18 Dec 2013

Some Sumner Shuffles

Economics, Federal Reserve, Inflation, Scott Sumner 25 Comments

I often focus on Krugman Kontradictions, but here at Free Advice we have given examples of Sumner Shuffles. Now Scott is much more slippery than Krugman, so to get a true Sumner Shuffle is difficult. I’ve been checking his blog lately to see what he has to say about the Fed’s “taper” announcement–it was expansionary since the stock market went up, of course–but in so doing I ran across something that doesn’t add up.

Two days ago, Scott wrote:

Over the years my critics, and even some of my supporters, have said; “Sumner is basically proposing higher inflation as a solution.” That was never really true, I was proposing 5% NGDP, level targeting. And even though inflation would have been above 2% in the 2008-11 period, a stable NGDP policy would have prevented that inflation from having the sort of negative effects that economists see as resulting from high inflation. My plan would not have hurt savers, in aggregate.

Even so, there was a grain of truth in the claim. The inflation rate would have run above 2% for a number of years under my proposal. But now even that is no longer true. There have been enough wage/price adjustments to the lower NGDP growth rate that the SRAS has been shifting to the right. This means that today even a 2% inflation rate would produce a robust recovery. [Bold added.]

Now surely the above means that under Scott’s proposal, price inflation would not rise above 2%, or at least it would not be above 2% for the next few years. Right? That’s the only possible way the above statements can make any sense.

And yet today, Scott writes:

Today’s decision did not drop the unemployment threshold, but it weakened it. It should simply be removed. If they promised to wait until expected inflation rose to 2.5% before raising rates, it would mean roughly 2% inflation over the next few years, as the inflation rate is now running below 2%. That’s still too tight, but much better than the old policy, and even better than today’s revision.

I kinda think the people who said “Scott is basically calling for more inflation” could be forgiven for saying that, in light of comments like this. And now it seems that 2% inflation is “too tight,” whereas two days ago it would have produced a robust recovery.

One thing in closing: If someone says, “But Bob, Scott is calling for level targeting of nominal GDP growth…” that person will be locked in a shed with Major Freedom and Lord Keynes for 3 days.

25 Responses to “Some Sumner Shuffles”

  1. Matt Tanous says:

    But Bob, the solution to reduced aggregate demand is aggregate demand targeting. Just like the solution to my fireplace going out is to just start lighting things on fire randomly until the temperature returns to what it was before!

  2. Major_Freedom says:

    ““But Bob, Scott is calling for level targeting of nominal GDP growth…” that person will be locked in a shed with Major Freedom and Lord Keynes for 3 days.”


  3. Ken B says:

    He says “even a 2%” rate would be good. That wording suggests he thinks more than 2% would be better. And today he says …. more than 2% would be better.

    Long day. I need a beer. Even a lager would do. I’ll check the fridge, maybe I’ve got an ale, which would be better. But even a lager would do.

    • Bob Murphy says:

      OK Ken try this:

      “It used to be the case that my optimal beer order would be stronger than a lager, but now that’s no longer true. My tolerance has dropped so much since college that even a lager gets me buzzed.”

      You’re telling me this guy wants to order a drink stronger than a lager?

      • Ken B says:

        If even a lager will get me buzzed I might still think an ale would do the job better. I might want to be more than just buzzed. (As I would had I won that 3 day special.) Sumner could prefer a roaring recovery to a robust one perhaps?

      • Ken B says:

        Incidentally your analogy is amiss on an important point. Sumner preferred a effective 2% rate in the past not because it was the most stimulating but because it was stimulating in a way that avoided the worst of the ill side effects of inflation. In other words you’re slightly messing with the meaning of optimal in your analogy.

      • Harold says:

        Lager not strong enough? Surely 12% will do the trick.

        • Ken B says:

          Uh uh uh. It was Bob who interpreted my comment as meaning stronger. I just meant better. I like ales more.
          Bob’s extrapolation is reasonable though since Sumner wants more inflation, so I went with it in my response.

          As an aisde, Belgian lager? Remind me to never let Harold buy the next round!

    • Bob Murphy says:

      I grant you Ken, when you only look at 50% of what someone says, you’re very good at constructing analogies to show why I’m possibly stupid.

    • martinK says:

      He says “even a 2%” rate would be good. That wording suggests he thinks more than 2% would be better. And today he says …. more than 2% would be better.

      Yes. But today he *also* says 2% is *not* good. Because “too tight” means not good.

      • Ken B says:

        No, it means not as good as it could be. He means it could and should be looser.

        • Major_Freedom says:

          If 2% would “produce a robust recovery”, isn’t it a bit strange to also say 2% is not enough?

  4. Ken B says:

    Apropos of nothing, I was looking at my blood pressure readings. It used to be that I had to take three blood-pressure pills a day to get down near normal. My doctor would’ve put me on four but he worried about the side effects. Three he said would avoid the ill effects usually associated with overdosing this medicine. Now that I’ve lost weight and started exercising, I can get down into the safe range with two. Doc says even just two is a good idea but it would be better if I still kept taking three. Normal he says is better than just in the safe range.

    Plus even a lager will get me buzzed with 3.

  5. Bill Woolsey says:

    Sumner’s first best policy is nominal GDP level targeting. While there was no explicit target, the U.S. came close to that during the Great Moderation. In 2007, growth in nominal GDP slowed and fell below that past trend. Summner’s view is that the best thing for the Fed to do then, and then in the second half of 2008, when nominal GDP fell, was to explicitly commit to retuning nominal GDP to that past trend, making it into a target for the future.

    If we had followed that sort of policy and stayed close to the 5% growth path, he believes inflation would have been higher than 2% over the period. I presume that he believes that potential output was growing less than the 3% trend during that period. The CBO’s estimates are less than 3% for potential output.

    However, to the degree that the policy was only implemented later, say in late 2008 or 2009, after a period of lower than trend inflation and the one quarter of modest deflation, then the return to trend for nominal GDP would have resulted in higher inflation.

    The slowdown in potential output growth would have reinforced that effect during the period.

    It is now the end of 2013. As time has passed, Sumner changed is view as to what should be done now. While staying on the 5% trend of the Great Moderation would have been a good idea, today, it would be better to have a couple of years of 7% nominal GDP growth and then go to 5%. This would leave the U.S. well below the trend of the Great Moderation.

    And then, since the Fed isn’t likely to change its target away from inflation, Sumner proposes that they adopt level targeting for 2% inflation. I think that was what he was proposing. I would call it a 2% growth path target for the price level. If inflation continued for a time at lower than 2% as seems likely, then a level target would require more than 2% inflation, maybe up to 2.5% for a time, to catch up to the target growth path.

    The business about the short run aggregate supply curve is that in order to keep prices on that 2% growth path, spending on output would need to grow at a sufficiently rapid clip to generate a strong recovery in real output.

    For a market monetarist, keeping nominal GDP on a target growth path is the goal. If you fall below, then you need to catch up. The likely effect of falling below is slower inflation and the catching up is higher inflation. It isn’t that the higher inflation is good, it is rather a likely consequence of getting nominal GDP back up to the target growth path.

    Given the growth path of nominal GDP, assuming production is below potential, the slower wage growth and inflation are good things, returning the economy to equilibrium.

    • Major_Freedom says:

      Equilibrium is not possible in human life.

      All non-market mechanisms of money production distort economic calculation and bring about the very supply and demand “shocks” that you all too often treat as inexplicable “exogenous” forces deeply hidden in the darkest corners of the market process.

      In a perfect world of individual private property rights enforcement and respect, where all interactions are voluntary and consensual, the asbtract sum of all individual spending that takes place over a given geopraphical territory would not rise at a constant rate. Forcing it to be constant through coercion in money production has detrimental economic effects which are rarely if ever even admitted, let alone seriously addressed, by market monetarists.

      There is no consideration from market monetarists regarding the balancing forces that fluctuating territorial spending allows. For example, if a group of people over territory X are low savers and investors, and high consumers, such that the rate of money outflow exceeds the rate of money inflow, then in order for that group of people to be efficently integrated into the division of labor with other territories, would be if domestic prices fell. In a free market of money, that is indeed what happens with a trade deficit. But with forcing total spending to remain growing, a high consuming region isn’t faced with falling domestic prices and wage rates. As a result, the imbalance persists. This imbalance is not eliminated with floating fiat currencies among regions.

      This is just one example. Others are distortions to interest rates.

      All this “first best” talk from market monetarists cannot be distinguished from an absolutist socialist who hates free markets and wants to keep money socialized. Communicating it as “It’s the best we can do with what we have”, is sloppy and makes one prone to errors in reasoning, which occurs quite often in your circles.

  6. RPLong says:

    At the risk of making some people upset, and potentially proving that I know absolutely nothing about macroeconomics, I have to say that this NGDPLT business feels a lot like a religion to me.

    I have never seen any economic rationale for 4% LT versus 5% LT versus 2% LT or anything else. Everyone who writes about this stuff is very loosey-goosey about how they choose their preferred target. They seem to choose it because it “feels right.”

    That’s relevant because if I don’t know specifically why 5% is preferable to 4% (or vice-versa) then I don’t really know why 2% is worse than 4%. And I’m not being stupid here – I know that 2% inflation is a lower number, which might potentially mean lower RGDP growth. But it also might not mean lower RGDP growth.

    NGDPLT is vaporware.

    • RPLong says:

      And please don’t harp on my sloppy language re: “2% inflation… might potentially mean lower RGDP growth.” We all know what I really meant was that 2% NGDPLT might potentially mean lower RGDP growth.” But this is just a case in point. The NGDPLT crowd is as bad as anyone when it comes to using a special private lexicon…

    • Yancey Ward says:

      I think your objection is right on the nose. I can’t see any real justification for why one number is better than another.

  7. Bob Murphy says:

    I’m going to say one last thing on this:

    (1) 3 days ago, Scott clearly said that his first-best policy proposal would NOT involve price inflation higher than 2% over the next few years. He admitted that it would have done so, had it been implemented in 2009 or 2010, but at this point such was no longer the case.

    (2) 1 day ago, Scott clearly said that 2% inflation for the next few years would be too tight.

    ==> The only way to reconcile these statements is:

    (A) Sumner admits that his first-best policy proposal is too tight.

    (B) If the Fed isn’t doing NGDPLT, but instead is targeting price inflation, then the second-best (or third-best, whatever it is) policy involves a higher price inflation rate than would occur under NGDPLT.

    ^^ If any of you had suggested (B), I would have been OK with it. But clearly that’s not what Ken B. is doing with his nonanalogous analogies, and I don’t think Bill Woolsey’s novella does it either.

    • Ken B says:

      I think you misread on 2. Sumner notes there are two triggers and the caveat that AT LEAST one must be met before the fed acts. He is concerned about the AT LEAST which he uppercased. He wants rid of the unemployment trigger. That would leave a 2.5% expected inflation before raising rates and 2% after. And he thinks THAT POLICY, a 2.5 trigger with no unemployment caveat, is too tight. He is not saying 2% for years is too tight. Because he says we’re well below 2% pi now. So he thinks, based on that below 2%, that even mild rate rises will curtail activity strongly. He’d prefer a policy where the trigger was say a forecast of 2.6% or 2.7% or some other magic number. That does not commit him to saying that long term the actual rate will have turned out to be much different from 2%

      • Bob Murphy says:

        Ken B. you are nuts. You wrote:

        And he thinks THAT POLICY, a 2.5 trigger with no unemployment caveat, is too tight. He is not saying 2% for years is too tight.

        Go read it again. He is saying that if the Fed says, “We will raise rates once we hit 2.5% CPI” then that means in practice we will have (average) CPI inflation of 2% for the next few years. And that will be too tight.

        But 3 days ago, he said 2% inflation would produce a robust recovery, and his first-best policy prescription would not need any higher inflation than that.

        The only way out is to say that the resulting price inflation rate under his first-best policy is lower than what it would be with inflation targeting. If you want to argue that, OK fine.

        But that’s not what you’re arguing. You are arguing that 2% inflation is optimal and too tight simultaneously.

    • Dan W. says:


      I think you are nitpicking. Scott’s argument is not about the numbers but the policy. That he promotes 2% inflation and 5% NGDP growth is because he likes those numbers, not because they are optimal. As such I doubt he cares whether inflation is 2% or 2.5%.

      I think there are many flaws with NGDPLT but the least of these is the numerical choice of targets.

      • Bob Murphy says:

        Scott’s argument is not about the numbers but the policy. That he promotes 2% inflation and 5% NGDP growth is because he likes those numbers, not because they are optimal. As such I doubt he cares whether inflation is 2% or 2.5%.

        If that’s true, is it too much to ask then that he stops saying things like, “Inflation of 2% would be better than the status quo, but it’s still too tight.” ? You’re saying he actually doesn’t believe what he’s typing up?

        (ADDED: That isn’t an exact quote from Sumner, I’m just paraphrasing what he clearly said in his recent post, for which I am being criticized now for reading and assuming Sumner believed what he typed.)

        • Ken B says:

          No, you’re being criticized for asserting that’s what he clearly meant when he certainly did not clarly mean it, and most likely mean something else. The deabte really is about what “that” refers to (Shades of “you didn’t build that”) and what “roughly” means. You are arguing that is crystal clear that he meant a particular combination of meanings for those, with the result he says something mildly contradictory. I, and maybe Dan W, are saying there are other readings, at least as plausible prima facie that are more consistent with what SS says generally.

          Since you paraphrased let me. first I note he explicilty compares “that” to a *policy* suggesting the referrant IS a policy

          ” If [the new policy were] to wait until expected inflation rose to 2.5% before raising rates, it would mean roughly 2% inflation over the next few years, as the inflation rate is now running below 2%. That’s still too tight [a policy], but much better than the old policy,” Note “old policy” is in the original.

          had he written “That’s still too tight, but much better than the old rate” we wouldn’t be having this debate.
          I’m not up on econ jargon, but is the adjective “tight” commonly used for rates or for policies? In normal parlance rates are high or low, policies are tight or loose.

          Next “roughtly”. I think SS envisions raising rates when the projected inflation, assuming the current rate, hits 2.5%. I think he imagines that will reduce the pi rate to somewhere near 2%. If he was previously claiming a level of precision to say the 5 th decimal place, “no higher than 2.0001” then maybe he’s shuffling.

          So to sum up. Your interpretation is not “clearly” correct. At best it is a reasonable reading, but I don’t think it deserves much more than that. On balance it looks probably wrong.

          • Ken B says:

            Actually he compares “that” to a policy twice. “old policy” and “today’s revision” What was revised today was the a policy.

Leave a Reply