Negative central bank interest rates

For any observer of central banks, the last few years have offered much to think about.  In the immediate aftermath of the financial crisis of 2008, we saw the aggressive use of monetary policy, as interest rates in almost all major economies were reduced to near zero, levels never before seen.  Then from about 2011 we saw central banks make active use of their balance sheets, buying financial assets outright under what has become known as Quantitative Easing (QE) and in the process growing their balance sheets to sizes far beyond the pre-crisis norm.  And since about 2012, we have seen official rates for some bank deposits with the central bank reduced below zero, thus breaking what central bankers and economists alike long believed was the “zero lower bound” or ZLB – in other words, the “iron fact” that interests rates could not go below 0%.

But now official rates have gone below zero;  indeed in Switzerland the official rate stands as low as -0.75%.  And this prompts a number of questions.  If low interest rates are good for struggling economies, does that mean that negative interest rates are even better?  And if yes, are they better the more negative they are?  And finally, how negative can they actually go – if zero is not the lower bound, is it -1%, or -2%, or is there in fact no limit?

Even the answer to the first of these questions is unclear. And the answers to the follow-up questions are even more uncertain.

John Maynard Keynes was very clear about the issue of negative interest rates, stating in his General Theory that “the rate of interest is never negative”.  This was the earliest and clearest statement of the ZLB theory, formulated by Keynes some 80 years ago in the midst of the Great Depression, and broadly speaking it has been the consensus view of most academics and central bankers ever since.  And if nominal interest rates did by some chance go negative, it was universally assumed that they couldn’t go below, say, -0.5%, because below that rate, households and companies would start hoarding cash.

Yet now five major central banks – the European Central Bank (ECB), the Bank of Japan (BoJ), the Swiss National Bank, the Swedish Riksbank and Danmarks Nationalbank – all have negative interest rates of some kind or other;  and in the Swiss and Danish cases, well below -0.5%.  Yet there is no rush to cash and the world goes on.  Consensus and the ZLB theory were clearly wrong.  But why?

There are actually a number of questions around negative interest rates.  Why use them?  How do they work?  What are the side effects?  And what will happen with them in the future?

The general thrust of central bank monetary policy for the last 7 years or so has been two-fold:  to stimulate economic activity and, more latterly, to lift economies out of the deflationary trap and to return them to positive inflation.  And negative interest rates are merely the latest salvo from central banks in this 7-year campaign.  In this sense they follow on from, and have the same objectives as, the initial easing of rates in 2008-2011 and the various rounds of QE from then.

There are theoretically two ways in which negative interest rates should work and should help central banks achieve their aims.  First, negative interest rates should encourage banks to deploy their large reserves with central banks, e.g. by spurring bank lending and risk taking, and also (where banks choose not to lend) by buying assets.  The theory here is that any of these uses of the reserves is better for the bank than paying the central bank to warehouse them, and all of them are, ceteris paribus, stimulative for the economy.  Unfortunately, the evidence that they work in any of these directions is as yet extremely vague – although it is true that negative interest rates have only been around for a comparatively short time, since 2012 in Denmark, since 2014 in Sweden, Switzerland and the Eurozone and since early 2016 in Japan.

However, there is another way in which negative interest rates are supposed to work – openly pursued by the Swedish, Swiss and Danish central banks, and although not mentioned by either the ECB or the BoJ, very much hoped for by them too.  This is by depreciating the currency.  Here, the evidence is more mixed, but there is clearly some impact.  Unfortunately, the evidence also implies that the effect is fleeting, and moreover, the more countries introduce negative interest rates, the more difficult it is for one country to use them to weaken its currency against everyone else’s.  As long as Sweden, Denmark and Switzerland (three small economies) were doing it on their own in Europe, the policy (broadly) worked. But once the mighty ECB joined in, it became much more difficult for anyone to achieve the same result.

Meanwhile, there is one major negative side-effect of negative interest rates.  This has to do with whether or not they are passed on to borrowers and depositors. For there to be economic benefit to the wider economy, banks must pass on the lower interest rates to borrowers, even though from the banks’ own perspective, to not pass the rate cuts through is advantageous, as they then benefit from higher net margins and improve their balance sheets.  This sets up a tension in the banks between benefitting the general economy and benefitting themselves;  but the tension is hardly a new one, as it is the same whether interests rates are -1% or +10%, and competitive pressures will mean that generally, interest rate cuts are passed through to borrowers.

But the problem is not so much borrowers as depositors.  Should banks pass the negative rates through to savers as well?  If they don’t, they squeeze their margins.  But households (in particular) have an aversion towards paying banks for looking after their money, and any bank that chose to pass on negative interest rates to its retail depositors would quickly find that they decamped to friendlier shores – i.e. banks that have not followed suit.  But not passing on the negative rates means that they actually erode banks’ balance sheets and thus work counter to any hope of stimulating further growth of credit.

To put this more clearly, for monetary policy to be able to influence the real economy, the banks have to act as a conduit, passing the changes in interest rates through to their customers.  An easing of monetary policy (ie a lowering of rates) only benefits the real economy if the banks pass the lower rates through to ordinary people.  If they do not, if the conduit is broken, then monetary easing ceases to act (positively) on the real economy and starts acting (negatively) on the banking system.

Central banks are not unaware of this danger, of course, and have tried to square this circle by introducing tiered structures, where they actually only levy negative interest rates on a small part of banks’ deposits.  Meanwhile, market rates tend to follow the lowest short-term rate, meaning that the impact on the exchange rate is hopefully achieved anyway.  However, while this overcomes the immediate problem, it creates further complications for when the policy will eventually be reversed.

In spite of this, negative interest rates, having arrived, are almost certain to remain.  There are at least two reasons for this.  First, for countries that openly target the exchange rate – e.g. Switzerland and Denmark – they clearly do work. That in itself is reason enough.  Second, for economies attempting to achieve an inflation target – Sweden, Japan and the Eurozone – while things are not so clear cut[1], it remains the case that where conventional policies have been tried and failed, unconventional policies must also be tried.  This is not just a case of “We must try something; negative interest rates are something; therefore we must try negative rates”. Although negative interest rates may not be a panacea, they should weaken the currency, which is inflationary, and they may, if deployed for long enough, also boost the growth of broad money and credit, and hence also activity and eventually inflation.

This being said, there is clearly some level below which interest rates cannot go. We don’t know exactly where it is, although studies by the Bank of Canada and by Danmarks Nationalbank imply perhaps -2%.  One reason is the impact on banks’ balance sheet.  Another is that, at this stage, the cost of keeping a deposit with the central bank finally begins to exceed the storage cost of cash.

Which brings us to the final point: Why has there been no rush to hold cash?  The answer is simple.  Cash does not lose its nominal value, true.  But it is quite inconvenient to use.  Yes, you can pay for your weekly food shop with cash.  But attempting to pay for anything costing – say – more than £200, let alone more than £20,000, would invite suspicion and becomes increasingly awkward.  And there are major costs involved in holding cash.  You need the space and you need the security.  For a family, or for a small company with one workplace, it may possibly be doable – but is your storage secure enough?  For a company with many workplaces, it becomes thoroughly inconvenient.  And you quickly need a lot of space.  The table below shows how much cash you can store in one cubic metre of space, using the biggest denomination banknote in each country; and also how many cubic metres you need for the equivalent of one billion dollars.

Amount of cash that will take up 1 cubic metre 
Highest denomination  1m3 $1bn = ?m3
USA $100 $85,951,400 11.3
Eurozone €500 €337,254,000 2.7
UK £50 £33,369,350 21.2
China ¥100 ¥74,147,900 88.0
Switzerland Fr1,000 Fr660,710,000 1.5
Japan ¥10,000 ¥7,277,590,000 16.1
Source: Oxford Economics

So, to sum up: having turned out to be possible, negative interest rates will remain part of central banks’ armouries.  But they work in a patchy way and they have substantial drawbacks as well as their benefits.  And this is before we have even tried exiting from a negative interest rate policy, which may prove very interesting – as in disruptive – as well.

The famous (although mythical) Chinese curse is “May you live in interesting times”.  Central banks, and central bank watchers, have not come to the end of their interesting times.


[1]  In the case of Sweden, the need for negative interest rates is in fact not clear cut at all, despite inflation which is well below the Riksbank’s preferred level of 2%. The economy is growing strongly, with real GDP growth averaging 0.9% per quarter since Q1 2014 and accelerating, the currency is not too strong, as shown by the large current account surplus, broad money and credit growth are rapid and household debt is surging. There is a build-up of imbalances which may well turn out to be far more serious than below-target inflation once interest rates begin to be normalised.


This article has been co-authored with Gabriel Stein, Director, Asset Management Services at Oxford Economics