Negative numbers "make dark of things which are in their nature excessively obvious and simple".
British Mathematician Francis Maseres, 1758
There is an obsession with negative nominal interest rates. People seem to think that they make no sense. And, there is a fixation with keeping track of the fraction of sovereign debt that is trading at negative nominal rates. (At this writing, the number is approaching one-third of the total outstanding.) Clearly many central bankers believe that setting the policy rate below zero is a legitimate use of this conventional instrument, a point that we have supported in the past. But the fact that people are so disturbed prompts us to ask why. In this post, we first discuss why we are confused by this reaction, and then try to identify what might account for it.
From a purely economic perspective, what matters is the real interest rate, not the nominal interest rate. The gap between the two is expected inflation. Borrowers and lenders ought to care about the purchasing power of the money they pay out and receive, not the number of currency units. That means that it is important to adjust nominal borrowing and lending rates for expected inflation: the more prices are expected to rise, the less the future purchasing power of any future loan repayment (or, for that matter, of any fixed nominal quantity).
And it is the real interest rate that is the primary means by which monetary policy influences the economy. Faced with the threat of a recession in a world of low inflation and slow potential economic growth, monetary policymakers probably will seek to drive the real interest rate below zero. For example, a simple, textbook Taylor rule with a neutral real rate of 1% and an inflation target of 2% suggests that if the unemployment rate were to rise even 2 percentage points above its full-employment level, policymakers would aim to drive the real interest rate to zero or below (see our earlier post here.) Today’s neutral real rate may be less than 1%, raising the odds that, with a 2% inflation target, policymakers will seek to drive real rates negative.
How frequently have real interest rates fallen below zero in the past? In the chart below, we show a simple estimate of the real interest rate for the United States, Germany and Japan. It is calculated by subtracting the past year’s inflation from the current one-year government yield (this is equivalent to assuming that expectations of next year’s inflation are equal to last year’s inflation). Notably, real rates measured this way have been frequently below zero, and especially over the past decade. Starting in 1995, the one-year real interest rate has been negative 31% of the time in Germany, 34% of the time in Japan, and 44% of the time in the United States. In other words, no one should be very surprised when the purchasing power of repayments turns out to be less than the purchasing power of loans.
Estimated One-Year Real Interest Rate, 1995-July 2016
Sources: FRED2, Bundesbank, Japan Statistics Bureau, Bloomberg, and authors’ calculations.
And, since interest income is normally taxed, these estimated real interest rates exceed the returns that investors will actually anticipate. Assuming a 30% tax rate, for example, the real after-tax returns on these one-year bonds have been negative 40% of the time in Germany, 35% of the time in Japan, and a whopping 56% of the time in the United States. In other words, from a rational economic perspective, there is absolutely nothing special about negative returns.
So, why all the fuss about negative nominal interest rates? We can think of a number of possibilities. The first is that subzero nominal rates might drive people to accumulate cash. In theory, there ought to be an effective lower bound (ELB) below which people would prefer cash, despite the transactions costs associated with it. But if they do, there is no significant evidence of it so far. In the euro area, for example, currency outstanding has grown by about three percentage points of GDP since mid-2014. And in Japan, the increase in bank notes over the past two years is a negligible one percentage point of GDP.
This is all consistent with our view, expressed in an earlier post, that the cost of holding large quantities of cash is substantial. While vault space may be cheap and plentiful, insurance, monitoring and compliance costs are another story. Anti-money laundering regulations make it complicated for individuals to withdraw large quantities of cash from their banks. Even non-bank firms might have a difficult time if they ask for tens of millions of dollars, or the equivalent. And, in some jurisdictions, if a bank were to request billions of currency units from the central bank, it is unclear whether the central bank would comply or, instead, try to lean against efforts to hoard cash.
That said, if people do start to withdraw large amounts of cash and store it in vaults both large and small, it would be quite damaging. Bank balance sheets could be forced to shrink, reducing lending. The result could easily be a decline in consumption, investment and growth. Put differently, any effort by a central bank to push the nominal interest rate below the ELB would be contractionary, not expansionary. That is why policymakers have not driven nominal interest rates very far below zero, and so long as cash offers a zero nominal return, they are not going to do so in the future.
Another possible explanation for people’s strong reaction is that they view negative interest rates as a signal of central bank desperation (see, for example, Kocherlakota’s concern about policy communication). Some observers may have reacted this way to quantitative easing (QE) when it was first introduced. But, while QE may have seemed abstract to most people, negative interest rates are anything but abstract to the public at large. They see that when they make a $1,000 loan or purchase a $1,000 bond, they receive less than $1,000 at maturity. To many people, this may suggest that something is very wrong. The resulting damage to consumer and business sentiment could very well be counterproductive.
If this were the case, we would expect to see both evidence of a shift in consumer attitudes and changes in saving behavior. On the latter, the saving rate of German households has been stable at around 10 percent since the mid-1990s. And, in Japan, saving as a percentage of disposable income has been low, and occasionally turned negative, in recent years.
Similarly, surveys of consumer confidence, which are available more frequently and are more up to date than savings data, exhibit no deterioration. The following chart provides monthly readings for both Germany and Japan, together with the OECD average. While confidence is lower than it was at its peak several years ago, we see no evidence that the recent move to subzero central bank policy rates has led to any decline. Considering the dimming outlook over recent years in these countries for long-term growth—a key reason that both real and nominal interest rates are so low—it is perhaps more surprising how well confidence has held up!
Consumer Confidence Index: Japan, Germany, and OECD (long-term average=100), 2009-July 2016
Source: OECD Main Economic Indicators.
The fact that there is no hoarding of bank notes, consumer confidence has not suffered, and savings rates do not appear to be rising leaves us with only one explanation for the reactions we are seeing: money illusion. That is, even people in the financial sector may not fully appreciate that, a significant proportion of the time, eroding purchasing power results in negative real interest rates even when nominal rates are positive. As we have written before, zero matters because people don’t like it when the nominal level of their incomes shrink—whether that income is coming from wages or interest payments.
Interestingly, there is a range of experimental research highlighting a role for nominal rigidities and money illusion. For example, in a laboratory asset market, Noussair et al find that people respond differently to inflation and deflation, adjusting rapidly to the former and slowly to the latter. Fehr and Tyran also find nominal inertia after a negative monetary shock (but not after a positive one). And Yamamori et al observe that price changes increase the tendency for people to over- or under-consume.
Anecdotally, we are reminded of some stories from the mid-1980s. U.S. inflation had fallen from near 15 percent to about 4 percent. As one would expect, nominal interest rates had plunged by about 10 percentage points, too. However, retirees living off their interest from their bonds and CDs were upset, as they saw their income falling. Any attempt to explain that they had been eating into their capital by spending all of their interest income during the high inflation period—and worse, having it taxed away—fell on deaf ears. We suspect that attempts to explain negative interest rates would meet with the same bewildered reaction. Instead, negative rates reinforce the dire warnings that have bombarded retirees regarding the impact of low nominal interest rates on their savings.
Economic policy is about trying to influence behavior. The whole point is to give people an incentive to change what they are doing. If the economy is slowing, policymakers will lower interest rates to entice people to consume and invest more. When the economy is overheating, they will raise interest rates to get them to reduce their consumption and investment.
Critical to the success is that households and firms react in a predictable way. Given the difficulty they have in understanding negative nominal interest rates, we need to ask seriously whether this is a policy that will have its intended effect. That is, policymakers who follow our earlier analysis and lower rates below zero (arguably before they engage in quantitative easing) ought to pay close attention to the impact on cash holdings, consumer confidence and savings patterns. And if they wish to use negative interest rates routinely as a stimulative policy tool, they need to educate the public that it is a normal tool—not one reserved only for desperate times. Otherwise, they could inadvertently stoke damaging fears.