Central bank e-cash

After my post last week, prompted by the Reserve Bank’s recent statement that

Work is currently under-way to assess the future demand for New Zealand fiat currency and to consider whether it would be feasible for the Reserve Bank to replace the physical currency that currently circulates with a digital alternative.

I exchanged notes with a few readers with some in-depth thoughts on the issue, and found my way to some other relevant material including the recent first report of the Swedish central bank’s e-krona project.    And I noticed that Phil Lowe, Governor of the Reserve Bank of Australia, was giving a speech on exactly that topic – “An eAUD?” –  yesterday.  I gather that among advanced country central banks this is now treated as quite a high priority issue.    But it is also interesting that –  contrary to the Reserve Bank of New Zealand comment about their work –  both the RBA and the Riksbank are only talking about the possibility of electronic retail cash as a a complement to physical currency, rather than a replacement for it (and Sweden already has one of the very lowest currency to GDP ratios of any country anywhere).

Lowe’s speech was interesting, but also unsatisfying and unconvincing in a number of important areas.    As a New Zealand reader –  from a country with many of the same banks (and presumably banking technology options) –  I was struck by the contrast in what has been happening to currency to GDP ratios in the two countries.   Lowe illustrates that the share of transactions being effected by cash is also dropping sharply in Australia.  But here is the New Zealand currency to GDP chart I ran last week

notes and coin

And here is comparable Australian chart from Lowe’s speech.

Aus currency to GDP
45 years ago, the levels of the two series were very similar.  Since then, the trends have been very different and now there are many more physical AUDs in circulation (relative to GDP) than NZDs.   But there is nothing in Lowe’s speech about just why so much physical currency continues to be held in Australia –  far more than any plausible transactions demands (supported by evidence from payments practices data) would support.    Ken Rogoff suggested, in a US context, that the bulk must be held to facilitate illegal activities, or tax evasion in respect of otherwise legal activities.   Perhaps Lowe felt it wasn’t his place to venture far into territory around lost tax revenue, crime etc, but it was still a surprise to see no mention at all, when the RBA seems largely content with currency physical currency arrangements.

I was also rather surprised to see no serious engagement with the issues around the near-zero lower bound on nominal interest rates, which arises because of the option to convert unlimited amounts of bank deposits etc into zero-interest physical currency, an option that would be likely to be exercised on a large scale if official interest rates were dropped much below, say, -0.75 per cent.  Like New Zealand, Australia hasn’t yet approached the near-zero bound.  Neither had the US, Japan, Switzerland, Sweden, or the euro-area, until they did.   But Australia’s official interest rate is now only 1.5 per cent.  Perhaps it will be raised a bit before the next serious recession hits, but no prudent central banker could be discounting the possibility that even the RBA will hit the effective floor –  and limits of conventional monetary policy –  when that next recession comes.    Dealing effectively with that floor  –  by significantly winding back access to physical cash –  should be one important consideration when central banks are considering e-cash options.  But Lowe doesn’t even mention the issue, and while the limits of monetary policy might not have been of much interest to his immediate listeners (the Australian Payment Summit), interest in his speech –  and the issue –  goes much wider than the immediate audience.   (Strangely, in the Riksbank’s work they also talk in terms of zero-interest e-cash options –  albeit with the flexibility to change that at a later date –  and thus don’t really grapple either with the near-zero bound problem.)

To me, the heart of Lowe’s speech was his discussion of the possibility of the Reserve Bank of Australia issuing one or other of two types of eAUDs.

  • An electronic form of banknotes could coexist with the electronic payment systems operated by the banks, although the case for this new form of money is not yet established. If an electronic form of Australian dollar banknotes was to become a commonly used payment method, it would probably best be issued by the RBA and distributed by financial institutions, just as physical banknotes are today.

  • Another possibility that is sometimes suggested for encouraging the shift to electronic payments would be for the RBA to offer every Australian an exchange settlement account with easy, low-cost payments functionality. To be clear, we see no case for doing this.

I’m not sure I have a particularly good sense of what the first option involves, but here is how Lowe describes the possibility

The technologies for doing this on an economy-wide scale are still developing. It is possible that it could be achieved through a distributed ledger, although there are other possibilities as well. The issuing authority could issue electronic currency in the form of files or ‘tokens’. These tokens could be stored in digital wallets, provided by financial institutions and others. These tokens could then be used for payments in a similar way that physical banknotes are used today.

But he doesn’t seem keen, and so I’m going to focus my discussion in the rest of this post on the second of his options.   The issues and risks are pretty similar for both options, and I favour (provisionally) something like the second option.

At present, central banks offer exchange settlement accounts to facilitate the interbank settlement of transactions (the RBNZ policy is here –  something they must be reviewing, as there was an RFP for work in this area a few months ago).   These accounts facilitate payments, but they also allow entities given access to such accounts to hold electronic claims on the Reserve Bank (that are free of credit risk).  Central bank physical banknotes are also credit risk-free claims on the central bank.   But one set of claims is newer technology, regularly updated, enabling banks to both easily make payments and store value, while the other is a declining technology.

Here is how Lowe describes the option in this area

Another possible change that some have suggested would encourage the shift to electronic payments would be for the central bank to issue every person a bank account – for each Australian to have their own exchange settlement account with the RBA. In addition to serving as deposit accounts, these accounts could be used for low-cost electronic payments, in a similar way that third-party payment providers currently use accounts at the RBA to make payments between themselves. Some advocates of this model also suggest that the central bank could pay interest on these accounts or even charge interest if the policy rate was negative.

I’m not sure anyone argues for this approach to “encourage the shift to electronic payments”, but rather to reflect the world we now find ourselves in, in which electronic payments media and (records of) stores of value overwhelmingly dominate.   If favoured banks and financial institutions are allowed access to risk-free overnigh electronic balances, why shouldn’t ordinary Australians (or New Zealanders) have such access?  After all, at the absurd extreme, central banks could still insist that to the extent banks wanted to deal with them, they did so in physical banknotes.  It would be wildly inefficient to do so, but it could be done.  But if it doesn’t make sense to restrict such “big end of town” transactions to physical currency, why does it make sense to restrict ordinary citizens’ access to central bank outside money?

But the RBA is firmly opposed to change of this sort.

On this issue, we have reached a conclusion, rather than just develop a hypothesis. The conclusion is that we do not see it as in the public interest to go down this route.

Why?   Lowe raises three concerns, of which two are substantive and one is mostly rhetorical.

If we did go down this route, the RBA would find itself in direct competition with the private banking sector, both in terms of deposits and payment services. In doing so, the nature of commercial banking as we know it today would be reshaped. The RBA could find itself not just as the nation’s central bank, but as a type of large commercial bank as well.

In times of stress, it is highly likely that people might want to run from what funds they still hold in commercial bank accounts to their account at the RBA. This would make the remaining private banking system prone to runs.

On both counts, I think he is largely wrong, and that any issues are quite readily manageable.

It isn’t at all clear why (many of) the public would want to use an RBA (or RBNZ) exchange settlement account for routine transactions services.  Revealed preference suggests that people are mostly very happy to run the modest credit risk associated with using private bank deposit and payment services.  Almost all of us now use bank deposits for most of our transactions –  even when physical cash is a perfectly feasible alternative (eg there is no additional cost in time or anything else to, say, taking out $400 from an ATM once a week rather than say $200).  And in the handful of places where private banknotes still circulate (eg Scotland) there doesn’t seem to be any unease about taking them, or transacting with them.

In addition, banks can offer bundled products –  cheaper fees for example where you have your mortgage, or term deposits, with the same bank as your transaction account.  No one proposes that central banks will be offering mortgages, term deposits or any of the rest of the gamut of products the typical commercial bank makes available.

I’m not aware that anyone is suggesting central banks should set out to out-compete banks.  The argument for making central bank e-cash readily available is about a fallback –  a residual option, much as cash is now for many purposes.   Central banks almost inevitably would lag behind commercial banks in their technology anyway, which wouldn’t make a central bank transactions account product particularly attractive.   And it could easily be kept that way –  don’t offer provision for regular direct debits etc, don’t allow overdrafts at all, keep the fees just a bit higher than those on commercial bank accounts, and –  of course –  be prepared to adjust the interest rates paid (or charged) on credit balances to limit potential demand.    What would be on offer would be a basic credit-risk free product –  something similar to the fairly basic products central banks provide to banks themselves.  Frankly, I’d be a bit surprised if there was much (normal times) demand at all (and I think back to the days –  decades ago –  when the Reserve Bank offered –  in direct competition with the private banks –  cheque accounts to its own staff; perhaps some people used theirs extensively,  but I used it hardly at all).

Lowe’s other concern –  and I’ve seen this concern in other places too –  is that provision of e-cash for ordinary citizens might destabilise the banking system.    As he noted earlier in his speech “it is likely that the process of switching from commercial bank deposits to digital banknotes would be easier than switching to physical banknotes. In other words, it might be easier to run on the banking system.”

Frankly, if the only thing that prevents runs on the banking system is that it is too hard to run to cash, central banks and regulators have bigger problems that they might need to address directly.  Runs are often quite rational –  there are real issues with the “victims” funding and/or asset quality.  If it really were easier to run with electronic central bank cash, banks – and their regulators –  might need to look to the size of the capital and liquidity buffers.   As it is, Lowe seems to be suggesting banks can free-ride on technical obstacles their (retail) depositors face.

But I’m not really persuaded that simply making available a basic retail e-central bank cash option would either increase the prevalence of runs or threaten the stability of the financial system.     When there is a concern about an individual bank (or non-bank) people “run” electronically anyway –  mostly they don’t withdraw their deposits into physical cash, but into liabilities of another private institution (and we seem to have been seeing such a quiet run on UDC in recent months).   Wholesale runs –  the sort that took down Bear Stearns and Lehmans –  all happen electronically.  Banks themselves can run straight to central bank cash, when they cut lines on each other.  Is the Governor really suggesting that it is just fine that wholesale investors should find it easy to run but not retail investors?  In practice, that is what he is saying.  In a systemic run –  or a period of heightened systemic unease – it is very easy for wholesale investors to find a safe asset (whether exchange settlement account balances for banks, or government bonds/ Treasury bills for others).  It isn’t for retail investors.  And recall that in New Zealand we have no deposit insurance.

If I’m uneasy at all about the idea of making available an eNZD (or AUD) for retail users –  a basic store of value/means of payment technology with no credit risk –  it is that demand would be very limited in normal times, and that if there ever was a systemic crisis it might prove very hard to scale the product quickly to adequately demand.   There are probably ways of resolving that concern, but it does need more work.

One other concern I’ve heard expressed if this if the central bank issued retail e-cash it would create a reinvestment problem –  what would the Reserve Bank buy and hold on the other side of its balance sheet (with associated credit and quasi-fiscal risks).  This is mostly a non-problem for several reasons:

  • normal times demand is likely to be low, and can be kept fairly low through pricing,
  • retail e-cash would probably go hand in hand with steps to reduce the stock of physical cash (and central banks already reinvest the proceeds of the sale of notes),
  • in a crisis, central banks have this issue anyway –  the substantial liquidity injections typically involve material credit risk anyway, and
  • in practice, many central banks typically reinvest the proceeds of note issue (or subscribed capital) in government bonds (predominant approach in New Zealand) or foreign reserves (typically mostly the government bonds of other countries).

With an integrated approach to gradually reduce the stock of physical currency, while making available a retail e-cash product, I would expect that if anything central bank balance sheets would shrink somewhat (especially in Australia, with a higher currency to GDP ratio) rather than grow.   Steps in that direction would:

  • help deal with the zero lower bound problem,
  • reduce the tax evasion etc issues apparently associated with large holdings of physical cash, and
  • provide ordinary citizens with the same sort of basic risk mitigant/payments product open to banks.

Finally, I said that one of Phil Lowe’s counter-arguments was mostly rhetorical. That was this one

The point here is that exchange settlement accounts are for settlement of interbank obligations between institutions that operate third-party payment businesses to address systemic risk – something that is central to our mandate. A decision to offer exchange settlement accounts for day-to-day use would be a step into a completely different policy area.

Well, yes, as conceived at present exchange settlement accounts are about interbank dealings.  That is a core part of the RBA’s (and RBNZ”s) responsibilities.  But the provision of basic “outside money” –  credit risk free –  has also long been a core part of both central bank’s responsibilitiies.  Retail e-cash helps fulfil that part of those mandates in a technological age.



Whither cash?

Last week the Reserve Bank released an interesting Analytical Note on “Crypto-currencies – An introduction to not-so-funny moneys” .    If, like me, you hadn’t paid a great deal of attention to Bitcoin and the like, it is a very useful introduction to the subject, from a monetary perspective (including some of the potential policy and regulatory issues).  At least for me, it struck just the right balance of detail and perspective.

Analytical Notes are published with the standard disclaimer that the material in them represents the views of the authors rather than, necessarily, of the Bank.  (That said, I’m pretty sure nothing has ever been published in one that the Bank was unhappy with.)  They are mostly written by researchers rather than policy people.  So it was interesting, and perhaps a little surprising, to get to the second to last page of this paper and find this

Work is currently under-way to assess the future demand for New Zealand fiat currency and to consider whether it would be feasible for the Reserve Bank to replace the physical currency that currently circulates with a digital alternative. Over time, analysis associated with this project will filter through into the public domain.

Interesting, because that is quite a radical and specific suggestion: to replace physical currency with a digital alternative.   And surprising because there was no hint of this work –  on a pretty major issue affecting all New Zealanders –  in the Reserve Bank’s Statement of Intent released only a few months ago.   Statements of Intent can seem like just another bureaucratic hoop to jump through, but the requirement to prepare and publish them was put in place for a reason: it is supposed to be the vehicle through which the Minister of Finance can inject his or her views on what the Bank’s work priorities should be, and is supposed to enable stakeholders and the public more generally to get a sense of what the Bank is up to.

I’m pleased the Reserve Bank is now doing this work on the future of currency.  Over the last couple of years I have been critical of the fact that, in published documents, there was no sign of any such preparatory work going on (including, more generally, around dealing with the problems of the near-zero lower bound, which will almost certainly become binding for New Zealand in our next recession).  In this year’s Statement of Intent, for example, published as recently as the end of June, there was 1.5 pages (pp 28-29) on the Bank’s currency functions, and not a hint of any work on the possibility of replacing physical currency with digital currency.   Perhaps doing the work is an initiative of temporary “acting Governor” –  but then he was required, by law, as Deputy Governor, to sign the Statement of Intent.  Or perhaps it was just the Bank deliberately keeping things secret?

As usual with the Bank, they talk loftily about how the analysis will eventually “filter through to the public domain”.  That isn’t good enough –  this is publicly funded work on a matter of considerable potential significance – , and I have lodged an Official Information Act request for the research and analysis they have already done.

I’ve come and gone for decades on what the best approach to physical currency is.  I’ve long been troubled by the monopoly Parliament gave to the Reserve Bank over the issuance of physical notes and coin.  There is no good economic reason for it (nothing about the efficacy of monetary policy for example) –  and for half of modern New Zealand history it wasn’t the situation in New Zealand.  For decades it may well have led to inefficiently low currency holdings: in a genuinely competitive market there is a reasonable chance that (eg) serial number lotteries would have provided a (expected) return to holders of bank notes.  In the high inflation years –  and especially as interest rates were deregulated –  holding as little currency as possible was the sensible thing to do.

notes and coin

As the chart shows, the ratio of notes and coins (in the hands of the public) to GDP troughed in the year to March 1988 –  when inflation and interest rates were both high (and, of course, returns to holding currency were zero).

At one level, the partial recovery in the amount of physical currency held isn’t too surprising.  Inflation has been low for decades, and interest rates are now very low too.  Holding physical currency isn’t very costly at all.

Then again, there have been huge advances in payments technologies.   Even when I started work, the Reserve Bank still offered to pay its staff (I think perhaps only the clerical and operational staff) in cash, and that wouldn’t have been too uncommon then.  ATMs didn’t exist then –  it was the queue at the local bank branch each Friday lunchtime –  let alone EFTPOS, internet banking and so on.   These days, by contrast, a huge proportion (by value) of transactions occur electronically.  Even school fairs –  often held out previously as the sort of place one really needed cash for –  have often gone electronic to some extent at least.

And although the ratio of cash to GDP is quite low in New Zealand (by international standards –  in many advanced economies something around 5 per cent isn’t uncommon –  there is still a lot of cash around.    The numbers in the chart are equivalent to a bit more than $1000 per man, woman, and child.    For a household like mine, more than $5000.    I’m a slow adapter, and almost always do have a reasonable amount of cash on me, but I’d be surprised if on an average day our household had more than $250 in cash in total (surveys from other countries suggest that isn’t unusual).   I’m not sure I’ve ever had a $100 bill, but Reserve Bank data suggest that on average each man, woman, and child has $400 in $100 bills.

As it happens, last week I was reading (Harvard economics professor) Ken Rogoff’s book The Curse of Cash.   As he notes, in the United States, there is around $3400 per man, woman, and child outstanding in US $100 bills –  while surveys of what ordinary consumers are actually carrying suggest that no more than 1 in 20 adults has a $100 bill on them at any one time.    Rogoff makes a pretty strong case that the bulk of physical currency holdings – even allowing, in say the US case, for the use of the USD in other countries – is held to facilitate illegality.   That could be outright illegal activities –  the drugs trade for example –  or tax evasion in respect of the proceeds of lawful activities.  The likely revenue losses, on his estimates, are very substantial.    The scale of the problem is probably smaller here, but there is no point pretending that the issue is specific to the United States (and, as Rogoff documents, a number of European countries have now put limits on the maximum size of cash payments –  although such rules seem more likely to catch those who comply with the law, rather than those who knowingly break it).

Somewhat reluctantly, Rogoff’s book has shifted my perspective on the physical cash issue.    As a macroeconomist, my main interest in this area in recent years has been to do something about the near-zero lower bound on nominal interest rates.  If the Reserve Bank cut interest rates to, say, -5 per cent, it would be attractive for people to pull money out of banks and hold it in physical currency in safe deposit boxes. If that happened to any large extent it would substantially undermine the effectiveness of monetary policy.  The fear that it might happen has already constrained central banks in various countries, and no one has been willing to cut official interest rates below 0.75 per cent (which was also about how far we thought the OCR could be cut when I led some work on the issue at the Reserve Bank some years ago).

Getting rid of physical currency altogether would solve the problem.  If there is no domestic cash, clearly you can’t hold any.  Of course, you could always seek out foreign cash, but the process of doing that would lower our exchange rate –  one of the ways monetary policy works, and thus not a problem.    But one doesn’t need to get rid of cash –  or even just large denomination notes –  to limit that risk.    There are various clever options that have been developed in the literature (effectively involving an exchange rate between physical and electronic cash), and as I’ve noted here previously, one could achieve the same result by simply putting a physical limit on the amount of currency the Reserve Bank issues, and then auctioning it to the banks (if demand surged this would, in effect, introduce an exchange rate or a fee).    It is disconcerting that, as far as we can tell, no country is properly prepared to use options like these in the next recession (which, in itself, risks exacerbating the recession because smart observers will recognise that governments have fewer options than usual) –  no one has (at least openly) done the preparatory legal work, or prepared the ground with the public.  Our Reserve Bank is, as as we can tell, no exception.

I’ve resisted the idea of getting rid of physical currency on both convenience and privacy grounds.  There is, as yet, no real substitute for cash if –  say –  one wants to send a child to the local dairy to buy the newspaper when one is on holiday.   And the ability to conduct entirely innocent transactions without the state being able to know what one is spending one’s money on (or one’s bank for that matter) remains a very attractive ideal.

And yet….and yet…..I wonder if it is a real freedom now to any great extent anyway.   We might not gone all the way –  yet –  to China’s “social credit” scoring system, but you have to be pretty determined to avoid the gaze of a government determined to find out what you’ve been up to.  Some of that is voluntary –  people choose to carry phones around, for example, which locate you –  and some of it isn’t (local councils put up CCTVs, and so do all too many retailers). AML provisions, and know-your-customer rules are ever more pervasive and intrusive.   Sure, using cash enables one to keep from a spouse what one spent on a birthday present, or where it was bought from, but it is a pretty small space left.

And so perhaps it is best for us to think now about serious steps towards phasing out physical currency.  Rogoff himself doesn’t recommend complete abolition at this stage, but rather ceasing to issue, and then over time withdrawing, high denomination notes.   Our largest note isn’t very large at all (NZD100 is only around USD70) but as I noted earlier a huge share of currency in circulation is in the form of $100 bills, even in New Zealand, which few people use for day-to-day transactions that are both lawful in themselves and where there is no intention to evade lawful tax obligations.   But if we were to amend the law to prohibit the Reserve Bank from issuing notes larger than (say) $20 –  and this is a decision that should be made by Parliament or at least an elected minister, not by a single bureaucrat –  we’d still make small cash transaction easy enough (school fair, or the kid sent to buy the newspaper, while greatly increasing the difficulty of a major flight to cash in the next serious recession, and increasing the difficulty of tax evasion and other criminal transactions.

If the government were to choose to go this way, it would still make sense for active precautions to be taken now to reduce the risk of the effectiveness of monetary policy being undermined even by a flight to $20 notes –   they take up roughly five times as much space as the equivalent amount in $100 notes, but you can still fit a lot of money in a secure vault.   Whatever the mix of measures, it is really important that the authorities –  Bank, Treasury, IRD, government, FMA –  adopt a greater degree of urgency.  No one knows when the next serious recession will be, but it isn’t prudent (ever) to assume it is far away.

And what of the Reserve Bank’s own scheme: the possibility of replacing physical currency with digital Reserve Bank currency?   We need to see more of what they have in mind.  My own long-held prediction is that they are two quite different products –  only the RB can issue physical notes, while anyone can issue electronic transactions media –  and that in normal times demand for a Reserve Bank retail-level digital currency would be almost non-existent.   That doesn’t mean they shouldn’t do it: there is something about the democratisation of finance, in enabling the public to hold the same sort of secure liability banks already can (in their case electronic settlement account balances), and  –  as we saw globally in 2008 –  banks runs can still happen.   Unless society decides to completely up-end the entire monetary system (and I have readers who favour that), we need an “outside money” that people can convert their bank liabilities into if/when they lose confidence in the issuing institution or system.    For most purposes, a digital Reserve Bank retail currency should be able to do that at least as well as physical banknotes.

Most….but not necessarily all (when serious people worry about EMP attacks on/by North Korea, there is no point pretending electronics is the answer to everything).   Those are the sorts of issues that need to be carefully examined, preferably in an open way, rather than with conclusions loftily filtered out to the public when it suits the officials.

Rogoff’s book is worth reading, especially (but not only) if you are new to the issue.  He covers a little of issues I didn’t have space for, including natural disasters (where cash might be more useful than cards, but most people don’t hold much cash anyway, so it actually isn’t that much of a help.)  Like the Reserve Bank paper, he also points out that things like Bitcoin offer a lot less effective anonymity than many people realise.