Last week the International Monetary Fund released a paper prepared in its Monetary and Capital Markets Department by five researchers (one a former RBNZer). The title? Negative Interest Rates: Taking Stock of the Experience So Far. It isn’t an official IMF view, but it seems unlikely that a paper of this sort would have been published if the senior management of the department were not broadly comfortable with the messages it contains. There is an accessible summary of the paper on the IMF’s blog.

As the authors note, a number of these sorts of survey papers have been done over recent years, but recent is almost always better when (a) the experiments with modestly negative interest rates are really quite recent themselves, and (b) there is a steady flow of new papers attempting to get or other angle of how negative (policy) rates might, or might not, have worked. And with policy interest rates now lower than ever – in New Zealand too – it is not as if the issue has no continuing relevance. Even if we get through this pandemic downturn without any more countries deploying negative policy rates, who knows when the next more-economically-founded downturn would be.

I won’t claim that the paper is an easy read for someone coming new to the issue, but by the standards of such papers much of it is pretty readily accessible, and there is plenty of summary material (arguably to the point of repetitiveness).

There seems to be quite a range of views among central bankers themselves on the potential of (mildly) negative policy rates, even across pairs of economies and financial systems that are otherwise very similar. In New Zealand, this was the latest from the Reserve Bank, in last month’s MPS.

Just a shame they hadn’t used the previous decade to sort out operational readiness to deploy the tool.

On the other hand, the Reserve Bank of Australia (and apparently especially the Governor) has been really quite dismissive on the possibility of a negative policy rate tool, for reasons that they have never really sought to articulate.

So what do the authors of the IMF paper have to say? These paragraphs are from their Executive Summary

It is not without nuance, and as the authors note in the text unpicking the effects is rarely easy, but overall it is a pretty story. It was music to my ears, having been championing the case for having negative rates in the toolkit here, and generally consistent with the (subset of the) papers that I had read and conversations I’d had, but I was still quite pleasantly surprised to see it in an IMF paper, especially perhaps when taken with the final paragraph of the Executive Summary

You might not like negative policy rates, but you might not have much choice. I found that conclusion particularly interesting because the authors are more confident than I am that central bank large scale asset purchase operations have had a material and useful macroeconomic stabilisation effect.

I’m not fully persuaded by some of the authors’ stories. For example, they claim the flow-through into corporate deposit rates has been greater than that for household deposit rates because “it is costlier for companies to switch into [physical] cash”. I don’t really buy that argument. You or I might find it easy to hold an extra $200 in our wallets, but storing securely $50000 or more of cash just isn’t that easy, and it is really the conversion of large scale holdings (as distinct from transactions balances) that is at issue here. By contrast, for big investment funds conversion to physical cash would be more feasible if central banks pushed policy rates “too deeply” negative. We don’t actually know how deep is “too deep” here, but as the authors note there has, so far, been no sign of large scale physical cash conversions yet. There has been a hunch that it would be unwise to push beyond about -0.75 per cent, but no central bank has yet been willing to push the point to find out. My own interpretation of why household deposits rates mostly haven’t fallen below zero is some mix of (what the authors report) material increases in fees charged by banks on household deposits, and (perhaps not unrelatedly) a sense that it isn’t worth facing the aggro that might come from charging a negative interest rate on household deposit when, at most, it is a few tens of basis points involved. Threshold effects sometimes matter.

The idea of a “reversal rate” has had some play in the literature and debate on negative rates, including being touted by some bank economists here. This is the idea that a move to a negative policy rate might actually have the paradoxical effect of tightening overall monetary conditions, perhaps by tightening lending margins and reducing the willingness of banks to lend. Generally, the IMF authors are not persuaded that this theoretical possibility has been a real world outcome in the countries (euro-area and Denmark, Japan, Switzerland, and Sweden) that have run negative policy rates. And some evidence has suggested that whatever banks do, corporates sitting on large deposits facing negative rates have been encouraged to increase physical investment (transmission mechanism working as one might hope).

Reflecting on the IMF paper and the wider issue of negative policy rates, three points strike me:

The first is a reminder of just what small changes in rates are being dealt with when researchers try to unpick the effects, and how few changes there are to study. The Swiss National Bank’s policy rate of -0.75 per cent is the lowest anywhere. By contrast, as the IMF researchers note, in studying the effect of cyclical swings in monetary policy we are often dealing with policy rate fluctuations of 500 basis points or more (RBNZ in the last recession -575 basis points), and whereas policy rates used to be adjusted quite often, there just have been many changes in the last decade. Somewhat related to this, one negative rate is not necessarily quite like the other: the various central banks that used the tool have also typically introduced tiering-type regimes to attentuate the effect (especially on returns to core holdings of settlement cash by banks). With a handful of countries, unavoidable selection bias in the choice of those countries, small adjustments and infrequent fluctuations, any conclusions are inevitably going to be provisional.

The second is to note that over the 12 months or so since Covid became an issue, although almost all central banks claim to have done quite a lot with monetary policy (a) no central bank that had not already had negative policy rates has moved to introduce them, and (b) none of the central banks with negative policy rates have cut them (even though all other advanced country central banks have cut their policy rates). I don’t purport to know why that is, and really hope some smart and careful researcher in the area has a paper in the works on the subject. In the case of the already-negative central banks, perhaps it really is that they think they have already reached the effective lower bound (ELB) and that, although cuts so far have been useful and stimulatory, any further cut at all would be too risky, and either ineffective or counterproductive. That might make some sense, although the IMF researchers nicely illustrate the absence of any systematic shift to physical cash thus far (although in New Zealand, coincident with the cut in the OCR to near-zero currency in circulation was 13 per cent higher in January than in January 2020). A year ago I would not have believed an “operational unreadiness” explanation for no further countries moving to use negative rates – given the 10 years or so advance notice they have all had – but the revealed failure of the RBNZ to be ready (even when amenabe to using the tool), and the Bank of England even now, suggests there might be more to this story than I had thought plausible. Another interesting piece of research for someone would be to dig into the experience of the negative rate countries and find out how, and how quickly, they came to have systems that were operationally ready.

The third and final point is related to the first. The greatest extent of negative policy rates is really only playing at the margin. Central banks that have used negative rates appear to have found them useful, and (for example) the IMF survey tends to back up such a view. And yet a decade on, not one central bank (or government – and it is likely to be in effect a joint responsibility) – appears to have taken any steps at all to remove, or sharply reduce/attenuate, the effective lower bounds, by a wedge to preventless the limitless conversion at par from settlement cash balances to physical cash. There is no sign any central bank had done so in the 2010s, and there has been no hint of any fresh urgency in the year since Covid dispelled any wishful thinking that macroeconomic conditions would mean rates could really only rise from where they had got to.

The issue here is not about deciding to cut the policy rate more deeply, but about optionality. If macroeconomic circumstances – weak inflation, probably hand in hand with above-normal unemployment – meant that much more macro policy action was warranted do monetary policymakers have all possible tools at their disposal. And do markets (and firms and households) believe they have? Believe in the efficacy of asset purchase programmes all you like (I don’t, especially when- as in New Zealand – it just comes to swapping one government liability for another) but no one has ever deployed a programme that purports to be as effective as 500 basis points of policy rate cuts, and it would be exceedingly rash to believe such recessions will never happen again. Perhaps the world’s central bankers are now all big fans of fiscal policy – not just as short-term income relief – but (a) even if so, they can’t ensure fiscal policy is used, and they still have macro stabilisation responsibilities, and (b) if they really want to give up on monetary policy they should probably surrender their autonomy and simply become operational branches of finance ministries. It seems negligent to have done nothing about easing an obstacle to using monetary policy that exists only because of a rather arbitrary series of state interventions in the first place (banning private notes, and the innovation that probably would have come with them, while insisting on invariant conversion at par between central bank notes and central bank deposit liabilities).

I’m genuinely puzzled why nothing has been done, either in the quiet times – the idea time to socialise these ideas and new rules and procedures – or in the difficult conditions of the last year. There is no obvious good explanation, leaving either subtle ones (too secret for the public to know) or negligent ones.

As it happens, our Reserve Bank came a bit closer to addressing the issue openly than I’ve seen from others. In a speech given a year ago yesterday – at a time when the Bank was still oblivious to the wave about to break over them, Orr included this in a discussion on tools under consideration

That second sentence was right to the point (and I recall welcoming it at the time). But we’ve heard not a word more from them either, even though only recently (see above) they have reaffirmed their view that a negative OCR has a valuable place in the toolkit. If a modestly negative OCR does, why not the possibility of a deeply negative one? Convince people that you have a credible tool of that sort, and would be willing to act aggressively to deploy it, and you are less likely ever to need it, since expectations will do some of the work for you. If you fail to do so, you risk recessions lingering longer than they need to, something inconsistent with the thrust of inflation targeting whether in its 1989/90 articulations or this government’s (cosmetically different) new one.