Why I’m deeply sceptical about deeply negative rates

Given that we work in the realm of the dismal science, where nothing is certain, and everything open to question and debate, a useful attribute when writing, is a large dose of humility, and a healthy respect for alternative views. Thus, when reading an article, even if you don’t agree with the content, you can at least engage intellectually with the author at the same level.

Kenneth Rogoff’s “The case for deeply negative interest rates” which is published as one of our Think Outside articles, and originally written for Project syndicate, is light on these attributes. And that doesn’t help the reader to entertain his basic premise, which is that substantially negative interest rates are the best way to drag the global economy out of the doldrums.

This is an opinion that I do not share. The alternative viewpoint, that such policies provide a mixture of positive and negative impacts, and that as rates fall, the negatives begin to outweigh the positives, is not a dogmatic standpoint, but one that has arisen from observation of low and negative rates in multiple economies over recent years. It is also supported empirically (for example, in this study from Bath University) though I concede that the empirical evidence is mixed. I will return to this later.

Here is the article summary as a backdrop for our subsequent critique: “Only monetary policy addresses credit throughout the economy. Until inflation and real interest rates rise from the grave, only a policy of effective deep negative interest rates, backed up by measures to prevent cash hoarding by financial firms, can do the job”.

And immediately alarm bells start ringing as in just two sentences all other policy options and opinions are dismissed while promoting substantial negative rate cuts as the only response worth considering.

This isn’t a great start in my opinion (others may disagree), but let’s take a closer look at some of the subsequent comments, and follow them with some remarks of our own.

That may well be the case. But equally, there have been times in recent economic history where monetary policy has been utterly impotent. For example when policy rates failed to have any impact on the rest of the yield curve, as we experienced for some time during the so-called bond yield conundrum. Sometimes, policies just don’t behave in practice the way the textbooks say they should, and that is usually not because the theory is wrong, but because textbooks oversimplify the reaction functions. Such deviations from expected outcomes are, in my view, much more likely when policies deviate substantially from their normal operating ranges, for example, substantially negative policy rates.

I see what is being implied, but is there perhaps a confusion between debt-service and debt repayment here? Negative rates can help with the former, but if you can’t pay the principal, it won’t help these entities much. Greece’s debt crisis is a good example of these differences.

While empirical evidence is mentioned to support the negative interest rate proposition, the article cites just one recent piece. I’ve done a bit of google searching myself, and I can find numerous pieces of research on both sides of the fence, including the Bath University piece I referenced earlier. But any empirical evidence that does exist can only be with respect to moderately negative rates as employed by the European Central Bank or Bank of Japan, since the sort of substantial negative rates Rogoff is championing haven’t been implemented anywhere. Indeed, much of the positive opinion on negative rates seems to come from research by, and on behalf of, the central banks that have undertaken negative rates themselves (for example, this piece by the ECB). Are these impartial peer-reviewed studies? The ECB’s recent use of other policy measures seems to say almost as much about what they think about negative rates as these published endorsements.

Again, I think this may miss the point. The issue that is facing the firms, cities and states that Rogoff believes will be helped by negative interest rates during this Covid-19 crisis is not debt-service costs, but basic cash-flow. What I believe is needed, and what seems to be happening in most economies around the world, is large-scale lending to alleviate the fact that corporate earnings have all but dried up, backed up by huge dollops of fiscal stimulus to support demand. The guarantees that the article decries, provide the protection for lenders to extend this necessary lending without having to worry about future capital losses. Reducing the price of money to negative might help a bit, but only at the margin, because this really isn’t the problem, as it doesn’t guarantee the lenders that they get their money back. Negative rates also have significant negative consequences for some parts of the economy – creditors and savers, for example.

I would very much like to see the justification for this claim. This doesn’t chime with our experience at all. So far, the evidence from credit markets is that financial institutions like pension funds have been piling into this stuff. Recent data from BofA indicates that cash holdings of financial institutions are now only marginally above historical norms, though they have been higher.

There is a good reason for the blizzard of objections, and that is the growing evidence of a so-called “reversal rate”, at which the negative impacts of low interest-rates outweigh the positives. This need not even require rates to turn negative but could happen at low positive rates. I don’t believe it is helpful to describe anyone who does not share your views in this complicated field as “fuzzy-headed”, and I don’t believe that citing his own 2016 book for support adds any weight to his argument. I’d add a similar disclaimer to any statement which starts “it is not rocket science”, which precedes some other comments. Had it been rocket science, it would all be very much simpler and less open to debate. Rocket science, unlike economics and financial markets, operates according to robust and predictable Newtonian physics.

The underlying assumption underpinning much of Rogoff’s argument seems to be the textbook assumptions that substitution effects of rate policy changes always and everywhere dominate the consumption (investment) / savings decision. Most textbooks represent this relationship as a downward-sloping straight line. But that is only a stylistic representation, and most of these books were written long before negative interest rates, or even very low nominal positive rates were even considered a possibility.

The more likely reality, in my view, is a non-linear relationship. An interest rate (which is the price of money) is basically the cost of having consumption (or investment) now instead of waiting to have it later. Alternatively put, it is the reward for waiting to have more consumption later (and so more consumption in aggregate). Higher rates reward saving with more consumption later, and more in total when simply aggregating the present with the future and not worrying about net present values.

But with all prices, if they fall far enough, another effect can come to dominate – and that is the income effect. As rates fall, there may be substitution of present consumption for future consumption, but if they fall far enough, falling expectations of future consumption may deter even present consumption. If this sounds familiar, then you may be in your 50s or 60s looking at how miserably performing your retirement savings pot is, and wondering how dismal an existence in retirement you will have on the predicted returns.

And all of this assumes that the financial industry would happily keep lending at negative rates even though sharply negative rates would undermine the entire maturity transformation model on which most bank lending is based.

Much of what I take issue with in this note is the veneer of certainty with which the author makes his claims without, it seems, all that much to support them. I think the article pays insufficient heed to the practical difficulties of the unintended negative consequences of negative rates on the financial system and relies on (very mixed) empirical support for something quite different to what he is proposing.

This is only my personal view. But why not read his article and decide for yourself.