In a perfect world

Regular readers are sick of this, so just take a vacation and come back next week.  I keep being told that no one’s interested even as I have record volume of viewers.  (Didn’t Yogi Berra say “nobody goes there anymore, it’s too crowded.”)  I’m interested, and I’m the one who decides.

In a perfect world I’d lay out a concise logical proof that Simon Wren-Lewis and Paul Krugman are wrong.  And number each point.  They’d respond saying which of my points were wrong, and why.  Then I’d reply.  It would be something like this:

1.  Wren-Lewis claimed actual C would fall, but would decline less than actual Y-T.  I claim that because he said it, and because he assumed consumption smoothing.

2.  Point one implies that actual private saving fell, because actual saving equals actual income minus actual consumption. 

3.  Wren-Lewis was considering the balanced budget multiplier, so if actual private saving fell, then actual national saving fell by an equal amount.

4.  Because actual saving fell, actual investment also fell.  It might have been less fixed investment, or less inventory accumulation.  But actual investment fell.  This is because even Wren-Lewis admits that S=I is an identity for actual saving and actual investment (not desired saving and investment.)

5.  Then Wren-Lewis said fiscal stimulus with consumption smoothing would work because G would rise more than C fell , and hence AD would rise. But this ignores the fall in investment.  He implicitly treated the extra saving [edit: I meant reduced saving] as if it just disappeared.  And that’s not just wrong, it’s wrong in a very important way for us non-Keynesians.  (Just as the Treasury view drives Keynesians insane.)

I certainly don’t expect Krugman to attempt to refute any of these 5 arguments (how could he?), but lesser lights over in the liberal macroeconomic blogosphere are welcome to take a stab at it.  In a few days I’ll report back on the number who took up the challenge. 

I now face an unprecedented flood of comments.  Some, like the first comment on the previous post, provide comic relief.  But I simply don’t have the time to respond to 5,000,000 commenters who say “I’ve never taken a macro course, but I wondered why S=I.  Suppose I just put money in the bank?”  Go look at all my previous responses, I’ve answered every question.

And no offense readers, I do welcome your interest.  And I take partial blame for not doing the 1-5 list earlier.  But I have simply run out of time.  So take any comments that you think actually address this issue, and bring them over here. 

The rules are as follows.  In this post you must start your comment:  “I disagree with point x.”  And then you must tell me why.  Otherwise I won’t respond.  Let’s do other topics on other posts.  This will streamline things and help me save lots of time.  I won’t have to repeat the same answer a hundred times.

People are telling me “Please stop, you’ve made your point.  You’re right, now just move on.”  Glad to hear I’m right, but I won’t stop until I hear that from one particular person, and he doesn’t teach at Oxford.

PS.  There’ll be plenty of monetary stuff next week with the big Fed meeting.  But I just can’t let this go.  I have a personality disorder where I get annoyed when people with big platforms make me look like a fool, even though I know I’m right.  And I’m just too busy right now to go to a psychiatrist to get it treated.

PPS.  Relevant links are in the previous two posts.


Tags:

 
 
 

226 Responses to “In a perfect world”

  1. Gravatar of There is no such thing as fiscal policy « The Market Monetarist There is no such thing as fiscal policy « The Market Monetarist
    18. January 2012 at 12:35

    […] fantastic family in Denmark – while my family is now asleep I see that Scott has yet another fiscal policy comment. Advertisement GA_googleAddAttr("AdOpt", "1"); GA_googleAddAttr("Origin", "other"); […]

  2. Gravatar of JKH JKH
    18. January 2012 at 12:37

    Disagree with points 3, 4, and 5:

    ……….

    S = I is an identity only in a closed, balanced budget economy.

    Consider the balanced budget multiplier in two stages – government spending injection; plus an equivalent tax hike.

    In the first stage, unless the government expenditure multiplier is zero, the consumption expenditure multiplication forces a government budget deficit at the margin. This comes out of the math, inevitably. So, if the starting point is a balanced budget, the post-multiplied GDP will feature a budget deficit, if the multiplier is positive.

    I.e. a balanced budget is an untenable constraint over the course of consumer expenditure multiplication, unless the multiplier is zero.

    At that point, the initial S = I assumption will not hold. With a positive multiplier, T will be less than G, and in effect, S will in part fund G to plug the deficit gap.

    So, in sequential context, the tax increase stage starts not with a balanced budget, but a deficit.

    The matching tax increase then reduces both C and S. Whether or not S returns to its original level, I does not change. S will be out of balance with G, unless the balanced budget multiplier turns out to be zero.

    And the change in AD will be explained by C + G, with no change in I.

    A change in I is not forced by a change in S, because the budget can’t necessarily be constrained to balance, post expenditure multiplier and post balanced budget multiplier.

  3. Gravatar of ChacoKevy ChacoKevy
    18. January 2012 at 12:49

    “I now face an unprecedented flood of comments.”

    I bet if you combine these posts with the treason post of last year, you could break the internet. It would be fitting on SOPA blackout day :)

    Condolences*petty gloating* on the Packers…

  4. Gravatar of D R D R
    18. January 2012 at 13:13

    “5. Then Wren-Lewis said fiscal stimulus with consumption smoothing would work because G would rise more than C fell , and hence AD would rise. But this ignores the fall in investment. He implicitly treated the extra saving as if it just disappeared. And that’s not just wrong, it’s wrong in a very important way for us non-Keynesians. (Just as the Treasury view drives Keynesians insane.)”

    And then you lose me.

    Wren-Lewis started his model with “If you spend X at time t to build a bridge, aggregate demand increases by X at time t.”

    Thus, Y rises by the same about as G. And C+I must remain constant. If there is consumption smoothing of a factor of F, this implies C rises by X/F and I falls by X/F.

  5. Gravatar of J Mann J Mann
    18. January 2012 at 13:21

    No need to respond to me, Scott, but just FYI – I love reading your responses, and encourage your current plan of focusing on highest value responses rather than cutting them out entirely.

  6. Gravatar of Neil L Neil L
    18. January 2012 at 13:27

    Why don’t you just call up Wren-Lewis or Krugman, and you all can sort out what is is exactly that you disagree on? It doesn’t seem like the current format of the debate has been enlightening towards any of the participants.

  7. Gravatar of Bob Murphy Bob Murphy
    18. January 2012 at 13:35

    Scott,

    I had a similar experience with the “can the federal debt burden our grandchildren?” debate a few weeks ago. Like you, I came up with what I thought was an ironclad demonstration that Krugman had screwed up. Unlike you, Krugman ignored me.

    For your debate, I think you are forgetting that nobody else on Earth is paying as careful attention to the back-and-forth as you are. So it helps to remind people on every post what your overall view is. For example, at one point I thought you were saying that you agreed with Cochrane that fiscal policy even in principle couldn’t stimulate AD this year, but now in retrospect it’s obvious you weren’t saying that.

    Here is the progression in the debate as I see it. You can hopefully cut Krugman some slack for not getting your (relatively subtle) contribution:

    1) John Cochrane uses Ricardian Equivalence-esque arguments to claim that fiscal policy, in principle, can’t stimulate aggregate demand this year.

    2) Krugman et al. go nuts, saying that even if you believed in Rational Expectations (which they don’t), Cochrane would still be wrong. To see how wrong he is, they reminded people about the phenomenon of consumption smoothing.

    3) Scott Sumner says, “Whoa! Krugman et al. are idiots! Consumption smoothing has nothing to do with it. If it *did* have something to do with it, then we’d violate S = I.”

    4) Krugman et al. go nuts (again). “Scott Sumner is an idiot! Of course we know that S = I. Look, we’re talking about comparative statics. Duh.”

    Do you see how this is really confusing? You were implicitly claiming, “Krugman, you used an invalid argument to attack Cochrane’s invalid argument.”

  8. Gravatar of Bob Murphy Bob Murphy
    18. January 2012 at 13:37

    * Oops in point (2) I meant to say, “…even if you agree with Ricardian Equivalence (which they don’t)…”

  9. Gravatar of D R D R
    18. January 2012 at 13:40

    I disagree with point 5.

    Wren-Lewis started his model with “If you spend X at time t to build a bridge, aggregate demand increases by X at time t.”

    Thus, Y rises by the same about as G. And C+I must remain constant. If there is consumption smoothing of a factor of F, this implies C rises by X/F and I falls by X/F.

    There lies the mysterious fall in investment. (But it is offset with a mysterious rise in consumption.)

  10. Gravatar of StatsGuy StatsGuy
    18. January 2012 at 13:40

    partly reposted from earlier thread:

    K writes: “First I ask how much this would increase savings, holding GDP constant; then I ask how much GDP has to fall to restore the equality between savings and investment.”

    The notion here is that there is a temporary disequillibrium between S and I. I think it’s fair to think that Krugman did in fact mean this… Here’s a 2009 post:

    http://krugman.blogs.nytimes.com/2009/01/27/a-dark-age-of-macroeconomics-wonkish/

    “after a change in desired savings or investment something happens to make the accounting identity hold. And if interest rates are fixed, what happens is that GDP changes to make S and I equal.”

    I (and others) accept that Krugman is sometimes sloppy with terms… I’m still not entirely sure what he means by Savings (desired resource savings?). But I do think he has a legitimate claim that he (not sure about wren lewis) really did mean desired savings, or at least savings as an intermediate construct before equillibrium is restored.

    I do understand that you have given him the benefit of the doubt just as much as he’s given you the benefit of the doubt – which is never.

    I think K’s larger argument is that, with consumption smoothing, increasing G and T can increase Y even in the presence of a CB targeting an interest rate. [at least when SRAS is flat…] You seem to actually agree this is possible, though empirically weak.

    I think a _real_ question instead of this he said/she said stuff – is whether it is even possible for increasing G and T simultaneously to increase AD. I would like to know whether Krugman agrees with you that this is impossible, or whether he can construct any plausible scenario where it might be possible.

    We know Krugman would argue that if the CB targets the interest rate at zero bound (aka, they are at the zero bound and are too chicken to “go negative”, but would like inflation higher) then increasing G and T can certainly increase AD (or Y) if an inflation increase is permitted (WWII proved that adequately).

    Perhaps you could run a post titled “What Krugman and I agree on”???

    I still cannot fathom why you feel compelled to defend Cochrane, other than loyalty to your alma mater. Although, admittedly, nothing boosts ratings like a good throw down.

  11. Gravatar of Hubert Hall Hubert Hall
    18. January 2012 at 13:41

    D.R. or somebody. I feel really stupid reading that. You (and Wren-Lewis) seem to be saying “Y rises by X, and some other stuff might change, therefore Y rises by X”. I don’t get it.

  12. Gravatar of RueTheDay RueTheDay
    18. January 2012 at 13:49

    “Because actual saving fell, actual investment also fell”

    Actually, according to Keynes, Investment determines Savings, not vice versa.

    This isn’t just a minor nit, it’s pretty key to how the S=I identity works: Investment creates the necessary Savings to make them identical.

  13. Gravatar of D R D R
    18. January 2012 at 13:51

    I disagree with point 1.

    Wren-Lewis only states that C falls in response to the tax cut. He does not say there is no corresponding rise in response to the bridge.

    (See prior objection to point 5)

  14. Gravatar of D R D R
    18. January 2012 at 13:57

    “This isn’t just a minor nit, it’s pretty key to how the S=I identity works: Investment creates the necessary Savings to make them identical.”

    Sorry, but I believe Scott means “having observed a fall in savings, we must also observe a fall in investment.” It’s not an argument about causality, but rather an argument about outcomes.

  15. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 14:02

    D R:

    “Wren-Lewis started his model with “If you spend X at time t to build a bridge, aggregate demand increases by X at time t.””

    “Thus, Y rises by the same about as G. And C+I must remain constant. If there is consumption smoothing of a factor of F, this implies C rises by X/F and I falls by X/F.”

    “I” does not have to remain constant in a consumption smoothing world.

  16. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 14:03

    D R:

    Because of that fact alone, C+I does not remain constant.

  17. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 14:05

    RueTheDay:

    “Because actual saving fell, actual investment also fell”

    “Actually, according to Keynes, Investment determines Savings, not vice versa.”

    Keynes said many contradictory things. Here, he sometimes said S=I, and sometimes he said I can lag S.

  18. Gravatar of Lee Kelly Lee Kelly
    18. January 2012 at 14:09

    It’s the same old story: Keynesians regularly assume that savings disappear into a black hole.

    In the spirit of Nick Rowe’s post about abolishing ‘saving’, I say we abolish the distinction between investment and consumption. The distinction isn’t unimportant, it’s just not all that important to monetary policy. It really doesn’t matter much how NGDP breaks down between investment and consumption spending (savings rates change, no?), so long as NGDP is held to some sensible target.

  19. Gravatar of Jon Jon
    18. January 2012 at 14:16

    I disagree with point 5. Investment need not fall because the decrease in desired savings causes an increase in Y, so disposible income can stay the same (or increase) and consumption is smoothed without decreasing actual saving.

  20. Gravatar of StatsGuy StatsGuy
    18. January 2012 at 14:21

    “I think a _real_ question instead of this he said/she said stuff – is whether it is even possible for increasing G and T simultaneously to increase AD.”

    That SHOULD say …”under a regime where CBs target the inflation rate”.

    “I would like to know whether Krugman agrees with you that this is impossible, or whether he can construct any plausible scenario where it might be possible.”

    That probably makes more sense.

  21. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 14:22

    Lee Kelly:

    “It’s the same old story: Keynesians regularly assume that savings disappear into a black hole.”

    And the flip side is that they regularly assume investment comes out of nowhere.

    “In the spirit of Nick Rowe’s post about abolishing ‘saving’, I say we abolish the distinction between investment and consumption. The distinction isn’t unimportant, it’s just not all that important to monetary policy. It really doesn’t matter much how NGDP breaks down between investment and consumption spending (savings rates change, no?), so long as NGDP is held to some sensible target.”

    I mean this in jest:

    It’s the same old story: NGDP pseudo-monetarists regularly assume that increased cash holdings and decreased nominal spending must be treated as if the cash holdings disappeared into a black hole, thus requiring the central bank to…increase people’s cash holdings(!) by way of taking assets out of the economy and replacing them with money, as if the people are made better off with fewer assets and more cash.

  22. Gravatar of D R D R
    18. January 2012 at 14:30

    Jon,

    I am way with you on that, but this does mean you better have some objection to an earlier point.

    By point 3 you have already accepted that actual savings has fallen.

    By “actual” Sumner is clearly driving at “observed” or “ex-post” or what-have-you.

    Plus, Sumner is doggedly slippery as to whether he is at any point discussing a particular sentence of Wren-Lewis’ argument, or the whole.

    In point 1, he seems to be discussing the tax cut in isolation. By implication, point 2 seems to be just the tax cut. In point 3 it’s the “balanced budget” etc.

  23. Gravatar of Lee Kelly Lee Kelly
    18. January 2012 at 14:34

    It’s the same old story: MMTers regularly assume that rising demand for cars and a falling quantity of cars sold (at a higher price) cannot be satisfied by swapping new cars for motorcycles they want less, as if people are going to be better off with more cars and fewer motorcycles.

    Swap ‘cars’ for ‘money’ and ‘motorcycles’ for ‘assets’.

  24. Gravatar of D R D R
    18. January 2012 at 14:36

    “‘I’ does not have to remain constant in a consumption smoothing world.”

    I did not even remotely suggest that. Is there some part of “I falls by X/F” that is unclear?

    C+I must absolutely remain constant if the bridge is taken in isolation (*and* the spending multiplier is 1.0). Y must by assumption rise with G, so C+I=Y-G must not change.

  25. Gravatar of Lee Kelly Lee Kelly
    18. January 2012 at 14:39

    Oops, I mangled the above comment. This is what I meant:

    ‘It’s the same old story: Market monetarists regularly assume that rising demand for cars and a falling quantity of cars sold (at a higher price) can be changed by swapping cars for motorcycles, as if people are going to be better off with more of what they want and fewer of what they don’t want.

    Swap “cars” for “money” and “motorcycles” for “assets”.’

  26. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 14:40

    D R:

    I know you didn’t suggest that. I am saying it is important to WL’s argument, because he didn’t take investment into account when he tried to disprove BBM = 0.

  27. Gravatar of Bababooey Bababooey
    18. January 2012 at 14:43

    “I disagree with point x.”

    Why?, because you told me to.

    Kidding, but I write to remind that it’s OK to read and not comment at all. If you have questions or oft-mentioned objections, definitely look at the FAQ or Scott’s publications first.

    Its especially good for noneconomists to remain silent when the economists are talking about how best to jam complexity into formalistic models; a method which isn’t helped by inductive reasoning from the observations of non-economists (or of economists unfamiliar with the model at issue).

  28. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 14:48

    Lee Kelly:

    “It’s the same old story: MMTers regularly assume that rising demand for cars and a falling quantity of cars sold (at a higher price) cannot be satisfied by swapping new cars for motorcycles they want less, as if people are going to be better off with more cars and fewer motorcycles.”

    HAHA, a player!

    It’s the same old story: Post Keynesians regularly assume that full employment is impossible in a competitive market by way of falling prices and wage rates, because investors are perpetually stuck in an ephemeral no man’s land of paranoia and fear without having any means of finding out what consumers want and where profitable investment opportunities reside, unless and until the government can successfully bring about an increase in aggregate demand that no investor invests into and no seller sells into anyway, as if McDonald’s integrates “aggregate demand” into their investment in hamburger meat.

  29. Gravatar of Daniel Daniel
    18. January 2012 at 14:48

    Bob Murphy’s analysis, above, seems exactly right to me about how Krugman came to misunderstand you.

    You and Krugman have something in common: you’re both unforgiving of illogic. Yet you’re both far more comprehensible when you analyze the economy than when you analyze others’ logical errors.

    That’s fine with me. We need better public economists far more than we need better public logicians.

  30. Gravatar of Paul Krugman to the Rescue (I’m serious…but even more so Milton Friedman to the Rescue) « Portrait of the Economist as a Young Man Paul Krugman to the Rescue (I’m serious…but even more so Milton Friedman to the Rescue) « Portrait of the Economist as a Young Man
    18. January 2012 at 14:50

    […] a bunch of Keynesian style examples. In his most recent post, Sumner summarizes his argument: http://www.themoneyillusion.com/?p=12753 and suggests that he won’t stop pushing this argument until Krugman acknowledges or at least […]

  31. Gravatar of D R D R
    18. January 2012 at 14:55

    Major,

    That is utter nonsense. The fall in investment respecting the bridge follows directly from explicit assumptions which Wren-Lewis made.

    Unless you insist that consumers don’t smooth in response to the bridge… but do you really want to go there?

  32. Gravatar of tom m tom m
    18. January 2012 at 15:13

    I disagree with:

    Wren-Lewis was considering the balanced budget multiplier, so if actual private saving fell, then actual national saving fell by an equal amount.

    Wren-Lewis’s point has nothing to do with the BBM. He is simply expressing the logic of Ricardian Equivalence. If there is a temporary increase G then private wealth goes down by the present value of the taxes financing that increase. Let G go up by $1 for one year. Wealth goes down by $1. The present value of consumption out to an infinite horizon must equal wealth. W=C/r, so DeltaC= r DeltaW=-r where r is the real interest rate. Net effect: demand goes up by 1-r. This is what Cochrane missed, and you seem to have also. Investment has nothing to do with it.

  33. Gravatar of Scott B. Scott B.
    18. January 2012 at 15:22

    I’m going to have to side with my imaginary view of what Yglesias might think about all this seemingly-neverending squabble, which is that I’m neither a Keynesian partisan nor a monetarist partisan, but I’m sure as heck not a Cochrane partisan. Cochrane and similar freshwater economists, to my ears, have come to the conclusion that nothing can be done but let the recession end itself, more Austrian than anything else.

    Whether the government builds some bridges or Ben Bernanke sends $1000 checks to every citizen, I don’t really care. But someone has to spend some money, and these days the only thing I ever seem to hear about is how well the rich are doing, either through investment income or tax cuts. And I know one thing for sure, rich people spend a good chunk of their money chasing scarce goods like beachfront property and old paintings, which doesn’t employ very many people.

  34. Gravatar of tom tom
    18. January 2012 at 15:31

    Number 5 is false.
    Krugman (via wikipedia): “Note, that as such, this does not imply that an increase in savings must lead directly to an increase in investment. Indeed, business may respond to increased inventories by decreasing both output and intended investment. Likewise, this reduction in output by business will reduce incomes, forcing an unintended reduction in savings. Even if the end result of this process is ultimately a lower level of investment, it will nonetheless remain true at any given point in time that the S=I identity holds.”
    In short: increase in savings may lead to a lower level of investment and still S=I all the time.

    And Woodford is even better because he considers different macro cases.
    You are wasting your time Scott because you can’t win this discussion.

  35. Gravatar of Cthorm Cthorm
    18. January 2012 at 15:41

    Scott,

    I’ve bitched about these posts already. I can’t help it, but my eyes just glaze over when I see C+I+G=Y and it takes a lot of willpower to keep reading. But the posts have actually been enjoyable because I love seeing Krugman twist himself into knots to keep his ego inflated. Kudos to you for keeping up with it, and I hope you finally get the mea culpa you’re looking for. With that said, I can’t wait to get back to our regularly scheduled programming.

  36. Gravatar of Mike Sax Mike Sax
    18. January 2012 at 15:45

    I don’t see how Krugman has twisted himself into knots Cthorm. He has written 2 or 3 posts on this, Scott has written 10 or more.

    The reason Scott is so hot and bothered is that it is his ego that has been deflated. He hasn’t stopped and I suspect he wont. For a long time.

  37. Gravatar of D R D R
    18. January 2012 at 16:21

    Sorry, Scott. I know you have things to do, so the following is not about responding to me personally.

    After how many posts… then writing “Glad to hear I’m right, but I won’t stop until I hear [‘You’re right’] from one particular person, and he doesn’t teach at Oxford.”

    Is this not an admission that Scott is now past the point of reason? It no longer matters how right or wrong he is, for he will press on in the face of any argument whatsoever… until “one particular person” tells him he is right?

    Wow.
    You know, I got a lot out of this, but that’s just… sad.

    I stupidly hold out hope, but… wow.

  38. Gravatar of ssumner ssumner
    18. January 2012 at 16:24

    JKH, Both Wren-Lewis and I are assuming a closed, balanced budget economy. That’s not the issue being disputed.

    So your complaint has no merit. I’m quite certain even Wren-Lewis would agree with me.

    DR, No, he doesn’t assume C rises, he assumes C falls.

    Statsguy, Krugman does not believe there is any lag in actual S and I. Desired is irrelevant for this debate.

    Sure we can make excuses–I did a post that excuses Cochrane and Lucas, and made their comments sensible in context. It’s easy to do.

    DR, He looks at an equilibrium outcome where C has fallen. Otherwise there’s no consumption smoothing involved. That excuse fails on two counts.

    Jon, Your comment makes no sense unless you disagreed with earlier points. Which ones?

    tom m, No, read all the Wren-Lewis and Krugman stuff. Both sides of the debate agree that RE shows that tax cuts may not work. The only issue at stake is the BBM, both Krugman and Wren-Lewis agree on that, at least in this current debate (they may have reservations elsewhere.) Wren-Lewis was evaluating the BBM.

    Tom, You are confusing dynamics with comparative statics. Both Krugman and I are looking at comparative statics here. We are considering a multiplier, for which you need actual outcomes of C, I, S etc.

    Mike Sax, I expected you to be able to attack at least one of the five. Can’t find a single flaw? I thought you told me that were a big time blogger, that all the other big time bloggers read?

    Everyone, Thanks, that was much easier than I expected.

  39. Gravatar of tom tom
    18. January 2012 at 16:30

    Scott I’m quoting Krugman so dont tell me that I’m confusing something. Just read his textbook or Woodford’s paper. Otherwise this debate will never end.

  40. Gravatar of hyperbollocks hyperbollocks
    18. January 2012 at 16:44

    2) is wrong. Because in the new equilibrium Y is higher, private saving does not fall. of course this assumes unemployment. Which we have.

    See every edition of Samuelsons intro textbook.

  41. Gravatar of tom tom
    18. January 2012 at 16:59

    Scott you want comparative statics here you go:
    In the first period savings fall and investment fall (by decreased invenories). S=I
    In the second period because of decreased inventories firms increase production and intended investment. Likewise, this increase in output by business will increase incomes, forcing an unintended increase in savings. S=I

    Is this so difficult Scott?

  42. Gravatar of D R D R
    18. January 2012 at 17:01

    Scott,

    Even if you make the unjustified claim that Wren-Lewis says consumption falls overall, your step 5 is still wrong.

    C -5 <- C falls, AND does so by less than Yd, which also fell (#1)
    I -20 <- Investment fell (#4)
    G +100
    T+100

    Y +75 <- AD still rises, *** #5 DOES NOT FOLLOW ***
    Yd-25
    Sp-20 <- Private savings fell (#2)
    Sg+0
    S -20 <- National savings fell (#3)

  43. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 17:01

    D R:

    “The fall in investment respecting the bridge follows directly from explicit assumptions which Wren-Lewis made.”

    No, it doesn’t. WL only said that consumption smoothing takes place, which reduces consumer spending and hence aggregate demand by less than the increase in spending financed by taxation.

    There is nothing “explicit” about investment there.

    “Unless you insist that consumers don’t smooth in response to the bridge… but do you really want to go there?”

    The point is that if consumers do smooth in response, that can’t disprove a BBM of zero.

  44. Gravatar of PrometheeFeu PrometheeFeu
    18. January 2012 at 17:03

    Like heck we are! Well, I actually had to take a break and am just catching up now because I didn’t have time to read all you were writing plus all the responses coming from everywhere and digest everything so I could figure out what I think. (Also, your posts read like policy debate speeches: hundreds of words per minute) But honestly, this is fascinating.

  45. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 17:03

    D R:

    “Even if you make the unjustified claim that Wren-Lewis says consumption falls overall, your step 5 is still wrong.”

    That is not what Sumner claimed. He claimed that WL incorrectly assumed that a fall in consumer spending that is less than the tax hike and spending, results in a rise in AD. He failed to take into account investment.

  46. Gravatar of ssumner ssumner
    18. January 2012 at 17:06

    DR, That was a joke. Not your fault, I have an unusual sense of humor. I’m actually about done.

  47. Gravatar of D R D R
    18. January 2012 at 17:08

    Major,

    Even if *you* are correct, step 5 still does not logically follow.

  48. Gravatar of D R D R
    18. January 2012 at 17:09

    Scott,

    Sorry I missed it. Care to address the deficiency in step 5?

  49. Gravatar of ssumner ssumner
    18. January 2012 at 17:12

    Tom, The fact that you read something doesn’t mean you interpreted it correctly. If I was wrong on that point lots of smart Keynesians I know would be hammering me. They are surprising silent.

    DR, I’m fine with that conclusion (Y went up) and always have been. I criticized his reasoning. Consumption smoothing doesn’t get you that result, it might accompany that result, or not.

    If we are judging people on whether the conclusion is right, then I’d give Cochrane a pass, as his conclusion (zero multiplier, it probably about right.

  50. Gravatar of JPIrving JPIrving
    18. January 2012 at 17:12

    Meanwhile MV still equals PY and Wren-Lewis/Krugman have made a not so meaningful observation even if they *were* right.

  51. Gravatar of D R D R
    18. January 2012 at 17:16

    Ha. Fool me once. That was clearly your humor showing.

    You have found no fault in his reasoning. You laid out your argument, and it did not hold.

  52. Gravatar of tom tom
    18. January 2012 at 17:17

    Ok Scott last question (I have to go to sleep it’s late here): Do you agree that if you consider 2 periods there are no problems with Wren or Krugman reasoning?

  53. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 17:42

    D R:

    “Major,”

    “Even if *you* are correct, step 5 still does not logically follow.”

    Step 5 isn’t a syllogistic assertion deduced from one or any of the steps 1 to 4. Step 5 is just what WL argued.

  54. Gravatar of rob rob
    18. January 2012 at 17:45

    Scott (or anyone else who knows the answer),

    Can you clarify #4 ?

    As a response to T people adjust the way they divide their now smaller disposable income between S and C. If they save less and consume more then must increase. This extra spending will reduce inventories. I understand that the claim is being made that these smaller inventories mean a reduction in I to bring it in line with the lower S. But which ever you way you count it I don’t see how one can deny that expenditure (and therefore AD) has increased.

    Can someone clarify ?

  55. Gravatar of TravisA TravisA
    18. January 2012 at 17:56

    D R: you have indefatigable energy in contesting this question :-)

    In your latest example, you show that Y can increase. Again, Scott doesn’t dispute that Y can increase under Keynesian assumptions. Scott is saying that Wren-Lewis’ explanation is flawed. WL says, “Put the two together and aggregate demand rises.” WL needed to say, “Put the *three* together and AD (still) rises.” WL omitted that under his assumptions, investment must fall. Here’s Scott’s point from a previous post:

    http://www.themoneyillusion.com/?p=12729

    Consider their assumptions:

    1. A tax-financed increase in G.

    2. Consumption declines by less than after-tax income (consumption smoothing.)

    That’s pretty general. I defy anyone to construct an example using those two assumptions, where the Wren-Lewis quotation is correct. Find an example where the final change in AD can be described as the change in C plus the change in G.

  56. Gravatar of D R D R
    18. January 2012 at 17:56

    rob,

    Sumner is arguing that
    “If you raise taxes by X at time t, consumers will smooth this effect over time, so their spending at time t will fall by much less than X”

    Actually means
    “If you raise taxes *and government spending* by X at time t, consumers will smooth this effect over time, so their spending at time t will fall by much less than X”

    I have no idea why he thinks that reading makes any sense, but he does.

  57. Gravatar of D R D R
    18. January 2012 at 17:59

    Actually, that it means

    “If you raise taxes *and government spending* by X at time t, consumers will smooth this effect over time, so their *consumption* at time t will fall *and do so* by *an amount* much less than X”

    Pretty tortured.

  58. Gravatar of D R D R
    18. January 2012 at 18:02

    TravisA,

    That is easy.

    C +0
    I +0
    G +100
    T+100

  59. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 18:03

    Sumner, if you have time, I would like to clarify step 5.

    Recall:

    “Then Wren-Lewis said fiscal stimulus with consumption smoothing would work because G would rise more than C fell , and hence AD would rise. But this ignores the fall in investment. He implicitly treated the extra saving as if it just disappeared. And that’s not just wrong, it’s wrong in a very important way for us non-Keynesians. (Just as the Treasury view drives Keynesians insane.)”

    You said that investment can be viewed as being composed of new capital purchases and build up of inventory.

    Your criticism of WL concerning investment only works if the investment is comprised of new capital purchases only. It does not work if any of the investment is composed of inventory build up.

    Why is this? Because the context here is aggregate spending.

    If consumers smooth their consumption in response to a tax hike, such that G rises by more than C falls, which then is supposed to mean (and this is the controversial assumption) that AD rises as well, then it is no argument against this to say that it ignores investment of the inventory build up type. This is because such inventory build up doesn’t have any additional spending associated with it. It is building up because there is a LACK of spending associated with it, namely, the lack of spending from the consumers.

    Thus even if WL is ignoring investment of the inventory build up type, then including it cannot contribute to AD.

    Only if the investment that is ignored is comprised of new capital purchases only, will the “don’t forget that investment can fall” argument be considered a valid counter-argument to WL. This is because if investors spend less on new capital goods, then that will be able to be used in conjunction with C to say “The effect of G on AD has been nullified by the drop in C and I (of the capital purchase type).”

  60. Gravatar of Random thoughts « Economic Sophisms Random thoughts « Economic Sophisms
    18. January 2012 at 18:03

    […] of my readers will be at least aware of the Sumner v.s. Wren-Lewis +Krugman debate. Once I saw that their fight was over fiscal multipliers I sort of stopped reading. I am too busy […]

  61. Gravatar of rob rob
    18. January 2012 at 18:11

    RE-POSTED AS I MANGLED MY FIRST ATTEMPT ABOVE

    Scott (or anyone else who knows the answer),

    Can you clarify #4 ?

    As a response to T people adjust the way they divide their now smaller disposable income between S and C. If they save less and consume more then expenditure must increase. This extra spending will reduce inventories. I understand that the claim is being made that these smaller inventories mean a reduction in I to bring it in line with the lower S. But which ever you way you count it I don’t see how one can deny that expenditure (and therefore AD) has increased. G has increased by more than C falls, and Investment in money stays the same even if the Inventory adjustment means that it has gone down in accounting terms

    Can someone clarify ?

  62. Gravatar of D R D R
    18. January 2012 at 18:19

    rob,

    Scott literally does not care what is and is not valid. His argument is as follows:

    Investment fell, and Wren-Lewis did not discuss it. Therefore, Wren-Lewis is wrong.

    I think Sumner is wrong, because when Wren-Lewis discussed the bridge, it follows immediately (from the multiplier and consumption smoothing) that investment must have fallen at that point.

  63. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 18:27

    rob:

    “As a response to T people adjust the way they divide their now smaller disposable income between S and C. If they save less and consume more then expenditure must increase. This extra spending will reduce inventories. I understand that the claim is being made that these smaller inventories mean a reduction in I to bring it in line with the lower S. But which ever you way you count it I don’t see how one can deny that expenditure (and therefore AD) has increased. G has increased by more than C falls, and Investment in money stays the same even if the Inventory adjustment means that it has gone down in accounting terms”

    This is what I had in mind too.

    This is why I think the only relevant investment to the issue of AD is the new capital purchases type of investment, not the build up of inventories type of investment.

    If you tax me $100, and then I respond by reducing my spending, then assuming I am consumption smoothing, which means I reduce my consumption by $20, then the only way that “our” aggregate spending can remain the same as before, is if I reduce my investment SPENDING by $80.

    If we just focus on the build up of consumer goods inventory that results from my $20 drop in consumption spending, and your spending of $100 which decrease inventory, then “our” consumption activity generated a net $80 decrease in inventory. But that $80 value of reduced inventory cannot be called a decrease in spending that nullifies your $100 tax financed spending, thus creating a BBM of zero.

    The only way that I can “nullify” your $100 spending, assuming I reduce my consumption by $20 (smoothing), is if I reduce my new capital purchases by $80.

    Having said that, the fact that WL ignored this, and assumed that a drop in C by itself can lead to a drop in AD, is still enough to make Sumner’s criticism correct, IMO.

  64. Gravatar of Again: I Does Not Equal S « The Traders Crucible Again: I Does Not Equal S « The Traders Crucible
    18. January 2012 at 18:33

    […] is Scott Sumner responding to JKH over his statement “you are assuming a closed economy with a balanced […]

  65. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 18:35

    D R:

    “I think Sumner is wrong, because when Wren-Lewis discussed the bridge, it follows immediately (from the multiplier and consumption smoothing) that investment must have fallen at that point.”

    No, that does not follow. It is possible that investment spending does not fall at all.

    If you take $100 from me, and then I reduce my consumption by $20, then it does not follow that I must have reduced my investment spending, by $80 or any other amount.

    But this is not even the correct way to view what happens anyway. If you tax me $100, then the correct way to do “comparative statics”, IMO, is to ask what I would have done with that money had you not taken it from me. Since you only tax what I earn through exchanges, and not my bank account, then taxing my transactions by $100 means that it is almost certainly the case that had you not taxed me, the entire $100 sum would have otherwise been in circulation, contributing to AD.

    So by you taxing me and then spending it, you’re almsot certainly not ADDING anything to AD that I would not have added myself had you not taxed me $100, which means tax financed stimulus almost certainly has a BBM of zero.

    That’s how I think this problem should be approached, not comparing the government taxing and spending of $100 on the one hand, and what the taxpayer does IN ADDITION to that, after the tax, on the other.

  66. Gravatar of D R D R
    18. January 2012 at 18:36

    rob,

    I apologize. I have no idea how to answer your question because the premise is all wrong from the start.

    Once you accept step three– that national savings declined (that when all is said and done that in the end it did decline)– step four immediately follows.

    However S and I rebalanced, if it involved a fall in S, it involved a fall in I.

    There simply isn’t anything more to that step.

  67. Gravatar of D R D R
    18. January 2012 at 18:45

    Major,

    Give it up. Seriously. Are you *literally* trying to waste my time?

    “No, that does not follow. It is possible that investment spending does not fall at all.

    “If you take $100 from me…”

    STRAW MAN. The bridge is treated completely separate from the tax.

    If you insist on reading Wren-Lewis so that first sentence is in the setting of a balanced budget, then the next two sentences are completely and utterly meaningless. End of story.

    It is plain from reading Wren-Lewis that he is trying to separate the effect of the bridge from the effect of the tax. That is why he combines the two effects at the end.

  68. Gravatar of rob rob
    18. January 2012 at 18:51

    D. R.

    If I had to put money on it I would bet on Scott being right – just can’t see past that step #4 right now, but I’m keeping my options open.

  69. Gravatar of D R D R
    18. January 2012 at 18:56

    rob,

    He’s going to be right in his own mind, because he doesn’t allow any interpretation of Wren-Lewis’ words than his own. It doesn’t matter that his interpretation is ridiculous– he won’t debate any point which does not accept that.

    Normally, one tries to be charitable, but Scott… he has dug himself in way too deep now to allow any charity. I guess they do teach game theory at Chicago.

  70. Gravatar of Mike Sax Mike Sax
    18. January 2012 at 19:12

    Scott it’s not about answering some menu that you try to pigeon hole the discussion about. To act as if the burden of proof is on me or anyone else to answer these 5 arbitary points of yours is not the point.

    You can say becauwse I don’t answer the question on your terms you’ve won the argument. I can point out that you don’t answer my argument on my terms-where does that get us. The larger point is that you’re trying to defend Cochran and Lucas who employ some lousy arguments to claim that fiscal stimulus doesn’t work.

    You claim that they miss something too but you object to Krugman calling them ludicrous, though they are.

    You can’t asnwer why Krugman is criticized as uncivil next to all the uncivil things you have said about Keynesianism.

    Then you come up with this false claim that savings means more investment-for some reason we have the words saving and investment but according to you they are the same thing.

    Now you have drawn up these 5 theses that whether they are true or false are besides the point.

    I agree with Tom. You haven’t won this debate. You can’t leave this disussion alone because you have lost something here. Your repuation certainly hasn’t benefitted from this.

    Trying to pigenhole the discussion into 5 points won’t solve this fact.

  71. Gravatar of math math
    18. January 2012 at 20:12

    [Some unorthodoxy below]

    Honestly speaking, I don’t think that Sumner is correct here (he was correct w.r.t. consumption smoothing only but here he makes a bigger claim).

    The difference is how Monetarists and Keynesians treat the models — the former take a bottom-up approach while the latter do a top-to-bottom approach.

    Here is what I mean. For Monetarists, identity S=I always holds as where else does the money go if not into investment? It can be seen as a dogma caused by the fact that they build their models from the ground up (somewhat irrespective of the reality).

    OTH, Keynesians don’t look at S=I as a dogma. They try to fit their models (not such rigorous as the Monetarists’ ones) to the surrounding reality. In reality, there is a gap between S and I (since there is banking system in-between serving as an intermediary where banks can “create” money in their books to satisfy the demand for investment). So, while S=I in the long run that identity can be violated in the short run. (But of course Keynesians have not created yet rigorous models to formulate that relation mathematically via rigorous equations. They prefer to play with the graphical charts and numerical examples.)

    So, I believe that no side wins here. There are some profound philosophical differences between them in their approaches to describe the economic reality in my view.

  72. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 20:24

    D R:

    “Major,”

    “Give it up. Seriously. Are you *literally* trying to waste my time?”

    DR, there is nothing for me to “give up” about. And if you feel your time is wasted, then you have the choice on what to do about that.

    “No, that does not follow. It is possible that investment spending does not fall at all.”

    “If you take $100 from me…”

    “STRAW MAN. The bridge is treated completely separate from the tax.”

    First, I wasn’t attributing to you any position that you did not have, so it can’t be a straw man. Second, the context of this whole debate is a taxed financed stimulus policy. It is not wrong to consider a thought experiment where there is taxation. Third, the bridge cannot possibly be separate from taxation if it is financed by taxation.

    “If you insist on reading Wren-Lewis so that first sentence is in the setting of a balanced budget, then the next two sentences are completely and utterly meaningless. End of story.”

    This makes no sense.

    “It is plain from reading Wren-Lewis that he is trying to separate the effect of the bridge from the effect of the tax. That is why he combines the two effects at the end.”

    The tax IS an effect itself that prevents a particular counterfactual world from taking place. That’s my personal argument on how to approach all this.

    On the other hand, whether the tax has a temporal effect on subsequent consumption and/or investment on the part of those taxed, is the debate between Sumner and WL.

  73. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 20:37

    Mike Sax:

    Your rants say nothing about the actual discussion at hand and smack of nothing but worthless antagonism and ad hominem.

    “Scott it’s not about answering some menu that you try to pigeon hole the discussion about. To act as if the burden of proof is on me or anyone else to answer these 5 arbitary points of yours is not the point.”

    Burden of proof? Sumner just laid out his side of the story. If you don’t agree, fine. But don’t mistake yourself to be given some unfair burden of proof.

    “You can say becauwse I don’t answer the question on your terms you’ve won the argument. I can point out that you don’t answer my argument on my terms-where does that get us. The larger point is that you’re trying to defend Cochran and Lucas who employ some lousy arguments to claim that fiscal stimulus doesn’t work.”

    So in other words, your beef isn’t with the argument Sumner is actually making, your beef is with the conclusion that goes against your religion that fiscal stimulus doesn’t work. For the millionth time, Sumner isn’t saying that fiscal stimulus doesn’t “work”, in the sense that it does not raise AD. You’re just so blinded by ideology that you can’t even see what he saying. He is just saying that WL’s argument about consumption smoothing can’t get to the conclusion that tax financed spending “works.”

    This isn’t about “the good guys who support fiscal policy” and “the bad guys who don’t support fiscal policy.”

    “You claim that they miss something too but you object to Krugman calling them ludicrous, though they are.”

    Oh, so your a raging ideologue who thinks those who question anyone who supports fiscal policy are “ludicrous.” Well, you’re ludicrous. Where does that get us?

    “You can’t asnwer why Krugman is criticized as uncivil next to all the uncivil things you have said about Keynesianism.”

    You sound mad. Maybe you need to chill out and realize that not everyone is brainwashed into believing in Keynesianism.

    “Then you come up with this false claim that savings means more investment-for some reason we have the words saving and investment but according to you they are the same thing.”

    Saving means investment as long as you understand the national accounting identity S=I, which both WL and Krugman accept.

    “Now you have drawn up these 5 theses that whether they are true or false are besides the point.”

    No, they are on the friggin point. This is Sumner’s blog. He makes the points, not you.

    “I agree with Tom. You haven’t won this debate. You can’t leave this disussion alone because you have lost something here. Your repuation certainly hasn’t benefitted from this.”

    LOL, as if you’re a judge on anyone’s credibility! And who are you? Editor of the JFE?

    Sumner’s reputation is not threatened by the likes of you raging Keynesian cementheads who can’t stand it when their precious ilk are questioned.

    “Trying to pigenhole the discussion into 5 points won’t solve this fact.”

    It’s not a pigeonholing. It’s his argument that people like you are failing to grasp because you’re so emotionally distraught that anyone would dare suggest fiscal policy might have a zero BBM.

  74. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 20:40

    D R:

    “He’s going to be right in his own mind, because he doesn’t allow any interpretation of Wren-Lewis’ words than his own. It doesn’t matter that his interpretation is ridiculous- he won’t debate any point which does not accept that.”

    That is EXACTLY what you are doing. You refuse to accept the English language interpretation of WL, because in your mind you’re fixated so much on your own interpretation. You won’t debate any point which does not accept that.

    “Normally, one tries to be charitable, but Scott… he has dug himself in way too deep now to allow any charity. I guess they do teach game theory at Chicago.”

    Take your own advice. Sumner hasn’t dug any hole other than the one you fell into.

    You deserve no charity at all.

  75. Gravatar of John Schultz John Schultz
    18. January 2012 at 20:51

    Scott, I think you have a typo in Point #5.

    “But this ignores the fall in investment. He implicitly treated the extra saving as if it just disappeared.”

    Did you you mean “reduced saving?”

  76. Gravatar of D R D R
    18. January 2012 at 21:06

    “That is EXACTLY what you are doing. You refuse to accept the English language interpretation of WL, because in your mind you’re fixated so much on your own interpretation. You won’t debate any point which does not accept that.”

    Is that so? Let’s test this out. When Wren-Lewis says “If you raise taxes by X at time t, consumers will smooth this effect over time, so their spending at time t will fall by much less than X.” do you think he is talking about the effect of an increase in both G and T? Or just an increase in T?

  77. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 21:52

    D R:

    “Is that so? Let’s test this out. When Wren-Lewis says “If you raise taxes by X at time t, consumers will smooth this effect over time, so their spending at time t will fall by much less than X.” do you think he is talking about the effect of an increase in both G and T? Or just an increase in T?”

    How is testing me in any way related to the demeanor that I attributed to you that compelled you to want to “test [it] out”?

    Any way, back to what I think you’re asking…

    What WL was talking about was this: Assuming consumption smoothing, the effect of a tax hike is that consumption spending does not fall by as much as the tax hike and subsequent government spending, hence AD must rise.

    Sumner showed that’s wrong because AD does not necessarily rise here, because investment might fall.

    Suppose I told you this:

    Taxation on derivative contracts is going to rise, and the effect might be that investors are going to put forth a reduced expenditure.

    Now, is it a reasonable interpretation that I am trying to say that in response to the tax hike, consumers are going to reduce their consumption spending and investment spending?

    Or is it more reasonable to interpret it as me trying to say that in response to the tax hike, investors are going to cut back on their investment spending?

    Assuming you’re honest, you should hold the latter interpretation.

    Well, substitute “investors” for “consumers”, and you can see why the reasonable interpretation of WL is that he was talking about consumer spending and not investors investing.

  78. Gravatar of John Schultz John Schultz
    18. January 2012 at 22:24

    FWIW, I’ve changed my mind and now agree that Sumner is interpreting Wren-Lewis’ argument as he meant it and that it is wrong (and bad) in multiple ways. Furthermore, I can’t see any issues with Sumner’s criticism of it.

    At first, I thought that Wren-Lewis was trying to make his argument solely based on Ricardian Equivalence like Krugman did here:

    http://krugman.blogs.nytimes.com/2009/04/06/one-more-time/

    I thought this because I initially read Wren-Lewis’ argument in the context of Krugman’s argument, as I actually got there from Krugman’s posts on Cochrane/Lucas + Ricardian Equivalence, and Krugman seemed to think that Wren-Lewis was making the same argument as him too.

    I didn’t notice that Wren-Lewis was making a different argument than Krugman. I thought when he said “consumers will smooth this [tax] effect over time” he was referring to the way that under RE the private sector permanently (i.e. – “smoothing over time”) contracts in response to an expansion in government, even a temporary one, like in Krugman’s argument.

    I only realized that Wren-Lewis was making a different argument than Krugman when I read Sumner’s criticisms and then worked through his logic several times that pretty much decimated Wren-Lewis’ argument.

    At that point, I thought, “Wait a minute. I imagine an economics professor at Oxford is a very smart guy. Furthermore, since he is criticizing someone else for making basic errors he would be damn careful that he was right in his counter argument. Is Sumner interpreting him correctly? If we slightly alter Sumner’s interpretation of Wren-Lewis does most of the criticism go away?”

    I noticed that if you interpret the word “consumers” in Wren-Lewis’ argument as “the private sector” then you got something very similar to Krugman’s RE argument where, again, the smoothing is not Friedman consumption smoothing but rather an effect of RE permanently contracting the private sector in response to increased government spending.

    So, I thought this entire episode was all just a semantic misunderstanding and that Wren-Lewis’ particular choice of the word “consumers” was just an unfortunate red herring and causing confusion.

    Finally, what convinced me that this RE reading of Wren-Lewis is wrong and that Sumner’s interpretation is correct was not certain posters repeatedly asserting that there is no other possible interpretation and it is not possible that Wren-Lewis spoke imprecisely, but rather in re-reading Wren-Lewis’ post I re-read this:

    “I cannot imagine anyone who teaches New Keynesian economics, or who talked to people who teach New Keynesian economics, making this mistake. This is because, in these models, we do have to worry about aggregate demand. We focus on consumption smoothing, and Ricardian Equivalence, and teach it from the start.

    Yeah, OK, I give. Wren-Lewis’ argument was definitely based on Friedman’s concept of consumption smoothing, which is completely insufficient to yield his conclusion.

  79. Gravatar of JamesH JamesH
    18. January 2012 at 23:34

    Scott,

    I disagree with point 5.

    Is it possible that simple tensing issues are causing wide-spread confusion here? The accounting identity S=I means money saved equals money invested past tense. Clearly the all world investments are not equal to all world savings right now: investments go bad or up in value and savings can be lost.

    If, in an economy consisting only of apples, everyone suddenly puts half their paycheck under their mattress. Savings goes up (it just does – this counts as savings in the identity). Investment goes up too – as all the goods that haven’t been bought count as investment in a ‘stockpile’. That makes intuitive sense when you calculate GDP – all the apples that haven’t been bought, are clearly still part of the national production in that period.

    But after this accounting, anything can happen. All the money that’s been saved could be lost in a huge fire- in that case you still have all the investment apples. Or all the apples could be lost to mould, in which case you still have all your money.

    The point I’m making here, is that national income accounting says nothing about whether consumers spending their savings will increase GDP or not. In your consumption smoothing example, if everyone spends money that they’ve saved that doesn’t have to decrease investment.

  80. Gravatar of bob bob
    19. January 2012 at 01:28

    I think that I understand what makes this so confusing. In IS/LM we leave out the market for bonds, and assert it must clear by Walras Law.

    If G rises, the supply of bonds rises, so interest rates rise. So now we have delta_G = -delta_I. Cochrane is confused because he thinks it ends here. But with higher interest rates, there’s an excess supply of money. As the money market clears, private spending rises (C and I). That’s why unfunded fiscal policy can work.

  81. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 03:48

    John Schultz:

    “Yeah, OK, I give. Wren-Lewis’ argument was definitely based on Friedman’s concept of consumption smoothing, which is completely insufficient to yield his conclusion.”

    And notice how accepting this doesn’t even require you to accept either Cochrane’s story or WL’s story to be true as is?

  82. Gravatar of Scott Sumner on Aggregate Demand « Modeled Behavior Scott Sumner on Aggregate Demand « Modeled Behavior
    19. January 2012 at 04:00

    […] says Regular readers are sick of this, so just take a vacation and come back next week.  I keep […]

  83. Gravatar of D R D R
    19. January 2012 at 04:59

    “What WL was talking about was this: Assuming consumption smoothing, the effect of a tax hike is that consumption spending does not fall by as much as the tax hike and subsequent government spending, hence AD must rise.”

    My numerical example was entirely consistent with what you say here. Are you leaving out something important?

  84. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 05:07

    D R:

    “My numerical example was entirely consistent with what you say here. Are you leaving out something important?”

    Your numerical example presumes its own assumptions are true. It does not take into account any fall in investment. You have I as +0.

  85. Gravatar of D R D R
    19. January 2012 at 05:08

    “Sumner showed that’s wrong because AD does not necessarily rise here, because investment might fall.”

    Sumner tried to prove that Wren-Lewis was wrong on the grounds that *he could not logically be correct* without having failed to mention a fall in investment

    I’m saying Sumner is wrong, because either:

    1) Consumption smoothing applies to the first sentence of Wren-Lewis’ argument, in which case, it logically follows that the spending leads to a fall in investment, or

    2) Consumption smoothing does not apply to the first sentence of Wren-Lewis’ statement, in which case, it is not logical to try to consumption smooth in the balanced-budget setting, as Sumner tries.

    Either Sumner’s premise is wrong, or his result does not follow from his premise.

  86. Gravatar of D R D R
    19. January 2012 at 05:13

    “Your numerical example presumes its own assumptions are true. It does not take into account any fall in investment. You have I as +0.”

    *You* didn’t say anything about investment.

    Also, I had in mind my first example, C-5, I-20…

  87. Gravatar of D R D R
    19. January 2012 at 05:16

    “Sumner tried to prove that Wren-Lewis was wrong on the grounds that *he could not logically be correct* without having failed to mention a fall in investment.”

    Bah. “without having mentioned a fall in investment.

  88. Gravatar of John Schultz John Schultz
    19. January 2012 at 05:33

    DR, I do agree that Wren-Lewis is trying to reason about what happens by examining the two changes each in isolation and then somehow trying to combine their results to get what happens if you did both changes simultaneously.

    That actually makes Wren-Lewis’ reasoning even worse than Sumner gave him credit for, which you correctly characterized, as trying to examine what happens with a simultaneous increase in G and T.

  89. Gravatar of John Schultz John Schultz
    19. January 2012 at 05:42

    Wren-Lewis replies:

    http://mainlymacro.blogspot.com/2012/01/consumption-smoothing-and-balanced.html

    He clarifies that his model has savings but not investment (savings is all money stuffed in mattresses), which makes his argument coherent.

    The primary criticism of his argument then is that his model is nowhere near complex enough to try to predict a result for the real world.

  90. Gravatar of D R D R
    19. January 2012 at 05:42

    John,

    What do you mean? Wren-Lewis is wrong because the effects cannot be separable? Or that his reasoning was faulty, because the separability was only implied?

  91. Gravatar of D R D R
    19. January 2012 at 05:50

    John,

    Looks like it. I take it back. I overread Wren-Lewis, and he never said anything about investment at all. If there is no investment at all then my argument against Sumner is wrong, and Sumner’s argument against Wren-Lewis is wrong.

    I am OK with your subsequent criticism. But so long as the argument is consistent, it is enough to disprove Cochrane.

  92. Gravatar of Kevin Donoghue Kevin Donoghue
    19. January 2012 at 06:40

    John Schultz,

    Woodford’s 47-page paper has no mention of investment either. Y=C+G. This is quite usual in the literature. There’s no compelling reason to drag I into it unless you want to model the growth of the capital stock, which complicates (modern) macro in a big way. (It was easy enough for the Old Keynesians to include I, since they were very realistic about long-run mortality. But modern models involve infinite-horizon optimization.)

  93. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 06:47

    D R:

    “Your numerical example presumes its own assumptions are true. It does not take into account any fall in investment. You have I as +0.”

    “*You* didn’t say anything about investment.”

    Of course I did. Since the start. Investment falling has been the sticking point the whole time.

    “Sumner tried to prove that Wren-Lewis was wrong on the grounds that *he could not logically be correct* without having failed to mention a fall in investment.”

    Yes, and I think that is because Sumner treats inventory build up on account of consumers abstaining from spending, to be an “investment”. While in the national accounting sense that may be true, it contributes nothing to aggregate demand or NGDP. Only if investment spending changes, will there be a change to aggregate demand and NGDP. However, WL excludes such investment in his model, so he can’t be logically refuted, but he certainly can be refuted on realism grounds, in that assuming away investment in a model of a modern economy is like assuming away money in monetary policy.

    “I am OK with your subsequent criticism. But so long as the argument is consistent, it is enough to disprove Cochrane.”

    It is not enough to disprove Cochrane by assuming away invesment. Cochrane didn’t. You can’t refute someone’s conclusion by merely assuming away the premises of their conclusion. You have to show how those premises are flawed/incorrect. Since investment in the real world exists, assuming away investment cannot possibly refute Cochrane’s zero BBM.

    At any rate, I still contend that all this confusion can be eliminated if we approached the issue of government spending and taxation from a completely different angle.

    Instead of isolating the 100 tax and government spending and then asking what do civilians do with their remaining money, after which we compare them so see some alleged “net effect”, we should instead be considering all possible counterfactuals that would have otherwise taken place had the government not taxed and spent that 100. This is the only rational approach to fiscal policy, because the whole intention of fiscal policy is for the government to divert the economy away from where it otherwise would have gone in the absence of fiscal policy, to where the government intends the economy to go with fiscal policy.

    We can’t be asking what the effect of fiscal policy is by comparing time = 0 statics to time = 1 statics. That is silly. For all we would be doing is comparing two completely incommensurate “worlds”, namely, the past and the present. Fiscal policy puts us on an entirely different temporal trajectory going forward, as compared to no fiscal policy going forward.

    Even if fiscal policy at time = 0 was then followed by some economic reality at time = 1 where NGDP is higher, it says nothing on whether fiscal policy was a good idea or bad idea. In order to know for sure, we have to know what the world would have otherwise looked like at time = 1 had no fiscal policy taken place at time = 0. Of course that is impossible to do, but that doesn’t mean we should take the alternative and very false route of judging fiscal policy by looking at the economy over two or more subsequent time periods.

    My position is that tax financed government spending must be approached by asking what is the most likely counterfactual that would have otherwise taken place in the absence of fiscal policy, given what we know to be taking place in the presence of fiscal policy. I said that the most likely counterfactual would have shown tax financed stimulus to have a BBM of zero, or very close to zero, on the basis that taxation is collected on transactions, and not cash balances. Thus, the most likely counterfactual that would have taken place in the absence of tax financed government spending, is that the 100 in (now counterfactual) tax would have been put into circulation anyway by those making the exchanges. They have shown a preference in this world for putting that 100 into circulation with the tax, so it is likely that they would have done the same without the tax. Since the 100 would have therefore otherwise have been in circulation anyway and counting towards NGDP, it follows that tax financed government spending is most likely not stimulative at all.

  94. Gravatar of D R D R
    19. January 2012 at 08:30

    “Of course I did. Since the start. Investment falling has been the sticking point the whole time.”

    Shall I quote?

    “What WL was talking about was this: Assuming consumption smoothing, the effect of a tax hike is that consumption spending does not fall by as much as the tax hike and subsequent government spending, hence AD must rise.”

    No mention of investment.

    “It is not enough to disprove Cochrane by assuming away invesment. Cochrane didn’t. You can’t refute someone’s conclusion by merely assuming away the premises of their conclusion.”

    Cochrane DID NOT assume investment. Cochrane did not assume anything at all, except for his conclusion. He made no argument on this point at all except to assert that he was correct. Tax-financed infrastructure was not a case he covered.

    So if you want to go that route, I can say Cochrane is wrong because tax-financed infrastructure exists and I therefore dispute his “assumption” that is does not. His only defense would be to deny the possibility of such a project. Which would be really weird, considering his introduction.

  95. Gravatar of D R D R
    19. January 2012 at 08:37

    “we should instead be considering all possible counterfactuals that would have otherwise taken place had the government not taxed and spent that 100.”

    OK. Here is mine

    C+10 I+50 G+100 T+100

  96. Gravatar of Eclair Eclair
    19. January 2012 at 09:01

    @ Major_Freedom, S.Sumner

    “If you tax me $100, then the correct way to do “comparative statics”, IMO, is to ask what I would have done with that money had you not taken it from me.”

    That is true if you believe that government needs to tax people before it can spend. I believe this is not the way the current US system operates..?

    On Jan 18, 18:03 you restated the point made by WL in one of is recent posts, that:

    “Your [Sumner’s] criticism of WL concerning investment only works if the investment is comprised of new capital purchases only. It does not work if any of the investment is composed of inventory build up.

    Why is this? Because the context here is aggregate spending.

    If consumers smooth their consumption in response to a tax hike, such that G rises by more than C falls, which then is supposed to mean (and this is the controversial assumption) that AD rises as well, then it is no argument against this to say that it ignores investment of the inventory build up type. This is because such inventory build up doesn’t have any additional spending associated with it. It is building up because there is a LACK of spending associated with it, namely, the lack of spending from the consumers.

    Thus even if WL is ignoring investment of the inventory build up type, then including it cannot contribute to AD.

    Only if the investment that is ignored is comprised of new capital purchases only, will the “don’t forget that investment can fall” argument be considered a valid counter-argument to WL. This is because if investors spend less on new capital goods, then that will be able to be used in conjunction with C to say “The effect of G on AD has been nullified by the drop in C and I (of the capital purchase type).”

    To me, this quote above is a rather complicated way of saying that when demand is low and capacity utilisation is weak, spending in new capital goods (which is the only activity I would recommend using the term investment for, it would spare us endless confusion) by the private sectors is not very likely. That observation of course is as spot on as it was when Keynes made it 80 years ago (German economists had this element in their theory of income determination – as opposed to the neoclassical endowment allocation theory that – in the early 1920s), and I would like to hear Sumner’s response to it.

    Here in Germany, the two stimulus measures that worked best during the 2009 crisis were the 6 bn spent on the cash-for-clunckers scheme, which allowed auto-makers to clear their inventories and hold on to their workforce despite the drastic drop in demand, as well as the other 6 bn the State channeled into the short-time working scheme, which served the same purpose and statistically maintained over 300.000 jobs that year. Both measures stabilised incomes that would have dropped significantly if the people had been fired and become dependent on unemployment benefits, which would have further lowered AD and ultimately GDP and made S equal I at a much lower level.

  97. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 09:10

    D R:

    “What WL was talking about was this: Assuming consumption smoothing, the effect of a tax hike is that consumption spending does not fall by as much as the tax hike and subsequent government spending, hence AD must rise.”

    “No mention of investment.”

    Awww, you’re confused again.

    You first told me:

    “*You* didn’t say anything about investment.”

    I responded to that and said yes I did talk about investment, as it is precisely investment that made me disagree with WL.

    To disprove my response, you then rhetorically asked “shall I quote?”, and so, naturally, I was expecting something that I said that you are going to dig up. But what do you do? You quote WL, not me.

    I hope you realize how I suspect you to be in a state of confusion.

    “It is not enough to disprove Cochrane by assuming away invesment. Cochrane didn’t. You can’t refute someone’s conclusion by merely assuming away the premises of their conclusion.”

    “Cochrane DID NOT assume investment. Cochrane did not assume anything at all, except for his conclusion.”

    False. Cochrane DID assume investment.

    http://faculty.chicagobooth.edu/john.cochrane/research/papers/stimulus_rip.html

    Please read.

    “He made no argument on this point at all except to assert that he was correct.”

    Haha, that’s just a silly caricature of what he said.

    “Tax-financed infrastructure was not a case he covered.”

    False. He specifically used the example of a tax financed project. He wrote: “Spending supported by taxes pretty obviously won’t work:” and then he started to give his explanation, which WL then jumped on.

    Can’t you keep up?

    “So if you want to go that route, I can say Cochrane is wrong because tax-financed infrastructure exists and I therefore dispute his “assumption” that is does not.”

    Cochrane didn’t deny that tax financed infrastructure “exists.”

    “His only defense would be to deny the possibility of such a project.”

    Wrong, he could defend against that baseless accusation by just tell you to read his article.

  98. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 09:13

    D R:

    “we should instead be considering all possible counterfactuals that would have otherwise taken place had the government not taxed and spent that 100.”

    “OK. Here is mine”

    “C+10 I+50 G+100 T+100”

    No, I meant a counterfactual to the tax based government spending scenario, not a counterfactual to no taxes and no spending.

  99. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 09:41

    Eclair:

    Major_Freedom: “If you tax me $100, then the correct way to do “comparative statics”, IMO, is to ask what I would have done with that money had you not taken it from me.”

    “That is true if you believe that government needs to tax people before it can spend. I believe this is not the way the current US system operates..?”

    Uh oh, MMTer alert.

    Actually, what I said does not depend on assuming that the government has to spend before it taxes. It just requires that the taxpayers have the money at time = 0, and then the government taxes them, thus leaving them with 100 less at time = 1.

    It makes no statement about the ultimate source of where that 100 “originally came from.”

    We are comparing the effects of tax and spend $100 versus no taxing and no spending of that $100. We are not considering the effects of the government not spending any money at all ever, in which case at some point of continuous taxation, there will no money left in the market to tax.

    “To me, this quote above is a rather complicated way of saying that when demand is low and capacity utilisation is weak, spending in new capital goods (which is the only activity I would recommend using the term investment for, it would spare us endless confusion) by the private sectors is not very likely.”

    Actually I am saying something a little different.

    But when you consider the context, which is a TAX financed spending program, then the government is taxing money that is in circulation. Is it really so hard to believe that because people have proven themselves able and willing to exchange that $100 in the actual world, our world, that they wouldn’t have done so in a counterfactual world where the only difference is that the government didn’t tax people of $100 at time = 1?

    The only way I can see the counterfactual world you are proposing transpiring would be if people would have otherwise kept the $100 as cash, but then because of the presence of the tax hike only, people then put that $100 into circulation where it can be taxed.

    That is possible, but I just don’t see that being at all likely. Maybe a million to one shot. I mean, when was the last time you responded to a tax hike by saying “Gosh, this means I should draw down my cash balance, and exchange more money on something, anything, so that the government can actually collect what it plans to get in taxes.”

    “That observation of course is as spot on as it was when Keynes made it 80 years ago (German economists had this element in their theory of income determination – as opposed to the neoclassical endowment allocation theory that – in the early 1920s), and I would like to hear Sumner’s response to it.

    I think I know what’s going on. You know about this portion of Keynes’ writings, and you saw some resemblance to that and what I said, so that you took advantage of that opportunity to write all that down. That’s cool and everything, but it doesn’t apply to what I am saying.

    “Here in Germany, the two stimulus measures that worked best during the 2009 crisis were the 6 bn spent on the cash-for-clunckers scheme, which allowed auto-makers to clear their inventories and hold on to their workforce despite the drastic drop in demand, as well as the other 6 bn the State channeled into the short-time working scheme, which served the same purpose and statistically maintained over 300.000 jobs that year. Both measures stabilised incomes that would have dropped significantly if the people had been fired and become dependent on unemployment benefits, which would have further lowered AD and ultimately GDP and made S equal I at a much lower level.”

    Ah, notice how you said “Both measures stabilised incomes that would have dropped significantly”? You’re talking about counterfactuals. While you are most likely correct about AD falling, it is not *necessarily* true that a lower AD will lead to a lower output. But even if a lower AD and lower output accompanied each other, that STILL wouldn’t mean it is “bad”. For consider. You said that there was a drastic drop in demand for output (cars). In a world where the market caters to individual consumer preference, then it would have been “good” if car production declined, and labor and capital released from the car making industry.

    Given that consumers didn’t collapse their demand for goods as such to zero, it means that the labor and capital released from the car industry could have been profitably allocated into other industries, provided of course that the prices fall.

    Now, you might find this solution “harsh”, but consumers ARE “harsh”, and capitalism is geared towards satisfying the harsh consumer.

    But there is nuance here, because the demand for cars collapsing in 2009 is due to the fragile nature of Germany’s, as well as the world’s, banking system. Our monetary system is built on a house of cards where any wind can knock the whole thing over. So I can appreciate it when people reject the free market solution as too harsh for something that is supposed to be in passing only, because AD “shouldn’t” have fallen like it did (if only the central bank were omniscient Gods, etc, etc).

    But the dramatic aggregate demand collapses we witness cannot be solved by more of what caused the fragility in the first place, which is artificial monetary stimulus and low interest rates.

    It’s tragic that the monetary system has become so unsound, so brittle, that it is almost as if the free market seems like a destroyer of economies, instead of their cure. But whatever, monetarism is going to collapse, no matter what the “target” happens to be, because the problem is the central planning of money and interest rates in the first place.

  100. Gravatar of D R D R
    19. January 2012 at 09:59

    “No, I meant a counterfactual to the tax based government spending scenario, not a counterfactual to no taxes and no spending.”

    [facepalm]

    C-10 I-50 G-100 T-100″³

  101. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 10:27

    D R:

    “C-10 I-50 G-100 T-100″³

    [Double facepalm]

    I don’t think you know how to do this.

    You should at least have G+0 and T+0, then allocate a 100 quantity of money to C(consumption) + I(investment) + S(cash).

  102. Gravatar of John Schultz John Schultz
    19. January 2012 at 11:02

    MF: “We are comparing the effects of tax and spend $100 versus no taxing and no spending of that $100.”

    Ok … but isn’t that exactly what economists are trying to do when they say things like: “Assume (my model of) our economy is in static equilibrium before time 0. At time 0, increase govt. spending by $100 and increase govt. taxation by $100. (My model of) Our economy will evolve over time like and eventually reach a new static equilibrium of ?”

    Then it comes down to a question of how realistic (empirically based) their model and assumptions are as to whether you should trust conclusions drawn from it or not.

    I’m not really sure to what you are objecting nor what you offering instead of this?

  103. Gravatar of John Schultz John Schultz
    19. January 2012 at 11:05

    Woops, I accidentally used reserved html characters in the above post. I wasn’t trying to link or shout. Here it is again:

    MF: “We are comparing the effects of tax and spend $100 versus no taxing and no spending of that $100.”

    Ok … but isn’t that exactly what economists are trying to do when they say things like: “Assume (my model of) our economy is in static equilibrium before time 0. At time 0, increase govt. spending by $100 and increase govt. taxation by $100. (My model of) Our economy will evolve over time like [A] and eventually reach a new static equilibrium of [B]?”

    Then it comes down to a question of how realistic (empirically based) their model and assumptions are as to whether you should trust conclusions drawn from it or not.

    I’m not really sure to what you are objecting nor what you offering instead of this?

  104. Gravatar of D R D R
    19. January 2012 at 11:40

    “I don’t think you know how to do this.”

    Do what? Again, you quote

    “”we should instead be considering all possible counterfactuals that would have otherwise taken place had the government not taxed and spent that 100.”

    You are asking me to consider what would have taken place “had the “government not taxed” (T-100) “and spent that 100” (G-100)

    Now you want me to…. consider increasing the money supply by 100?

  105. Gravatar of D R D R
    19. January 2012 at 11:43

    … I’m not even going to ask if I can allocate a different amount to S than I allocate to I.

  106. Gravatar of D R D R
    19. January 2012 at 11:47

    As far as I can tell, you’re asking me what would happen if GDP rose by 100. I’m going to guess that the multiplier on GDP is 1.0.

  107. Gravatar of ssumner ssumner
    19. January 2012 at 12:59

    DR, Your step five shows WL is wrong. He claimed the rise in AD was the net change in C+G, but you show it’s not. The fact that the multiplier is still positive shows Cochrane was wrong, but it’s doesn’t show that my criticism of WL’s argument was wrong.

    Tom, That’s a meaningless question, because I have no idea what would happen in period 2. Unlike Keynesians I don’t assume that “spending” drives AD, I think monetary policy does. The point is that the argument was wrong as stated, and if he’d ignored I in period 2, it would have been wrong then too.

    Rob, False, if G rises by 100 and C falls by 30 and inventories fall by 70 then no change in AD.

    Major freedom. Wrong–all economists consider inventory increases to be part of AD.

    Mike Sax, No I’m not trying to defend Cochrane and Lucas, I’m saying they are making the same sort of mistakes and K and WL. It’s “moral equivalence.”

    Math, Keynesian textbooks say S=I is an identity, That means it’s true by defintion, always and everywhere.

    Thanks John, That is a typo. For some reason I make that sort of typo often. I’ll change it soon.

    And thanks for supporting me. After all this effort it’s gratifying to see people agreeing with me. I should have made this stripped down argument right up front.

    JamesH, You are mixing stocks and flows. In econ savings is a stock, but saving is a flow, and equals the flow of new investment. The stock value of existing capital is another issue.

    bob, that’s right

    John and Kevin, It’s a no no to do saving w/o investment. Does Woodford’s paper have saving?

    And if they are saving cash, where the hell does the cash come from? Other people, then the aggregate S doesn’t rise. The Fed? Then it’s an expansionary monetary policy–a no no when trying to prove fiscal stimulus works.

    Eclair, You are wrong about inventories, see my earlier replies in this post.

    DR, If there is no I, then there cannot be consumption smoothing. So what the heck is all this about?

    Everyone, This is what WL really meant:

    1. An 100 tax-financed increase in G would raise Y by almost as much as G rose, but not quite. Say 85. So C+I would fall by 15. Since there is consumption smoothing most of the fall is less saving and investment, say a drop of 10, and C falls by 5.

    But he forgot investment.

    By the way you can’t draw conclusions about year 2 unless you know whether the G stimulus is repeated, or ended.

  108. Gravatar of rob rob
    19. January 2012 at 13:11

    “Rob, False, if G rises by 100 and C falls by 30 and inventories fall by 70 then no change in AD.”

    But why does a fall in inventories imply a fall in AD ? Surely the only reason that inventories have fallen is because people have increased AD to buy them ?

  109. Gravatar of rob rob
    19. January 2012 at 13:18

    Re: Inventory levels – not sure how we can rely on this but for what it is worth here is what Wikipedia says on inventories

    “Thus, only the planned or intended or desired part of investment (Ip) is counted as part of aggregate demand. (So, I does not include the ‘investment’ in running up or depleting inventory levels.)”

  110. Gravatar of tom tom
    19. January 2012 at 13:21

    Scott: “I have no idea what would happen in period 2” but you wanted comparative statics so I gave an example. We can merge these two periods into one when simultaneously these events takes place and we achieve the same result: positive multiplier.

  111. Gravatar of tom tom
    19. January 2012 at 13:26

    rob “But why does a fall in inventories imply a fall in AD ” actually Scott is correct BUT we have to remember that because of decreased inventories firms will replenish them and then Investment will rise and Y will rise so multiplier is positive. Here is formal proof (which I posted on Karl Smith’s site):
    1) dY=dC+dS+dT and 2) dY=dC+dI+dInventories +dG
    If dG=dT=2 and dC=dS=dInventories=-1 and dI=0 then:
    in the first period dY=0 BUT in the second period dI=1 (firms want to replenish inventories) and from 2) dY=1 and this implies from 1) that dS=1.
    So multiplier is positive

  112. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 13:27

    John Schultz:

    “MF: “We are comparing the effects of tax and spend $100 versus no taxing and no spending of that $100.””

    “Ok … but isn’t that exactly what economists are trying to do when they say things like: “Assume (my model of) our economy is in static equilibrium before time 0. At time 0, increase govt. spending by $100 and increase govt. taxation by $100. (My model of) Our economy will evolve over time like [A] and eventually reach a new static equilibrium of [B]?””

    No. That is temporal. I am talking counterfactuals. So I would rewrite that to say:

    “Assume (my model of) our economy is in static equilibrium before time 0. That economy will evolve over time like [X] and eventually reach a new static equilibrium of [Y]. Now assume an identical economy with the following changes: At time 0, increase govt. spending by $100 and increase govt. taxation by $100. (My model of) Our economy will evolve over time like [A] and eventually reach a new static equilibrium of [B]?”

    One of these economies cannot occur because choices have to be made, and making a choice rules out all other possible alternatives.

    “Then it comes down to a question of how realistic (empirically based) their model and assumptions are as to whether you should trust conclusions drawn from it or not.”

    What about all the possible worlds that could never exist because the required choices to make them a reality were not made?

    “I’m not really sure to what you are objecting nor what you offering instead of this?”

    This is what I think is the incorrect way to go about it:

    World 1, time = 0: G+0, T+0, C+250, I+850, S+100, Y+1100

    World 1, time = 1: G+100, T+100, C+200, I+800, S+100, Y+1100

    This is what I think is the correct way to go about it:

    World 1, time = 0 (factual): G+0, T+0, C+250, I+850, S+100, Y+1100

    World 2, time = 0 (possible counter-factual): G+100, T+100, C+200, I+800, S+100, Y+1100

  113. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 13:33

    ssumner:

    “Major freedom. Wrong-all economists consider inventory increases to be part of AD.”

    Woah, I am not a PhD (yet), but that can’t be right.

    Inventory increases are a part of GDP, not AD.

    AD is the sum total of all DEMANDS, in money. Inventory increasing is not an increase in anyone putting forth additional demand for goods. It is just a physical increasing of inventory. If it’s brought about by a fall in consumer spending and accumulating of cash balances, then AD will fall, not rise.

    Or did I miss something?

  114. Gravatar of math math
    19. January 2012 at 13:35

    Sumner: “Math, Keynesian textbooks say S=I is an identity, That means it’s true by definition, always and everywhere.”

    I know. I think that Keynesians are in schizophrenic position.

    On one hand, they want to have rigorous models as Monetarists do. So, they assume the identity S=I in their textbooks.

    On the other hand, they want to be closer to the market reality. So, they relax that condition when they are trying to analyze the real economy.

    So, they have two models in mind — one model is for the formal presentations (with the identity) and another one is for the economic analysis (where that condition is relaxed).

    Remember what Krugman said? “If your verbal reasoning led you to think that expansionary fiscal policy can’t be expansionary as a matter of logic, well, your logic was wrong.” (http://krugman.blogs.nytimes.com/2012/01/17/when-some-rigor-helps-mildly-wonkish/) So, he is talking that he (as a Keynesian guy) assumes here the priority of economic reality over the economic models.

  115. Gravatar of D R D R
    19. January 2012 at 13:38

    “DR, If there is no I, then there cannot be consumption smoothing. So what the heck is all this about?”

    Are you sure you don’t want to think a bit more carefully about that?

    Let C “rise” by dC. G+100, T+100, Y rises by 100+dC and disposable income rises by dC. Thus, consumption smoothing *implies* dC=0, and I’m pretty certain that 0 is a fair bit less than 100.

    Even then, that only means that the effects of individual smoothers cancel each other out. (Unless you *assume* that the tax on every agent was equal, that they smooth by the same factors, that they each profited equally from the sale of the bridge….)

  116. Gravatar of D R D R
    19. January 2012 at 13:41

    And I forgot to mention in case that wasn’t clear… There is no I, but that simply means the folks with rising Yd lend to the folks with sinking Yd. They can still smooth without inducing any investment.

  117. Gravatar of rob rob
    19. January 2012 at 13:44

    Thanks Tom,

    I see that if you define inventories as part of AD then what Scott says is correct. I just don’t understand the logic in why inventories are part of AD and was asking help in understanding that.

    (BTW I think the issue under discussion is current and not later periods.)

  118. Gravatar of D R D R
    19. January 2012 at 13:46

    “AD is the sum total of all DEMANDS, in money. Inventory increasing is not an increase in anyone putting forth additional demand for goods. It is just a physical increasing of inventory. If it’s brought about by a fall in consumer spending and accumulating of cash balances, then AD will fall, not rise.”

    Oh, burn. Especially considering the source.

    That is another point largely missed. Of course, the inventory changes still must be considered in order to go from AD to Y and compute the multiplier.

    Still…. burn.

  119. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 13:53

    D R:

    “we should instead be considering all possible counterfactuals that would have otherwise taken place had the government not taxed and spent that 100.”

    “You are asking me to consider what would have taken place “had the “government not taxed” (T-100) “and spent that 100″³ (G-100)”

    No, no tax and no spending means we have T+0 and G+0. T-100 to me sounds like a tax credit, and G-100 to me sounds like a reduction of spending from a particular lower level prior in time.

    But I am not talking temporally, I am talking cross sectionally.

    “Now you want me to…. consider increasing the money supply by 100?”

    No, I mean if in the real world the government taxes and spends 100, then in another counterfactual, possible world, the private sector would have had that 100 to spend. So when I say “allocate” the 100, I don’t mean create it out of nothing, or printed, I just mean select one possible world where the private sector spends the 100.

    The reason I am going through all this is that I think its important to understand that when the government taxes and spends 100, that is a product of people making choices. Fiscal stimulus is a choice. If that choice is not made, and we consider what if the choice is made not to tax and spend 100, what might have otherwise taken place?

    Going back to my original scenarios:

    World 1, time = 0 (factual): G+0, T+0, C+250, I+850, S+100, Y+1100

    World 2, time = 0 (possible counter-factual): G+100, T+100, C+200, I+800, S+100, Y+1100

    I have set this up so that a possible counterfactual world in the presence of tax financed spending has the following characteristics:

    Total spending is the same. Most Keynesians would believe the tax and spend economy is no worse off, because Y remained at 1100.

    But if we realize that we have I+850 in the no tax and spend world, and I+800 in the tax and spend world, then we can conclude that the no tax and spend world will be less productive in real terms compared to the tax and spend world. This is despite Y remaining the same. The reason it is more productive is because there is more investment spending relative to aggregate spending. With more capital in demand, more capital will be produced. With more capital, more consumer goods can be produced at lower prices. This is again despite Y remaining unchanged.

    All I really want to advance here is the argument that choosing tax financed government spending programs can potentially reduce economic productivity and hence standards of living, compared to choosing no tax and no spending instead.

    Since we can never see these counterfactuals, it means that nobody can claim to know that government stimulus improves economies by observing economies over time. The debate on economics must be in the intellectual logical sphere, not the empirical sphere. Looking at economies over time cannot enable us to know if what we are doing is hurting us or helping us, EVEN IF we observe standard of living metrics nominally rising over time. For it is possible that without the stimulus, those metrics could have risen EVEN MORE. But we can’t see it because we can only see the reduced standard of living growth in reality.

  120. Gravatar of tom tom
    19. January 2012 at 13:54

    rob
    “I think the issue under discussion is current and not later periods” .
    Actually it doesnt matter. For comparative statics we can separete this into two periods: one when invenories are decreased and second when they are replenished. But the end result is always the same i.e. Y increases because of increase in I caused by fiscal stimulus (increase in G). You can look at it separately or simultaneously when these events takes place but it doesnt change anything. That’s why Krugman schooled Scott about comparative statics.

  121. Gravatar of tom tom
    19. January 2012 at 13:58

    rob
    AD=Y=C+G+I+Inventories
    So if demand for inventories fall than agregate demand fall
    (Usually inventories are treated as part of I)

  122. Gravatar of D R D R
    19. January 2012 at 14:02

    Major,

    I’m sorry it wasn’t clear. I wan’t trying to suggest these were LEVELS, but CHANGES. You know, from whatever assumed baseline.

    G+100: Government spending rises by 100
    G-100: Government spending falls by 100

    Get it?

  123. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 14:02

    Dang, meant to say:

    “But if we realize that we have I+850 in the no tax and spend world, and I+800 in the tax and spend world, then we can conclude that the no tax and spend world will be more productive in real terms compared to the tax and spend world. This is despite Y remaining the same. The reason it is more productive is because there is more investment spending relative to aggregate spending. With more capital in demand, more capital will be produced. With more capital, more consumer goods can be produced at lower prices. This is again despite Y remaining unchanged.”

  124. Gravatar of rob rob
    19. January 2012 at 14:03

    I’m looking this from a purely common sense point of view.

    I’ve been hoarding cash for a while, I finally decide to spend it. Everyone’s else spending stays the same. I buy a new car and the garage had not been expecting to sell it so there inventory goes 1 below their planned level.

    Total expenditure has gone up by the value of 1 new car. Why is this not an increase in AD? (And why do we assume that the new car came from inventories, rather than the same number of cars being sold at higher prices ?)

  125. Gravatar of D R D R
    19. January 2012 at 14:11

    “Total spending is the same. Most Keynesians would believe the tax and spend economy is no worse off, because Y remained at 1100.”

    Um… I’m not sure how Keynesians fit in here.

    Are you saying “Keynesians would accept the premise that Y remained at 1100, and this would cause them to believe the tax and spend economy is no worse off.”

    Or are you saying “If Y remained at 1100, this would cause Keynesians to believe that the tax and spend economy is no worse off.”

    English is such a wonderfully ambiguous language.

  126. Gravatar of D R D R
    19. January 2012 at 14:14

    rob,

    “I’ve been hoarding cash for a while, I finally decide to spend it. Everyone’s else spending stays the same. I buy a new car and the garage had not been expecting to sell it so there inventory goes 1 below their planned level.

    “Total expenditure has gone up by the value of 1 new car. Why is this not an increase in AD? (And why do we assume that the new car came from inventories, rather than the same number of cars being sold at higher prices ?)”

    I think that is an increase in AD, but not Y. The immediate impact of your purchase is to reduce investment by exactly the amount of your added consumption.

  127. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 14:15

    D R:

    “I’m sorry it wasn’t clear. I wan’t trying to suggest these were LEVELS, but CHANGES. You know, from whatever assumed baseline.”

    Bingo. And I am proposing that the only rational baseline is one that is founded upon counterfactual reasoning, not a prior period in time.

    If we only consider the economy changing over time, then we only going to over observe the product of our choices and actions in the past, and we will never consider whether we could have done the entire thing differently.

    I think it is flawed reasoning to believe something like “The government engaged in fiscal stimulus at time = 0, and by time = 1, real GDP grew by 3% and unemployment fell to 5%. Ergo, fiscal stimulus works.”

    I think it is flawed because it doesn’t consider the possibility that real GDP could have otherwise have grown by 4% and unemployment could have otherwise have fallen to 3%, if fiscal stimulus were NOT utilized.

    I hold that we can never gain knowledge of correct theoretical reasoning in economics by observing past data and then trying to see rates of change of various statistics and whatnot.

    Let me give you an EXTREME example to show you what I mean. Suppose I kidnapped you and locked you into a room. Every day thereafter, I gradually increase the rate of food that I give you, the rate of sunshine I give you, and the rate of water I give to you.

    Now, if we ONLY go by past experience and past data, the sloppy minded person who only thinks in terms of what he sees, may conclude that kidnapping can bring about growing prosperity. For I kidnapped you at time = 0, and then we see all these economic indicators growing over time, to time = 1.

    We should of course say wait a minute, this can’t be right. Kidnapping makes people’s lives worse off.

    But then worse off compared to what? It can’t be the past, because compared to the past they are better off. It has to be compared to something else. Well, I say that something else should be reasoning in terms of counterfactuals, based on some universal standard of validating economic principles and theories. Again, it can’t be past history because past history is telling us kidnapping leads to growing prosperity, which through our intuition we know to be false.

    Is this making any sense or am I talking like a space cadet?

  128. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 14:17

    D R:

    “Are you saying “Keynesians would accept the premise that Y remained at 1100, and this would cause them to believe the tax and spend economy is no worse off.””

    “Or are you saying “If Y remained at 1100, this would cause Keynesians to believe that the tax and spend economy is no worse off.””

    The former.

  129. Gravatar of D R D R
    19. January 2012 at 14:19

    rob,

    To answer the second question, it’s because you don’t buy the car right off the line, and the car was counted toward GDP at that point as investment. Your purchase may induce a nearly-instantaneous rebuilding of inventories, but your purchase is likely out of investment.

    If you get a massage, on the other hand… I think that’s an increase in consumption services.

  130. Gravatar of tom tom
    19. January 2012 at 14:24

    rob,
    In this case dS=-1 dInventories=-1 dC=1 dG=0 dI=0
    So in total: Y=C+G+I+Inventories= 1+0+0-1=0, AD doesnt change.

    “why do we assume that the new car came from inventories, rather than the same number of cars being sold at higher prices ?”
    Because of output gap: first rises empoyment and production and prices rise later when we approach full employment.

  131. Gravatar of D R D R
    19. January 2012 at 14:25

    “The former.”

    Then I believe you are wrong. I am pretty certain that Keynesians believe that G+100 T+100 raises Y unless you start assuming full employment. Or, you know, that someone will tighten monetary policy to prevent Y from rising.

    I also do not understand how S fails to equal I.

  132. Gravatar of D R D R
    19. January 2012 at 14:30

    “Bingo. And I am proposing that the only rational baseline is one that is founded upon counterfactual reasoning, not a prior period in time.”

    I think you’ve missed the point of this entire discussion. I’m pretty sure everyone has been arguing relative to a baseline.

  133. Gravatar of tom tom
    19. January 2012 at 14:30

    Rob,
    In the second step (or you can analyze it simultaneously) because of lower inventories firms want to invest to replenish them so Y can rise and AD also.

  134. Gravatar of rob rob
    19. January 2012 at 14:34

    Thanks. And actually that leads me to see the reason why expenditure on inventory probably doesn’t count towards AD.

    expenditure = income

    so if a company pays wages to workers to produce a car that goes into inventory then you have to count the value of the car as income in that period to make the books balance. And it would be double counting to count it again when it actually gets sold.

    If that is what this whole debate hangs upon then I am a little underwhelmed

  135. Gravatar of On civility, and the purpose of the blogosphere « Shewing the fly On civility, and the purpose of the blogosphere « Shewing the fly
    19. January 2012 at 14:41

    […] Was Krugman a fool and a knave before he discovered what he thought were good and rigorous arguments for something he previously believed was a myth? Of course not. So just don’t be surprised when we demand a very high standard of argument to be persuaded, rather than having our intelligence continually insulted. The condescending tone of his awful post on comparative statics, in the context of what was ostensibly a “debate” with Scott Sumner was staggering to anyone who had actually, you know, been reading what Scott was writing. […]

  136. Gravatar of D R D R
    19. January 2012 at 14:42

    rob,

    Yes, you seem to have it.

    As for the debate, well, I haven’t quite followed why that would be the critical piece of the puzzle.

  137. Gravatar of tom tom
    19. January 2012 at 14:42

    Rob,
    “I am a little underwhelmed”, no read my last post. If you buy inventories firms have to produce to replenish them so this increases Y and AD. You act like fiscal stimulus:)

  138. Gravatar of D R D R
    19. January 2012 at 14:47

    It is also one thing to watch out for in watching the economy. If GDP rises because inventories are up, that might be because of an increase in anticipated sales, and it might be because of currently depressed sales. That kind of thing.

    Which I guess is the flip side of this debate… I mean, the real one, not the one which interests Sumner.

  139. Gravatar of D R D R
    19. January 2012 at 14:53

    tom,

    “In this case dS=-1 dInventories=-1 dC=1 dG=0 dI=0
    So in total: Y=C+G+I+Inventories= 1+0+0-1=0, AD doesnt change.”

    Because you are adding I to inventories in calculating Y, you must by “I” mean “Ip” (planned investment)

    So you have Y rising by dC-dInventories=0, but AD = C+Ip+G rises by 1.

    I am confused.

  140. Gravatar of tom tom
    19. January 2012 at 15:00

    DR

    AD=Y=C+G+I (with inventories)
    http://en.wikipedia.org/wiki/Gross_domestic_product
    http://en.wikipedia.org/wiki/Aggregate_demand
    http://en.wikipedia.org/wiki/Savings_identity

  141. Gravatar of tom tom
    19. January 2012 at 15:03

    DR

    AD=Y=C+G+I with inventories
    Appropiate links are awaiting moderation

  142. Gravatar of D R D R
    19. January 2012 at 15:03

    tom,

    continuing… yet I think I agree with your argument, if not your math. If I may…

    If you buy out of inventories, the immediate effect is
    dC=100, dIp=0, dInv=100, dG=0 so dY=dC+dIp+dInv+dG=0 and output is unchanged. However, dAD=dC+dIp+dG=100, so AD does rise.

    In response to this increase in AD, producers may respond by hiring/extending hours/increasing beatings/raising pay/whatever to increase production.

  143. Gravatar of tom tom
    19. January 2012 at 15:07

    DR
    Y=AD
    check wikipedia at aggregate demand and gross domestic product (or any textbook)

  144. Gravatar of D R D R
    19. January 2012 at 15:09

    tom,

    I am pretty certain the flaw there is AD=Y. I mean, it depends on who you ask, right? Lots of Keynesians do not count unexpected changes in inventory as AD. You know, because if I cut back on consumption and your inventories rise as a result, it makes sense to say that demand for stuff fell even though output did not change.

    Often, they will be used interchangeably, but it can clarify thought to make a distinction.

  145. Gravatar of D R D R
    19. January 2012 at 15:11

    tom,

    via your chosen reference,

    “This is the demand for the gross domestic product of a country when inventory levels are static. It is often called effective demand, though at other times this term is distinguished.”

  146. Gravatar of John Schultz John Schultz
    19. January 2012 at 15:15

    Sorry, I just found this amusing:

    MF: “Since we can never see these counterfactuals, it means that nobody can claim to know [any macroeconomic proposition] by observing economies over time. The debate on economics must be in the intellectual logical sphere, not the empirical sphere.”

    That is one reason why economics is not science, why there are so many models and why there is no much disagreement over even the most basic propositions.

    Economics (at least macro) is (virtually) unfalsifiable. That’s why it is called “the dismal science.”

    Maybe once we can model and simulate humans well enough within computers we can eventually run controlled experiments that will allow for repeated, controlled experiments. Even then, there will likely be endless disagreement on the models of humans and their environment.

  147. Gravatar of tom tom
    19. January 2012 at 15:17

    DR
    ok you are right but it doesnt change main conclusion: that fiscal stimulus is effective if there is significant output gap. We can include or exlcude inventories from AD. I will correct my example

  148. Gravatar of D R D R
    19. January 2012 at 15:20

    tom,

    Thanks. As I said, I agreed with your conclusion. Did my recap accurately reflect your view?

  149. Gravatar of D R D R
    19. January 2012 at 15:22

    John,

    Not to mention humans changing behavior in response to… Er… Maybe I shouldn’t go there.

  150. Gravatar of tom tom
    19. January 2012 at 15:27

    In this case dS=-1 dInventories=-1 dC=1 dG=0 dI=0
    So in total: Y=C+G+I+Inventories= 1+0+0-1=0, Y doesnt change
    AD= C+G+I= 1+0+0 =1 so it increases. Because AD>Y firms invest to replenish inventories so Y increases by dInventories=1 and AD doesnt change. So in the end AD=Y.

  151. Gravatar of D R D R
    19. January 2012 at 15:31

    tom,

    That sounds correct to me. I mean, firms don’t have to react to the increase in AD that way, and they may be unable to even if they wish to react that way, but it is reasonable to assert that they try.

  152. Gravatar of tom tom
    19. January 2012 at 15:33

    DR
    I agree. If you exclude inventories from AD it is even easier to follow Keynesian logic because after fiscal stimulus you result with Y 0. Just like in my example. Thanks

  153. Gravatar of tom tom
    19. January 2012 at 15:35

    DR
    “firms don’t have to react to the increase in AD that way”
    they should at ZLB

  154. Gravatar of tom tom
    19. January 2012 at 15:39

    DR correction
    I agree. If you exclude inventories from AD it is even easier to follow Keynesian logic because after fiscal stimulus you result with Y significantly less then AD. Just like in my example. Thanks

  155. Gravatar of John Schultz John Schultz
    19. January 2012 at 15:41

    “Is this making any sense or am I talking like a space cadet?”

    You are saying that observation can tell us next to nothing about macroeconomics and/or that it can lead to very wrong conclusions.

    Still, I think what you are after is exactly what economists are doing when they are playing with themselves … I mean their models of the economy. When an economist says something like “Assume an economy modeled like [A] with current state [B] …” he has established one “timeline” that the economy will follow.

    For example, he might say “Assume an economy described by … that is in static equilibrium.” If no shock is introduced to this economy, then it will continue from now to eternity doing as it has done by assumption.

    If our exemplar economist then introduces a shock at time t, shows how it reacts over time and ultimately establishes a new equilibrium (that by definition lasts forever). Then he has demonstrated another “timeline” for his economy that stems off from his first economy at time t.

    The economist has described two potential universes for the economy. He can then argue that one is better than the other for whatever reason he prefers (often higher GDP).

    Now what you might really be after is for our plucky economist to not just examine one particular shock picked out of the blue, but instead examine all the potential shocks that could be caused by conscious decision and maximize to find “the best” resulting universe. Then he could recommend those particular decisions / actions be implemented.

    Is that what you’re after?

  156. Gravatar of D R D R
    19. January 2012 at 15:49

    tom,

    I in no way whatsoever meant to imply that I thought they would not succeed in the current economic climate. There is a lot of slack in the economy, and I really doubt that Bernanke would tighten in response. Sumner notwithstanding, my read is that Bernanke is at the very least desirous of fiscal stimulus.

    Then again, maybe he just wants enough additional stimulus to allow monetary tightening. While this would do nothing for the economy, it would help get the inflation hawks off his back.

    Who knows. Anyway. Well done.

  157. Gravatar of Kevin Donoghue Kevin Donoghue
    19. January 2012 at 16:00

    Scott: “Does Woodford’s paper have saving?”

    Yes, but no Investment. Equation (1.3): Y = C + G.

    So G = S in equilibrium. This is not unusual in NK models.

  158. Gravatar of tom tom
    19. January 2012 at 16:06

    DR
    We agree. Here are my two proofs that multiplier is greater than 0. From Karl Smit’s site:
    Here is my argument:
    1) dY=dC+dS+dT and 2) dY=dC+dI+dInventories +dG
    If dG=dT=2 and dC=dS=dInventories=-1 and dI=0 then:
    in the first period dY=0 BUT in the second period dI=1 (firms want to replenish inventories) and from 2) dY=1 and this implies from 1) that dS=1.
    So multiplier is positive

    Here is another version of my proof:
    Y= C+G+I+Inventories AD=C+G+I
    If dG=dT=2 and dC=dS=dInventories=-1 and dI=0 then:
    in the first period dY=0 but dAD=1. Because dADis greater then dY firms invest to replenish inventories so Y increases by dInventories=1 and AD doesnt change. So in the end AD=Y at the higher level with positive multiplier

  159. Gravatar of tom tom
    19. January 2012 at 16:09

    Kevin
    It is not necessary to consider S to achieve conclusion that multiplier is positive. Read my previous post.

  160. Gravatar of D R D R
    19. January 2012 at 16:42

    tom,

    I’m not sure I understand your argument.

    What are the periods? Are you getting Major all confused by talking about a dynamic multiplier, or are the periods more abstract?

    Are you simply trying to prove a positive multiplier, or is this tied in with Wren-Lewis/Sumner?

  161. Gravatar of John Schultz John Schultz
    19. January 2012 at 16:58

    Tom, in your first proof I think your problematic assumption is that you assume the economy is not “happy” with a different level of inventories.

    In your second proof, you’ve redefined AD to make your argument, which is cheating.

  162. Gravatar of tom tom
    19. January 2012 at 17:09

    DR
    You can call it steps not periods. It just shows how economy achieves higher levet of equilibrium after initial fiscal stimulus. Actually we cant separate one period of decreasing inventories with another when firms replenish them because firms are doing it all the time. They dont wait until stocks are empty to invest. So we can merge these steps into one period and result will be the same.

  163. Gravatar of tom tom
    19. January 2012 at 17:16

    John
    “in your first proof I think your problematic assumption is that you assume the economy is not “happy” with a different level of inventories.”
    I disagree because we can always argue that fiscal stimulus should be sufficiently large to substantially deplete inventories so firms are forced to replenish them in order to be able to conduct regular business.

    “In your second proof, you’ve redefined AD to make your argument, which is cheating”
    That is not true read my link to wikipedia about agregate demand (actually it was thanks to DR). AD excludes inventories

  164. Gravatar of tom tom
    19. January 2012 at 17:40

    Ok here are talk only about one period:

    1) dY=dC+dS+dT and 2) dY=dC+dI+dInventories +dG
    If dG=dT=2 and dC=dS=dInventories=-1 and dI=0 then:
    in the first step dY=0 BUT in the next step dI=1 (firms want to replenish inventories) and from 2) dY=1 and this implies from 1) that dS=1.
    So multiplier is positive

    Here is another version of my proof:
    Y= C+G+I+Inventories AD=C+G+I
    If dG=dT=2 and dC=dS=dInventories=-1 and dI=0 then:
    in the first step dY=0 but dAD=1. Because dADis greater then dY firms invest to replenish inventories so Y increases by dInventories=1 and AD doesnt change. So in the end AD=Y at the higher level with positive multiplier

  165. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 17:55

    D R:

    “The former.”

    “Then I believe you are wrong. I am pretty certain that Keynesians believe that G+100 T+100 raises Y unless you start assuming full employment. Or, you know, that someone will tighten monetary policy to prevent Y from rising.”

    That’s not related to the argument I made. I was talking about Y being 1100 for two different possible economies, one with government taxation and spending of 100, the other without taxation and spending, and how I suspect most Keynesians would not consider the taxing and spending economy as worse off.

    I don’t doubt that they believe that an additional tax financed spending spree increases Y. I do however hold that believe to be incorrect.

  166. Gravatar of tom tom
    19. January 2012 at 17:56

    Last remark before I go to sleep (it’s nearly 3 a.m. here):
    in both cases savings identity is maintained so Scott’s critique is not justified.

  167. Gravatar of Major_Freedom Major_Freedom
    19. January 2012 at 18:19

    John Schultz:

    “You are saying that observation can tell us next to nothing about macroeconomics and/or that it can lead to very wrong conclusions.”

    I am saying that observation can tell us absolutely everything about economic history, but absolutely nothing about economic theory.

    If anyone makes any choices whatsoever, be it raising tax financed spending, or lowering corporate tax rates, or raising unemployment insurance, or starting a new business, or spending more on investment, or less on consumption, every single choice that can ever be made, will produce a single history. We will never be able to extract economic theory from history, because there is no way to hold a variable constant and then run history over again to see how other things changed.

    If we observe tax financed spending to be raised, and subsequent to that we observe AD to rise, then we cannot say we acquired knowledge that tax financed spending raises AD. Even if we’re in a liquidity trap, with idle resources, unemployment, etc, this conclusion cannot be made on the basis of historical experience.

    If you think this is too off the cuff to accept, then I’ll just say that even the positivist Keynesians don’t consider their core views as in principle falsifiable. No matter what happens to AD after government stimulus, no matter what is observed, the theory cannot be falsified. If AD falls, then the conclusion is that AD would have fallen further without fiscal stimulus, hence AD was still boosted by fiscal stimulus. If AD remains flat, then the conclusion is that AD would have fallen without fiscal stimulus, hence AD was still boosted by fiscal stimulus. If AD rises, then the conclusion is that AD would have risen by less without fiscal stimulus, hence AD was still boosted by fiscal stimulus.

    “Still, I think what you are after is exactly what economists are doing when they are playing with themselves … I mean their models of the economy. When an economist says something like “Assume an economy modeled like [A] with current state [B] …” he has established one “timeline” that the economy will follow.”

    LOL, with themselves…

    Yes, this is what theoretical economists do, and that’s where, I think, the debates are ultimately settled.

    “If our exemplar economist then introduces a shock at time t, shows how it reacts over time and ultimately establishes a new equilibrium (that by definition lasts forever). Then he has demonstrated another “timeline” for his economy that stems off from his first economy at time t.”

    “The economist has described two potential universes for the economy. He can then argue that one is better than the other for whatever reason he prefers (often higher GDP).”

    “Now what you might really be after is for our plucky economist to not just examine one particular shock picked out of the blue, but instead examine all the potential shocks that could be caused by conscious decision and maximize to find “the best” resulting universe. Then he could recommend those particular decisions / actions be implemented.”

    “Is that what you’re after?”

    Precisely.

  168. Gravatar of John Schultz John Schultz
    19. January 2012 at 18:35

    DR: “Not to mention humans changing behavior in response to… Er… Maybe I shouldn’t go there.”

    Obviously, the psychohistorians will have to keep all their resulting models, predictions, etc., top secret or there is no chance that they could be effectively applied to the real world.

  169. Gravatar of John Schultz John Schultz
    19. January 2012 at 18:57

    MF: “I am saying that observation can tell us absolutely everything about economic history, but absolutely nothing about economic theory.”

    I think that might be overstating it a bit.

    If a (macroeconomic) theory makes a prediction given the current state of the world (to whatever precision the model demands) and the prediction turns out to be wrong, then the theory is wrong or at least incomplete in that instance.

    A (macroeconomic) theory that keeps making predictions that keep coming true in the real world might be deemed very useful even though we have no way of truly knowing if the theory is right or just getting lucky so far.

    But this is a general problem for all science. For example, we don’t truly know that Newton’s “Laws” will continue being an accurate description of the universe in the future. We just know that they have (virtually) never been contradicted since they were first formulated hundreds of years ago.

  170. Gravatar of D R D R
    19. January 2012 at 21:09

    tom,

    Can’t think much of it now, but I think it works out fine with a BBM of 0.5. I think it might help clarify the assumptions behind the argument, but then again, I’m not the one you’re trying to convince.

    Three points…

    1. I don’t know if you meant to smooth consumption– if that’s why in the first step C falls by 1 while Yd=Y-T falls by 2. If so, you might have be a bit more clear about step 2.

    2. It might help if you use “Ip” for planned inventories to distinguish it from investment. Particularly the way these things get conflated, it would help make it clear that you’re not redefining or double-counting.

    3. Even here, the use of AD isn’t technically correct, as AD is actually a demand schedule, rather than a quantity demanded. But then again, you don’t have prices in the model… etc. I’m not saying it’s a bad way of looking at it. It should be understandable what you’re saying, but some people might start picking nits.

    Good luck.

  171. Gravatar of tom tom
    20. January 2012 at 01:46

    DR
    The only assumption that would be more emphasized is that firms want to maintain certain amount of inventories and fiscal stimulus drain them so they have to replenish them.
    Besides that my proofs are ok. And we can think that these quantities are measured in money so your remarks in point 3 are irrelevent.

  172. Gravatar of ssumner ssumner
    20. January 2012 at 07:41

    rob, See Tom’s reply.

    Tom, Yes, we can assume something completely different from what he said, and it might be right. Of course we can do the same for Cochrane, which was precisely my point.

    Major Freedom, Perhaps some people define AD differently, I’ve always seen it defined as C+I+G.

    Math, You said;

    “On the other hand, they want to be closer to the market reality. So, they relax that condition when they are trying to analyze the real economy.”

    I don’t think they do. They assume desired S and I differ, but actual are the same (in all the Keynesian analysis I’ve seen.)

    DR, If there is no change in income, there is nothing to “smooth.” In any case, WL said C fell, so it’s doubly wrong.

    Obviously he was trying to show that consumption smoothing proved Cochrane wrong, and he failed.

    Kevin, But there is no national saving in that model, private saving equals government dissaving.

  173. Gravatar of D R D R
    20. January 2012 at 07:59

    “DR, If there is no change in income, there is nothing to “smooth.” In any case, WL said C fell, so it’s doubly wrong.”

    You keep saying this, but it makes no sense. At all. Suppose our economy has two private agents. Each gets taxed 50, but one gets paid 100 to build a bridge. Aggregate income rises by 100, but aggregate disposable income is unchanged. This is NOT the end of the story.

    The disposable income of agent 1 has fallen by 50 and wishes to borrow 10. The disposable income of agent 2 has risen by 50 and wishes to save 10. So agent 2 loans 10 to agent 2. In this manner, each agent SMOOTHS the effect of the tax-financed infrastructure.

    Get it?

  174. Gravatar of D R D R
    20. January 2012 at 07:59

    Ach. Obviously, that’s a typo. I meant agent 2 loans to agent 1.

  175. Gravatar of D R D R
    20. January 2012 at 08:08

    One reason that you have trouble with the “C fell” thing is that you, along with me, misread Wren-Lewis, who says that he doesn’t include investment in his model at all.

    But his statement was “so their spending at time t will fall by much less than X.”

    Is zero much less than 100? I’m pretty certain it is.

  176. Gravatar of tom tom
    20. January 2012 at 08:13

    Scott, but the diference is that Keynesians (at least Krugman and Woodford) differantiate cases when fiscal stimulus can and cannot be appropriate. Cochrane claims that it’s bad in all cases and this claim he derives from theoretical reasons. If he just said that fiscal stimulus in theory can be ok but in practice no because for example government is a bad manager that would be a different situation.

  177. Gravatar of tom tom
    20. January 2012 at 09:39

    Scott,
    “Find an example where the final change in AD can be described as the change in C plus the change in G.”

    If you exclude inventories from AD as Keynesians do (you can check my links to wikipedia) then my examples fit your requirements because dAD=dC+dG and dI=0 and dInventories=dS.

    Do you agree Scott?

  178. Gravatar of Miguel Madeira Miguel Madeira
    20. January 2012 at 10:18

    “1. Wren-Lewis claimed actual C would fall, but would decline less than actual Y-T. I claim that because he said it, and because he assumed consumption smoothing.

    2. Point one implies that actual private saving fell, because actual saving equals actual income minus actual consumption. ”

    I disagree from point 2 – if “C” falls less than “Y-T”, but “Y” rises, it is possible that actual private saving rise (even if savings rate fall), or at least falls less than the increase in “C”

  179. Gravatar of Kevin Donoghue Kevin Donoghue
    20. January 2012 at 10:36

    Scott: “But there is no national saving in that [Y=C+G] model, private saving equals government dissaving.”

    Of course. I’m quite sure Simon Wren-Lewis understands that. Where I think you are misreading him is, you think he’s trying to prove a sweeping proposition, when what he’s trying to show is that Lucas and Cochrane are doing that; specifically, they are claiming that there is something quite absurd about Keynesian economics which can be grasped by anyone who thinks a bit about adding-up constraints.

    Lucas and Cochrane don’t need capital accumulation for the (dubious) argument they are making, so Wren-Lewis doesn’t need it to point out the flaw. Lucas says you must apply the same multiplier to T as you do to G, which is just wrong in a model with Ricardian equivalence. Of course there are lots of models in which dY/dG = 0 in any case, but Lucas clearly isn’t referring to a model of that kind.

    Wren-Lewis does mention saving later in the post, but it’s not needed for his debunking of Lucas and Cochrane. For that a Y=C+G model does fine.

  180. Gravatar of Major_Freedom Major_Freedom
    20. January 2012 at 11:03

    sumner:

    “Major Freedom, Perhaps some people define AD differently, I’ve always seen it defined as C+I+G.”

    I hold C+I+G to be in dollar spending terms, not physical terms.

    But by saying inventory build up contributes to AD, one is implying there is some sort of additional money spending associated with the inventory build up, when the opposite is the case (consumers reduce spending, accumulate cash, consumer goods inventory builds up, AD falls).

    If consumer goods are physically built up, if capital goods are physically built up, these are not contributive to AD because there is no additional money demand associated with them. If they are built up by way of productivity gains, then no fall in AD necessarily follows, because it’s selling more for lower prices, and unchanged demand and thus unchanged spending on “C”. If they are built up by way of reduced spending, then a fall in AD does necessarily follow, because spending on “C” is reduced.

    The “I” in the C+I+G total is in money spending terms. You can’t take “5 additional computers, 4 additional microwaves” and say “C” increased and thus C+I+G total increased, unless you relate additional money spending to these additional consumer goods. So in the case of consumers reducing their spending and accumulating cash, C is reduced because spending in C is reduced. But I is not increased because there is no increase in I spending. There is an increase in I in the physical sense yes, but C+I+G is not physical, its monetary. It’s adding up spending, not physical goods.

  181. Gravatar of tom tom
    20. January 2012 at 13:01

    Scot
    Final proof that you are wrong

    “First recall that C + I + G = AD = GDP = gross income in a closed economy”

    It is wrong because GDP-AD=Inventories so it can be positive

    “But Wren-Lewis seems to forget that saving is the same thing as spending on capital goods. Thus the public might spend $20 million less on consumer goods and $80 million less on new houses. In that case private aggregate demand falls by exactly the same amount as G increases”

    Again wrong because the public might spend $80 million less on houses from existing inventories. In that case private aggregate demand falls by less amount as G increases. So there was consumption smoothing that caused positive multiplier.

  182. Gravatar of Major_Freedom Major_Freedom
    20. January 2012 at 13:29

    tom:

    “Again wrong because the public might spend $80 million less on houses from existing inventories. In that case private aggregate demand falls by less amount as G increases. So there was consumption smoothing that caused positive multiplier.”

    That’s wrong. You are ignoring the fall in consumption and investment spending that could have been made but weren’t, because G increased. The money the government spends is itself money that was not spent by the private sector. So if government spending rises, then it sacrifices everything that could have been spent in the private sector.

    Thus, you cannot claim to know that there is a positive multiplier.

    Take $100 from me and spend it, and that makes it impossible for me to have spent it myself. You cannot say your spending brought about a higher aggregate spending than what I would have done had you not taken my money.

    You are also incorrectly treating consumer spending as the only spending that takes place.

  183. Gravatar of tom tom
    20. January 2012 at 13:35

    It is easy to understand it intuitively. If there is a lot of slack in the economy and government buys this not used stuff than certainly private aggregate demand will fall by less than after-tax income (Y-T). If we assume that this stuff is used productively than multiplier will be positive.

  184. Gravatar of tom tom
    20. January 2012 at 13:41

    Major
    I’m not ignoring anything. I have just changed that houses are bought from inventories. In this case there is consumption smoothing and if they are used productively than multiplier will be positive.

    “You cannot say your spending brought about a higher aggregate spending than what I would have done had you not taken my money.”
    Major I can say that because of slack. In other words because you dont want to spend I have to do it:)

  185. Gravatar of tom tom
    20. January 2012 at 14:12

    Another explanation:
    Let’s assume that public spends only on consumption goods and government buys only unsold stuff from inventories.
    Let’s use Scott’s numbers: G increases by 100 C decreases by 20 and S decreases by 80.
    In this case private aggregate demand falls by 20 even though S decreases by 80. People reduce their spending by less than government increases its outlays. Because how can you claim that public reduces its spending by more than 20? It’s against assumptions. And still savings identity is maintained.

  186. Gravatar of math math
    20. January 2012 at 14:14

    Sumner; “I don’t think they do. They assume desired S and I differ, but actual are the same (in all the Keynesian analysis I’ve seen.)”

    It’s very unfair. Krugman said to you that they equate through the “mechanism,” i.e., they can be different at some point in time.

    “Don’t tell me about how the identity must hold, tell me about the mechanism that induces the individual decisions that make it hold.” (http://krugman.blogs.nytimes.com/2012/01/16/mistaken-identities-wonkish/)

    Why are you distorting their views?

  187. Gravatar of John Fitzgerald John Fitzgerald
    20. January 2012 at 14:22

    When the gov’t removes funds from the individual wage earner in the form of taxes, spendable income decreases. With production constant and demand constant, either prices decline or wages increase such that the taxes now represent additional money in the economic flow.

    Now, when the government then spends that money on otherwise excess and untapped resources, like unemployed labor or excess agricultural capacity, it creates additional economic flow in addition to the already existing GDP.

  188. Gravatar of Kailer Kailer
    20. January 2012 at 14:35

    Tom,
    Houses that are in inventories are not new. They were new when they were added to inventories, but once there are no different than old homes. Buying a house from inventories isn’t savings since the company you buy the house from disaves by the exact same amount so in aggregate there’s no change in savings. So you reduce your spending on houses from inventory and that reduces the income of whoever owns the inventory and so now they have to reduce their spending on investment or consumption by the same amount.

  189. Gravatar of rob rob
    20. January 2012 at 14:52

    RE: rob, See Tom’s reply.

    Tom’s last reply appears to say that when firms inventories get used up they will increase investment to replenish them. But this will happen simultaneously with the switch from savings to consumption (as soon as inventories go down firms will start to replenish). Beyond the “splitting hairs” level this is exactly what Wren-Lewis said.

    Which is why I am underwhelmed

  190. Gravatar of tom tom
    20. January 2012 at 15:00

    Kailer

    You are wrong decreasing inventories decreases savings (check wikipedia or any textbook). And I havent said that buying house from inventories is savings. Your statement would be correct if you had noticed that in aggregate there’s no change in aggregate private demand because of this activity.

    “So you reduce your spending on houses from inventory and that reduces the income of whoever owns the inventory and so now they have to reduce their spending on investment or consumption by the same amount.”
    Again wrong. That doesnt reduce income if you notice that government buys stuff from inventories. But because of reduced private spending there is smoothing. And again if this stuff is used in a good way multiplier will be positive.

  191. Gravatar of D R D R
    20. January 2012 at 15:05

    math,

    I don’t think it is possible for the identity to fail to hold. Pick any values for C, I, G, and T and you find that…

    Y = C+I+G
    S = (Y-T-C)+(T-G) = Y-C-G = I

    Krugman’s point is that they may adjust in many different ways. In particular, an act of saving may leave investment unchanged.

    For example,
    1) If a car rolls of the line, investment (inventories), output and savings each have risen by the same amount.
    2) If a car is purchased, inventories falls by the amount consumption increases, so output is unchanged, but savings has fallen.
    3) If I suddenly decide not to buy a car after all, and hold on to my cash, then inventory investment rises by the amount my consumption falls. Output is unchanged, but my savings has risen to match investment.
    4) If I deposit money at a bank I am loaning money, so my savings is offset by the bank’s dis-saving, so nothing changes
    5) If the bank turns around and loans “my” money to a business, then again the banks’ savings is offset by the business’ dis-savings
    6) If the business then builds a factory with the “bank’s” money, then investment rises, and output and savings have risen to match investment

    Whatever you do, it has to show up somewhere else in the accounting– even if it is simply cancelled out on the other side of the transaction.

  192. Gravatar of tom tom
    20. January 2012 at 15:10

    Rob
    that is correct. I had to change a little my analysis because according to wikipedia inventories are excluded from AD. It doesnt influence conclusions but it is easier to show that Scott is wrong.

  193. Gravatar of rob rob
    20. January 2012 at 15:17

    When you think about it in terms of real goods rather than money flows then you can see why inventories count as savings/investment.

    Has Scott ever addressed specifically the reason you thing he is wrong (ie that inventory rebuilding will start almost immediately after t1) ?

  194. Gravatar of tom tom
    20. January 2012 at 15:28

    Rob
    I think the main problem with Scott is that he claims that C + I + G = AD = GDP = gross income in a closed economy and that is not true because GDP=Gross Income but not AD. Scott doesnt exclude possibility of positive multipliers in general but not in the case of consumption smoothing (i.e. when private spending falls less then after tax income). When we agree that AD can differ from Y than it is easy to construct examples with consumption smoothing and positive multipliers. Additionally it is easy to show that dAD can equal dC+dG something Scott said is not possible.

  195. Gravatar of tom tom
    20. January 2012 at 15:51

    The more I think about it the more I’m convinced that Krugman is right when he said that “it is very sad” if professors of economics dont understand basic Keynesians models. Because if they did they would knew that Y doesnt have to equal AD in these models and there would not be all these confusions.
    If you understand assumptions in Keynesians models they are internally consistent.

  196. Gravatar of D R D R
    20. January 2012 at 16:30

    tom,

    Part of the problem is that AD is actually a schedule, not a quantity. Krugman’s graph of intersecting “S” and “I” likewise is showing the intersection of savings and investment schedules– how much savings is desired at any quantity of output, and how much investment is desired at any quantity of output. Their intersection indicates the levels of S, I, and Y which are consistent.

    AD is technically a curve like the “S” or “I” in the graph.

  197. Gravatar of tom tom
    20. January 2012 at 16:31

    It is easy to explain why change in inventories doesnt influance AD. If one person sells something to someone else his net spending decreases by the amount of money he has received and the other person increases his net spending by the same amount. So aggregate spending has not changed.

  198. Gravatar of tom tom
    20. January 2012 at 16:37

    DR
    I agree. You can plot a graph where you can have Y on one axis and AD on the other. And the slope has not to be 45 degrees because dAD may differ from dY.

  199. Gravatar of tom tom
    20. January 2012 at 16:55

    I have just looked at wikipedia at keynesian cross. And there you have Keynesian cross diagram with the slope different than 45 degrees (just like I said). So how can Scott claim that he understands “1962-vintage Keynesian” because he is not “dim-witted second-grader”? If he understood this diagram he would knew that AD=Y only in equilibrium. But who can claim that currently our economy is in equilibrium ?

  200. Gravatar of Major_Freedom Major_Freedom
    20. January 2012 at 20:20

    tom:

    “I’m not ignoring anything. I have just changed that houses are bought from inventories. In this case there is consumption smoothing and if they are used productively than multiplier will be positive.”

    Again you are ignoring investment spending.

    “You cannot say your spending brought about a higher aggregate spending than what I would have done had you not taken my money.”

    “Major I can say that because of slack. In other words because you dont want to spend I have to do it:)”

    You can’t say that. If you tax me $100, then you don’t know what I would have done with it myself if you didn’t tax me of that money. And because tax falls on transactions, not cash balances, it means that the most likely counterfactual world would have been me spending that $100 myself. I mean, in this world, the world where you taxed me, I put that $100 into circulation. That’s how you were able to tax it!

    How can you say I am unwilling to spend it, if you don’t sit back and observe what I in fact do? By taking my money, you are preventing any chance of me showing you that I am going to spend it myself. You can’t claim to know I wouldn’t have otherwise spent it. That requires you to be able to read my mind.

  201. Gravatar of Morgan Warstler Morgan Warstler
    20. January 2012 at 21:56

    {Economics (at least macro) is (virtually) unfalsifiable. That’s why it is called “the dismal science.””

    Just run your model assuming there is only one global unprintable, infinitely divisible digital currency.

    Everything else is BS.

    You can still try and fiddle around with tax policy and income redistribution, but work from the premise that Mundell WINS, and if your ideas don’t exist anymore then bye-bye.

  202. Gravatar of tom tom
    21. January 2012 at 02:43

    Major

    We are considering model without investment where still savings identity is maintained. If economy is outside of equilibrium (i.e. AD is not equal Y) then we have consumption smoothing and positive multiplier.
    Again in this model because Y does not equal AD it means that there is not sufficient aggregate spending (so I was only joking that exactly you dont want to spend). That is why taxing in order to increase government spending is good.
    In order to refute my logic you would have to refute my claim that there is not sufficient spending in the economy. But in this case you would have to explain why there is high unemployment and output gap. Of cource you can always argue that governmnet is a bad manager and is not able to spend efficiently but it is practical argument not theoretical.
    To summarize: in Keynesan model there is consumption smoothing and positive multiplier if economy is outside of equilibrium (just like in current economy).

  203. Gravatar of tom tom
    21. January 2012 at 03:08

    Major and even if we included investments in this model it wouldnt change anything because we can always assume that change in savings is mirrored in change in inventories not investments. So the conclusions would be the same.

  204. Gravatar of Major_Freedom Major_Freedom
    21. January 2012 at 10:51

    tom:

    “We are considering model without investment where still savings identity is maintained. If economy is outside of equilibrium (i.e. AD is not equal Y) then we have consumption smoothing and positive multiplier.”

    “Again in this model because Y does not equal AD it means that there is not sufficient aggregate spending (so I was only joking that exactly you dont want to spend). That is why taxing in order to increase government spending is good.”

    “In order to refute my logic you would have to refute my claim that there is not sufficient spending in the economy.”

    “But in this case you would have to explain why there is high unemployment and output gap. Of cource you can always argue that governmnet is a bad manager and is not able to spend efficiently but it is practical argument not theoretical.”

    “To summarize: in Keynesan model there is consumption smoothing and positive multiplier if economy is outside of equilibrium (just like in current economy).”

    “Major and even if we included investments in this model it wouldnt change anything because we can always assume that change in savings is mirrored in change in inventories not investments. So the conclusions would be the same.”

    There is so much wrong here that it is almost impossible to know where to start.

    First, regardless of whether you include investment in the model, the point is that taxation falls on transactions, not cash balances. That fact is crucial. It refutes your logic as illogical.

    In order for ANY taxation to take place, there must be antecedent transactions to tax. Without transactions taking place, taxation cannot occur. The government doesn’t tax cash balances, they tax transactions.

    Therefore, because transactions are logically antecedent to taxation, because taxation requires existing transactions to take place, it means that the government taxing transactions and then spending the taxes on their own transactions cannot possibly RAISE aggregate spending from what it otherwise would have been anyway through the non-taxed transactions. It can only ever keep the same “spending” intact. The only difference is that instead of the receivers of the money getting 100% of the transacted money, they get say 90% and the government gets 10%.

    The government cannot bring about an increase in aggregate spending by taxing transactions. They can only ever redirect already transacting money away from where the private sector would have spent it, towards where the government wants to spend it.

    The only way that taxation financed government spending can bring about an increase in total spending, is if the government taxed people’s cash balances, and not transfers of cash balances. But since they only tax transfers of cash balances, it means transactions and hence additions to NGDP are logically antecedent to taxing and spending. Eliminating taxing and spending will just mean that the same NGDP will consist of 100% private transactions, instead of, say, 90% private and 10% public transactions. Taxation of transactions simply cannot increase AD, ever.

    Suppose you observe myself and someone else each own $1000 in cash balances, and in addition to this, we engage in a daily transaction of $50 per day. Every day, I give him $50 and in exchange he gives me a Keynesian lesson, and every day, he gives me $50 and in exchange I give him an economics lesson. What is NGDP? NGDP is $100 a day. NGDP does not include the $2000 total cash balances. It only includes the transactions.

    Now suppose you are government, and you want to introduce taxation. What money can you tax exactly? You can’t tax either of our cash balances, you can only tax our transactions. If we aren’t already spending, if we aren’t already exchanging $100 total a day, then you cannot tax us of anything.

    So let’s suppose you do introduce a 10% tax. What happens? Well, every time I spend $50, $45 goes to the guy giving me Keynesian lessons, and $5 goes to you. Every time the other guy spends $50, $45 goes to me who gives economics lessons, and $5 goes to you.

    Now at this point I can almost guarantee you are thinking that the government doesn’t just hold onto tax money, it spends the tax money. So what we should do is add $10 government spending. Well, OK, fine, let’s add it. We can add the $10 tax financed spending, but then I must ask you: Add the $10 to what exactly? Well, we have to add that $10 daily spending to the spending that myself and the other person are giving to each other. What are we giving to each other each day? Well, with you around taking a 10% cut, we are actually giving each other only $45 each day, for a total of $90 of spending each day. I get $45 a day, he gets $45 a day, for a total of $90 each day.

    What’s the new NGDP with you in the mix? NGDP (or AD), is now $10 + $45 + $45 = $100. It’s unchanged!

    The catastrophic flaw in your reasoning is that you are incorrectly taking the government’s taxing and spending and treating it as exogenous and given, then you look to the future and what people do after the fact. You can’t do that. You have to compare you taxing transactions versus you not taxing transactions. You are failing to take into account the fact that taxing falls on already existing transactions that would have otherwise been counted fully to NGDP. Taxation does not fall on cash balances that somehow turn idle cash into spending cash, and thus boosting AD.

    Again, I cannot repeat this enough: Taxation falls on transactions, not cash balances. Transactions and thus AD is logically antecedent to taxation. Taxation cannot enter the picture unless there is AD already there. Because of that, taxation can only EVER redirect existing AD, and cannot ever add to AD.

    If you didn’t tax our transactions, then NGDP or AD would have been $50 + $50 = $100 each day. With you taxing and spending, NGDP or AD becomes $10 + $45 + $45 = $100 each day. AD is the same! The only difference is that instead of two people receiving $50 each for a total of $100 each day, there is now a third person in the mix, acquiring and spending out of the same $100 total spending each day, so that you acquire and spend $10, while we acquire and spend $45 each, or $90, for a total of $100 AD.

    One final error you made is the claim that AD can somehow be different from Y. That is utterly impossible. AD is the sum of money transacted over a period of time. Y is the sum of money transacted over a period of time. Therefore, if you consider a period of time, and you want to find what AD is, then you will at the same time be finding out what Y is as well.

    The error you are making stems from the belief that Y has any meaning outside of actual transactions in money. The belief is that one can attach some sort of notional or monetary “value” to output, and then say “Hey, the total “value” of output is higher than AD! That means Y is lagging AD. Hey, that means government has to spend money to boost AD to match Y again.”

    This muddleheaded belief has its origin in the false and totally unwarranted assumption that there exists some Platonic, abstract “price” to goods that is independent of market prices in actual transactions. Now sure, in the accounting sense, business firms account for unsold inventory by attaching a price to them. But these valuations are not actual market prices. They are always expected market prices. This is true even if the inventory in question is physically identical to other inventory that is being sold and actually does have market prices.

    If AD lags the estimated, accounting “Y”, then why should AD be boosted by government? Why can’t the estimated, accounting “Y” be reduced via market forces, to match AD? After all, Y is borne out of market prices. Market prices are not borne out of Y. Total monetary demand isn’t set according to aggregate prices as if aggregate prices are primary. Aggregate prices are set according to aggregate monetary demand (and aggregate supply). Prices are established according to supply and demand.

    I mean really, suppose all business firms in the country accounted for their inventory in such a way that they expected to sell their goods for $1 trillion. Since there is not nearly enough money in the economy to even make possible the expected prices associated with this demand into actual prices, then does that mean someone should print enough money and spend it, in order to enable business firms to sell their goods at the prices they expected? Or should we want business firms to reduce their expected prices down to the market driven demands and prices?

  205. Gravatar of Major_Freedom Major_Freedom
    21. January 2012 at 11:05

    Small correction:

    “The error you are making stems from the belief that Y has any meaning outside of actual transactions in money. The belief is that one can attach some sort of notional or monetary “value” to output, and then say “Hey, the total “value” of output is higher than AD! That means Y is lagging AD. Hey, that means government has to spend money to boost AD to match Y again.””

    should read:

    “The error you are making stems from the belief that Y has any meaning outside of actual transactions in money. The belief is that one can attach some sort of notional or monetary “value” to output, and then say “Hey, the total “value” of output is higher than AD! That means AD is lagging Y. Hey, that means government has to spend money to boost AD to match Y again.”

  206. Gravatar of D R D R
    21. January 2012 at 12:01

    “Now suppose you are government, and you want to introduce taxation. What money can you tax exactly? You can’t tax either of our cash balances, you can only tax our transactions. If we aren’t already spending, if we aren’t already exchanging $100 total a day, then you cannot tax us of anything.”

    Very strange use of the word “can’t” regarding cash balances. Why do you make such an assertion?

  207. Gravatar of D R D R
    21. January 2012 at 12:03

    “Hey, the total “value” of output is higher than AD! That means AD is lagging Y. Hey, that means government has to spend money to boost AD to match Y again.”

    Or, you know, this means the private sector will want to expand output.

  208. Gravatar of D R D R
    21. January 2012 at 12:06

    *contract

  209. Gravatar of ssumner ssumner
    21. January 2012 at 13:45

    DR, You said;

    “But his statement was “so their spending at time t will fall by much less than X.”
    Is zero much less than 100? I’m pretty certain it is”

    I hope you aren’t being serious here. I think I know the meaning of the word ‘fall.’

    In your case, there is no aggregate consumption smoothing, so he’s wrong that consumption smoothing somehow proves Cochrane wrong. You’d get the same result w/o any smoothing.

    Tom, I agree Cochrane was wrong.

    Keynesians do not exclude inventories from change in Y, which is what the multiplier measures.

    Miguel, No because private saving is (Y-T) – C.

    Kevin, You said;

    “For that a Y=C+G model does fine.”

    Again, you cannot have consumption smoothing in a balanced budget context w/o investment, he’s simply wrong.

    Major freedom, I don’t follow your comment at all. Dollars vs volumes have nothing to do with what we are discussing.

    Tom, I believe AD is usually defined as C+I+G in textbooks. The fact that Wikipedia disagrees don’t make me wrong, they simply use a different definition. In any case, this is unrelated to the WL mistake.

    Math, First Krugman misunderstood me, now you are misunderstanding Krugman. He most certainly does accept that S=I.

    It’s a waste of time talking about what Krugman said in that post, because it has no bearing on any assertion in my post. That’s why he didn’t quote me, it would have been immediately obvious that he’d misinterpreted me.

    Tom, You said;

    “Scott doesnt exclude possibility of positive multipliers in general but not in the case of consumption smoothing (i.e. when private spending falls less then after tax income).”

    False, please don’t misrepresent my views.

    You said;

    “The more I think about it the more I’m convinced that Krugman is right when he said that “it is very sad” if professors of economics dont understand basic Keynesians models. Because if they did they would knew that Y doesnt have to equal AD in these models and there would not be all these confusions.”

    No, this sort of statement is just sad. Do you really think I’m a moron?

    You said;

    “It is easy to explain why change in inventories doesnt influance AD. If one person sells something to someone else his net spending decreases by the amount of money he has received and the other person increases his net spending by the same amount. So aggregate spending has not changed.”

    And you think I am a little slow! Now we are getting into the silly season.

    DR, Changes in inventories don’t necessarily tell us anything about future changes in output.

  210. Gravatar of tom tom
    21. January 2012 at 14:00

    Major
    It doesnt matter what government taxes because if your transactions are taxed you could be forced to take money from your cash reserves in order to do regular bussines. In this way cash reserves are indirectly taxed. Besides as DR said you can tax for examples deposits in banks and so effectively you are taxing cash reserves. As you said in this way taxation financed government spending can bring about an increase in total spending.
    You also claim that Y can not differ from AD so please explain high unemployment and negative output gap.
    Actually you are contradicting yourself because you admitted that AD can difer from Y and that “If AD lags the estimated, accounting “Y” than adjustment should take place by reduction in prices. Ok it is possible but it doesnt negate possibility that it can take place by government stimulation (at least in a model).
    And why Keynesians favore the second solution(fiscal stimulus). Because prices are sticky for example nominal wages never fall this is historical observation. So in Keynesian model adjustment is quicker via demand stimulation.

  211. Gravatar of tom tom
    21. January 2012 at 14:06

    Scott, you are answering selectively so I will ask you only one question: What is keynesian cross?

  212. Gravatar of tom tom
    21. January 2012 at 14:12

    And Scott I am very sorry but I have to tell you that you are behaving like a moron because I’m pretty sure that by now you have learnt that in Keynesian model outside of equilibrium fiscal stimulus causes consumption smoothing and positive multilpier.

  213. Gravatar of tom tom
    21. January 2012 at 14:24

    Scott how can you claim that “C + I + G = AD = GDP = gross income in a closed economy ” and that you understand “1962-vintage Keynesian cross”?
    It is contradiction. If you admit that it was your mistake then certainly you are not a moron.

  214. Gravatar of tom tom
    21. January 2012 at 14:40

    Scott if you were honest you would admit that in classical model because Y equal AD then consumption smoothing impies zero multilplier. But in Keynesian model Y may differ from AD so consumption smoothing implies positive multiplier.
    So Wren Lewis and Krugman are right and we can end this discussion.

  215. Gravatar of math math
    21. January 2012 at 16:44

    Scott Sumner,

    OK, I disagree with you here. I think you are rather playing potical games. And I don’t like it.

  216. Gravatar of Major_Freedom Major_Freedom
    21. January 2012 at 21:02

    tom:

    “It doesnt matter what government taxes because if your transactions are taxed you could be forced to take money from your cash reserves in order to do regular bussines.”

    But if you are already spending the gross amount, which is taxed, than that is transaction based, not cash balance based.

    “In this way cash reserves are indirectly taxed. Besides as DR said you can tax for examples deposits in banks and so effectively you are taxing cash reserves.”

    When does the government tax cash balances?

  217. Gravatar of tom tom
    22. January 2012 at 02:54

    Major dont be stubborn. If transaction A is taxed then either you dont have enough money to do transaction B and you have to decrease your economic activity or you have to take money from your cash reserves in order to maintain your economic activity on the same level. So evectively your cash reseves are taxed.

  218. Gravatar of Kevin Donoghue Kevin Donoghue
    22. January 2012 at 03:47

    “…you cannot have consumption smoothing in a balanced budget context w/o investment….”

    As so often before, I get the sense that you have your own definition of a key concept which is not the one I’m used to. To me, to say that a model incorporates consumption-smoothing simply means that households seek a stable path for consumption. If they have to pay out more in the current period (due to an increase in the current-period lump-sum tax, for example) they will seek to spread the impact on consumption over many periods. You can certainly have that in a Y=C+G model. See Woodford’s model for example.

  219. Gravatar of ssumner ssumner
    22. January 2012 at 09:27

    Tom, It’s a model where G and I are exogenous and C is a function of disposable income with an MPC less than 1. Y=C+I+G.

    Now will you please leave me alone?

    Kevin, You said;

    “To me, to say that a model incorporates consumption-smoothing simply means that households seek a stable path for consumption.”

    No, I think it’s much more specific than that. It means people smooth changes in after-tax income. If you have a balanced budget multiplier, than changes in Y-T equal changes in Y-G. If you have consumption smoothing, then C changes by an amount different from the change in Y-G.

    But in order for C to change by an amount different from the change in Y-G, you need a change in I.

  220. Gravatar of D R D R
    22. January 2012 at 09:33

    “You’d get the same result w/o any smoothing.”

    So what? To prove Cochrane wrong, requires only a single counter-example. A counter-counter-example does not make Cochrane correct.

    Furthermore, you are still parsing Wren-Lewis in the worst possible way.
    1) Wren-Lewis wrote, “If you raise taxes by X at time t…” and you took this to mean, “If you raise taxes *and spending* by X at time t…”

    Even granting your interpretation,
    2) Wren-Lewis continued, “consumers will smooth this effect over time, so their spending at time t will fall by much less than X” and you parsed this as “spending at time t will fall *by some nonzero amount but* by much less than X”

    Even granting *this* as well, let us continue. So suppose the amount of the change in consumption is 100<dC<0. Thus,

    C+dC G+100 T+100 Y+100+dC Yd+dC

    and in the aggregate there is now UNDER-consumption relative to the desired amount. (And there must be at least one agent under-consuming)

    So to the extent that Wren-Lewis is "wrong" it is that he understated his case.

  221. Gravatar of Kevin Donoghue Kevin Donoghue
    22. January 2012 at 10:25

    Scott: “If you have consumption smoothing, then C changes by an amount different from the change in Y-G.”

    I bet your definition will seem as strange to Krugman and Wren-Lewis as it does to me. Small wonder that you disagree; you’re not speaking the same language.

  222. Gravatar of Advice to Scott: Avoid Accounting Identities at ALL Costs « Uneasy Money Advice to Scott: Avoid Accounting Identities at ALL Costs « Uneasy Money
    22. January 2012 at 10:33

    […] Scott says […]

  223. Gravatar of Major_Freedom Major_Freedom
    22. January 2012 at 11:30

    tom:

    “Major dont be stubborn. If transaction A is taxed then either you dont have enough money to do transaction B and you have to decrease your economic activity or you have to take money from your cash reserves in order to maintain your economic activity on the same level. So evectively your cash reseves are taxed.”

    tom don’t be stubborn.

    If transaction A exists, then it’s because people are willing able and willing to put that amount of money into circulation. If they want to hold X cash, then taxing cannot decrease that quantity of money they want to hold as cash. It doesn’t make any sense to say “Given that people want to hold X cash, let’s pretend that they don’t really want to hold X cash. Then we can tax them more and they will put some of that X cash into circulation.”

  224. Gravatar of Major_Freedom Major_Freedom
    22. January 2012 at 11:38

    tom:

    “You also claim that Y can not differ from AD so please explain high unemployment and negative output gap.”

    Wage rates are too high relative to the clearing wage rates.

    Idle resources are investments that should not have been made, and should be liquidated.

    “Actually you are contradicting yourself because you admitted that AD can difer from Y and that “If AD lags the estimated, accounting “Y” than adjustment should take place by reduction in prices. Ok it is possible but it doesnt negate possibility that it can take place by government stimulation (at least in a model).”

    It’s not a contradiction because I don’t hold Y in the accounting sense. I said AD can lag “Y in the accounting sense.”

    And that argument isn’t the argument that does negate the efficacy of inflation and government spending in order to equate AD and Y in the accounting sense. Another set of arguments refutes it.

    “And why Keynesians favore the second solution(fiscal stimulus). Because prices are sticky for example nominal wages never fall this is historical observation. So in Keynesian model adjustment is quicker via demand stimulation.”

    Prices are sticky because of minimum wage laws, unemployment welfare, and generations of inflation induced psychology that makes people accustomed to increasing prices.

    It is precisely Keynesian policies that have made prices sticky.

  225. Gravatar of ssumner ssumner
    24. January 2012 at 08:48

    DR, You’re still making the same mistake as WL, forgetting the change in I. If C changes than I must change if there is consumption smoothing. (Assuming a BBM)

    Kevin, What’s strange about it? The change in Y-G is the change in after tax income in a balanced budget model.

    Read Noah Smith’s comments on my Noahopinion post. He gets it, and he’s a Keynesian.

  226. Gravatar of D R D R
    26. January 2012 at 10:19

    “DR, You’re still making the same mistake as WL, forgetting the change in I. If C changes than I must change if there is consumption smoothing. (Assuming a BBM)”

    No. I have read Noah’s post. I understand what he is arguing. You, on the other hand, are arguing that

    a) There is investment in Wren-Lewis’ model; and
    b) That a number *less* than a positive number must also be positive.

    Now, a) is just your read, but let’s put that aside. Now, b) is true once you restrict yourself to positive numbers, and there is no reason to impose such a constraint. But let’s put that aside as well.

    Now (we may presume) that if Wren-Lewis had *not* mentioned consumption smoothing, but rather assumed that consumption does not change, then C+0 I+0 G+100 T+100 Y+100 Yd+0 would have been just fine.

    He didn’t, so who cares? Well, if from a point of C+0 I+0 G+100 T+100 that individual consumers do smooth, the winners of government policy may loan money to the losers with no change in national savings or investment.

    In this way, we can say “Wren-Lewis said consumption fell, but he forgot that the consumption increase of the winners will compensate for the fall in consumption of the losers! In fact, consumption-smoothing does not result in a fall of consumption at all!”

    So yes, if you insist on your read of Wren-Lewis, he is “wrong”. However, he *understated* his case, rather than *overstated* it. I think that makes a difference.

Leave a Reply