non-backed pegged currencies
Alex Mizrahi
things) user currencies which track value of a specific assets without
being backed by that asset by issuer.
At first this seems to be counter-intuitive... You can't create
digital gold out of nothing.
However, something very similar happens on derivative markets. For
example, futures price converges towards spot price by the date of
settlement on futures market.
This happens according to no-arbitrage condition: difference in price
would create an arbitrage opportunity which would level this
difference.
If you consider cash-settled futures, exchange operator effectively
moves money from pockets of those who have "lost the bet" into pockets
of those who won during the settlement.
This forced transfer creates potential arbitrage opportunity at the
time of settlement, which forces price convergence.
If we can implement something like cash-settles futures on
decentralized market we'll be able to create a digital asset with
price pegged to physical asset price which won't be backed by physical
asset. Price will be stabilized by speculations.
However, using futures directly would be inconvenient.
J. R. Willett outlined a stabilization approach involving two
different assets linked together, let's call them GoldShares and
GoldCoins.
Price of GoldCoins is supposed to track price of gold, while
GoldShares is very unusual kind of derivative which lets people to bet
on demand for GoldCoins.
They are linked together in such a way that GoldCoin supply will be
able to fluctuate, compensating for changes in demand and making sure
that price is more-or-less stable.
However, what paper outlines is a long-term stabilization and I doubt
it would work very well. I'll quote the paper:
* “Proportional” gain of the PID loop governing value corrections
(i.e. “Increase GoldCoin
supply by 1% per year for every 2% GoldCoins are below target.
Decrease GoldCoin
supply by 1% per year for every 3% GoldCoins are above target.”)
* “Integral” gain of the PID loop governing value corrections (i.e.
“Increase GoldCoin
supply by 1% per year for every 1% GoldCoins stay below their target
value multiplied
by the number of years they have been below target. Decrease GoldCoin
supply by 1%
per year for every 0.5% GoldCoins stay above their target value
multiplied by the
number of years they have been above target.
I believe there is a more straightforward way to implement this:
GoldCoins can be created or destroyed instantly, forcing price to fall
into a very narrow corridor.
Moreover, this approach is compatible with colored coins which are not
as flexible as MasterCoins.
Here's how we do this: suppose we have separate order books for
GoldCoins and GoldShares.
Price pegging is implemented by issuer, he defines acceptable price
range for GoldCoins, fixing it each day, for example.
As long as price is within that range nothing special happens, order
are matched like with normal p2ptrade.
Now suppose that price range for GoldCoin today is 0.9 .. 1.1 BTC.
Somebody puts a bid at 1.2 BTC and it is not matched. What happens?
Issuer intervenes and creates news GoldCoins to satisfy this extra demand.
To keep things balanced he needs to destroy GoldShares. Effectively
GoldShares are converted into GoldCoins.
At what rate? Obviously converting 1 GoldShare into 1 GoldCoin ain't
going to work.
I think it works perfectly if we use this formula: one GoldShare creates
(GoldShare price) / (GoldCoin price)
GoldCoins.
Here's what it means: if GoldShares are in shortage and their price is
high, one GoldShare creates a lot of GoldCoins.
When GoldShares are abundant and their price is low, you need many
GoldShares to create one GoldCoin.
So supply is going to be very flexible, if demand is high we can
create practically unlimited amount of GoldCoins, driving GoldShares
ask price higher and higher.
If demand is low and GoldCoin price is falling, we'll create lots of GoldShares.
However, this scheme isn't perfect. When there is a massive sell-off
of GoldCoins, we are essentially limited to whatever liquidity is on
bid side of GoldShares.
If bid price goes very low, GoldCoin price will fall below lower
threshold, potentially to zero.
As J.R. Willett writes, in that case a reserve fund might help the
situation. Issuer gets some money (e.g. Bitcoins) when he sells
initial GoldShares.
He should use that money to buy them back. However, in colored
coin-based protocol we can't really force issuer to do anything.
So, to summarize, non-backed pegged currency price can be stabilized
via a speculative market to a certain degree.
However, in case of a catastrophic sell-off we are at mercy of issuer.
Perhaps there is a better way to do it, via futures-like instruments
or something like that.
J.R. Willett
Nice summary!
I wanted slow forced price convergence to reduce churn and to allow arbitrage traders to do some of the work for me, but I don't see any reason it couldn't work as you describe.
If somebody actually attempts something like this, I'd probably wet myself with joy. Just the fact that you, the #1 Dev working on this area, understand the core of my proposal already has me dangerously close to such am embarrassing display! :-)
--
Jorge Timón
bitcoin) won't ever collapse in price, which I don't think you can
assure of any money (not even gold, despite his "intrinsic value" as
an industrial commodity).
In it's core is somehow similar to beertokens, but more complex and
making active use of the arbitrage opportunities. I think the reserve
for beertokens was going to be usd, LRusd or a basket of fiat
currencies, can't remember.
In any case, this all reminds me to some ideas I've been discussing
with Impaler from Freicoin.
His aim is different: to have zero inflation and zero risk investment
instruments so that we can measure the basic interest/liquidity
premium/time preference.
I think this is the last thread where we discussed this:
http://www.freicoin.org/decentralized-auto-adjusting-demurrage-t51.html
I'm copying him, just in case he wants to intervene.
>
>
>
--
Jorge Timón
http://freico.in/
http://archive.ripple-project.org/
Yoni Johnathan Assia
I think that What is described as goldcoins and goldshares is a private case of multiple issuers to the same asset, each of them making a market allowing the buyers and sellers to define the spread between the 2 issuers.
An example of this in current market structures are but side and sell side market makers on exchanges, very common in the US.
I believe A more abstract and free market approach, would be for each issuer to decide if they are also a market maker which means they contniously place either buy or sell orders or both, and enable in p2p trade to define assets like gold which are supported by more than one color.
This translates the risk of the issuer or market maker into the spread between the bid ask, so the exchange rate of gold is as thin as there are more trustable issuers and market makers.
For example i am willinh to sell gold at market+100 pips, and buy gold at market-100 pips, as a single issuers the risk of holding the asset depends on trusting that I will always buy/sell in the market. If there are 20 different market makers then the risk is the is lower as all you need as holder is to trust that one of them will remain a market maker when you want to sell.
I think we are almost ready to issue goldcoins backed (partially) by bitcoins, who wants to buy some goldcoins ?
--