Close

This page had been redirected to a new URL, please update any bookmarks.

Font Size: A A A

macroblog

« Try, Try Again | Main | What the FOMC Said: More Clarification »

December 28, 2012

Nominal GDP Targeting: Still a Skeptic

In a few days the clock will run out on another year of disappointing economic growth in the United States and, generally speaking, in the world. It is inevitable and appropriate, then, that the year-end ritual of looking forward by looking backward will include an assessment of whether more or better policy can contribute to a pick-up in growth that failed to materialize in 2012.

To this discussion, Harvard professor Jeff Frankel brings some fresh thinking to the not-quite-fresh notion that the Fed should adopt a nominal gross domestic product (GDP) targeting approach as a replacement for existing central bank practice—described by Frankel and others as policy driven by an inflation-targeting framework. What I particularly like about Frankel's proposal is the fact that he offers up a practical roadmap for using the Fed's current communications tools to transition to an explicit nominal GDP targeting framework. If I were inclined to think such a move would be a good idea, I would view Frankel's proposal with some enthusiasm. Alas, I am not yet so inclined.

As to the case for skepticism on theoretical grounds, I commend to you this excellent post by Mark Thoma at Economist's View. But Professor Frankel suggests a case for nominal GDP targeting on practical grounds by appealing to this counterfactual:

A nominal GDP target for the US Federal Reserve might have avoided the mistake of excessively easy monetary policy during 2004-06, a period when nominal GDP growth exceeded 6 per cent.

Maybe. Average annual real GDP growth over those three years was just over 3 percent, compared to the Congressional Budget Office (CBO) estimates of potential GDP growth of just under 2.5 percent. That's not a big difference, but more importantly the average gap between the level of real GDP and the CBO estimate of potential was just 0.3 percent of average output—essentially zero. The importance of this so-called "output gap" becomes evident if you read the Michael Woodford interview referenced in the aforementioned piece at Economist's View. In that interview, Woodford says, "The idea was to talk about a price level, as opposed to the inflation rate, but a corrected price level target where you add to it some multiple of the real output gap." So for him, something like this measure would be a key element of his proposed monetary policy rule.

What if, rather than some measure of nominal GDP, the 2004–06 Fed had instead been solely focused on the inflation rate? You can't answer that question without operationalizing what it means to be "focused on the inflation rate," but for the sake of argument let's simply consider actual annualized PCE inflation over a two-year horizon. (In his press statement explaining the Federal Open Market Committee's (FOMC) latest decision, Fed Chairman Ben Bernanke suggested using a one- to two-year horizon for inflation forecasting horizon to smooth through purely transitory influences on inflation that the central bank would inclined to "look through.") Here's the record, with the period from 2004 through 2006 highlighted:

121228b

If you really do think that there was a policy mistake over the 2004–06 period—and in particular if you believe the FOMC during that should have adopted a more restrictive policy stance—I'm hard-pressed to see what advantage is offered by focusing on nominal GDP rather than inflation alone. In fact, you could argue that that the GDP part of the nominal GDP target would have added just about nothing to the discussion.

I add the observation in the chart above to my earlier comments on an earlier Frankel call for nominal GDP targets. To summarize my concerns, the Achilles' heel of nominal GDP targeting is that it provides a poor nominal anchor in an environment in which there is great uncertainty about the path of potential real GDP. As I noted in my earlier post, there is historical justification for that concern.

Basically, anyone puzzling through how demographics are affecting labor force participation rates, how technology is changing the dynamics of job creation, or how policy might be altering labor supply should feel some humility about where potential GDP is headed. For me, a lack of confidence in the path of real GDP takes a lot of luster out of the idea of a nominal GDP target.

Dave AltigBy Dave Altig, executive vice president and research director at the Atlanta Fed

 


December 28, 2012 in Federal Reserve and Monetary Policy, GDP, Monetary Policy | Permalink

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341c834f53ef017ee6b9785e970d

Listed below are links to blogs that reference Nominal GDP Targeting: Still a Skeptic:

Comments

I think you're right to be a skeptic on targeting NGDP. Nominal GDP bounced between 5-7% post the 2001 recession. And that's after all the revisions. Maybe more to the point is that the Great Recession occurred, in large part, because of hyper credit growth. It's likely that rates were kept too low for too long, which ultimately pushed the reach for yield process, which enabled the excess debt creation. We won't go into the abdication of the regulatory bodies.....

Posted by: stewart sprague | December 28, 2012 at 06:53 PM

Most people who advocate NGDP targeting today advocate level path targeting, not growth rate targeting. I don't believe that your "historical justification" applies in this case. Indeed, I think it makes the case for level targeting (of either the price level or NGDP, but there are reasons to prefer the latter) relative to the current system which centers on a growth rate target for the price level (in other words, an inflation target).

The question is whether the Fed should forgive itself for missing earlier targets. Under the current system, there is near-100% amnesty, which has the potential to make the nominal anchor ineffective, in the case where the Fed keeps making the same mistake over and over, as it did in the 1970's. A level path target during the 1970's would have forced the Fed to tighten when it missed its targets on the upside, in order to get back to the target path. This is true for either a price level path target or an NGDP level path target.

Of course the downside of using a level path target is pretty obvious: you need to produce "unnecessary" recessions and/or inflations to compensate for earlier misses. But I think the improvement in credibility (particularly when the zero bound comes into play, but also in a situation like the 1970's, where forecast errors were serially correlated) would be worth the cost. And I don't think there's much of a case to be made that price level path targeting is better than NGDP level path targeting, unless the price level is your only mandated objective.

Posted by: Andy Harless | December 28, 2012 at 07:21 PM

"To summarize my concerns, the Achilles' heel of nominal GDP targeting is that it provides a poor nominal anchor in an environment in which there is great uncertainty about the path of potential real GDP."

I think that most people who are blogging in support of NGDP level targeting would be puzzled by this comment. Yes, NGDP level targeting sets the nominal anchor in terms of nominal incomes, not the price level. But most of what are called "welfare costs of inflation" seem to correlate better with variations in nominal incomes than in prices. For instance, "natural" interest rates are better described as correlating with NGDP growth rates than with inflation rates; nominal wages correlate with NGDP rather than the price level.

George Selgin (in "Less than Zero") advances a rather complex argument that because variations in RGDP are generally due to firm- or sector-specific changes, nominal income targeting interacts better with price stickiness and similar imperfections.

In general, resiliency to supply-side instability is generally seen as a key _benefit_ of NGDPLT: and this should be all the more true when the RGDP path is uncertain.

Anyway, Sumner has come up with a 'compromise' proposal (dubbed a NGDP/inflation hybrid) which aims to stabilize NGDP in the short/medium run while still keeping a stable inflation target in the longer run: see www.themoneyillusion.com/?p=18145

Posted by: anon | December 29, 2012 at 10:13 PM

What matters for monetary policy is how it affects expectations. In that context, you're either missing or ignoring a couple of things about the Market Monetarist position.

First, a level target rather than a growth rate target really matters. The former is equivalent to the latter but with the additional assurance that misses will be made up for. That assurance means that over all but the shortest runs, the implicit growth rate target will be hit. That can only help when expectations are your real target.

In the Thoma post you link to, Woodford also points out that as a pratical matter, an NGPD level target is about the best the Fed can do. A policy rule has to (i) be simple enough that you can explain it to Congress and the public, and (ii) be straightforward enough that it doesn't arouse suspicions that the Fed is cooking the books. We have enough conspiracy theorists out there already. Let's not feed them even more by employing a target that looks like the Fed could be manipulating it via arcane calculations.

As you say, we don't know the future path of potential output. But that's hardly an excuse to add to the uncertainty by refusing to adopt clear and predictable Fed policies.

Posted by: Jeff | December 30, 2012 at 08:02 AM

Isn't there a point at which the policy stance should be to stop intervening in markets, to let the free market determine interest rates like any other price, and for the Fed to stop targeting ANYTHING? Why are we even entertaining never ending central planning as a solution to anything?

Posted by: Chris | January 01, 2013 at 09:49 PM

Isn't there a point at which the policy stance should be to stop intervening in markets, to let the free market determine interest rates like any other price, and for the Fed to stop targeting ANYTHING? Why are we even entertaining never ending central planning as a solution to anything?

Posted by: Chris | January 01, 2013 at 09:49 PM

David,
Just to add to Andy's point, advocates of NGDP level targeting argue that it's precisely becasue of uncertainty around estiamtes potential output is NGDP trageting should be adopted. They argue that has long as the central bank keeps nominal spending on, say, a 5% trend line, there will be niether demand side recessions (mass unemployment) nor high inflation. In other words AD will be stable and this will produce a stable macroeconomic environment. Whether inflation is 2% and real output growth at 3% or inflation is 3% and real output grows at 2% is of no concern.

Posted by: Gregor Bush | January 04, 2013 at 02:13 PM

Anybody here ever heard of Goodhart's Law?

Posted by: Thomas Esmond Knox | January 13, 2013 at 11:54 PM

Most people who advocate NGDP targeting today advocate level path targeting, not growth rate targeting

Posted by: myVegas hack cheats tool | February 07, 2013 at 03:15 PM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in