At its next meeting on 3 December, the European Central Bank (ECB) is likely to reduce its deposit rate further into negative territory, from -0.2% to perhaps -0.3%. How low can negative interest rates go?
One example is the experience of Sweden, Denmark, and Switzerland, which have seen money market rates around -50bp and below (Chart 1). But the comparison with the Eurozone is not exact because these are much smaller economies, and negative rates in Denmark and Switzerland have been implemented to counteract the possibility of significant currency appreciation.
In principle, the main constraint on negative rates is the level where it makes sense for banks to substitute banknotes, stored in a vault and yielding zero, in lieu of central bank reserves with a negative interest rate. Whether or not cash substitution proliferates should depend on four factors: the cost and inconvenience of storing large quantities of cash, how negative the deposit rate is, what quantity of reserves is subject to the negative rate, and how long negative rates are expected to persist. For any individual bank, the key question is: does it expect to hold a very large amount of central bank reserves for long enough to justify the cost of converting reserves into physical banknotes, and the inconvenience of not being able to use the reserves as a means of payment.
The Bank of England estimated that it could apply a -50bp negative rate for one to two years before cash substitution became a problem. The Swiss National Bank has adopted a -75bp deposit rate, but considers this to be pushing the boundaries of negative rates. Another point of comparison is that the largest gold ETF, which is essentially in the business of storing gold in a vault, charges an expense ratio of 40bp, while the largest silver ETF charges 50bp.
Interestingly, the Swiss franc has amongst the largest denomination banknotes in circulation: the 1000 franc note (EUR 925). In theory, vault costs should be lower there in comparison to the euro whose largest denomination is “only” EUR 500. A billion francs is about half as much physical paper as a similar value in euros. Lower storage costs should raise the true lower bound (i.e. the rate is less negative) all else equal.
The ECB will have thought carefully about these issues, because all else is not equal. Denmark and Sweden each bifurcated their policy rates, exempting certain portions of bank reserves from the negative rates, thereby allowing the lower bound to decrease further into negative territory. The ECB could proceed down a similar technical avenue. Though this would muddy the ECB’s message slightly, it would give the bank greater leeway to shift market rates further into negative territory, while mitigating the associated tax on banks, and the risks of cash substitution*.
The other important constraint is whether negative rates filter through to bank customers, which would be very unpopular. In practice, this hasn’t happened yet for bank deposit rates either in the Eurozone (Chart 2) or in the other negative-deposit-rate countries (Chart 3). Banks have drawn a distinction between large corporate and institutional deposits, which are subject to negative rates, and smaller and retail deposits, where the negative rate is not passed on, presumably for fear of losing customers. Instead, banks seek to protect their margins via other charges, higher lending rates, etc.
Overall, a level around -50bp seems like the most probable lower bound for the ECB, given reasonable assumptions on the technicalities. The smaller the amount of reserves subject to the rate, the more negative the lower bound becomes. And whatever the ECB does next month, it is set to leave open the possibility of some further reduction of the deposit rate, to encourage the market to price some probability of this.
*Conveniently setting aside a major civil liberty taboo, it is indeed fascinating to contemplate what heterodox policies might be enacted to thwart cash hoarding by banks. Could high-denomination banknotes be imprinted with radio-frequency technology that erodes their value at the same negative rate when the paper enters a bank vault? Simplistically, bank regulators could simply tax vault cash (the amounts are known and declared as a fungible substitute for reserves), or make it illegal to hold cash piles larger than say, 1 billion euros.