bounties and some questions

[tonyk]

What is really cool application of this method is that one can build an oracle, that constantly watches the price changes of BTC (and other collaterals) and dynamically adjust the insurance rate.

[tonyk]

The logic is as follows:

The insurance is equal to the amount needed to buy a put option [with exercise price equal to the black swan event price of BTC] – so you can sell all BTC held as collateral for at least that price if BTC goes below it. The option's theoretical price itself is a complete reflection of the probability of BTC price going below the strike price of the option.

 A mainstream (the most popular arguably) way to calculate that put price is presented in the  post above.

Interesting logic. So what initial insurance rate would you argue for? Once the system is running we are simply going to profit maximize for mkr, adjusting the insurance rate to optimize dai supply growth - but setting the initial insurance rate value is really difficult because its so arbitrary yet important to price correctly for early adoption.

If you run the formula you end up with a precise number. I can do it for you *. I will do it by plugin deferent numbers for the volatility of BTC - starting from price changes for say say 12h/ daily/ weekly periods and will collect the data for 3 - 6 - 12mo, up to several years of BTC trading.
My point is, there is not one and only straight correct answer (4% and that's it); but a reasonably correct one  is very achievable  - for example,before this last  rally, BTC had long uncharacteristically  low volatility period - generally the first 10 or so months of this year. Taking that period only (or even the last 12mo) will result in  a number too low, in my view.


*If you agree on principal that that is the correct answer, aka I won the bounty.

[Rune]

The logic is as follows:

The insurance is equal to the amount needed to buy a put option [with exercise price equal to the black swan event price of BTC] – so you can sell all BTC held as collateral for at least that price if BTC goes below it. The option's theoretical price itself is a complete reflection of the probability of BTC price going below the strike price of the option.

 A mainstream (the most popular arguably) way to calculate that put price is presented in the  post above.

Interesting logic. So what initial insurance rate would you argue for? Once the system is running we are simply going to profit maximize for mkr, adjusting the insurance rate to optimize dai supply growth - but setting the initial insurance rate value is really difficult because its so arbitrary yet important to price correctly for early adoption.

[tonyk]

The logic is as follows:

The insurance is equal to the amount needed to buy a put option [with exercise price equal to the black swan event price of BTC] – so you can sell all BTC held as collateral for at least that price if BTC goes below it. The option's theoretical price itself is a complete reflection of the probability of BTC price going below the strike price of the option.

 A mainstream (the most popular arguably) way to calculate that put price is presented in the  post above.

[tonyk]

Bounty 2: We are now beginning discussions on how to set the insurance rate of the Dai Credit System. It's a variable rate that will be optimized on as the system progresses, but it still needs an initial value that isn't completely bonkers. Currently the best argument for this is 4% per year (meaning if debt interest is 11% then dai yield is 7%), if someone comes up with a solid public argument on the Maker forum for a different initial value I'll pay 1000 USD for the effort. See my post here: https://forum.makerdao.com/t/argument-for-a-4-insurance-rate/279/4?u=rune

This assume: 1. you only use BTC  as collateral (adjust for each new asset used) and 2. you withdraw nothing from the ins fund for 30 days. (I do not know if/how/when and how much you do. 3 I would think you should be pretty safe as T^0.5 as true ins rate, but have provided the total 30 days needed to make your own call on that.

T - total insurance fund needed for the first 30 days (and any 30 days there after)

U - price of btc

t = 1/12

v = annual volatility of BTC [in decimals; i.e 18% as 0.18]

r = risk free interest rate [as decimal fraction]

N = normal distributing CDF

E - price of btc to cause black swan

ln - natural logarithm


H =(   ln (U/E) + (r+(v^2) /2)*t    ) / (v* t^(1/2)


T = - U * N(-H)  + E*e^(-r*t)*N(v*t^(1/2) - H)

[phillyguy]

I have 99 CFSCOIN sitting in my wallet. I'll gladly send them back to you for free - or if you want to send me one or two MKR as a tip, that's cool too :-).

[Rune]

Sorry for getting mad at you Tony, I was stressed and it made me late for a meeting with some investors and the yunbi team, since I obviously couldnt just wait with replying to such a post. Ultimately Im glad to hear your opinions on this, I just wish you would make it sound less like FUD and also stop assuming that we are idiots.

Also I should have just explained the debt ceiling mechanism instead of referring to the whitepaper. There are individual debt ceilings for each collateral type that have to be manually increased over time. So in the beginning the portfolio could be something like:

ETH: 1000k max debt
MKR: 100k max debt
EMAXBTC: 1000k max debt (emax is our E-Money licensed partner exchange who will do a btc IOU on ethereum)
EMAXBTS: 300k max debt
DGXGOLD: 500k max debt (digix is a company that puts fully audited and regulated physical commodity IOUs by singapore vaults on ethereum)
etc etc

These rates will then be manually increased over time, each increase a subject of discussion and review by the community and board of governors, and any action taken has a priming delay ensuring a bad decision can be reversed. We are going to be very careful and conservative with this as MKR holders insure everything and are thus highly incentivized to not fuck up.

Another thing to understand is that long term Maker will not just be a platform for margin trading, as that is just a small part of the market, it will simply be a general platform for collateralized credit and most of the CDP's will use DAO shares or company stock (delivered on the ethereum blockchain by services such as Emax or DACX). And rather than use the Dai they issue for margin trading, they will use them for day to day liquidity for financing operations, something most small companies currently pay 10-15% rates for even when the discount rate is negative.

Once publicly traded SME's have direct access to credit with as few middlemen as possible things will really start to change as the economies of scale in the current financial system is one of the biggest factors causing companies to be overly bloated and centralized, and thus needlessly inefficient.

[xeroc]

Additionally describing the stable coin creators as borrowers is something I have argued against for  a long time (at least 1 year now).  You are not the only one that finds (or used to find) the use of this terminology useful for his/her personal reasons and systems he/she tries to explain. I understand why it is very convenient - it is a great way to persuade the dai holders that they will receive interest on their stable coins. Same argument I heard when someone else was trying to pay interest on their own stable coin. The problem is this terminology brakes with negative interest rates - So what you are receiving interest on the money you borrowed?

So the stable coin creator is not a borrower - he just has a long position (how long depends on the collateral requirements) in the collateral against the stable coin (or dai in your terminology)! This is not only factually correct but gives an easy explanation on the 'negative interest on the borrowed money' phenomenon, present in systems that claim the 'dai creator as the borrower'

I tend to agree here and I think I made some remarks on that in the whitepaper. For me, whenever you put collateral into a contract and get issued another token (stable or not) ... you basically just relabel your funds ... from 150% ETH to 100% dai .. If you were to just close the contract again (assuming 0% interest) .. then you could simple reverse the whole process and get back to 150% ETH.
It's not that you 'borrow' from a decentralized entity (the protocol) .. but you enter a contract with that entity and gain access to a token that is under the sole (issuance)-control over this token.

Now, when you sell that token (dai in your case) to someone else .. that person is than LONG 'dai' .. and not a borrower at all .. it's a trade between two arbitrary assets ..

Anyway, just my 2 cents .. and I am not an economist (btw)

[liondani]

hope this conversation will continue...
I am sure it will be worth the time for all participants

[tonyk]

Additionally describing the stable coin creators as borrowers is something I have argued against for  a long time (at least 1 year now).  You are not the only one that finds (or used to find) the use of this terminology useful for his/her personal reasons and systems he/she tries to explain. I understand why it is very convenient - it is a great way to persuade the dai holders that they will receive interest on their stable coins. Same argument I heard when someone else was trying to pay interest on their own stable coin. The problem is this terminology brakes with negative interest rates - So what you are receiving interest on the money you borrowed?

So the stable coin creator is not a borrower - he just has a long position (how long depends on the collateral requirements) in the collateral against the stable coin (or dai in your terminology)! This is not only factually correct but gives an easy explanation on the 'negative interest on the borrowed money' phenomenon, present in systems that claim the 'dai creator as the borrower'

[tonyk]

Sigh, it just won't stop...

which in the beginning let's face it will consist of 90%+ MKR

Why haven't you bothered spending 20 minutes trying to figure out the thing your misunderstood genius is apparently able to so effortlessly pick apart? Maybe you would have read about something called the debt ceiling, the primary risk management mechanism that enforces diversity of the collateral, as we would obviously never be dumb enough to allow a 90% MKR collateral scenario. If you want me to concede that is impossible to insure an asset with itself and that doing so would be really dumb, then congratulations, you are right. You are also being completely irrelevant to the situation as that is literally rule number one of insurance and obviously not something anyone who spent more than a minute thinking about collateral insurance would ever do.

And calling it insulting that I am spending time here to patiently bang my head against a wall in an effort to explain why you're wrong, and it then turns out you clearly haven't even bothered with reading the white paper, is insulting. Just the fact that you seriously thought we could be stupid enough to not prevent an obviously stupid failure mode, is insulting. Why was your reaction "oooh, these guys are idiots, let me go tell them on a forum how stupid they are", rather than "hmmm, maybe I should read the white paper instead since I've clearly missed something".

Read the whitepaper AGAIN!  - as much as you do not believe in markets (i.e. believe that the most efficient way to control the supply of stable coins is by using interest rates. Interest rates if need be  fine tuned by a  board of people or oracles acting as governors ), the market will chose the most effective collateral to use in the current conditions (granted in the confines of your randomly picked [sorry scientifically and FED approved]  "debt ceiling"/"volatility grouping" requirements).

You might believe that you will manage to pick the 'debt ceiling' and volatility classification, so precisely that every day all the stable coins created will use a perfect mix of all available collaterals, BUT I think it will be one particular asset that the people are predominantly willing to go against at any given time. As a result that asset will be the biggest one held in collateral.  I also think that MKR will be the one chosen to begin with (it really does not matter which one it will be - even if it BTC it can also go down sharply against the dollar and cause black swan event).
[*Reading the white paper (Again), I did not see a mentioning of setting once and requiring forever a permanent mix (with fix %-ages) of collaterals. Which is of course a possibility but goes with its own big can of warms.]

PS
Reading the thingy again - I realized that your claims that the stable coins are fungible is open for debate. 

[cass]

Also, if I register a short UIA name for 500k BTS now, will it carry over to graphene without any change? How much will short UIA names cost post graphene?

yep no problem

[fuzzy]

Sucks the discussion ended here in this way. 

But before it did I was enjoying reading!

[Rune]

Sigh, it just won't stop...

which in the beginning let's face it will consist of 90%+ MKR

Why haven't you bothered spending 20 minutes trying to figure out the thing your misunderstood genius is apparently able to so effortlessly pick apart? Maybe you would have read about something called the debt ceiling, the primary risk management mechanism that enforces diversity of the collateral, as we would obviously never be dumb enough to allow a 90% MKR collateral scenario. If you want me to concede that is impossible to insure an asset with itself and that doing so would be really dumb, then congratulations, you are right. You are also being completely irrelevant to the situation as that is literally rule number one of insurance and obviously not something anyone who spent more than a minute thinking about collateral insurance would ever do.

And calling it insulting that I am spending time here to patiently bang my head against a wall in an effort to explain why you're wrong, and it then turns out you clearly haven't even bothered with reading the white paper, is insulting. Just the fact that you seriously thought we could be stupid enough to not prevent an obviously stupid failure mode, is insulting. Why was your reaction "oooh, these guys are idiots, let me go tell them on a forum how stupid they are", rather than "hmmm, maybe I should read the white paper instead since I've clearly missed something".

But the top of the stupidity list is your clear attempt to claim that posting something on a forum and if no-one has proven you are wrong with arguments to your satisfaction (see above stupid and more stupid) - That is the proof that your system is 100% solid!

There you go calling me stupid again. You're really being a dick.

Don't bother replying unless you're going to finally give me a solid argument for why the system doesn't work, now that I've shown you that your lazy assumption that the Dai will be primarily collateralized by MKR, was wrong. I have actual work to do and you are wasting my time.

[tonyk]

I thought I already won the second bounty by saying it does not really matter what % you choose. Real Insurances must have all the money upfront, what you are doing is collecting money for a rainy day. It really does not matter what % you chose.

On the same note - you can set aside from launch an amount that you dedicate for that purpose (i.e. insurance) from the total supply - not a perfect solution but the ONLY one that makes some sense.

Infinite makercoin is set aside from launch for insurance

The assertion that the system is fundamentally flawed in an obvious way that can be described with a few lines on a forum post is easily contradicted by the fact that MKR has a market value. (Unless you want to argue that all MKR investors/employees are idiots, which seems a bit harsh even for the crypto community)

If you think that argument sucks, I hope you'll be  encouraged to do more research and come up with arguments for your position, so I'll have something real to make a rebuttal of. If you manage to prove with real arguments (such as examples using microeconomics) that forced inflation doesn't work as an insurance fund (as long as the asset still has market value) then you'll definitely get the bounty . In fact this makes me think I should make another, much higher bounty for finding a flaw in the whitepaper. How much money would motivate you to do real research?

What research and more proof do you need Rune - As I said in the other thread that:
1. in the case of falling MKR coin prices from the very beginning, the shorters (or whatever they are called in you system) demand being paid interest (instead of paying for interest and insurance)  => no insurance fund.
2.  If black swan occurs at some point during this price decline you have no funds to cover the stable asset so you print more MKR. Driving the price further down.
The only thing needing proof in the above is the claim that at falling prices (and  expectations the price decline to continue) the shorters demand being paid interest. If that is the proof that you need I can gladly do it for you.

Interest rates going negative when there is bear market in  a single asset is just a blind assertion. Take a look at historical btc price vs interest rate on btc collateralized debt if you want to actually try to prove this (hint: reality won't fit your beliefs). Secondly if it was true its still completely disconnected from the second part of your argument. Interest rates can be negative while the insurance rate stays positive..

I think what's causing you trouble is wrapping your head around the negative interest with positive insurance rate. It appears counterintuitive so you conclude that the system would go into a weird failure mode with a damaging positive feedback on the insurance rate.

However, consider the scenario where yield is -5% and the insurance rate is 2%. Dai issuers would see their outstanding debt fall with -3% APR, achieving what you just thought impossible! Bear in mind that this scenario requires either absurd Dai demand, or a global economic downturn of cataclysmic levels. Despite that, the insurance rate and Makers income relative to outstanding Dai would remain the same.

So, I really did not see that - your Dai buyer will pay 2% insurance on top on the quoted rate? So if the shorter wants 5% premium the buyer will pay 7.1% premium just so he is 'insured'.... ohh OK. (1)

"Interest rates going negative when there is bear market in  a single asset is just a blind assertion." - which nobody did Rune. Asking for premium [cause the premium is the correct term here not interest]  in the case of the basket of collaterals  (which in the beginning let's face it will consist of 90%+ MKR) going down in price, is not blind assertion but a easily provable fact.

[EDIT]
(1) I realized you might have a hard time "wrapping your head around " that, this has no impact on the whole picture - Issuing 100K of stable assets even with 5% insurance paid is still 5K in the insurance fund. If the black swan happened at that point your assets are still 95% NOT covered.

PS
And It is really a stupid idea to post something as a serious discussion and the whole time to claim stuff that I have never said is wrong as your main argument. ( the rest of your arguments are – “it is not likely so we can ignore this scenario”)
Using insulting language to attack the things I never said is even more stupid.
But the top of the stupidity list is your clear attempt to claim that posting something on a forum and if no-one has proven you are wrong with arguments to your satisfaction (see above stupid and more stupid) - That is the proof that your system is 100% solid!