avatar

A Challenge to the Bitcoin Community

by George Selgin May 2nd, 2013 3:13 pm

Well, it seems that Freebanking.org may be having its own little Bitcoin bubble, what with all the recent posting on the topic by Brad, Kurt, and myself. Still I can't help adding one more, having planned to do so since my last post on the topic, and having as well a selfish motive in mind, to wit, that of supplying myself with some materials with which to revise and expand my paper on "Synthetic Commodity Money."

I wish to accomplish that end by challenging Bitcoin fans to propose means by which one might combine the same advantages as Bitcoin presently offers with a mining or production protocol that, instead of adjusting mining rewards so as to achieve some predetermined output rate and limit, adjusts them so as to automatically alter the rate of coin production as needed to achieve some other objective. For example, the objectives might be that of achieving a steady rate of cybercurrency appreciation (as measured by the cybercurrency-equivalent CPI), or a steady level or growth rate for the total value of cybercurrency payments. But there are many others also worth contemplating. The general question is whether it is possible to have a more "elastic" or otherwise macroeconomically friendly alternative to Bitcoin. (Even the output of gold, after all, tended in the long run to respond to changes in that metal's relative price, thereby ruling-out both persistent inflation and persistent (or sharp) deflation.)

In putting forward this challenge I don't mean to endorse any particular monetary or macroeconomic policy ideal. I'm only interested in exploring the extent to which it may be possible to have mining protocols other than those calling for a steady (or steadily declining) rate of coin output. If there are, then the question whether any of them would in fact supply the basis for a "better" cybercurrency is one that would then be worth taking up.

As food for thought to those pondering the challenge, here are a few of the very interesting suggestions I received concerning it in the comments to my original Bitcoin post:

(Peter Surda): you cannot design a synthetic commodity whose supply mirrors macroeconomic aggregates, because these aggregate variables are exogenous to the network (whereas the hashing parameter of networks like Bitcoin only depends on time and the number of blocks in the blockchain, which from a perspective of the network are endogenous), so they cannot be unambigiously measured.

(Derka): Yes you can certainly base altcoin production (supply inflation) on any parameter (nominal or real GDP, employment, etc.), the tricky part is doing it without centralizing trust. For the sake of demonstration, call the parameter of interest X.

Option 1) Centralize authority. Give the WSJ authority (via public/private key encryption) to report the value of X to the altcoin network. The altcoin network responds algorithmically to this news. Perhaps if X is NGDP and is down, increase production!

Option 2) Give the authority to the miners. Every miner could record the current value of X in the blocks they produce. Problematically, the miners will collude and devalue the currency to reap the hyperinflated rewards.

(Koen): but would it not be possible for the network to respond to patterns that occur in the network but that are reliably associated with (e.g. being the effect of or having the same cause as) certain (macro-)economic phenomena? I mean, for example, contractions in the money supply seem to have both a (macro-)economic reality outside of the network and specific effects within the network

(Justin Reitz): Off the cuff, it might be possible to use the Federal Reserve's FRED API to automatically pull data on a chosen economic aggregate. Of course there are all sort of issues: timeliness of Fed data, data updates, security of the the FRED API, how to reduce the virtual currency in circulation, etc.

(tdbtdb): Perhaps this is what you mean when you mention the security of FRED, but clearly having such a system depend on FRED would put it under the control of whoever controls FRED, nagating the whole point. So there are 2 risks, one that unauthorized hackers could compromise FRED, the other that the people feeding data to FRED might feel tempted to lie.

I wish I could offer some real prize for more thoughts like those above, let alone for any more substantial answer to my challenge. But as I am no fatcat but only a humble college professor, I can offer nothing save my gratitude, and the modest reward of doing whatever I can to credit those responsible for their contributions to one component of what H. G. Wells (himself no mean futurist) considered the "permanently effective task before Mankind," namely, that of "working out and applying...a Science of Currency."

35 Responses to “A Challenge to the Bitcoin Community”

  1. avatar Thrica says:

    In theory I could imagine a protocol tied to a different computational problem with (practically) infinite but unpredictable solutions, such that mining and transaction are a single activity. Easiest way of distributing would be lottery-like, so each transacted coin has a certain probability of generating another for the spender (which would have the added benefit of encouraging use instead of hoarding, as Bitcoin seems to suffer). This way the supply would increase with its velocity.

    Unfortunately it wouldn't adjust back downward. This seems like it would be the biggest difficulty with creating a cryptocurrency which adjusts itself without the aid of a banking structure: you can't put the mathematical genie back in the bottle, so you'd never be able to contract the base money supply.

  2. avatar alanx says:

    Since George cannot offer a prize, I will offer one for him. I will give $1 in Bitcoin (per the average price on MtGox on that date) to whoever he declares to be the winner. This only requires those that might wish to receive the prize to publish a Bitcoin address. (Generate one for yourself here for free: http://brainwallet.org/)

  3. avatar MichaelM says:

    I don't know if this is a resolvable challenge. As Peter Surda and Derka point out, a single bitcoin-like digital currency would require some kind of input in order to respond to outside conditions. This input would have to come from somewhere, so you're opening a vulnerability in the system that sacrifices the very qualities that make it attractive. The only reason it works with bitcoin is because the input is itself internal, an exercise in the most literal kind of navel gazing.

    Getting away from what other people have already had to say, it sounds to me like digital currencies would be better served by continuing to act like bit-coin, but becoming plural. That is, as the supply of a particular bitcoin like currency becomes more and more disconnected from the demand to hold it, new bitcoin-like currencies are allowed to pop up and take up the slack. Digital currencies would then take a role similar to Hayek's concept of denationalized fiat currencies without the price stabilizing central management.

  4. avatar alanx says:

    With a Crypto Currency, recreating the Federal Reserve is a piece of cake. Let's call our monster "USCoin".

    USCoin would have two possible sources, FedCoin and BankCoin.

    1) Transaction fees + a Fed set rate of reward in "Fedcoin" is paid to registered miners.
    2) Fedcoin is generated into a set of special "FedGen" accounts.
    3) FedGen accounts can fund loans to banks.
    4) Banks pay back FedGen loans to two accounts (set up for each loan). Principal is paid to a Principle account (and the funds destroyed). Interest to an Interest account (where funds are used to run the Federal Reserve).
    5) FedGen accounts can buy Treasuries.
    6) Treasuries are redeemed in the same way as bank loans. Principle is destroyed, and interest funds the Federal Reserve.
    7) FedGen accounts can be borrowed by the Federal Reserve itself, as they are doing with QE3 to buy securities.
    8) Any "self loans" would require the Principle to be paid back (and destroyed) and any shortfall made up with other funds.

    9) Banks would be allowed to loan against their base by generating FedCoin.
    10) Bank issued FedCoin could only be issued up to a percentage of the FedCoin held in reserve by the Bank, where that percentage is set by the Federal Reserve.
    12) Banks would be required to destroy principle and collect interest via the same rules as dictated for the Federal Reserve.

    Advantages:
    The Federal Reserve and Banks would be a snap to audit. If you require the Federal Reserve to only send funds to publicly labeled accounts, then all funds could be tracked at all the way from the Federal Reserve and through the banks. From that point, the funds would enter the wild, and be rather difficult to track (as Bitcoin is).

    Summary:
    You can absolutely recreate the Federal Reserve with Crypto Currency. I might have a few details wrong in my list above, but with a little effort they can be made to exactly model what we do with Dollars today.

    Problem:
    If FedCoin is as easily transacted as Bitcoin (and that is one of the assumptions above), then Bitcoin will gain in value against FedCoin over time. FedCoin's value as a currency would suffer from trivial conversion between FedCoin and Bitcoin.

    • avatar alanx says:

      Opps. Notation error: I debated distingishing between FedCoin and BankCoin (issued by banks) which collectively would be called "USCoin". But thinking it through decided that was useless. I can't edit, so just consider USCoin == to FedCoin.

    • avatar alanx says:

      Oh, and my Bitcoin Address: 1YvPtWNM5X59AfXMGf77wJYTpL4GzoWCH

      Just incase you decide I should pay myself :P

  5. avatar predius says:

    I think people are looking at this from the wrong angle. We're still thinking "one country = one coin" . Bitcoin will certainly not be the last cryptocurrency to gain acceptance in society, it is (already) the first of many.

    The problem is looking at the system as "an economy" with a "currency". We can't equate Bitcoin or Litecoin or PPcoin or any or the thousands that will follow to the Euro or the Sterling. When the cost of creating a currency which everyone can trust is (almost) 0, then the odds are that there will be more currencies than before 2009, since you pretty much needed to be a state and have control over a territority and an "economy" to have a currency. The money stock will be not dependent on how many Bitcoins (one of many) are created, but the market need for currencies and the subsequent market (and other meaningful human) processes that produce these currencies.

    These may look a lot like Bitcoin or very little, but we are about to find out.

  6. avatar michaelsuede says:

    I am a professional software developer, and I can tell you that virtually anything you can conceive of with software is possible to build. It is simply a matter of time and resources to construct it. Where there is a will, there is a way. Information can be manipulated in an infinite number of ways. If an idea is big enough, entire operating systems, network protocols, and hardware infrastructures can be created to support the end goal. Nothing is impossible. Perhaps infeasible, but not impossible.

    That said, I'm not convinced that Bitcoin is macroeconomically unfriendly. I view the supply fluctuation in gold, and the constant increase in supply, as a problem for the gold standard, not some kind of benefit. Those are reasons why people would want to hold Bitcoin over gold, rather than vice versa.

    The higher the price for Bitcoins becomes, the more people are going to want a piece of the action. Since Bitcoin is infinitely divisible, I don't see an issue here at all.

    • avatar Martin Brock says:

      People holding Bitcoin rather than gold, expecting Bitcoin to appreciate faster, are not using Bitcoin as money. People holding Bitcoin rather than dollars, expecting Bitcoin to appreciate faster, are using dollars as money instead of Bitcoin.

      The problem with preferring dollars as money to Bitcoin as money is that Bitcoin has persistent value only as money.

      • avatar michaelsuede says:

        I'm not sure what that has to do with my argument.

        • avatar Martin Brock says:

          You say that Bitcoin's less elastic supply, compared with gold, makes people more eager to hold Bitcoins compared with gold. At some level, I think you're right, but this preference for holding Bitcoin over accepting Bitcoin only to spend it in the short term is not a preference for Bitcoin as money, and Bitcoin ultimately has value if people prefer it as money.

          People already hold gold speculatively and as an inflation edge, but gold is not money for this reason. That holding gold this way does not make it money is not a problem for gold, because gold has many non-monetary uses. Bitcoin has no non-monetary use, or it's only non-monetary use is as vehicle for speculation. A vehicle for speculation with no other use is a Ponzi scheme.

          Bitcoin is not a Ponzi scheme if people use it as money, but the more people use it as money, the more valuable it becomes, and the more valuable it becomes, the more people hold it speculatively, and the more people hold it speculatively, the less people use it as money.

          Bitcoin 1.0 has two modes, the money mode and the Ponzi scheme mode, and I expect it to shift back and forth between these modes until the instability makes it so unattractive as money that it finally settles permanently into the Ponzi scheme mode.

          A Bitcoin 2.0 incorporating existing Bitcoins into a more elastic supply of Bitcoins can avoid this outcome, and George seems to advocate this course.

          • avatar alanx says:

            One of the functions of money is to store value for the future. If the value of money rises, encouraging people to save, then how is this not a function of money?

            If more people were holding Bitcoin, and merchants were accepting Bitcoin, then some level of savings would top off, and people would spend for goods that they want more than to save money.

            The idea that rising values of currency, and the falling prices of goods would cut off all spending is refuted by the fantastic drops in prices of technology over the last 40 years, which just hasn't seen a choking off effect. Or perhaps a healthy market just takes that in stride, given that some do wait to make purchases knowing features will go up, and bugs and prices will go down over time.

          • avatar alanx says:

            Furthermore, if saving (or speculating) is a ponzi scheme, what are capital markets? And what are banks who make loans against money they don't really have?

            And what is the use of money created from nothing to buy securities or treasuries (like the Fed is doing now) in competition with people that would like to buy those securities with money they had to work to earn?

            Hint: If saving is "Ponzi scheming, speculating, or hoarding" then it is completely fair to call expanding the money supply by the Fed "counterfeiting".

            The negative terms used to describe understandable choices people make doesn't prove the value of various monetary polices.

          • avatar Martin Brock says:

            @alanx

            A medium of exchange need only store value for a short time. A good held in expectation of appreciation is not money. It's an investment. Conflating money with an investment can only blind you to the many goods that people may use as money.

            People may hold Bitcoin and exchange something else for goods they want from merchants, and if they expect Bitcoin to appreciate more than another good that they possess and merchants accept, they will hold Bitcoin and use the other good as money. Won't they? If you permit yourself to answer this question, you'll understand the point.

            Technology with a falling price is not money, so its falling price refutes nothing about money. The issue is what people with options will use as money. Nothing I've said here implies that technology with a falling price should choke off either the supply of or the demand for the technology.

            A rising value of Bitcoin doesn't choke off demand for Bitcoin. It only chokes off Bitcoin's use as money, because Bitcoin's use as an investment competes directly with its use as money, and the supply of Bitcoin does not respond elastically to demand for its use as money. The problem is that Bitcoin has no other use.

          • avatar Martin Brock says:

            @alanz

            Saving is not a Ponzi scheme. I save by accumulating title to stocks, bonds, real estate, commodities and other durable goods reasonably expected to appreciate in value or otherwise offering a return on my investment.

            Banks offering promissory notes backed by valuable collateral (including the labor of borrowers) as loans of money that borrowers use to purchase goods can be profitable business models, and I have nothing against them. The problem with fiat money is not that banks operate this way. The problem is that a state imposes debts on its subjects by fiat.

            "Ponzi scheme" is not a simple pejorative.

  7. avatar Rob R. says:

    One way to achieve nominal target in a digitized currency would be to use transactions fees/subsidies.

    For example: If you wished to stabilize aggregate spending in the currency you could have a variable transaction fee. If aggregate spending was ahead of target you increase the transaction fee to slow things down. If you are behind target you reduce the fees and make them negative if necessary until you hit the target.

    For a growing currency targeting aggregate spending might not make much sense and you might want to stabilize value against a bundle of goods. I think transaction fees could also be used for this. Increasing transaction fees would reduce the currencies value and negative transaction fees would boost its value. Unlike targeting aggregate spending the feeds for stabilizing value would be exogenous.

    • avatar Rob R. says:

      Correction: "Increasing transaction fees would INCREASE the currencies value and negative transaction fees would REDUCE its value."

  8. avatar jmh530 says:

    I think there's substantial room for designing a better bitcoin (altcoin). The point raised by Peter Surda in your blog post is rather important, though I'm not sure it's entirely for the same reasons he thinks it is. Consider an altcoin that at the end of every day sets Mt=Mt-1Vt-1/Vt. In practice, you would need to estimate velocity Vt every day using the actual money, prices, and transactions of altcoin (i.e. not of the U.S. market more generally). This may not be in real-time, but it would still respond to the altcoin's aggregates. However, this requires some central body where all the transactions would need to be reported to. By contrast, bitcoin does not require any information about how many transactions take place, nor their size. That's a far more decentralized approach.

    It seems preferable to have a free banking framework where incentives of private actors drive money creation rather than a formula set by a central authority. There could be scope for bitcoins in such a system, in the same way that gold or U.S. dollars can serve as the monetary base.

  9. avatar JimM47 says:

    The obvious mechanism for a commodity (synthetic or otherwise) to respond to market conditions is through effort spent producing more of the commodity. I believe it was a paper by Prof. Selgin that pointed out that during the classical gold standard, changes in the price level would affect the economic feasibility of mining additional gold, in a way that was counter-cyclical. And analogously, a paper co-authored by Prof. White points out that the supply of counterfeited Somali banknotes appears to follow the economic feasibility of printing them at a profit.

    As I understand it, it is already the case that BitCoin "mining" follows a similar logic. At the current exchange value of BitCoins, and the current complexity of the computations required for mining, specific hardware is required to be able to mine BitCoins without costing more in electricity that is generated. But if the value of BitCoins goes up while electricity stays the same, then more mining activity is feasible.

    As I understand it, however, the *difference* between BitCoin mining and gold mining is that extra gold mining activity causes more gold to be mined, whereas extra BitCoin mining does NOT cause more BitCoins to be mined. So there isn't the same feedback mechanism. (Someone correct me if I am wrong about this.)

    But for the sake of a thought experiment, simply imagine that the amount of BitCoins produced were to be proportional to the amount of mining. Then, just as historically the supply of gold was partly a function of the value of gold, BitCoins would also increase in supply as their exchange-trade value went up, acting to stabilize their value. Changes in the supply of BitCoins would be a function of two exogenous variables -- the price level in BitCoins, and the cost and computational power of hardware -- and one endogenous network variable -- and the computational difficulty of mining.

    The very obvious problem that presents itself is that the cost and computational power of hardware will mostly reflect the state of engineering, and it is difficult to anticipate the future state of hardware when designing the rules of a system. Imagine if the production of BitCoins was tied to computational work done while mining and there was suddenly a huge leap in computational power (quantum computing, perhaps?). The result would be massive amounts of mining, and hyperinflation until a new equilibrium was produced -- the only saving grace of this hyperinflation would be that if it was caused by a one-time jump in computing power, you could probably determine from the start what the new equilibrium would eventually be and how quickly it would be reached. Even without a massive jump, continued engineering advances might still create a predictable level of inflation.

    But let's say you come up with a mechanism to make the computational complexity of mining and the computational power of hardware move in tandem. What you now have is a supply of a synthetic commodity that should act to stabilize the nominal cost (in terms of the commodity) of its non-synthetic inputs -- cost of electricity, cost of hardware, cost of some amount of technical support. So the question is going to be: how representative is that small basket of goods going to be of the larger basket of goods in the economy?

    Potentially it could be quite representative, but ONLY IF the state of automation advances to the point that the cost of labor is a very small part of most products (as the cost of labor would be a very small part of BitCoin mining).

    • avatar JimM47 says:

      It seems to me that, if the currency were especially widespread, in the hypothetical I imagine, you might also be able to "set an interest rate" by causing some mining activity to produce BitCoins in the future, rather than the present, at some set premium.

      • avatar JimM47 says:

        ... or, conversely, by offering miners a choice of times/premiums, the system might be able to determine interest rates without relying on a centralized data source.

  10. avatar Peter Šurda says:

    Dear professor Selgin,

    earlier today I found that you'll be having a lecture in Bratislava on the 13th, so I registered and if you'd like we can discuss the topic there.

  11. avatar Martin Brock says:

    Ultimately, money is a token accepted in exchange for a good that the recipient expects in turn to exchange for another good of equivalent value, so if prices are to be stable, the supply of these tokens must reflect people's demand for this sort of indirect exchange. A note credibly promising a widely consumed, standard good with an elastic supply seems a good choice.

    To stabilize the price of a standard basket of goods (like the goods in the CPI), the note might as well promise the basket of goods, so maybe I can take the note to WalMart and buy a week's worth of staple groceries, so many gallons of WalMart brand milk, so many pounds of flour and so on. Everyone doesn't want everything in this basket, but I can use everything in the basket somehow, if only by trading with my neighbors, so if I can trust the promise, the notes will do as money.

    This sort of money makes more sense to me than gold or bitcoins or notes promising either gold or bitcoins, and I'd trust WalMart with its issuance. WalMart already issues discount coupons which aren't so different. If WalMart might issue too many notes as the notes become generally accepted as money, permitting competing grocers to issue notes promising a comparable basket of goods seems to address this problem.

    I'm not sure why this doesn't happen, but I suppose the Congress has something to do with it.

  12. avatar JoelKatz says:

    It just comes down to who, or what, you trust to produce the index and introduce it into the system.

    With a Bitcoin-like system, the only way to agree on something is to include it in a mined block. The miners have a combined interest in devaluing existing currency. So likely you'd have to change it into something more based on proof of stake or combined in some way to balance the competing interests.

    Those who hold Bitcoins have an interest in assuring there's sufficient mining to keep the currency secure. And miners already have an incentive to include transactions that include fees. One can leverage this as follows:

    1) Bitcoin transactions can include a desired block reward changes.

    2) Bitcoin blocks can track the "votes" on the changes in the desired block reward.

    3) Vote can be weighed based on Bitcoin days destroyed to act as a proof of stake system.

    4) A miner can be bribed to include a block that includes a block reward vote he disagrees with just by ensuring the transaction fee is adequate.

    5) Part of the block validity test can be to ensure the weighted voting information is correctly updated based on votes in the block.

    6) Every so many blocks, perhaps at the same time the difficulty is changed, the reward would also be changed based on the votes in past blocks.

    This balances power between miners and Bitcoin holders and the voting algorithm can be adjusted to avoid rapid changes. All it would require was some extra fields in the block which could be held in a pseudo-transaction.

  13. avatar Eitan says:

    The way I imagine a digital monetary system with elastic supply working would be something like the idealized way gold with free banking works. You have a base currency with a predictable supply (say increasing at a decreasing rate like bitcoin). Then you allow people to issue their own redeemable substitutes with no built-in supply restriction at all, sharing the same address space as the base currency. The issuers reserve holdings and how much they issue would be publicly viewable in the ledger. Individual addresses could choose to accept or not the substitutes of any particular issuer, based on their reputation and publicly viewable reserve ratio. The issuers would want to incentivize this, probably by offering interest payments to people who hold their issue. If a user doesn't accept a substitute, then it automatically withdraws from the issuer's reserves instead, returning the substitutes to the issuer, thus acting as redemption. Also, issuers could accept or not the substitutes of other issuers and contract on how to run clearing transactions between them, in order to economize on reserve utilization. And the system would have a default procedure for issuers, automatically giving the reserves back to substitute holders in proportion to their holdings.

    • avatar Martin Brock says:

      Notes promising a standard of value can substitute for the standard in trade, and the supply of these notes can be more elastic than the supply of the standard; however, the elasticity of the supply of notes seems problematic when the supply of the standard is highly inelastic (like the supply of bitcoins).

      When the standard becomes scarce and thus has a rising value relative to the notes, everyone holding the standard has the same incentive to hoard it at the same time, so scarcity of the standard on the market has a feedback creating greater scarcity. Stability requires the opposite feedback in this scenario.

    • avatar JP Koning says:

      I like Eitan's point.

      I'm not sure how one would make the base quantity of bitcoin elastic. The elasticity of central bank money derives from the fact that when it is no longer wanted, it can be sucked back in by the issuer in return for redemption media (gold, bonds, whatever). Thus the quantity of Fed dollars contracts so that prices don't have to rise. Bitcoin has no issuer so when the demand for bitcoin is reduced, it can't be sucked back in.

  14. avatar apetersson says:

    ok. what you are looking for IS possible. i don't have time to exapand fully on this. but here is a hint:

    https://en.bitcoin.it/wiki/Contracts#Example_4:_Using_external_state

    "I have discovered a truly marvelous proof of this, which this margin is too narrow to contain."

  15. avatar Wesleybruce says:

    George since gold tracks with many basic commodities over the long run simply being able to buy and sell gold with bitcoin may be sufficient. If there was a market where you could short bitcoins etc that would also smooth the bumps. Several scholars have looked at the value of transactions recorded in Roman and older texts. Often the same weight of gold would get you the same quantity and quality of goods as the same weight of gold would get you today. Inflation is relative to gold.

    In Leviticus 5: 14 the bible requires offerings to be made and includes a weight of silver or a quantity of grain flour. This had the effect of locking the two together. When one was low it was used in the temple pushing its demand up and other demand down to balance the two. If payment were promised or required in either bitcoin or something else then it would have the same effect as a basket of goods. It means if bitcoin or a ton of beans were promised and beans were cheaper then it would be cheaper to buy beans and pay with that. The two would lock together. The catch is most commodities are not emailable. Your back to the thing Satoshi was trying to eliminate. Counter party risk. The need to trust someone.

    Escrow provisions are an option in the bitcoin code. Its not activated. It may be possible to create a short selling operation in the code its self and smooth out bitcoin. Where people pay extra conditional escrowed satoshi's for a longer escrow period on a pair of swapped bitcoins, a differential delay in the transaction clearing equalling a loan in a short/ long transaction. Essentially three transactions set to mimic the operations a pair of short and long transactions. The conditional transaction creates the profit to one and the risk of the other and may reference other goods in some cases. We may need a fourth or so and a third player in the trade to make it work but that's possible.

  16. avatar Onar Åm says:

    As already pointed out, a cryptocurrency like this cannot stabilize itself endogenously. It needs output from the outside, some kind of measure of stability. Second, it actually needs to be able to interact with the outside monetary world, i.e. purchase things. If it is able to do that then it is possible to have a managed currency (like any other) and link it to whatever desired measure of stable value that you wish. One obvious choice is putting the cryptocurrency on the gold standard and linking it to the price of gold.

    In order to achieve this you need the following:

    1) the cryptocyrrency protocol must be the legal owner of bank accounts with money in it (dollars, yen, gold, whatever)

    2) the cryptocyrrency protocol must have access to monetary exchanges where both bitcoin and other currencies are traded

    3) built into the protocol must be a mechanism for automatically creating new coins and selling them for other currencies whenever the coin appreciates in value compared to the benchmark (e.g. gold)

    4) the protocol must also automatically buy up and destroy coins whenever the value of the coin depreciates.

    If implemented correctly it will behave like an automated managed p2p cryptocurrency.

    Barring this scheme bitcoin provides the very best alternative.

  17. avatar Onar Åm says:

    I forgot to mention that no change is needed in bitcoin in order to put it on a gold standard (once it has seen widespread adoption and the price of bitcoin is say, 100,000 dollars per btc). All that is needed is for some non-profit organization to set up a btc and currency fund for the purpose of managing the price of BTC so that it correlates with gold (or some other measure of stability).

  18. avatar EricD says:

    I think there is a straightforward way to define a bitcoin variant that automatically stabilizes one macro variable in particular: total transaction volume.

    First a digression. As Eitan mentions above, bitcoin can be used as a reserve asset, held by bitcoin banks--an asset against which these banks could issue redeemable-on-demand claims, i.e. fractional reserve bank money. But there is no need to integrate a technology for the handling of such bank money into the bitcoin client/code itself. Indeed there is already a similar but separate technology, complementary to bitcoin, that can support digital/crypto bank money: Open Transactions https://en.bitcoin.it/wiki/Open_Transactions . One nice aspect of bitcoin in particular as a reserve asset is that one could digitally verify a bank’s claim to possess a given quantity of bitcoin reserves, without inspecting a physical vault or relying on a trusted third party.

    Under the scenario in which the medium of exchange is primarily bitcoin-backed bank money rather than bitcoins themselves, it is not clear how to meet Selgin’s challenge (creating a bitcoin variant that automatically stabilizes a desired macro variable, as against the total coin supply). But the challenge itself would be less important owing to elasticity in the supply of bank money.

    However the possibility still needs to be considered that people won’t deposit their coins into banks, and so Selgin’s challenge remains relevant. In the case of an arbitrary macro variable, I agree with Peter Surda, there’s no clean way to integrate its measurement into the bitcoin technology itself. This gets to the fact that macro variables are inherently hard to measure, with subtle judgements to be made that ought not be frozen into the monetary infrastructure.

    But automatic stabilization should be possible for one special macro variable--indeed the one presumably of most interest to free bankers: the total (say daily) volume of transactions. This is because each node in the bitcoin network has a full copy of the “blockchain,” the list of all verified bitcoin transactions that ever occurred, from which one can infer the entire history of total daily transaction volumes.

    Let me spell it out, because it seems pretty straightforward to me and I’m confused why it hasn’t been mentioned by previous commenters. In bitcoin itself nodes accept newly “mined” blocks into the blockchain by (among other things) verifying that the miner has solved a computational problem designed to be sufficiently difficult that the total supply of bitcoins will continue to grow along its predetermined path (asymptoting to 21 million). But suppose the threshold for computational difficulty were adjusted according to the total transaction volume revealed by the blockchain rather than to the total bitcoin supply. Bitcoin is set up now to, e.g., increase the difficulty when the blockchain indicates that it’s taking fewer than ten minutes to generate each new block (averaged over the last two weeks). Instead we can increase the difficulty when the total volume of coins traded (not the total number of blocks produced) exceeds some target volume. Just like the target supply for bitcoin, the target volume in this variant may be set as a predetermined function of time.

    Since, unlike the total supply of coins, the transaction volume depends on other things beside how many new blocks are accepted into the blockchain, more care should be taken in setting the feedback behavior between realized transaction volume and block acceptance. But this is a more generic macro problem about the transmission mechanism by which additional base money fuels more spending.

    It’s interesting that transaction volume (i.e. total spending) is the one macro measure that can be cleanly and automatically stabilized by a bitcoin variant. It is total spending that Selgin originally argued to be stabilized under free banking, rather than, e.g., inflation or the short rate. Indeed it is NGDP (spending on final goods) that serves as a proxy for total spending in Selgin’s rationale for NGDP targeting, rather than vice versa. Stabilizing macro targets like inflation or the short rate, or even NGDP, would require information about the goods side of a subset of transactions, while stabilizing total spending requires only money-side information.

    A technical problem with this proposal for stabilizing transaction volume arises if each node ceases to hold the entire blockchain in some future version of bitcoin, due to network scalability issues. Still, I think, this is easily solved by a technical modification to block syntax. Each block currently includes a hash of the block it’s trying to build on top of. If we just append to this hash one number, the running total transaction volume, then a block only needs to see the last block on the chain in order that this running total is maintained and can be used to trigger updates to the computational difficulty setting.

  19. avatar EricD says:

    Resend (original submission date 5/6/2013):

    I think there is a straightforward way to define a bitcoin variant that automatically stabilizes one macro variable in particular: total transaction volume.

    First a digression. As Eitan mentions above, bitcoin can be used as a reserve asset, held by bitcoin banks--an asset against which these banks could issue redeemable-on-demand claims, i.e. fractional reserve bank money. But there is no need to integrate a technology for the handling of such bank money into the bitcoin client/code itself. Indeed there is already a similar but separate technology, complementary to bitcoin, that can support digital/crypto bank money: Open Transactions https://en.bitcoin.it/wiki/Open_Transactions . One nice aspect of bitcoin in particular as a reserve asset is that one could digitally verify a bank’s claim to possess a given quantity of bitcoin reserves, without inspecting a physical vault or relying on a trusted third party.

    Under the scenario in which the medium of exchange is primarily bitcoin-backed bank money rather than bitcoins themselves, it is not clear how to meet Selgin’s challenge (creating a bitcoin variant that automatically stabilizes a desired macro variable, as against the total coin supply). But the challenge itself would be less important owing to elasticity in the supply of bank money.

    However the possibility still needs to be considered that people won’t deposit their coins into banks, and so Selgin’s challenge remains relevant. In the case of an arbitrary macro variable, I agree with Peter Surda, there’s no clean way to integrate its measurement into the bitcoin technology itself. This gets to the fact that macro variables are inherently hard to measure, with subtle judgements to be made that ought not be frozen into the monetary infrastructure.

    But automatic stabilization should be possible for one special macro variable--indeed the one presumably of most interest to free bankers: the total (say daily) volume of transactions. This is because each node in the bitcoin network has a full copy of the “blockchain,” the list of all verified bitcoin transactions that ever occurred, from which one can infer the entire history of total daily transaction volumes.

    Let me spell it out, because it seems pretty straightforward to me and I’m confused why it hasn’t been mentioned by previous commenters. In bitcoin itself nodes accept newly “mined” blocks into the blockchain by (among other things) verifying that the miner has solved a computational problem designed to be sufficiently difficult that the total supply of bitcoins will continue to grow along its predetermined path (asymptoting to 21 million). But suppose the threshold for computational difficulty were adjusted according to the total transaction volume revealed by the blockchain rather than to the total bitcoin supply. Bitcoin is set up now to, e.g., increase the difficulty when the blockchain indicates that it’s taking fewer than ten minutes to generate each new block (averaged over the last two weeks). Instead we can increase the difficulty when the total volume of coins traded (not the total number of blocks produced) exceeds some target volume. Just like the target supply for bitcoin, the target volume in this variant may be set as a predetermined function of time.

    Since, unlike the total supply of coins, the transaction volume depends on other things beside how many new blocks are accepted into the blockchain, more care should be taken in setting the feedback behavior between realized transaction volume and block acceptance. But this is a more generic macro problem about the transmission mechanism by which additional base money fuels more spending.

    It’s interesting that transaction volume (i.e. total spending) is the one macro measure that can be cleanly and automatically stabilized by a bitcoin variant. It is total spending that Selgin originally argued to be stabilized under free banking, rather than, e.g., inflation or the short rate. Indeed it is NGDP (spending on final goods) that serves as a proxy for total spending in Selgin’s rationale for NGDP targeting, rather than vice versa. Stabilizing macro targets like inflation or the short rate, or even NGDP, would require information about the goods side of a subset of transactions, while stabilizing total spending requires only money-side information.

    A technical problem with this proposal for stabilizing transaction volume arises if each node ceases to hold the entire blockchain in some future version of bitcoin, due to network scalability issues. Still, I think, this is easily solved by a technical modification to block syntax. Each block currently includes a hash of the block it’s trying to build on top of. If we just append to this hash one number, the running total transaction volume, then a block only needs to see the last block on the chain in order that this running total is maintained and can be used to trigger updates to the computational difficulty setting.

  20. avatar George Selgin says:

    I am much encouraged by, and extremely grateful for, all of these thoughtful responses to my "challenge," which I plan to consider carefully in planning a follow-up to my work on "Synthetic Commodity Money."

    One thought that already occurs to me is that some of the comments appear to take the challenge to be one of mimicking the outcome of a gold standard. That understanding wasn't what I intended. The idea is to be able to design a synthetic money commodity with properties superior to those of gold, including immunity to supply shocks. If the best we can do synthetically is to mimic gold, we might as well just return to using gold, were that possible. In any event, the usual arguments against returning to gold could serve to oppose any equally imperfect synthetic equivalent.

Leave a Reply

You must be logged in to post a comment. Login now.