Jason Furman on NGDP targeting

David Beckworth has a new tweet that caught my eye:

David has some good follow-up tweets that further explain his argument.

I have several problems with Furman’s claim. Let’s start with policy alternatives such as simple inflation targeting, average inflation targeting, and price level targeting. If you use 2019 as the starting point, all three of those alternatives call for even tighter money than NGDPLT. That’s because while NGDP is still modestly below a 4% trend line from 2019, current PCE inflation is well above 2%, and the price level is somewhat above a 2% trend line from 2019.

Now you might argue that the Fed is a “flexible” average inflation targeter, taking into account not just inflation but also output gaps. But as David suggests in a follow-up tweet, the “dual mandate” policy is actually quite similar to NGDPLT. When policy is too expansionary by the NGDPLT criterion it is usually too expansionary by the dual mandate criterion as well. The main advantage of NGDPLT is its greater clarity and transparency.

My second complaint with Furman’s claim is that he’s mixing up NGDPLT with NGDPLT plus a mechanical formula analogous to the Taylor Rule, But NGDP is mostly advocated by a bunch of market monetarists and New Keynesians that are quite skeptical of these sort of mechanical policy rules.

For example, consider the market monetarist view that policy should be set at a position where market expectation of future NGDP is roughly equal to the policy goal. Right now, interest rates are near zero and markets expect them to remain near zero for an extended period of time. And yet the 10-year TIPS spread is consistent with roughly 2% PCE inflation (a bit more with the CPI.) We don’t have a direct market measure of 10-year NGDP growth expectations, but if markets expect 2% PCE inflation then it’s a good bet that 10-year NGDP growth expectations are not significantly above 4%. So it is not obvious that monetary policy is too tight, and there’s no suggestion in market prices that short-term rates should be 4% right now.

The same people who advocated NGDP targeting in 2008 and 2009 were also saying that money was too tight in late 2008 precisely because the Fed was setting interest rates too high on the basis of flawed formula that relied too much on past inflation and too little on market forecasts of future inflation.

Please do not assume that NGDPLT can only work with a mechanical backward-looking formula for central banks to peg interest rates.


Tags:

 
 
 

22 Responses to “Jason Furman on NGDP targeting”

  1. Gravatar of Effem Effem
    4. August 2021 at 10:39

    If i’m not mistaken, NGDP forecasts put it well above a medium-term trend with in the next few quarters. So let’s say we overshoot the NGDP trend but forward inflation expectations are still roughly 2%…what would NGDP targeting call for?

    One issue is that forward inflation measures may always be well-behaved despite any current over/undershoot because the market assumes the Fed will get back on course. In in the interim you could deviate rather substantially from target.

  2. Gravatar of ssumner ssumner
    4. August 2021 at 10:54

    Effem, I’m not sure that NGDP is likely to go “well above” trend, but if it does then you always want to “target the forecast”. So policy would depend on what instrument setting is likely to lead to on-target NGDP a couple years out in the future.

    Your second point is basically about the “circularity problem”, which I discuss in some of my NGDP targeting papers. Basically you want the market forecast of the instrument setting that will lead to on course NGDP.

  3. Gravatar of Ram Ram
    4. August 2021 at 12:40

    Survey of Professional Forecasters is predicting ~10.0% NGDP growth next quarter, and ~7.3% the quarter after that. Given that trend NGDP growth is on the order of 4-4.5%, and that we’re already essentially back on the trendline, this seems like a forecast of big time overshooting. Furman’s critique is silly, however, since if NGDP has fully recovered but unemployment has not, this suggests the natural rate has risen (probably temporarily, due to all the perverse incentives in the various goodie bags doled out over the last couple of years). The Fed keeping its foot on the gas won’t do anything about that except stoke inflation.

  4. Gravatar of ssumner ssumner
    4. August 2021 at 13:05

    Ram, Those figures are dramatically different from Hypermind, and I tend to think the latter is more likely.

    I’ll wager that we are very close to the trend line in Q4 (we are about 1% below right now.) To get there, we’d need two quarters of 6% NGDP growth.

  5. Gravatar of ssumner ssumner
    4. August 2021 at 13:07

    I should add that Hypermind is probably too low–the truth is somewhere in between.

  6. Gravatar of agrippa postumus agrippa postumus
    4. August 2021 at 14:42

    hercules 13th labor was trying to explain neutrality of money to sumner; he failed and the gods have been angry ever since.

  7. Gravatar of Ram Ram
    4. August 2021 at 15:10

    My model says we need 6.0% this quarter to catch up fully, and 4.2% to maintain next quarter. I’m not particularly confident in the forecast either, but I suspect we’re going to see numbers well above these in the next two quarters unless there’s some marked change in policy.

  8. Gravatar of Spencer Bradley Hall Spencer Bradley Hall
    4. August 2021 at 15:24

    We’ll see in the first quarter of 2022 how good the forecasts are. The economy falls off a cliff if the FED doesn’t counteract the decline in N-gDp.

  9. Gravatar of Michael Sandifer Michael Sandifer
    4. August 2021 at 16:57

    For what it’s worth, having heard a few interviews in which Furman has discussed NGDP targeting, he doesn’t see forward indicators available, so thinks it must rely on lagged data.

    The Hypermind NGDP forecast for Q4 2022 seems about right. It’s close to what I see as the implicit market forecast at 4.4%, and the forward earnings yields for the S&P 500 have fallen to around 4.5%. For example:

    https://ycharts.com/indicators/sp_500_pe_ratio_forward_estimate

    The implicit expected mean NGDP growth rate going out to infinity is a bit over 4% right now.

  10. Gravatar of marcus nunes marcus nunes
    4. August 2021 at 17:08

    Neutral, Natural & Potential are “imaginary” numbers. I´m trying to gauge the state of the economy every month. NGDP has been back on the post GR trend since March. Is the Fed only waiting for supply constraints to “relax” before “defining” a higher trend level path?
    https://marcusnunes.substack.com/p/state-of-play-june-2021

  11. Gravatar of ssumner ssumner
    4. August 2021 at 21:22

    Ram, NGDP has risen by a total of 4.7% since 2019:Q4. If it had grown at a 4% rate, it would be up by a bit over 6% by 2021:Q2. So we are more than 1% below trend, and need two quarters of 6% plus to get back on trend, more like 6.6% per quarter to be exact.

    Where did I go wrong?

    Michael, If Furman is just relying on historical data, then inflation targeting would be even worse.

  12. Gravatar of ssumner ssumner
    4. August 2021 at 21:26

    I just realized that Hypermind is forecasting only 2.4% annualized NGDP growth in the second half. Is my math correct?

  13. Gravatar of Effem Effem
    5. August 2021 at 03:16

    Scott, finally got around to listening to your podcast with Beckworth – good stuff. One question: you suggest that fiscal spending wasn’t much of a driver of the 1970s inflation. But much like reasoning from a price change, isn’t looking at deficits in isolation sort of missing the point? The size of deficits normally track the amount of slack in the economy. Larger deficits with less slack produce inflation.

    See this chart from John Hussman for example:

    https://www.hussmanfunds.com/wp-content/uploads/comment/mc190805m.png

  14. Gravatar of Ram Ram
    5. August 2021 at 05:44

    Scott,

    Nothing wrong with your approach, I just use a more complicated one:

    For a given quarter q, I first compute the annual average NGDP growth rate since q – 1, q – 2, q – 3, etc. I then compute a weighted average of these. I say we’re on target in quarter q if this weighted average equals the target growth rate. If not, I compute what NGDP should have been in quarter q for that to be the case, and compute the percent by which actual quarter q NGDP needs to grow or shrink to achieve that optimal level. That percentage is my (retrospective) measure of monetary policy stance in quarter, with neutral being a value of 0.

    For weights, I give 5 years prior to q a weight of 1, and have the weight decay exponentially in both directions from that peak weight, reaching 0.05 in the most recent quarter. This ensures that recent quarters don’t change the target much (a la level targeting), which quarters more than 10 years old don’t either (a la average growth rate targeting), but the particular number choices are a bit arbitrary. Finally, I found that a target of 4.2% results in this algorithm indicating we had basically achieved a neutral stance in Q2 of 2018, a time when inflation was about 2%, inflation expectations were about 2%, unemployment was near its nadir and didn’t fall much further, and growth was reasonably brisk. So that felt like it gave the right answer.

    Per this algorithm, in Q2 NGDP needed to increase by 0.4% (absolute rate) to achieve neutrality, but will achieve neutrality by Q3 if we see 6.0% NGDP growth next quarter (annual rate). So on this model, ideally we’d see 6.0% next quarter, and then maintain close to 4.2% thereafter. But again, I expect we’ll see a lot more than that unless the Fed changes course soon.

  15. Gravatar of Ram Ram
    5. August 2021 at 05:48

    Here is what the algorithm looks like applied to the last 16 years or so:

    https://ibb.co/HLZx6tG

  16. Gravatar of ssumner ssumner
    5. August 2021 at 07:51

    Effem, Fair point, but slack in the economy during the 1970s wasn’t significantly different from other recent decades. There were two recessions and the average unemployment rate in the 1970s was not particularly low.

    Ram, That’s certainly reasonable. I still believe that NGDP growth in the 2020s will be about the same as in the 2010s, roughly 4%.

    Are you investing in Hypermind, where you can profit if NGDP growth averages more than 2.4% in the second half?

  17. Gravatar of Spencer Bradley Hall Spencer Bradley Hall
    5. August 2021 at 09:16

    There’s a significant “echo” in money flows due to stimulus injections. The rise in the y-0-y increase in prices will be reversed in FEB 22 (I guess that’s what people refer to as a “base” effect). But that’s not what it is.

    There’s a huge FOMC problem in the 1st qtr. of 2022.

  18. Gravatar of Ram Ram
    5. August 2021 at 12:30

    I’m not willing to bet, partly cuz I’m unreasonably risk averse, but also because my prediction is conditional on Fed policy staying the course. I’m not sure how to predict what the Fed will do between now and the end of the year, and that could change everything.

  19. Gravatar of Spencer Bradley Hall Spencer Bradley Hall
    5. August 2021 at 12:47

    re: “that could change everything”

    They can’t predict anything. Everything they do is after the fact.

  20. Gravatar of David S David S
    5. August 2021 at 14:49

    And now Joe Manchin wants the Fed to tighten and taper to protect us from the horror of temporary inflation. Maybe 6 or 7 quarters of NGDP growth below 3% will restore America and set us on the righteous path for a decade.

  21. Gravatar of Spencer Bradley Hall Spencer Bradley Hall
    6. August 2021 at 07:26

    The FED has already tightened. See: marcus nunes. The problem is that economists are a stupid lot. If you want real growth you gradually drive the banks out of the savings business. Nothing else will work.

  22. Gravatar of Spencer Bradley Hall Spencer Bradley Hall
    6. August 2021 at 13:26

    How about a little Elliot Wave theory. The Wilshire index is @46,143.35 – 225 points away from the 5th of the 5th wave. And futures are up 156. Time is running out.

Leave a Reply