Simon Wren-Lewis digs a deeper hole

Simon Wren-Lewis, Professor of Economics at Oxford, responds to my recent post with 5 arguments, none of which are persuasive:

Your penultimate move: I did not forget that saving is the same thing as spending on capital goods, because it is not.

Yes it is, check Krugman’s textbook.  S=I and I is spending on new capital goods.

Your final move: two mistakes here. You have output/income rising by 80, so after-tax income falls by 20, which is the same as the fall in planned consumption when income was expected to fall by 100, so you say no consumption smoothing and checkmate.  First, consumption smoothing is about consumption plans, not outturns.

You are trying to refute John Cochrane and Bob Lucas, who work with rational expectations models.  Plans are going to reflect the outcomes predicted by the model.  So this point is irrelevant to Cochrane’s claim.

Second, you are wrong to stop there. Consumers now change their plans, because they smooth: keeping proportions the same, they cut consumption by 4, not 20 and we iterate on.

But then there is no reduction in income, so no consumption smoothing occurs.  In any case, why make that assumption?   I provided a specific example with declines in consumption and home purchases that was 100% consistent with your consumption smoothing model.

 Anyone who has done first year macro will recognise the balanced budget multiplier.

So you are going to refute Lucas and Cochrane’s criticism of Keynesian economics by referring to simplistic undergraduate economics books that simply assume the most vulgar form of hydraulic Keynesianism is true?  The sort that was abandoned by the new Keynesians years ago?  Yes, we know the balanced budget multiplier is one in the textbook Keynesian model, the issue is whether that model is useful.

We get expansionary fiscal policy because of consumption smoothing.

No, because your argument here doesn’t rely on consumption smoothing, there is no change in income.  Again, I provided a very plausible specific example where C + I fell by the amount that G increases, and I did use consumption smoothing.  Spending on C fell by 20 and spending on I fell by 80.  It featured exactly the consumption smoothing you (wrongly) assumed somehow refuted Cochrane.  Consumption smoothing is part of the new classical model.  It still confuses me as to how you think you could refute the new classical model by relying consumption smoothing.


Tags:

 
 
 

46 Responses to “Simon Wren-Lewis digs a deeper hole”

  1. Gravatar of D R D R
    13. January 2012 at 15:55

    No, Scott, your specific example is only one possible result. Rather than *assuming* C-20, I-80, you can instead *assume* consumption smoothing, which implies C and I are each unchanged.

  2. Gravatar of D R D R
    13. January 2012 at 15:59

    Sorry, I meant to say, autonomous investment and consumption smoothing. Obviously, consumption smoothing alone and I-80 implies C-20.

  3. Gravatar of Andy Harless Andy Harless
    13. January 2012 at 16:06

    No, because your argument here doesn’t rely on consumption smoothing, there is no change in income.

    Scott, AFAICT you are falling into the fallacy of composition here. There is no change in aggregate income. There are changes in individual incomes, and individuals are smoothing. Some people have higher incomes because they are the ones experiencing the higher aggregate demand (for example, the people who have been hired by the government), and for them this more than offsets the increase in taxes. These people are smoothing their consumption, consuming a little bit more but not a lot more. Other people have lower disposable incomes because their taxes are higher (and because they, and the resources they own, were already fully employed and therefore aren’t affected by the increase in aggregate demand). These people are also smoothing, consuming a little bit less but not a lot less.

  4. Gravatar of Benjamin Daniels Benjamin Daniels
    13. January 2012 at 16:07

    Let’s walk through the proof then.

    Y(1) = c + MPC *(Y-T) + I + G
    Y(1) = (c + I + G – MPC*T)/(1-MPC)
    G -> G+a
    T -> T+a
    Y(2) = (c + I + (G+a) – MPC*(T+a))/(1-MPC)
    dY = Y(2) – Y(1) = a

    Note that this holds only if the distribution of income from the additional bridge-building is exactly the same, holding aggregate MPC constant across periods. If aggregate MPC is increased by a transfer to those with greater MPC, then the effect is increased further.

    “One recalls the sentence by James Meade, ‘that in classical economics a dog called savings wags a tail called investment, whereas in Keynes General Theory, it is a dog called investment, that wags a tail called savings.'”

    S=I may be a long-run identity but causality runs from I to S, not the other way around. In order to assert falling I you have to have a reason for it that is not a decline in savings.

  5. Gravatar of D R D R
    13. January 2012 at 16:41

    Ding! Well done, Andy!

    Let’s consider the transfer of $100 from A to B

    A reduces consumption by $7.50
    B increases consumption by $32.50
    total consumption rises by $25.
    investment rises by $100.
    total income rises by $125

    Now, if A and B have an equal claim to pre-tax income, they get $62.50 apiece. Thus,

    A’s disposable income falls $37.50, or 5 times $7.50.
    B’s disposable income rises $162.50, or 5 times $32.50.

    private savings, of course, rises by $100, etc.

  6. Gravatar of ssumner ssumner
    13. January 2012 at 16:50

    DR, Sure, that’s a possible result (I said so) but:

    1. It doesn’t involve any consumption smoothing as C doesn’t change.
    2. It doesn’t in any way refute Cochrane, it simply assumes the Keynesian model is correct. So what?

    But Wren-Lewis claimed that consumption smoothing can be used to refute Cochrane.

    Andy, I have no problem with income distribution issues, but that has nothing to do with consumption smoothing, as you get the same results with or without consumption smoothing, as aggregate income is unchanged in that case.

    See my previous response, I think you are missing my main objection–Wren-Lewis isn’t refuting Cochrane, he’s simply assuming the simple balanced budget multiplier is one. But that’s the issue we are debating. Cochrane’s assertion is 100% consistent with consumption smoothing (as I showed), contrary to Wren-Lewis.

    Benjamin, I can give you 100 reasons for it. There is no doubt that if you assume the Keynesian model is true, it can be shown to be true. But my example is 100% consistent with mainstream macro. Assume M and V don’t change, C falls by 20 and I falls by 80 and S falls by 80 and G and T rise by 100. It could easily happen. It also could easily not happen.

  7. Gravatar of D R D R
    13. January 2012 at 16:51

    Um… I have no idea how I got that $100 boost to investment. Recalculating route…

  8. Gravatar of D R D R
    13. January 2012 at 16:55

    ARGH.

    It *assumes* consumption smoothing, which constrains the results. In this example, it *happens* that there *must be* no change in disposable income to smooth.

    IF we assume smoothing and autonomous investment, THEN disposable income must not change.

  9. Gravatar of D R D R
    13. January 2012 at 17:07

    My apologies on that praise. I was in the middle of worrying about the bridge, rather than the transfer and jumbled things up.

    I do believe that if you assume autonomous consumption and consumption smoothing, then a transfer from A to B gets you nowhere. A must reduce consumption by $20 as B increases by $20. Hence, no change in C, blah blah blah.

    However, everything changes if we tax folks to pay for a bridge.

  10. Gravatar of marcus nunes marcus nunes
    13. January 2012 at 17:10

    First Krugman, then you and now Thoma. Wren Lewis is becoming a “household name”:
    http://economistsview.typepad.com/economistsview/2012/01/ideology-and-demand-denial.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+EconomistsView+%28Economist%27s+View+%28EconomistsView%29%29

  11. Gravatar of TravisA TravisA
    13. January 2012 at 17:35

    Scott wrote:

    Wren-Lewis isn’t refuting Cochrane, he’s simply assuming the simple balanced budget multiplier is one. But that’s the issue we are debating. Cochrane’s assertion is 100% consistent with consumption smoothing (as I showed), contrary to Wren-Lewis.

    I think this should go in a PS to this post. It is the key point. Wren-Lewis assumes the multiplier is one and says it is due to consumption smoothing while you have shown that the multiplier can be zero, even with consumption smoothing. Consumption smoothing isn’t sufficient to create the multiplier. You need other Keynesian assumptions, which of course Cochrane doesn’t necessarily accept.

  12. Gravatar of D R D R
    13. January 2012 at 18:03

    Scott says,

    “See my previous response, I think you are missing my main objection-Wren-Lewis isn’t refuting Cochrane, he’s simply assuming the simple balanced budget multiplier is one. But that’s the issue we are debating. Cochrane’s assertion is 100% consistent with consumption smoothing (as I showed), contrary to Wren-Lewis.”

    As I have said, repeatedly, I have no bone to pick over questions of different models. I quibble over three items:

    1) Your claim that Wren-Lewis failed accounting.

    2) Cochrane’s claim that he satisfactorily addressed the question of stimulus when he only addressed a forced transfer from A to B… no debt financed transfers from the government, no debt-financed government consumption/investment, and no tax-financed government consumption/investment.

    3) The notion that Wren-Lewis did not address Cochrane. As I wrote earlier,

    Wren-Lewis presented a counterexample of how stimulus *might* succeed, directly answering Cochrane when he wrote “The question for the ‘multiplier’ is not whether it is greater than one, it’s how on earth it can be greater than zero?”

    Because Cochrane implied his proof worked for all stimulus, there is nothing wrong with presenting a model which addresses tax-financed government investment. Futhermore, because Cochrane’s argument was so outwardly case-specific, it provided no framework for addressing any other question.* It is unfair to dismiss Wren-Lewis’ argument for trying.

    * If you can demonstrate how Cochrane’s argument applies to tax-financed government investment– based on a fair reading of his text– then go for it. But
    if you say:

    dC = -20
    dI = -80
    dG = +100
    dT = +100
    dY = 0
    dYd = -100

    then I still need to see how Cochrane would have come to that answer based on what he wrote. Otherwise, you’re speculating just as much as Wren-Lewis. Of course, based on his claim that it was a complete proof, you may infer that his model must produce this result. But that is begging the question, not applying his model.

  13. Gravatar of D R D R
    13. January 2012 at 18:20

    TravisA:

    Yes, that is a serious issue. We have no idea how Cochrane would argue against tax-financed government investment.

    Because tax-financed government investment is a legitimate stimulus plan (putting aside whether or not it works) it falls under the scope of Cochrane’s claimed conclusion that defenders of “stimulus” must “tell us which of the ‘ifs'” they disagree with.

    Nobody can point to Wren-Lewis’ argument and say which of his ‘ifs’ are contrary to Cochrane, because Cochrane never addressed that case at all, and so we have no idea what ‘ifs’ are up for debate.

    Cochrane’s proof was incomplete. That’s a serious matter.

  14. Gravatar of DW DW
    13. January 2012 at 18:55

    Is this debate civil? I don’t mean to be sarcastic, I genuinely can’t tell.

  15. Gravatar of Dave Dave
    14. January 2012 at 06:05

    I never thought I’d feel sorry for politicians, but when they say they’ve consulted with an economist I think I now will 🙂

  16. Gravatar of Andy Harless Andy Harless
    14. January 2012 at 06:12

    Scott, you seem to be attributing much weaker statements to Lucas and Cochrane than what they actually said.

    Cochrane: “Spending supported by taxes pretty obviously won’t work…

    Not “It might not work” but “It will not work.” Not “It will not work in my reasonable model” (which Cochrane never describes with any clarity anyhow) but “It obviously will not work.” The implication is that even a dumb-ass Keynesian should be able to understand why it won’t work.

    Lucas: “You apply a multiplier to the bridge builders, then you’ve got to apply the same multiplier with a minus sign to the people you taxed to build the bridge.

    Lucas is using the Keynesian term “multiplier” and arguing that the “multipliers” cancel out. He’s not saying that the concept of a multiplier doesn’t make sense (which might be a reasonable argument). He’s saying that there are two quantities to which to apply that multiplier, and they offset each other. The reasonable interpretation of what he says is that, even in a Keynesian model, if it’s done properly, the balanced budget multiplier is zero. Which is clearly wrong.

    Now Wren-Lewis goes back and makes the obvious interpretation and looks for the most obvious reason that someone might reach the fallacious conclusion that is present in that obvious interpretation. The defenses of Lucas and Cochrane are invariably based on the premise that they meant something very different from what they said.

    Nobody will deny that there are models (and only a few will deny that there are reasonable models) in which the balanced budget multiplier is zero. And if you want to make an argument based on such a model, fine. But you need to give some kind of explanation (for example, “There is no reason to expect that this would affect the amount of money that people choose to hold relative to their income or that it would affect the supply of money available…” which is actually a bad argument, but at least it is a coherent one). If you give a pseudo-Keynesian explanation, your reasonable expectation should be that it will be judged in terms of the Keynesian model.

  17. Gravatar of ssumner ssumner
    14. January 2012 at 06:20

    DR, The Keynesian model doesn’t sometimes assume “autonomous investment” although certainly more sophisticated Keynesian models don’t. But why in the world would a non-Keynesian like Cochrane assume autonomous investment.

    TravisA, It’s more than a question of consumption smoothing not being enough, it has nothing to do with the argument for Keynesian balanced budget multiplier. The simple Keynesian model doesn’t even have consumption smoothing, and the BBM is still one.

    DW. I hope so.

    DR, You disagreed with:

    “Your claim that Wren-Lewis failed accounting.”

    He suggested that a decline in spending is equal to a decline in C. But it’s not, the fall in private spending is the decline in C+I.

    As far as Cochrane is concerned I’ve criticized his reasoning in other recent posts, so that accusation of yours is incorrect. I have not given him a pass. My point was that Wren-Lewis and by implication Krugman made accounting errors, and they did. They suggested that the fall in after-tax income would lead to a smaller fall in consumption for consumption smoothing reasons. But that ignores the fall in saving! The fall in saving leads to less investment, unless saving didn’t fall, in which case their entire argument is completely incoherent.

  18. Gravatar of ssumner ssumner
    14. January 2012 at 06:41

    Andy, I completely agree with your comments on Lucas and Cochrane–those were bad arguments. But Wren-Lewis also had a really bad argument. He claimed that because of consumption smoothing a 100 dollar fall in after tax income would lead to only a 20 fall in C (I’m using my own numbers here.) But if that is so, then Cochrane was completely right, as C plus I would have fallen 100 in that case. So his own example show’s Cochrane is right.

    Now it’s certainly possible that Cochrane is wrong for other reasons, and the argument Cochrane used is very weak. But you should refute it by talking about changes in velocity, not by talking about consumption smoothing, which only occurs if income falls, and in that case Cochrane is right.

    Wren-Lewis seems to have falsely assumed that if consumption fell much less than G, then total spending would rise. But if the MPC is 20%, and consumption falls by 20, then it’s simple arithmatic that investment must have fallen by 80. I’d say that’s a really bad argument, which doesn’t seem aware of the fact that declines in “spending” are falls in C+I, not just C.

    Again, I am defending nothing of what Cochrane wrote here. Elsewhere I criticized him. The point is that Wren-Lewis’s argument was wrong, and Krugman endorsed it. It seems to make the same sort of EC101 error they accuse Cochrane of making. If he said Cochrane would be wrong if the balanced budget multiplier was one and everyone was a redhead it would be about as meaningful as saying Cochrane is wrong if the balanced budget multiplier is one and there is consumption smoothing. Basically Wren-Lewis is saying that Cochrane is wrong is you assume the balanced budget multiplier is one. Period.

    Now one can rescue Wren-Lewis by suggesting he must have meant something different from what he said, and that’s possible. Similarly I’ve tried to rescue Lucas and Cochrane by pointing out that elsewhere in the Cochrane paper DeLong criticized, Cochrane said that it’s theoretically possible that fiscal stimulus could raise V, but in that case the problem is monetary and the obvious solution is for the government to issue more risk free assets (money plus T-securities.) So it’s also true that Cochrane’s actual beliefs are much more sophisticated than that remark makes them seem.

    I’d also be interested in your views on Wren-Lewis’s odd statement that saving is not spending on capital goods. If not, then what is saving? It’s certainly not money set aside and not spent, as we know that a $100,000 cash purchase of a new house is “saving.”

  19. Gravatar of foosion foosion
    14. January 2012 at 09:41

    >> saving is not spending on capital goods >>

    Aren’t residential (housing) and inventory parts of I?

  20. Gravatar of TravisA TravisA
    14. January 2012 at 09:41

    Thanks, Scott. Your response to my point and also to Andy has made it clearer. With each iteration, it gets better.

    Neither Cochrane/Lewis nor Krugman/Wren-Lewis are covering themselves in glory, here. Nevertheless, some very good posts by you and Nick have come out of this that tremendously clarified the savings-investment definition and problems. I am curious whether you’ve read The Foundations of the Theory of National Income by Lipsey that Glasner has mentioned as being really good. Is it worth reading?

  21. Gravatar of ssumner ssumner
    14. January 2012 at 09:55

    Foosion, Of course they are. He made a very basic mistake there, forgetting that S=I.

    TravisA, You said;

    “Neither Cochrane/Lewis nor Krugman/Wren-Lewis are covering themselves in glory, here.”

    Yup.

  22. Gravatar of D R D R
    14. January 2012 at 11:10

    Scott says:

    “Now it’s certainly possible that Cochrane is wrong for other reasons, and the argument Cochrane used is very weak. But you should refute it by talking about changes in velocity, not by talking about consumption smoothing, which only occurs if income falls, and in that case Cochrane is right.”

    I fail to see why Scott feels the need to restrict how anyone refutes a thing. Either the logic is correct or it is not. More importantly, however, Scott again skips the main point about consumption smoothing.

    If the private sector smooths consumption with respect to *disposable* income, then

    1) if Yd falls, then C falls by a lesser amount;
    2) if Yd *rises*, then C *rises* by a lesser amount;
    3) if Yd *does not change*, then C *does not change.*

    Scott says consumption smoothing implies falling income. This is plainly indefensible, as rising income may lead to smoothing. This is so obvious, I have to assume this is not what Scott meant.

    Apparently, what Scott meant was if consumption smoothing and falling consumption, then falling income. But this is not logical, and pay be proven by counterexample.

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

    Y +50
    Yd -50

    Voila! Income up 50, even though consumption fell 10. So that’s not it, either.

    What Scott really meant was… if consumption smoothing and falling consumption, then falling *disposable* income. But then it does not directly follow that Cochrane is correct that income does not rise.

    Wait, you say… wasn’t Cochrane’s argument about a balanced budget? Well, yes and no. He gave an example of how the economy might so respond, but he claimed it addressed the general case. I have no idea why he thought it a sufficient proof.

    But let’s go back to Sumner. He may just as easily have stated… if consumption smoothing and non-increasing consumption, then non-increasing disposable income.

    OK…

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

    Y +100
    Yd +0

    So non-increasing disposable income and non-increasing consumption. Obviously, the economy is consumption-smoothing. Again, true that there is no change in disposable income and consumption, but that doesn’t mean there is no smoothing. The smoothing assumption disallows

    C +10
    I -10
    G +100
    T +100

    Y +100
    Yd +0

    Get it?

  23. Gravatar of D R D R
    14. January 2012 at 11:14

    Scott,

    “Foosion, Of course they are. He made a very basic mistake there, forgetting that S=I.”

    No, Scott misinterprets Wren-Lewis and then infers that the error is on Wren-Lewis’ part. If Scott accepted the plain and obvious interpretation of what Wren-Lewis wrote, there is no such error.

  24. Gravatar of e e
    14. January 2012 at 11:15

    Scott, I’m not sure why you’re last argument was a checkmate. I think you assumed that there was a fixed amount of production in a closed economy, but couldn’t they claim that by increasing the total production in a time period the Gov’t can increase consumption that would have come out of leisure or unemployed people. Wouldn’t that additional consumption not require reduced savings?

    I realize this isn’t what SWL is saying directly but I was curious.

  25. Gravatar of D R D R
    14. January 2012 at 11:30

    e,

    Scott does allow this is a possibility. (See http://www.themoneyillusion.com/?p=12636&cpage=1#comment-126894)

    I cannot explain why he seems to disallow it from entering his arguments.

  26. Gravatar of CA CA
    14. January 2012 at 20:45

    Wren-Lewis has a new post up.

    http://mainlymacro.blogspot.com/2012/01/savings-equals-investment.html#comment-form

  27. Gravatar of KRG KRG
    14. January 2012 at 21:27

    “Yes it is, check Krugman’s textbook. S=I and I is spending on new capital goods.”

    How can you tell what is C and what is S then, until well after the current moment and you’ve seen which spending was successfully employed to produce returns and what resulted in a loss? (Heck, by this interpretation, most of the food that each person eats is really S/I because it’s effectively an investment in their ability to work and produce income)

    But using an S that lumps capital and financial spending into one lump- when you say that S is reduced by 80, on what are you basing the assumption that it’s the capital spending “I(C)” portion and not the financial “I(F)” portion that drops? (Especially if I(C) can be written off as business expenses, and thus offers an avenue to avoid taxation completely, increasing velocity and letting more of the the tax burden fall on money taken out for I(F). Heck- with a sufficient increase in velocity from properly applied tax policies, you can effectively double down on potential multipliers by increasing the number of times that money from I(C) cycles back in until the desired amount is skimmed out of money diverted to I(F) instead.

  28. Gravatar of foosion foosion
    15. January 2012 at 04:48

    >>I is spending on new capital goods.>>

    Inventory is a new capital good? Residential housing is a new capital good? Aren’t capital goods means of production?

  29. Gravatar of ssumner ssumner
    15. January 2012 at 06:48

    DR, Wren-Lewis clearly rejects the S=I, he said so in his reply.

    As far as all your mathematical examples, none address my point. Suppose Wren-Lewis had said that Cochrane was wrong because the moon orbits around the Earth, which shows the balanced budget multiplier is one. I think we’d all agree it was mistaken. That doesn’t make the multiplier one. But of course the moon does revolve around the Earth, and it’s very possible the BBM is one for other reasons. And consumption smoothing does occur, and it’s very possible the BBM is one for other reasons. But not because of consumption smoothing.

    You show examples where the BBM is positive, but consumption smoothing plays no role in this examples. You simply assume a positive BBM. Wren-Lewis was saying Cochrane was wrong because he ignored consumption smoothing. That’s flat out wrong, as my counterexample showed. You can do all the consumption smoothing you want and still get a zero multiplier. Or a multiplier of one. Or 1/2. So Wren Lewis’s argument was clearly false, even if his conclusion later turned out to be true.

    e, See my response to DR, His conclusion certainly might be right, but his reasoning is wrong. Consumption smoothing is not an argument against Cochrane, whereas he claimed it was. In my second example there is no consumption smoothing. Consumption smoothing wasn’t even part of the original Keynesian multiplier model, but it is part of the new classical model. How could anything think it disproves the classical model?

    Thanks CA.

    KRG, You are right about one thing, the only sensible way to think about Keynesian multipliers is their impact on V (and M, I would add.) It certainly might increase V.

    But you are wrong about investment. Economists define investment as the construction of new capital goods. Financial investment doesn’t count, because it would be double counting. Imagine adding up the value of common stocks and the value of corporate physical assets, you’d be double counting.

    Yes, there is a gray area as to what should be considered capital goods and what should be considered consumer goods, but it doesn’t affect my argument here in any way.

    foosion, Yes, I believe that’s the standard definition used by economists. Housing is more “capital-like” than most factories, as it lasts longer. Houses produce a steady flow of shelter services over many decades, so that’s obviously capital. (Think of an apartment building.)

    Inventories are not consumed this period, but are held over until next period, allowing higher future consumption. That’s why they are considered capital.

  30. Gravatar of cthorm cthorm
    15. January 2012 at 07:35

    @DR

    You don’t get someone in to check mate by using a “good enough” counterpoint.

    @scott

    Very entertaining read throughout. It’s not often I laugh out loud from two grown men arguing about extremely abstract economics models, let alone while throwing out the “anyone who took basic” jabs.

  31. Gravatar of D R D R
    15. January 2012 at 08:17

    cthorn,

    That is EXACTLY my point. I am not claiming checkmate… Scott is, and the ability to remove the check by any legal means is enough to disprove checkmate.

  32. Gravatar of foosion foosion
    15. January 2012 at 08:23

    Scott, I increases due to inventory accumulation does not seem to fit well with your argument. It seems much closer to hoarding than spending on the means of production. It also looks much closer to production smoothing (cf. consumption smoothing). You can’t keep increasing inventory forever.

  33. Gravatar of D R D R
    15. January 2012 at 09:14

    foosion,

    Actually, it makes fine sense.

    We are trying to measure current production, so if something is produced by not sold, it still has to count somewhere. It may be “future” consumption, but it would not make sense to count it as current consumption. Instead, when the inventory eventually gets purchased, it counts at that time as both an increase in consumption and a decrease in investment.

  34. Gravatar of foosion foosion
    15. January 2012 at 10:57

    DR, I see that Wren-Lewis has posted on inventory accumulation http://mainlymacro.blogspot.com/2012/01/savings-equals-investment.html (apologies to CA for not seeing it sooner)

  35. Gravatar of D R D R
    15. January 2012 at 11:05

    Scott,

    You and Wren-Lewis are talking past each other on S=I.

    Yes, if everyone collectively (including the government) manages to *successfully* increase savings, then investment must rise. That is the meaning of S=I.

    But if I put in some overtime, earning me $100 of income as a result, and I smooth my consumption so I spend $20 more, I am increasing my “savings” by $80. That is not “the same thing as spending on capital goods.”

    Now, it *may* be that someone else will borrow my $80 to buy capital goods. Or maybe they will not. Maybe someone else will take time off, costing them $100 of income, and dissave $80.

    Yes, *collectively* we have not changed savings or investment. But that does not mean that savings is the same thing as spending on capital goods.

  36. Gravatar of D R D R
    15. January 2012 at 11:18

    Consuming less is saving.
    Cutting government expenditures is saving.
    Increasing taxes is saving.

    Any of these *may induce* additional investment, but that does not mean that savings IS the same thing as spending on capital goods.

  37. Gravatar of ssumner ssumner
    16. January 2012 at 11:55

    Thanks cthrom.

    DR, So is it legal for Cochrane to claim he was holding monetary policy constant (inflation targeting?)

    Or are only the “good guys” allowed those sneaky moves?

    foosion, I completely agree that inventory changes would only happen for a while. My point was different, that inventory changes are changes in investment. They are a form of “spending.”

    DR, You said;

    “But if I put in some overtime, earning me $100 of income as a result, and I smooth my consumption so I spend $20 more, I am increasing my “savings” by $80. That is not “the same thing as spending on capital goods.””

    We’ll never get anywhere talking at the individual level, as it depends whether someone else dissaves. If they don’t offset the individual saving, then investment MUST rise.

  38. Gravatar of D R D R
    16. January 2012 at 22:27

    *laugh*

    Is it legal? Are you asking if it is logical to say “the fiscal multiplier is zero because the central bank will offset any intended effects of fiscal policy?”

    First, I’m not convinced that the Fed has that much room to maneuver on the loosening side if there was a fiscal tightening. I would need to see how this would be possible before accepting that as a reasonable ground rule.

    Second, even if it’s theoretically possible, I don’t think I’ve seen a real argument that the Fed has been holding back on account of fiscal policy. They did double the monetary base in three months, cutting the effective fed funds rate to 0.15 by January 2009. I’m having trouble imagining what they would have done if fiscal policy had been austere in 2008 and 2009.

    I’m ignoring your next question. Clearly not serious.

    As to the last, I’m afraid it matters more than you let on. I had praised Andy for raising the specter of a composition fallacy. I then took it back. I now re-praise him. (Sorry, I can be persuaded by argument.) Ignoring for the moment the “irrelevance” of consumption smoothing, your consumption smoothing arguments worked only in the aggregate. Andy, I’m sorry I didn’t get it sooner.

    I see quite clearly now how the private sector agents may smooth consumption yet wind up with zero change in aggregate disposable income and zero change in investment, plus a BBM of 1.0. I can see quite clearly how Wren-Lewis can get a BBM of 1.0 even though “their spending at time t will fall by much less than X” (Protip: If the consumers look at their finances, having cut only $20 in consumption, they will realize they saved more than necessary.) Thus, Wren-Lewis could be said to have been overly generous.

    I am pretty sure this has all been explained elsewhere, so I won’t rehash.

    Also important, the recalculation story leading to a BBM of 1.0 is in fact dependent on the assumption of consumption smoothing.

    Thus, I owe not just Andy, but Wren-Lewis an apology. The consumption smoothing is relevant. There is a perfectly good story there.

    Of course, I still see a hidden assumption required to make sense of Wren-Lewis. Specifically, the government must crowd out a private-sector bridge, yet there must be slack elsewhere in the economy. (Before anyone jumps on this, the BBM of 1.0 also appears if the bridge-builder is unemployed… but the story is um… well… simpler, just not clearly consistent with Wren-Lewis.) But it does NOT require an assumption that the BBM is 1.0. Or 0.5 as I had suggested.

    In any case, it absolutely points out that Cochrane’s proof was deficient.

  39. Gravatar of Doc Merlin Doc Merlin
    17. January 2012 at 06:21

    The most vulgar form of Keynesianism is what they will turn to when their backs are to the wall. Its what happens in recessions.

  40. Gravatar of Major_Freedom Major_Freedom
    17. January 2012 at 06:22

    D R:

    “Ignoring for the moment the “irrelevance” of consumption smoothing, your consumption smoothing arguments worked only in the aggregate.”

    The aggregate was always the context under discussion.

    If, in the aggregate, you admit that consumption smoothing isn’t enough for Wren-Lewis to refute Cochrane (which Sumner has shown by counter-example), then you have conceded Sumner’s only point that he wanted to make.

    Your posts have been extremely muddled, error-prone, and smack of someone who is falling over himself antagonizing someone he has ideological differences with, regardless of whether logic is thrown out the window.

    You keep trying to turn this into a football match of Team Sumner/Cochrane VS. Team Wren-Lewis/Krugman. Any time an individual is critiqued, it somehow means their team won, and the other team lost. This is not what Sumner is doing and yet you keep trying to shoehorn people into an us versus them schoolyard fight.

    Sumner has made it clear that he does not necessarily agree with everything Cochrane said, and he made it clear that he does not necessarily disagree with everything Wren-Lewis said. His ONLY point, that you rabble rousing hippity hop at the barber shop antagonizers are clearly unable to understand, is the one, small, tiny, but nevertheless TRUE argument that consumption smoothing cannot be used to refute Cochrane’s point about AD. That’s ALL Sumner is saying. Sumner showed that the error that underlies Wren-Lewis is a conflation of consumer spending with spending qua spending.

    You’re going off the bloody handrails, making a fool of yourself trying to one up “Team Sumner/Cochrane”, rather than address the ACTUAL point Sumner is making. The point is as clear as day to me, and I am not even a professional. You’re bringing in to this discussion a series of hysterial hand wavey throw everything but the kitchen sink at “Team Sumner/Cochrane”, believing that Sumner is trying to start a war of us versus them, so we all must pick sides or else risk death on the battlefield. Relax, and think.

  41. Gravatar of D R D R
    17. January 2012 at 06:50

    “The aggregate was always the context under discussion.”

    Oh ha ha.

    Just in case that’s not a joke, however, the point is that to say “the private sector” smooths consumption is an outcome. To say that “actors in the private sector” smooth consumption is a behavioral assumption.

    Of course, the two are repeatedly conflated. I admit to having made this error in comments. Still, you can have the latter absent the former.

    Yes, I make errors. It’s very hard not to, when the discussion has been so muddled. At least I admit to them. I fail to see how you can defend the rest of your comments.

  42. Gravatar of Major_Freedom Major_Freedom
    17. January 2012 at 08:00

    D R:

    “The aggregate was always the context under discussion.”

    “Oh ha ha.”

    Is feeling the need to laugh at people’s comments something that you consider to be a mature demeanor?

    “Just in case that’s not a joke, however, the point is that to say “the private sector” smooths consumption is an outcome. To say that “actors in the private sector” smooth consumption is a behavioral assumption.”

    This is not Sumner’s point. Sumner’s point is that consumption smoothing is not a valid foundation in Wren-Smith’s criticism of Cochrane’s argument.

    Why can’t you understand this? It’s trivial (now, anyway).

  43. Gravatar of D R D R
    17. January 2012 at 08:43

    “This is not Sumner’s point. Sumner’s point is that consumption smoothing is not a valid foundation in Wren-Smith’s criticism of Cochrane’s argument.”

    I do understand this. If my income rises by 100 and yours falls by 100, then we may each consumption smooth by my lending you $80. In the aggregate, we see no change in consumption and no change in disposable income. Yet there is still consumption smoothing.

    To say, “there is no [aggregate] income to smooth, therefore, no smoothing” is invalid. Of course, even if my income does not change, an assumption of consumption smoothing is still necessary to show that my consumption does not change.

    Now, I grant Sumner this much– *if* Wren-Lewis is claiming that in the end that aggregate consumption does fall, *and* that the BBM is 1.0 [he didn’t say this in his original argument] then there is a problem. But *if* Wren-Lewis is claiming that in the end, because of consumption smoothing aggregate consumption does not fall [note that a fall of zero is less than a fall of X>0], resulting in a BBM of 1.0 then that is legit.

    Of course, why would Wren-Lewis say “fall of less than X” if he already knew it implied in the end a fall of zero? This was a unhelpful. But Wren-Lewis tried to explain later, there is an initial fall due to smoothing actions, but as actors realize they saved too much (considering the fall in disposable income), they will spend more. It is unnecessary to his case to make this point, as aggregate demand is rising, not falling, with each correction to consumption.

    On the other hand, Cochrane’s text has exactly the same problem– saying “B can spend $1 more” and then implying that “B *does* spend $1 more. This is an assumption.

    Need I even get into the “can” part of that statement? Could B *not* spend $1 more if not for the forced transfer? Was B borrowing-constrained? I understand B might not have wanted to spend another $1 before the transfer, but I’m not clear on how this part of Cochrane’s argument hinges on the transfer.

    And… once more… it’s nice to be able to argue over Cochrane’s discussion over the forced transfer. I would have been nicer if he made an argument surrounding tax-financed infrastructure.

  44. Gravatar of D R D R
    17. January 2012 at 08:51

    “Is feeling the need to laugh at people’s comments something that you consider to be a mature demeanor?”

    Um… reread your comment and then ask yourself if your own comments were entirely within the bounds of reasoned discussion. Or perhaps you think name-calling is a mature response.

    So nanny nanny billy goat!

  45. Gravatar of ssumner ssumner
    17. January 2012 at 15:38

    Everyone, I have way too many old comments going back to previous posts to answer them all. Let’s move the action to the most recent post, where people finally seem to be accepting that I am right. I will read all comments, but won’t respond to stuff like whether S really equals I anymore. Read any intro text if you want to know why. If I miss an important question of yours, bring it to the new post after reading the comments that begin to accept my argument.

    DR, It’s standard new Keynesian theory that the fiscal multiplier is zero when the central bank inflation targets. If you laugh at me you are laughing at all of NK theory. I must have 100 posts on that subject. I do think the Fed is doing something close to inflation targeting, but not exactly. That’s why I don’t think the multiplier is precisely zero.

  46. Gravatar of I Figured Out What Scott Sumner Is Talking About « Uneasy Money I Figured Out What Scott Sumner Is Talking About « Uneasy Money
    18. January 2012 at 13:11

    […] who could touch off the sort of cyberspace fireworks triggered by his series of posts (this, this, this, this, this and this) about Paul Krugman and Simon Wren-Lewis and their criticism of Bob Lucas and […]

Leave a Reply