Monday, April 13, 2015

A libertarian case for abolishing cash



Last week Citi's Willem Buiter published a note on the three ways to get rid of the effective lower bound to nominal interest rates, one of which is to abolish cash. He goes on to say that
politically, the abolition of currency would run into opposition from some of the legitimately cash-dependent poor and elderly, from those for whom the anonymity of cash is desired because they are engaged in illegal activities and from libertarians. The first constituency can be helped, the second can be ignored and the third one should take one for the team.
I think that Buiter is wrong to characterize libertarians as necessarily opposed to the abolition of cash. Their take on cash is probably (or at least should be) a bit more nuanced. Since libertarians generally advocate government withdrawal from lines of business like health care or liquor retailing, an exit of central banks from the cash business should be a desirable outcome.

What Buiter is advocating is a bit more extreme than just government exit, however. An across-the-board banning of cash would not only take the government out of the cash business but also prevent individuals and businesses from entering the product niche. The participation of the private sector in the provision of cash isn't just science fiction. Historically, commercial banks were intimately involved in the production of paper currency. In modern times, the majority of banknotes that circulate in Scotland are issued by three private banks—the Bank of Scotland, the Royal Bank of Scotland, and the Clydesdale Bank, while in Hong Kong, the major commercial banks are charged with issuing currency.

Buiter would probably object to private banknotes. After all, if private banks are able to issue negotiable bearer instrument that pay a zero nominal interest rate, a central banker will continue to be plagued by the problem that he/she can't reduce interest rates below zero—instead of fleeing into 0% government paper, the public will hide in private banknotes. It's the same liquidity trap as before, with private currency in the place of central bank currency.

However, there would be one key difference. Private banks must abide by the Darwinian calculus of profit and losses, central banks don't have to. Take a world with privatized cash. A recession hits and the rate of return on capital falls plummets. At the same time, the central bank drops its deposit rate deep into negative territory. As a private bank tries to match with deposit rate reductions of its own, say to -2%, customers will convert negative yielding deposits into the bank's higher-yielding 0% bank notes. The bank, whose survival depends on a healthy spread between the rates on borrowing and lending, faces a sudden spike in borrowing costs to 0%, the rate on their cash base. Spreads will shrink, even invert. Bankruptcy looms.

In order to avert this disaster, private issuers will quickly institute limits on their cash business. This could involve adopting any one of Buiter's three remedies: 1) cancel their note issue; 2) impose a fee on cash, or; 3) remove the fixed exchange rate between deposits and cash. Thus,the lower bound probably wouldn't be a problem in a banking system characterized by privatized paper issuance. The necessity of maintaining a spread would force private banks to rapidly innovate any one of these three escapes come recession and negative nominal rates. Upon recovery, they can remove these limitations and continue with their regular cash business.

Imagine that private banks all choose the first option when nominal rates fall below zero, cancellation. With cash no longer in existence, banks will have succeeded in restoring their margins to health. The population, however, will have effectively lost their ability to make anonymous transactions. This puts a libertarian in a tough philosophical position. On the one hand, a cashless world poses a serious threat to personal liberty. John Cochrane calls it an "Orwellian nightmare," and Chris Dillow has referred to banning cash as "a grossly illiberal measure - the banning of capitalist acts between consenting adults."

On the other hand, if cash threatens a bank's existence, no libertarian would advocate the use of force to prevent said bank from exiting the business of cash provision. Capitalistic acts cannot be forced upon non-consenting adults, or, put differently, Jack's desire for anonymity-providing products doesn't justify Jill being put into chains in order to provide those products. Therefore, a withdrawal of cash by banks as nominal rates fall below zero, and the loss of anonymity that comes with it, is consistent with libertarianism.

So oddly, Buiter's proposed end point—a cancellation of cash in order to rid the world of the lower bound—is very similar to what a libertarian end point could look like. Both institutions will elect to withdraw cash from circulation because it interferes with their institutional prerogatives. For a central bank, this mission boils down to the targeting of some nominal variable like inflation while in the case of a private bank it is its ability to earn a competitive return. That's not to say that a libertarian ought to support Buiter's abolition, only that the subject is more nuanced than it might seem upon a superficial reading.  

As a postscript, it's worth noting that neither Buiter's central banker nor a libertarian's private banker need go as far as abolishing cash in order to remove the effective lower bound. Buiter provides two other options, the best of which (in my opinion) is removing the fixed exchange rate between cash and deposits. Miles Kimball has gone through this option exhaustively. I've outlined some even less invasive, though not as effective, options here.



Related links: 

Does the zero lower bound exist thanks to the government's paper currency monopoly? (link)
Is legal tender an imposition on free markets or a free market institution? (link)
Bill Woolsey on how the private sector would withdraw cash at negative rates (link | link )
FTAlphaville: Buiter on the death of cash ( link )

Note: I changed some wording on September 26, 2015. The message remains the same.

29 comments:

  1. An interesting paradox. The flaw in your argument though is the assumption that in a libertarian world a bank would have to have to hold deposits at the central bank. Remove this requirement and you remove the need to eliminate cash for the bank. Negative interest rates at the central bank would simply cause the private bank to withdraw its funds from the central bank.

    ReplyDelete
    Replies
    1. The fact remains that the bank will have to invest those funds somewhere, if not the central bank, and earn a negative return on capital. As the bank reduces its deposit rate into negative territory to protect its spread, its customers will convert those deposits into cash.

      Delete
    2. "As the bank reduces its deposit rate into negative territory to protect its spread, its customers will convert those deposits into cash." Right, which is why the private bank that issues cash will still be subject to a zero lower bound and will not go into negative territory. Absent the requirement to hold reserves at the central bank, the cash issuing bank lacks an incentive to go into this negative territory. Your story only works because the central banker forces banks to keep reserves in the central bank, which cannot be squared with a libertarian view of the world.

      -Same anonymous.

      Delete
    3. "Absent the requirement to hold reserves at the central bank, the cash issuing bank lacks an incentive..."

      The overriding incentive to go into negative territory is due to the recession and the collapse in the return on capital. A central bank is just following behind as it cuts its rate. When the return on capital collapses, 0% notes (or 0% yielding deposits) are providing a superior return to the public while imposing a burdensome funding cost on the bank. Reducing the deposit rate to some negative level isn't sufficient since everyone will simply convert to notes. Cash has to be attacked too.

      Delete
    4. You're thinking like a central banker, not an individual consumer. Consumers would like a unit of exchange and a unit of account that has a non-negative interest rate. You presume that it would cost the bank money to allow the cash conversion. If the bank denominates their private currency according to a particular type of bond(s) or more likely precious metal they can always allow the conversion to the underlying asset, no matter its value in dollars. You effectively presume that banks would retain the dollar as the unit of account even if they were no longer allowed to use it as the unit of exchange in cash form because that interferes with the zero lower bound. So long as there is a stable type of bond or precious metal with a non-zero yield (or at least a yield above that of the central bank) the private bank can use that as a unit of account for its currency and undermine the central bank's lower bound. That obviously imposes costs on the consumer, namely the risk on the underlying asset. But absent coercion the barrier for the central bank to going significantly below zero is unavoidable.

      Delete
    5. You seem to be describing a change in the unit of account function of money. There are huge network effects that are at work here. For instance, it took a number of years of hyperinflation before Zimbabweans finally switched to pricing in dollars and rand. Negative rates are an infinitely-less burdensome on the population than hyperinflation --- in fact, they are designed to ensure a stable price level rather than deflation. So I would expect to see little in the way of currency substitution or redenomination.

      Delete
    6. Network effects can certainly help explain why a central bank can go (at least a litlle bit) into negative territory without consumers abandoning the currency, but your example of Zimbabwe is highly flawed. The problem with that example is that your post is about a libertarian world but your example involves a coercive state. Obviously a currency under extreme hyperinflation would be abandoned immediately if the state would allow it. The collective action problem is who will abandon it first if the state shoots the first store owner who switches to the dollar, but cannot make everyone abandon the dollar once everyone has already made the switch.

      Delete
    7. Think about another scenario, de-dollarization. States that have undergone high inflation will eventually experience dollarization. Long after the state's domestic unit has stabilized, nations will experience difficulties in de-dollarizing, even when they try to force the change upon the populace by an explicit policy. This 'hysteresis' shows the degree to which existing standards are resistant to change, even coercive change. So no, a negative interest rate policy -- which is far less of an inconvenience than a hyperinflation -- will not cause a regime switch.

      Delete
    8. To be very clear, I don't deny that there are network effects. But the scenario you describe does not reflect "hysteresis" and is not entirely a network effect. Even if the domestic currency has "stabilized" the government in such a country cannot credibly commit to be more hawkish on inflation than the Federal Reserve. So even without network effects many if not most local capital holders will prefer dollars to donuts. The local government might be better off if people switch back and you could even make an argument that the local economy might be better off making the switch, but that's in large part because that society might benefit from higher inflation than the Federal Reserve and those local people holding dollars instead of donuts.

      Delete
    9. It's referred to as hysteresis in the dollarization literature.

      Think about it this way. In normal times, Switzerland's central bank might set an interest rate of 1.5% and the ECB 2.75%. Despite the return on European exchange media exceeding that on francs by 1.25%, we wouldn't worry about euros displacing francs. The same goes when the Swiss set negative rates of -0.75% and the ECB sets rates at 0.5%. It's the same 1.25% spread.

      Delete
  2. Another possible outcome besides currency disappearing completely is currency becoming riskier, since a risk of default requires a rate above the risk-free rate. (I think I got this idea from Bill Woolsey, who's a fan of private currency).

    ReplyDelete
  3. The things which interest me about "getting rid of cash" are

    1 - How the legal tender law would be re-cast, especially bearing in mind (3) below. If cash and deposits are no longer of equal value, this becomes awfully messy.

    2 - How lawful-but-fraud-prone transactions such as the purchase of valuable used cars would proceed, given (3) below.

    3 - Bank transfers are reversible/reclaimable if fraud is claimed, and significantly risky (compared to the risk of accepting forged banknotes, which in the real world are still spendable) if the amount is large and the other party untrusted. Making bank transfers irreversible would significantly increase the attractiveness of bank fraud such as theft of online banking account details.

    4 - There are large illegal/borderline-legal industries here (the UK) such as prostitution, sale of drugs, and performing minor building work without the necessary paperwork/taxes. These industries deal exclusively in cash, will not go away, they are large and some of them extremely popular. If we do away with cash they will switch to some other medium of exchange - we should consider what that would be.

    A central banker's fear will presumably be that it is something outside their currency - either foreign banknotes or worse still, gold. It is not obvious to me that a banknote which loses value over time will be less attractive than a foreign one which doesn't.

    ReplyDelete
    Replies
    1. " If we do away with cash they will switch to some other medium of exchange - we should consider what that would be."

      Good points. Say that the grey/black market adopts the 500 euro note. As long as they continue to set prices in pounds while using euros as the medium of exchange, central bankers shouldn't be too concerned, since they exercise their power through the unit of account function of money. There are large network externalities that must be overcome for a full unit of account switch.

      Yes, legal tender is an issue. See Miles Kimball:

      http://blog.supplysideliberal.com/post/60337533206/the-path-to-electronic-money-as-a-monetary-system

      Delete
  4. Who cares? It does not have to be paper currency, you know. Libertarian positions would also liberalize gold and silver for the means of exchange. Hold that instead of paper.

    Besides, this whole ZLB is a farce of a chimera. Remember, low interest rates are generally a sign that money has been tight, as in Japan; high interest rates, that money has been easy.

    The world needs more currency and higher NGDP, not less currency and lower interest rates. Exactly what is the goal here, anyway?

    ReplyDelete
  5. Anonymous 1

    It is the ability of banks to hold deposits at the central bank, at zero or even positive interest rates, that allow a bank to issue currency at a zero nominal rate.

    When credit demand falls, the banks have to lower the interest rates they charge to maintain lending. When interest rates on earning assets fall, banks must lower the interest rates paid on deposits. If credit demand falls enough, then issuing zero interest currency becomes a losing proposition.

    If people want to invest in gold or silver, that is fine. The market price of these metals will rise enough to clear the market. And, of course, they might later fall. Investing in these metals would be as risky as instead purchasing equities, if not more.

    Only if a gold or silver standard is somehow enforced would the ability to accumulate these metals put a lower bound on the nominal interest rate.

    ReplyDelete
  6. So what you're saying is that everyone should have to use the central bank monetary system, but only private banks should be allowed to have access to central bank money.

    Yeah that sounds really "libertarian".

    Actually, no. It's very obviously authoritarian and, worse that that, it's oligarchic.

    ReplyDelete
  7. I don't follow the loss of spread argument. Tell me where I go wrong. I'm a bank. I lend at 0 (my notes if need be, redeemable against me for CB reserves). I borrow over night at the base rate -2, either at the CB window or in the market. I clear away so my reserve account is nil (if necessary I buy any financial asset above -2). I roll over each night at -2. I pay less than I borrow. I keep my spread.

    ReplyDelete
    Replies
    1. You've got it flipped around, I didn't explains well. A bank funds itself by notes. That is to say, it borrows by issuing notes at 0%. By holding its notes, the public is investing, or lending, to the bank.

      The bank itself invests in loans to others. It can lend to the public or other banks. When it holds reserves overnight at the central bank, it is lending to the central bank and earning interest.

      So if a central bank lowers its deposit rate to -2% and a commercial bank's entire portfolio is comprised of loans to the central bank, then the bank is effectively earning -2%. That would be fine if its funding, or borrowing, costs were -4%, for a spread of 2%. But if its entire borrowing base is comprised of note, then that cost will be 0%, for a net spread of -2%.

      Delete
    2. Thank you for your reply. That makes sense for a bank that is attracting deposits, (i.e. Seeking to counter a reserve drain) but what about a situation, arguably more relevant, when the bank is writing new liabilities. It's not borrowing from the public, and provide it doesn't kend to the CB at -2, say by working to keep its reserve balance 0 overnight, it holds its spread. I mean the base rate works both ways. Yes I lend at -2, but I also borrow at near -2. The whole rationale for the CB going negative is surely to promote this kind of behaviour, but does a bank even need to make new loans to make money here, just borrow as much as possible from the CB at -2.

      Delete
    3. As a sector banks issue 0 % notes (their liability) and as an asset they have -2 % reserve deposit (they "lend" to the CB).

      Delete
  8. As a longtime reader, I want to say that I appreciate these posts that revisit previous topics. Although the gist of what you're saying isn't much different from the sarlacc pit post, the wording here is somehow easier to understand. In particular, the paragraph starting with the following sentence was extremely helpful:

    "However, there would be one key difference. Private banks must abide by the Darwinian calculus of profit and losses, central banks don't have to."

    I haven't had time to think through all possible objections to your argument, but I definitely see some new subtleties.

    ReplyDelete
    Replies
    1. Thanks, John. I figured a re-visitation was worthwhile given all the recent banter about banning cash. Either that or I'm running out of ideas ;)

      Delete
  9. If neither the central bank nor commercial banks issue cash, what would the commercial bank deposit be redeemable for? If nothing, then commercial banks would be defrauding the customers. Also, what would be legal tender?

    ReplyDelete
    Replies
    1. They'd be made redeemable for central bank deposits. Bank deposits could become legal tender, or central bank deposits could. Or it could be that there is no legal tender, as in Scotland.

      Delete
    2. Ok, however, in such a case commercial banks and payment processors may object when the central bank is competing with them, and on the other hand, central bank will incur costs by providing these services (infrastructure, security, marketing, customer support, dispute resolution, ...). Also, I am not sure how it is in other countries, but in the US, federal institutions (e.g. the FED) are exempt from FinCEN's oversight, so they do not need to identify their customers or report suspicious transactions. This creates a market distortion and policy problems.

      Delete
    3. I agree. Redemption into central bank deposits would only be available to banks, in the same way that only a select group of actors can redeem ETF units for the underlying. Arbitrage by this group would ensure that 1 bank deposit = 1 central bank deposit.

      Delete
  10. I'd like to comment on the more general premises rather than the conents of the proposal. Based on my experience, the libertarian positions are often misunderstood (and probably also not well explained by many libertarians). I have presented and debated on the topic so I am a bit familar with it.

    The core libertarian issue with central banks is not merely that commercial banks cannot issue paper notes. Rather, this is a symptom of a much deeper issue, which is the existence of legal privileges in general. For example, in your model, even if commercial banks may issue bank notes, they still need to hold reserves in central bank deposits. If you try to run a bank without having the appropriate amount of central deposits, or if you issue a competing instrument for banks to hold, this is deemed illegal and you are prosecuted. There are other restrictions, such as how to do tax accounting (which I mentioned some time ago), that favour a particular unit of account. Recently I found other more obscure ones, for example, EU regulations require you to charge VAT on electronic services based on the country where the buyer is located, which penalises the use of cryptocurrencies in the EU (since the transfer of a cryptocurrency does not include data on the country of the payer). This probably wasn't intended this way, but has the effect nevertheless.

    Of course, as Ludwig von Mises insighfully pointed out, without a monopoly on the production of the monetary base, you can't conduct monetary policy. In other words, the whole concept of monetary policy is a statist one (anti-libertarian). In anarchy, monetary policy is impossible.

    Mises explained as well that as money is only demanded for its liquidity (something that you should find familiar), changes in the money supply are a purely redistributive measure. In fact, this is true for all media of exchange to a certain extent, because a proportion of their market price contains a liquidity premium. If you have a monopoly on a production of a liquid asset, this grants you also a great redistributive power, so no wonder libertarians object to it.

    ReplyDelete
    Replies
    1. I disagree that monetary policy is a statist concept. Competing banks will naturally coordinate their activity at one or two private clearinghouses. In a libertarian society, it is likely that this clearinghouse will be governed as a not-for-profit utility owned by its member banks. The members will have to determine rules for issuing clearing balances and how to peg the price of those balances. If they choose to peg to gold, than that is the clearinghouse's monetary policy.

      Delete
    2. I'll begin from the end. There is a difference between pegging a monetary unit to gold (which mainstream calls "gold standard") and when the monetary unit is a weight of gold (which the Austrians call "gold standard"). The latter is not a monetary policy. The former can be, but only if there is a monopoly provider of the monetary unit (competitors would otherwise issue their own competing alternatives). And to deny competitors entry, you need a legal privilege (= state).

      While banks could try to form a cartel in anarchy, they can't prevent competitors from creating their own systems. They can try to deny competitors access to their cartelised clearing system (just like they do now), and that is a nuisance, but that may not end work in the long term.

      You could also have an open-access clearing system (similar to Ripple, but not Ripple itself because that is not open-access anymore), which would prevent the possibility of cartelisation (as that requires the ability to deny access to unwanted competitors).

      Furthermore, it is possible that free market would separate lending and deposit functions of banks, preventing them from affecting the money supply at all. Some argue this would happen because fractional reserve banking is illegal, some claim that this would happen if monetary base has sufficiently low transaction costs and thus there is insufficient monetary demand for demand deposits (e.g. Bitcoin).

      A lot of the competition that would exist on a free market is now either illegal or has high barriers to entry, whether this is intended or not. For example, if you want to found a bank or a payment processor, you need to comply with the regulations whether you want access to the existing clearing houses or not.

      Delete