Beyond the zero lower bound on nominal interest rates

Rajiv Shastri

14 August 2010



With the threat of deflation looming, attention is once again focused on monetary policy. Many economists have made suggestions through which negative interest rates can be applied to bank reserves and consumer accounts. The primary hurdle appears to be the ability of banks and consumers to hold currency which would be protected against negative nominal interest rates (NNIR), defeating their efficacy and diluting transmission to the real economy.

Banks are the interface between a nation’s monetary authority and its citizens. As such, they fulfil two primary responsibilities in an economy. These are:

  • Provide banking services encompassing savings and credit, and
  • Ensure the availability of fiat currency.

In undertaking the second efficiently, banks need to hold a certain stock of physical currency which, in a positive nominal interest rate environment, is a cost. Recognising this, in 1959, the Fed allowed banks to set off such stock of physical currency against their regulatory reserve obligations.

But even without this facility, banks are free to hold as much currency as they deem fit, thereby presenting the first hurdle to NNIR. If the Fed imposes NNIR on reserve deposits, it is entirely possible that banks would switch over to hoarding currency. In the absence of empirical data on bank behaviour in this environment, perhaps our best option is to consider evidence of opposing behaviour in a positive nominal interest rate environment. Bennet and Peristianii (2002) provide valuable insights in their analysis of the manner in which banks managed vault cash in a positive nominal interest rate environment. They conclude that an increase in the federal funds rate resulted in a decline in the vault cash holdings of “unbound” banks. It would be fair to assume the inverse in a NNIR environment, especially since most banks hold excess reserves and can be considered “unbound”.

This can be alleviated somewhat by excluding currency holdings from required reserves. In the current environment however, when bank reserve deposits with the Fed far exceed regulatory requirements, this would only result in banks holding required reserves with the Fed while holding the rest as currency to preserve its value.

Assuming a solution to the above is possible, it is likely that NNIR incidence on banks would lead to its incidence on consumers as well. In the absence of any legal or regulatory binding, consumers would resort to hoarding currency to preserve its value as well. This presents the second hurdle to NNIR.

Proposals to make negative nominal interest rates a reality

So far, suggestions on NNIR have looked at both these hurdles as one, proposing solutions that discourage hoarding of currency regardless of the hoarder’s constitution. Greg Mankiw suggests a possible solution on his blog (Mankiw 2009), whereby the Fed announces that, a year from now, it will randomly pick a digit from 0 to 9 and cancel all currency notes ending with that number. This suggestion, attributed to an unnamed student, would result in an effective nominal interest rate of negative 10%. This would allow banks to charge a NNIR of less than 10% per year to consumers who would prefer depositing their physical currency in a bank to avoid a 10% loss. However, apart from Willem Buiter’s assertion that taking away legal tender status from currency notes need not have an impact on their value, this system can easily be circumvented by holding physical currency for 364 days and depositing it in the bank for just one day thereafter. This would impose a nominal cost (one day’s NNIR) unless banks charge NNIR for the full year on the day the Fed picks the number. In this case as well, assuming banks would be permitted to deposit currency with the Fed to avoid the 10% loss, competition between banks would ensure negligible, or zero, loss to consumers.

Willem Buiter offers three further suggestions in his column (Buiter 2005).

  • Abolishing currency
  • Taxing physical currency holdings, and
  • Decoupling the unit of account from the currency

On the second point, Buiter himself admits that taxing currency holdings is too administratively cumbersome to be undertaken effectively. As for points one and three, one can safely say that three is a variation of one. Under proposal one, if currency is abolished the dollar would continue to be the unit of account. Its physical form would be replaced by private alternatives which represent the dollar in part or completely. In this case, the Fed would lose control over currency in circulation; something that may not be desirable from a long-term perspective. Under proposition three as well, the dollar would remain the unit of account but would be abolished as currency to be replaced by another – in this case “rallod”. Since this would be controlled by the Fed, it would continue to control currency in circulation. Despite this, Buiter’s belief that the non-physical dollar may be subject to NNIR with the rallod interest rate at zero relies on the Fed’s ability to “announce a credible appreciation of the dollar in terms of rallod”. This, much like the Fed’s ability to announce a credible inflation target in the current environment, is extremely doubtful. If it wasn’t, the need for NNIR would not arise. Moreover both propositions one and three would suffer from Buiter’s criticism of Mankiw’s proposal – unless accompanied by threat of confiscation or penalty.

Jumping the hurdles

As outlined above, these suggestions deal with currency holdings as a single problem in an NNIR environment. To achieve both efficiency and social acceptance, a solution must recognise the differences between banks and consumers. In dealing with the first hurdle, banks’ ability to hoard currency, it is suggested that bank vaults (including ATMs), and the currency within, be treated as the property of the Federal Reserve System, with banks given credit for their currency holdings in their reserve account with the Fed. The Fed would rely on bank declarations about the quantity of vault cash, as it currently does for the purpose of maintaining reserves. The second measure, to strengthen the efficacy of the first, would prohibit banks from holding cash on an overnight basis for any purpose. The combined effect would free the Fed to charge a NNIR on bank reserve accounts.

Consumers on the other hand, with no regulatory and legal bind, present a far greater challenge. The social and legal acceptability of any measure would, therefore, depend on its ability to discourage currency holdings without infringing on their fundamental rights. Instead of Buiter’s dramatic suggestion of abolishing all currency which could have significant unintended consequences, it is proposed that the Fed demonetise currency notes of denominations greater than $1 giving advance notice of say, 90 days. Within this 90 day period, citizens may deposit such currency notes in banks in exchange for their full value, either in $1 notes or a credit in their account. Buiter’s concern regarding persistence in value of “cancelled” currency still holds, but if partial demonetisation is undertaken in a manner that does not subject currency holders to a loss, it is expected that currency holders would surrender the demonetised denominations rather than risk lack of persistence in their value. Demonetisation of high denomination currency notes may also bring significant social benefits.

Lower denomination currency would increase the inconvenience, and cost, associated with storage dramatically – and thereby discourage hoarding by consumers. This would free banks to charge a NNIR on customer account balances.

This would, like Mankiw and Buiter’s suggestions, need to be protected against emergence of foreign currency notes as an alternative means of exchange, an outcome which can be avoided only in two ways. The first would be for all reserve currency nations to co-ordinate execution of this measure. The second would be to criminalise the possession of foreign currency notes. The former is definitely the preferred option.

While this may not be as watertight as abolishing currency, the possibility of it being socially and legally acceptable is far higher. It avoids unnecessary hardship to consumers, is achieved without infringing on their fundamental rights and does not result in random losses.


Bennet, Paul and Stavros Peristiani (2002), “Are US Reserve Requirements Still Binding?”, Federal Reserve Bank of New York Economic Policy Review, 8, 1.

Buiter, Willem H (2005), “Overcoming the Zero Bound on Nominal Interest Rates: Gesell's Currency Carry Tax vs. Eisler's Parallel Virtual Currency”, Discussion Paper 96, Hitotsubashi University Research Unit for Statistical Analysis in Social Sciences

Buiter, Willem H (2009a), “Monetary Policy Under Conditions of Financial Turmoil”, European Central Bank’s Invited Speakers Seminar, Frankfurt, 8 May.

Buiter, Willem H (2009b), “The wonderful world of negative nominal interest rates, again”,, 4 June.

Goodfriend, Marvin (2000), “Overcoming the Zero Bound on Interest Rate Policy “, Journal of Money, Credit, and Banking, 32(4):1007‐1035.

Mankiw, N Gregory (2009), “Observations on negative interest rates”, Greg Mankiw’s Blog, 19 April.



Topics:  Global crisis Monetary policy

Tags:  monetary policy, zero lower bound

Independent Macroeconomic Consultant