Fractional Reserve Banking With Bitcoins

I recently got into an argument over on the Reddit Bitcoin boards where I held the position that fractional reserve banking with Bitcoins was not possible.  I’ve repeatedly made this claim before due to the fact that I think no one would accept receipts for Bitcoins (bank notes) over actual Bitcoins.  However, the argument on the Reddit boards led me to see the error of my previous logic.  What can happen is that a Bitcoin bank could use its deposited accounts as “receipts” for Bitcoins because the digital account balances could be detached from actual Bitcoin wallets.

A bank could keep a digital ledger of accounts and pool all their deposited coins into a single wallet that it lends from.  Obviously this is the most likely scenario anyways, since we see this same process with current Bitcoin exchanges.  So using a 10% reserve ratio as an example, a bank could have 100 Bitcoins in demand deposits while lending out 90 of those Bitcoins to other accounts.  It would be able to get away with this because of the digital account balances.  Those who deposited money in the bank would not see their balances decrease when the money is lent out unless the bank explicitly setup its system to operate this way.

This could lead to the unstable situation described in this Khan Academy video on fractional reserve banking:

The one thing that would NOT be possible under Bitcoin banking would be the creation of a central bank or FDIC type “insurance” scheme to backstop bank reserves.  This is because there is no way for new Bitcoins to be created other than the mining process and because people directly conduct transactions using Bitcoins themselves.  Obviously I find this to be a very healthy thing because it limits the fraud and forces individual bank depositors to be responsible for losses instead of tax payers.

The implications of this are that Bitcoin banks will be viewed with a high degree of skepticism by depositors.  Depositors will view depositing their money at a bank as taking the same kind of risk as they would view buying stock in a company.

In thinking about what prevents bank centralization in an all Bitcoin economy, it dawned on me that the use of Bitcoins in direct transactions is vital to preventing bank centralization.  Say if Bitcoin banks all operated on a fractional reserve banking system, in theory, bank centralization could occur in much the same way it did under the gold standard if Bitcoins were not used directly in exchange.

What really made bank centralization possible under the gold standard was the use of bank notes to represent gold, and the fact that gold was a physical item that could be easily seized.  Once banks got it in peoples’ heads that bank notes were “as good as gold,” the path toward monetary destruction was set.  The ability for people to conduct wire transactions anonymously using Bitcoins is a key component in preventing the centralization of banks under a Bitcoin system.

Hmmm…. This raises some more food for thought.  Let’s suppose a state tries to seize control of the banking system.  It could order all banks to freeze their accounts at gun point, and then order all banks to transmit their wallets to a central banking authority that would take control of all the Bitcoins in the banking system.  Peoples’ digital account balances would remain unchanged, but all the Bitcoins would be gone.

The only thing that would prevent the state from being able to get away with this is the fact that all the retail payment systems are setup to accept Bitcoins directly, rather than accepting ledger transfers from banks as they do today.  Keep this fact in mind when people start advocating for instant transfer mechanisms between banks and retailers that rely on ledger transfers rather than actual Bitcoins.  It could lead down a dangerous road.

It will be interesting to see how Bitcoin banks operate in the future.  I suspect people will be leery of banking with a bank that does not require time deposits for interest bearing accounts since this would be an obvious sign of fraud.

If I had to speculate, I think we might see banks operate with several various account choices.  A bank might offer a secure demand deposit account where it hosts an individual wallet for each account.  It might also offer an unsecured demand deposit option, which would allow for instant transfers of payment between bank members.  And then it might offer an interest bearing time deposit type of account that it would use to make loans from (depositor assumes risk as well as reward).   The risk of fraud to depositors would be if the bank started using its unsecured demand deposits to lend from or backstop its reserves.

  • Pingback: Fractional Reserve Banking With Bitcoins | Libertarian News | BitJack21 Blog()

  • “Keep this fact in mind when people start advocating for instant transfer mechanisms between banks and retailers that rely on ledger transfers rather than actual bitcoins.”

    That is indeed a dangerous road! We all must demand that any business that accept bitcoin-bank-transfers also accept actual bitcoins. As soon as someone stops accepting actual bitcoins and only bank transfers denominated in Bitcoin is when we have a huge problem.

  • Pingback: Bitcoin | Pearltrees()

  • Thanks Michael. People should remember and file this article. This will become a big issue in the near future and I’m confident that it can be an overall plus for bitcoin. It is really no different than the free banking debate around the gold standard and the gold-exchange standard. Bitcoin is digital gold; and gold is analog bitcoin.

  • simi tong

    I would like to thank you for the efforts you have put in writing this blog. I’m hoping
    the same high-grade web site post from you in the future also. Actually your
    creative writing abilities has encouraged me to get my own web site going now.
    Really blogging is spreading its wings and growing fast. Your write up is a great example.

  • mikeriddell62

    @jonmatonis:twitter I’m no expert in currency or fractional reserve banking, let me be clear. But what backs Bitcoin? It seems to me that it lacks intrinsic value. Gold has intrinsic value so how does that real value transfer over to Bitcoin in the way you suggest? It also seems to me that there aren’t any moral values that back Bitcoin since it’s used for a variety of purposes none of which appear to contribute to the interests of humanity.  Isn’t it just one crap currency replacing another crap currency?

    • Asking “what backs Bitcoin” is like asking “what backs gold?”

      Bitcoin has value because it acts like a digital commodity.  Bitcoin has all the same properties that make gold a money.  Bitcoin is fungible, divisible, easily identifiable, and scarce.  Gold isn’t a money because it looks pretty.  Investment gold sits in bank vaults in bars that no one ever looks at.   Bitcoins are like a “gold” accounting system without the gold.

      • mikeriddell62

        It may be all of the things you say it is, but that doesn’t make it valuable. Currencies seem to me to be underpinned by trust. What if people in the marketplace don’t trust bitcoin?

        Will you accept bitcoin for your gold?

        • Actually it does make it valuable. Hence, that is why it will cost you 5 bucks to get a hold of one.

          What if the marketplace loses trust in the dollar? Have you ever heard of Weimar Germany?

          • mikeriddell62

            I think the market is likely to lose trust in the dollar, just as the world loses trust in the US. What their brand stands for is what gives them their value. I’m not saying Bitcoin isn’t a suitable replacement I’m just saying that as a brand it needs working on.

            Other digital currencies are being developed as i’m sure you are aware and those that are backed by the trust of real communities will, in my opinion, find it easier to hold and grow their value.

            Here in the Uk we have several such currencies being developed in Wigan, Brixton and Bristol which might end up being exchanged with Bitcoin at some point in the future.

          • It’s an interesting thought, but I don’t think any other digital currency can compete with Bitcoin because Bitcoin is a global currency.  Local currencies that have some kind of commodity backing generally have difficulty expanding the further they get from the place of issuance.  With Bitcoin, redemption is not a problem, which means it has the same value everywhere.

          • mikeriddell62

             Our currency is only ever earned into existence by people volunteering time to their local community. One hour generates 10 £Wigan (1,000 points). It is underpinned by the principles of timebanking and backed by time. It can be exchanged for unsold seats at sports games or for excess stock that a retailer is looking to get rid of.

            We think we have a chance of persuading the municipal authority to accept the points in part payment for local taxes.

          • Yeah, and that’s great.  But again, the problem I’m specifically addressing is that redemption of those tickets becomes a problem the farther away from the community you get.  This isn’t a problem for Bitcoin.

  • Bitcoin E-wallet services are just begging to turn into banks one day. And if scalability problems will not be solved – they will.

  • Andrew Skretvedt

    The Khan Academy video is nice, but a bit muddled, so for anyone who’s not yet heard of it, Murray Rothbard did a much clearer job of explaining the situation in his book, “The Mystery of Banking.” You can buy a copy, or download a free PDF version, on

    Rothbard also illustrated what was left out of this video, how you get to the “modern” banking system from the situation of an individual bank making loans to clients out of deposits, retaining a reserve whether gold-pieces or paper money or checks submitted which draw upon a foreign bank. I very highly recommend this if you’ve not yet read it. Really. Go read it now.

    Also, I think if bitcoin does ultimately develop into a modern alternative currency, people will demand the same convenience features they get from the existing financial system. So, this will mean swipe-instant or button-instant transactions without waiting the many minutes for blockchain confirmations of transactions. I’m aware of OpenTransactions as one software framework to help achieve this. Its principal developer thought that bitcoin would become the underlying medium backing the transactions in substitute media/contracts his framework would account for. You get an instant payment inside an OT federated network of “bank”-like entities, these entities could settle accounts among one another via bitcoin transfers.

    So with OT, you’d gradually develop toward a free-banking financial ecosystem, where for the sake of convenience, people commonly transact in a bank-entity provided bitcoin substitutes with modified properties from actual bitcoin. I’d expect that initially (discounting scheming wildcats) the system would be 100% reserve to earn trust and users and foster a culture of usage. Later, some fractional-reserve expansion could set in, motivated by profit, and kept in check as a function of the number of bank-entities and the size of their clientele, as Rothbard explained for gold-based fractional reserve banking. A central bank could operate here, with fractional-reserve credit expansion and inflation limited only by the extent to which people still believed balances were able to be redeemed in bitcoin.

    The check on this credit-expansion situation getting out of hand would be the extent to which the actors in the economy still transact directly in bitcoin. Bitcoin doesn’t lend itself very well to live-in-person sorts of cash transactions, so I might expect a well developed physical economy around bitcoin to make heavier use of substitutes that can be passed in person like cash or as cash with a minimum of fuss, extra equipment, or access to computing and network resources that the other party may not have conveniently to hand. This might make direct bitcoin transactions by individuals rare on the street and in most places, except online near a browser window. Even online, I could see where people might want the option of a mutually trusted payments processor where chargebacks are possible as a surety on contract performance and protection against certain forms of fraud.

    If bitcoin is digital gold, perhaps bitcoin could follow the same trajectory that gold has (in some sense, if so, then bitcoin already will never develop beyond a fringe curiosity due to existing legal tender laws and that taxes must be paid in legal tender…there would first have to be mass repudiation of the government’s right to power by the public).

    • Yeah, you basically re-wrote my article LOL.  I agree with what you are saying.  What I think this means is that there is room for improvement within crypto-currencies.

      The holy grail is instantaneous confirmation in a distributed ledger.  In my opinion, the development of Bitcoin should proceed in such a way as to minimize the chances of the currency being replaced by bank substitutes.  Obviously bank substitutes are highly dangerous.

  • Peter Surda

    Hello Michael,

    I analyse Bitcoin FRB in my upcoming paper about Bitcoin. I don’t want to paste it all here, so just a brief summary.

    In order for FRB to expand the money supply, the claims to Bitcoin need to act as as a medium of exchange. I was able to determine three possible ways this can be achieved:

    – reduction of transaction costs
    – payment of interest
    – government intervention (or other use of force)

    First two can be achieved natively, without a claim to Bitcoin being a medium of exchange. Bitcoin already has low transaction costs, and there are methods of transacting it outside of the blockchain without a debt instrument. Like Casascius coins. Digitally, this can be done (if I understand it right) with OpenTransactions. There are also methods like the green-address approach, or hybrid solutions like the Bit-Pay merchant tools. With respect to interest, Bitcoinica already provides interest without offsetting it by loans: they use the deposit balance for hedging the trades. The hedging positions can be liquidated immediately, so there is no maturity mismatch. Obviously, in the end these two points are empirical and we can’t predict them completely, but at least they conceptually they present obstacles for expansion of the money supply, unlike with gold/fiat.

    The last one is unclear. You argue that the government can monopolise Bitcoin banks. There is a second approach possible too: government fixes the exchange ratio between Bitcoin and claims on Bitcoin, fuelling the outflow of Bitcoin from the economy (Gresham’s law). In my opinion, it is difficult to do either effectively unless Bitcoin is a legal tender or at least a widely deployed medium of exchange, AND these claims on Bitcoin are already used as a medium of exchange. Otherwise, the result could simply be that people start a new blockchain and the whole effort of government would end up pointless (also the market price of the old blockchain would drop and the monopolised banks would go bankrupt). This is similar to forking of open source projects, and we see that historically it can happen despite the network effect favouring the encumbents (e.g. libreoffice vs. openoffice). Again, this is an empirical question, of unclear consequences.

    So while on one hand I am a bit more pessimistic than your prior positions on Bitcoin FRB, I am also more optimistic than your current one.

    • “here is a second approach possible too: government fixes the exchange ratio between Bitcoin and claims on Bitcoin, fuelling the outflow of Bitcoin from the economy (Gresham’s law).”

      That would be difficult for them to do considering that Bitcoin is an international currency. I would simply exchange my coins for another currency and then trade those against the dollar if I really wanted to.

      I’m not saying that the government could, at this time, seize control of the Bitcoin economy. I’m just pointing out that fractional reserve banking leads down the path toward centralization. I agree that all sorts of laws and regulations would need to be imposed in order to make this happen. The question is whether people could easily by-pass those laws making centralization impossible.

      • Peter Surda

        I agree with you in that having a fixed exchange rate would be difficult, and also that FRB leads to centralisation. I guess my point is that FRB does not come into existence magically, it must follow economic rules. With gold (and similar commodities, FRB comes to existence for these reasons:
        on the demand side: there is a demand for money substitutes, because they provide something that money proper does not (see the three aforementioned points)on the supply side: money substitutes carry maintenance costs for the issuer (e.g. storage of gold) and these need to be offset somehow. The issuer can charge on holding (e.g. demurrage of bank notes), transacting (e.g. cheque clearing), or, obviously, externalise the costs through fractional reserves. From the point of view of an individual user, FRB appears to be the least costly alternative (it might or might not be when you consider inflation, I’m not yet sure), so obviously, FRB wins.

        Putting it together: if there is a general demand for money substitutes, this leads to FRB (unless it’s illegal, then it might not). Solution: have money which does not lead to the creation of money substitutes. Bitcoin shows that at least hypothetically, this is possible. I might even go a bit further and make this statement: if on a free market money substitutes do not develop even though there is no legal or technical obstacle for them, it means that the choice of money is pareto-optimal (since no change in the monetary system leads to an increase in utility). I hope it does not sound too technocratic 🙂

        • Peter Surda

          Now that I think about it, there is a fourth way of offsetting the costs: using the deposited money proper in highly liquid financial markets, just like Bitcoinica does. I suspect that in the more distant past, this wasn’t possible, so FRB won. I need to think about it a bit more.

          • Yeah its a tough call.  I agree that the solution is to  have money which does not lead to the creation of money substitutes. I’m not sure if Bitcoin completely fills this bill though because of the transaction delay that is necessary to confirm a transaction – which is what I’m pointing out in the article. Tough to say.

            I still think the bottom line is that if the general market retail system for products and services is established to accept bitcoins directly, it will go a long way toward preventing any kind of attempt at centralization no matter what system of banking is used.

            So far, it looks like retailers are progressing along this track, which is obviously a good thing.

          • Peter Surda

            Transaction delay is not that big a problem. There are solutions for it that are not based on credit, for example transacting keys (e.g. Casascius coins, OpenTransactions) or adding new layers (e.g. green address approach).

            The acceptance by market participants is the ultimate question, obviously. This is a complicated topic, because competition between payment systems is unusual. The economic term is “two-sided market”. Also the possibility of hybrid solutions (economic term “multihoming”), such as Bit-Pay allowing the merchant to receive fiat instead of bitcoin, can lead to piggybacking/hijacking.

          • Taulant Ramabaja

            I’ve been thinking about this since your first blogpost concerning FRB and bItcoin.

            It seem to me that the transaction delay and the self-interest of the holder to not spend it, might lead to an interesting new phenomenon. Maybe my thinking is completely off, but I’d love to know your thoughts on this:

            I think that OpenTransactions will become the defacto standard when it comes to money-like representations of bitcoins. One should be able to lockin a certain amount of bitcoins into an OT account, and instead give an OT token to somebody else, representing that amount of bitcoin. The OT token can be send in an instant (an completely anonymously, not just pseudonymously), while the new holder can be absolutely certain that the bitcoin which backs the OT token has not moved within the blockchain.

            Now, the incentive to use OT instead of btc directly just for anonymity reasons or the speed of transaction might not be enough to become a viable secondary layer on top of btc. However, there is the second issue within the btc economy that the current holders of btc don’t have any reason to spend their btc, making it very difficult for btc to turn into anything which can truly be backed up by products and services within an economy.

            Now, if however the current holder of the btc could lockin the btc into an OT account, but put a demurrage on it, effectively making himself a bank, then there would be no reason for anybody who already wants to spend a bitcoin not to use this possibility, simply because they will get some if not most of their btc back at some point. In a sense they would become their own bank, and spend the OT instead of the btc due to grasham’s law.

            From a receiver’s perspective (a service business for example) there would be no reason not to accept such a token either because a) the value of the represented btc is increasing fast enough due to deflation as to cover the demurrage in real value terms and b) the demurrage token should be redeemable for the real btc at any point when they need it for a preset fee.

            Anybody who accepts btc already has either discounts or better offers for the spenders of btc (I’ve seen up to 30% discounts), which makes sense within the deflationary economy. Accepting btc backed OT tokens would similarly be calculated by the market, however, real running businesses would try to cover their expenses by directly spending the OT token instead of first redeeming it into btc and then covering their costs.

            Added to that real world transactions which are currently covered by cash could finally happen with btc backed OT tokens.

            The economic dynamic between the OT demurrage token, btc, the fee structure, and the open market could have a positive feedback effect, increasing the value of btc due to economic activity, therefore evening out the demurrage effect for most businesses.

  • Since Bitcoin has no central bank, Bitcoin fractional reserve banking by definition does not exist. Fractional reserve banking means, that banks are required to hold a certain fraction (e.g.1%) of their client’s deposits as a reserve in a central bank account. The remainder can be lent out to borrowers. Bitcoin lending operations do not face a reserve requirement because there is no Bitcoin central bank.

    However, the Bitcoin money supply ( grows with Bitcoin lending operations and the effect is almost identical to FRB. And as your article says, holding deposits at a Bitcoin bank is riskier than at a fiat currency bank, because there is no deposit insurance. I think this is good because then people question more what they do with their money. I personally belive that the best way to operate Bitcoin lending is through peer to peer lending as we do on You always lend over a specified time. Therefore the golden rule of banking (always match maturities in borrowing and lending) is obeyed and a run to the bank along with deposit insurance becomes obsolete.


  • Nanofuture

    Creation of FDIC-style insurance is entirely possible. Every bank would simply pay fees into it and then the insurance would pay out in case of failure, like how it works now. You just wouldn’t have the additional backing by the full faith and credit of government x.