Vindication! David Glasner (unwittingly) proves my point

Now even fellow market monetarists are beating up on me.  But don’t worry readers, just like in those Hollywood movies where your hero is being beaten to a pulp, one last lunge will save the day and insure victory.  And David Glasner has just put the sword in my hand.

For those who don’t know, this was all set off when I complained about this quotation from Simon Wren-Lewis (which has been endorsed by Paul Krugman):

Both make the same simple error. 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.

I have two objections.  First, they are simply assuming the Keynesian model is true; all zero multiplier models that I can think of are perfectly consistent with consumption smoothing.  An inflation-targeting economy would be one such case.  Consumption smoothing tells us nothing about which model is true.  Indeed that’s my main point.

But my most controversial charge was that Wren-Lewis (and by implication Krugman) forget that S=I.  They constructed an example that implicitly assumed a fall in saving, but forget that because S=I (always and everywhere and immediately and in all possible universes and without any exception), then any fall in saving implies a fall in investment.  (Obviously I’m talking about private domestic saving and investment in the balanced budget case where G-T is fixed and there is a closed economy.)

David’s new post tries to refute me with a Keynesian model full of all the usual hydraulics.  Just to repeat, I never denied that if you assume the Keynesian model is true, then it is true.  But the example he provides actually proves my point.  I won’t go through all the technical details, but here’s the bottom line:

1.  David starts with an economy where C = 200, S = 200, T = 0, G = 0, and Y = 400.

2.  Then he does a 100 tax-financed government spending project.

3.  In his model the new equilibrium is C = 170, S = 150, T = 100, G = 100 and Y = 420.

So the balanced budget multiplier is positive, 0.2 to be specific.  David thinks this refutes my argument.  But I never denied that it’s possible to get a positive balanced budget multiplier if you simply assume the Keynesian model is true.

What David doesn’t realize is that he also proved my point about the need to always remember that S = I.

Go back to the Wren-Lewis quotation above.  He argues that when after-tax income falls, consumption will fall by a smaller amount due to consumption smoothing.  And that’s exactly what David’s example shows.  Then Wren-Lewis argues that in this case the increase in AD will reflect two factors, the change in consumption and the change in government.  I quote once again:

Put the two together and aggregate demand rises.

Note that he said put the TWO together, not put the THREE together.

Why is this important?  Consider the example provided by David.  G rises by 100 and C falls by 30.  That’s certainly consumption smoothing, after all, after-tax income fell by 80.  Wren-Lewis (and Krugman) would claim that the net change in AD is the change in C plus the change in G, which is plus 70 in the example provided by David.  However David calculates a change in AD of plus 20, not plus 70.  What could explain this discrepancy?

Wren-Lewis forgot that if consumption falls by less than after-tax income falls, then saving must also fall.  After all, saving is the part of income not spent on consumption.  But it’s not just the part of income not spent on consumption, it’s also precisely equal to investment.  So the fall in saving equals the fall in investment.

Indeed unlike Wren-Lewis and Krugman, David did remember to change investment; it falls by 50 in his example.  So in fact the correct answer is that the change in AD is equal to the change in C + I + G, not the change in C + G.  That’s because less saving implies less investment.

Some of you may think there is something weird about David’s example, and that there are other cases where Wren-Lewis and Krugman could be correct.  Perhaps, but I don’t see how.  Consider their assumptions:

1.  A tax-financed increase in G.

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

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

That’s why I claimed they forget the S=I identity.  Now obviously I understand that both these guys are very smart, and certainly understand that S=I.  I’m claiming that people often forget that identity when thinking about specific cases.  They forgot that more saving implies more investment.  Just as Lucas and Cochrane temporarily forget that fiscal stimulus might affect M*V by affecting V.

There is one other possibility that someone suggested in a comment section.  Perhaps Wren-Lewis misspoke when he said consumption fell less that after-tax income; perhaps he meant consumption didn’t change at.  And perhaps after-tax income also didn’t change at all.  Maybe, but that’s not what he said, and it would certainly be a weird example of “consumption smoothing.”  And David, who takes Wren-Lewis’s side in the dispute, certainly didn’t read it that way.

To conclude, the simple assumption of consumption smoothing does nothing to refute Cochrane’s claim that the balanced budget multiplier is zero.

PS.  Nick Rowe has a new post that claims:

Therefore, consumption smoothing and Ricardian Equivalence will almost always make the government spending multiplier smaller than it otherwise would be.

PPS.  I have a feeling I made some simple error.  I guess I’ll find out soon.


Tags:

 
 
 

135 Responses to “Vindication! David Glasner (unwittingly) proves my point”

  1. Gravatar of Kevin Donoghue Kevin Donoghue
    17. January 2012 at 07:02

    Anyone who thinks he has figured out what Scott Sumner is on about should lie down until the feeling goes away.

  2. Gravatar of D R D R
    17. January 2012 at 07:14

    “To conclude, the simple assumption of consumption smoothing does nothing to refute Cochrane’s claim that the balanced budget multiplier is zero.”

    I will say it. This is a straw man. Cochrane did not claim that the BBM is zero. He claimed to have proven that the BBM is zero.

    Unless I am wrong. In which case, Cochrane merely asserted it to be true, and any valid counterexample– no matter how farfetched– suffices to disprove his assertion.

  3. Gravatar of Nick Rowe Nick Rowe
    17. January 2012 at 07:24

    Scott: interpret Simon as saying that the increase in desired saving takes the form of an increased (flow) demand for money. Then it all works out. And, directly or indirectly, that is the key mechanism underlying all Keynesian models (whether Keynesians realise it or not, and some do). And also interpret Paul Krugman as having his simple (yet far more general than it looks) babysitting model at the back of his mind.

  4. Gravatar of Alex Alex
    17. January 2012 at 07:25

    Danth’s Law.
    http://rationalwiki.org/wiki/Danth%27s_Law

  5. Gravatar of Kevin Donoghue Kevin Donoghue
    17. January 2012 at 07:28

    Wot D R said.

    Cochrane wrote: “Before we spend a trillion dollars or so, it’s important to understand how it’s supposed to work.”

    He’s not describing a New Classical model or an RBC model. He’s presenting his understanding of how Keynesian models work. It’s quite plain that he doesn’t understand at all.

  6. Gravatar of rob rob
    17. January 2012 at 07:29

    Scott: I think its time to move on. Enough already. At the end of the day, the question of the value of multipliers is an empirical one. As Nick Rowe points out, there is a range of values that these multipliers can assume and they are dependent on the assumptions involved. Period. End of story. Let’s move on to something else.

  7. Gravatar of Nick Rowe Nick Rowe
    17. January 2012 at 07:31

    The only thing really wrong with keynesians is that they aren’t explicit about Sd-Id=(directly or indirectly)excess flow demand for the medium of exchange/account. It all makes sense once you add that in.

  8. Gravatar of David Glasner David Glasner
    17. January 2012 at 07:34

    Scott, It’s always a pleasure for me to be of service to you in any way that I can. I actually was not trying to refute you or to take Wren-Lewis’s side in this little dust-up, I was just trying to make sense of what you were saying and correct the mistakes that I saw in your argument. I don’t particularly care one way or the other what value the balanced-budget multiplier has. Also, my understanding of the point of the exercise was to find out the following: if we stipulate that the balanced budget multiplier is zero without consumption smoothing, would the introduction of consumption smoothing tend to make it positive? I was not assuming that the Keynesian model as such is true; I was just using it as a platform on which to conduct the analysis. The assumption that the model has to be adjusted to take into account a reduction in investment (or some other variable) as a result of an equal increase in G and T means that I was not taking the Keynesian model at face value. There may be other factors to take into account, but there seems to be a tendency for consumption smoothing to increase the implied balanced-budget multiplier. That was my only substantive conclusion.

  9. Gravatar of Tommy Dorsett Tommy Dorsett
    17. January 2012 at 07:35

    This dispute has been beaten like a dead horse, but I think Scott was more clear here than he was in earlier posts. Perhaps all parties would agree that when the velocity-adjusted money stock falls, AD goes down. Keynesian hydraulics or not.

  10. Gravatar of tom tom
    17. January 2012 at 07:36

    Scott: “So the fall in saving equals the fall in investment.”

    But it doesnt imply fall in investment spending (because of inventories).
    Why you are not able to understand this Scott?

    Scott: “They forgot that more saving implies more investment”

    It is not true
    Krugman: “Note, that as such, this does not imply that an increase in savings must lead directly to an increase in investment”
    http://en.wikipedia.org/wiki/Savings_identity

  11. Gravatar of KRG KRG
    17. January 2012 at 07:39

    Some portion of I is cash reserves, idle bank reserves, government bonds, money under the mattress, etc… basically, non-capital investment savings. As long at it’s that portion of I that’s taking the brunt in the drop in I, the implicitly increased overall velocity leads to a positive multiplier as the purely financial market contracts and money moves over toward the more deductible productive capital type of investment.

    To me what seems like it’s missing in your analysis, but assumed as obvious in the more Keynesian analyses- that there is a huge buffer of non-productive savings; investments in mattress padding- that will be cleared by the drop in S before there’s any danger to the portion of I that represents capital purchases.

  12. Gravatar of KRG KRG
    17. January 2012 at 07:47

    And, actually, if unsold inventory is part of S, then a fall in S suggests increased pressure for increasing productive S/I in the future to replace those inventories.

    It would seem to me that the analysis should be looking at expected dS/dt and DI/dt -what pressure such actions put on future velocity- than on any current state.

  13. Gravatar of bill woolsey bill woolsey
    17. January 2012 at 08:26

    Scott:

    I finally get it.

    Consumption smoothing means that tax financed increases in government spending imply a reduction in planned saving. Lewis Wren is wrong to say nothing about the process by which planned saving is brought into equilibrium with planned investment. We cannot ignore investment, because investment and saving are equal by definition.

    I really don’t think the “equal by definition” adds anything but confusion to the argument. What is important is the planned saving and investment must be brought into equilibrium.

    The Keynesian story is that investment would immediately fall by unplanned inventory disinvestment. This signals firms to produce more output. When output increases, planned saving (and consumption) both rise bringing them into equilibrium with unchanged planned investment.

    The traditional story would be that reduced planned saving results in fewer funds being lent, higher interest rates, and so less planned investment. The consumption smoothing just means that the increase in government spending is matched by a decrease in spending on consumer goods and capital goods.

    The reason why the first might happen rather than the second is going to be related to the demand to hold money and the quantity of money. I must admit, that I would like everyone to aways mention that and Lewis-Wren, in particular, failed to tease that out.

    I think you are being unfair to Krugman and Lewis-Wren, just a bit. And that is because Cochrane seemed to say that Ricardian Equivalance suggests that consumption immediately falls by an amount equal to the increase in government spending. If that is true, we never get any further.

    You would never get to the point of arguing that now, the interest rate doesn’t rise enough to reduce investment enough, when combined with the decrease in consumption, to offset the decrease in investment. That you are left with negative inventory investment (or just shortages which don’t show up at all in the accounting.) And so, a motivation to expand production and raise prices.

    If consumption just falls to match the increase in goverment spending, then we never get that far.

    And I am very sure that the point of this is Ricardian equivalance and not really tax financed increase in government spending.

    I know for a fact that there are free market economists who use Ricardian Equivalence as just one of a laundry list of arguments as to why fiscal policy doesn’t work to cure recessions. And they don’t even take it seriously. They are dead set againt budget deficits and government debt believing that it does crowd out private investment. And further, some don’t even think aggregate demand is ever a problem anyway.

  14. Gravatar of ssumner ssumner
    17. January 2012 at 08:29

    Kevin, This isn’t rocket science. Wren-Lewis claims the increase in AD is the change in C+G, It’s actually the change in C+I+G. How complicated is that?

    DR, Yes, for the millionth time, Cochrane didn’t prove anything. So what? I never said he did. How about talking about the argument I did make.

    Nick, That doesn’t help. Because if it took the form of increased demand for money, and if you assume this increased demand for money is not investment, then actual saving would not have declined and consumption would not have been smoothed. Consumption would not have declined less than after tax income. So he’s still wrong. (The details are very complex, as you’d need to make assumptions about whether cash hoarding is a liability of the government, and hence whether the government saving didn’t change (which is the BBM assumption.))

    Can you come up with a set of numbers along the lines of David Glasner, where Wren-Lewis is right? If so, what is the change in consumption, saving, and after-tax income in your example?

    I think your mistake is that your trying to come up with an interpretation where an attempt to save less boosts income enough so that saving and investment don’t fall. Sure, there’s Keynesian models where that can occur, but Wren-Lewis’s explanation implies saving and investment do fall. Indeed he should have said that because investment is fixed, a tax increase will not lead to any decline in after-tax income at all. So consumption won’t decline. At least that would have been technically correct. But he thought he could prove much more, and failed.

    rob, I’ll move on when someone tells my why I’m wrong. So far no one has addressed the points raised in this post.

    Nick, You said;

    “The only thing really wrong with keynesians is that they aren’t explicit about Sd-Id=(directly or indirectly)excess flow demand for the medium of exchange/account. It all makes sense once you add that in.”

    Yes, that’s fine if you assume fiscal policy has no impact on the quantity of money. But why would that be an interesting assumption?

    But you’re still not dealing with the issue in this post. Wren-Lewis doesn’t present the Keynesian model correctly, Glasner does. Glasner says the change in AD is the change in C+I+G, whereas Wren-Lewis says it’s the change in C+G. So Wren-Lewis inaccurately describes his own model. You are trying to defend the Wren-Lewis by assuming he correctly defended the Keynesian model, but he didn’t. The quotation I provide show Wren-Lewis saying something inaccurate–and it’s a very revealing error.

    David, Two points; Nick Rowe says exactly the opposite. I’m in no position to say which of you is right. But until you convince Nick I’ll keep an open mind. More importantly, that was never the issue. There were two issues:

    1. Does consumption smoothing disprove zero multiplier models? No, it’s easy to think of zero multiplier models where consumption smoothing occurs. Yes, if you choose Keynesian hydraulics, you can get the Keynesin result, I never denied that.

    2. Second, I claimed that Wren-Lewis made a basic mistake by ignoring S=I. Your example shows he did. And that shows you were wrong in criticizing me for using the S+I identity in my analysis. It proved very useful in showing Wren-Lewis and Krugman were wrong.

    Tommy, I’m glad it’s clear to you-it seems you and I are the only ones. BTW, the velocity adjust money stock equals AD by definition.

    Tom, Wren-Lewis says S+I is an identity. He’s right. That means it’s always precisely true. Are you saying Wren-Lewis is wrong? Your mistake is to assume that a fall in inventories is not a fall in investment. It is. Changes in inventories are a part of AD. So even if it’s true, as you claim, that the change in investment shows up in inventories, not in fixed investment, I’m still right. AD changes by an amount different from the change in C+G. The change in inventories is a part of the change in AD.

  15. Gravatar of ssumner ssumner
    17. January 2012 at 08:45

    KRG, Well that’s a new one. Is the Keynesian fall back position now that AD is no longer equal to C+I+G?

    I feel like I’m playing whack-a-mole.

    You may be right about the future, but that has nothing to do with my criticism of Wren-Lewis. If he had made that argument I wouldn’t have called him out on it.

    Bill, You said;

    “I finally get it.

    Consumption smoothing means that tax-financed increases in government spending imply a reduction in planned saving. Wren Lewis is wrong to say nothing about the process by which planned saving is brought into equilibrium with planned investment. We cannot ignore investment, because investment and saving are equal by definition.”

    Eureka!!! Someone finally gets my point. I must have been really bad at explaining it. Perhaps I should just give up blogging.

    Later you said;

    “I think you are being unfair to Krugman and Lewis-Wren, just a bit. And that is because Cochrane seemed to say that Ricardian Equivalance suggests that consumption immediately falls by an amount equal to the increase in government spending. If that is true, we never get any further.”

    I strongly disagree. If you read his writings in detail he’s very clear that the fiscal stimulus will reduce I. He says spending falls, not just consumption. Don’t forget that home purchases by “consumers” is investment. Even cars are, (although the government considers them consumption.) Cochrane talks a lot about how big government discourages investment. Remember he’s not a Keynesian, it’s the Keynesians who always false equate spending and consumption.

    Having said that, I’m NOT trying to defend the specific quotation of Cochrane, which we all agree did not prove what he claimed it proved. But Cochrane being wrong doesn’t make Wren-Lewis and Krugman right. They are also wrong.

    I certainly agree that using RE to argue the BBM is zero is highly questionable. I prefer the inflation targeting argument, which seems far more realistic in the modern world.

  16. Gravatar of David Glasner David Glasner
    17. January 2012 at 08:48

    Scott, OK. My new homework assignment is to figure out what Nick is saying and explain why we are (if indeed we are) coming up with different answers. I will report back when I have some results for you. Before I do that, I will repeat that S = I is not an identity, it is an equilibrium condition. That’s non-negotiable!

  17. Gravatar of Morgan Warstler Morgan Warstler
    17. January 2012 at 08:52

    “I know for a fact that there are free market economists who use Ricardian Equivalence as just one of a laundry list of arguments as to why fiscal policy doesn’t work to cure recessions. And they don’t even take it seriously. They are dead set againt budget deficits and government debt believing that it does crowd out private investment. And further, some don’t even think aggregate demand is ever a problem anyway.”

    lulz.

    Bill, pls. note “free market economists” and “actual non-rent seeking businessmen” all think the same thing. It isn’t until you get into the agency problem, where economists don’t want to be the bitches of businessmen (through gvt., through monetary, through college eggheadedry) that economics deviates from that which business does.

    Look, ANYONE who has ever been on the losing end of a deal that got goosed along by the gvt. KNOWS gvt. is only in the way.

    And that is basically EVERYONE in business. Forget the local gvt. meddlers. Just look at The Internet since 1994. From Netscape running on to beg for rent to SOPA, everyone I know in this industry thinks gvt. is wildly inefficient and a net pain in the ass.

    That doesn’t mean people want no govt.

    It just means that that ECONOMICS as practiced by non-free market “economists” is really about “desired inefficiency” and “desired unproductivity”

    People take Econ to figure out how to make money, not to figure out how the meek inherit the earth.

  18. Gravatar of ssumner ssumner
    17. January 2012 at 08:52

    David, Mankiw, Krugman, and Wren-Lewis, all say it’s an identity. So I want you to bash them in a new post, showing how foolish they are to claim it’s an identity. 🙂

    Alex, Quit? Now? When I am converting people like Bill Woolsey? You’ve got to be kidding! I’m going to beat this dead horse until it’s nothing but bloody pulp.

  19. Gravatar of D R D R
    17. January 2012 at 08:58

    “Eureka!!! Someone finally gets my point. I must have been really bad at explaining it. Perhaps I should just give up blogging.”

    Um… so this all boils down to “Wren-Lewis didn’t spell this out sufficiently for me.”

    Okay. I can live with that.

  20. Gravatar of tom tom
    17. January 2012 at 09:04

    Scott have you read this: http://en.wikipedia.org/wiki/Savings_identity

    “Your mistake is to assume that a fall in inventories is not a fall in investment.” But it is not a fall in investment spending.

    “AD changes by an amount different from the change in C+G. The change in inventories is a part of the change in AD.”

    It is true but it doesnt imply that change in output is equal to change in spending (because of inventories).

  21. Gravatar of HBK HBK
    17. January 2012 at 09:07

    “Obviously I’m talking about private domestic saving and investment in the balanced budget case where G-T is fixed and there is a closed economy.”

    Why is everyone talking about a closed economy, when clearly any economy in which such stimulus would take place would be at least somewhat open? Does that not mean that his entire debate is an intellectual curiosity at best? I am sure I am missing something, so please enlighten me…

    Full disclosure: I was taught by Wren-Lewis during my undergrad and am therefore biased in his favour.

  22. Gravatar of Neal Neal
    17. January 2012 at 09:12

    I am a non-economist reader who has been struggling to understand this debate. (By the way, Professor Summer, it would be very helpful for us struggling non-Economist readers if you would identify your variables when using them in discussions or charts rather than assuming that all of your readers are familiar with the common usage of the letter in economist jargon.) Its difficult for me to follow the references to different economic theories, so I’m trying to understand the debate in simple terms.

    Let’s suppose that before the recession, I make $800 per week. I consume $600 of that and the remaining $200 goes into my bank account. The bank then lends this money to businesses, which use $50 of it to make capital improvements and $150 of it purchase inventory. So my $800 weekly paycheck is producing $800 of demand which is broken down as $600 individual consumption, $50 capital improvements and $150 inventory.

    Now let’s suppose there’s a recession. Let’s assume that I manage to keep my job. I continue to consume $600 per week and send the remaining $200 to my bank account. However, a lot of other people have lost their jobs, and they aren’t buying as many things, which means that businesses now have excess inventory. So what does the bank do with my $200? Not being an economist or a banker, I’m not really certain. I don’t think the full $150 would continue going to inventory, because the businesses have excess. I also don’t think the full $50 would go to capital improvements because businesses don’t usually make capital improvements during a recession. So where does the difference go? Reserves? Let’s assume that the bank is now holdng half of my $100 in reserves and distributes the remainder by the same ratio as before the recession ($25 to capital improvements and $75 to inventory).

    Now let’s suppose the government intervenes with a bridge project that costs $125 million and pays for it with an immediate tax increase which is evenly distribution among 1 million tax payers; that makes my share $125. As a cost-saving measure, I stop buying coffee at Starbucks every morning, reducing my consumption by $25 per week. The rest of the difference comes out of the money that I was sending to my bank. I am now sending $100 to the bank each week instead of $200. That means that the bank now has $100 of my money to distribute rather than $200.

    So what has the bridge project done to demand? The bridge project creates $125 million worth of demand, but is that demand entirely offset by the taxes? If you compare what the bank did with my $200 before the rescession to the government-intervention situation, then the new demand does look entirely offset: the bank has $100 less to distribute to businesses and when aggregated over 1 million tax payers that would mean $100 million less demand, which must be combined with $25 less consumption aggregated over 1 million consumers for a total offset of $125 million.

    That’s not really the correct comparison, though, is it? We should instead compare what the bank was doing after the recession to the government intervention situation. If my uneducated assumption that after the recession the bank began holding $100 of my weekly contribution in reserve is correct, then the result looks very different. I assume that the bank would continue supplying businesses with the $25 for capital improvement and $75 for inventory but would now stop accumulating $100 of reserves. The means that the $110 worth of demand created by the bridge project is now offset by only $25 million in reduced consumption. Overall, demand is increased by $85 million.

    So to my mind, the real question comes down to this: what do banks do with consumer savings in a recession economy absent government intervention? The answer to that question strikes me as more relevant than how many economic text books state that S=I.

  23. Gravatar of David Glasner David Glasner
    17. January 2012 at 09:14

    Scott, Please send me references for Mankiw, Krugman, and Wren-Lewis, and I will add that to my growing list homework assignments.

  24. Gravatar of Benjamin Cole Benjamin Cole
    17. January 2012 at 09:16

    The main point to remember is that federal deficits may have some stimulative value, but that value can be snuffed out by monetary policy. And probably monetary policy alone is better. I think even monetizing the debt is better (through QE).

    I say we give Sumner the TKO.

  25. Gravatar of SK SK
    17. January 2012 at 09:43

    David,

    S=I where I=I(planned)+ I(unplanned) This is an identity.

    S=I(planned) This is an equilibrium condition.

    Assuming no G or T.

    No?

  26. Gravatar of Brito Brito
    17. January 2012 at 09:55

    Finally, this post actually makes sense.

  27. Gravatar of Brito Brito
    17. January 2012 at 09:56

    David Glasner

    Read this article: http://en.wikipedia.org/wiki/Savings_identity

    It explains why it holds at all times, but that’s mainly because investment is defined probably not how you think it is defined.

  28. Gravatar of ssumner ssumner
    17. January 2012 at 10:03

    DR, Wren-Lewis assumed that a fall in spending is equivalent to a fall in consumption. Forgeting that saving would also change. That’s a serious error made by lots of Keynesians. It’s the mistake Keynesians make when they call for more consumer spending to boost the economy. That only works if S & I don’t fall by the amount that C increases. (It might or might not). I’d say it’s a very serious error.

    If you take your line we could just say “all Cochrane did was forget to say he was holding M*V constant,” after all later in the paper criticized by DeLong he said something like “of course V could change after fiscal stimulus, but then you’d simply print more money.” It’s easy to make excuses for those you like.

    Tom, Inventory adjustments most certainly are a part of investment spending, every single textbook that I’ve seen considers inventory changes as a part of I. But it doesn’t really matter, because either way it’s a part of spending. Spending includes money spent on inventory accumulation.

    People act like I have novel definitions, open up any undergraduate textbook.

    David, Scroll down a few posts and I quote Wren-Lewis saying:

    “So S=I by definition.”

    https://www.themoneyillusion.com/?p=12696

    The others are in their intro textbooks (macro split for Mankiw) and indeed in virtually every intro macro text ever published.

    By the way, S=I always held in your example. Can you do it in a way where S doesn’t equal I, I’d be curious as to what it looks like.

    Thanks Ben, But first I want to hear my opponents cry “No mas.” When they are so sick of hearing about this that they concede my point out of sheer despair 🙂

  29. Gravatar of o. nate o. nate
    17. January 2012 at 10:10

    Since both sides (Sumner and Krugman) agree that Cochrane made a mistake in his reasoning, it seems to me that this whole controversy comes down to the question of where exactly he went off the rails. Since his original argument was convoluted to the point of being self-contradictory, I would humbly submit that this question couldn’t be resolved by anyone other than Cochrane himself, and assuming that he doesn’t agree that he made a mistake, probably not even by him.

    To me the important thing to remember is that S=I is an identity in a static equilibrium model, so one should be very careful of trying to reach any conclusions about possible movements to a new equilibrium using that identity.

  30. Gravatar of marcus nunes marcus nunes
    17. January 2012 at 10:23

    So I say “No mas”! Everyone will profitably read chapters 3-6 of Richard Froyens Macroeconomics Theories and Policies 9th edition.

  31. Gravatar of ssumner ssumner
    17. January 2012 at 10:24

    Brito, Don’t say you weren’t warned, this quote is from “About this blog:”

    “Welcome to a new blog on the endlessly perplexing problem of monetary policy. You’ll quickly notice that I am not a natural blogger,”

    And of course I agree about S=I.

  32. Gravatar of ssumner ssumner
    17. January 2012 at 10:26

    o. nate, Yes, and Sumner and Cochrane both agree that Krugman made a mistake, so are we equal?

    Thanks Marcus, But there’s only one guy I really want to hear “no mas” from, and he doesn’t teach at Oxford.

  33. Gravatar of D R D R
    17. January 2012 at 10:36

    I am sorry. I thought we had just agreed that “Wren Lewis is wrong to say nothing about the process by which planned saving is brought into equilibrium with planned investment.”

    Which says that if he had anything to say about the process, he did not make his reasoning known.

    But then you wrote “Wren-Lewis assumed that a fall in spending is equivalent to a fall in consumption.”

    Which implies you know exactly what he had to say about the process.

    So once again, we are back into the murky waters of not understanding your point.

    Also, you wrote “If you take your line we could just say ‘all Cochrane did was forget to say he was holding M*V constant…'”

    You don’t need me as an excuse to make that statement. Go ahead. The difference is that Cochrane has made no effort to clarify his thinking, but Wren-Lewis has. Would you like me to re-tell Wren-Lewis story as I understand it? I’m sure you’ll be pleased how I point out some hidden assumptions.

  34. Gravatar of tom tom
    17. January 2012 at 10:40

    Scott: “Spending includes money spent on inventory accumulation.”
    This is wrong. Inventory is accumulated because people dont buy stuff from firms. Just because firms spend money on production doesnt mean that this production is automatically sold.

    http://en.wikipedia.org/wiki/Savings_identity

    “should consumers decide to save more, and spend less, the fall in demand would lead to an increase in business inventories”

  35. Gravatar of johnleemk johnleemk
    17. January 2012 at 10:46

    Scott,

    “Eureka!!! Someone finally gets my point. I must have been really bad at explaining it. Perhaps I should just give up blogging.”

    While I don’t agree at all with your final sentence, it’s true that I really didn’t understand the crux of the debate until I read this post and Bill’s excellent comment summarising the nub of the problem. Now it’s quite obvious you are right.

    Where I think a lot of critics are missing your point is that it seemed you were arguing Cochrane was right or defending Cochrane’s original quotation which Wren-Lewis criticised. And then when it emerged neither was true, people then thought you were making a mountain out of a molehill; sure Wren-Lewis and Krugman made a minor mistake in their criticisms but so what?

    It takes this post + Bill’s comment to explain that the minor mistake isn’t so minor; you saw this as an opportunity to explain that Keynesians often overlook the impact of S and I on GDP. Had this main thrust been clearer I think general reception would have been different.

    Then again, maybe this isn’t your main thrust and I’m still misreading you.

  36. Gravatar of D R D R
    17. January 2012 at 10:47

    tom,

    That seems awfully pedantic. Scott clearly means that the goods are produced (an increase in inventory) and if they are consumed as planned, then consumption rises as investment falls, while if they are not consumed as planned, then… nothing.

  37. Gravatar of FormerSwingVoter FormerSwingVoter
    17. January 2012 at 10:49

    Oh my… it looks as if Scott is new to this whole “arguing with people on the internet” thing…

    “I feel like I’m playing whack-a-mole.

    Yeah, that’s how it works. You shoot down one bad argument, and two more arise in its place. Because when you argue against multiple people at once, you’re arguing against multiple people at once. The folks you’re arguing against aren’t tapped into a single hive-mind. On the plus side, once you shoot down enough of them you’ll start seeing repeats, so shooting them down takes less thought if not time. 😉

    If it’s any consolation, this back-and-forth has been really helpful for us non-economists – it’s made more explicit some of the implicit assumptions that different economists make, and thus made these assumptions more evident.

  38. Gravatar of ssumner ssumner
    17. January 2012 at 10:52

    Neal, Don’t take this the wrong way, but I’m getting swamped with comments (in back posts) and so I don’t have time for in depth answers. One quick answer. Look at the asset side of the bank’s balance sheet. If it’s not financing investment it’s a loan to a person or the government (bonds, reserves, etc) Those are dissaving by the borrower.

    SK, That’s right.

    DR, I haven’t seen Wren-Lewis explain why he overlooked the fall in investment, when discussing the impact on AD. Where did he do that?

    How do you know Cochrane hasn’t clarified his story? I think he has.

    Tom, The definition of “spending” in macro is not consumer spending in macro, it’s C+I+G spending. It includes government and investment spending. That’s in all the textbooks. Even Wren-Lewis includes G in AD, so why not I?

  39. Gravatar of ssumner ssumner
    17. January 2012 at 10:54

    DR, Thanks for the reply to Tom, you seem fair-minded. I’ll try to dig up some Cochrane clarifications later today and do a post.

  40. Gravatar of o. nate o. nate
    17. January 2012 at 10:55

    Scott, The way I see it, the attempt to use S=I as an argument for a fiscal multiplier of zero (ascribed to Cochrane) has been shot down. On that, I think you and Krugman agree. However, as I understand it, you still argue for a fiscal multiplier of zero, due to the monetary reaction function. AFAIK, Krugman has not directly addressed that argument.

  41. Gravatar of Brito Brito
    17. January 2012 at 11:04

    Scott, off topic but: I’ve been given an econometrics project to do on a topic of my choosing, I think I want to do it on something about monetary policy. Any ideas for a model or theory I could test?

  42. Gravatar of John Schultz John Schultz
    17. January 2012 at 11:09

    Scott, I think you are making your argument far more complex than it has to be.

    “Officially,” “consumption smoothing” is the concept that given a temporary increase (decrease) in income that (non-liquidity constrained, perfect) consumers will try to maintain their current consumption by adding onto (eating into) their savings. This is a result of the permanent income theory, etc., we got from Friedman, which is fairly broadly accepted these days.

    Consumption smoothing itself says *ABSOLUTELY NOTHING* about what the effect of a temporary increase in government spending (and taxes) will be on output, which is what Wren-Lewis wanted to talk about. A particular consumption smoothing function can only tell you how C, I and S will change in a particular year *GIVEN* changes in G, T and Y in that year.

    Wren-Lewis essentially assumed that government stimulus increases output and then set out to prove that government stimulus increases output because of consumption smoothing. He assumed his conclusion in his premise. That’s a circular argument and is a logical fallacy.

    QED

    =======================

    The above is only true if you buy into the idea that Wren-Lewis is referring to the specific concept of “consumption smoothing” as used by Friedman et al when he says “If you raise taxes by X at time t, consumers will smooth this effect over time” and isn’t describing the RE idea that consumers (actually the private sector) will reduce spending by an amount equivalent to the present day value of future expected taxes. Because the tax is temporary, the present day value of all future expected taxes will go up far less than the stimulus itself. Therefore, the private sector contracts, but much less than the stimulus in year 0 and overall output increases.

    ===============

    In the end, sadly, this looks like a semantics argument to me.

    Wren-Lewis used the phrase “consumers will smooth this effect over time,” which points to a very specific idea for Sumner that Wren-Lewis actually didn’t mean.

  43. Gravatar of marcus nunes marcus nunes
    17. January 2012 at 11:28

    Scott
    You will never hear “no mas” from “the other guy”. He´ll never give in because he truly believes he´s RIGHT (and I´m not saying he is or isn´t, that´s beside the point here).
    Like the “liquidity trap” thing, it´s a question of “religious faith” with him.
    On top, he doesn´t answer or discuss his commenters…

  44. Gravatar of John Schultz John Schultz
    17. January 2012 at 11:37

    PS – I think your arguments are getting on very shaky ground when you say Keynesians, including the likes of Krugman, usually assume I, S as fixed and/or ignore any effect on I and S and try to make that the “new and improved” thrust of your argument.

    You’d be far better off trying to stick to the line that Wren-Lewis doesn’t know what you mean by consumption smoothing, which is pretty weak tea in and of itself.

  45. Gravatar of John Schultz John Schultz
    17. January 2012 at 11:58

    “1. A tax-financed increase in G.

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

    That’s pretty general. I defy anyone to construct an example using those two assumptions, where the Wren-Lewis quotation is correct.”

    The Wren-Lewis quotation in question being:

    “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.”

    Wren-Lewis’ first sentence is simply the first step in computing a new equilibrium after an increase in G disturbs the previous equilibrium and before he accounts for the reactions of the other variables. Krugman wrote a post about how he thought that Sumner didn’t understand what Wren-Lewis was doing here. Needles to say, that post deeply offended Sumner as this is truly Macro 101 stuff.

    The second sentence is where Sumner thinks he’s nailed Wren-Lewis. Sumner latches onto the words “consumer” “smooth” and “spending” to assert that Wren-Lewis is ignoring any impact on S and I (alternatively assuming they are constant) and to note that by the official freshwater definition of “consumption smoothing”, a specific phrase Wren-Lewis never directly uses, btw, no smoothing is actually occurring here if you do assume S and I are constant.

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

    Yeah, I think you are reading too much into his specific choice of words and then deriving your own meaning from them. Replace the word “consumer” with “private sector” and your entire argument falls to pieces and Wren-Lewis and Krugman are obviously correct:

    “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, [the private sector] 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.”

    Wren-Lewis is referring to the notion of Ricardian Equivalence that Krugman and DeLong had been using to beat up Cochrane and Lucas for a long time already. You simply read him wrong.

  46. Gravatar of D R D R
    17. January 2012 at 12:02

    Scott,

    Thank you. I appreciate that. (Oh, wait, is civility going to make a mess of things?)

    Since we agree on that point on inventories, let us make an assertion that people smooth consumption. (Uh oh…) Now, suppose there is a tax hike of $100. People say, “my disposable income is falling, so I better not consume as much today” and C falls by $20. This shows up as an immediate increase in inventories, so I rises by $20. The net result of which is no change in income, but a fall in disposable income of $100– just as consumers expected.

    At first blush, then, it looks like the [NOTE: non-BB] tax multiplier is zero. The question is what happens next. Do businesses see the increased inventory as a sign they should slow production and start cutting labor and income? Possibly. More assumptions are needed to move on from here.

    Now, what if the $100 is used to buy a bridge? Well, from where does the bridge come?

    1) It comes out of inventories. Then investment falls by $100 as G rises by $100. On net then, C-20, I-80, G+100, T+100. Thus, Y+0 and Yd-100. All is well and the BBM is 0. Call this case “Sumner.”

    2) The bridge comes from increased production. So first, inventories rise by 100,
    (C-20, I+120, G+0, T+100) then the bridge is sold. C-20, I+20, G+100, T+100. Thus, Y+100 and Yd+0. Oops… whoever built the bridge has another $100 income and must now smooth by increasing consumption by 20. Say this consumption comes out of inventories, so in the end, C+0, I+0, G+100, T+100. Hence Y+100, Yd+0. All is well, and the BBM is 1.0. Call this “Wren-Lewis.”

    [Note, here, that the only difference between this and Wren-Lewis’s text is that Wren-Lewis failed to mention that consumption must smooth back UP to its original level when the bridge purchase is actually combined with the tax cut, and so there are no missing savings. Also note that the assumption which kicked this off also is implicit, but when combined with an assumption of consumption smoothing, is *equivalent* to Wren-Lewis’ assertion that the spending multiplier (just the bridge) is 1.0. Proof is left to the reader…]

    3) It comes out of consumption– that is, some consumer doesn’t get to purchase that bridge. So consumption falls by $100 as G rises by $100. On net then, C-120, I+20, G+100, T+100. Thus Y+0 and Yd-100. Oops. This is not consistent with consumption smoothing. Specifically, the consumer who wanted the bridge must now smooth by substituting other consumption goods for the bridge. Now what?

    3a) The additional consumption comes out of inventories. Thus, we wind up with C-20, I-80, G+100, T+100. We wind up exactly the same as in “Sumner”.

    3b) The additional consumption comes out of increased production. That is, first inventories rise by 100 (on net, C-120, I+120, G+100, T+100) then the consumer makes the purchase. Now, C-20, I+20, G+100, T+100. But… Y+100 means Yd+0. Oops. Whoever got paid for that production is under-consuming. If that person’s additional consumption comes out of inventories, then… C+0, Y+0, G+100, T+100. We wind up with Y+100, and Yd+0, so smoothing is working just fine. And we wind up again with “Wren-Lewis.”

    How are we doing?

  47. Gravatar of Philo Philo
    17. January 2012 at 12:26

    To quote David Glasner: “There is simply no basis for saying that saving is the same thing as spending on capital goods, just as there is no basis for saying that eggs are chickens, or that chickens are eggs. Eggs give rise to chickens, and chickens give rise to eggs, but eggs are not the same as chickens.” This suggests a time-lag between saving (first) and investment (afterward). (I don’t understand the suggested time-lag in the other direction–investment preceding saving. Mightn’t investment give rise to consumption rather than saving? Indeed, unsuccessful investment might give rise to neither.) I think Glasner regards saving as accumulating money, which might then be spent on consumption (this would be dissaving, canceling the prior saving; so we might just ignore the combination of initial saving and then almost immediate dissaving, since the net quantity zero is achieved quickly) or on capital goods (investment). But there is a third possibility: accumulating cash (“hoarding”). So according to Glasner there can be saving without spending on capital goods; there can be hoarding. Even in macroeconomics this would be possible, if the money supply increased.

    My question: Is this really how economists use the term ‘investment’, so that it does not include the mere accumulation of extra cash? If so, it is clearly wrong for Scott (along with Glasner, Mankiw, Krugman, Wren-Lewis, and . . .) to say Saving = Investment (unless the money supply is assumed to be held constant; but in a reply to a comment, Scott asks, “why would that [or something close to that] be an interesting assumption?”).

    My comment: One should reject the assumption I have attributed to Glasner, that saving = accumulating cash. Consider the situation in which I have a roast in the refrigerator, but I don’t cook it and eat it tonight: I save it for tomorrow night. That really is saving: I postpone consumption from tonight to tomorrow night. But no cash is involved. (Would Glasner reply that this sort of non-cash saving is trivial in a macroeconomic context, and so may be ignored? Would that be correct?)

  48. Gravatar of math math
    17. January 2012 at 12:29

    Perhaps as a matter of personal taste, but I like Krugman’s position more than Sumner’s one.

    Yes, Krugman didn’t present his view accurately and perhaps made a mistake. But at least he has not tried to change his position back and forth after fact as Sumner has done here.

    First, Sumner said that “consumption smoothing” is irrelevant concept there at all, then he said in reply to me that it is relevant but in other form than Wren-Lewis implied, now he is saying that “simple assumption of consumption smoothing does nothing to refute Cochrane’s claim that the balanced budget multiplier is zero.” What does this expression of “simple assumption” mean and how is it different from the “complex assumption”? He probably should have said that while Krugman made a mistake I myself is not perfect and made a mistake also. So, it’s a draw in my view.

  49. Gravatar of D R D R
    17. January 2012 at 12:39

    Philo,

    Yes, investment does not include mere accumulation of cash. Investment is a category of production. The “cash” in investment is that money which is used to pay the producer.

    Putting money in the bank is not “investment” but rather “saving” and the immediate impact of this is (as you have loaned the bank money) is that the bank has dissaved by an equal amount, so S does not change, and there is no adjustment in the balance of S and I required.

    Now, the bank *may* also turn around and loan out your money to a business to build a factory. Thus, the bank “saves” while the business “invests.” In this, S goes up and I goes up. But that’s a different transaction.

    Why did the business invest? Well, maybe your act of saving lowered interest rates, making it more attractive for the business to expand. But the key there is a lowered rate of interest. Just one example…

    Hope that helps.

  50. Gravatar of rob rob
    17. January 2012 at 12:57

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

    – G increase by 100m
    – T increase by 100m
    – C falls 20 (100m – effect of consumption smoothing)
    – S falls 80m (because C+S must the fall to match T)
    – I falls 80m via decreases in inventories
    – Y increases as firms rebuild inventories
    – S and C increase as Y increases
    The process stops when I = S at the old level of I when inventories have been fully rebuilt and we are back in equilibrium.

    AD has changed by G + C, all other variables are the same as they were at the time t.

  51. Gravatar of D R D R
    17. January 2012 at 13:20

    rob,

    Sounds like my #2 in reverse order of inventory changes. (I just had the business increase inventory in anticipation of the sale.)

    The problem is that as you have ordered things, the third step doesn’t make sense. Once you start with G+100 and T+100, you have so far increased Y by 100, but disposable income is unchanged. There is no reason for C to fall.

    There is something missing before you get to that step. I think you want to move your fall in inventories up next to step 1 where the purchase is actually made.

  52. Gravatar of rob rob
    17. January 2012 at 13:30

    C falls because of the tax increase. C+S together have to fall by the amount T.

    I would not be surprised if I made others beginners mistakes though.

    Of course having I not being affected by the change in S is all a bit dubious in my view anyway, but I think that is what Kenysian’s assume.

  53. Gravatar of D R D R
    17. January 2012 at 13:39

    rob,

    Y = C+I+G
    Yd = Y-T = C+I+(G-T)
    Sp = Yd-C
    Sg = T-G
    S = Yd-C-(G-T)
    C+S = Yd-(G-T)

    So assuming a balanced budget, C+S has to fall by Yd, not T.

  54. Gravatar of D R D R
    17. January 2012 at 13:44

    Or, more simply, balanced budget requires S=I, so

    Y = C+I+G = C+S+T

    so

    C+S = Y-T

  55. Gravatar of ssumner ssumner
    17. January 2012 at 14:15

    Johnleemk, Thanks for that comment. I regret not making myself clearer. Sometimes I make the mistake of equating “simple” and “obvious” My point was a fairly simple one, but lots of simple things aren’t obvious at all. I needed to do a better job of explaining where my focus was, and not just assume readers would follow along.

    former swing voter, Good point.

    o. nate, Yes, and I wouldn’t expect him to agree with me on that. It’s certainly a very debatable point. BTW, there are other factors that also reduce the fiscal multiplier below the norm of the Keynesian model, but I see the monetary reaction as the only one that could make it zero. RE plus crowding out just make it smaller. Ditto for the disincentive effects of taxes.

    Brito, I really have little knowledge of econometrics. Some people do money demand. I suppose PPP as a long run proposition might be interesting (although only loosely related to money.)

    Here’s one possibility. Try to refute MMT by showing that prices levels around the world are much more correlated with the monetary base than the level of national debt (prior to 2008) I did a post comparing Canada and Australia a while back that looks at this issue–it could be found through search. If so, then open market purchases matter when rates aren’t at zero.

    John, Your first comment makes no sense to me. You seem to be criticizing me, but the passage you criticize is not something I wrote. I can’t tell which views are yours and which aren’t. Use quotation marks.

    Marcus. I’d guess you are right.

    John, You said;

    “I think your arguments are getting on very shaky ground when you say Keynesians, including the likes of Krugman, usually assume I, S as fixed and/or ignore any effect on I and S and try to make that the “new and improved” thrust of your argument.”

    Didn’t Krugman assume I was fixed in his reply to me? Even though the Wren-Lewis quote I attacked assumed (implicitly) that I was changing.

    You said;

    “and to note that by the official freshwater definition of “consumption smoothing”, a specific phrase Wren-Lewis never directly uses, btw, no smoothing is actually occurring here if you do assume S and I are constant.”

    This is comical. I don’t even know that the official freshwater view is. Everyone understands the term smoothing, I just used the common sense definition that C changes by less than disposable income. That’s all you need.

    You said;

    “Replace the word “consumer” with “private sector” and your entire argument falls to pieces and Wren-Lewis and Krugman are obviously correct”

    Nice try, but I already found another Krugman post with the same mistake. It’s very clear they are talking about CONSUMPTION, not C+I smoothing. Indeed I’ve never even heard of C+I smoothing.

    DR, I don’t think that’s what people mean by consumption smoothing. I thought they meant that if you took a $100 hit to wealth (in say higher taxes) that you’d cut consumption by something like $5 a year for twenty years. Thus no near term bounce back. I actually thought about that inventory explanation for Wren-Lewis, but even that doesn’t seem to work for me, because consumption smoothing ought to last much longer than any near-term inventory adjustment. But this is provisional. I am sure that Wren-Lewis is wrong as written, but don’t doubt that people could tweak his argument to make it effective, just as I once tweaked the Lucas/Cochrane argument in a previous post.

    After that you kind of lost me on the math, but that may be burnout on my part. First let’s resolve whether consumption smoothing allows that up and down movement in C.

    Philo, Most of the problem here come from confusing individual with aggregate. In an economy with no investment at all there can be lots of saving by individuals. You can put money in the bank. But 100% of the saving is someone else’s dissaving, and hence no aggregate national saving. Aggregate saving only occurs when investment occurs. Otherwise you are just lending money to someone else, or to the government (in the case of T-bonds and cash.) That’s not aggregate saving.

    Regarding the roast, it must carry over until next year to to saving in the national income accounts. And even then only in theory–the government doesn’t include your fridge in “inventories.”

    Math, Nice try. I guess that means I’m winning. I meant it may be important in some contexts in the field of economics, but not in the way Wren-Lewis claimed. I’ve been pretty consistent.

    rob, No, In your example AD doesn’t change, and yet C+G rises by 80. So it actually supports my point.

  56. Gravatar of D R D R
    17. January 2012 at 14:35

    “I don’t think that’s what people mean by consumption smoothing. I thought they meant that if you took a $100 hit to wealth (in say higher taxes) that you’d cut consumption by something like $5 a year for twenty years. Thus no near term bounce back.”

    Ah, that’s it exactly. We do agree on consumption smoothing. The question is why you thought they took a $100 hit to wealth. It’s true, they lost $100 to the government in taxes, but also someone got paid $100 by the government. It depends on the story you tell of production. Is it a “Sumner” or a “Wren-Lewis?”

  57. Gravatar of D R D R
    17. January 2012 at 14:47

    … To be clear, somebody *built* the bridge, increasing wealth by $100, and traded it to the government for cash. The question next question varies by the story you tell, but one is… “Would they have built the bridge anyway to sell privately?” To a Keynesian, the answer might today be “no, because the bridge would be stuck unsold in inventories” That is, they might build that bridge, expanding output, but the builder would see no revenue for all that labor.

  58. Gravatar of math math
    17. January 2012 at 15:19

    “Math, Nice try. I guess that means I’m winning. I meant it may be important in some contexts in the field of economics, but not in the way Wren-Lewis claimed. I’ve been pretty consistent.”

    OK, in that narrow sense you probably won. Wren-Lewis was wrong. I give up.

  59. Gravatar of rob rob
    17. January 2012 at 15:27

    Scott,

    I guess I’m missing something – but why do you say that AD stays the same ? AD is the same after the initial fall in C and S cancel out the rise in G , but won’t AD then increase as firms rebuild inventories ?

  60. Gravatar of D R D R
    17. January 2012 at 15:30

    Scott,

    I realize I may be jumping ahead. Sorry. I believe we agree on the way that consumption smoothing works. I don’t understand what you mean by “up and down movement.” To which of my scenarios are you referring?

  61. Gravatar of Brito Brito
    17. January 2012 at 15:36

    “Here’s one possibility. Try to refute MMT by showing that prices levels around the world are much more correlated with the monetary base than the level of national debt (prior to 2008) I did a post comparing Canada and Australia a while back that looks at this issue-it could be found through search. If so, then open market purchases matter when rates aren’t at zero.”

    Well in the interest of being objective, I will see which model best explains the price level rather than actively trying to refute one model, but yeah that sounds like a good idea. I’ll also see if both have explanatory power simultaneously too.

  62. Gravatar of Mike Sax Mike Sax
    17. January 2012 at 16:06

    Math you give up way too easily. Scott you play an impressive game of 3 card monte-or at least a dogged one. When push comes to shove you still haven’t proven anything.

    You keep claiming that your opponents-who you castigate as “Keynesians” even those like Glasner aren’t fair to your arguments. But you haven’t even touched theirs much less done any justice to them.

    No one has forgotten S=I. The troulbe is that your applying it wrong and thinking you can prove something profound at the behavioral level based on an accounting identity. No matter how many times you claim that S=I means that savings means an increase in capital goods it doesn.

    Nor is the heavy weather you are making about whether or not non-Keynesians believe in consumption smoothing proving anything much less that Cochrane is right about anything.

    You have not proved Simon-Wren wrong not even close. If you wanted to you should start with the fact that nobody is denying the accounting identity S=I just the way you are applying the argument to make behavioral conclusions.

    While you have written I’m counting at least half a dozen posts about this you aren’t any more right than your were at the start. I can argue all day about the earth beling flat it remains round.

    No matter how many posts you write about this until you admit that no one denies the accounting identity S=I just your use of it you won’t even have started to make a point much less defend Cochrane and Lucas.

  63. Gravatar of John Schultz John Schultz
    17. January 2012 at 17:01

    I’ll try again. As I understand your argument in boiled down terms it is:

    (1) Wren-Lewis set out to prove that an increase in G with an equal increase in T in a single year leads to an increase in Y in that year.

    (2) He tried to do this based on the combination of ideas that (a) “If you spend X at time t to build a bridge, aggregate demand increases by X at time t” and (b) “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.”

    (3) The primary problem with this reasoning is that it (a) assumes exactly what he set out to prove, which is circular reasoning!!!

    (4) Then based on that (illegitimate) assumption he argued that consumers will smooth the effect of the tax over time, so their spending will fall by much less than X. He seemingly ignored any impact on I, S or assumed them constant (a common fault amongst Keynesians, apparently) and therefore concluded that his assumed increase in output remains largely in tact in that year.

    (4a) Even with all of Wren-Lewis’ arguable assumptions, C drops by an amount exactly equal to the reduction in after tax income, which doesn’t seem to fit any definition of “smoothing over time.”

    (5) Essentially, Wren-Lewis assumed (most of) his conclusion in his premise, which is circular reasoning.

    Is this basically your counter argument to Wren-Lewis or not?

    ========================

    Assuming that I’ve fairly summarized the core of your argument above, I think your whole argument is based off a misreading / misunderstanding of what Wren-Lewis wrote and meant.

    You saw the words “consumer” “smooth” and “spending” in the same sentence and latched onto them as meaning as “C changes by less than disposable income.” (Btw, the “official definition” of “consumption smoothing” comes from Friedman’s permanent income theory: consumers will attempt to maintain their current consumption in the face of a temporary decrease [increase] in income by eating into [adding onto] their savings.)

    In fact, Wren-Lewis was trying to make Krugman’s Ricardian Equivalence argument that the affect of a temporary increase in T on the present day value of all future expected taxes is far less than the expenditure itself. Since the total expenditure is concentrated in a single year, but the total response to the expected change in taxes in the private sector is spread out over many years, then the stimulus does stimulate in year 0.

    The question of exactly how much C or I accounts for the decrease in the private sector activity is completely besides the point. The point is that C + I decreases by less than G increases, therefore, Y increases.

    So, basically, your whole argument is based on a misreading. Wren-Lewis was trying to make Krugman’s RE argument, which you have astutely not tried to refute yet because it is absolutely correct.

  64. Gravatar of SK SK
    17. January 2012 at 17:02

    Scott–

    Krugman plainly thinks we are below the long-term equilibrium levels of Y and employment. In your view, does Cochrane think we are in continuous LT equilibrium or that we are incapable of using monetary and fiscal policy to move from a current ST equilibrium to the LT equilibrium? If the later, what markets are clearing too slowly to get us to LT equilibrium and why is policy, according to him, ineffective to get us there?

    I take your view as an explicit commitment to remove uncertainty about monetary policy by targeting NGDP and thus get us to the long-term technologically determined real GDP growth rate(3.0%ish) while building in a 2% inflation rate to avoid the problems associated with deflation and downward nominal rigidities. Is this about right?

    Finally, I think using the planned/unplanned investment distinction will help with the Krugman issue when discussing moving from one equilibrium to another in a comparative statics or dynamic sense. Relying on identities can get one into trouble by masking the behavioral issues. (Of course, the identities need to hold.)

  65. Gravatar of John Schultz John Schultz
    17. January 2012 at 17:06

    In my last post I said, “then the stimulus does stimulate in year 0.”

    More precisely: “then the increase in G does lead to an increase in Y in that year.”

  66. Gravatar of Morgan Warstler Morgan Warstler
    17. January 2012 at 17:08

    Sax,

    1. I don’t think econ eggheads should be allowed to use terms outside boundaries of common conception…

    2. But I’m also spent a life-time in organized debate, and on POINTS it isn’t close. As I said, Scott caught DeKrugman sans pants, on a real point – and nothing DeKrugman says will get him around it. The fact the DeKrugman most recently continues to argue without answering the consumption smoothing argument is TRULY all you need to know.

    3. The jews call it pilpul, its where debaters make their chops. Think chess match.

    4. Obviously, at a grand theory level, neither side can whip the other, because the each side has their own internally consistent positions…

    5. BUT when DeKrugman stumbles on the pilpul, it’s a big deal, because AS WE SEE, he can’t just quickly go back and admit the foul. He’s TERRIFIED of knocking over a king, and will try to play to draw.

    6. Stop trying to equivocate, the AFF (bearded commies) put forward a non-responsive critique of Cochrane, the NEG (Scottie) doesn’t have to prove all of Keynes is wrong, he just gets to draw out the blood letting as long the idiots on the other side will keep trying to SQUIRREL.

    Dude, it the grand theory level – what really happens – we get hegemony trumps macro (ie money is the servant exclusively of the money holders).

    At the competitive theory level, you get movements (MM, Neo-commies, Austrians etc.)

    At the blog level, the reputation level, you get pilpul.

  67. Gravatar of math math
    17. January 2012 at 17:18

    Mike Sax: “Math you give up way too easily.”

    Sumner is saying that while Wren-Lewis’ ultimate conclusion (i.e., fiscal stimuli do work) is right, the economic apparatus (i.e., consumption smoothing) he (Wren-Lewis) had employed was wrong. So, Sumner is only asserting that chosen apparatus to prove it was wrong.

  68. Gravatar of johnleemk johnleemk
    17. January 2012 at 17:36

    John,

    If Wren-Lewis had meant to make the Ricardian equivalence argument, fine. But to me, his original post doesn’t give any indication that that is what he meant. He just says that consumers smooth their spending (i.e. consumption; no mention of savings or investment) across time, ergo Cochrane is wrong. Cochrane is wrong, but to take Wren-Lewis’s illustration, any student who tried to prove him wrong by writing down “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” would not get full marks. There is no explicit mention of Ricardian equivalence.

    Moreover, there is no mention of the impact of a fall in I on GDP, even though as you observe, S will surely fall and thus by accounting identity, so will I. Wren-Lewis talks purely about “[consumers’] spending at time t will fall by much less than X”. He is talking only about C + G. Not even C + I, or, correctly, C + I + G. Again, in an exam, the examiner would at best assume the student is a poor/unclear writer and at worst, assume the student doesn’t correctly understand the components of national income. Either way, since the two can’t really be distinguished when the examiner is just reading a single paper, the student would get marks taken off. Sumner is just holding Wren-Lewis to the same standard Wren-Lewis held Cochrane and Lucas to.

    Did Cochrane and Lucas make egregious mistakes? I suppose so, since it’s I think a mistake to make as unequivocal statements as they did about the balanced budget multiplier being 0. Their mistake was quite elementary in the sense that they seem to be applying layman economics (even if behind their statements they have a more sophisticated model of the economy). Wren-Lewis however made an equally elementary mistake. It’s not egregious in the sense that it’s not one a layman could easily make, but he tried to take down Cochrane and Lucas without laying out how fiscal stimulus affects national income in his model. He talked purely about the impact of stimulus on C and G. When you don’t talk about what happens to I, there is no earthly way you can arrive at national income. You just have C + G.

    (On another note I might be misunderstanding something, but I also think Wren-Lewis is quite explicit about the increase in T and increase in G both occurring at the same time (time t). I am not sure how to square this with your assertion that “Since the total expenditure is concentrated in a single year, but the total response to the expected change in taxes in the private sector is spread out over many years” — but I could be misunderstanding it.)

  69. Gravatar of Mike Sax Mike Sax
    17. January 2012 at 17:44

    Math you mean this is what Scott has been quibbling about for a week-that Wren-Lewis is right that fiscal stimulus works but he chose a wrong apparatus-“consumption smoothing?”

    I don’t think so. Scott if you agree with Math let me know-that you think that fiscal stimulus works but that WL used the wrong apparatus.

    Sumner’s goal is to somehow convince people that fiscal stimulus doesn’t work-typical of his style he will claim not that it defintiely won’t work but in all likelihood won’t. But the goal is to discredit fiscal stimulus.

    Usually he employs the idea that the Fed will undo any fiscal stimulus with tighter money. What he never mentions though is that even if this were true it is only based on current Fed policy. A less inflation phobic Fed this would not be true.

    This argument certainly wouldn’t be relevant with a Marriner Eccles at the Fed.

    Scott right now I’m reading a book about FDR during the run up to WWII. You seem to be committing Hitler’s error when he started on another front in the middle of invading Russia. The one thing he didn’t want to do was declare war on the U.S.-of course Japan had something to do with it.

    Before you seemed to want to engage the Keynesians to get their support for NGDP. Maybe you feel that’s secure now lately you have spent most of the time throwing stones at them. It’s funny that Tyler Cowen accuses Krugman of being overly polemical when you had a post called “Why I Hate Keynesian Talk” and that Keynesianism is “mind numbingly stupid?”

  70. Gravatar of johnleemk johnleemk
    17. January 2012 at 17:46

    math,

    You’re saying that examining the premises that lead to a conclusion is at best pedantic. Let’s see what happens if we do this with other issues in the sciences. Maybe Cochrane and Lucas said, say, that climate change is not real. Wren-Lewis responded that it is real because temperatures have changed significantly in his backyard over the last 60 years. Sumner responded that Wren-Lewis forgot to take into account changes in the rest of the world; however it still turns out that climate change globally is true. One cannot argue simply on the basis of incomplete/faulty models or data and expect never to be called out on it.

    I am exaggerating in my illustration, but Sumner isn’t calling Wren-Lewis and Krugman out on a typo. He’s calling them out on forgetting one of the three basic components of the national income accounting identity. This isn’t like a physics problem where the student doesn’t have to account for air resistance or can assume a perfect sphere; it’s more like a physics student misrepresenting a basic formula.

  71. Gravatar of Mike Sax Mike Sax
    17. January 2012 at 17:51

    Morgan there is nothing about Sumner’s quibbling about consumption smoothing that “terrifies” Krugman.

    How’d the Jews get into this discussion? Oh well, I did mention Hitler-in an unrealted way after you mentioned the Jews. Krugman I think is Jewish is that your point?

  72. Gravatar of Anthony T Anthony T
    17. January 2012 at 17:53

    When you’re crazy, you should just go with it and see what happens. We can probably milk this topic for another week.

  73. Gravatar of D R D R
    17. January 2012 at 17:55

    “even though as you observe, S will surely fall and thus by accounting identity, so will I”

    Ah, yes. The “surely” gambit. Care to explain the assumptions underlying that assertion?

  74. Gravatar of johnleemk johnleemk
    17. January 2012 at 17:57

    Mike Sax,

    As I read it, Scott believes fiscal stimulus works in certain models of the economy, including many mainstream models. However he is skeptical that it works very effectively, since many of these models in response to the 2008 economic crisis called for stimulus much larger than what was passed by Congress (a stimulus which was really on the border of political feasibility). He also (I think) thinks the major pitfall of these models is that they don’t account for what response the monetary authorities will have, and take the monetary actor for granted.

    The point you seem to be missing is that if we got a better Fed, we wouldn’t need fiscal stimulus, because we’d get enough monetary stimulus to render it irrelevant. (The Fed’s power to print money dwarfs the US government’s power to spend money.) And as long as there’s a crappy Fed in place, even if attempts for stronger fiscal policy succeed, they will likely be cramped by an inflation-phobic Fed. Any push for fiscal stimulus is pointless as long as there’s a crappy Fed in place.

    What you’re saying essentially is that the central bank has the power to prevent the economy from growing out of a demand crisis, but not the power to help it grow out of a demand crisis.

    Scott doesn’t need to convince people that fiscal stimulus won’t work. He just needs to convince them that it won’t work as well as monetary stimulus probably will. And even if he fails at that, it sounds like he’s at least convinced you that monetary accommodation/stimulus is a prerequisite for fiscal stimulus.

  75. Gravatar of johnleemk johnleemk
    17. January 2012 at 17:57

    “Ah, yes. The “surely” gambit. Care to explain the assumptions underlying that assertion?”

    John Schultz:

    (Btw, the “official definition” of “consumption smoothing” comes from Friedman’s permanent income theory: consumers will attempt to maintain their current consumption in the face of a temporary decrease [increase] in income by eating into [adding onto] their savings.)

  76. Gravatar of D R D R
    17. January 2012 at 18:04

    Um… I didn’t mean, “what is consumption smoothing?” I meant, “what assumptions underlie your assertion that savings must fall?” Sorry if that wasn’t clear.

  77. Gravatar of johnleemk johnleemk
    17. January 2012 at 18:11

    When after-tax income falls, consumers can choose between consuming less, saving less, or some combination of the two. Assuming they opt to smooth a fall in consumption across time, which is I understand what Wren-Lewis would have us assume, they must save less than they planned to.

  78. Gravatar of D R D R
    17. January 2012 at 18:14

    And, why do you think that after-tax income falls?

  79. Gravatar of Major-Freedom Major-Freedom
    17. January 2012 at 18:20

    D R:

    In your “wren-smith story”, you said that the bridge could have been created out of production (contra the “Sumner story” where the bridge is created out of inventory).

    You said that if production increases to make the bridge possible, that Inventories must first increase by 100. But that isn’t necessarily the case. Increased production tends to DECREASE prices. So you cannot create 100 of new spending out of thin air like that.

  80. Gravatar of johnleemk johnleemk
    17. January 2012 at 18:28

    It’s built into Wren-Lewis’s discussion: “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.”

    The only way an increase in taxes can directly force consumer spending down is if it results in a fall in the income consumers have to spend or save, i.e. after-tax income. It is possible to say that Wren-Lewis is taking the assumptions of Lucas and Cochrane, but in any case, I am merely following the path laid down already; if Wren-Lewis accepts them, to respond to his argument one must either point out their mistakes or accept them and reason further from them.

  81. Gravatar of RueTheDay RueTheDay
    17. January 2012 at 18:29

    Back to “it’s essential to think of saving as spending on capital goods”. The problem here is that the definition of I is “spending on capital goods”. So that can’t also be the definition of S. Or you have a circulus in probando fallacy.

    If you want to get formal, the derivation of the S=I identity is:

    1. Income = Expenditure (Y=Y)
    2. Expenditure = Consumption + Investment (Y=C+I)
    3. Income is always either consumed or saved (Y=C+S or S=Y-C)
    :.
    S = I

    We can apply this to a Robinson Crusoe economy:

    Robinson picks 10 apples. Y=10.
    Robinson eats 7 apples. C=7. S=3.
    Robinson adds the remaining 3 apples to his stock of goods. I=3.
    3=3=S=I

    This demonstrates why S always equals I. However, it does not imply a direction of causation. Just because there is an attempt to increase S does not mean we must increase I in the GDP equation. The attempt to increase S, in the absence of an equal desire to increase I, will simply result in a lower Y in which lower values of S and I once again equal one another.

    The problem though, is that this is a barter economy model in which all resources are fully utilized because there is no other outlet in such an economy.

    Introduce money and finance and we now have an economy in which there can be unemployed labor and idle capital and it’s possible, through increased government spending, for example, to simultaneously increase Y, I, and S.

    See:

    http://www.disequilibria.com/blog/?p=208
    http://www.disequilibria.com/ppfa.jpg
    http://www.disequilibria.com/blog/?p=47

  82. Gravatar of D R D R
    17. January 2012 at 18:40

    OK. So here we go.

    Taxes go up by 100, lowering disposable income by 100. This results in consumption smoothing whereby consumption falls by 20 and investment rises by 20. Private spending falls by… zero, which is certainly less than 100.

  83. Gravatar of John Schultz John Schultz
    17. January 2012 at 18:43

    @johnleemk: “If Wren-Lewis had meant to make the Ricardian equivalence argument, fine. But to me, his original post doesn’t give any indication that that is what he meant. He just says that consumers smooth their spending (i.e. consumption; no mention of savings or investment) across time …”

    I think this reading does not give him his full due. He (and Krugman elsewhere) is talking about the reaction of the private economy to an increase in taxes (under full Ricardian Equivalence). In this context, he sloppily used the term consumer as a stand in for the more precise “private sector.” Put another way, if you change his quote to read:

    “If you raise taxes by X at time t, [the private sector] will smooth [their reaction] over time, so their spending at time t will fall by much less than X.” Put that together with an increase in G of X at t and you get an increase in Y at t.

    That is almost exactly the RE argument that a temporary stimulus can increase Y in the short run that Krugman makes here:

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

    (Although Krugman also seemingly restricts the reaction solely to consumption in his example, but I think he does that solely for understandability reasons. His argument is obviously / easily more general and is talking about Y = Private + Government rather than decomposing it further into Y = C + I + G as Sumner assumes they do).

    So, you can either assume that Wren-Lewis was slightly sloppy with his choice of words or that he (most Keynesians actually) automatically assumes his model is true and ignores any impact on S, I.

    Apply Occam’s Razor and it’s fairly clear to me that Sumner’s argument is based entirely on a misreading / misunderstanding of Wren-Lewis (and others).

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

    Sumner:

    I asked:

    “Tomorrow, people have decided to increase their cash holdings to such a degree, that NGDP falls by 99%.

    The NGDP model calls for a governmental agency, the central bank, to go on a buying spree, buying up everything from treasuries, to derivatives, to stocks, bonds, and mortgages. NGDP is still under “target.” So the central bank starts buying cars and television sets and computers. The NGDP target is almost there. So the central bank buys up clothes, and food. NGDP target has been reached.

    Question: In this world, which you must agree is POSSIBLE, I must ask you whether you will reject the model, or reject the government taking over the economy. You can’t have both. You must pick one. Which will it be?”

    You responded:

    “major, Good question. I’d increase the trend rate of inflation so we could avoid a takeover. I see that as the lesser of evils.”

    How can the Fed increase inflation in the scenario above if not through buying up assets, and thus owning of assets, and thus taking over the economy? Are you suggesting straight up checks being drawn on the Fed and then mailed to citizens to goose nominal spending?

    rob:

    You said

    G increase by 100m
    – T increase by 100m
    – C falls 20 (100m – effect of consumption smoothing)
    – S falls 80m (because C+S must the fall to match T)
    – I falls 80m via decreases in inventories
    – Y increases as firms rebuild inventories
    – S and C increase as Y increases
    The process stops when I = S at the old level of I when inventories have been fully rebuilt and we are back in equilibrium.

    AD has changed by G + C, all other variables are the same as they were at the time t.

    Your Y is being treated interchangeably as both in spending terms and in real terms.

    If spending doesn’t change, then a rebuilding of inventories into a given spending, will lead to falling prices.

    You can’t assume C and S in spending increases as Y in real terms increases.

  85. Gravatar of Mike Sax Mike Sax
    17. January 2012 at 18:51

    Johnleem-appreciate the feedback. Sumner’s point about NGDP sounds good or if there is something wrong with it I don’t know what it is.

    It does seem to me that he spends an inoridnate amount of time these days knocking fiscal stimulus.

    In reality the truth is that our debate of this issue is academic anyway as there won’t be meaningful stimulus out of this divided Congress.

    If the Democrats do bettter than anyone dare’s to hope right now possible there could be more in that regard.

    So I don’t really see what Sumner is spending so much time criticizing fiscal stimulus as there is little danger of it being done at the present.

    Even Krugman’s attitude as far as NGDP is concerned is why not-worth a try as we can’t get fiscal stimulus anyway let’s try monetary stimulus.

    Interesting point you make here, “What you’re saying essentially is that the central bank has the power to prevent the economy from growing out of a demand crisis, but not the power to help it grow out of a demand crisis.”

    It makes me wonder what I really do think. LOL. I do believe that the Fed showed it’s ability to negatively effect the economy greatly as it did under Volcker who traded almost 12 percent unemployment for low inlfation.

    A different way to look at it I guess is the MMT way which sees the distinction between the Treasury printing money and the Fed kind of more form than substance.

    In that case a stimlus by the Fed or the Treasury may not be so different.

    Not necessarily wholly endorsing MMT just pointing out that’s a wholly different way to look at it.

    Overall my economic views are a work in progress, and I haven’t wholly dedicded some pretty important qauestions yet though I think about them a lot.

    I did find Sumner’s views on NGDP at least thought provoking but it seems that he has become such a relentless Keynesian basher lately. Again I thought the important thing was stimulus though obviously he prefers monetary stimulus.

    Lately he is more upset that there are people that believe fiscal stimulus works than those who think we need no stimulus.

  86. Gravatar of o. nate o. nate
    17. January 2012 at 18:54

    I think the most charitable explanation to Wren-Lewis is that he meant to invoke the RE argument but stated it clumsily. Krugman probably scanned it a bit too quickly before reposting it on his blog. In his original post, Scott correctly pointed out this error, but in so doing invoked S=I in a way that was prone to misinterpretation. Krugman pounced, again perhaps a bit hastily. I think the bottom line here is that Krugman is not always a close reader of other bloggers, and he has a tendency to assume those who support his conclusions are reasoning correctly and those who don’t aren’t.

  87. Gravatar of D R D R
    17. January 2012 at 19:02

    By the way, I realize I did make an assumption generous to Wren-Lewis. There is no need for investment to rise. Perhaps the fall in consumption leads to a fall in production.

    In that case, the tax cut results in
    C -25
    I +0
    G +0
    T +100
    Y -25
    Yd-125

    Ok, so private spending fell by 25, not 20 or zero. Still, way below 100.

  88. Gravatar of D R D R
    17. January 2012 at 19:04

    … and please take note that the fall is all on the consumption side!

  89. Gravatar of John Schultz John Schultz
    17. January 2012 at 19:15

    Or put another way,

    “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.”

    Sumner interprets this statement as Wren-Lewis assuming that only consumers (i.e. – consumption) will change in response to the increase in government and taxes.

    But why on Earth would Wren-Lewis (and others) assume that private sector spending on capital goods (i.e. – I, such as cars) would be unaffected by increased taxation?

    I see no good reason to assume that they would and see it far more likely that Wren-Lewis simply used the word “consumer … spending” as an imprecise stand in for “private sector … spending.”

  90. Gravatar of johnleemk johnleemk
    17. January 2012 at 19:15

    D R,

    Taxes go up by 100, lowering disposable income by 100. This results in consumption smoothing whereby consumption falls by 20 and investment rises by 20. Private spending falls by… zero, which is certainly less than 100.

    Maybe I’m missing something here but it’s not clear to me where the fall in consumption of 20 and rise in investment of 20 is coming from. Is this going off the example from David Glasner?

    John Schultz,

    Perhaps that’s the case. It doesn’t change the fact that talking about consumers when you mean the private sector is very sloppy thinking. I would say Krugman makes exactly the same mistake in his post, talking entirely about consumption and consumer spending; the only reason we can understand what he’s getting at is he explicitly refers to Ricardian equivalence. If you are writing for an audience that includes economists (which I think is virtually always the case when you use the term “Ricardian equivalence”), you should either use consumption or consumer spending as economists understand them, or clarify what exactly it is you mean when you use these terms as shorthand for non-government spending.

    Krugman and Wren-Lewis are not wrong if you take them to mean non-government spending instead of consumer spending, but the very fact that they gloss over investment so consistently seems arguably meaningful. If a student made this mistake repeatedly in his or her exams, the teacher would surely make a note of it.

    Mike Sax,

    Academics have devoted themselves to the search for truth, so if they see something they think is untrue, it’s hard to begrudge them shouting it from the rooftops. I think he’s always been a Keynesian basher, FWIW. Keynesians generally make the mistake of believing monetary stimulus can do nothing in the current crisis, either because they believe it can never do anything, or because they believe monetary authorities are impotent when rates are at 0%. Those are mistakes Scott has always pounced on.

  91. Gravatar of math math
    17. January 2012 at 19:19

    Mike Sax,

    “Sumner’s goal is to somehow convince people that fiscal stimulus doesn’t work”

    Sumner said to me directly that the fiscal stimulus might work (I guess of course he meant with the proper investment and economy at the under capacity).

    Sumner’s point is smaller. He just wanted to belittle Krugman’s ego. He said about it explicitly.

  92. Gravatar of Brian Brian
    17. January 2012 at 19:24

    Just wanted to interrupt the Krugman wars to relay that I heard an economist for a fortune 100 company state today that the fed was capping growth because of interest rate risk to the treasury. That is, paying interest on reserves, etc., is part of intentional policy to put a lid on growth in the low rate environment. Approx 2.5% NGDP is the cap. No job growth at that level. This economist has pretty close ties to one or more regional fed banks.

  93. Gravatar of D R D R
    17. January 2012 at 19:25

    “But why on Earth would Wren-Lewis (and others) assume that private sector spending on capital goods (i.e. – I, such as cars) would be unaffected by increased taxation?”

    Well, why *should* I be affected? Are we not all in agreement that S=I? Well what does happen (immediately) to S when taxes are raised? Private savings falls, and public savings rises by the same amount. Thus, an increase in T does not *immediately* have any impact on S. So why should it have any impact on I?

  94. Gravatar of math math
    17. January 2012 at 19:25

    johnleemk,

    “Sumner isn’t calling Wren-Lewis and Krugman out on a typo. He’s calling them out on forgetting one of the three basic components of the national income accounting identity.”

    I disagree. Sumner found inconsistency in their logic and offered the plausible explanation of the omission from his point of view. It is just his guess (and nothing more) that they omitted the identity. It is a supportive argument on his part.

  95. Gravatar of D R D R
    17. January 2012 at 19:30

    “Maybe I’m missing something here but it’s not clear to me where the fall in consumption of 20 and rise in investment of 20 is coming from. Is this going off the example from David Glasner?”

    Nope. I’m afraid I haven’t read Glasner. The C-20/I+20 starts from an assumption that there is consumption smoothing in response to the tax cut. Then, as consumers cut back on consumption this shows up first as a rise in inventories.

    It could also show up as a fall in production, in which case, you get C-25 alone. (They have to cut back more in this case, because the fall in production costs income.)

    Make sense?

  96. Gravatar of Major_Freedom Major_Freedom
    17. January 2012 at 19:38

    I wrote this at David Glasner’s blog:

    I don’t understand this fixation on goosing “spending” by government deficits.

    A higher nominal GDP through government deficits doesn’t mean the economy is being “stimulated” in the sense of leading to economic growth.

    Some incomes are consumptive, while other incomes are productive. It is possible for nominal GDP to rise and yet the economy shrinks in real terms, if a rise in consumption spending comes at the expense of investment spending.

    Since we cannot observe counterfactual worlds of what “could have happened”, single minded Keynesians and single minded Monetarists are going to be arguing at each other’s throats forever, debating what “could” have otherwise happened. Keynesians say the money would have just been sitting around, while Monetarists say the money would have been invested.

    The terms of the debate are all messed up.

    Even if nominal GDP is goosed by government deficits, it very well can come at the expense of productive investment spending. This is what Sumner is saying is possible, and yet you and so many other “No, all government spending comes from idle cash” ideologues simply deny it, and you’re falling over yourselves trying to manipulate stupid accounting tautologies all day long trying to refute Sumner on this incredibly obvious point.

    Just step back and think. If the government taxes people, then that means people have less money at their disposal. Agreed? OK, well then surely you must also agree that it is POSSIBLE that the people who are taxed, MAY have otherwise invested that money if they had control over their own money. Sure, they could have stuffed it under their mattresses too, but it’s a damn counterfactual for crying out loud. Nobody can say exactly what people would have done. If any of you deny that investment MAY have been sacrificed, then sorry, Sumner wins, because he’s clearly debating a bunch of blind moles who don’t know when they’re whacked. This is just basic verbal logic that the fixation on math and equations seems to confuse people over. Krugman ignorantly believes that it is verbal logic that leads people astray. But that’s just his positivist brainwashing. The OPPOSITE is the case. One sentence verbal logic wins:

    “If people are taxed, then they have less money at their disposal. Since money can be invested in principle, it means that it is possible that the taxpayers might have otherwise invested with it. Ergo, taxation MAY reduce investment.”

    QED

    At any rate, I have to ask why in the bleeding world is Sumner’s point about investment POSSIBLY being reduced on account of government taxation, how can that obvious point be challenged in any way by anyone who doesn’t have a hole in their head? It’s so incredibly obvious! You take my money, and boom, you don’t know what I would have done with it. That world is forever lost. But a possible world is me investing with it. Surely you cannot claim to know with 100% certainty that I would not have invested it. I mean I’ve heard dogmatic beliefs from people before, but that takes the cake. For it requires the person to be able to have knowledge of a world that doesn’t even exist. Are we economists or are we religious mystics with en eye to other dimensions?

    Please, David, for the love of everything holy, just concede Sumner’s painfully obvious point that taxation can POSSIBLY reduce investment. It’s not hard. Don’t worry, Keynesianism isn’t going to implode by making that concession. Yes, it requires you to view people as POSSIBLY non-hoarding forward looking abstainers from consumption. It’s so hard isn’t it? Government taxation and spending simply must have zero affect on investment, doesn’t it? Talk about dogmatic.

  97. Gravatar of John Schultz John Schultz
    17. January 2012 at 19:43

    In general, I think it is highly uncharitable to always interpret the phrase “consumer spending” as talking solely and only about consumption (C) in the national accounting sense.

    For example, I personally am a consumer in every common sense of the word.

    However, sometimes when I spend money I am actually engaging in investment in the national accounting sense. For example, when I spend money to send my son to college, or purchase a new house or a new car or any other new capital good.

    Let’s say we posit that an increase in my taxes will cause me, a consumer, to reduce my spending.

    Does that mean I will likely reduce my spending on consumption? Does that mean I will likely reduce my spending on investment?

    Most people would think it means that I am likely to reduce my spending on both consumption and investment because all you are really doing is decomposing and classifying what type of goods I am spending my money on. The key point is that my *total* spending on consumption and investment goes down. How that total decrease is decomposed and classified is really not the point.

  98. Gravatar of D R D R
    17. January 2012 at 19:46

    “OK, well then surely you must also agree that it is POSSIBLE that the people who are taxed, MAY have otherwise invested that money if they had control over their own money.”

    Yes. From an accounting standpoint, it is POSSIBLE. But… er… taxes do not immediately reduce national savings, so it is not obvious that taxes induce a fall in investment.

    As murky as all this has been, I find it very very hard to believe that is Sumner’s point.

  99. Gravatar of D R D R
    17. January 2012 at 19:50

    John,

    I don’t see how this matters. The point is, that under very plausible assumptions *total* private spending will fall less than *public* spending, thus raising national income.

  100. Gravatar of John Schultz John Schultz
    17. January 2012 at 19:58

    Me: “But why on Earth would Wren-Lewis (and others) assume that private sector spending on capital goods (i.e. – I, such as cars) would be unaffected by increased taxation?”

    DR: “Well, why *should* I be affected? Are we not all in agreement that S=I? Well what does happen (immediately) to S when taxes are raised? Private savings falls, and public savings rises by the same amount. Thus, an increase in T does not *immediately* have any impact on S. So why should it have any impact on I?”

    That assumes that the tax comes “out of the blue” or that private sector spenders are not forward looking at all.

    When the tax increase is announced, targeted private sector spenders will reduce their total spending in anticipation of having to pay higher taxes / having less after tax income. Now, you can decompose private sector spending into C and I before the tax is announced and into C’ and I’ after the tax is announced and see the composition of the changes in private sector spending, but that is really besides the point. The point is that total private sector spending decreased in response to higher taxes.

  101. Gravatar of John Schultz John Schultz
    17. January 2012 at 20:04

    DR: “I don’t see how this matters. The point is, that under very plausible assumptions *total* private spending will fall less than *public* spending, thus raising national income.”

    Yes, that’s exactly my point!

    I was saying that when Wren-Lewis says

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

    it is very uncharitable to interpret him as saying that only C will decrease and I, S will remain fixed.

    Basically, I’m saying Wren-Lewis is right but used sloppy terminology and Sumner is pedantically trying to pin him to the wall because Wren-Lewis used the word “consumers” rather than being more precise and saying something like “the private sector.”

    Is that clear now?

  102. Gravatar of Major_Freedom Major_Freedom
    17. January 2012 at 20:05

    D R:

    “Yes. From an accounting standpoint, it is POSSIBLE. But… er… taxes do not immediately reduce national savings, so it is not obvious that taxes induce a fall in investment.

    [Facepalm]

    Seriously, like is there something in the Keynesian borg mind that prevents thinking straight?

    DR, you claim that “taxes do not immediately reduce national savings”. But how in the world can you claim that, when the taxpayers never got the chance to do anything with that money…since they were taxed of it?

    You say “reduce” as if the reduction is temporal. But the terms of the debate is based on counter-factuals, not actual history from time = -1 to time 0. Over time, there are constant counter-factuals not taking place because the government is taxing people.

    Sumner is in essence saying that if the government taxes and spends, then it is unwise to ignore investment. You agree that it is possible for investment to fall, from where it otherwise would have been of course. That means you have sided with Sumner on the only point he wanted to emphasize. He is saying Wren-Smith was wrong to ignore investment in his critique of Cochrane.

    If you admit that taking my money and then spending it MAY have reduced investment from where it otherwise would have been, since I could have invested with it, then we’re done. Sumner wins.

    Why is this so incredibly painful to admit? Is it because you are so sensitive about the Keynesian position that refuse to accept that Keynesianism can reduce investment and hence can reduce our prosperity, if the deficits come at the expense of investment?

    Since we can never know what COULD have happened, we can only surmise that there is an incredibly high probability that stimulus reduces investment. After all, the last time I checked, the government didn’t tax just consumption transactions. They tax transactions in the investment sector as well. Surely if the government didn’t tax those investment transactions, then it is POSSIBLE that investment would have been higher.

  103. Gravatar of Major_Freedom Major_Freedom
    17. January 2012 at 20:09

    D R:

    “As murky as all this has been, I find it very very hard to believe that is Sumner’s point.”

    When Sumner sees what I wrote, I can with 99.9% certainty that if he responds to it, he will say that yes, that is what he is saying.

  104. Gravatar of D R D R
    17. January 2012 at 20:13

    John,

    You are making assertions here which are not defensible.

    “That assumes that the tax comes ‘out of the blue’ or that private sector spenders are not forward looking at all.”

    Absolutely not. I start by assuming that disposable income falls by the amount of the tax. I then say that in response, people work less. That is, national income falls by 25% of the amount of the tax.

    C -25
    I +0
    G +0
    T +100
    Y -25
    Yd-125
    Sp-100
    Sg+100
    S +0

    “The point is that total private sector spending decreased in response to higher taxes.”

    As you can see, I have said this is true.

  105. Gravatar of D R D R
    17. January 2012 at 20:17

    “it is very uncharitable to interpret him as saying that only C will decrease and I, S will remain fixed.”

    You have it backwards. If you change the assumption that output was lost to an assumption that lost consumption increases inventories, then investment goes UP, not down.

    C -20
    I +20
    G +0
    T +100
    Y +0
    Yd-100
    Sp-80
    Sg+100
    S +20

  106. Gravatar of D R D R
    17. January 2012 at 20:24

    “If you admit that taking my money and then spending it…”

    Wait just right there. These examples I gave were for JUST the tax increase. (Notice how G doesn’t change?) Don’t get carried away.

    Now continuing:

    “If you admit that taking my money and then spending it MAY have reduced investment from where it otherwise would have been, since I could have invested with it, then we’re done. Sumner wins.”

    Not even close.

  107. Gravatar of Morgan Warstler Morgan Warstler
    17. January 2012 at 20:42

    Sax,

    You REALLY need to stop listening for dog whistles. I said pilpul, all you saw was jews.

    but it gets worse with Keynes.

    I keep reminding you (and everyone else) Scott’s a full bore small government conservative.

    Boxing up the Fed and making it run by futures market is just a FAR TRICKIER way of doing Friedman’s computer. And Scott thought of it! Here’s a hint: getting rid of conservative leadership at fed, doesn’t change under 4.5% NGDP – or the 4% or 3% I want like Woolsey – it gets MORE CONSERVATIVE and is run by the market.

    If you pay attention to the argument flow…. all these eggheads had pretty much decided the monetary trumped fiscal stimulus like 30 yeas ago, but then DeKrugman argued liquidity trap (special circumstances 2008), and confused things – either way that isn’t going to actually work (I prefer the knowledge problem to explain it), but Scott has this nifty response about fiscal gets neutered.

    But somehow, to you Keynes = more free shit for poor people, or social justice, or whatever – which ISN’T what Keynes actually said… nobody every pays attention to the small print with economic theories – Keynes said, when times are good, public employees can’t have raises, when times are good, social benefits can’t be extended – precisely so when times are bad, we can go into hawk, and not gut them.

    Note: Scott’s plan has small print too, and I can’t get him to talk about it. But it is his show, he calls the tune. I did say he’d be smart to keep hitting DeKrugman on this, and it is only smart.

    It really is like one of many chess matches that always play to draw. Now, Dekrugman HAS to lose this one, but instead of ending it quick, he’s literally going to get chased around the board – which benefits Scott.

    John Schultz,

    Let the Oxford man and NYT social justice crusader just say “we mispoke!” You don’t have to channel them – blame them.

  108. Gravatar of John Schultz John Schultz
    17. January 2012 at 21:06

    Me: “That assumes that the tax comes “out of the blue” or that private sector spenders are not forward looking at all.”

    DR: “You are making assertions here which are not defensible.”

    What I meant is that private sector entities will begin changing their spending in response to the tax the moment it is signed into law (maybe even prior). They will anticipate the tax with their actions long before the tax is actually collected.

    Targeted private sector spenders will reduce their spending (i.e. – their consumption and investment) to increase their savings in anticipation of having to pay the tax. The reduced personal spending will lead to a rise in inventories. So, C goes down by X and S (and I) increases by X. Private sector spenders with increasing inventories respond to the build up by reducing their output, so Y falls, with both C and I likely falling. All this happens before the tax is actually collected.

    I agree with you that when the actual tax hits it doesn’t have any effect on S because the private drop in savings is exactly counterbalanced by an increase in public savings. However, I see no good reason to assume that S (and I) will end up exactly where it was before the tax was announced. That is certainly one possible outcome out of the infinite possible ones, but it would likely require some strange assumption about the economy (i.e. – I, S is always ultimately unaffected by changes in Y) to get it to work out exactly that way.

  109. Gravatar of John Schultz John Schultz
    17. January 2012 at 21:23

    WL: “”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.”

    Me: “it is very uncharitable to interpret [WL] as saying that only C will decrease and I, S will remain fixed.”

    DR: “You have it backwards. If you change the assumption that output was lost to an assumption that lost consumption increases inventories, then investment goes UP, not down.”

    I’m not understanding you here.

    Sumner is the one accusing Wren-Lewis of ignoring any possible reaction in S, I due to an increase in Y and G, T. Sumner is saying that Wren-Lewis is saying that any change in Y will only be due to changes in G, T and C — that I, S are assumed constant or unchanged.

    I’m saying Sumner’s interpretation of Wren-Lewis’ english is uncharitable because he is interpreting “consumers … spending” as solely talking about consumption expenditures and not at all about investment expenditures.

    As I said, I’m a consumer by most any definition you care to try on and yet in the real world I actually both consume and invest. If Wren-Lewis said:

    ‘If you raise John’s taxes by X at time t, John will smooth this effect over time, so John’s spending at time t will fall by much less than X.’

    Would you assume that he was only talking about my consumption expenditures and not at all about my investment expenditures?

  110. Gravatar of D R D R
    17. January 2012 at 21:37

    “What I meant is that private sector entities will begin changing their spending in response to the tax the moment it is signed into law (maybe even prior). They will anticipate the tax with their actions long before the tax is actually collected.”

    OK. So if C-25, I+0 (or C-20 I+20) are reasonable responses to an actual tax collection, and we assume that in advance of an actual tax collection that private actors behave *just as if* they had actually been taxed, then,

    C-25
    I +0
    G +0
    T +0 <– acting as though actually +100
    Y -25
    Yd-25 <– acting as though actually -125
    Sp+0
    Sg+0
    S +0

    And even with the fall in output, S still hasn't changed, despite the fall in output. I'm still not seeing your point.

  111. Gravatar of John Schultz John Schultz
    17. January 2012 at 22:10

    This is my behavioral argument for what happens:

    “Targeted private sector spenders will reduce their spending (i.e. – their consumption and investment) to increase their savings in anticipation of having to pay the tax. The reduced personal spending will lead to a rise in inventories. So, C goes down by X and S (and I) increases by X.”

    C: -X
    S: +X
    I: +X

    “Private sector spenders with increasing inventories respond to the build up by reducing their output, so Y [ultimately] falls [by Z], with both C and I likely falling.”

    Y: -Z
    C: -X – aZ
    S: +X – bZ
    I: +X – bZ

    Where a and b are the coefficients (a + b = 1) of the drop in Y that ultimately show up, respectively, in consumption and investment.

    Finally, the tax actually hits but has no further effect on Y = C + I + G because all the reactions already happened in anticipation and the decrease in private savings is exactly equal to the increase in public savings.

    You gave one example where I, S is unchanged and another where I, S is changed. I’m not saying either of those is impossible.

    I’m saying that Sumner reading Wren-Lewis as assuming that I, S is always ultimately unchanged is uncharitable to Wren-Lewis. I, S ultimately not changing is one possibility out of the infinite possible ones and you have to make some very specific assumption about the way the economy works to get it play out exactly that way.

    Sumner interpreting Wren-Lewis (and most other Keynesians) as (always) making that particular assumption is an uncharitable read by him.

  112. Gravatar of D R D R
    17. January 2012 at 22:30

    John:

    “I’m not understanding you here.

    “Sumner is the one accusing Wren-Lewis of ignoring any possible reaction in S, I due to an increase in Y and G, T. Sumner is saying that Wren-Lewis is saying that any change in Y will only be due to changes in G, T and C “” that I, S are assumed constant or unchanged.”

    OK. First of all, the part of Wren-Lewis which has everyone’s attention explicitly refers to the effect of the tax increase in isolation. If you need further convincing that the bridge is in this regard irrelevant, let me offer the following:

    With respect to the bridge (in isolation of the tax cut) Wren-Lewis asserted a multiplier of 1.0. Thus, an increase in G of 100 induces (by assumption) an increase in Y of 100. Which implies that C+I does not change. Now, the increase in Y implies an increase in Yd of 100 as well, so by consumption smoothing we know ex-post that C must have risen by 20 and I must have fallen by 20. The important thing here is that Y goes up by 100, and savings and investment are all accounted for.

    Now at this point you may say, “Gotcha! To tax is to spend, and everyone knows the government is going to have to tax somewhere along the line. Don’t folks act as if they were taxed right now?” That’s a fair argument. So let us go through it again.

    The increase in G of 100 still induces (by assumption) an increase in Y of 100. Which still implies that C+I does not change. Now, the increase in Y (because it came from deficit-spending) effectively imposes a tax of 100, so Yd is now unchanged. Consumption smoothing, ex-post, implies that C must not have changed, nor I. Again, Y goes up by 100 and savings and investment are all accounted for.

    The fact is, none of this is important because Wren-Lewis asserted the spending multiplier of 1.0, so it doesn’t really matter what you pick for the change in C. It must be that C+I=0 so that Y=C+I+G=G. All private spending is accounted for.

    Ok. So what about the tax cut? There are two ways to proceed. First, we may *presume* that because Wren-Lewis did not mention investment when he went over this section, that he had assumed investment did not change. That is fair to me. So let us fix I. Well, now if C falls by 25 then disposable income falls by 125. That gets us the proper ratio in the consumption smoothing. But can C fall by more? No, because if I is fixed, then a fall in C reduces disposable income 1:1, and the consumption smoothing assumption will be violated. This, Wren-Lewis must have meant C fell by 25, which is much less than 100.

    What is the other way to proceed? Well, we can say “disposable income fell by 100, and smoothing consumption implies a fall in consumption of 20.” Now, it matters whether or not the consumption came out of inventories or production. Suppose that businesses adjust to any change in inventories by adjusting production so that inventories do not, in the end, change. So again we must wind up with I unchanged and the only result consistent with this and consumption smoothing is for C to fall by 25, just like before.

    On the other hand, if the fall in consumption results in an increase in inventories (note that’s a positive change in investment) then saying Wren-Lewis failed to count changes in I points out that private-sector spending is even larger than Wren-Lewis was claiming. For example, C-20 I+20 results in a fall in disposable income of 100. Everything works out and is accounted for.

    Hope that helps.

  113. Gravatar of D R D R
    17. January 2012 at 22:33

    Tax hike. Obviously I was referring to a tax hike. Sorry.

  114. Gravatar of D R D R
    17. January 2012 at 22:43

    “I’m saying that Sumner reading Wren-Lewis as assuming that I, S is always ultimately unchanged is uncharitable to Wren-Lewis. I, S ultimately not changing is one possibility out of the infinite possible ones and you have to make some very specific assumption about the way the economy works to get it play out exactly that way.”

    Then I don’t understand what you and I are debating. So long as Wren-Lewis made *any* consistent argument that the multiplier for tax-financed infrastructure is positive, then it refutes Cochrane’s assertion that the such a multiplier must not be positive.

  115. Gravatar of John Schultz John Schultz
    18. January 2012 at 00:12

    DR: “Then I don’t understand what you and I are debating. So long as Wren-Lewis made *any* consistent argument that the multiplier for tax-financed infrastructure is positive, then it refutes Cochrane’s assertion that the such a multiplier must not be positive.”

    Sumner criticizes Wren-Lewis as starting off by assuming that an increase in G, T yields an increase in Y when that is exactly what Wren-Lewis is trying to prove. Effectively assuming your result is generally not considered good logic.

    Sumner claims that there are plenty of internally coherent models of the economy where an increase in G, T would leave Y unchanged: dG = -d(C+I). For example, when the additional government spending perfectly crowds out private sector spending. In that case, all consumption smoothing tells you is how the drop in C+I is proportioned between C and I — it doesn’t yield that tax balanced increased govt. spending increases output.

    Sumner is saying that Wren-Lewis is claiming he is right because he assumed he is right.

    Furthermore, Wren-Lewis used the phrase “consumer … spending” in describing the reaction to increased taxation, which Sumner interpreted to mean that Wren-Lewis also assumed that I, S were fixed, which apparently is a common failing of Keynesians according to Sumner.

    ============

    I’m arguing that Wren-Lewis was actually trying to make the Ricardian Equivalence argument that Krugman made. Here’s my version of that argument:

    Assume full Ricardian Equivalence. Assume that the private sector will perfectly counterbalance any permanent increase in government spending to leave output unchanged (i.e. – fiscal multiplier of zero).

    Given a new govt. policy that deficit spends $X additional dollars this year, what will be the effect on Y be during this year?

    By RE, the private sector will assume the total cost of the expenditure must eventually be paid for through increased taxes and will alter their behavior immediately in response. By the counterbalancing of permanent expenditure assumption, the private sector will permanently reduce its spending by the change in the present day value of all expected future taxes.

    Because the expenditure is a “one time charge” the change in the present day value of all expected future taxes will be far less than the expenditure itself. Because the expenditure adds $X to G during the year it is spent but private sector spending only declines by far less, therefore, Y increases.

  116. Gravatar of Jon Jon
    18. January 2012 at 03:53

    Scott,

    It seems to me that you are confusing changes in desired saving with changes in actual saving (I think that’s what Krugman was getting at in his “comparative statics” post). Consumption-smoothing implies (in this case) a fall in desired saving. However, this does not imply that actual saving increases, and it is not necessary for consumption to fall by less than after-tax income (or, indeed, at all). Instead, the desire to smooth consumption (to save less) may increase demand and hence raise output. The increase in Y increases desired saving again, so investment doesn’t change. Thus the change in AD is equal to the change in C + G.

    Now, obviously you can write down models where this isn’t true, as Glasner demonstrates. It depends on your assumptions about prices and interest rates. In a classical, flexible price model, output is not determined by demand, but by supply, and increases in G are crowded out by higher interest rates (reducing I). But Lucas and Cochrane weren’t assuming such a model, there were asserting some sort of universal truth based on accounting identities. That’s a fallacy.

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

    D R:

    “Yes. From an accounting standpoint, it is POSSIBLE. But… er… taxes do not immediately reduce national savings, so it is not obvious that taxes induce a fall in investment.”

    Given that you admit it is possible that I is reduced on account of the taxation, then “national savings” is reduced to whatever extent I is reduced, since wherever there is I, there is S.

    Your mathematical examples consistently fail to take this in account, which means you are merely assuming the Keynesian model true.

  118. Gravatar of JKH JKH
    18. January 2012 at 05:07

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

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

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

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

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

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

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

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

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

  119. Gravatar of Adam (ak) Adam (ak)
    18. January 2012 at 05:43

    Let me make the second and final attempt… there is a fundamental error made in the reasoning contained in the sentences: “there are plenty of internally coherent models of the economy where an increase in G, T would leave Y unchanged: dG = -d(C+I). For example, when the additional government spending perfectly crowds out private sector spending”.

    I am not questioning the statement that “effectively assuming your result is generally not considered good logic”. I am not a follower of the desexed Keynesianism preached by prof Krugman and the others from his camp.

    I firmly reject the idea of financial crowding out but this is not the point I want to make.

    Why hasn’t anyone discovered that even if there is no Keynesian multiplier in action nor any investment acceleration effect then still the value of the balanced multiplier will be closer to 1 than to 0? We add a flow at the firms sector, that additional flow goes to the households sector and then we drain that extra flow via taxation. GDP is increased but the government balance is unchanged.

    The key point is that the government drains by taxation every extra dollar it spends. We cannot say that the consumption will fall because of the increased taxation as the disposable income will not be altered at all by the similar increase in G and T.

    Let me give you a silly example. Imagine the government is communist (I spent a half of my life living in a communist country so I know what I am talking about). The government orders everyone to work every Saturday for free and clean up the streets (which are dirty by default in a communist country). Will this alter the value of GDP? Certainly the value of services produced in the country will increase by the value of the so-called “voluntary social work”. Will the unpaid work on Saturday change the value of C or I? I am assuming no crowding out of real resources (nobody cares about cleaning up the streets so cleaning companies won’t complain about unfair competition and nobody works on Saturday on his/her own). There is absolutely no reason to assume that C, S or I change much (OK in a command economy these are determined by the Communist Party so the example is really silly). Now let’s assume that the government has reformed itself, started listening to IMF and is no longer communist but fully market-oriented. It now pays the workers on Saturday evening $25 for the cleaning work they have just completed. However the government doesn’t want to run a budget deficit so they confiscate these $25 from the workers on Monday morning. This confiscation is called taxation. Does this change C and I? Certainly the value of GDP increased by $ (52*25*number_of_workers). The value of G and T has increased by the same value but this is it. Yd should not change. People may get angry but so what? Aren’t they more angry in the real-life Greece?

    So what’s the difference between ordering a new bridge and cleaning up the streets provided that both activities engage idle resources? (There is no need for a stimulus when the capacity utilisation is close to 100% and I am not considering this case).

    As I stated before, even if financial crowding out effect exists (I reject that idea), the precondition for crowding out is not met in our case (a balanced budget multiplier) as every dollar taken from the financial markets is returned back. The same applies to Ricardian Equivalence – even if such an effect existed, our case would not have anything to do with borrowing money now and fretting about increased taxes later. The extra money is not net-borrowed by the government. Similarly there is no persistent increase in the net position (balance) of the government as (G-T) remains the same. Since Yd of the private sector remains the same unless C and I change autonomously there is no reason to apply the consumption smoothing theory (again I disagree with this “invention” but this is another issue).

    Someone may argue that the money for the services ordered by the government is not paid to these people who are taxed. This may cause some changes in the redistribution of the income between C and I but claiming that this will lead to a dramatic decrease in Y is baseless.

    Of course knowing G,C,T and I before the stimulus and relying on the sectoral balance equation only will not allow us to predict the value of C, S and I afterwards. More information is needed to close the model as it is under-determined. If C, S and I don’t change on their own then Y will increase by the value of the stimulus, Yd will remain the same and the multiplier will be equal to 1. In theory one can claim that the level of irrationality of the private sector agents is such that they will reduce their consumption by exactly the same amount as extra government spending but I really don’t see any reason why would anything like that happen.

    Let me finish this comment by quoting my Chinese colleague from the office whom I told about this blog. I asked him whether he would think that paying unemployed workers to build a bridge and then draining the extra income would upset the private sector. “Of course the Americans don’t understand anything. We haven’t got a crisis in China for the last 30 years because our government knows how to look after the economy. You cannot just leave it to the markets because you will have a crisis every 10 years or so”.

    Is he right or is he wrong? Is Keynes more dead than Milton Friedman?

    Let me remind everyone that these algebraic equations are similar to what my daughter studies at high school – this is not rocket science. If several professors of economics cannot agree whether ordering a road upgrade will screw up the private sector in the US and deepen the crisis or maybe help reduce the unemployment a bit and fix the crumbling infrastructure – something is seriously wrong. This is not just theorising about rational expectations and we cannot separate economics and politics. 10 maybe 15 more years of the stagnation/austerity in the US/Europe and I will have to learn the third “first language” in my life that is Mandarin.

  120. Gravatar of TheMoneyIllusion » And a cherry on top! (Why Cowen is wiser than Krugman) TheMoneyIllusion » And a cherry on top! (Why Cowen is wiser than Krugman)
    18. January 2012 at 06:39

    […] I realized that his example actually supported my argument.  So I did a post explaining why, and I think it’s fair to say that things turned around (at least in the comment section a […]

  121. Gravatar of D R D R
    18. January 2012 at 07:02

    “Why hasn’t anyone discovered that even if there is no Keynesian multiplier in action nor any investment acceleration effect then still the value of the balanced multiplier will be closer to 1 than to 0? We add a flow at the firms sector, that additional flow goes to the households sector and then we drain that extra flow via taxation. GDP is increased but the government balance is unchanged.”

    I think this is because there are different questions herein considered. First, is the multiplier positive; second, is the multiplier greater than 1.0?

    It is certainly the case that the multiplier *can* be zero. If the economy is running at full capacity– both labor and capital are at maximum utilization and cannot physically produce anything more in a given period– then nothing whatsoever can result in an increase in GDP. If you are at the limit right now, then by definition you are at the limit.

  122. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 07:07

    Jon:

    “Instead, the desire to smooth consumption (to save less) may increase demand and hence raise output.”

    That doesn’t follow. A higher nominal demand does not raise output.

    Only if there is continued saving and investment specifically, can output be raised.

    It is no argument against this to say “Well without demand for output, there won’t be investment in producing output.” While true in itself, it is not a counter-argument, because the demand for output and the demand for factors of production are in fact in competition with each other.

    Putting forth a demand for output immediately implies that one is prevented from putting forth a demand for everything else, including demand for input factors. It is in principle possible for the demand for output to be raised to such a degree, that there is no demand for input factors at all. Just imagine if every business took 100% of their revenues and consumed with them. The entire spending stream would go to demanding consumer goods, and there would be zero demand for either labor or capital goods. Imagine what the economy would look like if that happened. Aggregate demand is no lower, there is still “spending”, and yet the economy collapsed and wage employment went to 0%.

    No, aggregate demand doesn’t tell us the whole story. Aggregate demand rising does not rule out a shrinking real economy. If too much consumption spending takes place, and not enough saving and investment spending, then even though aggregate demand is not falling, what happens is that capital goods are used up and worn out in production without being replaced (or more than replaced as is necessary in a progressing economy). In order to replace used up and worn out capital goods, there must be an adequate demand for them. Too little demand and they won’t be produced at an adequate enough scale.

    Thus, government policies focused on stimulating “aggregate demand” via consumption spending are quite likely destroying the health of the economy, if the consumption spending they bring about is too high relative to saving and investment. The greater the deficits, the greater the chances that real wealth is being destroyed on net.

    It should not be surprising to those who actually understand how economies grow, that the US economy went from being an industrial powerhouse based on saving and investment and small deficits, to a consumer driven economy based on going deeper into debt and large deficits.

    People who focus on “aggregate demand” as being the source of prosperity are going to be in for a rude awakening once the debt financed consumption era is over, where even the government, who took over the debt financed consumption binge, can’t borrow any more to keep it going.

  123. Gravatar of There is no such thing as fiscal policy « The Market Monetarist There is no such thing as fiscal policy « The Market Monetarist
    18. January 2012 at 07:11

    […] for children to see their parents fight. I feel a bit like that when I see two of my heroes Scott Sumner and David Glasner discussing fiscal policy. The whole thing started with Scott picking a fight with […]

  124. Gravatar of John Schultz John Schultz
    18. January 2012 at 07:31

    “It now pays the workers on Saturday evening $25 for the cleaning work they have just completed. However the government doesn’t want to run a budget deficit so they confiscate these $25 from the workers on Monday morning. This confiscation is called taxation. Does this change C and I? Certainly the value of GDP increased by $ (52*25*number_of_workers). The value of G and T has increased by the same value but this is it. Yd should not change. People may get angry but so what? Aren’t they more angry in the real-life Greece?”

    I think this reasoning is flawed.

    The government could order everyone to work virtually every waking hour that they aren’t already working, pay them, and then turn around and (immediately) confiscate what they paid them through taxation. By your reasoning, C and I would remain unchanged because after tax income would remain constant but G and, therefore, Y would go through the roof.

    The flaw in your reasoning is that we have real world experience with the fact that this doesn’t work in practice as you described precisely because “people get angry.” Actually, they get sullen because most slaves resent their position and produce further away from their full potential than if you allowed the market to set their compensation and let them determine their own work. The end result of your “papered-over” govt. mandated slave labor is that actually Y increases nowhere near the increase in G, implying that C+I does fall.

    This is the core argument of why in the real world we observe that free market economies achieve far better output than communism.

  125. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 07:50

    D R:

    “It is certainly the case that the multiplier *can* be zero. If the economy is running at full capacity- both labor and capital are at maximum utilization and cannot physically produce anything more in a given period- then nothing whatsoever can result in an increase in GDP.”

    This is incorrect. You are conflating spending with real productivity.

    Even if the economy is “below capacity” in real terms, then the BBM can still potentially be zero. The same reason why 100 in tax financed spending CAN reduce consumption and investment together by 100 in a full capacity world, is the same reason why 100 in tax financed spending CAN reduce consumption and investment together by 100 in a less than full capacity world.

    Taxation comes from utilized resources and labor, not idle resources and labor. Taxation is charged on transactions, not cash balances, which means taxation falls on non-idle resources and labor.

    “If you are at the limit right now, then by definition you are at the limit.”

    Astoundingly enlightening.

  126. Gravatar of D R D R
    18. January 2012 at 07:56

    Again, I agree. My point there was that it is pointless to argue that the multiplier cannot possibly be zero. Why are you arguing?

  127. Gravatar of Major_Freedom Major_Freedom
    18. January 2012 at 08:21

    D R:

    “Again, I agree. My point there was that it is pointless to argue that the multiplier cannot possibly be zero. Why are you arguing?”

    I am not arguing with you over whether or not the BBM is zero or positive. Pay attention.

    In that particular set of assertions you made, you said that the BBM can be zero, if there are no idle resources or labor. But the presence or idle resources and labor are not even decisive or necessary to have a BBM of zero.

    It would be like saying “The BBM can be zero, if more people visit Germany this year than France.”

  128. Gravatar of Morgan Warstler Morgan Warstler
    18. January 2012 at 08:36

    adam, I smell and MMTer.

    John Schultz, your “people get angry” is the BEST SMARTEST thing you have said.

    In reality, econ is much more like herding cats, whats more increasingly intelligent cats.

    The cats of the 1930’s are not the same cats as today. And frankly, specific policies that you could do to the functionally retarded populace in the 1930’s would NEVER fly today.

    I take this “people get angry” to its natural conclusion.

    All that matters is voting and money.

    Only some people vote.
    Only some people have money.

    And what I call the A Power (the Tea Party), those people in the private sector who expect to spend part of their life in the top 20% and ALWAYS VOTE – they are the deciders. And oh, BTW, they own 200M guns.

    The A Power truly has veto authority on all monetary policy and all political policy – everything has to at minimum do them no harm.

    They are the hegemony.

    I think therefore, economists should spend there time working within the box of “what do the hegemony want?” instead of “what can we try to do to them?”

    Any economist not working in service to the Tea Party (those with money and votes), I think is just tilting at windmills, wishing in one hand and shitting in the other.

    Scott, doesn’t see the world this way.

    He sees 1930, and thinks we are repeating it. That the rousted up bottom half will RUIN US if the Fed screws up monetary policy.

    Even in the face of the Tea Party 2010, he thinks that Obamacare and 4 more years of Obama is the price we are going to have to pay for not handling monetary policy correctly in 2008.

    To Scott, Obama is our hair shirt.

    I suspect that this “nothing has changed since 1930” mindset comes from Scott studying the GD, and that he’s such a technophobe.

    Anyhoo, when I win our bet in November, I will grab Scott’s conch and do a snoopy dance. Until then, let’s enjoy watching DeKrugman get taken down a few pegs.

  129. Gravatar of Cam Cam
    18. January 2012 at 08:37

    We can even remove the tax part of the equation, and try to answer the following question:
    -if the government pass a law forcing every consumer to buy a bridge, will it stimulates the economy?

    Cochrane argument seems to be that no, because consumers will reduce their non forced consumption by the same amount they spend on the bridge (and by Ricardien equivalence that would be the case even if they didn’t need to pay the bridge cash), so their individual amount of consumption and saving will be unchanged, and there will be no global effect.

    Krugman and Wren-Lewis seem to say that Cochrane is wrong, the first step is not null even for rational consumers, because consumers will tend to smooth change of their non forced consumption; so on the first year, their total consumption (forced and non forced) will increase (decreasing their saving). They are not saying it is the end of the story; but simply that the first effect would be to push up consumption, that is, a stimulative effect. Which equilibrium the economy will converge in reaction to this stimulus is model dependent (that’s were the effect of reduction in saving will play for them), and the final effect on GDP will generally not be equal to the change of consumption from this first step; it could even be zero if the economy is already at full capacity or if the Fed react to offset this increased in desired consumption (thought that’s not the point here).
    The important point is that Krugman and Wren-Lewis argument is just for the first round of effect (in which direction the government action will tend to push the economy), but not directly on where the economy will settle in reaction to this initial push (that’s how I understand Krugman explanation on Comparative statics method). But more or less implicitly they suggest that, in the condition of this particular argument from Cochrane (a depressed economy and a central bank that does not try to offset the stimulus), such increase of consumption will be globally stimulative.
    I understand Sumner argument here as saying,firstly, that this is a strong assumption, already assuming Keynesian model to be true; however the example given (inflation targeting central bank) make me wonder if he speaks generally or in the particular case of a depressed economy (once again,no one I think would deny central bank can offset the stimulative effect if it wants,that’s not the point).
    The second argument is that the final equilibrium depends on the effect of reduction in saving and does not correspond directly to the change of consumption in the first step of Krugman and Wren-Lewis; as it is what Krugman is saying in his explanation on Comparative statics method, I don’t see that as contradiction!

  130. Gravatar of Adam (ak) Adam (ak)
    18. January 2012 at 12:24

    I think that the debate is shifting more towards addressing the real world issues – money and power. At least the veil of useless models not describing the reality has been partially lifted.

    John Schultz – of course mass coercion works only for a while. The reasons “orthodox” communism failed were a bit more complex though (see “From Marx to the market:
    socialism in search of an economic system.” WÅ‚odzimierz Brus, Kazimierz Laski, 1989).

    I dare to say the notoriously inefficient system of Real Socialism has been fixed enough to be able to outcompete the neoliberal financial capitalism and that theories of WÅ‚odzimierz Brus played a significant role in early reforms of Deng Xiaoping.

    Please consider the essence of the change which occurred in China in the 1980s and compare it with Russia or Central Europe. This is the starting point of my analysis – I spent several months trying to understand what kind of economic policy is pursued in China. My conclusion is that they use the economics of Keynes and Kalecki as a tool of choice to achieve their goals – but they are totally pragmatic state capitalists otherwise.

    Is anything applicable to the current situation in the US there? I am not arguing for copying their model but maybe learning something about the models they use to manage the economy.

    Does the fact that people hate coercion (or rather love capitalist free market coercion such as forcing a double-digit number of people out of job the market but hate state coercion) mean that BBM==0 rather than 1? I disagree. Keynesian polices have nothing to do with command economy. This is a typical “McCarthian” argument. Whether a certain policy will be accepted is another issue.

    Morgan Warstler, yes I agree with the statement that “the A Power truly has veto authority on all monetary policy and all political policy – everything has to at minimum do them no harm. They are the hegemony.” That’s why I think that it is quite likely that the mechanism of the state self-destruction has already been activated in the US and to some extent, the EU. The “1%” in the US has very limited interest in winning the race with the mercantilist BRIC countries in terms of developing productive capacities and new technologies. Please consider the role of magnates in 18th century Poland or the role of the Janissaries in Turkey in the similar period.

    Free market worship and libertarianism will fix as much as “golden freedom” and liberum veto in Poland I am afraid.

    Once the Chinese grab a significantly larger chunk of non-renewable resources everything will become “rare earths” to the West I am afraid.

    They will be able to do it precisely because the Chinese know how to use spending multipliers and they herd cats in the right (or rather “left”) direction.

  131. Gravatar of Paul Krugman’s Been Hanging Around in Bars Paul Krugman’s Been Hanging Around in Bars
    18. January 2012 at 21:30

    […] forgive Dr. Krugman for hanging around in bars. After all, within the past month Nick Rowe and now Scott Sumner* have totally blown up his positions with irrefutable numerical illustrations, in two different […]

  132. Gravatar of Morgan Warstler Morgan Warstler
    18. January 2012 at 22:15

    Adam,

    The 1% are beatable daily… They are the B power, the VERY BEST they can do is play to a draw… IF they have the C power (the Democrats that manipulate the bottom half).

    If you look at income distribution over past two decades, the A power (Tea Party) lost nothing, even as Big Business (1%) partners with and screwed with their pants on the C power (the bottom 60%).

  133. Gravatar of Morgan Warstler Morgan Warstler
    18. January 2012 at 22:17

    Rare earths, don’t worry – we just need the C power to get so desperate they do whatever the A power says.

    Environmentalism is a luxury the 1% enjoys, and the bottom 60% pays for.

  134. Gravatar of ssumner ssumner
    19. January 2012 at 13:50

    SK, You asked;

    “Krugman plainly thinks we are below the long-term equilibrium levels of Y and employment. In your view, does Cochrane think we are in continuous LT equilibrium or that we are incapable of using monetary and fiscal policy to move from a current ST equilibrium to the LT equilibrium? If the later, what markets are clearing too slowly to get us to LT equilibrium and why is policy, according to him, ineffective to get us there?”

    Obviously not, if you read his longer papers Cochrane clearly understands that too little money, or money plus gov bonds, can be a problem.

    math, You said;

    “Sumner’s point is smaller. He just wanted to belittle Krugman’s ego. He said about it explicitly.”

    Your words–I say I want him to be nicer.

    Everyone, I skipped al the debates about WL, as that’s being continued over at the “Points 1 to 5 post.”

  135. Gravatar of Why Am I Arguing with Scott Sumner « Uneasy Money Why Am I Arguing with Scott Sumner « Uneasy Money
    19. January 2012 at 20:56

    […] of advocating – horrors! – tax and spend policies as the way to stimulate the economy. In fact, Scott himself seems to think that what I am trying to do is defend what he calls the hydraulic Keyne…That’s a misunderstanding; I am simply trying to enforce some basic standards of good grammar […]

Leave a Reply