In another post, I argued that two factors contributed to monetary policy mistakes:
1. Having the wrong target.
2. Having an instrument setting unlikely to hit the target.
In the comment section, John made the following claim:
I think this was a good post and the right post to respond to Tyler with, but it does raise a conundrum. Your argument is basically that policymakers shouldn’t rely on inflation forecasting models and should instead rely on market-based expectations of inflation (or nominal GDP). But how should the market participants forecast inflation (or nominal GDP)? I don’t think nominal GDP is high because the Fed let nominal GDP be high is satisfactory for this purpose (maybe if the Fed actually had the objective to target NGDP).
I believe there are cases where “nominal GDP is high because the Fed let nominal GDP be high” is a useful way of understanding the problem, even if it’s not a complete explanation in the deepest sense of the term.
Consider the following analogy. Fred is depressed. One day while driving down the road, he decides to accelerate his car and plow into a tree. Now consider two possible explanations for the fatal crash:
1. Fred felt depressed, and committed suicide.
2. Fred’s right hand suddenly moved from a position at 12 o’clock on the steering wheel to a position at 3 o’clock on the wheel, causing the car to suddenly veer toward a big tree.
There is a sense in which the second explanation is more scientific, more like “physics”. But most people would find the first explanation to be more useful.
Insiders suggest that by late 2021 the Fed fully understood that NGDP was about to rise well above trend. Outside the Fed, the same perception was widespread. So why did the Fed make this mistake? In a recent comment, Rajat reminds me of an exchange between David Beckworth and Jason Furman, which occurred on mid-2021. Here’s Rajat’s comment:
From when I started reading your blog in 2011, it took me at least a couple of years to grasp the importance of level targeting to your approach. The NGDP aspect was what brought you to prominence, perhaps because it was so intuitive in the midst of a supply shock, but in (my) hindsight is less important. I think the reason that policy-makers and people like Tyler have such difficulty accepting level targeting is that it removes a major policy escape valve. I constantly come back to David Beckworth’s interview with Jason Furman on Macro Musings (from June 2021) where Furman said:
“You in 2019, put down a really elegant framework for nominal GDP targeting. If we were following it now, we would already have lifted off interest rates. And we’re going to, with extreme likelihood, overshoot the nominal GDP target we were on.
So under your framework, you’d have to make up for that with a sustained period of lower than trend on nominal GDP growth. I don’t mean that to pick on you, this experience has destroyed anyone’s plans that they wrote down before. It’s such a weird period. But to me, that says, “I’d like the Fed, if the unemployment rate a year from now is still 5.5%, I’d like the Fed to take that into account, regardless of what’s happening to nominal GDP or prices as an independent problem and issue that they need to take into account.” So I think that anything has to have a dual mandate, but do you look at nominal GDP and the like, instead of inflation? Maybe.”
As it turned out, had the Fed tightened in 1H 2021, the US probably would have avoided most of the excess price level increase it has experienced with little reduction in employment growth. But as you noted, the supply shocks have mostly been reversed in the US. What if that never happened and the Fed tightened in 1H 2021 anyway? Then maybe employment would not have recovered as quickly, and policy-makers would have been under a lot of pressure from people like Furman.
I have enormous respect for Furman, who is an outstanding economist. But in this particular case he got it wrong; policy did need to tighten in order to prevent a big NGDP overshoot. The Fed knew what it was doing. It set rates at zero and did extensive quantitative easing, despite clear signs of above trend NGDP growth. If we go back to the two types of mistakes outlined at the top of this post, the late 2021 error was clearly an example of “having the wrong target.”
More broadly, almost all important Fed policy mistakes are of this type. Commenters occasionally point out that the inflation surge of 2021-23 ended up being much larger than the TIPS market expected in early 2021. That’s true, but this fact doesn’t have the implication that many assume. It does not mean that the Fed was doing the best job it could, and just got unlucky. Under a level targeting regime the inflation surge would have been far milder, some transitory supply-side inflation but no permanent NGDP overshoot. The primary cause of high inflation over the past 5 years has been NGDP overshooting its 4%/year trend line by 11%. That’s a lot!
If I would weight my policy advice in terms of relative importance, it would be:
A 10% weight on using markets to guide policy. (Tactics)
A 90% weight on doing NGDP level targeting. (Strategy)
If you are not aiming for the right target, it doesn’t help much to be a skilled navigator. Market guidance is useful, but it’s not a panacea.
READER COMMENTS
Garrett
Jan 2 2025 at 3:57pm
Removing the “escape valve” is a feature, not a bug
Rajat
Jan 2 2025 at 10:39pm
Thanks Scott. Trying to piece together what happened, it seems that the TIPS and share markets in 2021 believed that the NGDP overshoot in 2H 2021 would mostly result in higher-than-trend real growth, but then realised their mistake over the course of 2022. That led to a sharp rise in TIPS yields and a 25% fall in US equities. Then from late 2022, the share market came to the view that the Fed was in no rush to return inflation to target, and was unlikely to risk a recession to do so. Subsequently, the share market rallied.
In that context, I thought about this:
Assuming the TIPS market is roughly informationally efficient, wouldn’t the fact that it was ‘wrong’ in 2021 imply that the way a level targeting regime would’ve avoided as large an inflation overshoot over 2021-23 would’ve been to tighten much more rapidly in 1H 2022 than otherwise and quite possibly instigate a recession later in 2022? If policy instrument markets get it wrong for whatever reason, what level targeting regimes can achieve it seems to me is either a smaller recession and/or a faster return to target than growth targeting regimes. Would you agree? If so, that’s the risk that central banks concerned about their future autonomy may be wary of.
Scott Sumner
Jan 3 2025 at 4:08pm
I believe under level targeting the policy would have been tighter in 2021, as it was already clear that NGDP was overshooting the trend line. There’s some risk of recession in any policy, but I believe that level targeting is the way to minimize that risk.
The Fed has yet to get inflation down to 2%, so there’s still some risk of recession in 2025.
Rajat
Jan 3 2025 at 8:45pm
It depends on the level target. If NGDPLT, then I agree that policy would’ve been tighter in early 2021. But if price level targeting, then I think less so. Your post refers to both, but because of the reference to the error made by the TIPS market, I was thinking mainly of PLT. I agree even PLT would have been an improvement over what we had. And I agree that there is still a risk of recession if the Fed wants to get inflation back to 2%. But human nature is such that it will seem to policymakers that the risk is more manageable than entrusting monetary policy to a rule and markets.
Scott Sumner
Jan 5 2025 at 1:18am
But it was “flexible” price level targeting, which as Bullard pointed out was effectively close to NGDP level targeting, if it had been symmetrical.
In any case, even the PCE was well above trend by mid-2021 (and rising fast), so policy would have been tighter.
https://fred.stlouisfed.org/series/PCEPI
Thomas L Hutcheson
Jan 4 2025 at 8:16am
But is the under-target inflation that level NGDP targeting calls for optimal? [And there is an implicit inflation target in NGDPLT targeting] If the purpose of inflation target is to facilitate adjustment to ongoing average level of shocks, why would under-target inflation _ever_ be the right policy.
This seems to me to be completely separate from what data on what time scale the Fed might use to set its policy instruments to achieve whatever objective it is perusing.
John Hall
Jan 10 2025 at 9:45pm
As it relates to my comment, I think what you’re arguing is that actual NGDP is determined by the Fed. My argument is more that the Fed can control expected NGDP, but not necessarily actual NGDP.
So for instance, 2021 NGDP rises above trend, you argue that the Fed implicitly set an NGDP target too high. I argue that the Fed implicitly policy such that NGDP growth would be at a more normal pace at the start of the year (without looking up the actual numbers), but they got things wrong based on what else was going on and NGDP ended up coming in higher than they had expected.
Of course, I don’t dispute your other comments on the importance of level targeting.
Comments are closed.