Forking the Blockchain for Bonds (25 BTC Bounty)

[cunicula]

[edited post to replace term futures with bonds and add bounty]
I think the bitcoin system should allow mining for bitcoin bonds. Bitcoin bonds would be tradable coins with a fixed maturity date.
Upon reaching the maturity date, bitcoin would become identical to regular bitcoins. In my view, this is a much more efficient solution than exchange-based loans that require third parties to hold collateral.

For example, miners would have the option of mining instantly maturing bitcoins or bitcoins maturing on Jan 1st of 2012, 2013, 2014, 2015, 2016. (one to five year maturity)
Average difficulty for all maturities would be pegged to hold the total number of all types of coins constant according to Satoshi's schedule. Difficulty for each individual maturity would be initially set to the average difficulty and then allowed to float (like difficulty is handled right now). Supply and demand would determine how coin generation was distributed across the various maturities. This would smooth discontinuities in the coin generation rate that exist under the current system.

Bonds could be exchanged for regular bitcoins on bankless electronic exchanges, much like the Namecoin/Bitcoin exchange at bitparking. These exchanges have extremely low costs because the exchange does not need to hold collateral for a prolonged period. The exchange rates would indicate bitcoin holders' expectations about the future of the bitcoin economy.

Bonds would give wealthy individuals a stronger incentive to develop the bitcoin economy. For example, suppose Mary holds bitcoins and has a business idea that would add significant value to the bitcoin economy within one to two years. It would be in Mary's interest to sell some of her current bitcoins for bonds and some for USD to finance her idea. Effectively, Mary would be loaning her bitcoins out to raise USD capital with perfect bitcoin collateral. Mary would internalize some benefit from any bitcoin appreciation resulting from her business. If Mary just held USD, she might want to purchase bitcoin futures for this reason as well. This mechanism could encourage larger companies to enter the bitcoin economy because value-adding companies would earn more from purchasing bitcoin bonds rather than regular bitcoin.

Many will object to revision of the bitcoin generation process, but I think this is misguided. Whether the bitcoin economy will be valuable several years from now depends on the currency remaining more innovative than potential competing currencies. The emergence of a competitor is a much more important concern than whether bitcoin will be inflated through revised currency generation. Bitcoin users wouldn't accept an inflationary revised currency generation anyway.

Offering 25 BTC bounty for this, see post below.

[BubbleBoy]

The futures market will not solve Bitcoin's inherent instability. There's no point in taking risks now to develop the Bitcoin economy, when you can leave somebody else to do it and reap the gains. If you don't have any faith in the long term viability of the Bitcoin economy you liquidate now, again zero risk. Efficient markets baby.

[cunicula]

Market efficiency depends on market organization. My proposed reorganization would create a deep futures market because of miner arbitrage across currency creation with different difficulty levels. It would offer much lower transaction costs than a third party based system. A deep futures market with low transaction costs would make it much easier for companies to reap benefits from long-term entry into the bitcoin economy.

[relative]

either you or I don't understand futures.
futures are by definition the exchange of one good for another (or a equivalent monetary settlement), here BTC for USD.
how can this NOT be done by an exchange but by bitcoin only?

[Stefan Thomas]

I've sold Bitcoin options and made good money with it. You can sell mining contracts, futures, etc. for Bitcoin just like silver or gold or any other commodity. There doesn't need to be any fancy mechanism in the software itself.  :-)

[cunicula]

A bitcoin now and a bitcoin delivered next year are distinct goods, but perhaps calling this a bond would be better. I will edit the original post to reflect the name change. I will also place a 25 BTC bounty on development of a bitcoin client which offers the 'bond' features I am proposing within one year. Happy to deliver the bounty in advance to Gavin Andresen to judge contributions. If he accepts this responsibility, he would judge contributions to successful development of the proposed bond features or return the BTC to me in one year if they are not successfully developed.

[cunicula]

@ Stefan

Transacting with you (or anyone else) creates default risks that radically drive up the cost of transaction. Negotiating terms with you increases the costs of transactions, evaluating your rep and monitoring changes increases the costs of transaction. Profit you earn from creating one side of a market increases the costs of transactions. Transferring a contract with you to a third party is difficult and this increases the cost of transaction. Transaction costs are so high that at present the bond market is so illiquid that it almost doesn't exist.

My proposal solves all of the above problems in their entirety. Allowing private parties or exchanges to handle this trade is an extremely inefficient alternative. The current lack of a bond market suggests that it is not a viable option.

[relative]

A bitcoin now and a bitcoin delivered next year are distinct goods, but perhaps calling this a bond would be better. I will edit the original post to reflect the name change. I will also place a 25 BTC bounty on development of a bitcoin client which offers the 'bond' features I am proposing within one year. Happy to deliver the bounty in advance to Gavin Andresen to judge contributions. If he accepts this responsibility, he would judge contributions to successful development of the proposed bond features or return the BTC to me in one year if they are not successfully developed.

I'm afraid you dont understand bonds either. bonds are issued by one party and bought by another, as a whole forming a loan agreement. if bitcoin succeeds there will be a bond market.

what has that to do with the monetary system and why do you need to "mine" them?

[cunicula]

Mining the bonds is just an efficient way of organizing the bond market. Establishing this market should appeal to hoarders. This is the core proposal. The semantics are irrelevant and a waste of time to debate.  

Perhaps it will help to think of the bonds as bitcoin-denominated treasury bonds issued by the bitcoin software. Trading current bitcoin for bonds of identical bitcoin value is equivalent to borrowing money using
your bitcoin as collateral.

For example, I would like to hold bitcoin on the chance it is successful in a few years time. Since I plan to hoard, I would like to convert my bitcoin holdings into bonds. I don't want to invest in a company that might rip me off.  A bond-holding agent who needs to buy things with BTC would want to sell their bond holdings to me in exchange for current bitcoins. I can't transact with this agent because the market doesn't exist at present. The proposal would establish exceptionally low cost markets for this transaction. I doubt that a more efficient solution than embedding the market in the bitcoin client exists.

[Stefan Thomas]

Average difficulty for all maturities would be pegged to hold the total number of all types of coins constant according to Satoshi's schedule. Difficulty for each individual maturity would be initially set to the average difficulty and then allowed to float (like difficulty is handled right now). Supply and demand would determine how coin generation was distributed across the various maturities.

You say that the difficulty for each maturity is determined "like difficulty is handled right now" i.e. self-adjusting so generation remains constant in terms of units generated per time period. But if generation of bonds of each maturity is constant, then the last sentence in the quote above makes no sense, the relative amount of each bond issued couldn't fluctuate.

You can either let relative hashing power determine the difficulty for each bond type or the amount of each bond type, but not both.

To do what I think you want, you'd need to find out what people are paying for those bonds. Then you would have enough information to adjust the difficulty such that the "(present value at market) per (unit of work)" is the same for all maturities. The problem is you cannot find out easily and securely how much the bonds are trading for - at least it would get very complex at that point.

Let me know if I misunderstood your proposal.

[relative]

For example, I would like to hold bitcoin on the chance it is successful in a few years time. Since I plan to hoard, I would like to convert my bitcoin holdings into bonds. I don't want to invest in a company that might rip me off.

...and there is your logical fallacy.
it doesnt add to the bitcoin economy if you buy a bond without risk generated by the monetary system.

a paper ("bond") without any risk and with no interest is known as cash, or BTC holdings in this case. why dont you just hold BTC for 2 years?
if you want to convert your cash into capital by buying a bond, there is risk involved, and that is the way it is supposed to be.
riskless bonds is certainly not a goal of a viable monetary system, neither is endless "capital".

the latter point is the main fallacy of our current monetary policy: the assumption that the injection of money creates capital and grows the economy.

[cunicula]

@Stefan. I really appreciate your question. The short answer is that the miners would observe prices on markets and their consequent mining decisions would determine difficulty. Here are some algebraic details of how it might work.

Here is the approximate existing system which provides a starting point and is left basically intact:

A(t) is the actual total coins of all maturities per hour during period t
H is the  target number of aggregate coins of all types per hour. This decreases according to those strange discontinuities introduced in Satoshi's system, but I'll just pretend it is constant for simplicity.
D(t) = aggregate difficulty needed to approximate H at time A(t)

D(t)= [D(t-1)A(t-1)]/H   Note that this is determined without knowing anything about the past difficulties of specific maturity bonds.
    
Now I graft some bonds on to the existing system. Incorporating multiple bond types should be easy.

Ab(t) is the number of coins of one year maturity per hour during period t.
Ac(t) is the number of coins of instant maturity per hour during period t.
A(t) = Ab(t) + Ac(t) (by definition)

Db(t) is the difficulty for coins of one year maturity
Db(t) = [Db(t-1)Ab(t-1)] / D(t)  

Dc(t) is the difficulty for coins of instant maturity
Dc(t) = [Dc(t-1)Ac(t-1)] / D(t)

Rearranging the equations the ratio of the two difficulties is:

Dc(t)/Db(t)= Dc(t-1)Ac(t-1) /  Db(t-1)Ab(t-1)

Bounds on one period maximum upwards and downwards movements in the difficulty ratio would prevent sudden movements across maturities from screwing with the aggregate generation rate A(t).
e.g. Dc(t)/Db(t) = 0.8 [Dc(t-1)/Db(t-1)]           if       [Dc(t-1)Ac(t-1) /  Db(t-1)Ab(t-1)]   <=    0.8 [Dc(t-1)/Db(t-1)]
                       = [Dc(t-1)/Db(t-1)]               if       1.2 [Dc(t-1)/Db(t-1)] >  [Dc(t-1)Ac(t-1) /  Db(t-1)Ab(t-1)]   >    0.8 [Dc(t-1)/Db(t-1)]
                       = 1.2 [Dc(t-1)/Db(t-1)]          if       [Dc(t-1)Ac(t-1) /  Db(t-1)Ab(t-1)]   >    1.2 [Dc(t-1)/Db(t-1)]  
 
Initially the two difficulties should just be seeded as equal, Dc(t)/Db(t)=1. Over time Dc(t)/Db(t) would increase until it reaches a price that clears the mining market (i.e. miners are indifferent between maturities).
In the medium to long-term, Dc(t)/Db(t) would fall (increase) if markets became more (less) confident about the future of bitcoin.
  

[cunicula]

@ relative

Very glad you are not on the FED reserve board. Of course, cash dominates bonds if the bonds offer a zero percent or negative interest rate. This is the FED's 0% lower bound on nominal interest rates that is often referred to in the popular press. Bonds would not sell at a zero percent interest rate because other people (not me) would rather spend their bitcoins now.  

[cunicula]

This doesn't appear to be generating much well-reasoned discussion here. Could someone help me move this to Project Development and Technical Discussion?

[lemonginger]

I'm afraid you won't get much traction there either because forking the chain is a Big Deal and there is nothing you want to do that can't be done with already existing bitcoins and payment agreements.

[Vince Torres]

[edited post to replace term futures with bonds and add bounty]
I think the bitcoin system should allow mining for bitcoin bonds. Bitcoin bonds would be tradable coins with a fixed maturity date.
Upon reaching the maturity date, bitcoin would become identical to regular bitcoins. In my view, this is a much more efficient solution than exchange-based loans that require third parties to hold collateral.

For example, miners would have the option of mining instantly maturing bitcoins or bitcoins maturing on Jan 1st of 2012, 2013, 2014, 2015, 2016. (one to five year maturity)
Average difficulty for all maturities would be pegged to hold the total number of all types of coins constant according to Satoshi's schedule. Difficulty for each individual maturity would be initially set to the average difficulty and then allowed to float (like difficulty is handled right now). Supply and demand would determine how coin generation was distributed across the various maturities. This would smooth discontinuities in the coin generation rate that exist under the current system.

Bonds could be exchanged for regular bitcoins on bankless electronic exchanges, much like the Namecoin/Bitcoin exchange at bitparking. These exchanges have extremely low costs because the exchange does not need to hold collateral for a prolonged period. The exchange rates would indicate bitcoin holders' expectations about the future of the bitcoin economy.

Bonds would give wealthy individuals a stronger incentive to develop the bitcoin economy. For example, suppose Mary holds bitcoins and has a business idea that would add significant value to the bitcoin economy within one to two years. It would be in Mary's interest to sell some of her current bitcoins for bonds and some for USD to finance her idea. Effectively, Mary would be loaning her bitcoins out to raise USD capital with perfect bitcoin collateral. Mary would internalize some benefit from any bitcoin appreciation resulting from her business. If Mary just held USD, she might want to purchase bitcoin futures for this reason as well. This mechanism could encourage larger companies to enter the bitcoin economy because value-adding companies would earn more from purchasing bitcoin bonds rather than regular bitcoin.

Many will object to revision of the bitcoin generation process, but I think this is misguided. Whether the bitcoin economy will be valuable several years from now depends on the currency remaining more innovative than potential competing currencies. The emergence of a competitor is a much more important concern than whether bitcoin will be inflated through revised currency generation. Bitcoin users wouldn't accept an inflationary revised currency generation anyway.

Offering 25 BTC bounty for this, see post below.

Just go ahead and write code for this.

[cunicula]

I would be happy if someone could come up with a technical solution that didn't involve mining and forking the block chain. However, using a contract where a third party holds collateral does not accomplish what I want done. The idea is to reduce the cost of contracting and make bitcoin a more efficient medium of exchange.  This should be the primary goal of bitcoin developers.

I think the main problem is that the post readers aren't sufficiently intelligent and/or ignorant of economics.  I understand that almost no one here understands economics well, but the computer scientists are at least intelligent. As far as me writing the code, I am an economics professor, not a computer programmer.

[MeSarah]

Ive been thinking about something like this but from a different angle. I havent thought it all the way through but here are my ideas. One of the problems I see with btc is the whole decimal issue. From my prespective 0.000003 or 0.0005 is error prone. I also belive that it has an impeding effect on the adoption of btc.

So here is my unthought through idea. We add multiple coin types to the block chain. For example, if btc is our fixed base of 1 regarding quantity and dificulty, we might add a coin type that is 10 times the quantity of btc and 1/10 of the dificulty. If the orginal btc is the gold standard we might call this the silver btc. We might want to add a copper btc as well. That might look like 100 times the quantity of silver and 1/100 the dificulty of silver. We could go furth and add an aluminum btc. It might look like 1000 times the quantity and 1/1000 of the dificulty of copper btc. I think it would be important that these different coin types live in the same eco-system.

The key would be that they draw reference from the same dificulty, that of the orginal btc. With all coin types using the same reference dificulty it would add a measure of stablity between the coin types. I believe this would bring back cpu miners. That would broaden the user base. I believe it would also keep gpu miners mining gold btc instead of the aluminum because they would in essence be making the same but in a more mature market. But of course the market would decide the value of each coin type. If it becomes more profitible to mine aluminum btc then gold btc miners would transition. This would decrease quanities of new gold btc eventually making it more profitible to mine gold btc again. The affect of adding these new coin types would in essence increase the total btc quanity.

Well, there it is. Please dont be to harsh on me.

[cunicula]

@MeSarah
The redomination you suggest (which I think is a good one though though USD like terms might be better than WOW currency like terms) could be achieved by developing an improved GUI. I think that developers will do this in the near future. I don't think that creating new currencies from scratch is a good idea, however. For the reasons described below:

Eventually a digital currency will emerge (call it Bettercoin) which allows for more p2p contracting possibilities than are currently possible with Bitcoin. If bitcoin is widely used before Bettercoin emerges, than bitcoin will likely remain dominant despite its manifest technological inferiority (think QWERTY keyboards vs. superior alternative key arrangements, or Paypal vs. Bitcoin, or Windows vs. Linux). This is what current holders of bitcoin would like to see happen. If Bettercoin emerges before bitcoin becomes widely used, Bettercoin will gather a larger user base and bitcoin will collapse in value. Bitcoin holders would not be happy about that. The issue with creating new currencies from scratch (ones that don't grandfather in existing holdings) is that these currencies would fragment bitcoin in to several smaller interest groups each of which would like to see their currency take off. Collectively, the currencies would lose value due to uncertainty over which one would eventually take over. Bettercoin would be better positioned to step in and take over the market, and existing asset holders would be wiped out. 

I am doubtful that they will pursue options like the one I am suggesting precisely because of this fragmentation issue, but think that this is misguided. What I am suggesting would require significant modifications to the underlying currency generation rules. Most likely it would entail a splitting bitcoin into two currencies. However, provided that
1) one branch had superior features (for example was faster, more secure, or allowed for more contracting possibilities) 
2) this branch protected the value of incumbent bitcoin holders by grandfathering in their holdings within a certain time frame.
Then everyone would rapidly convert to the superior currency branch and fragmentation would be short-lived. I am worried that a reluctance to experiment with more radical modifications will leave bitcoin in a weaker position relative to future competitors. If everyone is too fixated on preserving Satoshi's original plan, then bitcoin will be unable to adopt features offered by superior p2p platforms. If bitcoin experiments with new features now, then bitcoin will be ready to copy features offered by competitors, effectively killing them off before they became threatening.

[ByteCoin]

...
For example, miners would have the option of mining instantly maturing bitcoins or bitcoins maturing on Jan 1st of 2012, 2013, 2014, 2015, 2016. (one to five year maturity)
Average difficulty for all maturities would be pegged to hold the total number of all types of coins constant according to Satoshi's schedule.
...

So, given that the client instantly converts bitcoin bonds on maturity into bitcoins and given that the number of bitcoins in the economy meets a fairly strict schedule, in the run up to the maturity of one set of bonds the bitcoin client would have to know how many bonds were going to mature and adjust the difficulty of the normal block generation so that after they mature, the right number of bitcoins is in circulation. This would probably mess up the block generation rate in the time before maturity and so prevent this implementation of bonds. Block generation provides timely transaction confirmation and anything which harms this cannot be tolerated.

ByteCoin