More NGDP Targeting Concerns: Gavyn Davies Edition

Gavyn Davies has a new column where he responds to Michael Woodford's call for a NGDP level target.  Davies is always a good read, but this time he raises some concerns about NGDP level targeting that are unwarranted.  He claims  that the Fed may be uncomfortable with a NGDP level target because it might unmoor inflation expectations, it might be seen as time-inconsistent with the Fed's long-run objectives, and finally it may be too late to return NGDP to its pre-crisis trend.  While understandable, the first two concerns are without merit under NGDP level targeting.  This approach to monetary policy actually anchors long-run inflation expectations and provides a credible way to commit.  The last concern has more merit, but even here it is not a clear-cut case. Let's look at each of these concerns in turn.

Here is Davies on the first concern :

[T]here is one obvious reason why the Fed might feel uncomfortable with some aspects of the Woodford approach: it might unhinge inflation expectations from the 2 per cent anchor which has been in place for more than a decade, and which was so costly to establish in the 1980s and 1990s... And greater uncertainty about the future rate of inflation could damage the real economy.

This same concern was raised earlier by John Cochrane in his response to Michael Woodford's speech.  As I noted in my reply to Cochrane, this concern assumes that the temporarily higher inflation would be viewed as a discretionary, ad-hoc surge in inflation.  But that is  not the case under a NGDP level target:

A level target anchors long-run inflation expectations, but allows for temporary catch-up growth or contraction in NGDP so that past misses in aggregate nominal expenditure growth do not cause NGDP to permanently deviate from its targeted path.  Woodford notes that currently NGDP is anywhere from 10-15% below its trend (and thus expected) growth path.  Any increase in inflation under this target, therefore, would not be some ad-hoc temporary increase but part of a systematic approach that would return NGDP to its trend. 

The point is that these temporary increases (or declines in the case of a boom) in inflation would be seen as part of a systematic response of monetary policy that simply returned nominal incomes to some stable growth path.   And the expected stable nominal income path would anchor long-run inflation expectations as I noted in my reply to Cochrane:

 

The black line has NGDP growing at a 5% annualized rate.  Then, at time t a negative aggregate demand (AD) shock causes NGDP to contract through time t+1.  There is now an a NGDP shortfall.  To make up for it, the Fed must actually grow NGDP  significantly faster than 5% to return aggregate nominal spending to its targeted level.  For example, if NGDP fell 6% between t and t+1 it is now 11% under its trend.  Next period the Fed must make up for the 11% shortfall plus the regular 5% growth for that period.  In short, the Fed would need to grow NGDP about 16% between t+1 and t+2 to get back to trend.  There might be temporarily higher inflation as part of the rapid NGDP growth, but over the long-run a NGDP level target would settle back at 5% growth.  Nominal and thus inflationary expectations would be firmly anchored. 

 Woodford, himself, made this point in his paper:

[S]uch a commitment would accordingly require pursuit of nominal GDP growth well above the intended long-run trend rate for a few years in order to close this gap. At the same time, such a commitment would clearly bound the amount of excess nominal income growth that would be allowed, at a level consistent with the Fed’s announced long-runt target for inflation.

In short, Woodford is saying a NGDP level target firmly would anchor long-run inflation expectations and other nominal variables. So this first concern is not an issue.

What about Gavyn Davies' second concern?  Here it is:

A second reason for possible Fed concern is more institutional in nature, concerning the long term credibility of the Fed. Professor Woodford’s recommended target framework risks being seen as time inconsistent. The optimal policy for the Fed today is not optimal for a future FOMC to pursue, once the economy has recovered, and inflation is above 2 per cent. The pressure on the future FOMC to renege would be enormous. The markets know this today, so would the commitment to the NGDP framework be credible in the first place?

No, under a NGDP level target any temporary easing (or tightening) in the short-run is very time consistent, because it is returning nominal income to the Fed's long-run growth path target.  That's kind of the point of level targeting, it coordinates short-run policy moves with long-run policy objectives. Under a NGDP level target, everyone understands the "catch-up" nominal income growth is temporary and tied to an objective.  The public would also understand that once the target path was hit, nominal income growth would slow to trend.  They would come to expect it.  In other words, there are no inconsistencies between the short-run and long-run.

Moreover, a NGDP level target would create so much credibility that the markets would end up doing the heavy lifting.  If the markets know the Fed will do whatever is necessary (i.e. buy or sale as many assets as needed) to return nominal income to its target growth path, than this expectation causes the public to start adjusting spending and investment choices today.  For example, let's say the Fed announced that QE3 was now aimed at raising NGDP 10% to hit its pre-crisis trend and that it would start buying every week enough securities to make sure that it happened.  This would be a huge slap to the face of the market and cause a major rebalancing of portfolios.  This rebalancing would create wealth effects, improve balance sheets,and ultimately spur nominal spending.  The Fed would not have to purchase that many securities.  Just the threat of doing so would be sufficient.

Because a NGDP level target would create such credibility and thereby improve management of expectations, this approach would also  tend to automatically reduce the occurrence of AD shocks in the first place.  Michael Woodford agrees:

A commitment not to let the target path shift down means that, to the extent that the target path is undershot during the period of a binding lower bound for the policy rate, this automatically justifies anticipation of a (temporarily) more expansionary policy later, which anticipation should reduce the incentives for price cuts and spending cutbacks earlier, and so should tend to limit the degree of the undershooting. Such a commitment also avoids some of the common objections to the simple Krugman (1998) proposal that the central bank target a higher rate of inflation when the zero lower bound constrains policy.

Credibility is not an issue for well-implemented NGDP level target. Angst

Gavyn Davies' final concern is that the price level, one component of NGDP, is already back to trend and that the trend path of potential real GDP has declined since the crisis. Therefore, returning NGDP to trend would ultimately result in an permanently higher price level path than expected, since closing the real GDP gap would not suffice by itself to close the NGDP gap.  That is a reasonable concern that others have brought up before.  The problem with this concern is that it is based on Davies' estimation of a trend line.  Others, like Tim Duy, have fitted lines that put the price level currently below its trend.  Similarly, the decline in potential real GDP may prove to be temporary once the economy starts roaring again, a point made by Mark Thoma.  If so, then the closing the real GDP gap may suffice for closing the NGDP gap.   Still, there is a lot of uncertainty over how much of the NGDP gap can be closed. One simple solution suggested by Scott Sumner is to set a slightly lower target growth path for NGDP than the pre-crisis trend.  This concern, therefore, should not be a game changer for whether the Fed adopts an explicit NGDP level target. 

Overall, then, Gavyn Davies concerns about implementing a NGDP level target are unwarranted.  If the Fed decided to follow an explicit NGDP level target then inflation expectations should not become unanchored and credibility should not be an issue.  Now one might reply that the Fed could never really commit to a NGDP level target in the first place.  That is a different argument and if true points to deeper problems at the Fed that would affect the effectiveness of any targeting regime the Fed adopted, not just a NGDP level target.  This too, however, could be addressed by Congress changing the Fed mandate to NGDP level targeting.

P.S. Scott Sumner makes the overall best case for NGDP level targeting here.