Level vs. growth rate targeting

Here’s Lars Christensen:

In fact I am pretty sure that if somebody had told Scott in July 2009 that from now on the Fed will follow a 4% NGDP target starting at the then level of NGDP then Scott would have applauded it. He might have said that he would have preferred a 5% trend rather than a 4% trend, but overall I think Scott would have been very happy to see a 4% NGDP target as official Fed policy.

Actually I would have been very upset, as indeed I was as soon as I saw what they were doing.  I favored a policy of level targeting, which meant returning to the previous trend line.

Now of course if they had adopted a permanent policy of 4% NGDP targeting, I would have had the satisfaction of knowing that while the policy was inappropriate at the moment, in the long run it would be optimal.  Alas, they did not do that.  The recent 4% growth in NGDP is not the result of a credible policy regime, and hence won’t be maintained when there is a shock to the economy.

However I do agree with Lars that the Fed has done much better than the ECB.

HT:  TravisV.



19 Responses to “Level vs. growth rate targeting”

  1. Gravatar of Major.Freedom Major.Freedom
    12. September 2014 at 19:49

    Shock caused by too much inflation now and since 2009.

  2. Gravatar of Chase Chase
    12. September 2014 at 20:33


    I am sure you have already broached this question in prior posts given the time spent on the topic. I assume you will ignore if you don’t feel it worth your time.

    I am struggling to get my head around how the NGDP targeting method deals with productivity driven unemployment or offshore driven unemployment. If we assume the level at which the Fed is targeting suggests a growth rate of 5% and it is hitting target with acceptable employment and inflation. Then car manufacturers decide to mechanize the entire vehicle production process and lay off all the workers. All else being equal NGDP is materially unchanged while unemployment has increased. If only looking at NGDP the Fed assumes all is well and does nothing to deal with the added unemployment.

    The same could be said of unemployment driven by offshoring I suppose.

    It seems to me in these cases the dual inflation and unemployment mandate would be superior. What am I missing here?

    In fact, given that NGDP per worker has likely increased not linearly, but exponentially over time wouldn’t we expect that the fed in an NGDP targeting scheme would have unnecessarily tightened, or failed to loosen, monetary policy anytime NGDP increased faster than the logarithmic expectation for appropriate structural reasons?


  3. Gravatar of benjamin cole benjamin cole
    12. September 2014 at 22:45

    Why not 5-6 percent NGDP growth targeting?
    In practise, I think 5-6 percent NGDP growth is close to the Reserve Bank of Australia’s 2-3 percent inflation band, btw. In either case, there is no runaway inflation, but room for growth.
    Is not it better to have inflation run a (as officially measured) a percent or two higher to get robust economic growth? Who cares if inflation is (officially) 2 or 3 percent? What if Market Monetarists win the day, only to select a too-low rate of NGDP growth (or level)?

  4. Gravatar of Michael Byrnes Michael Byrnes
    13. September 2014 at 03:37


    Consider what happened in the early stages of the home construction bust (in 2006-2007). Housing construction collapsed… and unemployment merely ticked up from 4.7% to 4.9%. Clearly the loss of home construction employment was largely offset by a gain in other forms of employment.

    It was not until much later, when money tightened, that we saw a huge rise in unemployment.

  5. Gravatar of ssumner ssumner
    13. September 2014 at 05:42

    Chase, If you target NGDP, then technological progress does not cause unemployment. Instead it causes RGDP to speed up and inflation to slow down. The speed up in RGDP growth offsets the loss in jobs due to technology.

    This was also a fear during the 1930s, but as soon as NGDP recovered, the “technological unemployment” problem went away.

    Ben, In my view the exact number is much less important than the policy itself. Four, five, or six percent would all be fine.

  6. Gravatar of Peter K. Peter K.
    13. September 2014 at 06:57

    “Four, five, or six percent would all be fine.”

    In other words, depending on the desired inflation rate.

    They have different implications for employment levels, which is maybe why the Fed wouldn’t announce a target that is not determined by Congress. The Fed’s job is stable prizes and maximum employment.

    A five percent NGDP path level target since 2009 would have entailed more QE and a quicker recovery. Unemployment would have been reduced more quickly. They unilaterally chose an unspoken 4 percent NGDP level which is more beneficial for creditors and less so for debtors, the unemployment and your average worker all people who don’t donate very much to political campaign. Conservatives assert that liberal politicians demagogue to get their votes but that’s naive and not how the world works. Those potential voters are apathetic and cynical about voting and, say, fail to show up for the mid-term elections. So we get a 4 percent NGDP path level target.

  7. Gravatar of Peter K. Peter K.
    13. September 2014 at 07:24

    My point in other words is that the ECB undemocratically chose a one or two percent NGDP path level target out of its inflationphobia. The Fed is more democratically responsive and is seeking to do more about unemployment, given its charter.

  8. Gravatar of Lars Christensen Lars Christensen
    13. September 2014 at 07:38


    Here is my – far too long – answer.

    Cheer up my friend – you have had a more positive impact than you give yourself credit for.


  9. Gravatar of ssumner ssumner
    13. September 2014 at 08:38

    Peter, You said:

    “In other words, depending on the desired inflation rate.”

    No, because inflation doesn’t matter. It should be; “Depending on the desired NGDP rate.” And that depends on the tax rate on capital, and the degree of downward wage stickiness.

  10. Gravatar of TravisV TravisV
    13. September 2014 at 09:16

    Prof. Sumner,

    Haven’t NGDP growth expectations been falling to 3.0% or perhaps even less? Isn’t that a problem for Lars’s argument?

  11. Gravatar of TravisV TravisV
    13. September 2014 at 10:37

    Lars Christensen in the comments section of his latest post:


    I think it is important to differentiate between the one-year outlook for NGDP and the longer term outlook. What Scott has been arguing is that long-term yields are relatively long and that that is indicating low NGDP growth over the longer run. I on the other hand is less worried. First of all, long-term inflation expectations are around 2.5% in the US. Second, I am much less worried about the long-term growth outlook for the US. I have no problem seeing 2-3% potential real GDP over the coming decade. Scott is much more negative.”

  12. Gravatar of Major.Freedom Major.Freedom
    13. September 2014 at 10:50

    “The bubble probably needs to continue in order to sustain the current global financial system and the necessary future deleveraging.” – Jim Reid, Deutsche Bank.

  13. Gravatar of TravisV TravisV
    13. September 2014 at 14:02

    Lars Christensen thinks this reasoning by Prof. Sumner is too pessimistic:

    “Danny, about 0.2% labor force growth and about 0.5% productivity growth. They add another 0.5% in case I underestimated one of the two.

    Note that RGDP growth has averaged 2% during a 5 year period where the unemployment rate has been falling fast. When the unemployment rate stops falling, RGDP growth will slow sharply. That means around 1.2% RGDP growth, or less.”


  14. Gravatar of ssumner ssumner
    13. September 2014 at 19:34

    Travis, I’m a bit more pessimistic than Lars, although that difference doesn’t really explain our differences on this issue.

  15. Gravatar of Don Geddis Don Geddis
    14. September 2014 at 10:46

    @Peter K: You have a very odd way of looking at possible policy choices. You say, “[4%/5%/6% NGDPLT] have different implications for employment levels.” That’s not true in general; the chosen growth rate has essentially no effect on employment. So why do you conclude it does? Because you think only about a “level target since 2009”. But of course, the critical point of a level target is to correct for errors, and 2008 had the biggest NGDP error since the GD. But you start your analysis in 2009, in the middle of a severe recession. And then conclude, “choosing a higher growth rate would have been better.”

    You really seem to have been missing the whole point of a level target. Once you have a level target, based on a trend line from a HEALTHY ECONOMY, *then* the growth rate number doesn’t matter.

    Talking about NGDPLT policy alternatives starting from a base economy in recession, makes no sense. That’s not how you compare possible policies.

  16. Gravatar of Chase Chase
    15. September 2014 at 07:30

    Sorry, tough finding time with kids these days.

    Michael – The house collapse was accompanied with a boom in commercial construction that started well before the residential collapse. this transition was underway already and was of course, temporary. I don’t see this the same as a dramatic increase in productivity. Something more appropriate are the vehicles that are being built that can essentially pave road as they drive with far less construction workers needed.

    Scott – I don’t disagree that a large part of the time productivity increases are borne out in more output for less cost leaving NGDP unchanged (although to say this is always the case and always will be the case I believe to be optimistic), but does NGDP stay constant because the reallocation of labor or the increased output of product at a lower price is driving up sales of that product effectively lowering the savings rate? i.e. is it not possible for productivity to take employees out of an industry and drive up sales (due to lower cost) causing no change in NGDP without the re-employment of those laborers? Do we need the workers to be reallocated to hit the NGDP target again?

  17. Gravatar of Chase Chase
    15. September 2014 at 08:44

    Correction, I indicated that you suggested technology enhancements always accelerate RGDP and leave NGDP unchanged, but I should have stipulated that this is only true when NGDP is targeted. My argument/question remains the same. I question whether this is ‘always’ true. Can NGDP be left unchanged with increases in productivity while unemployment ticks up?

  18. Gravatar of ssumner ssumner
    15. September 2014 at 19:08

    Chase, My claim was that if we targeted NGDP then when productivity increased we’d get more output, as the stable NGDP would help stabilize the labor market–keep it in equilibrium. That may not always be true, but it’s more likely under NGDPLT than inflation targeting.

  19. Gravatar of Chase Chase
    16. September 2014 at 06:47

    Thanks Scott

Leave a Reply