03 Oct 2018

Last One I Promise: Krugman on DeLong on Murphy on Potential GDP

DeLong, Krugman 21 Comments

I’m just trying to get my ducks in a row here “among friends” before I go out into the cruel world. I think I have a very strong case so I don’t want to commit an unforced error.

A few of you who have no axe to grind (such as Capt. Parker, who is outranked by Major Freedom just to be clear) are disputing my claim that both DeLong AND Krugman were saying potential GDP wasn’t much lower than the pre-recession trajectory would have implied, as of August 2013. And yet, DeLong goes through a whole post saying I didn’t do my homework, because if you looked at various investment considerations and other factors, you’d know that potential GDP in August 2013 was 1% below trend.

Then Krugman linked to that post and opened with: “Brad DeLong gets very annoyed at Robert Murphy, who ridicules him for not taking into account the effect of low investment on potential output. Brad notes that Murphy apparently hasn’t done the math, which indicates that even the sustained shortfall we’ve had since 2008 (mainly in residential investment) should not have had a major effect. (I added the bold.)

So help me out, guys. You’re saying that *really* what Krugman was saying there was, “For all I know, Murphy is right when he says potential GDP is substantially down because of shortfalls in investment since the recession struck”?

One last potential loophole: It occurred to me that perhaps Krugman merely meant, “Sure, potential GDP could be way down–who knows?–but if it *is*, it’s not because of investment patterns. Rather it’s because unemployed workers see their skills atrophy through disuse.”

But that’s not the story at least in the papers Krugman links in his latest post. I haven’t seen anybody make that particular point; instead I see them talking about “procyclical R&D and other investment” that gets hit when governments foolishly engage in austerity.

Either way, I’ll move on with my life (here on the blog) after this post, but I really have no idea what you guys are talking about. Any normal person would be certain Krugman was agreeing with DeLong in August 2013, when he claimed all the data suggested potential GDP was maybe 1% below trend.

21 Responses to “Last One I Promise: Krugman on DeLong on Murphy on Potential GDP”

  1. Transformer says:

    I suspect you will not like my analysis but here goes.

    If the issue was purely about: ‘both DeLong AND Krugman were saying potential GDP wasn’t much lower than the pre-recession trajectory would have implied, as of August 2013′ then as CBO data shows this to be untrue this would be a very short discussion with a clear winner.

    A number of factors make it more complex than that.

    – DeLong certainly gets off to a bad start when he says ”There are no signs in the pace of technological progress, in the level of investment, in the pace at which the American labor force educates itself, in measures of capacity utilization, in signs of upward wage pressure due to labor quality bottlenecks, or in surging commodity prices due to supply bottlenecks to suggest that the path of growth of U.S. sustainable potential GDP is materially lower today than was believed back in 2007.’. CBO data contradicts this.

    But then the plot thickens.

    – You challenge him not by pointing to CBO data on falling potential GDP, but by looking at investment levels. It is this that upsets him. He says ‘ a 5% real rate of return on capital the reduction in potential as a result of the post-2008 investment shortfall is now (19%-14% fall in investment share) x 4 years x 5%/year return on capital = 1% reduction in potential GDP,’. Here he is is explicitly talking about the contribution of investment shortfalls (which is the way you framed the issue in your post) to falls in potential GDP and its possible the falls already seen by 2013 were attributed to other factors than investment – so this is not as clearly a smoking gun against him as I first thought.

    Krugman then jumps in:

    – He makes the seemingly correct point that potential GDP already adjusts for cyclical investment changes

    – He is clearly giving support to DeLong but (as it often the case) steers clear of actually saying anything that is factually wrong. When he says ‘Brad notes that Murphy apparently hasn’t done the math, which indicates that even the sustained shortfall we’ve had since 2008 (mainly in residential investment) should not have had a major effect.’ he uses the conditional “should not” rather then “did not”.

    – Jump forward to 2018 and Krugman has a post explicitly talking about the falls in potential GDP including a chart that (presumably unintentionally) throws DeLong under the bus in regards to the initial 2013 post. That post is quite nuanced and comes up with numerous theories for the fall in potential GDP including a suggestion that ‘a big slump could hurt long-run prospects in other ways, for example by depressing business investment’.

    So bottom line:

    – DeLong is wrong about potential GDP but suspect he would not budge on his claim that falling investment was only a small part of the reason it fell, and would stick to his guns that you had miscalled this in your initial post.

    – Krugman walks away scot free.

    • Bob Murphy says:

      Transformer, suppose the following debate happened right before Einstein published his work on special relativity:

      DELONG: There’s no reason to suppose a rocket ship accelerating constantly at 1m/s/s couldn’t eventually reach 3x the speed of light.
      MURPHY: No because the mass increases with velocity.
      DELONG: Murphy hasn’t done his homework. There’s no reason to think mass increases with velocity. No reason the rocket ship shouldn’t reach 3x the speed of light.
      KRUGMAN: DeLong rightly is annoyed with Robert Murphy. There’s no reason issues of mass should keep the rocket from reaching 3x the speed of light.

      Is it your position, Transformer, that in the above exchange Krugman walks away scotfree? Because after all, he correctly pointed out what “should” have happened, given the model DeLong had in mind?

      • Transformer says:

        No, in your example he would clearly be giving support to an incorrect view.

        But if he said ‘DeLong is annoyed with Robert Murphy for suggesting that issues of mass should keep the rocket from reaching 3x the speed of light.’ he would just be reporting the facts, right ?

        Admittedly its a bit ambiguous if when he says ‘which indicates that even the sustained shortfall…..’ he is just channeling DeLong or if its his own editorial comment. But I am sure they both would defend the position that ‘even the sustained shortfall we’ve had since 2008 (mainly in residential investment)’ would not have had a major effect on whatever fall in potential GDP (probably 6%) that had occurred at that time.

        • Bob Murphy says:

          Transformer OK when we add in the very next sentence Krugman wrote: “But it’s actually much worse than even Brad seems to realize.”

          …I really have to say you are being unreasonable. You’re saying Krugman meant, “For all I know, DeLong is totally wrong and Murphy is right, but let me bring up a completely separate point”?

          • Transformer says:

            With Krugman its often not so much what he wants the reader to think he means as trying to pin down down that meaning in the exact words he writes.

            I’m sure his intent was to leave the reader thinking ‘Not only is Murphy an idiot for the reason that Brad gives, but he’s an doubly an idiot for this totally different reason’ .

            But all we have is the words he actually wrote and nowhere does he explicitly endorse a claim that ‘potential GDP wasn’t much lower than the pre-recession trajectory would have implied’

            • Transformer says:

              And (at the risk of being annoying) for DeLong at least the discussion was not about potential GDP being lower than the pre-recession trajectory would have implied, its about whether the data you showed in the chart from your initial response to him would lead to that outcome – that’s what his later post is about.

            • Bob Murphy says:

              Transformer wrote: “But all we have is the words he actually wrote”

              I agree, and he actually wrote “But it’s actually much worse than even Brad seems to realize.”

              So you’re saying Krugman might have meant, “Murphy’s post is really good”?

              • Transformer says:

                My point is that it doesn’t matter what he “meant”, all that matter is what he wrote.

                And he wrote “But it’s actually much worse than even Brad seems to realize.” (which is pretty vague) and not “potential GDP wasn’t much lower than the pre-recession trajectory would have implied, as of August 2013” (which would have been pretty specific).

              • Bob Murphy says:

                Transformer wrote: “And he wrote “But it’s actually much worse than even Brad seems to realize.” (which is pretty vague) and not “potential GDP wasn’t much lower than the pre-recession trajectory would have implied, as of August 2013” (which would have been pretty specific).”

                OK I think I have to walk away from this one. If you chime in more, I’ll read what you say, but I think we are at an impasse. It sounds like your position is, even if Krugman had said, “I give DeLong a high-five in this post about Murphy,” you would still say we can’t be sure whether he was endorsing DeLong’s claims about potential GDP.

              • Transformer says:

                Yes, I agree its time to stop.

                Final thoughts: Up until Capt Parker published the chart showing the various revisions of potential GDP yesterday I had thought the discussion really was about whether or not potential GDP was much lower than the pre-recession trajectory would have implied, as of August 2013.

                But as that argument could have been settled instantly in 2013 from the CBO data it had to be about more than that.

                Rereading everything I concluded it actually was (at least for DeLong) about the extent to which falling investment in and of itself would cause potential GDP to fall.

                And with that understanding DeLong’s back of the envelope ‘with a 5% real rate of return on capital the reduction in potential as a result of the post-2008 investment shortfall is now (19%-14% fall in investment share) x 4 years x 5%/year return on capital = 1% reduction in potential GDP,’ started to make some sense.

                Of course if potential GDP fell by 8% but only 1% of it was due to falling investment there is still some explaining to do. But I think that is tangential to DeLong’s point.

                I still think DeLong and Krugman were unacceptably hostile and should have apologized for that. But having looked at this issue in way more depth than it possibly deserves – I now have some perspective on where they were coming from in their posts on the topic.

              • Bob Murphy says:

                I’m sorry Transformer for doing one more post on this, but it’s only because your latest thoughts make me better understand you.

                I think there is one enormous flaw in your logic: This was the original paragraph from DeLong that set the whole thing off:

                “There are no signs in the pace of technological progress, in the level of investment, in the pace at which the American labor force educates itself, in measures of capacity utilization, in signs of upward wage pressure due to labor quality bottlenecks, or in surging commodity prices due to supply bottlenecks to suggest that the path of growth of U.S. sustainable potential GDP is materially lower today than was believed back in 2007.”

                So no, there is no possible way DeLong thought potential GDP was way below trend, and he was just saying, “Murphy is wrong for thinking investment alone could explain it.”

                He clarified in later comments that “materially lower” would mean 5%, but he reiterated that he thought it was 1%.

                I agree with you, it’s astonishing DeLong could be taking me to task when the CBO itself, at that very moment, confirmed potential GDP was 3.6% below trend, but I made that point myself. DeLong never answered me.

                So now, when you’re realizing the CBO itself *at that time* could’ve shown it was crazy for DeLong to say potential GDP was no more than 1% below trend…I agree with you. I agreed with that point back in August 2013.

                I think you are giving DeLong too much of the benefit of the doubt here.

              • Transformer says:

                If you look at my very first comment in this comments section you will see I acknowledge that DeLong got off to a very bad start with his first paragraph.

                And it was that paragraph that prompted your initial post.

                But you hone in one sentence ‘there’s nothing “in the level of investment…to suggest that the path of growth of U.S. sustainable potential GDP is materially lower today than was believed back in 2007.’.

                And those 5 words ‘in the level of investment’ tipped things back in his favor. The 1% calculation and the 5% threshold for materiality appear clearly to be in relation to the level of investment and not other factors that affect potential GDP.

                Things definitely get fuzzy in the comments on Daniel’s and your blog and people reading those alone might think the discussion was about ‘the path of growth of U.S. sustainable potential GDP is materially lower today than was believed back in 2007.’. And of course you as one of the main protagonists are better placed than I to know what the issue at hand was.

                But that is not validated from Brad’s response to your original post where its pretty clear the issue is investment and potential GDP not potential GDP in general.

                Is it possible that you had one view and DeLong another of what was being discussed ?

  2. Capt. J Parker says:

    Dr Murphy said:

    “A few of you who have no axe to grind (such as Capt. Parker, who is outranked by Major Freedom just to be clear) are disputing my claim that both DeLong AND Krugman were saying potential GDP wasn’t much lower than the pre-recession trajectory would have implied, as of August 2013.”

    No, I am not! I’m in total agreement with your claim “that both DeLong AND Krugman were saying potential GDP wasn’t much lower than the pre-recession trajectory would have implied, as of August 2013.”

    I’m disputing this:

    “A) Then: Since potential GDP hasn’t fallen, the actual drop in GDP is due to lack of Demand, so Keynesians are right.”

    Specifically I am disputing that Krugman said, in 2013, that “because” potential GDP has not fallen that this “proves” the output gap is all lack of aggregate demand. Krugman repeatedly said that the way to prove or disprove the theory that there is a structural problem is to look at inflation. If you are doing monetary and fiscal stimulus because you think you have an aggregate demand problem when you really have a structural problem then that stimulus will cause inflation. No inflation, then no structural problem. I not claiming that Krugman’s inflation argument is airtight. I’m just saying Krugman never offered the level of potential GDP as “proof” of anything other than that there was and output gap.

    Moreover, my casual understanding of potential GDP is that it really isn’t much more than a trendline. Anyone that want’s to claim that the path of potential GDP tells you something profound that you can’t otherwise infer by looking at the path of GDP does so at their peril.

    I found the potential GDP discussion here informative: https://research.stlouisfed.org/publications/economic-synopses/2012/04/20/what-is-potential-gdp-and-why-does-it-matter/

    • Tel says:

      I’ve long been a PGDP skeptic and I vaguely remember back in 2013 when this first came up I was confused as to why any Austrian would even buy into this kind of debate. Anyway, I followed your link and now I’m just as skeptical as before… so that’s comforting.

      But this theoretical “perfect world” is not the real world, and the scenario just described is not the concept of potential GDP monetary policymakers typically use when setting monetary policy. Instead, they estimate potential GDP by constructing measures of the trend in actual GDP that smooth out business cycle fluctuations. Looking back in time, potential output is relatively easy to measure because we have reliable methods to extract smooth trends from historical data. However, measuring potential output in real time is more difficult because only past data are available to estimate the trend.

      When I look at their PFD chart (linked about halfway down their page) they don’t appear to use any “reliable method” other than just drawing in what they want to see. They only start from 2001 which seems kind of silly. if you are going to use any of the well accepted curve fitting techniques you want to have a large chunk of data, if you can get it.

      https://fred.stlouisfed.org/graph/fredgraph.png?g=luaI

      There’s RGDP (presuming you trust the Fed’s deflator and seasonal adjustments) and it has the fairly evident characteristic that Australian engineers call “a banana” … and by that I mean the rate of growth has been steadily reducing over time. Thus, a most basic curve fit like a quadratic would be a reasonable place to start, if you just wanted to “let the data speak” and you don’t have any particular theory as to why the curve would look like that.

      The curve would end up going roughly through the middle of the data, which should be below the booms and above the busts. Then you can extrapolate boldly by a small amount, knowing you have something at least plausibly backed up by a methodology… admittedly a simplistic one.

      But why does potential GDP matter? How do we use it? Potential GDP is important because monetary policymakers use the difference between actual and potential GDP—the output gap—to determine whether the economy needs more or less monetary stimulus.

      OK, but that in itself proves that PGDP has no concrete definition, given that “monetary stimulus” could be anything at all. Maybe “monetary stimulus” means low interest rates, but then again it could also mean Obama’s shovel ready projects, or it could mean Trump’s wall building, or Bush/Hillary warmongering. Believing in the concept of PGDP implies that you also believe it doesn’t matter where the new money gets dumped into the economy, and therefore it doesn’t matter which projects people are working on. Digging holes and filling them back up again, in the Keynesian mindset. Economics is no longer related to decision making.

  3. Major_Freedom says:

    Nobody knows potential GDP, least of all those looking to reduce the output of others for any reason.

  4. Transformer says:

    I just noticed an odd thing,

    In Krugman’s response he says:

    ‘So the CBO already takes into account the effect of a smaller capital stock on potential output. That’s part of the reason CBO’s projections of future potential have in fact been marked down since the crisis began.’

    He seems to be pretty much saying that since 2007 the CBO has reduced potential GDP because of a smaller capital stock. Isn’t that pretty much your point all along ?

  5. Transformer says:

    DeLong: I see no reason to think potential GDP has fallen, why is the fed thinking of tapering ?
    Murphy: No reason why potential GDP has fallen? Look at the huge drop in investment levels.
    DeLong: Your an idiot Murphy, that drop is small beer – it would only causer a 1% drop in potential
    Krugman: Your even more of an idiot than Delong thought, Murphy. The CBO has already reduced potential as a result of the fall in the capital stock !
    Murphy: Your both idiots. The CBO numbers show that potential growth has dropped 25% since 2007 and that has had an impact on potential GDP of about 3.6%
    DeLong: Unless the fall in investment led to at least a 5% fall in potential I’m still right and your still an idiot.
    Krugman (in 2018): Actually the fall in potential was 8% but of the various theories none see the drop in investment levels as big enough to explain even most of that 8%.
    Transformer: Why did I waste hours of my life trying to get to the bottom of this crap ?

    • Bob Murphy says:

      That’s pretty good Transformer except this: “Krugman (in 2018): Actually the fall in potential was 8% but of the various theories none see the drop in investment levels as big enough to explain even most of that 8%.

      You just made that up, right? I’m not saying that with hostility, I’m genuinely asking. Forget our current dispute, if Krugman said that somewhere, I would like to know.

      • Transformer says:

        He says:

        ‘….CBO’s projections try to take the incentive effects of tax policy and the effects of deficits on investment into account, but these are small change compared with what you would need to link the Great Recession to the massive downgrading of long-run growth projections.’.

        As you called me on it – I checked more careful and he later also says in the section on “hysteresis”:

        ‘But you can come up with various stories about how a big slump could hurt long-run prospects in other ways, for example by depressing business investment, including spending on R&D.’

        I think overall its fair to say that none of his three theories center much on the capital stock or investment levels (that I certainly would assume would be pretty key in assessing potential growth!).

        • Bob Murphy says:

          Transformer I really don’t know what to say. You literally quote him saying the hysteresis story works for example by depressing business investment, and then you conclude that none of his theories centers on investment.

          • Transformer says:

            Mentioning something in passing , and making it the center of a theory are sort of different.

            But whatever. I think this is a sign that its time I stopped thinking about this any more and moved on.

Leave a Reply to Bob Murphy

Cancel Reply