22:08 pm
monsterer, in writing the whitepaper you have seen, I considered this possibility. It's actual the most basic way of implementing a stable currency, but it's suboptimal because it creates massive volatility in the supply of the native token.

First let me summarise what I understand is being proposed here, although I have not read all the detail. To put this in terms that Bitshares would understand, when bitUSD demand rises, the block-chain would auction off freshly created bitUSD for BTS and the BTS would be burned. When bitUSD demand falls, the block-chain would auction off freshly created BTS for bitUSD and burn the bitUSD. This maintains a flexible supply allowing the price to be pegged.

Now critically this is un unfunded approach, with no defined collateral pool backing the bitUSD. There is only a commitment to inflate the BTS supply as much as necessary to meet the sales of bitUSD when they occur. All is great while bitUSD demand is rising, but if bitUSD started getting sold back, and particularly at depressed valuations of BTS, it would dilute BTS into the ground. Of course practical measures could be used to constrain this (we've actually touched on this topic before monsterer), but only at the expense of compromising liquidity and exchangeability.

In thinking through all the options for my whitepaper, you will see I came up with a number of pre-conditions to the optimal architecture. See the bottom of this post for my whitepaper... https://bitsharestalk.org/index.php/topic,15880.msg203777.html#msg203777.

To reiterate, those preconditions are:

Pre-conditions

The following are pre-conditions to the required architecture:

(i)   The currency must be redeemable on the block-chain to eliminate counterparty risk,
(ii)   The currency must be backed by a pool of native tokens to meet redemptions, to avoid inflation of native tokens for redemptions
(iii)   Exchange of [edit] native tokens for creation or destruction of currency [end edit] must be facilitated at the feed price (near instantly) to ensure maximum pegging in all markets
(iv)   There needs to be a single floating variable to ensure supply and demand can equilibrate at the peg, the most recognized being interest
(v)   The interest rate must be determined by market forces between buyers and sellers, to ensure markets always clear (no shortage or surplus at the feed price)
(vi)   Markets at varying terms should be available, so that mismatches in term preferences between buyers and sellers can be priced along the yield curve, rather than stifling an isolated market.

The proposed approach passes (i), but fails at (ii) as I discussed above. Thus it also fails (iii)-(vi). All the pre-conditions are satisfied by my white-paper approach. [Incidentally, BitAsset 3.0 on its own satisfies (i)-(ii) and now (iii), but does not meet (iv)-(vi)].

Another issue he talks about is the price feed. He basically reiterates my view that a price feed must be exogenous if it refers to an external valuation anchor. However, he says that an endogenous price feed is possible if the goal is simply stability rather than pegging. This seems to depend however on maturity in the native token, by which time a stable alternative is irrelevant because the native token itself will be much more stable.

Title: Re: Ethereum's bitUSD equivalent - has this been discussed and dismissed?
Post by: bytemaster on April 27, 2015, 06:14:34 pm
monsterer, in writing the whitepaper you have seen, I considered this possibility. It's actual the most basic way of implementing a stable currency, but it's suboptimal because it creates massive volatility in the supply of the native token.

First let me summarise what I understand is being proposed here, although I have not read all the detail. To put this in terms that Bitshares would understand, when bitUSD demand rises, the block-chain would auction off freshly created bitUSD for BTS and the BTS would be burned. When bitUSD demand falls, the block-chain would auction off freshly created BTS for bitUSD and burn the bitUSD. This maintains a flexible supply allowing the price to be pegged.

Now critically this is un unfunded approach, with no defined collateral pool backing the bitUSD. There is only a commitment to inflate the BTS supply as much as necessary to meet the sales of bitUSD when they occur. All is great while bitUSD demand is rising, but if bitUSD started getting sold back, and particularly at depressed valuations of BTS, it would dilute BTS into the ground. Of course practical measures could be used to constrain this (we've actually touched on this topic before monsterer), but only at the expense of compromising liquidity and exchangeability.

In thinking through all the options for my whitepaper, you will see I came up with a number of pre-conditions to the optimal architecture. See the bottom of this post for my whitepaper... https://bitsharestalk.org/index.php/topic,15880.msg203777.html#msg203777.

To reiterate, those preconditions are:

Pre-conditions

The following are pre-conditions to the required architecture:

(i)   The currency must be redeemable on the block-chain to eliminate counterparty risk,
(ii)   The currency must be backed by a pool of native tokens to meet redemptions, to avoid inflation of native tokens for redemptions
(iii)   Exchange of [edit] native tokens for creation or destruction of currency [end edit] must be facilitated at the feed price (near instantly) to ensure maximum pegging in all markets
(iv)   There needs to be a single floating variable to ensure supply and demand can equilibrate at the peg, the most recognized being interest
(v)   The interest rate must be determined by market forces between buyers and sellers, to ensure markets always clear (no shortage or surplus at the feed price)
(vi)   Markets at varying terms should be available, so that mismatches in term preferences between buyers and sellers can be priced along the yield curve, rather than stifling an isolated market.

The Ethereum approach passes (i), but fails at (ii) as I discussed above. Thus it also fails (iii)-(vi). All the pre-conditions are satisfied by my white-paper approach. [Incidentally, BitAsset 3.0 on its own satisfies (i)-(ii) and now (iii), but does not meet (iv)-(vi)].

Another issue he talks about is the price feed. He basically reiterates my view that a price feed must be exogenous if it refers to an external valuation anchor. However, he says that an endogenous price feed is possible if the goal is simply stability rather than pegging. This seems to depend however on maturity in the native token, by which time a stable alternative is irrelevant because the native token itself will be much more stable.

Very good analysis.   I think BTA 3.0 + Bond Market will allow the market to adapt to all interest rates / terms / yield curves / and collateral ratios/leverage requirements. 
Title: Re: Ethereum's bitUSD equivalent - has this been discussed and dismissed?
Post by: Ander on April 27, 2015, 06:19:47 pm
While I think the Ethereum system might work for what they are going for, in BTS we do NOT want BTS to be created and auctioned off when bitUSD is sold, since BTS is the equity in the entire system!  (Note that in Ethereum, they are not having Ether be created and auctioned off in this system).

If we did this with BTS, then in any market pegged asset, if people manipulated the system correctly and obught/sold at a profit, they could actually cause new BTS to be generated, causing inflation.  They could do this in every single asset. It would be terrible!
Title: Re: Ethereum's bitUSD equivalent - has this been discussed and dismissed?
Post by: toast on April 27, 2015, 07:27:49 pm
IMO the better effort is the eDollar which is heavily modeled after bitUSD, built by our own "ethereum guy" Rune.

"Ethereum's bitUSD" is a bad way to describe it, because this is one single stablecoin proposal for their platform. If they succeed at their mission there will be plenty of people who build on ethereum who want no particular association with the ethereum project/team.
Title: Re: Ethereum's bitUSD equivalent - has this been discussed and dismissed?
Post by: Pheonike on April 27, 2015, 11:52:22 pm

eDollar full presentation with slideshow and demo

https://www.youtube.com/watch?v=39vuJDtqY2c
Title: Re: Ethereum's bitUSD equivalent - has this been discussed and dismissed?
Post by: starspirit on April 28, 2015, 02:12:49 am
monsterer, in writing the whitepaper you have seen, I considered this possibility. It's actual the most basic way of implementing a stable currency, but it's suboptimal because it creates massive volatility in the supply of the native token.

First let me summarise what I understand is being proposed here, although I have not read all the detail. To put this in terms that Bitshares would understand, when bitUSD demand rises, the block-chain would auction off freshly created bitUSD for BTS and the BTS would be burned. When bitUSD demand falls, the block-chain would auction off freshly created BTS for bitUSD and burn the bitUSD. This maintains a flexible supply allowing the price to be pegged.

Now critically this is un unfunded approach, with no defined collateral pool backing the bitUSD. There is only a commitment to inflate the BTS supply as much as necessary to meet the sales of bitUSD when they occur. All is great while bitUSD demand is rising, but if bitUSD started getting sold back, and particularly at depressed valuations of BTS, it would dilute BTS into the ground. Of course practical measures could be used to constrain this (we've actually touched on this topic before monsterer), but only at the expense of compromising liquidity and exchangeability.

In thinking through all the options for my whitepaper, you will see I came up with a number of pre-conditions to the optimal architecture. See the bottom of this post for my whitepaper... https://bitsharestalk.org/index.php/topic,15880.msg203777.html#msg203777.

To reiterate, those preconditions are:

Pre-conditions

The following are pre-conditions to the required architecture:

(i)   The currency must be redeemable on the block-chain to eliminate counterparty risk,
(ii)   The currency must be backed by a pool of native tokens to meet redemptions, to avoid inflation of native tokens for redemptions
(iii)   Exchange of [edit] native tokens for creation or destruction of currency [end edit] must be facilitated at the feed price (near instantly) to ensure maximum pegging in all markets
(iv)   There needs to be a single floating variable to ensure supply and demand can equilibrate at the peg, the most recognized being interest
(v)   The interest rate must be determined by market forces between buyers and sellers, to ensure markets always clear (no shortage or surplus at the feed price)
(vi)   Markets at varying terms should be available, so that mismatches in term preferences between buyers and sellers can be priced along the yield curve, rather than stifling an isolated market.

The Ethereum approach passes (i), but fails at (ii) as I discussed above. Thus it also fails (iii)-(vi). All the pre-conditions are satisfied by my white-paper approach. [Incidentally, BitAsset 3.0 on its own satisfies (i)-(ii) and now (iii), but does not meet (iv)-(vi)].

Another issue he talks about is the price feed. He basically reiterates my view that a price feed must be exogenous if it refers to an external valuation anchor. However, he says that an endogenous price feed is possible if the goal is simply stability rather than pegging. This seems to depend however on maturity in the native token, by which time a stable alternative is irrelevant because the native token itself will be much more stable.

Very good analysis.   I think BTA 3.0 + Bond Market will allow the market to adapt to all interest rates / terms / yield curves / and collateral ratios/leverage requirements.

Thanks BM. I think it is heading in a good direction - I would add:

- "BTA 3.0" (but with some modifications), plus
- An at-call interest-based deposit market
- Bond Market

I'll explain further in a subsequent post. You may see better ways of doing some of the things I've tried to do in my draft white-paper, but I would love for you to discuss it with me, given I think it covers more ground than I've seen covered elsewhere.