In the first and third of his Munich Lectures in Economics (and in an earlier oped in the FT), Kenneth Rogoff argued in favour of phasing out cash, at least high denominations and in some developed economies, see my post. Rogoff emphasised two beneficial consequences. First, the abolition of the zero lower bound on nominal interest rates and thus, the relaxation of a constraint on monetary policy. And second, the abolition of a means of payment that guarantees anonymity and thus, facilitates criminal transactions, money laundering, tax evasion and the like.

Both Rogoff and other academics have discussed the topic before. More than in academic papers, the end of cash has been the subject of intense debate in the blogosphere. By far the clearest discussion I know (and a very comprehensive one) is due to Willem Buiter in a blog post I summarise here. But the list of authors that have contributed to the discussion is much longer. Here is a selective overview:

• As far as solutions to the zero lower bound problem are concerned, Buiter in his post referred to several academic contributions, namely Eisler (1932), Goodfriend (2000), Buiter and Panigirtzoglou (2001, 2003), Davies (2004) and Buiter (2004, 2007). Rogoff in his lectures referred to Silvio Gesell as well as writers in the blogosphere including Mankiw, Buiter and Kimball.
• Concerning the loss of tax revenue due to anonymous currency holdings, Rogoff referred to his own earlier work (Rogoff 1998).
• On April 19, 2009, Gregory Mankiw discussed the zero lower bound in the New York Times. He reported a proposal by a graduate student to relax the bound by taxing currency: The Fed should announce that all notes whose serial number ends in a particular digit would cease to be legal tender within a certain time period; and the digit should be determined by a lottery. (According to Buiter, Charles Goodhart made the same proposal earlier.)
• On May 7 and 19, 2009, Willem Buiter strongly argued in favour of negative nominal interest rates in his FT maverecon blog (see my post). He identified currency’s status as a bearer security as the cause of the zero lower bound and discussed three strategies to relax the bound: Abolishing currency; taxing it (difficult); and separating the medium-of-exchange role of money from the unit-of-account function by creating a unit of account dollar (think of reserves) on the one hand and a medium of exchange dollar (think of currency) on the other. The former would pay positive or negative interest, the latter would pay no interest. Both would trade at an exchange rate, and interest parity conditions would hold in equilibrium.
• Other FT bloggers took up Buiter’s proposal. An early post, on May 20, 2009, is due to Izabella Kaminska in FT Alphaville.
• On April 19, 2012, Matthew Yglesias argued in Slate that the abolition of the zero lower bound would facilitate expectations formation about monetary policy.
• On November 5, 2012, Miles Kimball took up the issue in a blog post. In another post, he discussed Marvin Goodfriend’s (2000) contribution to the debate.
• On April 15, 2013, Izabella Kaminska suggested in FT Alphaville that direct access of consumers and investors to government provided electronic money would allow central banks to bypass commercial banks, improve the monetary transmission mechanism and help end a shortage of safe assets.
• On April 16, 2013, Jean-François Groff argued in FT Alphaville that electronic money should be provided by the government instead of private companies (“digital legal tender”). Governments then could (re-)gain seignorage and consumers would benefit from lower fees and user costs.
• On July 27, 2014, John Cochrane discussed Sheila Bair’s opposition against letting the broader public hold reserves. On August 21, September 17 and September 22, 2014, he approvingly discussed (here, here and here) the Fed’s balance sheet policy from a financial stability/narrow banking perspective (see my post on narrow banking proposals). On November 21, 2014, he interpreted minutes of an FMOC meeting as suggestive evidence of plans to establish segregated cash accounts.

When evaluating the merit of these discussions, it is important to distinguish between (i) introducing government provided electronic money and (ii) doing so in combination with an abolition of currency. Consider first the former option, namely to have the government grant the broad access to central bank reserves. This could be useful as it opened up the possibility to eliminate the risk of bank runs and as a consequence, abolish the fragile and costly system of deposit “insurance.” If, that is, most savers opted to move their deposits to the central bank rather than keeping them with commercial banks. If they didn’t, then governments would most likely feel obliged to continue bailing out depositors in failing commercial banks.

Another advantage of introducing government provided electronic money would be to eliminate a disgraceful contradiction in public policy. Mostly for reasons related to the deterrence of tax evasion, governments increasingly force citizens to use electronic means of payment although these are not legal tender and declare the use of currency illegal although currency is legal tender. In effect, governments force citizens to use liabilities of private companies for their transactions and in doing so, expose citizens to various financial risks. (These risks are partly borne by the public sector, due to deposit insurance, but that insurance creates other problems.) This absurd situation would end if the government provided a legal tender for electronic payments.

But granting the public access to central bank reserves could also create new problems. Inducing savers to move their deposits from commercial banks to the central bank would undermine a central activity of the former, namely deposit financed credit creation. Douglas Diamond and Philip Dybvig (1983) have shown in a classic article that the insurance characteristics of a deposit contract help improve outcomes relative to a situation without such a contract. How large are those benefits? And how large are they relative to the social costs of bank deposits, namely inefficiencies due to deposit insurance (moral hazard) and costs of run-induced fire sales and defaults?

There are other open questions. One concerns the transition from the current system where savers hold deposits at commercial banks, to a new system where they hold central bank reserves. Would the central bank assist commercial banks and convert deposits into reserve holdings? And if not, how could runs be avoided?

In addition, questions of a more technical nature would have to be addressed. Should banks (in the interbank market of reserves) and the general public (when paying their bills) use the same payment system? Or should the existing system linking the central bank and commercial banks be kept separate from a new, to be designed, system that serves consumers? How would monetary policy in this new world look like and how would the monetary transmission mechanism work? Would the central bank lend funds to households, and would it set the same policy rates for banks and the general public?

Turn next to the more ambitious proposal, namely to augment the introduction of government provided electronic money with an abolition of currency. This suggestion is more problematic, because the promised benefits are likely overstated and the costs misjudged. Consider first the benefits. As far as the relaxation of the zero lower bound is concerned, the fundamental objective—to lower real interest rates in order to incentivise earlier consumption and investment—cannot only be achieved through monetary policy but also by tax policy. A trend increase in consumption or value added tax rates acts like a low or negative real interest rate. And even if the objective is to relax constraints on monetary policy rather than relying on fiscal policy, this is feasible without eliminating cash altogether (and without moving to a higher inflation target which is costly for other reasons). As explained by Buiter, all that is needed is a floating exchange rate between reserves and cash. Killing currency amounts to an overkill unless one fears negative consequences due to such a floating exchange rate (see, e.g., Goodfriend, 2009, fn. 23).

As far as the second objective—limiting tax evasion as well as criminal and black economy transactions—is concerned, the elimination of currency is not a sufficient measure. True, those seeking anonymity would need to incur additional costs to secure it. But these additional costs would likely be mostly fixed costs (e.g., fees for incorporating a shell company in Nevada and hiring a lawyer). The implicit tax on black market activity due to the abolition of currency thus would be a regressive one and the revenues it generated would likely be smaller than hoped for. Professional criminals directing large operations could easily afford the higher cost of securing anonymity while the tax dodging middle class plumber in a badly run country could not.

Turning to the disadvantages, eliminating currency has severe consequences for privacy. (Buiter’s suggestion of ‘cash-on-a-chip cards’ could limit those consequences somewhat.) This point is widely acknowledged in the debate but it is not given sufficient weight. Related, forcing savers to hold means of payment—and a significant share of their savings—exclusively with a branch of the government (the central bank) might cause concern, particularly in countries with a history of expropriation.

Finally, there is a completely different reason to be worried about the prospect of putting an end to currency; when pointed to the proposal under question, some mothers I talked to immediately articulated it: In a world without physical money it is harder to acquire basic financial literacy skills. This might appear like a third-order problem, but is it?