Is the Fed Getting Warmer?

The Fed might be warming up to NGDPLT in the eyes of some, but George Selgin says otherwise.

Relax: the Fed isn't about to catch fire or melt. This isn't about that sort of warming. It's about a different, more benign sort of Fed warming that my pals at the Mercatus Center claim to have discerned. Still, I don't believe them. Call me a Fed warming skeptic if you like, but so far as I'm concerned, it's all fake news.

Since Jay Powell announced the Fed's new average inflation targeting (AIT) strategy last week, both Scott Sumner and David Beckworth have welcomed it as a step, albeit only a tenuous one, toward their own (and my) preferred policy of NGDP level targeting.  Scott calls "average inflation targeting…a tiny step forward," though one that will allow the Fed more discretion than a move to price-level targeting would. David likewise observes that, although it isn't quite an NGDP level target, AIT "is a step in that direction."

Scott and David also seem to agree that a move to strict price level targeting (PLT) would take the Fed still closer to an NGDPLT ideal. Were they to play "Hot and Cold" with Jay Powell, with the intent of steering Fed policy toward NGDPLT, they'd shout "getting warmer!" any time he moved Fed policy further away from strict inflation targeting (SIT), with its "bygones are bygones" response to inflation over- and undershooting, and closer to price level targeting (PLT), as he did last week.

I only wish I could join that chorus! Alas, I can't, because I don't think it's correct to say that a move from SIT to AIT (and hence toward PLT) is also a step toward NGDPLT. It could turn out that way. But it could also turn out to be a step the other way, or one that leaves the Fed neither warmer nor cooler than it was before. So while I agree with Scott that the Fed has granted itself more scope for discretion, I'm not convinced that its new policy has a bright side.

Supply and Demand

Regular Alt-M readers know that I agree with Scott and David 99 and 99/100ths percent of the time. So why not now? Unless I'm mistaken, in arguing as they do Scott and David are implicitly downplaying supply shocks; that is, they're assuming that any price-level changes the Fed encounters reflect disturbances to aggregate demand. When that's so, PLT and NGDPLT are equivalent, and it logically follows that any change that takes the Fed closer to PLT also takes it closer to NGDPLT.

But if demand shocks were really all that mattered, there wouldn't be any reason to recommend NGDPLT targeting at all. Instead, Scott and David and other Market Monetarists could spare themselves a lot of effort by simply pleading for price-level targeting. Again: if supply shocks aren't important, NGDPLT and PLT are practically one and the same. And of the two, PLT is far easier to market as a special case of inflation targeting.

But supply shocks do matter! And now is no time to be arguing as if they didn't. As I type this, the U.S. economy is reeling from what may well turn out to be the mother of all adverse supply shocks! That makes the case for preferring NGDPLT to PLT as strong as ever. It's also why a step away from SIT and toward PLT can't really be said to be a step toward NGDPLT.

Just how do NGDPLT and PLT differ in practice when supply shocks occur? Consider the lockdown supply shock, but suppose for argument's sake that it didn't involve a concurrent shock to aggregate spending: people keep spending on stuff; but because fewer are working, less stuff is being produced. Under PLT, the Fed would either have to keep the price level from rising at all in response to reduced real output, or it would have to act quickly to restore an increased price level to its original path. Either policy would require a reduction in NGDP. In contrast, NGDPLT would simply let prices rise. In the case of a positive supply shock, NGDPLT would let the price level decline, while PLT would call for offsetting monetary expansion.

Clearly, a world in which supply shocks matter is no longer a world in which PLT and NGDPLT amount to the same thing. Instead, as I pointed out in my last post, when it comes to supply shocks, NGDPLT resembles strict inflation targeting (SIT), with its bygones are bygones attitude toward unanticipated price level movements. In a "supply shocks only" world, a step from SIT and toward PLT is a step away from NGDPLT.

Of course, in the real world, central banks have to deal with both supply and demand shocks. For that reason, the Fed's move from SIT to AIT and hence toward PLT isn't obviously a move away from NGDPLT. But neither is it obviously a move toward it. Whether it turns out to be one or the other or neither will depend on the shocks the Fed confronts. And the one thing about shocks that we know for certain is that we don't know which ones are in store.

An Asymmetric Target

Many critics of the Fed's move to average inflation targeting observe that the new policy commitment lacks credibility. In particular, they note that, should the Fed ever find it has exceeded its long-run inflation target, it may be unwilling to expose the economy to below-target inflation—and correspondingly lower employment—for the sake of making up for its mistake.

It's a perfectly plausible argument—so much so that I've made it myself! But during his recent Macro Musings interview with me, David Beckworth argued that it also reinforces the argument that the Fed's new strategy will take it closer to NGDPLT.

In making this particular argument, David doesn't ignore supply shocks. On the contrary: his argument is strongest for the case of an adverse supply shock, for if the Fed is likely to allow upward price shock bygones to be bygones, its response to such a shock will indeed resemble that of an NGDP level-targeting central bank.

So far, so good. But the hitch this time is that a Fed that lacks the fortitude needed to compensate for inflation overshooting from adverse supply shocks is no less likely to lack the fortitude needed to make up for inflation overshooting that results from adverse demand shocks. In the latter case, "asymmetric" AIT would be less like NGDPLT than symmetric AIT. In short, we once again have a policy innovation—not making up for above-target inflation only—that could take the Fed further away from, rather than closer to, an NGDPLT ideal, depending on the sorts of shocks the Fed ends up facing.

No Cigars

So, my fellow Market Monetarists, let's not be breaking out the cigars and champagne and confetti just yet! Things could indeed get better, NGDPLT-wise. Or not. It's just too early to tell. We must wait and see. In the meantime, instead of telling the Fed that it's getting warmer, I suggest we keep putting the heat on it.



  1. Thanks. Good analysis of the relationship between the various alternative targeting policies and NGDPLT.
    But for a long time I have been reading your articles and following the links from them to find any economics rationale for preferring NGDPLT to every other policy, given the assumption of a monetary regime of fiat money.
    I have seen no such justification. A proof that a given policy will optimize the ability of government to achieve its arbitrarily determined aggregate measurements is not a proof that the policy will optimize social benefit.
    Could you give such a link?

    1. I don't know about "optimizing": I doubt any real-world arrangement can do that; and of course noting that any particular achievable policy isn't "optimal" is no argument against it unless you can show that another policy is. I have given my reasons for doubting we can re-establish an (imperfect) gold standard in various places. As for why I think NGDP targeting is the best among fiat strategies, I spell my basic reasons for that out at greatest length in my '97 book, Less Than Zero. New edition here:

      1. Prof. Selgin,


        "As for why I think NGDP targeting is the best among fiat strategies, I
        spell my basic reasons for that out at greatest length in my '97 book,
        Less Than Zero. New edition here…"

        …is the answer to the question I asked, thanks.


      1. I've not yet read "Less than Zero", but I read the above paper. It is very helpful in explaining your theory.

        My interest is in placing your theory in the context of Machlup, Hayek, and Wicksell. (To keep the list short).

        Do you take their criticisms into account, or do you generally consider them too weak to consider (or…invalid)?

        Hayek and Machlup criticize the theories of Keynesians and Monetarists because, to start with, they ignore the effects of the temporal dimension of the productive structure. This, plus a positivist methodology, supposedly lead to them assigning too much importance to statistically constructed "aggregate", non-discounted prices, at the expense of looking at distortions in real, time-discounted prices caused by a deviation of market and natural interest rates.

        Also, to them ignoring the resultant unsustainable distortions in "time-of-production" caused by the distorted interest rates and real prices.

        Would they criticize you on this basis?

        Perhaps in your book you do address their theory, but in this paper and everything else I've read from you, you never mention distortions in the temporal structure of production, nor in real prices, as opposed to artificially constructed aggregate prices. (An example of a real, discounted price would be the present implicit price of producing a new coal mine today, in terms of a Hershey bar in 10 years.)

          1. Thanks much! I am re-reading "Theory of Free Banking" now. It's been a while, and I didn't remember this section.

            (I think at the time I'd not read any of Machlup. Maybe not even Prices and Production by Hayek, which is critical to understanding him! You were the one who assigned that homework to me a few years back, in some thread or article 🙂

            I am starting to recall this controversy now, about whether buying bank liabilities is or is not voluntary savings. I must have read the debate on or somewhere, but I don't recall how I resolved the question in my mind, if I ever did.

            Odd, because it is at the very heart of the question, 'is Austrian theory correct about the trade cycle?'

          2. Not really: there are Austrians on both sides of the debate here. (But Machlup and Hayek have it right IMHO. Machlup is better for being more precise. Mises couldn't make up his mind. The current Mises Institute crowd adhere to his incorrect statements while disavowing his correct ones.

          3. Re: The Theory of Free Banking: Money Supply under Competitive Note Issue and

            The "'Productivity Norm' versus Zero Inflation in the History of Economic Thought"

            Both of these address a different question than I was asking about. Sorry I wasn't clear.

            (But the question they DO answer is really interesting: your position in relation to Machlup, Hayek, Mises, and others about the relationships between gross aggregate quantities of spending, credit, employment of factors, and aggregate "price level")

            I will try to clarify.

            Hayek and Machlup analyzed money, banking, the trade cycle, and aggregates under the assumptions of modern capital theory, pioneered by Böhm-Bawerk, in which the temporal structure of the production system is of critical importance: it doesn't just matter that someone invested 100 mu in labor, land, and production products THIS YEAR in SOME processes making SOMETHING. It matters WHAT consumer goods they (ultimately, and unknowingly for the most part) are producing and WHEN HOW MUCH OF those will be delivered, and HOW MUCH investment WHEN will be required to keep the altered processes going till the consumer goods are delivered. All this in relationship to the changing time preference of consumers.

            They reject the possibility of understanding the effects of different money and banking regimes under the more or less "classical" capital theory, resurrected by Keynes and the monetarists, as well as the modern "mathematical economists": the positivists) where investment is treated single lump value, and all production processes create final household consumption goods at the same time, after no delay, or after the same delay.

            Including these two references (I could only read title and abstract for the latter, which was behind a paywall), none of your explanations of why NGDP targeting is better than the alternatives you've considered, and according to what criteria, makes ANY mention of a temporal structure of capital and credit, nor of inter-temporal imbalances.

            Lacking any mention of the importance of these "Bawerkian" factors, I am assuming that you reject Hayek and Machlup specifically with regard to capital theory and the trade cycle (NOT with regard to the non-temporal aspects you mention, where you agree with Machlup and Hayek partly agrees with you.)

  2. "it may be unwilling to expose the economy to below-target inflation—and correspondingly lower employment—for the sake of making up for its mistake."

    It was certainly willing to expose the economy to that risk 2008-2016 and again 2020-02 to the present.

  3. "it may be unwilling to expose the economy to below-target inflation—and correspondingly lower employment—for the sake of making up for its mistake."

    It was certainly willing to expose the economy to that risk 2008-2016 and again 2020-02 to the present.

Comments are closed.