The not-so-new (but very uncertain) neutral

The not-so-new (but very uncertain) neutral

Add to Reading List
Add to Reading List

There has been great debate about when and how central banks in developed nations will start raising rates. There is general agreement that rates should be raised gradually in anticipation of a stronger economy, rather than suddenly and aggressively in response to fears of inflation. This would tend to encourage the start of a tightening cycle sooner rather than later. The FT, discussing this a year ago, suggested that central banks might wait for market tightening at the long end of the curve:

What the central bank wants to see is a period where long-term rates go up first, as the economy improves, before it raises the cost of short-term borrowing. That way banks and financial institutions can still make money by borrowing cheaply and lending it out at higher rates.

That comment was made in mid-2013, when long rates were beginning to rise. But that rise was short-lived, and the trend is now downward. This chart shows that there isn't a lot of room for the Fed to raise short rates:

Main article image

Of course, raising short rates might encourage long rates to rise too. But it didn't in the mid-2000s, when 30-year mortgage rates actually fell as the Fed raised rates. And I suspect that would happen again. In a world full of noise and bubbles, where long-term investment that can ride the waves looks increasingly attractive, the term premium might be very low, if indeed it still meaningfully exists at all. Flattening of the curve is the last thing a central bank embarking on a tightening cycle would want – but it might be what they get.

But some have questioned the rationale for raising rates at all. Pimco's Richard Carida & Bill Gross have postulated a “new neutral” regime in which neutral interest rates remain very low, at or near zero. Carida shows that global leverage since the financial crisis has actually increased, not decreased. He particularly highlights China, whose corporate debt is an astonishing 200% of gdp and whose banking system looks increasingly wobbly, and the Eurozone, whose current low yields belie its high sovereign leverage and poor growth. And he argues that although US private sector deleveraging in recent years has been effective and the energy revolution looks promising, headwinds from the rest of the globe may make it hard to restore growth – a point also made by US policy makers. He summarises his global growth outlook thus (his emphasis):

There is a risk that the global economy not just the U.S. – will be unable to grow and generate inflation at pre-crisis levels for many years to come – even if monetary policy rates remain at zero in nominal terms and negative in real terms.

And Bill Gross then develops it (my emphasis):

The reality is that now, five years after the global financial crisis,average growth in the global economy is modest and the level of global GDP remains below potential. The global economy has not as of today found a growth model that can generate and distribute global aggregate demand sufficient to absorb bountiful global aggregate supply. Unless and until it does, we will be operating in a multi-speed world with countries converging to historically modest trend rates of potential growth with low inflation. 0% neutral real policy rates for many developed and some developing countries will likely be the investment outcome.

But Lombard Street Research (LSR) fundamentally disagree. In a research note, they argue that current expectations of low interest rates for years to come have little to do with the eventual path of interest rates. Central banks do not know what level of interest rates they will reach, but they have every interest in keeping markets calm:

As it starts to raise interest rates in mid-2015, the Fed has every incentive to downplay the extent of the tightening cycle. It does not want a repeat of 1994, or even last summer’s ‘taper tantrum’. But these assurances are not particularly valuable – ultimately, the central bank doesn’t know what level of rates it will reach.

And they point out that we have played this scene before:

...when the Fed started raising rates in 2004, it thought the neutral real rate was in a range of 0.5-2.2%. This was consistent with the view, widely held by investors, that the economy’s indebtedness made it more sensitive to higher rates. Some commentators even argued it would only take a small number of hikes to cause deflation. Yet the Fed eventually raised rates by more than 400bps. And as it did so, it revised up its estimate of neutral, reaching a range of 2.2-3.9% in June 2006.”

LSR argue that investors should expect central banks to behave as they have done in the past, namely raising rates more than investors anticipate. I fear they may be right – and I fear for the outcome.

This chart shows that every recession since 1954 has been preceded by the Fed raising rates to levels higher than NGDP:

Main article image

In most cases they continue raising rates even when NGDP is falling.

Interestingly, this happens regardless of the monetary policy regime. From 1945 to 1971 the policy regime targeted the exchange rate, not the inflation rate, using interest rate rises to maintain the dollar's gold parity. That regime famously ended with the Nixon Shock in 1971, after which interest rates fell and NGDP recovered. Exchange rate targeting was eventually replaced with inflation targeting, after a nasty inflationary shock in the “no-man's land” of the 1970s and early 1980s. David Beckworth shows that the high interest rates of that time were because of high risk premia, which came down once inflation targeting was established:

Main article image

It seems to be important that the central bank targets SOMETHING. High inflation occurred when central banks didn't know what to target: Bretton Woods was broken, exchange rate targeting didn't work any more and inflation was regarded as a fiscal problem (hence prices and incomes policies).

But the high inflation of the 1970s has left deep scars – and policy makers seem to be the worst affected. I read through all the FOMC minutes from 2004-8, and two consistent themes sprang out to me:

  • the FOMC was constantly worried that inflation expectations would become unanchored and inflation would “fail to moderate”

  • conversely, the FOMC was continually upbeat about growth. Even in 2008, when growth was falling sharply due to the coming financial crisis, FOMC members believed that growth would recover and the real risk was inflation.

The result of this was that the FOMC over-tightened. You can see that clearly here:

Main article image

It raised interest rates too late – hence the fast rise in NGDP – and too much – hence its collapse. I exonerate the Fed's monetary policy management for the depth of the 2008-9 crash and recession: in my view the problem there was inept Treasury management and an implicit sovereign default (Lehman). But the US was already in recession long before Lehman.

Inflation fears can be toxic. FOMC minutes from 2008 show that members were aware that growth was falling – but they were still concerned about inflation from spiking oil prices. Consequently they held interest rates too high for too long, as the chart shows. One member even wanted to raise rates in August 2008.

I am not convinced that the typical central bank dynamic – pessimistic about inflation, optimistic about growth – has changed. Growth forecasts from most developed country central banks, as well as from supra-national organisations such as the IMF, have persistently over-estimated growth. Sometimes the effect is laughable, as with the IMF's continual growth revisions for Greece. But it is not the only aficionado of what I call the “hockey stick” approach to forecasting (“never mind the trend, there will be a significant uptick in growth ANY DAY NOW”): the ECB has produced similar forecasts, as has the UK's OBR. To its credit the Fed has so far not joined in the hockey game. I hope it doesn't.

But for this reason, although I think Carida & Gross are correct in their view that the neutral interest rate for developed countries will remain very low for a long time, I think LSR is also right that central banks will get interest rate policy wrong. They will raise rates too late, causing a bubble, then panic and raise them too much for too long, causing a crash - just as they have in every business cycle since WWII. They will do this regardless of what they target. And that is because they are attempting to forecast medium-term future paths from short-term, backward looking information. The FOMC minutes showed me just how hard FOMC members find it to distinguish between signal and noise – and how often they allow their own priors to inform their judgement, at times overriding both their own staff and the markets.

But even a flawed and biased FOMC making its best judgement and getting it wrong now and then is better than a rigid rule-based approach. There is a serious lack of accurate information: all the economic indicators the FOMC uses to make its decisions are subject to revision, sometimes by substantial amounts. And even when the information is accurate, its significance is not always understood: for example, it is painfully evident from the minutes of the meeting of September 16 2008 - the day after the fall of Lehman - that the FOMC did not appreciate what the economic consequences of that catastrophe would be. Furthermore, the effects of monetary policy changes take time to feed through into economic indicators and are not always clearly identifiable. A rigid rule-based approach – or worse, an algorithm - cannot possibly be better than human decision-making when information is inadequate, its significance unclear and the effects of policy changes both delayed and uncertain. Humans at least have hunches. Computers do not. 

So in my view economists are right to reject the GOP's proposal to force the Fed to use a Taylor rule regardless of the circumstances. Rather, there should be extensive improvements to information gathering and forecasting techniques. It may be that as information improves and short-term forecasting becomes more reliable (as it has with weather forecasting) automation of day-to-day monetary management might be possible. But crucial policy decision-making in uncertain and turbulent times should still remain the province of humans. Despite their flaws, they do it best. 

Related reading:

Weird is normal

Slaying the inflation monster - Coppola Comment

No, you can't have your risk-free returns back - FT Alphaville

Ever-shifting Phillips curves - Giles Wilkes

Should Congress legislate so that the Fed is forced to follow policy rules? - Long and Variable

Image of FOMC meeting from Wikipedia. 


Twitter Feed

RT @pmarca: We @a16z were proud to have @vgr as Philosopher In Residence last year. Read his new content series "Breaking Smart"! http://t.…

RT @Frances_Coppola: New Forbes post. The wheels are already coming off the Greek deal: