Over the past 10 years, there’s been increasing support for NGDP targeting in academia, the media, and among policymakers in DC. But there seem to be two sticking points that prevent it from being adopted as the Fed’s new policy target:
1. What if trend GDP growth changes? Won’t inflation become unanchored?
2. Isn’t inflation targeting much easier to communicate? Most people don’t even know what NGDP is.
It turns out that both of these objections can be easily addressed with a slight tweak of the NGDP targeting regime. This tweak might reduce the effectiveness of the policy very slightly, but it would still be a substantial improvement over the current regime.
I’ve previously suggested a fixed NGDP target growth path, say 4%/year. Instead, the Fed could adjust the target path every few years to reflect changes in the estimated rate of trend NGDP real GDP growth. The target growth path would be estimated trend real GDP growth plus 2%. This would keep long run inflation quite close to 2%, while allowing inflation to move up and down in the short run to help stabilize the economy when there are supply shocks. Fortunately, the trend rate of growth tends to evolve gradually over time. Thus you might have a 4.0% NGDP target for 5 years, then a 4.2% for the next 5 years, and then a 3.9% target path for the next 5 years. Those modest adjustments would not create large business cycles, but would keep long run inflation close to 2%.
Here’s how the Fed should sell its new policy regime in January 2020, when their current policy review is completed:
1. We will continue to have a 2% long run PCE inflation target.
2. We will continue to allow short run deviations from the 2% inflation target as needed to address the other side of the dual mandate, high employment.
3. However, we will no longer use “output gap” models to determine our short run deviations from 2% inflation. Rather we will stabilize NGDP growth at roughly 4% as the preferred method for achieving the employment side of our mandate.
4. While the target rate of NGDP growth will initially be set at 4%, the target growth rate path will be adjusted periodically, as needed to keep long run inflation close to 2%.
This is no harder to communicate than the current regime. Consider the fact that the current regime has a simple and understandable 2% long run inflation target and a complex method for addressing the employment side of the dual mandate, which most people don’t understand. The new system would have a simple and understandable 2% long run inflation target and a slightly less complex method for addressing the employment side of the mandate, which many people still won’t understand. The all-important financial markets can understand either target, but NGDP is simpler.
How can they be so similar? It’s simple. Even under a non-NGDP targeting regime, a successful implementation of policy, i.e. success on the dual mandate, will as a side effect produce a stable path of NGDP. So the advantage of NGDP level targeting is not so much that you are trying to achieve a radically different outcome, rather it’s primarily a method for better achieving what you are already trying to do. (It is substantially different from a simple inflation target, but the Fed now treats employment as equally important in the dual mandate.)
Yes, my modified proposal is not ideal. But it’s less far from ideal than you might assume. George Selgin has persuasively argued that the optimal NGDP growth rate should vary with changes in the growth rate of the labor force–but not productivity. But much of the recent slowdown in trend GDP growth in the US is due to a sharp slowdown in the growth rate of the working age population. So it’s quite possible that this modified regime, which appropriately responds to changes in the labor force growth rate and inappropriately responds to changes in productivity growth, is actually superior to a simple NGDP targeting regime that appropriately does not respond to productivity changes but inappropriately does not respond to changes in the labor force.
Also keep in mind that only long run changes in trend productivity would change the target path, not short run (cyclical) changes such as an oil shock. So it really is almost as good as a simple NGDP growth rate level target.
Over time, as people became comfortable with the new system it could be tweaked again, closer to what Selgin (correctly) argues is the optimal regime. Changes in the NGDP target could be made only on the basis of changes in the growth rate of the labor force, not productivity.
I understand that the Fed may be reluctant to leap into a radically new regime like NGDP targeting. But if they actually believe it’s as good a system as David Beckworth and I think it is, there are ways of putting their toe in the water and getting there in stages. One needs to think about what sort of gradual steps in that direction can be most easily communicated. In this post I’ve tried to show that there are very useful steps that would both substantially improve monetary policy and yet be easy to communicate, even if they fall a bit short of the ideal regime.
Don’t let the perfect be the enemy of the good.
PS. And I didn’t even mention another problem with inflation targeting, it’s much harder to communicate than most people assume. The public thinks it’s about keeping inflation low, and doesn’t understand the “symmetrical” nature of the target. Hence Bernanke faced a firestorm of criticism in 2010 when he announced the intention of trying to raise core inflation from 1% to 2%.
READER COMMENTS
John Hall
Oct 8 2019 at 2:48pm
Good post. I like that you are giving more attention to this hybrid version!
David’s paper made the argument that this hybrid version should underperform classical NGDP targeting. However, his argument hides an assumption about the credibility of monetary policy. At one point (in the Appendix), he argues that monetary policy is credible and that expected inflation equals the NGDP target minus long-run growth. Later when the NGDP target is changed, the expected inflation component of the Philips curve remains the same and does not adjust. This implies that monetary policy is not credible and inflation expectations remain the same. If monetary policy is credible, then the Philips curve should change.
Scott Sumner
Oct 8 2019 at 3:43pm
Does he advocate changing the expected NGDP target? And why would it matter if the expected inflation component of the Phillips curve changed?
John Hall
Oct 11 2019 at 3:10pm
I meant to reply earlier, sorry.
Beckworth does not advocate changing the NGDP target in that recent paper. He argues that a stable NGDP target is better than a varying one.
On the changing Phillips Curve, this is a little more difficult for me to explain without some context…
The appendix to the paper details the implications of a temporary or a permanent supply shock. He compares and contrasts an NGDP target with inflation targeting when there is a temporary supply shock, and then does the same with a stable NGDP target and a varying NGDP target for a permanent supply shock.
If you look at Figure A.3 with the varying NGDP target, he assumes that the Phillips Curve does not shift when the NGDP target changes. Earlier, he had defined the Phillips Curve as having a component for expected inflation. Higher (lower) inflation expectations cause it shift up (down). So since it is does not shift at all, then this means he assumes that expected inflation does not change. However, earlier in the appendix, he also defines expected inflation as equal to the NGDP target minus potential real GDP growth when monetary policy is credible. If you lower the NGDP target and monetary policy is credible, then inflation expectations should fall. This should flow through to the Phillips Curve (though I think David would agree that it does in the long-term) and cause it to shift downward. Since the Phillips Curve does not shift in his model, then he is implicitly assuming that monetary policy is not viewed as credible because expected inflation is not adjusting in response to the change in the NGDP target.
If you don’t assume that inflation expectations change, then it makes perfect sense for the varying NGDP target approach to look worse than a stable NGDP target approach. In his model, the varying NGDP approach only would make sense when monetary policy is viewed as credible and shifts the Phillips curve.
Kevin Erdmann
Oct 8 2019 at 3:22pm
The communications issue seems like a hurdle of perception rather than reality. If the switch in regimes can be made, then communication won’t be necessary. The only reason inflation fears demand constant conversation is because the current Fed regime isn’t trusted to maintain stable prices. And the only reason interest rates demand constant conversation is because the Fed insists on communicating about them in a way that suggests low rates are related to stimulated growth.
Once a credible NGDP growth path is in place, I don’t think anyone will care about those communications. If NGDP was growing at 4% or 5% annually with some stability, eventually I just don’t think people would pay much attention to inflation moving temporarily up and down between, say, 0% and 4% from year to year. That’s not really noticeable in daily life, and few will be talking about it because the primary fear that inflation runs at 5%+ for a decade will be off the table.
Scott Sumner
Oct 8 2019 at 3:41pm
Kevin, You said:
“Once a credible NGDP growth path is in place, I don’t think anyone will care about those communications. If NGDP was growing at 4% or 5% annually with some stability, eventually I just don’t think people would pay much attention to inflation moving temporarily up and down between, say, 0% and 4% from year to year.”
That’s also my view.
Thaomas
Oct 9 2019 at 7:29am
The key is how to establish the new expectation of PL or NGDPL targeting.
Announcing a policy is not enough; the Fed will need to be seen acting on the policy. Clearly people right now are not very confident that the Fed has and will keep to a 2% symmetric inflation target (= PL target starting when?) and not an inflation ceiling for the “prices” portion of its mandate. When have people observed the Fed “doing what it takes,” or doing much of anything, to get inflation above 2%? Credibility for the change to a PL target would need to begin to be established by announcing a target that implies needing to raise inflation above 2% for a while and then the Fed acting to make that happen.
Ditto for how to achieve credibility for a 4% NGDP target. (And I still would like to see some discussion of what the optimal rate of inflation to add to whatever rate of real GDP (TFP I suppose) trend is assumed.) If that target were announced from today’s base we would not see the Fed “doing what it takes” until we get a recession, and then, whatever if did, would not be seen as a change of policy until sometime in the future when GDP was growing faster than 4% for long enough to make up for the recession shortfall.
Thaomas
Oct 8 2019 at 4:04pm
Still seems like PL targeting is a better first step, leaving the “employment” portion of the mandate to be determined ad hoc as at present. Also if they are going to adopt a new targeting system, I’d hope they have a look at whether 2% is the optimal PL target, given (or not given) their reluctance to use anything but ST interest rates as instruments of intervention.
LK Beland
Oct 8 2019 at 5:39pm
“Isn’t inflation targeting much easier to communicate? Most people don’t even know what NGDP is.”
I’d argue that most people understand nominal national income better than they understand inflation.
Scott Sumner
Oct 8 2019 at 5:41pm
I agree.
Benoit Essiambre
Oct 8 2019 at 7:15pm
I also find PLT more realistic for practical reasons.
Targeting NGDP properly means recording every single transaction in the economy. This task has a totalitarian flavor, might become unreliable in turbulent times and seems vulnerable to manipulation.
It would be very difficult for a third party to verify that GDP data hasn’t been tampered with. With price, you can take a small sample of the economy and get a good estimate.
George
Oct 8 2019 at 10:06pm
“George Selgin has persuasively argued that the optimal NGDP growth rate should vary with changes in the growth rate of the labor force–but not productivity.” That’s one of two possibilities I like. The other sets NGDP growth to expected weighted factor (labor + capital) input growth.
Todd Ramsey
Oct 9 2019 at 9:41am
For the millions of Americans without Economics education, NGDP is easier to understand than RGDP.
Justin
Oct 10 2019 at 2:04pm
The best part of this strategy is that, if the NGDP target path is set at 5-year intervals, then we get to talk about 5-year plans. People underestimate how much tongue-in-cheek fun that would be. We could make ironic Soviet style propaganda posters about achieving the NGDP path, plus tinfoil-hat libertarians would lose their minds.
Lorenzo from Oz
Oct 12 2019 at 8:30pm
I still find it startling that Australia having the longest ever recorded economic expansion without a technical recession gets as little attention as it does. We are the 20th largest economy in the world by GDP (ppp) and the 13-14th by NGDP. About the same size as Spain’s, not much smaller than Russia’s. But it seems as if you live below the equator, you are invisible, or something. Or we are just lucky, or something.
Yes, we are a quarry with lots of sheep and cows. We have always been a quarry with lots of sheep and cows. We have not ever before been the poster economy for flat business cycles. On the contrary, our business cycles used to be unusually steep.
And the Reserve Bank effectively stabilises total spending. Averaging inflation between 2-3% pa over the business cycle is stabilising total spending.
Yes, I know, preaching to the converted. Yes, I know the problem is the culture among macroeconomists which changes but slowly. But still, this is supposed to an at least notionally scientific discipline where evidence counts.
Comments are closed.