The importance of level targeting

I have a new podcast interview with David Beckworth. In the interview I keep bringing up the importance of level targeting. My fear is that the Fed will eventually get there, but fail to announce the policy in advance. Most of the benefit from level targeting comes from the initial announcement effect, not from the subsequent attainment.



14 Responses to “The importance of level targeting”

  1. Gravatar of Ray Lopez Ray Lopez
    20. May 2020 at 11:29

    Sumner believes in the power of a press release? It’s a standing joke on Wall Street that an ineffective company decision that gives a short pop to the price of a stock is a mere “press release effect” But for Sumner, the ‘press release’ is the entire value-add?

  2. Gravatar of ssumner ssumner
    20. May 2020 at 11:44

    Ray, Yeah, why didn’t I think of that?

  3. Gravatar of Jim Jim
    20. May 2020 at 13:11


    If IOER went severely negative, what is the risk that reserves are exchanged for physical currency to avoid the penalizing negative rate? Any negative rate where you would start to worry?

  4. Gravatar of MJ MJ
    20. May 2020 at 14:45

    I’m starting to think the Fed is full of cowards. A price level target would obviously better help them achieve their mandate, but I think they’re scared of the criticism. Most of the population wouldn’t even notice or care if a price level was announced, but they’ll inevitably get heavily critisized from a small subset.

  5. Gravatar of Thomas Hutcheson Thomas Hutcheson
    20. May 2020 at 18:16

    @ MJ

    It’s certainly a mystery.

    The process of getting onto 2% pa unless it is strung out over years, would certainly be of real benefit.

  6. Gravatar of ssumner ssumner
    20. May 2020 at 18:37

    Jim, That might happen to some extent, but in theory banks could also put a negative IOR on all reserves, including vault cash. They won’t do that, but it’s theoretically possible.

    MJ, I agree that they fear criticism from a small but influential subset.

  7. Gravatar of dtoh dtoh
    20. May 2020 at 22:52


    Or…. the Fed doesn’t have any legal obligation to accept deposits in excess of the reserve requirements, and as I have suggested many times, they could just charge of a withdrawal fee for incremental cash withdrawals beyond a target rate of cash growth (presumably the target rate for cash growth would be approximately equal to the target rate for NGDP growth.)

  8. Gravatar of Postkey Postkey
    21. May 2020 at 00:27

    ‘“Sumner: Right, so we started with a basic principle that the Fed has two ways to boost aggregate demand or nominal spending and those are increasing the supply-base money and reducing the demand for base money.’
    When QE ‘started’ in Nov. 2008, what was the percentage increase in supply-based money?
    What ‘happened’ to the demand for base money?

  9. Gravatar of Lawrence Lawrence
    21. May 2020 at 02:34

    Scott, I would be interested in your thoughts on the idea of targeting cumulative NGDP rather than the level? Let’s make the assumption that we will see a V-shaped recovery (just an assumption). NGDPLT still would see a large one-off fiscal worsening (a shift up in the debt:GDP ratio) due to reduced tax receipts. But targeting cumulative NGDP, those taxes would turn up later because the total stream of NGDPLT that is taxed would remain the same. This would also be great for asset markets – the total flow of NGDP for paying debts would be the same. Holding the labour share constant, the future flow of profits would be also unchanged in total, which would be great for equity markets. I know this is a big departure from NGDPLT but what’s your view on this idea?

  10. Gravatar of ssumner ssumner
    21. May 2020 at 08:29

    dtoh, Yes, that’s possible. But don’t put a target rate on cash growth, just do enough so that NGDP expectations are on target.

    Postkey, You can do the math—what’s your point? Don’t confuse instrument settings with “the stance of monetary policy”.

    Lawrence, Interesting question. I’m not certain which approach is better. But recall that if you do cumulative on an undershoot you also have to do it on an overshoot. That could be painful.

  11. Gravatar of Postkey Postkey
    21. May 2020 at 14:31

    ‘Postkey, You can do the math—what’s your point? Don’t confuse instrument settings with “the stance of monetary policy”.’

    No answer was the strange reply!

  12. Gravatar of dtoh dtoh
    21. May 2020 at 15:37

    If you want to make it expensive for banks and individuals and firms to hoard cash by putting a fee on net cash withdrawals, you have to provide some allowance for increased need for cash as NGDP grows. That’s why you would put the fee on net withdrawals above the target NGDP growth rate.

  13. Gravatar of Matthias Görgens Matthias Görgens
    21. May 2020 at 21:42

    Lawrence, I had such thoughts recently as well.

    Scott, correcting for an overshoot of cumulative NGDP might not be the painful. The argument is essentially the same as for level targeting in the first place.

    You can correct for an overshoot in cumulative NGDP by just growing below target for a while. You don’t need to activately cut NGDP. Very similar to the plain level target.

    Overshoots and undershoots would be mild, thanks to the power of expectations.

    Ray, if all you care about is the stock price, then all the effect will be seen from the first credible press release onwards.

    You still have to deliver. But if everyone expects your delivery to succeed, successful delivery will not show up in the stock price at all.

  14. Gravatar of Thomas Hutcheson Thomas Hutcheson
    22. May 2020 at 10:12

    Do you have a theory about why the Fed allows inflation expectations to lie below 2% PCE even if they have not adopted NGDPL targeting? Do they fear, for some reason, the amount of assets they believe it would take to reach their target.

Leave a Reply