Noahpinion has an opinion

I suppose I should be flattered by the nonstop attention my S=I post has received.  But I guess I’d slightly prefer the attention to be acclaim at my brilliant insight:  “S=I?   Why didn’t we think if that!!”

Paul Krugman devoted three posts to telling me that identities are definitions, not behavioral relationships.  And also explaining how the Keynesian cross model works.  I’ve been blogging for three years and I defy anyone to find a place where I confused identities and behavioral relationships.

In one post he cited Noah Smith, who warned of the danger of confusing accounting identities with behavioral relationships:

Accounting identities are mostly just definitions. Very rarely do definitions tell us anything useful about the behavior of variables in the real world. The only exception is when you have a very good understanding of the behavior of all but one of the variables in an accounting identity, in which case the accounting identity acts like a budget constraint. But that is a very rare situation indeed.

Of course I agree with Noah, as (I assume) does every other economist on the planet.  Last night I saw this in a post by David Glasner:

Have a look as well at Brad DeLong who has a new post quoting Paul Krugman quoting Noah Smith on the dangers of accounting identities, and also quoting moi.

And I found that Brad DeLong also quotes Noah Smith, although Brad doesn’t mention me in his initial post.  (Which is interesting, as he certainly has not been shy about calling me an idiot in previous posts.)  Then Brad DeLong added an update with a quote from an earlier Glasner post.  Here’s the first sentence of the Glasner quotation that Brad provides:

Why Am I Arguing with Scott Sumner? « Uneasy Money: When Scott says he can derive a substantive result about the magnitude of the balanced-budget multiplier from an accounting identity between savings and investment, he is making a theoretically ungrammatical statement. . . .

Most of my disagreement with David is about the S=I identity.  Paul Krugman, Simon Wren-Lewis and I think it’s an identity for actual saving and investment, albeit not for planned saving and investment.  In fact, that distinction lies behind the paradox of thrift.  But David disagrees, claiming that S=I is not an identity at all.  We’ll just have to agree to disagree on that point.  But the sentence DeLong quoted is flat out wrong.  I never would claim that an identity could tell us anything about the size of the multiplier.  That would be nonsense.  Definitions prove nothing about causation.

As I said, David strongly disagrees with Krugman and Wren-Lewis about the identity question, but he and Krugman and DeLong all seem to like Noah Smith’s post (or at least I inferred that from the fact that they all cited him.)  So let’s see what Mr. Smith thinks of these arguments that I’m confusing accounting identities with behavioral relationships:

David –

I really should clear this up. I do NOT think that Scott was trying to make an “argument from accounting identity.” In the first post where he mentioned S=I, Scott was obviously *not* trying to claim that S=I implies a fiscal multiplier of zero. He was using S=I to help him demonstrate that consumption smoothing does not imply the smoothing of total private expenditure. That point is correct. It was not an argument from accounting identity, but rather the use of an accounting identity to illustrate the difference between two concepts (smoothing of consumption and smoothing of total private expenditure), and thus does not violate my Principle #4

Best,
Noah

Case closed?  Probably not.   So as long as others keep challenging my critique of Wren-Lewis, I’ll keep responding.  And I welcome these challenges, as I think this is a very important point.  Identities don’t prove anything, but proofs must be consistent with identities.  Krugman endorsed Wren-Lewis’s faulty “proof” and as far as I know has never retracted that endorsement.

PS.  I prefer the ultra-polite criticism by people like David Glasner, to the snarky stuff from you-know-who.


Tags:

 
 
 

122 Responses to “Noahpinion has an opinion”

  1. Gravatar of Gene Callahan Gene Callahan
    23. January 2012 at 06:41

    “I defy anyone to find a place where I confused identities and behavioral relationships.”

    Well, Scott, you certainly *seemed* to do so in the first S=I post. If you recognize that this is a matter of definition, then what the heck is all the “PERIOD. END OF STORY” BUSINESS”? That is not the way definitions work.

  2. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 06:49

    ” I prefer the ultra-polite criticism by people like David Glasner, to the snarky stuff from you-know-who.”

    You’re alway polite right? Please you write posts like “I Hate All Keynesian Talk” and claim that you find Keynesianism mindnumbingly stupid and yet you play like your miss manners. It’s a copout. It’s a crutch. You and your fellow freshwater guys are no more polite, in fact you are much less.

    After all it was Lucas and his friends who used to giggle at Keynesian arguments in class-his words. Get off your high horse Scott you don’t belong on it.

    Even with the oh so “polite” Robert Glasner you are putting words in his mouth that he denies an accounting indentity.

    “I suppose I should be flattered by the nonstop attention my S=I post has received. But I guess I’d slightly prefer the attention to be acclaim at my brilliant insight: “S=I? Why didn’t we think if that!!”

    You claim that Krugman doesn’t do justice to your profound subtle economic arguments you are wholly misconstruing those who disagree with your attempt to intevent on behalf of Cochrane and Lucas against Krugman and Simon-Wren.

    I wouldn’t be too flattered. The nonstop attention is due to your errors-again not in claiming that S=I as an accounting idnetity but the heavy weather you tried to make with it. Your quibbling about consumption smoothing has also failed to somehow prove Cochrane either right or at least to be making a non-absurd argument.

  3. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 08:03

    You want a case of you confoudning accounting identities with behavioral relationships you got it.

    “Wren-Lewis seems to be the one making a simple logical error (which is common among Keynesians.) He equates “spending” with “consumption.” But the part of income not “spent” is saved, which means it’s spent on investment projects. Remember that S=I, indeed saving is defined as the resources put into investment projects. So the tax on consumers will reduce their ability to save and invest.”

    https://www.themoneyillusion.com/?p=12596

    You also confound it in a later post with the title, “Saving is “setting mone aside” it’s BUYING CAPITAL GOODS.”

    By the way to a comment you made to me in your last post, “A few weeks ago you asked me why I didn’t read your blog more often. How can I put this politely . BTW, How old are you Mike?”

    Scott if you don’t read my blog that’s to your own detriment. I read your blog even though I clearly disagree on some important things.

    However, don’t cry for me Argentina. Today guess who else visited my blog-Unlearning Econ.

    Clearly he feels differntly about my blog and is now following me. check out his blog here

    http://unlearningeconomics.wordpress.com/

    I guess the one thing I should compliment you on is not playing the game other people have like Jane Hamsher at Fireoglake who banned me-she’s a flaming liberal for the record in case you don’t know her. I get into just as many fights with so called liberals as the Right wingers.

    Though I have criticized you in a polemical style you haven’t done that at least. I appreciate that.

  4. Gravatar of Benjamin Cole Benjamin Cole
    23. January 2012 at 08:37

    I dislike this diversion into extended arguments with Keynesians. Why?

    We had ’em. We had ’em when Krugman admitted that monetary policy can snuff out fiscal stimulus (obviously, but anyway).

    From that point we should have co-opted the Keynesians, and built up the idea that fiscal stimulus needs a major boost from monetary policy—-indeed the more aggressive the monetary policy the less you need fiscal stimulus, and perhaps only minor fiscal stimulus.

    From that platform it is not too far to pure Market Monetarism.

    Instead, we have segwayed into a backwater of arguments, in which Sumner reminds or educates economists that the New England Patriots cannot win the Super Bowl by showing nerves of steel combined with better relief pitching.

  5. Gravatar of Noah Smith Noah Smith
    23. January 2012 at 08:43

    I went back and read Wren-Lewis’ original argument:

    “If you spend X at time t to build a bridge, aggregate demand increases by X at time t. 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. Put the two together and aggregate demand rises.”

    This is clearly wrong in the *general* case, since private expenditure does not equal consumption. Saying “S=I” was clearly just a way of pointing out that the change in private investment must also be taken into account when calculating the total effect on aggregate demand.

    Also, Wren-Lewis argument does not take into account the *rise* in private income due to the fact that private individuals are being paid to build the bridge. In a Keynesian Cross model this will precisely cancel out the impact of the tax hike, leaving a balanced-budget multiplier of 1.

    However, reading Cochrane, it seems clear that Cochrane’s claim – that the multiplier must be identically zero – is not right without a whole bunch of assumptions (which Cochrane fails to specify). So Krugman is right to criticize Cochrane’s statement.

    As for all the stuff about S=I, plans, etc., I think it is basically a tangent.

  6. Gravatar of JL JL
    23. January 2012 at 08:49

    @Benjamin Cole.

    Full agreement here.
    I agree with Krugman and Sumner.
    The goal should be to end unemployment and return the economy to full productive capacity, whatever the means.

    The Fed moves last, so we need the Fed.
    But if Democrats want to spend a few billions so they can take credit for ending the Lesser Depression with a ‘New New Deal’, then it’s a small price to pay in the grand scheme of things.

    Politicians will be politicians: they need their token action so that they can claim credit in the eyes of the public and win the next election.
    Sure, democracy sucks, but it’s better than serving a king.

    The only people I don’t agree with are the austerity-loving inflation hawks: Fischer, Kocherlakota and Plosser.

    Treasonous bastards…

  7. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 09:11

    As it is a subject that has generated some interest-I see many from Money Ilusion stopped by without me putting the link in here it is for interested readers

    http://diaryofarepublicanhater.blogspot.com/2012/01/scott-sumners-economic-canards.html

  8. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 09:16

    Benjamin Cole, I think you are right. I was a lot more receptive to Scott’s arguments before every day he seemed to throwing more mud at Krugman.

    The idea of NGDP targeting still sounds good as best I can tell.

    Here though is a real question for you or Scott or anyoone else: Unlearning Econ made this following comment to me at Twitter.

    Speaking of Scott, “…he thinks macroeconomic policy can only influence nominal GDP. But it can have real impacts.”

    Is there not some truth in this? To take an extreme example to bring it home if tomorrow a metoer destroyed Europe and Asia what level of NGDP targeting would bring us back to normal growth?

    I would sincerely like Scott, Benjamin or anyone to answer this

  9. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 09:36

    “Full agreement here.
    I agree with Krugman and Sumner.
    The goal should be to end unemployment and return the economy to full productive capacity, whatever the means.”

    Agree with you too JL. The troulbe lately is it has seemed to be the monetarists and the austerity loving inflation hawks against the Keynesians. Which makes it seem like the real goal is just to ensure there is no fiscal stimulus

  10. Gravatar of johnleemk johnleemk
    23. January 2012 at 10:38

    “To take an extreme example to bring it home if tomorrow a metoer destroyed Europe and Asia what level of NGDP targeting would bring us back to normal growth?”

    Anyone seriously asking this question has not properly understood the distinction between aggregate demand and aggregate supply, and which of the two stimulus can affect.

    On another note, it seems as though people are taking this debate a lot more personally than it ought to be taken. This “you’re either with us or against us” logic seems to be implying that even when advocates of stimulus make mistakes in their case for stimulus, others on their “side” should ignore it.

    Krugman doesn’t hold back when Scott messes up. (I am trying to think of other monetary stimulus proponents, but Scott is probably the only economist in the blogosphere currently defined best by his advocacy of monetary stimulus.) It’s fair game either way.

  11. Gravatar of D R D R
    23. January 2012 at 10:56

    “Also, Wren-Lewis argument does not take into account the *rise* in private income due to the fact that private individuals are being paid to build the bridge.”

    I think this requires a bit more explanation, given that Wren-Lewis’ argument begins, “If you spend X at time t to build a bridge, aggregate demand increases by X at time t”

    What are you trying to say?

  12. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 11:07

    “man doesn’t hold back when Scott messes up. (I am trying to think of other monetary stimulus proponents, but Scott is probably the only economist in the blogosphere currently defined best by his advocacy of monetary stimulus.) It’s fair game either way.”

    Actually there are plenty others of them and am not even a Market Monetarist-how about NIck Rowe, Lars Chrisenssen, Woolsley, could go on all day.

    As to your claim that Krugman doesn’t hold back when Scott messes up I don’t know that I agree. Krugman doesn’t write about Scott much at all. Scott lately attacks Krugman or Keynesianism every day.

    I’m not “gettin personal” I just haven’t seen the errors Scott has been trying to hand around his neck.

  13. Gravatar of Kevin Dick Kevin Dick
    23. January 2012 at 11:27

    As a matter of logic, De Long is wrong about being able to derive something from an identity. It’s known as Proof by Contradiction. If you assume a conclusion and show that this assumption leads to the violation of an identity, you have disproven the conclusion.

    I’m not saying that’s the case in this argument. But it is in fact possible.

  14. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 11:46

    johnmleenk you’re probaly right about the distinction between AD and AS. But I think the real difference is this: to the extent that Scott, you or any other Market Monetarist claims that macroeconomic policy can influence influence only nominal GDP but not realy GDP you are mistkane and I think this is a big part of the disagreement on what efficacy fiscal stimulus has

  15. Gravatar of JL JL
    23. January 2012 at 11:46

    @johnleemk,

    I agree that the “Scott Sumner doesn’t understand comparative statics” was bollocks.
    I’m sure Krugman knows this: it was just a cheap and effective way to discredit Scott.

    Krugman knows that Scott was right, but you don’t win popular opinion and political debates by openly admitting your errors.

    Krugman made a gesture when he admitted that monetary policy can snuff out fiscal stimulus.

    The correct response would have been something like: “Glad we can agree. Though not fully convinced, I can understand that fiscal stimulus may seem necessary if monetary policy fails. Let’s work on proper monetary policy and perhaps some extra fiscal stimulus on infrastructure if unemployment remains high”.

    And then you publish a paper on why fiscal stimulus is bollocks. Because academic debate about politicized subjects should be made within academia, not in the public arena.

    > This “you’re either with us or against us” logic seems to be implying that even when advocates of stimulus make mistakes in their case for stimulus, others on their “side” should ignore it.

    In the political arena that is indeed the case: never discredit your allies, even when they goof up.

  16. Gravatar of Becky Hargrove Becky Hargrove
    23. January 2012 at 11:50

    You go Benjamin!
    Let’s build some bridges to solutions.

  17. Gravatar of D R D R
    23. January 2012 at 11:58

    “Krugman made a gesture when he admitted that monetary policy can snuff out fiscal stimulus.”

    Guns are harmless because, you know, Kevlar.

  18. Gravatar of ssumner ssumner
    23. January 2012 at 12:05

    Gene, Seemed doesn’t count, I want you to point out the paragraph where I made that mistake. I can’t be blamed if people read my posts to get “general impressions” rather than the specific arguments I make.

    Which paragraph?

    Mike Sax:

    You quoted me as saying:

    “I suppose I should be flattered by the nonstop attention my S=I post has received. But I guess I’d slightly prefer the attention to be acclaim at my brilliant insight: “S=I? Why didn’t we think if that!!”

    That was obviously a joke aimed at myself. I guess it went over your head, or else it was a lousy joke.

    You said;

    “Though I have criticized you in a polemical style”

    So that’s what you call taunting and insults: “Polemical.”

    The paragraph you quote is followed by two paragraphs, which consider multipliers of 0 and 1. So I don’t assume anything about the multiplier from the accounting identity. I acknowledge that the stimulus may boost income, and in that case consumption and savings are unchanged. But of course in that case there is no consumption smoothing. In additon, that paragraph comes right after a Wren-Lewis quote that assumes after tax income falls. So I’m actually not making the assumption that after tax income has to fall (as I point out two paragraphs later), but just following his assumption that it did.

    Ben, You’re still mixing baseball and football metaphors.

    Noah. Exactly. The argument WL should have made was that the balanced budget multiplier was one, and hence neither C nor S nor I would change at all. But he got greedy and thought he could use consumption smoothing against Cochrane, but ended up shooting himself in the foot.

    JL, Yes, monetary stimulus is needed:

    Mike Sax: You said;

    “Speaking of Scott, “…he thinks macroeconomic policy can only influence nominal GDP. But it can have real impacts.””

    That’s right, I don’t think macro policy can have any real effects on the economy. (That was a joke, in case you missed it.)

    DR, You said;

    “I think this requires a bit more explanation, given that Wren-Lewis’ argument begins, “If you spend X at time t to build a bridge, aggregate demand increases by X at time t”

    What are you trying to say?”

    Perhaps I shouldn’t speak for Noah Smith, but I think he was referring to the extra spending on consumption of those who built the bridge, which (according to the Keynesian BBM=1) should just balance out the reduced spending on consumption of those who pay taxes to build the bridge. Leaving GDP higher by the amount of the bridge itself.

    Mike Sax, You said;

    “I just haven’t seen the errors Scott has been trying to hand around his neck.”

    I agree, you haven’t seen them.

    Kevin, If you look at the first Smith quotation I provided, especially the last sentence about a “very rare situation” that applies exactly to this case. Wren-Lewis discussed everything except I. And he implicitly assumed I didn’t change even though he was implicitly assuming that S fell in equilibrium. The “budget constraint” is a good analogy. I would be like assuming Robinson Crusoe was consuming beyond his PPF.

  19. Gravatar of D R D R
    23. January 2012 at 12:45

    “Perhaps I shouldn’t speak for Noah Smith, but I think he was referring to the extra spending on consumption of those who built the bridge, which (according to the Keynesian BBM=1) should just balance out the reduced spending on consumption of those who pay taxes to build the bridge. Leaving GDP higher by the amount of the bridge itself.”

    … meaning that Wren-Lewis “erred” by understating his case?

  20. Gravatar of Kevin Dick Kevin Dick
    23. January 2012 at 13:02

    Scott, I agree with you that you probably caught Wren-Lewis in a proof by contradiction. I just don’t feel confident enough to pronounce it as so–I’m not “even” a professor at Bentley :-). But I do feel confident saying that De Long is wrong about dismissing discussions of identities.

  21. Gravatar of Lee Kelly Lee Kelly
    23. January 2012 at 13:55

    ‘1 + 1 = 2’ is only an identity, so don’t let it stop you making bogus economic arguments. Math is so bourgeois.

  22. Gravatar of Benjamin Cole Benjamin Cole
    23. January 2012 at 15:54

    Mike Sax-

    My point of view is that Market Monetarism is the best approach, whatever structural impediments are in place in a modern economy. (Your scenario of an asteroid smashing Europe would probably call for a socialized, government-controlled economy for a while, like postwar Japan. In that scenario, perhaps MM would not work).

    Yes, I would like less structural impediments, I would like taxes on consumption, not income—any number of ideas most economists agree with.

    But in the real world we have a modern industrial economy–one probably with less structural impediments than in the 1970s, btw. Global trade, a de-unionized workforce, and deregged industries, such a telecom and transportation.

    MM can work in the USA.

  23. Gravatar of Mike Sax Mike Sax
    23. January 2012 at 16:04

    My point is that if I understood what you said it would help the cause of MMers not to throw so much mud in the eyes if the Keynesians for no real good reason.

    All the major New Keynesians had jumped on the NGDP targeting bandwagon. But by attacking them every day you weaken any “co-opting” you might hope for.

  24. Gravatar of Noah Smith Noah Smith
    23. January 2012 at 17:52

    In the classic Keynesian model of a “balanced budget multiplier,” government taxes people X and spends X on building a bridge. People’s income goes down X from the taxes and goes up X from being paid to build the bridge. Hence, there is no change in their after-tax income, so their investment and consumption behavior are unchanged. Meanwhile, output rises by X because now we have a bridge that we didn’t have before (or, alternatively, because now people’s pre-tax incomes are higher by X). Thus, the balanced-budget multiplier in this model is 1. Taxing and spending X increased output by exactly X.

  25. Gravatar of Noah Smith Noah Smith
    23. January 2012 at 17:57

    So, basically, Wren-Lewis’ example does not correspond to the classic Keynesian model; he understates his case by forgetting about the extra income that people will receive from being paid to build the bridge, and arrives at a balanced-budget multiplier of 0.8 instead of 1.

  26. Gravatar of Noah Smith Noah Smith
    23. January 2012 at 18:03

    In fact, I believe that Scott made exactly this point in his first “S=I” post, where he said that if income doesn’t fall, then no consumption smoothing actually occurred, or something like that.

    The more general point is, I can’t think of any Keynesian type of model where the multiplier is inversely related to the elasticity of intertemporal substitution. Multipliers in old Keynesian models would be bigger if people consumed a fixed percent of their income instead of smoothing it. In the main New Keynesian models I don’t think it matters…

  27. Gravatar of Morgan Warstler Morgan Warstler
    23. January 2012 at 21:26

    $100K invested sells for $1.1M for $1M capital gain.

    Scott, please explain how the $1M has already been taxed.

  28. Gravatar of Marcelo Marcelo
    23. January 2012 at 21:43

    Scott,

    I was watching the Republican Debate tonight and Newt Gingrich was asked about the effectiveness of the Bush Tax Cuts on economic growth. He argued that the tax cuts helped the economy recover after 9/11 and that the recession would have been worse without them. This is a claim I hear from many of my friends further on the right than myself, then they will tend to go on and harp about the ineffectiveness of government stimulus.

    I usually try to explain to them that the government multiplier on a tax cut is still zero, as it is for any other form of government stimulus, particularly if the interest rate is above zero. Moreover, the effect of 9/11 on the economy was mostly a demand-side shock, as it decreased confidence in NGDP growth in the near future, having no supply side effects. Hence, even if one believes that the Bush tax cuts had supply side benefits, those are entirely unrelated to the events of 9/11…

    I know this is unrelated to your post, but I would like to know if you have any other commments on this kind of logic, though I think I have given a pretty Sumnerian-style answer to this question. I want to know your thinking on this because I think explaining this kind of thinking is vital to better explaining the effects of supply-side policies, as well as exposing the paradox of being anti-Keynesian but supporting not-payed-for tax cuts to end a recession…

  29. Gravatar of tom tom
    24. January 2012 at 04:21

    In keynesian model in case of recessionary gap if government buys unused stuff (so inventories are falling) then savings are falling. But how can you claim that investments have also to decline?

  30. Gravatar of TheMoneyIllusion » A picture tells 1000 words TheMoneyIllusion » A picture tells 1000 words
    24. January 2012 at 07:26

    […] here’s several observations that Noah Smith left in my comment section yesterday. In the classic Keynesian model of a “balanced budget multiplier,” government taxes […]

  31. Gravatar of Major_Freedom Major_Freedom
    24. January 2012 at 07:48

    “So as long as others keep challenging my critique of Wren-Lewis, I’ll keep responding. And I welcome these challenges, as I think this is a very important point.”

    I think a challenge was raised before but was not, IMO, adequately answered.

    It was about your position that a fall in consumer spending, and accumulating cash balances, results in an unchanged AD on the basis that “inventories rise.”

    The thinking goes: A fall in consumer spending results in a rise in inventories, and since inventories are included in I, we can say that because the rise in I offsets the fall in C, it means AD (which equals C+I+G) is unchanged.

    I challenged this by saying that aggregate demand is measured in terms of money spending, not real goods accumulating in warehouses. Unless you can show a “replacement” spending for the decline in consumer goods spending, then you cannot say AD remains unchanged. AD will fall because there is no additional spending that replaces the decline in consumer spending.

    C+I+G would go from, say, 100+100+50=250, to, say, 50+100+50=200. “I” does not increase because an accumulation of inventory on the basis of lower consumer spending and higher cash balances, does have an additional money spending component attached to it.

    Only if “I” increases by 50, can AD remain unchanged. In other words, only if consumers reduced their consumer spending but did not accumulate cash balances, and invested that money instead, can AD remain exchanged.

  32. Gravatar of ssumner ssumner
    24. January 2012 at 09:32

    DR, It’s worse, in the Keynesian model there’s no change in consumption (in the BBM case) and hence no consumption smoothing. See his newer comments.

    Of course all this simply assumes the Keynesian model is correct. As Karl Smith said, if it’s true, it’s true. But why should we believe it’s true?

    Kevin and Lee, Yes, identities are very useful in certain contexts.

    Mike Sax, I’ve praised Krugman many times on this blog. I once recommended he be put on the Fed. I’ve been far nicer to him than he has been to me (not that I’m complaining.)

    Morgan, Check out this post:

    https://www.themoneyillusion.com/?p=7091

    Marcelo, I wasn’t a big fan of the Bush tax cuts, but they probably had a small positive impact on the supply side. I would have preferred tax reform, which would have helped a lot more. 9/11 reduced both AS and AD, but not by very much. The recession ended 2 months after 9/11

    Tom, In the Keynesian model inventory adjustments are a part of investment.

    Major Freedom, You said;

    “I challenged this by saying that aggregate demand is measured in terms of money spending, not real goods accumulating in warehouses.”

    Money is also spent on inventories, isn’t it? And AD doesn’t include money spent in intermediate goods. So money spent is not a good way to estimate AD.

  33. Gravatar of Time to Move On – But Not Before I Explain (Definitively) What it all Means « Uneasy Money Time to Move On – But Not Before I Explain (Definitively) What it all Means « Uneasy Money
    24. January 2012 at 09:41

    […] Signs of fatigue are clearly evident and multiplying rapidly, so we had all better figure out and start executing our exit strategies from this convoluted, and at times acrimonious, debate about consumption smoothing. Things got started (two whole weeks ago!) when Simon Wren-Lewis picked on Robert Lucas for the following statement: […]

  34. Gravatar of Ron Ronson Ron Ronson
    24. January 2012 at 10:05

    I’m just trying to think this through – does the following make sense ?

    The Keynesian model assumes that I is unrelated to the level of Y and (other things being equal) will remain constant as either Y and S change. This means that if S falls for a given level of AD then the model dictates that Y must increase (via additional spending presumably from cash holding) until Y is at a sufficiently high level to bring I and S back into line. As this only requires additional spending (not inventory adjustments or any changes affecting real physical goods) this new equilibrium would be reached quite quickly. The amount of additional spending to reach the new equilibrium will be entirely dependent upon the consumption function – from where the multiplier is derived

  35. Gravatar of tom tom
    24. January 2012 at 10:47

    Scott, it is a matter of definition because not all of unsold stuff must be of investment type. But even if it is it can be bought by government. So from where come conclusion that consumers must reduce their spending by more than they reduce their consumotion? If they reduce spending by exactly the amount of consumption we still have three things being true: consumption smoothing, positive multiplier and maintained savings identity.

  36. Gravatar of D R D R
    24. January 2012 at 11:13

    “The Keynesian model assumes that I is unrelated to the level of Y and (other things being equal) will remain constant as either Y and S change.”

    1) What are you defining as “the Keynesian model”?

    2) The rest is unclear

    One may write down a Keynesian model where I does not change– just assume Y=C+G. In such a (closed economy) model, S = (Y-T-C)+(T-G) = Y-C-G = 0 always and everywhere.

    One may also write down Keynesian models where I does change. I may change in response to changes in Y, and/or Y may change in response to changes in I.

    So I don’t know how to answer your question.

  37. Gravatar of D R D R
    24. January 2012 at 11:19

    “DR, It’s worse, in the Keynesian model there’s no change in consumption (in the BBM case) and hence no consumption smoothing. See his newer comments.”

    I will say it again. No-*aggregate*-consumption-smoothing does not imply no-consumption-smoothing.

  38. Gravatar of Morgan Warstler Morgan Warstler
    24. January 2012 at 12:22

    Scott,

    This is not even close to correct:

    “Rather it reflects deferred consumption. The $200,000 received at 65 is exactly equal in present value to the $40,000 saved today.”

    So many assumptions, and not even close to the way I view entrepreneurial (job growth) investing.

    My REAL example is:

    You invest $100K, one year later you sell for $1.1M, you have a capital gain of $1M.

    Show me where the $1M was taxed.

  39. Gravatar of D R D R
    24. January 2012 at 13:09

    Look, regarding Wren-Lewis’ first statement regarding the bridge, there are three ways to interpret what he wrote.

    1) The bridge is considered in isolation and the private sector does not respond to the tax (real or anticipated in the future)

    1a) There is no consumption smoothing here. C+0 G+100 T+0. [I’m including this possibility out of completeness.]

    1b) There is consumption smoothing here. C+25 G+100 T+0. [In this case, Wren-Lewis is “wrong” because the spending multiplier is 1.25, not 1.0– that is, he understated his case. This is my understanding of Noah’s argument.]

    2) The bridge is NOT considered in isolation, and the private sector smooths consumption. C+0 G+100 T+100. [In this case, note that the bridge-builders are loaning money to the other taxpayers, so there is in fact actual consumption smoothing.]

    3) The bridge is not considered in isolation, but the private sector is not actually taxed– they merely anticipate the tax by attempting to smooth consumption. Say that 1 percent of the people split the bridge money, but everyone gets taxed. Then the 99% try to reduce their consumption by a collective 19.80 while the 1% try to increase their consumption by a collective 19.80. There is no change in aggregate consumption. Actual income rises, and disposable income along with it, but everyone is acting as though disposable income is unchanged (in preparation for the tax.) So, C+0 G+100 T+0, but everyone is smoothing even though there is no change in consumption. Actual private saving does rise, but it takes the form of money-hoarding and is offset by the actual decline in public saving.

    Now… what happens in each case with the tax?

    1) We are considering the tax in isolation, but there is definitely smoothing here. C-25 G+0 T+100. Thus, in combination,

    1a) C-25 G+100 T+100. Consumption and disposable income both fall by the same amount. This leads to two possibilities:

    1a(i)) This the end of the story. The odd result is is NOT because there is no smoothing, but rather because the assumption underlying this scene is that agents smooth tax hikes, but NOT income. Again, strange, but offered out of completeness. We finish with a BBM of 0.75.

    1a(ii)) Once the whole thing is put together, the private sector does eventually smooth. Seeing that their consumption has fallen by as much as disposable income, they increase consumption, which increases disposable income, etc in a virtuous cycle until they have in the end not reduced consumption at all. The final result– as a *consequence* of *additional* smoothing is C+0 G+100 T+100. [That is, Wren-Lewis was “wrong” because he left out the additional smoothing and understated his case.]

    1b) Put together, C+0 G+100 T+100.

    2) Nothing changes, as the tax was already included.

    3) The imposition of the tax has no effect other than reducing disposable income. Because the private sector hoarded money to pay for the tax and has already adjusted for the tax, in the end we still wind up with C+0 G+100 T+100.

    To summarize:

    1a(i)) Wren-Lewis made a weird assumption which threw folks for a loop. BBM=0.75

    1a(ii)) Wren-Lewis understated his case by ignoring the additional consumption resulting from the *combining* of the bridge and the tax effects. BBM=1.0.

    1b) Wren-Lewis understated his case by ignoring the additional consumption resulting from the bridge income. BBM=1.0

    2) There is consumption smoothing as bridge-builders loan money to taxpayers. BBM=1.0

    3) There is consumption smoothing in the form of money-hoarding, which is undone when the tax is actually imposed. BBM=1.0

  40. Gravatar of Major_Freedom Major_Freedom
    24. January 2012 at 18:58

    ssumner:

    I said: “I challenged this by saying that aggregate demand is measured in terms of money spending, not real goods accumulating in warehouses.”

    You said: “Money is also spent on inventories, isn’t it?”

    Money is spent on inventories yes, but no ADDITIONAL money is spent on accumulated inventory brought about by consumers reducing their consumption spending and accumulating cash.

    It would be like you making $1000 in investment expenditures each year, and there is $1100 in consumption expenditures at the end of each year for your goods, but then one year, after you make your investment of $1000 in inventory, there is a reduction in consumer spending for your goods from $1100 down to $550. Consumers reduced their consumption down to $550 and accumulated $550 in cash at the end of the year, but that doesn’t mean you made $550 in additional inventory spending at that time, such that AD remained unchanged.

    The inventory spending was already included in the activist “I” in the past that is distinct from consumer spending changes in the present.

    “And AD doesn’t include money spent in intermediate goods.”

    I thought AD includes spending for EVERYTHING, hence the “aggregate.” If a good is intermediate, but is sold, then isn’t that a capital good and isn’t the spending for it included in “I”? Isn’t it the case that buyers of those intermediate goods are investors, and thus should be considered to have made an investment expenditure, thus contributing to AD?

    As long as an expenditure earns a business its revenues, and hence potential profits, that is a place where people are spending their money on “things”, and thus should be included in AD.

    Sounds like you’re defining aggregate demand as something that only includes consumer goods. But then what is the “I” in C+I+G if you’re omitting intermediate, i.e. capital, goods?

    “So money spent is not a good way to estimate AD.”

    This is getting weird. If money spent isn’t aggregate demand, then what is the “demand” in aggregate demand?

    Isn’t NGDP an aggregate spending measure akin to AD, is just a more encompassing measurement of total money spent?

  41. Gravatar of Ron Ronson Ron Ronson
    25. January 2012 at 06:33

    I was very late to this discussion and just want to make sure I have got it.

    The issue is what effect the reduction in savings caused by consumption smoothing will have on investment and on AD, right ?

    If that is the right question then the surely the answer depends upon how Investment will respond to the lower pool of (new and old) savings. In a healthy economy then the reductions in savings will cause a reduction in I but not to the full amount of the fall in S because the higher ensuing interest rates would cause more savings to be forthcoming (I think Scott make this point early on)

    However in an unhealthy economy where the loans market is not clearing even at low interest rates then a reduction in S would lead to a small (or no) reduction in I and overall AD to rise. I assume this is is the scenario that Wren-Lewis envisioned.

  42. Gravatar of ssumner ssumner
    25. January 2012 at 07:07

    Ron, You said;

    “The Keynesian model assumes that I is unrelated to the level of Y and (other things being equal) will remain constant as either Y and S change.”

    No, the Keynesian model assumes S=I. The rest of your comment sounds about right.

    You should have said I is unrelated to changes in desired saving.

    Tom, You said;

    “If they reduce spending by exactly the amount of consumption we still have three things being true: consumption smoothing, positive multiplier and maintained savings identity.”

    No, if they reduce spending on C+I by the same as they reduce spending on C, then there has obviously been no consumption smoothing.

    DR, Wren-Lewis was CLEARLY talking about aggregate changes, otherwise his “proof” that Cochrane was wrong would make no sense.

    Morgan, Imagine a no tax scenario where you earn $200,000, invest if for 10 years, then end up with $2.2 million. Now lets put a 50% wage tax in place. Now you have $100,000 after paying the wage tax, you invest it for ten years and have $1.1 million. The tax effectively cuts in half the future value of your investment, even with only a wage tax. But our system is worse, taxing your wage income, and then taxing your capital income. So now you must pay 1/2 of the million dollar gain, leaving you only $600,000, compared to $2.2 million in the no tax case. That’s the sense in which your future consumption is double taxed. You can consume barely a fourth of the amount you could consume without taxes.

    DR, You continue to claim consumption smoothing where there is no aggregate consumption smoothing. So even if those cases were true, they’d have no bearing on Cochrane’s argument.

    Major freedom, You said;

    “Sounds like you’re defining aggregate demand as something that only includes consumer goods. But then what is the “I” in C+I+G if you’re omitting intermediate, i.e. capital, goods?”

    I’d suggest reading an intro to econ textbook. Investment goods are not intermediate goods. We are just talking past each other. Lots of my commenters seem to think it’s possible to understand macroeconomics just through common sense, without having studied the concepts. Not true. You are misusing all sorts of concepts in national income accounting.

  43. Gravatar of ssumner ssumner
    25. January 2012 at 07:08

    Ron, See my earlier reply to you.

  44. Gravatar of tom tom
    25. January 2012 at 10:29

    Scott,

    I thought that definition of consumption smoothing was that consumption C falls by less than income after tax Y-T. So why do you claim that “if they reduce spending on C+I by the same as they reduce spending on C, then there has obviously been no consumption smoothing”? It doesnt make any sense to me.

  45. Gravatar of D R D R
    25. January 2012 at 10:42

    Scott,

    “Wren-Lewis was CLEARLY talking about aggregate changes, otherwise his ‘proof’ that Cochrane was wrong would make no sense.”

    He was clearly talking about the aggregate changes in the third sentence. If he was talking about aggregate changes in the first sentence, what, I ask you, was the point of the other two sentences?

    “You continue to claim consumption smoothing where there is no aggregate consumption smoothing. So even if those cases were true, they’d have no bearing on Cochrane’s argument.”

    This makes no sense. No-aggregate-consumption-smoothing does not preclude actual acts of smoothing consumption. Proof: If the government takes from A and gives to B, then they can both smooth consumption with a loan of money from B to A. Yet both aggregate consumption and aggregate disposable income are unchanged.

  46. Gravatar of D R D R
    25. January 2012 at 11:13

    “I challenged this by saying that aggregate demand is measured in terms of money spending, not real goods accumulating in warehouses. Unless you can show a “replacement” spending for the decline in consumer goods spending, then you cannot say AD remains unchanged. AD will fall because there is no additional spending that replaces the decline in consumer spending.”

    First of all, AD is basically a schedule of purchases, while Y is a quantity produced. If you want to think of AD as a single quantity– i.e., shorthand for aggregate quantity demanded, it doesn’t have to balance exactly aggregate quantity produced at any time. That is, inventories may change unexpectedly. (Planned changes in inventories sensibly do count as “demand”)

    So if there is an unexpected increase in final sales, this does swap a bit of I for C in the national accounts. However, the fall in inventories being unexpected, the seller now desires a larger inventory than remains, and so demands one more unit of production.

    Thus, the unexpected increase in final sales does not *immediately* increase Y, but does *induce* an increase in Y.

  47. Gravatar of Major_Freedom Major_Freedom
    25. January 2012 at 16:30

    ssumner:

    I said: “Sounds like you’re defining aggregate demand as something that only includes consumer goods. But then what is the “I” in C+I+G if you’re omitting intermediate, i.e. capital, goods?”

    You said: “I’d suggest reading an intro to econ textbook. Investment goods are not intermediate goods. We are just talking past each other. Lots of my commenters seem to think it’s possible to understand macroeconomics just through common sense, without having studied the concepts. Not true. You are misusing all sorts of concepts in national income accounting.”

    Holy mother of pearl, more condescension, more fallacies, and more insinuation straw men. This time I have to open up intro textbooks, and believe that I have been arguing not from studying the concepts but from “common sense”, and believe I am allegedly “misusing” concepts in national income accounting. Yeah!

    I never argued investment goods are intermediate goods. I said intermediate goods are included in “I” in C+I+G. That is why I asked what do you define “I” to be if you’re EXCLUDING intermediate goods that every intro economics textbook includes in “I” if they are bought, i.e. invested in by business firms. “I” does in fact include intermediate goods, in C+I+G, if there is a demand for them in the money spending (investment) sense. For example, if an auto firm makes an investment in an intermediate good like steel sheet, then that would count towards “I” in C+I+G. You ask me to check an intro textbook? How about you check any intro textbook and realize it shows intermediate goods can be used, AND BOUGHT AND SOLD. If a firm buys an intermediate good for the purposes of making subsequent sales, that is an addition to “I” in C+I+G. The “I” in C+I+G is gross investment. “I” includes replacement purchases plus net additions to capital assets plus investments in inventories. If the auto firm buys an intermediate good like steel sheet, that is an addition to their capital assets. To say “I” does not include intermediate goods is to ignore a huge aspect of what actually constitutes “I”. You’re conflating accounting for value on the one hand, and aggregate demand in money terms on the other, and failing to understand what is included in “I”, what is included in AD, then accusing me of not knowing national income accounting, not having read intro books, and not studying concepts to boot! Oh the hilarity.

    Today I learned that Sumner refuses to answer criticisms and accuses his critics of not being educated. Wonderful. This coming from someone who believes such pearls of wisdom such as:

    “Money spending is not a good measure for aggregate DEMAND.”

    “Credit is not money” (my personal favorite).

    “Credit expansion could not have fueled the housing bubble because you can’t go into Wal-Mart and buy anything with bonds.” (that one made me spit out my coffee, requiring me to clean my PC)

    “Inventory build up caused by consumers decreasing their consumer spending and accumulating cash balances, contributes an addition to aggregate demand” (this one proved to me you have no clue what planet you’re on).

    This is breathtaking. I am seriously considering whether I should treat this blog as educational or for entertainment purposes.

  48. Gravatar of Major_Freedom Major_Freedom
    25. January 2012 at 17:01

    D R:

    First of all, AD is basically a schedule of purchases, while Y is a quantity produced.

    If you define AD in terms of purchases, and Y in terms of quantity of goods produced, then that means they are incommensurable units, and so you can’t even say “AD lags Y”. You cannot say “$14 trillion lags [30,000 cars + 500,000 t-shirts + 40,000 PCs, + …]”.

    Incommensurable units cannot be subtracted or added to each other to result in a third quantity.

    The only way that “AD lags Y” can be coherent is if they are of the same unit. If one uses dollars as the unit, then what “AD lags Y” means is that “Aggregate spending lags [supply of good A multiplied by some given “value” or price of A + supply of good B multiplied by some given “value” or price of B + supply of good C multiplied by some “value” or price of C + …]”.

    In other words, the only way that “AD lags Y” can exist is if aggregate spending is less than the spending that would be required to “clear the market” at prevailing ask and bid prices.

    But my view is that prevailing ask and bid prices are DETERMINED by supply and demand, so what is happening when “AD lags Y” is that owners of resources and labor are withholding their individual supplies and refusing to sell them at the lower prices which would clear the market given prevailing aggregate demand.

    The flaw in the Keynesian view is that they attack the price system based on subjective utility and preferences, and call for “AD to equal Y” on the side of demand, by way of government bringing about inflation and spending. This “solution” usurps the sovereign consumer in determining which firms are profitable and which firms are unprofitable. In other words, Keynesianism transforms the economy from a consumer driven economy, to a politically driven economy.

    In a sane world, we would all be producing and working and striving to achieve ends (consumption). In the insane Keynesian world, the ends become inhuman aggregate statistics controlled by the government, which is a cover for political ends.

    If you want to think of AD as a single quantity- i.e., shorthand for aggregate quantity demanded, it doesn’t have to balance exactly aggregate quantity produced at any time. That is, inventories may change unexpectedly. (Planned changes in inventories sensibly do count as “demand”)

    I consider aggregate demand to be a single quantity yes, but not aggregate quantity demanded. I consider aggregate demand as aggregate expenditures, or, if you will, the “intro macroeconomics textbook” definition.

    Planned inventories do not count as demand. The money expenditures associated with planned inventories count as demand.

    So if there is an unexpected increase in final sales, this does swap a bit of I for C in the national accounts. However, the fall in inventories being unexpected, the seller now desires a larger inventory than remains, and so demands one more unit of production.</i

    "Demands" one more unit of production? If there is an investment expenditure in inventory, then that counts towards "I". If there is merely a delay in the sale of inventory, brought about by a reduction in consumer spending, then the physical inventory accumulation is NOT an extra addition to "I" that is apart from the expenditure already associated with the inventory.

    If you make an expenditure that increases your inventory, then that is added to "I". If I buy that good, that is added to "C". C+I. If on the other hand you invest in the inventory, but I don't buy it, then "C" is reduced, but there is no extra addition to "I" that is besides the "I" you already made. So it will be I, and not, say, 2I.

    Thus, the unexpected increase in final sales does not *immediately* increase Y, but does *induce* an increase in Y.

    Only if the firm makes actual expenditures for more inventory. But if the accumulating inventory is due to a lack of consumer demand, then then that fact alone cannot be said to increase “I”. Only if there are more expenditures in inventory can you increase “I”.

  49. Gravatar of D R D R
    25. January 2012 at 17:27

    Major,

    I don’t know if you didn’t understand what I wrote, or don’t know what you are talking about, but I’m beyond worrying about it. There’s just far too much to unpack.

    For what it is worth:
    “In macroeconomics, aggregate demand (AD) is the total demand for final goods and services in the economy (Y) at a given time and price level. It is the amount of goods and services in the economy that will be purchased at all possible price levels.” — Wikipedia http://en.wikipedia.org/wiki/Aggregate_demand

  50. Gravatar of ssumner ssumner
    26. January 2012 at 10:18

    Tom, I’m assuming a balanced budget multiplier, so S&I must fall any time Y-T falls.

    DR, Yes, but in that case aggregate C doesn’t fall. So his “proof” has no bearing on Cochrane’s argument.

    Major Freedom, You said;

    “But then what is the “I” in C+I+G if you’re omitting intermediate, i.e. capital, goods?”

    and:

    “I never argued investment goods are intermediate goods. I said intermediate goods are included in “I” in C+I+G.”

    I can’t make heads or tails of anything you say. Maybe someone else can answer your question. GDP includes only final goods.

  51. Gravatar of D R D R
    26. January 2012 at 10:25

    “DR, Yes, but in that case aggregate C doesn’t fall. So his “proof” has no bearing on Cochrane’s argument.”

    I’m sorry… what? Please elaborate.

  52. Gravatar of ssumner ssumner
    27. January 2012 at 06:40

    DR, This will also be my response to your “In a perfect world comment”

    You seem to be assuming that WL is doing nothing more than showing GDP might change. But everyone including Cochrane agrees with the following:

    “Suppose C+I+G went up from 300 to 400. In that case GDP would have risen.”

    Cochrane would have agreed–if C+I+G rises, then GDP rises. His claim was that GDP would not increase. Everyone agrees that if it does increase then it does increase. WL was making a more ambitious claim, that it had to increase because of consumption smoothing. But in the example you cite you get exactly the same result (assuming a BBM of 1) with or without consumption smoothing. Yes, if you assume I is fixed, and if you assume Y goes up by as much as G, then C will be constant. But you’ve simply assumed an answer, you haven’t shown anything at all. And you certainly haven’t used consumption smoothing to prove Cochrane wrong.

    I assumed WL was actually trying to show something, not just stating a tautology.

  53. Gravatar of D R D R
    27. January 2012 at 09:26

    Scott,

    I don’t know if you don’t understand my argument or you deliberately misrepresent it. Just because I arrive at a conclusion does not imply I assumed a conclusion. Since you still don’t seem to understand that there is a nontrivial argument here, I’ll try once again to explain how a closed economy with no investment

    Let us break out bridge-building in the national accounts

    Y = C + G = (Cb+Co) + (Gb+Go)

    where “b” is for “bridges” and “o” is for other goods and services.

    1) Assume there is slack in the economy. Let there be underutilazation of capacity in bridge-building of 0<Ub100-Ub. (That is, there are at least 100 units of slack overall.)

    2) Assume further that increased demand for a good or service is accommodated by expanded utilization, in as far as that is possible– but that it must otherwise crowd out production of the same.

    Now, if Gb rises by 100, the immediate impact is for Cb to fall by 100-Ub. That is, Cb+Ub-100 Co+0 Gb+100 Go+0, so Y+Ub.

    Now, anticipating a tax hike of 100, the private sector figures their disposable income not to have risen, but rather perceive that Yd is falling by 0<100-Ub<100.

    3) Assume that consumers smooth consumption by a factor of 5. That is, if a consumer perceives an increase in disposable income of X, the consumer will increase consumption by X/5.

    4) Though this complicates matters, assume that consumers attempt to smooth consumption as individuals. [We can alternately go down a game-theory path which leads to the same result, but let us put that aside.]

    Now, perceiving an fall in disposable income of 100-Ub, consumers will want to reduce total consumption by 0<20-Ub/520-Ub/5 as a result of crowding-out so consumers want to (now) raise their consumption relative to the depressed level. Of course, there is no slack left in bridge-building, so this means that consumers must substitute other for other goods and services. That is, given the perceived fall in disposable income of 100-Ub, they want Co to rise by (Ub/5-20)-(Ub-100) = 4*(100-Ub)/5. Note that 4*(100-Ub)/5<100-Ub, so the capacity to meet all this additional demand does exist. As they proceed along this path, they find…

    Cb+(Ub-100) Co+4*(100-Ub)/5 Gb+100 Go+0, so that Y+100-(100-Ub)/5 and perceived Yd+(Ub-100)/5.

    Now, we *could* end the story here, with total consumption falling by 0<(100-Ub)/5<20. This would imply a BBM of 0<1-(1-Ub/100)/5<1.

    However, consumers' perceived Yd has risen by more than it had before they adjusted their consumption, and so now increase their consumption so that Co+24(100-Ub)/25. Again, there is slack to accommodate this increase, so we find a BBM of 0<1-(1-Ub/100)/25<1

    Et cetera. In the end, the increase in Co approaches 100-Ub and the BBM approaches 1.

    Now, did I just state a tautology? Or did I model the economy in a particular way and arrive at a particular result? Yes, I assumed that producers would (if possible) expand output to meet increased demand for their goods and services. Yes, I assumed that a certain amount of capacity to do so existed. And yes, to arrive at this result I assumed that consumers would smooth consumption.

    To cite Cochrane, "you have to tell us which of the "ifs" you disagree with. That discipline changes everything."

  54. Gravatar of D R D R
    27. January 2012 at 10:00

    I’m afraid some parts of my post vanished.

    For example:

    1) Assume there is slack in the economy. Let there be underutilazation of capacity in bridge-building of 0 “LT” Ub “LT” 100, and underutilization ofcapacity in building of other goods be Uo “GT” 100-Ub. (That is, there are at least 100 units of slack overall.)

  55. Gravatar of ssumner ssumner
    28. January 2012 at 07:40

    DR, So you are assuming that WL’s argument is based on a distinction between perceived income and actual income? Two problems. He didn’t say that, and it wouldn’t refute Cochrane.

  56. Gravatar of ssumner ssumner
    28. January 2012 at 07:42

    DR, BTW people who are extremely inept at explaining their ideas shouldn’t direct insults against those they are trying to convince. Take a look in the mirror, or show another economist your “proofs” to see if anyone else can understand them better than I can.

  57. Gravatar of D R D R
    28. January 2012 at 08:09

    “DR, So you are assuming that WL’s argument is based on a distinction between perceived income and actual income? Two problems. He didn’t say that, and it wouldn’t refute Cochrane.”

    No. I am absolutely not saying that. I am merely trying to explain to you the process of consumption smoothing so that you might understand Wren-Lewis’ argument.

    I am saying that that the process of consumption smoothing can be thought of in this fashion– that agents continually try to smooth consumption, and this results in changes to their disposable incomes which then result in further smoothing.

    I am also saying that this process, if *not* followed through to the end, results in a positive BBM. I am also saying that this process, if followed through to the end, results in a BBM of 1.0. I am suggesting that a game-theoretic approach to consumption smoothing would result (immediately) in a BBM of 1.0.

    I am saying that you choose to read Wren-Lewis is a most uncharitable way, and that even despite that uncharitable read his argument is only “wrong” to the extent that he understated his case.

    Now, what is your criticism of Wren-Lewis again?

  58. Gravatar of D R D R
    28. January 2012 at 12:03

    Stupid, stupid, stupid me.

    If G and T both rise by some fixed amount then consumption smoothers DO NOT WANT consumption to fall.

    Ideally, from the consumer’s perspective,

    C +dC
    G +dG
    T +dG
    Y +dC+dG
    Yd+dC

    The fact that dC = dYd means that consumers WANT dC = 0.

    That does not mean they can get it. If there is any slack in the economy, but not enough to support the increase in G, then consumers are stuck MINIMIZING the fall in dC. This results in a positive BBM and an overall fall in consumption.

  59. Gravatar of ssumner ssumner
    29. January 2012 at 12:42

    DR, I’m just not following your argument. If C falls than S must fall.

  60. Gravatar of D R D R
    29. January 2012 at 13:55

    Why? Cannot Y fall, leaving S unchanged?

  61. Gravatar of ssumner ssumner
    30. January 2012 at 06:07

    DR, Not with consumption smoothing. In that case C and S both fall, or neither fall.

  62. Gravatar of D R D R
    30. January 2012 at 07:01

    So it is your argument that in a closed economy with no investment and balanced government budgets and consumption smoothing…

    If G and T rise by the same amount, then C does not change.

    Do I have that right?

  63. Gravatar of ssumner ssumner
    31. January 2012 at 14:04

    DR, In an economy with no investment and a balanced budget and no trade you can’t have consumption smoothing, it’s impossible.

  64. Gravatar of D R D R
    31. January 2012 at 14:13

    Why not? Cannot private agents lend to one another?

  65. Gravatar of ssumner ssumner
    1. February 2012 at 17:16

    DR, Obviously, But it’s obvious that WL’s argument referred to aggregate consumption, otherwise it would have been a complete non sequitor–having no bearing at all on Cochrane’s argument.

  66. Gravatar of D R D R
    1. February 2012 at 18:04

    Okay, with that out of the way, it *is* your argument that in a closed economy with no investment and balanced government budgets and consumption smoothing that If G and T rise by the same amount, then C does not change.

    Now, why would that be a non-sequitur? That reads like a very direct argument.

  67. Gravatar of ssumner ssumner
    2. February 2012 at 13:02

    DR, Yes, under all those assumptions C would not change. But that has no bearing on Cochrane’s argument. WL said Cochrane was wrong because C would change less than disposable income.

  68. Gravatar of D R D R
    2. February 2012 at 13:49

    Actually, Wren-Lewis does NOT say that C changes by *less* than disposable income. He says it falls by less than X (the amount spent on the bridge.)

    I quote, “their spending at time t will fall by much less than X.”

    So what is wrong with his argument?

  69. Gravatar of ssumner ssumner
    3. February 2012 at 18:31

    DR, But combine that with consumption smoothing and it clearly means C fell less than Y-T. What else could it mean?

  70. Gravatar of D R D R
    3. February 2012 at 19:09

    Well, for one it could mean C fell by zero because Y-T fell by zero.

  71. Gravatar of ssumner ssumner
    4. February 2012 at 18:43

    DR, If WL means “spending doesn’t fall” when he says “spending falls” then he should say so and I’ll respond Until then I’m not going off on wild goose chases as to what he might have meant. The meaning was quite clear to me, and I think to most readers.

  72. Gravatar of D R D R
    4. February 2012 at 21:00

    Scott,

    The important point here is that even with *your* interpretation of what Wren-Lewis said, the obvious criticism is that he said spending fell, when it plainly doesn’t.

    By *your* interpretation of what he wrote, then, he understated his case.

  73. Gravatar of ssumner ssumner
    6. February 2012 at 06:45

    DR, No, I believe he believed exactly what he said, and forgot that if consumption fell then with consumption smoothing investment also had to fall. I believe he made a mistake, sort of analogous to the error he accuses Cochrane of making.

  74. Gravatar of D R D R
    6. February 2012 at 09:32

    Wow.

    So, given a complete explanation which perfectly jibes with everything Wren-Lewis says, but possibly errs in understatement, you choose to ignore it and “believe” he made a more error in overstatement by ignoring investment.

    And even though he never mentions investment, you “believe” this for the purpose of “not going off on wild goose chases as to what he might have meant.”

    Oh, and you do so even though this bizarre interpretation doesn’t actually pin down the level of consumption or investment at all.

    Oh, and what is your behavioral explanation for the fall in investment?

    Absolutely incredible.

  75. Gravatar of D R D R
    6. February 2012 at 10:04

    To be clear,

    You might argue that Wren-Lewis did not completely think through his model and that as a matter of fact in his model consumption does not fall.

    Instead, you avoid “going off on wild goose chases” by asserting that Wren-Lewis did not completely think through his model and as a matter of fact that investment must have fallen.

  76. Gravatar of ssumner ssumner
    7. February 2012 at 07:36

    DR, No he said consumption fell and your “explanation” has it not fall. So it’s not consistent with what he said.

    In addition his example claimed consumption smoothing proves Cochrane wrong, and in your “explanation” consumption smoothing exists, but plays no role in proving Cochrane wrong.

  77. Gravatar of D R D R
    7. February 2012 at 08:18

    Scott,

    *You* say he said it fell. Therefore, *you* might say he was wrong to say it fell. This is a perfectly legitimate criticism. I don’t understand why you have to go bringing investment into it.

    If consumption does not fall, then Cochrane is wrong. You get a BBM of 1.0.

  78. Gravatar of D R D R
    7. February 2012 at 08:28

    “In addition his example claimed consumption smoothing proves Cochrane wrong, and in your “explanation” consumption smoothing exists, but plays no role in proving Cochrane wrong.”

    If no investment and yes consumption smoothing then no fall in consumption, so BBM of 1.0.

    If no investment and maybe consumption smoothing then… who knows?

    So yes, consumption smoothing is an important part of the example which disproves Cochrane.

  79. Gravatar of ssumner ssumner
    8. February 2012 at 14:53

    DR, No if consumption doesn’t fall then Cochrane might still be right. Suppose I falls as much as G rises. I know it doesn’t in the Keynesian model, but Cochrane thinks the Keynesian model is wrong.

    If no investment then the BBM is one using Keynesian assumptions (with or without consumption smoothing)

    If you aren’t using Keynesian assumptions, then there is no reason to assume no change in investment. Hence WL’s “proof” would make no sense. His proof only makes sense if he assumes I doesn’t change. But in that case the BBM is one with or without consumption smoothing, assuming the Keynesian model.

  80. Gravatar of D R D R
    8. February 2012 at 17:14

    Scott,

    “if consumption doesn’t fall then Cochrane might still be right. Suppose I falls as much as G rises. I know it doesn’t in the Keynesian model, but Cochrane thinks the Keynesian model is wrong.”

    This is wrong. Cochrane claimed a “theorem” that showed a BBM of zero, but he did not address the case where G rises. Therefore, I may prove him wrong merely by saying the multiplier on G is one million. You may argue that my assumption is idiotic, but you may not argue that my assumptions contradict Cochrane, because he simply never addressed that case.

    “If no investment then the BBM is one using Keynesian assumptions (with or without consumption smoothing)”

    What do you mean by “Keynesian assumptions?”

    “If you aren’t using Keynesian assumptions, then there is no reason to assume no change in investment.”

    What? Are you asserting there are no modes without investment? I don’t see where Cochrane discussed investment in his “theorem” either. Furthermore, I have shown previously that there is no need for investment to change even if you include it in the model. Do we need to go over that yet again? I’ll be happy to do it.

    “Hence WL’s “proof” would make no sense. His proof only makes sense if he assumes I doesn’t change.”

    1) I can make sense of I not changing even if C does actually fall. But let’s not get bogged down in further complications.

    2) His example makes perfect sense if you understand that zero is less than any positive number.

    3) Even if you don’t care about the fact that zero is less than any positive number and you insist that Wren-Lewis says that actually C fell, then it makes sense to argue that Wren-Lewis was WRONG ABOUT C FALLING.

    “But in that case the BBM is one with or without consumption smoothing, assuming the Keynesian model.”

    That depends on what you mean by “Keynesian model.” If I do not assume consumption smoothing, I need some other assumption in its place. Otherwise, I don’t know what happens to C. But what difference does it make if some “Keynesian model” shows a BBM of 1.0? That doesn’t make the first model wrong.

  81. Gravatar of D R D R
    8. February 2012 at 17:24

    Oh what the hey.

    Say person A is employed and person B (a widget-builder) is unemployed. The government taxes A $100 and pays person B $100 to build a widget.

    The (permanent) disposable income of person A has fallen by $100, and person A would now like to reduce consumption by $20.

    The (permanent) disposable income of person B has increased by $100, and person B would now like to increase consumption by $20.

    Thus, person B loans person A $80, and everyone is happy.

    In the aggregate, C is unchanged, I is unchanged, G and T each rise by $100. The BBM is 1.0.

    If Wren-Lewis implied the BBM to be *less* than 1.0, then he was wrong.

    Are we in agreement yet?

  82. Gravatar of ssumner ssumner
    10. February 2012 at 09:16

    DR, You keep arguing that there is a world where WL’s conclusion is correct. I agree.

    You keep arguing that Cochrane is wrong. I agree.

    You aren’t giving me any reason to think WL’s specific argument wasn’t wrong. The meaning of “C fell” is very clear, it cannot mean “C didn’t fall.” And if he really did claim C didn’t fall, his proof would have everyone scratching their heads. Just re-write what WL said so that because of the tax increase, and because of consumption smoothing, consumers would not change C at all. And Thus Cochrane is wrong. What do you think people would make of that argument? I say they’d regard it as nonsensical. They’d say “what does that have to do with whether GDP changed?”

  83. Gravatar of ssumner ssumner
    10. February 2012 at 09:17

    I should add, they’d also ask WL “how does consumption smoothing prove C doesn’t change at all?”

  84. Gravatar of D R D R
    10. February 2012 at 10:24

    Scott,

    “how does consumption smoothing prove C doesn’t change at all?”

    Again, I repeat myself. If Y=C+G and G and T rise by the same amount, then Yd=C+G-T=C. If we assume (successful) smoothing of consumption, then C cannot change. Proof: If C changes, then Yd changes by exactly the same amount and consumption is not smoothed. Therefore, C cannot change.

    “You aren’t giving me any reason to think WL’s specific argument wasn’t wrong. The meaning of “C fell” is very clear, it cannot mean “C didn’t fall.”

    Point 1: This is only confusing if you keep misquoting him. Do not put “C fell” in quotes when referring to Wren-Lewis. The actual quote is “spending at time t will fall by much less than X.” (Where there is a hidden assumption that X is positive.)

    Point 2: Even if you do not accept point 1, then his “error” is one of understatement.

  85. Gravatar of ssumner ssumner
    11. February 2012 at 07:12

    DR, So you think WL was trying to show Cochrane was “wrong” by using a model where he simply assumes there is no saving! And even worse, a consumption smoothing model where there was no saving! What kind of proof is that? The consumption smoothing would then prove absolutely nothing, as aggregate consumption wouldn’t change!

    When most sensible people hear the term of “consumption smoothing,” they immediately think of saving as a shock absorber, so why in the world assume no saving?

    Yes, there are a set of assumptions where the BBM is one, you can just assume that. But don’t think it has any bearing on WL’s argument. If you assume no saving, then obviously the BBM will be one. But why assume that? And what does that have to do with WL’s argument. If there’s no saving, the the BBM is one with or without consumption smoothing. Surely WL was trying to actually PROVE something.

  86. Gravatar of D R D R
    11. February 2012 at 08:27

    “What kind of proof is that? The consumption smoothing would then prove absolutely nothing, as aggregate consumption wouldn’t change!”

    Wait… if consumption doesn’t change then it *doesn’t* disprove Cochrane? I must assume this is not what you mean.

    “And even worse, a consumption smoothing model where there was no saving! What kind of proof is that?”

    That is just not true. There is saving in the model. Agents with increased disposable incomes smooth their consumption by lending to agents with reduced disposable incomes. And it would make no difference if we did include investment in the model, because agents would smooth consumption in exactly the same way. Taking out investment just makes it easier to see how it works.

  87. Gravatar of ssumner ssumner
    12. February 2012 at 09:22

    You said;

    “Wait… if consumption doesn’t change then it *doesn’t* disprove Cochrane? I must assume this is not what you mean.”

    Wrong, investment might fall to offset the higher G. But then you’ll reply that you or WL assumes investment doesn’t change. Fine. So here’s WL’s “proof” according to you.

    WL: “Assume G increases. Then assume no change in investment. Then assume C doesn’t change. In that case C+I+G must rise.”

    How is that a “proof” of anything? It’s stating a mathematical relationship between components of GDP. And how is that “proof” in any way related to consumption smoothing? Believe me, Cochrane knows that C+I+G =GDP.

    On your second point; I meant there is no change in aggregate saving in the model. Obviously one person can lend to another.

  88. Gravatar of D R D R
    12. February 2012 at 10:13

    Scott, that’s patently untrue. Making assumptions which lead to a conclusion is not the same as assuming the conclusion. You don’t get to keep arguing that anyone is assuming a conclusion by ignoring the argument, which is that consumption smoothing *disallows* a change in C.

    In any case, I refer you to https://www.themoneyillusion.com/?p=12819#comment-133752 for an example which includes Y=C+I+G, consumption smoothing and a BBM of 1.0.

    What is your argument against what I wrote? That I might fall? Yeah, and pigs may fly, but unless you give a reason why they should, then there is no reason to care.

    So please, make a criticism which is actually specific to what I wrote, or stop arguing there is anything wrong with it.

  89. Gravatar of ssumner ssumner
    13. February 2012 at 12:15

    DR, If WL had argued that consumption smoothing disallowed a change in C I would have responded “so what?” That proves nothing about Cochrane’s argument, as Cochrane never claimed anything different.

    Instead, WL did NOT argue that consumption smoothing disallowed a change in C. So it’s a moot point anyway, because WL chose not to make the argument you are making.

    There are certain ground rules to arguments, if you won’t accept them then I can’t change your mind. If I see WL making the arguments you are making I’d be glad to respond with new posts, but I doubt he’d open himself up to such obvious criticism.

  90. Gravatar of D R D R
    13. February 2012 at 14:23

    Scott,

    First, if Wren-Lewis argued that consumption smoothing disallowed a change in C then that itself does not disprove Cochrane– the fact that such a situation implies that Y rises in step with G does disprove Cochrane’s “theorem” that the BBM must be zero.

    Second, your unstated “ground rules” may as well require me to dig my grandmother out of her grave and play you piano.

    You claimed that Wren-Lewis “ignores the fall in investment” and “implicitly treated the [reduced] saving as if it just disappeared.”

    I don’t believe your claims are defensible. In support of my contention, I have offered an explanation of how Wren-Lewis may have done neither of these things you claim.

    You have offered no specific criticism pointing to the illegitimacy of my explanation, so *of course* I have no idea how to defend my position.

    I don’t think that asking for a specific criticism is beyond “certain ground rules to arguments.”

  91. Gravatar of ssumner ssumner
    14. February 2012 at 17:56

    DR, You said;

    “First, if Wren-Lewis argued that consumption smoothing disallowed a change in C then that itself does not disprove Cochrane- the fact that such a situation implies that Y rises in step with G does disprove Cochrane’s “theorem” that the BBM must be zero.”

    No it wouldn’t even if he had done so (which he didn’t.) Discussing G + C tells us nothing about Y.

    No you haven’t showed I was wrong. I don’t agree with your claim that WL was describing a case where saving in the aggregate was unchanged. If you can get WL to admit that’s what he was doing, then I’ll gladly write another post bashing him, as that argument would be totally implausible in my view.

  92. Gravatar of D R D R
    14. February 2012 at 21:35

    If I do this thing, you will bash him with another post. That’s awfully sweet of you.

    Will you please explain how my example is implausible?

  93. Gravatar of ssumner ssumner
    16. February 2012 at 11:55

    DR, Again, it’s not your examples that bother me–anything is possible, it’s his argument. You examples don’t match his claims.

  94. Gravatar of D R D R
    16. February 2012 at 12:08

    Right. I’m granting for the sake of argument that Wren-Lewis says that consumption falls by some nonzero amount.

    Is there anything to indicate (apart from the fact he is therefore “wrong”) that my example is not what he had in mind?

  95. Gravatar of ssumner ssumner
    17. February 2012 at 07:00

    DR, Obviously Wren-Lewis “had in mind” the right model, as did Cochrane (if you read his entire piece he talks about how fiscal policy might affect velocity.) They are both very smart guys. And obviously both people stated their arguments poorly. That was my whole point. Both temporarily forget an important point.

  96. Gravatar of D R D R
    17. February 2012 at 08:04

    As I have repeatedly stated, the question is what *kind* of error Wren-Lewis made.

    You have said repeatedly that Wren-Lewis erred because he MUST have forgotten something which specifically cut *against* his argument. I am saying he was “wrong” by “forgetting” something which strengthens his case.

    If you acknowledge that your argument was incomplete, we may move forward and argue over which alleged error makes sense to attribute to Wren-Lewis. But until you either find actual fault with my argument or accept that it is legitimate, then I must maintain that you are wrong to insist that Wren-Lewis forgot any “important point.”

    So, do you accept that my argument is consistent with Wren-Lewis apart from any claim that “C fell?”

  97. Gravatar of ssumner ssumner
    18. February 2012 at 10:30

    DR, You said;

    “You have said repeatedly that Wren-Lewis erred because he MUST have forgotten something which specifically cut *against* his argument. I am saying he was “wrong” by “forgetting” something which strengthens his case.”

    That’s defensible, but that’s also what Cochrane did, from a certain perspective. Elsewhere he said monetary policy drives NGDP. So he should have said fiscal stimulus can’t help if you hold NGDP (i.e. monetary policy) constant.

    As far as your second question, I really can’t understand the claim you are making. It seems to me that all you are doing is showing an argument that WL could have made but didn’t. But I was reacting to what I thought he was trying to argue, and I still think the most “friendly interpretation” of his mistake was that he temporarily forgot that consumption smoothing plus falling aggregate C means falling I.

    If his mistake lay elsewhere, then I would have been much tougher on him. Suppose he was really trying to show that Cochrane was wrong because C and I would not have fallen at all. Then I would have been much more critical, as I think that’s just assuming the answer.

  98. Gravatar of D R D R
    18. February 2012 at 12:16

    “It seems to me that all you are doing is showing an argument that WL could have made but didn’t.”

    By your interpretation, there is a flaw in his argument.

    “I still think the most “friendly interpretation” of his mistake was that he temporarily forgot that consumption smoothing plus falling aggregate C means falling I.”

    That can only be true if my interpretation is not valid. As you refuse to explain how my argument is not valid then I must maintain that you are wrong to insist that Wren-Lewis forgot any “important point.”

    “Suppose he was really trying to show that Cochrane was wrong because C and I would not have fallen at all. Then I would have been much more critical, as I think that’s just assuming the answer.”

    There you go AGAIN. For the umpteenth time: that C and I do not fall is not an assumption in my example, but FOLLOWS from consumption smoothing. So you are not addressing my argument at all.

  99. Gravatar of D R D R
    18. February 2012 at 12:49

    (Sorry for the weird first part….)

    “It seems to me that all you are doing is showing an argument that WL could have made but didn’t.”

    So what? The same criticism applies to everything you have written. If YOU say that investment must fall, then you are showing an argument that Wren-Lewis could have made but didn’t.

    By your interpretation, there is a flaw in his argument. If we accept that “C fell” the question is where in the argument lies the flaw. You say the flaw is that “C fell” implies I fell and yet Wren-Lewis didn’t mention a fall in I, and was thus negligent, or misunderstood his model, or was lying. I say the flaw was in saying that “C fell” period. I say this because consumption smoothing seems to IMPLY that C did not fall.

  100. Gravatar of ssumner ssumner
    19. February 2012 at 18:29

    DR, No, I don’t think consumption smoothing implies C did not fall. If anything, the reverse. Why even mention consumption smoothing if C isn’t changing?

    And I disagree that consumption smoothing gets you a BBM of one.

    I agree you can construct an example where:

    1. There is consumption smoothing (but not aggregate).
    2. I doesn’t change
    3. The BBM is one.

    But that is meaningless, as it’s simply the Keynesian BBM=1 assumption, in which consumption smoothing plays no role. If you redo the example without consumption smoothing, but no change in I, you’d still get a BBM of one regardless of the MPC.

  101. Gravatar of D R D R
    19. February 2012 at 20:07

    “Why even mention consumption smoothing if C isn’t changing?”

    C isn’t changing BECAUSE of consumption smoothing, not the other way around! One needs to make *some* assumption about behavior, or just assume a result. It just so happens that that assumption leads to no change in (aggregate) consumption

    “And I disagree that consumption smoothing gets you a BBM of one.”

    Then please explain where I am wrong!

    “But that is meaningless, as it’s simply the Keynesian BBM=1 assumption, in which consumption smoothing plays no role.”

    That’s totally untrue. It plays an absolutely critical role. If I do not assume consumption smoothing, then I must make some other assumption– say C rises by ten trillion times the rise in G and T. That would be a ridiculous assumption, so better to assume consumption smoothing. It just so happens that consumption smoothing leads to no change in C or I.

    I don’t assume that the BBM=1. It is a result of the model. Just because my model *results* in a BBM of 1.0 does not mean I assumed it.

    Some disposable incomes fall by T while at the same time some disposable incomes rise by G=T. If as a result of these changes, everyone tries to smooth consumption, then aggregate savings will not change.

    “If you redo the example without consumption smoothing, but no change in I, you’d still get a BBM of one regardless of the MPC.”

    So what? I didn’t assume fixed investment. I just see no reason for it to change because there is nothing changing aggregate disposable income. If you think I should change, then it’s up to you to explain why. What would cause investment to change?

  102. Gravatar of ssumner ssumner
    21. February 2012 at 18:56

    DR, You said;

    “C isn’t changing BECAUSE of consumption smoothing, not the other way around! One needs to make *some* assumption about behavior, or just assume a result. It just so happens that that assumption leads to no change in (aggregate) consumption”

    For the 100th time, C doesn’t change in the Keynesian model with the BBM, whether you have consumption smoothing or not.

    You can run a Keynesian model with a MPC of 1%, 50%, or 99%, and the BBM is always one and C never changes. Consumption smoothing is a recent development, totally unneeded to get the standard Keynesian result. The only case where consumption smoothing gets you a new and interesting aggregate result is when aggregate disposable income CHANGES. But it doesn’t change in your example, because if it did WL would be wrong (investment would also change.)

  103. Gravatar of D R D R
    21. February 2012 at 20:22

    “For the 100th time, C doesn’t change in the Keynesian model with the BBM, whether you have consumption smoothing or not.”

    And for the 100th time, that it a non-sequitur. Why do you keep bringing “the Keynesian model” into the discussion? What does that have to do with what I wrote, except that you say it comes to the same result? This is totally and hopelessly irrelevant.

    “The only case where consumption smoothing gets you a new and interesting aggregate result is when aggregate disposable income CHANGES.”

    What difference does that make? So you can get the same result using different assumptions. That doesn’t mean that the assumptions are identical.

    “But it doesn’t change in your example, because if it did WL would be wrong”

    Ignoring whether this statement is factual or not, it is not a criticism of my example at all. In fact, the second part of your statement is:

    If Yd changes in my example, then WL would be wrong.

    This is logically equivalent to saying

    If WL is correct, then Yd does not change in my model.

    I don’t understand how this implies that Yd doesn’t change in my model. Of course it doesn’t. But that follows from the assumptions of my model and has absolutely nothing to do with Wren-Lewis being correct. In fact, I’m arguing that if my example is correct, then Wren-Lewis is “wrong”

  104. Gravatar of D R D R
    22. February 2012 at 04:55

    Scott,

    I really don’t understand what you’re hung up about, but I wonder if you’re still stuck on the argument that Wren-Lewis didn’t include investment.

    But that is irrelevant to my example. The very point of my example (again: https://www.themoneyillusion.com/?p=12819#comment-133752) is that there is no obvious reason why investment should change if it were included in the model.

    The *immediate* impact of government purchase, coming with an equal tax, is for G+100, T+100, Y+100, Yd+0, Sp+0, Sg+0, S+0. So why would this induce a change in investment?

    Then the private sector smooths consumption in response to the government intervention. No matter the incidence of taxes and payments for the bridge, the total of all *desired* changes in consumption is zero. The desire to increase lending on the part of those who gain from the government intervention is met with an equal desire to increase borrowing on the part of those who lose from the government intervention. So why would this induce a change in investment?

    I don’t see anything in this story which suggests that investment would obviously change. Do you?

    It’s true I could get the same (aggregate) result by saying that all agents fix their consumption rather than smooth. Or that they un-smooth– increasing consumption by $1 or $2 for every dollar of disposable income lost. In that sense smoothing doesn’t matter because the savers will still match up dollar-for-dollar with the dissavers. Thus, a more general argument may be made. But I don’t see anything in *any* model of this class which suggests an obvious inducement to change investment. Do you?

    So I argue that Wren-Lewis was “wrong” to argue that “C fell”

  105. Gravatar of ssumner ssumner
    23. February 2012 at 19:48

    DR, I am increasing confused about what you are claiming, and I think we need to reach a resolution. You said:

    “It’s true I could get the same (aggregate) result by saying that all agents fix their consumption rather than smooth. Or that they un-smooth- increasing consumption by $1 or $2 for every dollar of disposable income lost. In that sense smoothing doesn’t matter because the savers will still match up dollar-for-dollar with the dissavers.”

    And then you said:

    “So I argue that Wren-Lewis was “wrong” to argue that “C fell””

    Those are exactly my two points. The argument about C falling was flawed (although I think flawed in a different way from what you think.) And the consumption smoothing assumption in no way advances his case. You get the same results with other assumptions.

    The difference between you and I is that you think WL erred in claiming C fell, and I think he intended it to fall but erred in discussing the implications of it falling. It seems to me this long futile discussion revolves around what sort of thing he really meant when he poorly expressed his argument, and that can never be resolved without a mindreader.

  106. Gravatar of D R D R
    23. February 2012 at 22:43

    “the consumption smoothing assumption in no way advances his case. You get the same results with other assumptions.”

    Right. This is utterly irrelevant. He put forward a model which as far as I can tell, results in a BBM of 1.0. What else could he possibly need to do to disprove Cochrane’s case?

    Perhaps you are hung up on that old allegation that Wren-Lewis argues Cochrane was wrong *because* of consumption smoothing? But he said no such thing. Cochrane is wrong because of the entire argument. The critical piece of the argument being not consumption smoothing but the fact that the taxes were assumed to go to G and not to transfers.

    In fact, Wren-Lewis explicitly stated later “All I want to suggest is one way of reinterpreting the balanced budget multiplier that is consistent with consumption smoothing. I think it is also nice that all this stuff ends up with the same result, a multiplier of one.”

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

    But it does not matter why he chose to use consumption smoothing in his model. All that matters is what happens when you have an G and T each rise by the same amount under an assumption that the private sector smooths consumption.

    You allege that I must fall because Wren-Lewis says that “C fell.” I think this is entirely wrong for two reasons.

    (1) I do not accept that Wren-Lewis said that “C fell.” Even if he thought it, I do not accept that he wrote it. And as you agree that we are without a mindreader, we must look to his words. Specifically, he wrote:

    “spending at time t will fall by much less than X.”

    Now, you wish to parse this as (spending at time t will fall) (by much less than X) while I parse this as (spending at time t) (will fall by much less than X)

    Now, clearly, my parsing is more charitable than yours, but even accepting your less-charitable parsing, that only brings us to…

    (2) I don’t care that “C fell” implies I fell, because I do not accept that C falls as a result of consumption smoothing, or anything else Wren-Lewis wrote.

    If I make a proposition (P) and it happens that P implies Q but I say that P implies not-Q, you can argue that I was wrong to say P implies not-Q but that does not change the fact that P implies Q. It matters not at all what not-Q implies.

    That is– even if Wren-Lewis wrote “C fell”– I don’t see how that statement follows from what he wrote. You have repeatedly passed on offering any explanation as to why C should fall as a result of anything Wren-Lewis described– let alone address my demonstration. This only increases my confidence that I am correct and that C does not fall.

    So if I say to you, “Wren-Lewis was wrong because `C fell’ does not follow from what he wrote” then have I or have I not made a valid criticism?

  107. Gravatar of D R D R
    23. February 2012 at 22:52

    And if that isn’t clear enough, we can take it one step at a time. In case you haven’t noticed, I’m willing to engage, so if you have a *question* about anything I wrote, then I will be happy to respond.

  108. Gravatar of ssumner ssumner
    25. February 2012 at 11:26

    DR, You said;

    “Right. This is utterly irrelevant. He put forward a model which as far as I can tell, results in a BBM of 1.0. What else could he possibly need to do to disprove Cochrane’s case?”

    I can’t imagine what you mean here. What model does he put forward? Cochrane knows that the BBM is one in the Keynesian model, that’s not what’s at stake. The question is whether the Keynesian model is true. What arguments does WL offer for it being true? He seems to suggest that consumption smoothing makes a positive BBM more likely, even inevitable. But why?

    As far as “parsing” his meaning. Perhaps in isolation it might be read either way (although I think my reading is more natural) but combined with consumption smoothing there’s only one way to parse that phrase. He says:

    1. Disposable income falls.
    2. Consumption falls by less.
    3. Because consumption smoothing.

    Hmmm, how can we put all that together?

    Someone claiming consumption wouldn’t fall at all would not combine those three claims.

  109. Gravatar of D R D R
    26. February 2012 at 08:14

    1. “The question is whether the Keynesian model is true. What arguments does WL offer for it being true?”

    I disagree with the premise of your question. Cochrane claims a “theorem” that the BBM is zero.

    “The question for the ‘multiplier’ is not whether it is greater than one, it’s how on earth it can be greater than zero?… These statements are a theorem not a theory.”

    As for your question itself, I already answered that in my last full comment https://www.themoneyillusion.com/?p=12819#comment-137691

    2. “He seems to suggest that consumption smoothing makes a positive BBM more likely, even inevitable. But why?”

    I again disagree with the premise of your question. He is attempting to show that the BBM may be positive *despite* consumption smoothing, not *because* of it.

    3. “Hmmm, how can we put all that together?”

    And again I disagree with the premise of your question. Where do you get “disposable income falls” out of what Wren-Lewis wrote?

    “Someone claiming consumption wouldn’t fall at all would not combine those three claims.” And again, so what? If I accepted this premise, I would say his error is in saying “C fell”

    I repeat my question from my comment above,

    So if I say to you, “Wren-Lewis was wrong because `C fell’ does not follow from what he wrote” then have I or have I not made a valid criticism?

    Yea or nay, sir?

  110. Gravatar of ssumner ssumner
    28. February 2012 at 06:06

    DR, It’s irritating to have to respond to the exact same points over and over again. If you don’t agree let’s move on. I said previously that Cochrane’s mistakes have no bearing on errors of logic in WLs argument. If you don’t agree with me fine, but don’t keep raising the same points over and over again.

  111. Gravatar of ssumner ssumner
    28. February 2012 at 06:17

    DR, You said;

    “I again disagree with the premise of your question. He is attempting to show that the BBM may be positive *despite* consumption smoothing, not *because* of it.”

    I completely disagree. I think it’s obvious that he thinks consumption smoothing means less of a fall in C than with non-smoothing. It’s obvious to me that he thinks it advances his positive multiplier argument. I think we’ll just have to agree to disagree, as I find your interpretation of his comments to be increasing bizarre, as if you are reading a different language. I literally have no idea what you are talking about.

    You said;

    “And again I disagree with the premise of your question. Where do you get “disposable income falls” out of what Wren-Lewis wrote?”

    Seriously, how many times do I have to answer this question? Go read my previous replies, it’s all there.

    You said;

    “So if I say to you, “Wren-Lewis was wrong because `C fell’ does not follow from what he wrote” then have I or have I not made a valid criticism?

    Yea or nay, sir?”

    I say nay. He said C fell, I assume he meant C fell, so I assume he was wrong about something else. It’s possible he didn’t mean to say C fell, but I think that unlikely.

    And please don’t keep repeating the same arguments.

  112. Gravatar of D R D R
    28. February 2012 at 07:42

    “I literally have no idea what you are talking about.”

    Perhaps if you asked a specific question it would help. As I have no idea where you lose the thread, there is not much I can do about it. I’m sorry you are irritated.

    “It’s obvious to me that he thinks it advances his positive multiplier argument.”

    I’m glad you are a mind-reader, despite your previous denial. Otherwise, we’d have to argue about what he wrote.

    “Seriously, how many times do I have to answer this question? Go read my previous replies, it’s all there.”

    No, I have no idea where you have Wren-Lewis saying disposable income falling. I see you arguing implicitly that disposable income falls as a result of consumption falling (and therefore investment falling) But then you can’t use the fall in disposable income to argue that consumption must have fallen. That would be circular. What am I missing?

    “I say nay. He said C fell, I assume he meant C fell, so I assume he was wrong about something else.”

    This fails basic logic. One premise of my proposed statement is that Wren-Lewis said C fell. The proposed criticism cannot be wrong on the grounds that the premise is true.

  113. Gravatar of D R D R
    28. February 2012 at 09:13

    “I think it’s obvious that he thinks consumption smoothing means less of a fall in C than with non-smoothing.”

    Then I would argue that he “thinks” wrong. I see no reason for C to fall more with “non-smoothing” than with smoothing.

    But that depends on what you mean by “non-smoothing”. If consumers DO NOT smooth, then it seems to me that when G rises by the same amount as T, then consumers do… nothing. So C is unchanged, I is unchanged, Y rises by G, Yd is unchanged, and savings are unchanged.

    The way I see it, when G and T rise by the same amount, then consumption smoothing redistributes consumption among private agents.

    I see no good reason why, given sufficient initial economic slack, we should see aggregate consumption fall– smoothing or no, Wren-Lewis correct or no.

  114. Gravatar of ssumner ssumner
    29. February 2012 at 19:43

    DR, You said;

    “This fails basic logic. One premise of my proposed statement is that Wren-Lewis said C fell. The proposed criticism cannot be wrong on the grounds that the premise is true.”

    This completely mischaracterizes what I said. You claim he made a mistake in saying C fell. I say he intended to say C fell, and made his mistake elsewhere. So I’m not guilty of what you claim.

    And no, I’m not a mind-reader, I’m a reader. I read what other people say. Is there a problem with that? Didn’t WL read what Cochrane had to say?

    And I didn’t use the fall in DI to argue consumption had fallen. WL said it fell. That’s an exact quotation. It “fell.” If consumption smoothing occurred, then I presume he meant DI also fell, otherwise there’d be no reason to claim C fell by less than the tax.

    DR, You said;

    “Then I would argue that he “thinks” wrong. I see no reason for C to fall more with “non-smoothing” than with smoothing.”

    Then you and I must define consumption smoothing totally differently. For me that’s virtually the definition. If you smooth, then C changes less when it does change.

    As far as you seeing no reason why C would change, that depends on all sorts of issues that have nothing to do with the debate. I don’t believe the BBM is one, but I suppose anything is possible.

    I’ve started a new policy of only answering comments from the 5 most recent posts, so I’ll give the the last word. Since this is an important topic for you I will return one more time to read your comments if you’d like to respond. (So you don’t fell you are talking to a wall–although you might already feel that way!) But I no longer have time to keep so many open threads going.

  115. Gravatar of D R D R
    29. February 2012 at 21:45

    “This completely mischaracterizes what I said. You claim he made a mistake in saying C fell. I say he intended to say C fell, and made his mistake elsewhere. So I’m not guilty of what you claim.”

    Let us review.

    1. You argue that he said C fell.
    2. For the purposes of argument, I am agreeing that he said C fell.
    3. I argue that consumption smoothing does not imply a fall in C, so Wren-Lewis must be wrong.
    4. You say my criticism is invalid because he said C fell.

    This last statement (4) is nonsense because I have already agreed in (2) that he said C fell. How can my criticism be wrong on the grounds that we agree on that point? It makes no sense.

    I understand you “assume he was wrong about something else” but that’s the whole problem. If you ASSUME AWAY all other possible criticisms, then you haven’t proven Wren-Lewis wrong at all. Your argument is incomplete. And your argument is incomplete in a very important way– namely, in a way which cuts against you, and in favor of Wren-Lewis.

    “And no, I’m not a mind-reader, I’m a reader.”

    Didn’t you just write, “I say he intended to say C fell, and made his mistake elsewhere.” Either you are mind reading, and he only “intended” to say C fell, or he DID say C fell, and my criticism holds.

    For now, I can care less what he intended. You believe he said that C fell. For now, I accept that. So why do you say my criticism is invalid?

    “If consumption smoothing occurred, then I presume he meant DI also fell, otherwise there’d be no reason to claim C fell by less than the tax.”

    And I am arguing that if he thought C fell as a result of consumption smoothing, he thought WRONG, regardless of what he wrote. What is so difficult to understand about that?

    But when I asked you about Yd, you responded “there’s only one way to parse that phrase. He says: 1. Disposable income falls…”

    The problem is exactly that. Yd falling is an inference you made. He didn’t say it. You can say he implied it when he said C fell, but I argue the premise (C fell) is itself false because it doesn’t follow from consumption smoothing.

    “Then you and I must define consumption smoothing totally differently. For me that’s virtually the definition. If you smooth, then C changes less when it does change.”

    YES. WHEN IT DOES CHANGE (in the aggregate). I see no reason for C to change in the aggregate with or without smoothing. There is no misunderstanding on my part.

    “As far as you seeing no reason why C would change, that depends on all sorts of issues that have nothing to do with the debate.”

    It is entirely relevant. If C doesn’t change as a result of consumption smoothing then the statement “C fell” is WRONG. What could be more relevant than that?

    Anyway, I really would like to understand what’s wrong with my argument, so let me know if you’d like to take this private. I do believe we can make ourselves clear to each other if we take it in small steps and sort out what it is we do agree on.

    Thanks,
    D R

  116. Gravatar of ssumner ssumner
    2. March 2012 at 07:03

    DR, I don’t think we’ll ever resolve anything, but let’s see if we can at least agree on each others opinions:

    1. You seem to believe that he basically got things right, and understood what he was trying to say, but made one error–saying or at least implying that C fell, when the actual implication of his model/thought process was that C wouldn’t fall.

    Does that correctly state your view?

    2. I am saying he really meant to say C fell, and furthermore his consumption smoothing comment suggests he really meant to say DI fell by more than C. Then his mistake was forgetting that if C fell by less that DI he needed to address the S&I question, which he ignored.

    Does that make sense?

    Then if you asked me why I don’t like your interpretation, I don’t see why he would have tried to make the rebuttal to Cochrane on consumption smoothing grounds, if that’s what he really meant. Rather he would have used traditional balanced budget multplier analysis, as consumption smoothing doesn’t seem to add anything to the BBM argument. The other reason I like my interpretation, is that I could easily see how someone could make this mistake if not thinking clearly. We all know the Keynesian model typical views money saved as disappearing down a rat hole, but that’s because of the desired/actual saving difference. I.e. attmepts to save don’t actually lead to saving, which is why the BBM is one. Then I think he forgot that if he did a consumption smoothing example where C ACTUALLY DID FALL he could no longer make the savings goes down a rathole assumption, because actual saving and investment would fall. That’s the way I read his error and nothing you’ve said makes me think your interpretation (while theoretically possible) is even close to being as plausible.

    I’m not committing to carry on this conversation as I don’t think we are iterating closer to agreement (and I’m not sure whose fault that is. (I zip around between dozens of commenters, and have trouble remembering the long trail of previous points.) But I at least wanted to make one final attempt to summarize our two views, to see if we could at least agree where we disagree. If we both understand the other side’s argument.

  117. Gravatar of D R D R
    2. March 2012 at 13:38

    Thanks for your patience.

    “Does that correctly state your view?”

    No, not really.

    1) I don’t actually believe that he said that C fell by a nonzero amount, regardless of what was going on in his head. However, for the purposes of discussion, I am willing to accept that he did say that. Also, for the purposes of discussion, I am willing to accept that he included investment in his model– that is, that Y=C+I+G, and not Y=C+G.

    2) Solely within the context of that discussion, I don’t know one way or another that “he basically got things right.” I merely propose that his statements up to that point do not seem to me consistent with “C fell.” In this context, I don’t see a point in trying to understand him past that point, because that would require accepting a falsehood as true. Once I say that 0=1, the fact that 1=0+1=1+1=2 follows does not mean that my error is in saying that 1=2. My error is in saying 0=1. But maybe he erred even earlier than “C fell”. All I know is that I notice his troubles at this point. Did he basically get things right? In this discussion, I have no idea, because it all falls as a whole.

    3) I have no position on what he believed when he said C fell by a nonzero amount. It does not follow from what he wrote whether he believed it followed from his model, or instead he drew an erroneous conclusion from his model, or his model is something entirely apart from what he had written up to that point.

    “Does that [my (Scott’s) view] make sense?”

    That depends on how your reasoning progresses. Even within the context of discussion wherein I agree he actually said C fell, I do not understand how you come to say “he really meant to say C fell.”

    This statement suffers from ambiguity. Are you just restating the context for the discussion? (That is, “Whatever he really believed, he meant to write that C fell”) Or are you saying that whatever he wrote, he really believed that his model told him that C fell?

    In the first case, he may be wrong having written that C fell. In the second case, he may be wrong having believed that C fell. The rest of your argument I do not follow because I don’t understand this starting point. (Or, importantly, if this is your starting point at all.)

    I would appreciate it if you made clear which, if any of those statements in your “2)” serve as basis for your reasoning, or if, alternately, they are all conclusions drawn from other premises.

    As for the rest, maybe that needs to go on hold until we have the rest sorted out. But there is one comment here to which I cannot resist mentioning. Specifically, “We all know the Keynesian model typical views money saved as disappearing down a rat hole.” I don’t want to get into a whole long distraction here– unless we reach a point where it is our foremost stumbling block– but we very much do not all know that. (It may depend on what you mean by “the Keynesian model”– as if there is just one– and what you consider “a rat hole.”)

  118. Gravatar of ssumner ssumner
    4. March 2012 at 05:52

    DR, We are just going around in circles now. If his mistake was saying C fell, then he would have really wanted to make the following claim:

    “When government spending rose by X, consumption would not fall at all because after tax income would not fall at all. And that means saving and investment would not change. Ergo Ccochrane is wrong”

    Suppose he had used that argument to “disprove Cochrane.” Everyone would have laughed at him, and rightly so. And adding consumption smoothing wouldn’t help at all, as people would just laugh even harder, pointing out that if DI didn’t fall then obviously C wouldn’t fall, whether there is consumption smoothing or not. So one could hardly argue that “Cochrane was wrong because he forgot about consumption smoothing,” which is WLs actual argument. In other words, I interpret him the way I do because every other interpretation seems laughable.

    You seem to want an alternative interpretation where the numbers add up. I certainly agree that the numbers add up in the argument I just ridiculed. But IT’S NOT AN ARGUMENT AT ALL. It’s just stating “Cochrane is wrong because I think the BBM is one,” with no argument presented as to why it should be one. Cochrane agrees that if you assume the BBM is one, then it’s one. I thought WL was actually trying to make an argument based on behavioral relationships like consumption smoothing. But if he assumed C didn’t change, he simply would be assuming the Keynesian model is true. That’s no argument at all.

  119. Gravatar of D R D R
    4. March 2012 at 09:48

    This does not progress the argument one bit. You are simply misstating my argument, and making up quotes which you attribute to Wren-Lewis.

    “If his mistake was saying C fell, then he would have really wanted to make the following claim…”

    I DON’T CARE WHAT CLAIM HE WANTED TO MAKE. He made an argument, and that argument may either stand or fall. The argument cannot fall because a different argument may be silly.

    You say that “as people would just laugh” you interpret him to say that C fell by a nonzero amount.

    I SAY THAT EVEN IF HE SAID C FELL THEN HE WAS WRONG.

    “You seem to want an alternative interpretation where the numbers add up. I certainly agree that the numbers add up in the argument I just ridiculed. But IT’S NOT AN ARGUMENT AT ALL. It’s just stating ‘Cochrane is wrong because I think the BBM is one,’

    That’s just not true. I nowhere make an assumption that the BBM is 1.0.

    I assume that there is sufficient economic slack for Y to rise by X when G and T both rise by X. I assume that in response to individual changes in Yd, that consumers smooth their consumption. Am I missing anything?

    I have told you repeatedly that this requires no *assumption* that the BBM is 1.0. As I have explained to you previously that I am NOT assuming that the BBM is 1.0, I must infer that you either do not read my arguments, you quickly forget my arguments, or you are lying about my arguments.

    Either way, I would appreciate it if you stopped misrepresenting my arguments.

    It simply doesn’t matter why I make the argument I make, just as it doesn’t matter why Wren-Lewis makes the argument he makes. Just as it doesn’t matter why Cochrane makes the argument he makes. The argument stands on its own, and unless you explain how it is flawed, then I have to assume it stands.

    “But if he assumed C didn’t change, he simply would be assuming the Keynesian model is true. That’s no argument at all.”

    If he ASSUMED C didn’t change? So what? He didn’t so assume, and neither one of us is arguing he did. I infer that you are not, because then in fact his ERROR would be in saying that “C fell” when he ASSUMED otherwise.

    Just because one reaches the same conclusion, does not mean the assumptions are the same.

    Person A: Assume P1. P1 implies Q. Therefore Q.
    Person B: P2 implies Q, therefore Person A would be assuming P2.

    Person B is not refuting A. Person B is failing at logic.

  120. Gravatar of ssumner ssumner
    8. March 2012 at 06:50

    DR, I am completely confused as to what you are doing here, you seem to be repeatedly misinterpreting my point.

    If you don’t think your argument is assuming a BBM of one, then please do another mathematical example where the BBM is not one, and yet you use identical assumptions. If you cannot, then obviously your assumptions amount to just assuming the BBM is one.

  121. Gravatar of D R D R
    8. March 2012 at 08:52

    Scott,

    You are saying I am wrong because my assumptions lead inevitably to my conclusion? I’m pretty sure “a logical progression by which assumptions lead inevitably to a conclusion” is the definition of a proof.

    The words “amount to” suggest to me that you do understand that I didn’t assume my conclusion, but rather that the proof required so few steps as to be so obvious I may as well have.

    Then again, you are “not a natural blogger” so maybe I don’t misunderstand. Irrespective of your view of the simplicity or complexity of my argument, do you or do you not understand that I did not *actually* assume my conclusion?

  122. Gravatar of Krugman Calls Out Sumner: Hornets Nest! | Last Men and OverMen Krugman Calls Out Sumner: Hornets Nest! | Last Men and OverMen
    21. February 2017 at 05:50

    […]      https://www.themoneyillusion.com/?p=12819 […]

Leave a Reply