Nobel Prizes for alchemy?

Paul Krugman recently praised Simon Wren-Lewis’s attack on Bob Lucas and John Cochrane:

Imagine a Nobel Prize winner in physics, who in public debate makes elementary errors that would embarrass a good undergraduate. Now imagine other academic colleagues, from one of the best faculties in the world, making the same errors. It could not happen. However that is exactly what has happened in macro over the last few years.   Where is my evidence for such an outlandish claim? Well here is Nobel prize winner Robert Lucas:

“But, if we do build the bridge by taking tax money away from somebody else, and using that to pay the bridge builder — the guys who work on the bridge — then it’s just a wash.  It has no first-starter effect.  There’s no reason to expect any stimulation.  And, in some sense, there’s nothing to apply a multiplier to.  (Laughs.)  You apply a multiplier to the bridge builders, then you’ve got to apply the same multiplier with a minus sign to the people you taxed to build the bridge.”

I’m not sure if Wren-Lewis knows this, but Nobel Prizes are frequently awarded to people who don’t believe accept the Keynesian model (Lucas, Prescott, Friedman, Hayek, etc.)  So if this is the right analogy, then Wren-Lewis (and Krugman?) is suggesting that the academy in Sweden is continually awarding Nobel Prizes for the economic equivalent of alchemy.  In that case the Nobel Prize is a joke, and Wren-Lewis was wrong to use it as an indicator of academic prestige.

Lots of people believe the fiscal multiplier is roughly zero, including Lucas, Friedman and me.  That doesn’t mean one is ignorant of basic economics.

Wren-Lewis makes the same complaint about John Cochrane.  Here’s Cochrane:

Before we spend a trillion dollars or so, it’s important to understand how  it’s supposed to work.  Spending supported by taxes pretty obviously won’t work:  If the government taxes A by $1 and gives the money to B, B can spend $1 more. But A spends $1 less and we are not collectively any better off2.

“Stimulus” supposes that if the government borrows $1 from A and gives it to B we get a fundamentally different result, and we all are $1.50 better off.

But here’s the catch: to borrow today, the government must raise taxes tomorrow to repay that debt. If we borrow $1 from A, but tell him his taxes will be $1 higher (with interest) tomorrow, he reduces spending exactly as if we had taxed him today! If we tell both A and B that C (“the rich”) will pay the taxes, C will spend $1 less today.

.  .  .
2 Yes, I’m aware that old Keynesian models do give a multiplier to tax financed spending. Also, some new Keynesian models such as Christiano, Eichenbaum and Rebelo (2009) predict huge government spending multipliers whether financed by taxes or by borrowing. However, tax-financed spending is usually thought to have a weaker (if any) effect, which is why the current policy debate is only about borrowing to spend.

3 Advocates will go nuts here, and complain that A might be putting money in the bank, and “banks aren’t lending,” or stuffing the money in mattresses. As you can tell, this line of argument leads us into “something’s wrong” with the banking system, and confusion between fiscal and monetary policy.

So Cochrane is dividing debt-financed government spending into two components, a tax-financed spending increase and a deficit-financed tax cut.  I don’t see any problem with that thought experiment, but maybe I’m missing something.  Then he argues the debt-financed tax cut will do nothing, because of Ricardian equivalence.  I don’t entirely agree, but it’s certainly a respectable argument.  Then he suggests that the balanced budget multiplier is zero.  Obviously the Keynesian model says it’s not, but why assume that model is correct?  Do we have lots of empirical evidence of tax financed expansions of the state boosting RGDP?  Cochrane clearly indicates that he understands that the old Keynesian model implies a positive balanced budget multiplier, so it’s not a question of Chicago economists not having studied Keynesian economics.  He simply doesn’t agree.  (BTW, I learned the Keynesian model in Lucas’s macro class.)

Why did Wren-Lewis think this argument is silly?

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.

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

So now let’s consider two possibilities.  In the first, the fiscal stimulus fails, and the increase in G is offset by a fall of $100 in after-tax income and private spending.  In that case, consumption might fall by $10 and saving would have to fall by about $90.  That’s just accounting.  But since S=I, the fall in saving will reduce investment by $100 $90.  So the Wren-Lewis’s example would be wrong, the $100 in taxes would reduce private spending by exactly $100.

I’m pretty sure my Keynesian readers won’t like the previous example.  So let’s assume the bridge building is a success, and national income rises by $100.  In that case private after-tax income will be unchanged.  But in that case with have a “free lunch” where the private sector would not reduce consumption at all.

Either way Wren-Lewis’s example is wrong.  If viewed as accounting it’s wrong because he ignores saving and investment.  If viewed as a behavioral explanation it’s wrong because he assumes consumption will fall, but that’s only true if the fiscal stimulus failed.

Now that doesn’t mean the balanced budget multiplier is necessarily zero.  Here’s the criticism that Wren-Lewis should have made:

Cochrane ignores the fact that tax-financed bridge building will reduce private saving and hence boost interest rates.  This will increase the velocity of circulation, which will boost AD.

Those who read my blog know I don’t think much of this argument, because the Fed doesn’t tend to keep the monetary base fixed.  Rather they tend to adjust the base to offset changes in base velocity.  But at least it’s a respectable criticism of Lucas and Cochrane.  As it is, Wren-Lewis and Krugman are showing they don’t understand that not everyone agrees with the Keynesian model, and also that they don’t even know how to defend their own model.  It does no good to “refute” Cochrane with an example that implicitly accepts the crude Keynesian assumption that savings simply disappear down a rat-hole, and cause the economy to shrink.  Non-Keynesians aren’t likely to be impressed with that sort of argument, especially when presented with arrogant assertions that anyone who doesn’t buy into the Keynesian approach is some sort of ignoramus.

Again, the point is not that Wren-Lewis and Krugman are necessarily wrong about fiscal stimulus, but rather that the argument they present is incredibly weak.  I also find Cochrane’s critique of Keynesianism to be very weak (and equally arrogant), because he has no model of NGDP determination.  Indeed at times he seems to imply that if NGDP is the problem, then we are talking about monetary policy, not fiscal policy.  That’s my view as well, but if you plan to persuade Keynesians you need to do so with an explicit model of NGDP determination.  He needs to explain why fiscal stimulus won’t boost velocity, or why any boost would be offset by a lower money supply.  Cochrane doesn’t do that.

HT:  Marcus Nunes


Tags:

 
 
 

118 Responses to “Nobel Prizes for alchemy?”

  1. Gravatar of Kevin Donoghue Kevin Donoghue
    11. January 2012 at 06:46

    “…saving is defined as the resources put into investment projects.”

    Not in any textbook I’ve ever seen and certainly not in Keynes’s GT. S is defined as Y-C, basically. Obviously one can complicate the story by distinguishing between gross income and disposable income.

  2. Gravatar of Lars Christensen Lars Christensen
    11. January 2012 at 06:54

    Scott, I think the problem is basically the way “we” (the profession) is teaching micro and macro. In a recent post on my blug I suggested that we tell student first about Arrow-Debreau in micro. Then students understand that we introduce money and sticky prices and call that macro. See here: http://marketmonetarist.com/2011/12/23/how-i-would-like-teach-econ-101/

    A student that learned about macroeconomics in that fashion would ever argue that fiscal policy could move the AD-curve unless it impacts money velocity. Hence, the multiplier obviously is zero – and yes, no Nobel Prize to me…

  3. Gravatar of ssumner ssumner
    11. January 2012 at 07:00

    Kevin, Yes, saving is Y-C. And investment is Y-C.

    Or with government added then private saving is Y – C – T and government saving is T-G. So total saving is Y – C – G.

    And investment is Y – C – G.

    Lars, I think it’s quite possible that fiscal stimulus can impact V, although I don’t think that makes fiscal stimulus effective.

    (Is there a typo in your last paragraph?)

  4. Gravatar of Team Old Keynesian versus Team Chicago « Portrait of the Economist as a Young Man Team Old Keynesian versus Team Chicago « Portrait of the Economist as a Young Man
    11. January 2012 at 07:20

    […] http://www.themoneyillusion.com/?p=12596 […]

  5. Gravatar of Vivian Darkbloom Vivian Darkbloom
    11. January 2012 at 07:36

    It amazes me how frequently people (even economists) get into the argument of whether Keynes said “this” or Keynes said “that.” The *mere* issue of what Keynes did or did not say (or write) would be of academic importance only to a biographer or historian. To my knowledge, very little of what he actually did say or write did he himself ever put to any rigorous empirical test or controlled experiments. Keynes came with theories, little more. And, how often do you hear similar arguments as to what other, less famous, economists did or did not say?

    The whole point of quoting Keynes, or arguing what Keynes did or did not say, or even in arguing that so-and-so does not understand what Keynes meant, is an intellectual cop out. Logicians call this fallacy “the appeal to authority”. The subtext to these arguments is that because Keynes said something, it must be true. This is a particularly useful rhetorical device when selling policy to the general public, all in the name of Mr. Keynes.

    Economists need to de-emphasize debates over what Keynes said and focus on the imminently more important question: Was Keynes right? If you want to make an assertion as to what Keynes said, fine, but please move on quickly to the issue of whether the idea behind that attribution is actually correct.

  6. Gravatar of Lars Christensen Lars Christensen
    11. January 2012 at 07:47

    Scott, there was a typo – it “never” rather than “ever”…but I guess you where thinking about the NO Nobel Prize for me;-)

  7. Gravatar of Ram Ram
    11. January 2012 at 07:48

    Economic models, in my view, do not constitute successful scientific theories, because they make many predictions that fail to meet ordinary scientific standards of accuracy, precision, reliability, etc. We do not have a scientific theory of recessions, for example. Instead, we have a number of models (which may be more or less compelling) that enable us to study recessions in greatly simplified settings. Until one or more of these models passes ordinary scientific muster, there will continue to be disagreement about whether a given model captures enough, captures too much, represents the system under study appropriately, etc. Different economists, based upon different interpretations of the evidence, not to mention different priors, will invariably come down on different sides of these debates.

    Economists would do well to acknowledge, when discussing a topic as fraught with controversy as the business cycle, that none of their models qualify as successful scientific theories, that other economists find alternative models with different implications more compelling, and that their policy prescriptions are better thought of as policy experiments. A convincing model makes some policy experiments appear more promising than others, but nothing short of a successful scientific theory ought to give economists the confidence to neglect or ridicule the alternative perspectives of some of their peers.

    Personally, I find New Keynesian DSGE models, a la Woodford, to be the most compelling models of business cycles, including the present one, which is part of what draws me to NGDP path targeting. On the other hand, I realize it is just one class of models out of many. Some alternative models make such a policy appear less desirable, which is why I support experimenting with, but not betting the farm on, NGDP path targeting. Why is it so hard for prominent economists to discuss their policy prescriptions in the way I just did, without conveying not only excessive confidence in their preferred models, but also no respect for alternative models?

  8. Gravatar of Kailer Kailer
    11. January 2012 at 07:49

    Quoth Krugman:
    “[I]f you ask a liberal or a saltwater economist, “What would somebody on the other side of this divide say here? What would their version of it be?” A liberal can do that. A liberal can talk coherently about what the conservative view is because people like me actually do listen. We don’t think it’s right, but we pay enough attention to see what the other person is trying to get at. The reverse is not true. You try to get someone who is fiercely anti-Keynesian to even explain what a Keynesian economic argument is, they can’t do it. They can’t get it remotely right.”

  9. Gravatar of Foster Boondoggle Foster Boondoggle
    11. January 2012 at 07:53

    The point of these arguments – debates over what is considered “Keynesian” aside – is whether or not it’s possible to put idle resources to work and thereby increase net economic productivity and growth. Talking about money obscures the underlying economic process, which I thought was part of the point of the title of this blog. In the current environment the output gap can be reduced either by fiscal or monetary means – a first order effect. Debates about current saving in the face of future taxes are about second order effects.

    Perhaps the output gap cannot be reduced, because it doesn’t exist, because we’re measuring it wrong: if the economy’s measured output was in effect overstated by including things we no longer need or want (military hardware, houses in the desert), then we might be facing the ZMP worker situation they talk about over at Marginal Revolution. But that seems to be a different issue than the one being debated here.

  10. Gravatar of foosion foosion
    11. January 2012 at 07:53

    >>the Fed doesn’t tend to keep the monetary base fixed. Rather they tend to adjust the base to offset changes in base velocity>>

    How do the Fed’s recent actions fit into this?

  11. Gravatar of dwb dwb
    11. January 2012 at 08:00

    the “other multiplier” by the way, is the money multiplier … which is often argued to be essentially zero at the ZLB. One see’s Krugman argue both for more QE and inflation, and that monetary policy is ineffective at the ZLB. I don’t understand how that is supposed to work.

    Krugman here: “The liquidity trap “” which is in effect a special case of IS-LM analysis “” has some special properties. Notably, even large government borrowing won’t drive up interest rates (not unless the borrowing is enough to restore full employment), and you can print as much money as you like without causing inflation.”

    http://krugman.blogs.nytimes.com/2011/12/14/interest-rates-inflation-and-the-way-the-world-works-slightly-wonkish/?gwh=682BBD796526616099671E9D19773CDD

    Krugman here: “As I see it “” and as I suspect many people at the Fed see it “” the basic point is that to gain traction in a liquidity trap you must either engage in huge quantitative easing, raise the expected rate of inflation, or both. ”

    http://krugman.blogs.nytimes.com/2011/10/30/a-volcker-moment-indeed-slightly-wonkish/?gwh=0CBC565A751518DD8D9BA71220C6C337

    The only logical way both can be true in my opinion is (and Krugman is sometimes careful to add as a qualification) is that monetary policy is ineffective because it “wont” be delivered, not because it “cant” be delivered. So it logically seems to me the “liquidity trap” is a statement about policy failure, not some underlying technical issue with the ability of the Fed to create inflation (the evidence is against the abiility of the Fed to stimulate the economy right now anyway – markets react to any hint of further stmulus).

    Now, this begs the question, If i don’t trust the Fed to deliver optimal monetary policy, why should I expect the leviathan federal govt to deliver optimal fiscal policy? All I can say from my vantage point is that once the federal government starts spending, it is nearly impossible to stop. I could give scores of examples…”Temporary” programs and subsidies become permanent. local communities and businesses are very sensitive to the “loss” of jobs from spending programs. there is no such thing as temporary stimulus.

  12. Gravatar of Lars Christensen Lars Christensen
    11. January 2012 at 08:01

    …and concerning velocity. We have the same model in the head I think – yes, fiscal policy can influence velocity, but with a central bank that target NGDP or inflation then there is obviously no impact of monetary policy.

  13. Gravatar of Nick Rowe Nick Rowe
    11. January 2012 at 08:03

    Scott: remember that *desired* saving includes desired hoarding.

  14. Gravatar of Lars Christensen Lars Christensen
    11. January 2012 at 08:08

    Ram, “Why is it so hard for prominent economists to discuss their policy prescriptions in the way I just did, without conveying not only excessive confidence in their preferred models, but also no respect for alternative models?”

    Excellent point. You could argue that different policy rules should be tested in different types of models – afterall we really have no clue which model is the “true” model of the world. Bennett McCallum has in fact done exactly that he. Has test different policy rules in different models. His conclusion is that NGDP GROWTH targeting works the best across models. I am not sure he have done the same for NGDP LEVEL targeting…

  15. Gravatar of William William
    11. January 2012 at 08:24

    I am thinking the same thing as Kevin Donoghue, but I am much more confused about it. Isn’t the paradox of thrift about how increased saving decreases total spending, and is thus self-defeating? If saving were just investment spending then they never would have come up with the paradox of thrift, which I know you aren’t a fan of but it is quite famous. Where does it come from?

  16. Gravatar of orionorbit orionorbit
    11. January 2012 at 08:35

    Scott, first of all, in the bridge building example, it is not clear that M. Friedman would disagree with the Keynesian view given that he wasn’t opposed to government spending under all circumstances, in fact he did say that the FDR/Keynes fiscal strategy was at the time a good thing and that it wasn’t Keynes who was wrong but his over-zealus disciples. Also Bob Lucas’ opinion on fiscal stimulus has not been consistent, I remember Noah Smith quoting him both in support and in opposition of the Obama stimulus, so count him out as well. But above all, since some of my economics teachers were… ahem… members of the Nobel Committe, I have to protest on their behalf. The Nobel prize is NOT an endorsement of an economic theory or a statement on what is science and what is alchemy or ANYTHING remotely related to assigning a status of scientific authority to somebody. The econ Nobel committee does NOT follow the same standards as say the physics one (Which requires empirical valridation of theory), the econ Nobel is best seen as a recognition of somebody’s contribution that might be right or might be wrong. If you get a Nobel prize in physics it means that the committee thinks your theory is right. If you get a Nobel in econ, it does NOT mean that they think your theory is right, it means they think your theory is a useful contribution. In fact I would say that the biggest turn-off for Swedes is to see people that should know better not really getting what the Nobel prize is about, for instance, citing who’s got how many Nobel prizes in support of some position is exactly what the committee would like to avoid.

    But since you cite alchemy and science, I would say that the correct test of whether Cochrane or Krugman is right, is whether a model that assumes S=I in all circumstances predicts the observed data better than a model that assumes that S does not equal I in the liquidity trap. Krugman is not right because S is not equal to Y, he is right because by assuming this, he predicted that interest rates would stay low and the Cochrane/Fama crowd are wrong not because S=I is wrong but because by assuming that S=I they predicted rising rates because of crowding out. Models are not judged on the plausibility of their assumptions but on the accuracy of their predictions. The PHYSICS Nobel committee would agree with this principle and by the way, Milton Friedman would too.

  17. Gravatar of Kevin Donoghue Kevin Donoghue
    11. January 2012 at 08:36

    William,

    The way to avoid confusion is to distinguish clearly between accounting identities and equilibrium conditions. It’s fine to say that in a closed economy Y=C+I+G is an accounting identity, but you must remember that I includes unwanted accumulation of inventories. If we’re talking about demand and supply schedules however then aggregate supply (Y) is equal to agggegate demand, C(Y,r,P)+I(Y,r,P)+G only in equilibrium. Away from equilibrium when S(.)>C(.)+I(.) the paradox of thrift kicks in. If P is insufficiently flexible, Y and r must adjust to equilibrate S and I.

  18. Gravatar of Adam Adam
    11. January 2012 at 08:42

    This is one of the examples that bothers me the most about economics discussions: “But here’s the catch: to borrow today, the government must raise taxes tomorrow to repay that debt.”

    This is not a factual statement. It’s true under the right assumptions, but those assumptions don’t reflect reality. Government debt does not shrink due to sustained periods of surplus (i.e. higher future tax receipts). It shrinks relative to GDP via GDP growth, both nominal and real.

  19. Gravatar of Kevin Donoghue Kevin Donoghue
    11. January 2012 at 08:50

    Vivian Darkbloom: “Economists need to de-emphasize debates over what Keynes said and focus on the imminently more important question: Was Keynes right?”

    All economists that I know of agree that Keynes was right sometimes and wrong at other times. The arguments are about which times were which. Since Scott likes to take issue with Keynesian economics it’s rather difficult to entirely exclude JMK from the discussion.

  20. Gravatar of Vivian Darkbloom Vivian Darkbloom
    11. January 2012 at 08:54

    @ Kailer

    Is that the same Krugman who wrote this?:

    “Some have asked if there aren’t conservative sites I read regularly. Well, no. I will read anything I’ve been informed about that’s either interesting or revealing; but I don’t know of any economics or politics sites on that side that regularly provide analysis or information I need to take seriously. I know we’re supposed to pretend that both sides always have a point; but the truth is that most of the time they don’t. The parties are not equally irresponsible; Rachel Maddow isn’t Glenn Beck; and a conservative blog, almost by definition, is a blog written by someone who chooses not to notice that asymmetry. And life is short …”

    http://krugman.blogs.nytimes.com/2011/03/08/other-stuff-i-read/

  21. Gravatar of David K David K
    11. January 2012 at 08:58

    Cochrane says,
    “But here’s the catch: to borrow today, the government must raise taxes tomorrow to repay that debt. If we borrow $1 from A, but tell him his taxes will be $1 higher (with interest) tomorrow, he… will spend $1 less today.”

    This is where the mistake is, according to Wren-Lewis. He will not consume $1 less today, because he presumably smooths his consumption. A consumption-smoothing consumer under Ricardian equivalence will cut his consumption by $r per year in perpetuity, assuming r is the real interest rate. He would do the same if he were taxed $1 this year.

    If we leave the model here, we have stimulus, which is why the zero bound is so important to Keynesians. In the Keynesian model, the zero bound is serving as a binding price floor on capital, and thus the borrowing has no effect on the interest rate of savings and thus shouldn’t affect private savings at all.

  22. Gravatar of Vivian Darkbloom Vivian Darkbloom
    11. January 2012 at 09:04

    Kevin Donohue,

    I think my earlier comment clearly distinguished between necessary references to what Keynes may (or may not have) said for the purposes of putting those statements to the truth test, and *mere* (emphasis also in original) references that constitue an appeal to authority. I don’t think the distinction I drew was overly nuanced, but you seem to have missed it.

    As far as the “economists you know”, you are likely keeping very good company. There is an economist that I do not know, but whose blog is likely the most read of all and which, even I, read, who tends to make this appeal to authority on a weekly, if not daily, basis. He even has a Nobel Prize.

  23. Gravatar of Kevin Donoghue Kevin Donoghue
    11. January 2012 at 09:14

    Vivian,

    “Mr A was right when he said X” is not an appeal to authority, I’m sure you’ll agree. Nor of course is it a compelling argument, on its own. I’ll leave it at that.

    William,

    For “S(.)>C(.)+I(.)” in my earlier reply, please read S(Y,r)>G+I(Y,r).

  24. Gravatar of Mike Sax Mike Sax
    11. January 2012 at 09:26

    If you learned about the Keynesian mulitplier in Lucas’ class mabye that’s why you often say you don’t get Keynesian arguments.

    It must have been hard to learn with all that giggling going on (That’s what Mankiw remembers in Lucas’ classes).
    “the part of income not “spent” is saved, which means it’s spent on investment projects”

    There is a major dissonance of your theory and the Keynesian and I’m not at all clear that you are on the better side of it.

    So if I stuff money in my bank account it is in an investment project? Right now there are plenty of investment projects going on.

  25. Gravatar of bill woolsey bill woolsey
    11. January 2012 at 09:29

    The Wren argument is narrower. It assumes Ricardian equivalence (for the sake of argument.) Government spending increases (rather than taxes decrease.) The government spending is funded by borrowing. Private saving does increase this year, but by less than the increase in government spending and the increase in the deficit. People save over several years to accumulate enough wealth to pay the interest and principle on the added government debt.

    Since saving rises this year by less than the deficit this year, then national saving decreaces this year. In other words, consumption falls by less than government spending rises.

    Now, it could be that the decrease in national saving supply raises the interst rate (it does raise the wicksellian natural interest rate.) If the market interest rate rises, then this would tend to reduce consumption spending (increases the quantity of private saving supplied) as well as decrease investment spending.

    But if he natural interest rate is negative or there is some other version of the liquidity trap, or the central bank is pegging the market interest rate, or even if it is controlling the quantity of money, then the market interest rate doesn’t rise (or not as much as the natural interest rate) and so it doesn’t tend to reduce consumption or investment spending (enough for a full “crowding out.)

    Now, back to Ricardian equivalence. Suppose it wasn’t an increase in government spending, but rather a decrease in taxes funded by an larger budget deficit. The increase in disposable income will be saved (plausibly anyway) to pay the interest and principle on the national debt. National saving is not impacted, and consumption remains the same.

    The reason is consumption smoothing. Why consumer alot this year and reduce it in the future when taxes must be paid? The tax cuts provides a bonanza just in time to avoid the future reductions in consumption that will be needed to pay the taxes.

    If this doesn’t work, and for one reason or another, people don’t save more to pay the interest and principle of the national debt, then national saving falls. Assume no one thinks about future taxes and so private saving doesn’t change. National saving falls due to the higher budget deficit, annd consumption rises.

    Again, this increases the natural interest rate, and if the market rate rises too, then this will tend to raise the quantity of private savings supplied and reduce investment spending. But if the market rate doesn’t rise because of a liquidity trap or interest rate targeting, then that doesn’t happen.

    In summary, Ricardian equivalence means that debt financed tax cuts don’t impact consumption, saving, or the natural interest rate. They don’t raise aggregate demand, and so won’t tend to raise either real output and employment or prices and wages.

    On the other hand, debt financed government spending reduces national saving, and raises the natural interest rate. If there are liquidity trap issues or interest rate targeting or even quantity of money targeting, this would tend to raise aggregate demand and either real output and employment or prices and wages or some of both.

    By the way, the insulting tone of this follows at least partly from an assumption that _all_ free market economists really believe Ricardian Equivalance rather than the actual continuum from considering it implausible, to interesting possibility, to an important partial effect, to being the only sound way to do economics.

  26. Gravatar of Kevin Donoghue Kevin Donoghue
    11. January 2012 at 09:35

    Scott: BTW, I learned the Keynesian model in Lucas’s macro class.

    Make Sax may be onto something. Somewhere along the line I got the impression you were taught macro by David Laidler, which made me wonder why you seemed so perplexed by Keynesian models. Laidler was never a Keynesian, but he knew his stuff.

  27. Gravatar of Vivian Darkbloom Vivian Darkbloom
    11. January 2012 at 09:48

    Kevin Donohue,

    “I’m sure you’ll agree”? Why would you be so sure? Is this the bandwagon argument? If one *merely* asserts “JM Keynes was right when he said X” without further argumentation, that would be an appeal to authority. Some argue that is not, strictly speaking a *fallacy* if Mr. A is an “expert”, but others differ, and I agree with those others:

    “Another common fallacy is the appeal to authority, which consists of arguing a point by invoking the opinion of an expert. However, experts may be wrong, they may be expressing an opinion outside their area of expertise or they may have been incapacitated or joking when making the point. It is the expert’s reasons that are valuable, not the fact that they were announced by an expert.”

    (Daniel Sokol, “The Right Way to Argue.” BBC Magazine, December 20, 2006)

    JM Keynes was reputed to have said on his deathbed: “I should have drunk more champagne”. Not only may Keynes have simply been wrong, but that suggests to me that 1) he was no stranger to champagne; and/or 2) he may have been joking.

  28. Gravatar of UnlearningEcon UnlearningEcon
    11. January 2012 at 10:21

    Scott,

    You mistakenly assume that, because S=I, S creates I. It is actually the opposite which is true, because when banks lend they create savings in the process. Keynes knew this.

  29. Gravatar of ssumner ssumner
    11. January 2012 at 10:39

    Vivian, I agree.

    Ram. I strongly agree, except I’d quibble with your use of “science.” It’s probably more useful to think of science in terms of a style of inquiry rather that accuracy of models. Many areas of the physical sciences are also fraught with uncertainty.

    Kailer, That’s almost comical. I learned almost nothing about conservative economics at the very liberal University of Wisconsin, and lots about liberal economics at the University of Chicago. For God’s sake Krugman brags about the fact that he doesn’t read conservative blogs. Let’s get serious.

    Foster, There are idle resources and M-policy is by far the most effective way of putting them to work.

    foosion, In 2009 I’d say not well at all. But the assumption of no change in the base was 10 times further off course. So I’d stick by my generalization.

    Nick, I agree, but I don’t think that affects any of my arguments here. In terms of abstract models, fiat money is sort of like paper gold, it’s a capital good that generates a flow of liquidity services.

    William, No, that’s not the paradox of thrift. That paradox argues that attempts to increase saving will reduce income. And fail to actually increase saving.

    Orionorbit. It’s true that Friedman thought fiscal stimulus could work early in his career, but by 1996 he’d concluded that it was ineffective.

    I agree that the Nobel tells us nothing about whether theories are right, but it certain suggests that anti-Keynesian models are respected by the profession. That was my point.

    S=I is a tautology. If Krugman being right depends on it being wrong, then he is in deep trouble.

    Adam, I don’t agree, there is no free lunch with government debt.

    David K, The problem with your argument is that unlike Mr. Wren-Lewis, Cochrane actually understands that “spending” does not equate with “consumption.”

    Mike Sax, S=I is not a theory, it’s an identity. As far as I know it’s assumed true in every single macro textbook.

    Bill, I do understand all that. My point is that if the government spends an extra $100 on a bridge, there is nothing in “consumption smoothing” that would preclude C+I falling by exactly $100. It may not, as V may increase. But talking about consumption smoothing doesn’t advance the argument at all.

    Kevin, Yes, I’ve read lots of stuff by Laidler. My problem with Keynesianism isn’t because I haven’t studied lots of it. It’s either that I’m really dense or that the theory is illogical. I’m not sure which.

  30. Gravatar of ssumner ssumner
    11. January 2012 at 10:41

    UnlearningEcon. No I don’t assume that. I’ve spent three years arguing against that sort of thinking. The level of investment is determined by the interaction of S and I.

    Unless of course you mean an increase in the quantity of savings, in which case it’s true by definition that investment increases.

  31. Gravatar of Mike Sax Mike Sax
    11. January 2012 at 10:46

    S=I is an idenity in the way in accounting that debits=credits. This doesn’t mean that debits and credits are the same thing or that savings and investments are.

    Savings is one thing, investment is another. If I own a company and invest in new equipment or hire workers that is an investment. If I leave it in retarined earnings that is saving not investment. An accounting idenitity is not to say they are identical in a nonaccounting way.

  32. Gravatar of Mike Sax Mike Sax
    11. January 2012 at 10:47

    In accounting debits=credits + owner’s equity. But debits, credits, and owner’s equity are distinct things.

  33. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    11. January 2012 at 10:48

    ‘Imagine a Nobel Prize winner…who in public debate makes elementary errors that would embarrass a good undergraduate.’

    Like (just to take one example of many that popped into my mind) comparing the performance of public school students in Wisconsin to those in Texas without bothering to control for the rather large demographic differences between the two states?

  34. Gravatar of StatsGuy StatsGuy
    11. January 2012 at 10:57

    I am so tired of the Keynesian/Monetarist p!ssing match, particularly since both approaches seem to ignore issues that the vast majority of people seem to think matter:

    1) Debt. Does or does not debt matter? Both say no. The rest of the world is beginning to think K&M are obsessed with squabbling over details, and are irrelevant to reality – in which debt does matter.

    2) Distribution of Income. Does or does not income distribution matter? Scott would say income/wealth are irrelevant, since what matters is distribution of consumption. Yes, well, that might be true if present consumption was a _guarantee_ of future consumption, but what if you think the future is a nastier place than the present? You might want a bit of private wealth to preserve future income in case the state decides those transfers are not sustainable…

  35. Gravatar of Andrew’ Andrew'
    11. January 2012 at 10:59

    I admit to a limited data set to work from but when Krugman said Keynesians could discuss both sides (as if there are only two sides) in my mind I compared Sumner and Cowen to DeLong and the guy who dismissed Austrian theory as the hangover theory as worthy of thinking about as phlogiston theory of fire (he’s obviously thought more about the phlogiston theory of fire than about Austrianism). Those guys don’t even realize when they are taking shortcuts in their own wheelhouse.

  36. Gravatar of Merijn Knibbe Merijn Knibbe
    11. January 2012 at 11:03

    We have ‘classical’saving: people saving money and either lending it to a bank or to somebody else. This can lead to an increase of the ‘inventory’ of money. An increase of the ‘inventory’ of money at the banks or wherever is, in National Accounting (the basis of Keynesian theory) not accepted as an investment. The money can also be spend on ‘stuff’, like brdiges, that does count as a real investment. But we all should say goodbye to the idea that the intesrest rate is an price which equilibrates ‘money’ savings with investments. I call this money savings.

    And we have ‘national accounts’ savings. These do include an unwanted increase in inventories (including in fact unsold new houses). They do however not include an increase in ‘money inventories’ which are for instance put in the European Central Bank by the other banks. This means, that from the National Accounts and the Keynesian perspective money savings do not count as investments and the only real savings are indeed equal to investment. Investments are the real buildup of wealth by a society, resources not used for present but for future consumption.

    This, in it turn, means that if there is a buildup of money-inventories without a compensating creation of new money total spending goes down. In this situation, taxing these inventories directly or indirectly leads to an increase of total demand and therewith production.

    Guys, please check the National Accounts more thoroughly, those are the real macro model, fully estimated, based upon micro data, with all due accounting restrictions, including flow of funds data, including balance sheets, including ‘non produced assets’, showing all sectors and the connections between them instead of only a silly ‘representative consumer’which can’t be unemployed as heshe is the only kid in town and who owns/rents the only ‘representative house’- it’s so much more sohpisticated than all this bickering.And they indeed show that an increase in government spending can add as much to total demand as an increase in consumer spending or investments or exports.

  37. Gravatar of Kevin Donoghue Kevin Donoghue
    11. January 2012 at 11:15

    Scott, you say in one comment: “S=I is a tautology.” In the next you say: “The level of investment is determined by the interaction of S and I.” Now, I think I understand what you are trying to say. But really I pity your students.

  38. Gravatar of Mike Sax Mike Sax
    11. January 2012 at 11:22

    Kevin do you agree with my gloss on S=I? “S=I is an idenity in the way in accounting that debits=credits. This doesn’t mean that debits and credits are the same thing or that savings and investments are.”

    “Savings is one thing, investment is another. If I own a company and invest in new equipment or hire workers that is an investment. If I leave it in retarined earnings that is saving not investment. An accounting idenitity is not to say they are identical in a nonaccounting way.”

    “In accounting debits=credits + owner’s equity. But debits, credits, and owner’s equity are distinct things.”

    To me it seems like this is an apples to orange argument. S does equal I like debits equal credits in accounting but this does not imply they are the same thing in either case.

  39. Gravatar of Matt Rognlie Matt Rognlie
    11. January 2012 at 11:27

    Accounting identities function through the price system. If suddenly there’s an increase in the supply of savings, then we ordinarily expect investment to increase as well—but there’s more to the process than just S=I. The real interest rate falls to equilibriate the market, and a lower interest rate means more investment.

    In principle, you can analyze the savings and investment markets in a completely decentralized fashion. On one side, there are “savers” with a particular supply curve for savings in r-S space (interest rate/savings), and on the other side there are “investors” with a demand curve in r-I space (interest rate/investment). The two parties only interact through the equilibriating real interest rate. Now, in the real world this isn’t exactly true—there are intermediaries that service both parties, not an idealized Walrasian auctioneer. But I don’t think it’s so far from true either. The banking sector is pretty competitive, and so long as it’s reasonably healthy most banks have similar costs of obtaining financing. There is indeed an “r” identifiable in the data that roughly corresponds to the idealized price in an Econ 101, supply-demand diagram.

    But now, what if the standard mechanism for reaching equilibrium isn’t functioning? In particular, what if the nominal interest rate is pegged at a certain level, and inflation expectations are anchored enough that real interest rates are pretty much pegged as well? Then there is no reason to expect a movement in the savings supply curve to result in a higher equilibrium level of investment. How could it? How are the incentives for investment being changed in any way? Borrowers seeking to finance a project face the same cost–an interest rate that won’t move. This is the essence of the Keynesian analysis, and I don’t think Cochrane or Lucas have really acknowledged it.

    Now, there are a lot of caveats here. In particular, the assumption of a “fixed interest rate” may be wrong. First of all, firms’ cost of financing is a step removed from the federal funds rate, and includes various premia that can plausibly be affected by the Fed, even assuming a constant federal funds rate. For instance, if the Fed starts buying up a lot of BBB corporate debt, you can be sure that the rate will fall.

    Second, anyone seeking finance for an investment cares about interest rates far into the future, not just rates today, and it is possible (depending on the details of the Fed’s reaction function) that those rates will depend on fiscal policy today. You can imagine an extremely conservative Fed altering its policy commitments for the rate 2-5 years from now in order to offset any increase in inflation (and thus any increase in AD) that arises from fiscal policy. In that case, the multiplier is indeed much closer to zero.

    And in general, if the Fed isn’t truly “powerless” at the zero lower bound, you have to ask why it isn’t doing more now. If the answer is “it thinks that AD is currently at the right level”, then presumably it will act to offset any fiscal stimulus. This, I take it, is your fundamental argument, and I think it’s a good one. But if, on the other hand, the answer is “the Fed is influenced by erratic public opinion and regional presidents with incoherent models” or maybe “the Fed feels that it faces a fundamentally intractable time inconsistency problem that leaves it unable to shape expectations”, the Keynesian analysis does apply.

  40. Gravatar of foosion foosion
    11. January 2012 at 11:36

    >>And in general, if the Fed isn’t truly “powerless” at the zero lower bound, you have to ask why it isn’t doing more now.>>

    A popular answer is that the Fed cares more about inflation than unemployment. “the Fed is influenced by erratic public opinion and regional presidents with incoherent models” is not necessarily inconsistent with that answer.

  41. Gravatar of UnlearningEcon UnlearningEcon
    11. January 2012 at 11:44

    If two things are true by definition it makes no sense to say that something has to adjust to equilibrate them.

    In a Keynesian framework (not New Keynesian) the rate of interest is not determined by S or I but by liquidity preference. The rate of interest then determines the rate of investment, which creates its own savings.

    I don’t think it’s worth discussing whether this is true or not here, Scott, I’m just letting you know where I’m coming from.

  42. Gravatar of Matt Rognlie Matt Rognlie
    11. January 2012 at 11:45

    By the way, though I understand your reasoning, I think you should be more cautious about saying that the multiplier is literally “zero”, as opposed to “much lower than Keynesian economists think”. A formal treatment of these issues will almost never find that the answer is exactly zero, although this depends on the exact nature of the game that the fiscal and monetary authorities are playing, along with their policy instruments, degrees of commitment, etc.

    (A) Suppose that the fiscal authority credibly and irrevocably decides to increase the path of government spending over the next few years. Suppose also that the monetary authority has complete power to influence demand. The monetary authority optimizes a social welfare function, conditional on the path for government spending laid down by the fiscal authority. Is the multiplier “zero”? No, because unless government spending is a perfect substitute for private spending, the optimal level of private spending does not decrease one-for-one with an increase in government spending. The monetary authority tries to implement this optimal level. (Note that this is consistent with strict inflation targeting. People are willing to work more now that the higher trajectory of government spending has made them poorer, and along the optimal trajectory this exactly offsets the inflationary impact of higher spending, resulting in zero net change in inflation.)

    (B) Suppose that the fiscal and monetary authorities cooperate to achieve the best possible outcome, and are able to credibly commit to any policy trajectory. Their only constraint is that the monetary authority faces a binding zero lower bound for some period. Is there fiscal stimulus in the optimal joint fiscal/monetary plan? It turns out that the answer is yes: although the central bank can also increase AD when the zero lower bound is binding by committing to low rates after the trap ends, this is costly. The optimal policy consists of both stimulus through monetary commitment and increased spending during the trap. Ivan Werning does all this formally in a recent paper: http://dl.dropbox.com/u/125966/zero_bound_2011.pdf

  43. Gravatar of Matt Rognlie Matt Rognlie
    11. January 2012 at 11:53

    “In a Keynesian framework (not New Keynesian) the rate of interest is not determined by S or I but by liquidity preference. The rate of interest then determines the rate of investment, which creates its own savings.”

    This is true in a New Keynesian framework too. It’s implicit in the Fed’s ability to set the policy rate through open market operations: the demand for base money is determined by the spread between base money and a one-period bond (the nominal interest rate), and the Fed changes the supply to move along that demand curve. Yes, sometimes this is dispensed with because (1) a paper doesn’t bother with the routine issue of how the Fed sets the policy rate or (2) the model is explicitly “cashless” (not just a cashless limit) and posits that the Fed sets the policy rate by paying interest on money. But generally it’s in the background.

    This liquidity-preference view of the interest rate is compatible with a worldview where the interest rate is also the equilibriating force in the savings-investment market, because given many sensible objective functions the Fed will adjust policy so that these two values coincide. (i.e. it’s trying to hit the Wicksellian “natural rate”.) This only breaks down when the Fed is either unwilling or unable to adjust policy in response to events so that the natural rate corresponds with the actual one that is realized in money markets.

  44. Gravatar of OojOoj OojOoj
    11. January 2012 at 11:57

    With the added mistake of confusing utility with wealth, and ignoring the fact of diminishing marginal utility, taught somewhere between minutes 5 and 25 of the first day of Economics 101.

  45. Gravatar of Kevin Donoghue Kevin Donoghue
    11. January 2012 at 12:00

    @Mike Sax: “Kevin do you agree with my gloss on S=I?”

    I broadly agree but I won’t endorse every detail of your comment. For example, in my language it’s not right to say that hiring workers is investment. I presume what you have in mind is that the workers produce goods which then become part of the firm’s stock of capital, in the form of fixed assets, software, inventory or whatever. In those cases I think of the firm as buying the product from itself and that’s the sense in which I (the market value of the product) is a component of aggregate demand. If that’s what you mean when you refer to hiring workers then we’re in agreement.

    To repeat what I said to William: the way to avoid confusion is to distinguish clearly between accounting identities and equilibrium conditions. It’s fine to say Y=C+I (let’s forget G for now) is a tautology in a case where it’s clearly understood that I may include additions to inventory which entrepreneurs did not plan and are not happy with. If I represents an investment function (a schedule), as in I=I(Y,r) and C is the consumption function then the equation Y=C+I is not a tautology at all. Ordinarily it will hold for just one value of I, the equilibrium value. (Of course a model might have multiple equilibria but most textbook models don’t.)

    Since S=Y-C it follows that S=I may not be a tautology either. It depends on the context. That is where Scott is going wrong.

  46. Gravatar of James James
    11. January 2012 at 12:30

    I am only an economics student, so perhaps I simply haven’t been made privy to all of the dark secrets that are only revealed with full induction into the lodge. but the largest fault I find in nearly all of the arguments, on all sides of the issue is that everyone seems to be making the cognitive error of assuming a static model. Even in text books examples are given as “a change in G will result in …. all else being held equal”. I understand the need to abstract from reality in order to construct a functional model, but ignoring the effects of time when discussing macroeconomics, from any philosophical point of view, is roughly equivalent to ignoring the effects of gravity when discussing aerodynamics. I suppose you could do so, but personally I don’t believe I’d care to have you at the controls the next time I fly.

    Depending on the situation, a properly targeted stimulus package could conceivably erase itself by spurring sufficient growth, although in my opinion it would be borderline miraculous for it to do so, on the other hand allowing for something as simple as the time value of money, makes it quite clear that $100 in spending today will very rarely require $100 in taxation tomorrow, it might be $90, it might be $110, or it might be $10, if the Keynesians have a really good day, but it almost certainly will not be $100.

    Abstraction is necessary, but static models, while conceptually easier, are also nearly useless for modelling real world phenomenon. Its the difference between looking at a still life snapshot of an athlete competing, or watching the same performance on high-def video.

  47. Gravatar of Morgan Warstler Morgan Warstler
    11. January 2012 at 12:37

    DeKrugman is not an economist. What he practices is not the study of businessmen running the world. What he practices is not the study of unfettered free markets and infinite consumer wants faced with physical atomic scarcity and infinite digital replication.

    His goal is not to understand the above, his goal is to get free atomic shit for the weak and less accomplished. This is a fine thing to want, but the process of arguing for it, literally means you are no longer doing economics.

    The larger point is that…

    Government and money are weaker servants, literally and truly servants, to the businessmen that run the world. If governments or money fail to serve the interests of the people who matter (businessmen who run the world), the money will fall, the government will fall, and the businessmen that run the world, will keep running the world.

    It boggles my mind that we even have to do all of the above.

    Just say out loud that DeKrugman has MOTIVES that are premised on trying to make sure that businessmen don’t get to be the boss of everything and everybody – which is a HOPELESS pursuit.

    He is 100% predictable – look at his nod to MMT, the only reason he like it, is because it PRESUMES that businessmen do not run the world.

    First principles, prime directive, call it whatever you like: if you believe in freedom, in personal choice, it is impossible to view government directed anything (by force of gun) as being as valuable as business.

    DeKrugman wants the wrong kind of people to matter more.

    It isn’t complicated. The man who invented the paperclip is more important than anyone in government ever. Power derived from rousting up the bottom half to get some free shit from the top half is a VIABLE strategy for the ones doing the rousting, but it does nothing for the bottom half…. it just makes them wait longer for someone to invent another paperclip.

    The only reason DeKrugman isn’t being chased by a mob is because people don’t yet know what cool shit they could of had 5 years ago, but still won’t have for 4 more.

    The Internet could have started in the 1970’s and it is DeKrugman’s fault it came so late.

  48. Gravatar of Adam Adam
    11. January 2012 at 12:42

    Scott – It’s not a question of free lunch. It’s a question of what actually happens. Maybe instead of higher future taxes, there will be future spending cuts (in fact, reductions in the rate of growth of spending that exceed the rate of growth of debt is arguably the only way that government debt goes down in relation to GDP).

    But generic point is that there are innumerable fiscal, tax and monetary policies between the deficit today and the tax increase in the future such that assuming that a rational actor will anticipate a tax increase and reduce present consumption seems highly questionable. At least in the real world.

  49. Gravatar of Adam Adam
    11. January 2012 at 12:46

    Sorry for the multiple comments, but I forgot something unrelated.

    I understand that S=I is a definition, but is all I created equal? That is, might not the I that results from the S be something with little value? Like excess bank reserves or money under the mattress?

    And isn’t that where the zero lower bound thing comes in? More S means more money available for I which means lower interest rates, but if rates are at zero and there is still more S than demand to borrow it, again, what kind of I do you get?

  50. Gravatar of Charlie Charlie
    11. January 2012 at 12:52

    Cochran just did a post about the Fed and balance sheet risk. I have no idea about how to think about balance sheet risk. What would be the opposite of printing money and dropping it from a helicopter, levying taxes and lighting bills on fire?

    If I tend to think of the Fed needing to commit to a policy response that prints the right amount of money to keep NGDP stable, am I missing important balance sheet implications?

    http://johnhcochrane.blogspot.com/2012/01/worlds-biggest-hedge-fund.html?m=1

  51. Gravatar of Adam Adam
    11. January 2012 at 12:53

    You have strange timing, statsguy, as we’ve just been through a multi-week argument over whether debt matters. And, of course, Keynesians (see Stiglitz, Krugman, Reich (if he counts)) absolutely think distribution matters.

    I guess Mr. Rognlie said what I was trying to say above but more completely.

  52. Gravatar of Michael Kogan Michael Kogan
    11. January 2012 at 13:31

    Scott,

    Krugman ( et. al. ) obviously disagrees with your S=I position, etc. See here http://krugman.blogs.nytimes.com/2011/01/20/barbarous-economics/

    So basically for someone who is not an economist it sounds like a pure political argument

    Krugman says occasionally when people are awash in liquidity but risk averse for whatever reason, they don’t make investments, even though those investments have positive expected return. As a result, economy suffers, and so you can use government infrastructure spending + fed quantitative easing to fill that investment hole

    you don’t like government spending as a matter of course so you don’t like the above. To me it sounds like it is a political matter more than a pure disagreement about models.

  53. Gravatar of Benny Lava Benny Lava
    11. January 2012 at 13:42

    I have nothing useful to add here, I just want to point out that many people think the Nobel prizes are all worthless. A lot of people are still pissed off that James Joyce never won the prize in literature, or that Henry Kissinger won the peace prize. Maybe prizes aren’t good credentials. Maybe there aren’t any?

  54. Gravatar of Mike Sax Mike Sax
    11. January 2012 at 13:46

    The one thing that’s clear is that S=I does not imply the Treasury View

  55. Gravatar of StatsGuy StatsGuy
    11. January 2012 at 13:56

    “And, of course, Keynesians (see Stiglitz, Krugman, Reich (if he counts)) absolutely think distribution matters.”

    So where is wealth distribution in the baseline keynesian model? Krugman often argues that AD is raised if the federal govt. gives money to poor people because it will get spent. In the baseline model? Only in so far as it’s a tool to raise AD, but not at all the only tool to raise AD – in the core keynesian model it doesn’t really matter which tool you use. I suppose it’s the “answer” to ricardian equivalence, but at no point does anyone consider wealth distribution as a core source of the liquidity trap.

    AND YET….

    http://finance.yahoo.com/news/conflict-between-rich-poor-strongest-191040045.html

  56. Gravatar of James James
    11. January 2012 at 13:58

    Adam said:
    “I understand that S=I is a definition, but is all I created equal? That is, might not the I that results from the S be something with little value? Like excess bank reserves or money under the mattress?”

    That’s actually a point that I hear Keynesians make all the time in discussions of spending versus tax cuts as an approach to stimulus programs. If the government spends a $100bn on say expanding high-speed internet access, then we know exactly what it is being used for, and can at least attempt to predict the results. If, on the other hand, it chooses to give $100bn in tax breaks, who knows? Some of it will probably be well spent, some it will go into mattresses, and some of it will go up some addicts nose in all probability, but the point is we really don’t know the distribution and therefore cannot even begin to realistically predict the consequences.

    I think it comes down to a trade off between personal freedom and control, with control being more or less synonymous with predictability, at least in this context.
    There are arguments that can be made that creativity (which tends to be synonymous with personal freedom) generates the most growth, and that tax cuts are therefore more reliable as a means of stimulus, but those are arguments rooted more in philosophy than in economics.

    Of course the same can be said of much of macro theory, from all schools.

  57. Gravatar of bill woolsey bill woolsey
    11. January 2012 at 13:58

    Investment is generally the production of machines, buildings and equipment. However, there is also inventory investment. Firms can buy inputs for later production. They can also produce products for later sale.

    Now, sometimes firms produce products for current sale, but they are in error, and so they hold them for later sale. That is unplanned inventory investment.

    Income equals output (what is produced.) Both are represented as Y.

    Households buy most of the output. That is consumption, or C. (Firms do not do consumption.)

    Firms buy machines, buildings and equipment. That is I. But there is also the inventory investment.

    Y = C + I is an identity when unplanned inventory investment is included. What is produced is equal to the consumer goods households buy, plus the capital goods that firms buy, plus the inventories that firms accumulate because they want to sell them later, plus the inventories firms accumulate because they produce goods that they couldn’t sell. Those last three things are investment.

    Now, saving (S) = Y – C. It is that part of income not spent on consumer goods and services. It is possible that households directly buy capital goods. They might accumulate money and not spend it. They might pay down debts. They might buy stocks or bonds. They might put money is a savings account or get a CD at a bank.

    Remember, Y = C + I, So, Y – C = I

    And S = Y – C too.

    So, S = I. This is true by definitions (given above.) But I isn’t firms buying capital goods only. It includes goods firms accumulate because they couldn’t sell them. It includes this unplanned inventory investment.

    Now, the interest rate mechanism that Rognlie described, is supposed to get saving equal to planned investment (that is, buying capital goods or accumulating inventories on purpose.)

    That means income equals output which equals what the households want to buy and what the firms want to buy.

    There is another process however, that applies when interest rates are not at that level, at it is that firms cut production and so income falls. This reduces saving (because people are poorer and can’t afford to save as much.) And so, saving falls to a level of planned investment.

    Anyway, banks accumulating reserves is not investment. Also, hiring workers doesn’t count as investment. It is employment of labor.

  58. Gravatar of bill woolsey bill woolsey
    11. January 2012 at 14:12

    Scott:

    If MV is constant, then nominal income doesn’t change–agreed.

    If V (which is related to money demand) depends on the opportunity cost of holding money, then budget deficits might impact nominal income by impacting velocity through interest rates.

    That is all true.

    But, if people expand saving to match any increase in the budget deficit, then there is no impact on the interest rate, no impact on velocity, and no impact on nominal GDP.

    And so, “Ricardian Equivalence” is often taken to mean that debt financed government spending cannot impact nominal GDP.

    However, consumption smoothing means that debt financed government spending results in less national saving, higher interest rates, higher velocity, and more nominal GDP–even with ricardian equivalence.

    If velocity is not impacted by interest rates, or the Fed is able to perfectly target nominal income, then nominal income isn’t effected by debt financed government spending. True!

    But it will tend to raise interest rates. If people increased saving to match the increase in debt financed government spending immediately, then there would be no impact on interest rates. I know, you generally don’t care about interest rates, but, you know, it is part of economics. Does Ricardian equivalance mean that government budget deficits have no impact on interest rates? Well, if it is used to fund bridges, it does. If it is used for tax cuts, it doesn’t.

    This might not be relevant to nominal income.

    But this is what has Krugman and company all worked up. It seems that some of the Chicago people are ignoring this distinctino between budget deficits that fund tax cuts and budget deficits that fund additional government spending.

    Those Chicago people don’t even understand their own model. Or else, they are lying–it is all anti-goverment propaganda.

    I admit that I never paid much mind to this distinction because I think ricardian equivalance is implausible. Like Ricardo, I don’t think people really work that way.

  59. Gravatar of UnlearningEcon UnlearningEcon
    11. January 2012 at 14:32

    Matt Rognlie,

    What you’ve described is not at all the same as the old/post-Keynesian view, where monetary policy is *not* about the demand supply for base money, which is deemed largely irrelevant.

    And savings and investment don’t need to be equilibrated – they are the same by definition. That’s my problem with the conventional view of S=I.

  60. Gravatar of D R D R
    11. January 2012 at 14:55

    Huh?

    “Either way Wren-Lewis’s example is wrong. If viewed as accounting it’s wrong because he ignores saving and investment. If viewed as a behavioral explanation it’s wrong because he assumes consumption will fall, but that’s only true if the fiscal stimulus failed”

    What? This is all wrong. As a matter of accounting, all sorts of things might happen.

    For example, if
    C -$10
    I +$0
    G +$100
    T +$100

    Then…
    Y=C+I+G +$90
    Sp=Y-C-T +$0
    Sg=T-G +$0
    S=Sp+Sg=I +$0

    Voila! We have an increase in income of $90 without zero crowding out of investment, and with zero increase in the budget deficit. What happened? Output increased, hence income. The private sector saved 111% of that additional income (by lending it to the government)

    There is nothing in the accounting preventing such an outcome.

  61. Gravatar of D R D R
    11. January 2012 at 15:03

    Whoops on the last line of analysis. Obviously, the private sector didn’t lend anything… it was taxed away.

    But the point remains.

  62. Gravatar of D R D R
    11. January 2012 at 15:20

    Note that the above example results in a “multiplier” less than 1.0. But the accounting works no matter what the multiplier.

    It works with a terrible multiplier

    C -1000
    I -1000
    G +100
    T +100

    Y -1900 (multiplier of -19.0)
    Sp -1000
    Sg +0
    S -1000

    or a very impressive one

    C +1000
    I +1000
    G +100
    T +100

    Y +2100 (multiplier of 21.0)
    Sp +1000
    Sg +0
    S +1000

    Neither outcome seems likely, but I fail to see how the accounting prevents either one.

  63. Gravatar of Jim Glass Jim Glass
    11. January 2012 at 18:00

    Imagine a Nobel Prize winner … who in public debate makes elementary errors that would embarrass a good undergraduate.’

    Such as…

    “the revenue that will be lost because of the Bush tax cuts … would have been more than enough to ‘top up’ Social Security and Medicare, allowing them to operate without benefit cuts for the next 75 years.”

    … that? A mere 14-digits worth of dollars mistake that was never corrected in his column (and which he refused to correct when it was pointed out to him, expressing on his web site his indignation at the thought).

    And while it is not strictly economics, might one also imagine the societal perspective of an economist who uses his economics as a springboard to write about society at length, when he writes things such as …

    “Is this the same country that we had in 1970? I think we have a much more polarized political system, a much more polarized social climate …we’re probably not the country of Richard Nixon …”

    … lamenting the loss of the good old days of National Guardsmen shooting students dead on campus, race riots burning down inner cities, home-grown terrorists like the Weather Underground and SDS blowing up buildings, Spiro Agnew smoothing troubled political waters in the Nixonian era of political good feeling, etc. etc.

    As Patrick said, there are an awful lot of examples of such things, for Krugman to be raising this particular objection about others.

  64. Gravatar of ssumner ssumner
    11. January 2012 at 18:27

    dwb, Yes, that’s also my criticism of Krugman.

    Mike Sax, Aggregate investment MUST equal aggregate saving. But for any individual they are not necessarily equal.

    Patrick, Good point.

    Statsguy, Neither of those assertions correctly describe my views. I agree with Nick about the burden of the debt.

    And I agree that the present value of future consumption matters, not just current consumption.

    Merijn, I mostly agree, if I understand you correctly.

    Kevin, That was shorthand for the interaction of the S and I schedules. I assumed that I didn’t need to speak here at the same level I’d address my students. My students seem happy with my teaching.

    Matt, You said;

    “This is the essence of the Keynesian analysis, and I don’t think Cochrane or Lucas have really acknowledged it.”

    If you read the longer paper by Cochrane that DeLong cites, you’ll find that he does recognize this issue, but addresses it in what seems a strange way to me. By the way, you don’t need to assume anything about zero bounds or interest rates being stuck, the real problem is that when S shifts right, I shifts left in response–indeed even faster. Hence equilibrium S and I fall.

    I agree with your first comment.

    UnlearningEcon, I thought it was both–i.e. IS-LM.

    Matt, I never argue the multiplier is precisely zero, I might say it’s zero in some models, but in the real world I say it’s roughly zero, perhaps positive or perhaps negative. I also say there are some types of fiscal stimulus that have positive multipliers, and some monetary regimes where the fiscal multiplier is positive. My analysis is much more nuanced than you seem to indicate.

    You said;

    “It turns out that the answer is yes: although the central bank can also increase AD when the zero lower bound is binding by committing to low rates after the trap ends, this is costly.”

    I agree it’s costly, but that’s not the preferred policy of us market monetarists. We want to have them peg NGDP futures contracts, and make the base endogenous. In that case interest rates might immediately rise above zero. The liquidity trap is not a problem to overcome, it’s an indication of excessively tight money. That’s what people like Woodford get wrong. FDR raised prices immediately in 1933, and promising to hold rates at zero for a long time had nothing to do with his success.

    So I completely reject the notion that fiscal stimulus allows for a more efficient stimulus, than monetary alone. Money is easily capable of getting the job done.

    Kevin, Nothing you say about savings and investment functions is in any way inconsistent with this post. And as for this:

    “Since S=Y-C it follows that S=I may not be a tautology either. It depends on the context. That is where Scott is going wrong.”

    No. You are wrong and I am right. The variables in those equations are equilibrium levels.

    James, I don’t think anyone can accuse me of “ignoring the effects of time” Most people think I’m far too obsessed with the expectations channels.

    Morgan, You said;

    “The Internet could have started in the 1970’s and it is DeKrugman’s fault it came so late.”

    Are you trying to come across as ridiculous, or just doing so accidentally?

    Adam, Benefit cuts are identical to tax increases in these economic models–and they certainly are anticipated. When I ask my students most say they expect to get less Social Security than promised because of the big deficits. They also say it leads them to want to save more as a result. I think they are fairly typical.

    Some people put cash under the mattress, but it’s not a big factor in the macroeconomy. ERs pay interest, so they are basically debt today, not cash.

    Charlie, I haven’t looked at the Cochrane post yet.

    More to come . . .

  65. Gravatar of ssumner ssumner
    11. January 2012 at 18:49

    Michael, You said;

    “you don’t like government spending as a matter of course so you don’t like the above. To me it sounds like it is a political matter more than a pure disagreement about models.”

    This has nothing to do with it. I love tax cuts, I just don’t think they are effective stimulus (except employer side payroll tax cuts–but I actually like the payroll tax on ideological grounds.) So you are completely wrong about ideological bias coloring my views.

    Benny, The Peace prize is the worst, then literature, then economics, then ? ? ?

    Mike Sax, You said:

    “The one thing that’s clear is that S=I does not imply the Treasury View”

    Finally we agree on something!

    James: You said;

    “If the government spends a $100bn on say expanding high-speed internet access, then we know exactly what it is being used for, and can at least attempt to predict the results. If, on the other hand, it chooses to give $100bn in tax breaks, who knows?”

    That’s either an argument for communism, or (more likely) a false argument. Government spending decisions should be based on efficiency grounds. Adding stimulus to the economy is not an additional argument for G spending unless it’s believed more effective that private spending. But then it should be done even if we don’t need stimulus!

    Bill, I certainly understand all that. But, you said:

    “Those Chicago people don’t even understand their own model. Or else, they are lying-it is all anti-goverment propaganda.”

    I don’t see how you can say that. Cochrane admits that the Keynesian model implies the balanced budget multiplier is positive, i.e. that Ricardian equivilence is not enough. Read his footnote 2 again–it’s very clear. He thinks that that assumption is wrong, or that the effect is small relative to the argument about deficits, but he’s not ignorant of that fact that RE applies to debt, not spending.

    I’ve never claimed that more G can’t raise V, it quite likely can. But that doesn’t mean the multiplier is positive.

    UnlearningEcon, You said;

    “And savings and investment don’t need to be equilibrated – they are the same by definition.”

    What people mean is that the interest rate adjusts until the amount people want to save is equal to the amount people want to invest.

    DR, But that’s not consistent with his consumption smoothing argument. You have private income and private consumption both falling by $10.

    Jim Glass. Didn’t he warn that the relatively small Bush deficits were quickly leading us to a fiscal train wreck, and the far bigger Obama deficits? Well they need to be far bigger still.

  66. Gravatar of Jim Glass Jim Glass
    11. January 2012 at 18:52

    “I understand that S=I is a definition, but is all I created equal? That is, might not the I that results from the S be something with little value? Like excess bank reserves or money under the mattress?”

    That’s actually a point that I hear Keynesians make all the time in discussions of spending versus tax cuts as an approach to stimulus programs. If the government spends a $100bn on say expanding high-speed internet access, then we know exactly what it is being used for, and can at least attempt to predict the results. If, on the other hand, it chooses to give $100bn in tax breaks, who knows?

    Some of it will probably be well spent, some it will go into mattresses, and some of it will go up some addicts nose in all probability, but the point is we really don’t know the distribution and therefore cannot even begin to realistically predict the consequences…

    So on the one hand we can have stimulus effects promptly distributed via the market, and who can know what the effects of that will be?

    On the other hand, we can have a politician-directed spending stimulus very predictably not long delayed and distorted by political wrangling and logrolling, and designed to serve politicians first via paying for political pork, patronage, vote buying, supporting interest groups, and making investments such as Solandrya?

    Predictably, none of that will happen … or all of it will? Since this course is predictible, what is the predictible cost of all that?

    Just as “not all I is created equal”, neither is all government spending. And declining returns to society from it is endemic, according to the considered predictions of many.

    I think it comes down to a trade off between personal freedom and control, with control being more or less synonymous with predictability … arguments rooted more in philosophy than in economics.

    Oh, I think as to economic policy, the arguments are pretty well rooted in economics.

  67. Gravatar of Nick Rowe Nick Rowe
    11. January 2012 at 19:11

    Statsguy: “1) Debt. Does or does not debt matter? Both say no. The rest of the world is beginning to think K&M are obsessed with squabbling over details, and are irrelevant to reality – in which debt does matter.”

    This cuts across Keynesians, Monetarists, and even Austrians! Debt *is* a burden on future generations (unless offset by Ricardian Equivalence or offset by investment in schools etc. that helps future generations. The old “we owe it to ourselves” argument is fallacious. See my 4 posts last couple of weeks.

    Look everyone. We should just stop talking about “saving”. It’s the stupidest concept in macro. I got so frustrated in reading all this that I just finished writing:

    http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/01/why-saving-should-be-banned.html

  68. Gravatar of michael gordon michael gordon
    11. January 2012 at 19:18

    Scott:

    1) A very stimulating post, and the exchanges in the comments are almost as stimulating too. That said, there does seem to be some confusion between different definitions of savings and investment — including the accounting equivalence of S=I in national income accounts and desired (or intended or ex-ante) S and I . . . both calculated over a certain period of time (quarterly, annually).
    .
    2) Desired-savings at the start of the period and throughout much of it will not necessarily equal desired investment by business firms or the government. Obviously.
    By contrast, at the end of a given period — focus on annual GDP if you want — Actual Savings must equal Actual Investment.
    .
    3) The crux issue is how the Actual S and Actual I causally come together.
    * — In Classical Economics, the adjustment mechanism that makes the gap between desired S and desired I — carried out by different economist agents (households on one side and business firms and governments on the other) — is the interest rate. If desired S exceeds desired I at various points in the yearly period — at any rate for more than a few transient hours or days — then presumably the real interest rate will fall.
    The outcome? Households will save less, and simultaneously business investment will increase. By the end of the period, a year, the national income accounts will show the Actual S = equals Actual I.
    .
    *–In Keynesian economics of the orthodox sort, the causal mechanism that brings together desired S and desired I isn’t changes in the interest rate, but rather changes in national income during the course of the year.
    In particular, if desired S at the start of the year is insufficient to match desired I by business firms, then business firms can draw on credit through financial intermediaries and expand investment anyway. (Or the firms might draw on their own reserve-funds too or issue new stock etc. or borrow from banks).
    Expanded I that exceeds desired S will then raised Y (GDP) and as a result households will find their own income raised and save more . . . assuming no external disruptions suddenly jar GDP’s rise, such as a huge spike in oil prices. In turn, the firms’ investment will possibly lead to higher sales. In the upshot, households will put more money into savings accounts, or buy more firm’s stock, or firms’ or government securities . . . just as firms’ increased profits will enable them to meet their interest payments on their borrowed funds.
    ..
    4) CONCLUSIONS:

    —1. CREDIT AND FINANCIAL INTERMEDIATION
    As the last example shows, Keynes was one of the first to note that it was the growth of financial intermediaries — banks of all kinds, the stock market, and so on — and liquid financial investments and credit schemes that disrupted Say’s law. Depending on the money-standard as you note — gold, central bank targeting interest rates, or inflation-targeting — central banks could in principle use inflation-targeting to overcome all these clever financial intermediaries and financial instruments that helped cause the troubles of the last decade.
    In reality? As policymakers, could they be counted on to do this? You yourself, Scott, are generally skeptical on this score, n?

    —2. ANIMAL SPIRITS
    There was another causal mechanism in the Keynesian model that further diluted the role of interest rate changes in influencing the levels of Desired Investment and Desired Savings rates in a given period (a year, say). Business firms didn’t carry out investments, especially on a large scale that had to be paid back over a lengthy future period, by comparing interest rate costs with predicted returns.
    There was, in the Keynesian view, too much uncertainty in business decisions about the future during especially downturns (though “irrational exuberance” helps explains unsustainable stock-market booms). The result? Investment decisions made by firms would be influenced by big swerves in mood-changes about the future of the economy: specifically, swings between optimism and marked pessimism, over the short, mid-, and maybe long-term. (Keynes himself, according to Hicks, thought the short-term could be just a week or as long as a year, though Hicks — whom I studied with briefly — said he always thought it was just a week in duration).
    And so human foible could add to the causes that led to a macro-equilibrium of the national economy that fell far short of full employment.

    –3. THE LIQUIDITY TRAP AND THE SUPPLY OF AND DEMAND FOR MONEY.
    Then, too, Keynes was the first to identify the liquidity trap, which —on his and his followers’ views (which you rightly contest, Scott, or so I believe) — reinforced the inability of monetary policy to reduce real (and nominal) interest rates to a desirable equilibrium between Actual S and Actual I. Especially, of course — as Ralph Hawtrey himself acknowledged in 1937 (according to David Laidler’s “Fabricating the Keynesian Revolution” on p. 286) — because Keynes was the first to explain that the underlying cause of the rate of interest was found in the interaction of the supply of money and the demand for it. From which, needless to add, the liquidity trap emerged. And, signified among things (as Hawtrey also conceded), by the banking system letting reserves build up and so creating a “credit deadlock” (Hawtrey’s term)holding back the transformation of savings into investment. (Hawtrey was the most prominent exponent of the British Treasury view that fiscal policy could not help deal with a recession because of “crowding out”.

    —4. FOREIGN SAVINGS AND THE CURRENT ACCOUNT IN AN OPEN ECONOMY
    So far, the analysis in Scott’s post and comments refer to a closed economy . . . no doubt for ease of analysis.. But, as everybody probably knows, when Y (GDP) = C+I+(G-T)+X-M, the deficient desired national savings fall short of matching desired investment — say, for reasons of hoarding (as Nick Rowe properly observed) or just plain excessive consumption — a current account deficit for a country like the USA can be compensated for by a corresponding inflow of foreign currency. That allows, as it did in our country, National Spending (consumption and investment) to exceed National Output for years on end.
    With what results we all learned about, say, in the financial crisis of 2008 both here and elsewhere. The “Great Recession” that it markedly helped cause globally itself set off by “irrational exuberance” over supposedly safe securitized mortgage bonds and a huge global chain of credit-swaps and delusive efforts to pass on risk-taking to the next guy in that worldwide network.

    —5. KEYNESIANISM AND WAGES AND PRICES
    Finally, of course, Keynesianism of any sort always depends on sticky wages or sticky prices (or both) to cause a bad recession. Does this matter in policy-making? It depends, no? on the model an economist adheres to. My view? Your own persuasive policy prescription for monetary policy that targets NGDP will, if adopted, at least mitigate the depth and duration of a serious recession.
    (Whether a model and theory behind it is logically sound and applicable to a serious problem depends, really, on political and central banking heads applying its prescriptions in policymaking. Only then, given the impossibility of doing experimental work of a persuasive sort in economics, would we know which model is best. And as Krugman and the other exponents of big fiscal stimuli have argued, the Obama stimulus wasn’t big enough to refute their theory and IS-LM model . . . itself repudiated by Hicks in 1980.)

    P.S. I should add that though I have a joint-Ph.D. in both economics and political science, I became a practicing political scientist and hope I didn’t insult anybody’s intelligence here. Scott’s commentaries and the quality of those in his comments section strike me as amounting to the most effective economic blog on the web.

  69. Gravatar of sam sam
    11. January 2012 at 19:48

    Is S=I even close to true when excess reserves are at current levels?

  70. Gravatar of Jim Glass Jim Glass
    11. January 2012 at 20:11

    Jim Glass. Didn’t he warn that the relatively small Bush deficits were quickly leading us to a fiscal train wreck, and the far bigger Obama deficits? Well they need to be far bigger still.

    He certainly did. Krugman I (deficit 3.4% of GDP, debt held by public 36%):

    “I terrified… the looming threat to the federal government’s solvency … because of the future liabilities of Social Security and Medicare, the true budget picture is much worse than the conventional deficit numbers suggest … the conclusion is inescapable … the task is simply impossible. The accident, the fiscal train wreck, is already under way.

    “How will the train wreck play itself out? … my prediction is that politicians will eventually be tempted to resolve the crisis the way irresponsible governments usually do: by printing money, both to pay current bills and to inflate away debt…

    “… investors still can’t believe that the leaders of the United States are acting like the rulers of a banana republic. But I’ve done the math, and reached my own conclusions — and I’ve locked in my rate.”

    Krugman II (deficit 10% of GDP, debt held by the public 53%):

    in 1950, federal debt in the hands of the public was 80 percent of GDP, which is in the ballpark of what we’re looking at for 2019. By 1960 it was down to 46 percent “” and I haven’t heard that anyone considered America a debt-crippled nation when JFK took office….

    So, to review: to make the debt look scary, you have to dismiss the post-World -War II experience, even though it turns out that the 50s offer a quite good lesson…

    While transmorphing from “I’m terrified” into “to make this look scary, you have to…” the obvious $60 trillion flip-turn PK II pulled was to suddenly ignore SS and Medicare liabilities while comparing the fiscal numbers of today to those of the 1950s, when no such things existed. Forgetting the alarm PK I set off: “because of the future liabilities of Social Security and Medicare, the true budget picture is much worse than the conventional deficit numbers…”, which PK I said would be “simply impossible” to deal with.

    Since then PK II and his acolytes have admitted that PK I indeed made a mistake — but only a small and understandable one, as PK I was correct that the Bush deficits incurred while the economy was growing were indeed bad, but nobody could forsee the liquidity trap and zero interest rates coming. He was talking about the siutation existing when he was writing, short-term.

    But this is very disingenuous. PK I was writing about the long term — about protecting himself at the end of his new mortgage, towards 30 years away, when the politicians would “inflate away debt”. Absolutely nobody considered that a near-term risk in 2003.

    The irony is PK I was entirely right about this one. Just look at the CBO projections for 2030. (What happens well before then when the recovery comes and federal interest rates goes back up to 6%?)

    PK II applies Underpants Gnomes logic (steal underpants, ????, profit) to the fisc. Stimulate the economy by running up a lot more debt, ???????, entitlements are contained, ‘to make this look scary you have to dismiss the experience of the 1950s’.

    And then he rips into everyone who agrees with PK I.

  71. Gravatar of Scott Sumner Goes Too Far « Uneasy Money Scott Sumner Goes Too Far « Uneasy Money
    11. January 2012 at 20:47

    […] the monetary authority’s reaction function to changes in fiscal policy. However, Scott in a post today has gone further, accusing Keynesians of confusion about how fiscal policy works unless they accept […]

  72. Gravatar of Understanding Chicago Macro: Misdirecton from Scott Sumner… | FavStocks Understanding Chicago Macro: Misdirecton from Scott Sumner… | FavStocks
    12. January 2012 at 00:25

    […] TheMoneyIllusion » Nobel Prizes for alchemy?: I’m not sure if Wren-Lewis knows this, but Nobel Prizes are frequently awarded to people who don’t accept the Keynesian model…. Lots of people believe the fiscal multiplier is roughly zero, including Lucas, Friedman and me. That doesn’t mean one is ignorant of basic economics…. Cochrane is dividing debt-financed government spending into two components, a tax-financed spending increase and a deficit-financed tax cut….Then he argues the debt-financed tax cut will do nothing, because of Ricardian equivalence. I don’t entirely agree, but it’s certainly a respectable argument. Then he suggests that the balanced budget multiplier is zero.  Obviously the Keynesian model says it’s not, but why assume that model is correct?… [I]t’s not a question of Chicago economists not having studied Keynesian economics. He simply doesn’t agree… […]

  73. Gravatar of UnlearningEcon UnlearningEcon
    12. January 2012 at 02:03

    Scott: that’s what I reject. My position is that the interest rate is determined *solely* by liquidity preference and then determines the investment rate. It does not adjust at all based on S and I.

    Why? Mostly because of empirical evidence from periods with low long term rates, which precipitated growth and high investment.

  74. Gravatar of ssumner ssumner
    12. January 2012 at 05:33

    Nick, Nice post.

    Michael, Good description of the Keynesian model, but I don’t find it a very useful model. It doesn’t determine the price level or NGDP.

    Sam, Yup, exactly true.

    UnlearningEcon, Interest rates tend to be procyclical, low during depressions and high during booms.

  75. Gravatar of Nooooooooooo | Blog Pra falar de coisas Nooooooooooo | Blog Pra falar de coisas
    12. January 2012 at 05:56

    […] But I do remember a few thing of economics to understand that, at least the last time I checked, savings was not defined as “as the resources put into investment projects”. Scott Summer says also that S=I, […]

  76. Gravatar of D R D R
    12. January 2012 at 07:14

    RE: Consumption smoothing

    You mean after-tax income? That’s not an accounting problem, it’s a modeling problem.

    Say you want
    A) the government to spend
    B) no change in government savings, so taxes rise by the same amount as government spending
    C) fiscal stimulus to “work” (net positive multiplier despite the tax hike)
    D) after-tax income to fall
    E) consumption to fall by a lesser amount than after-tax income

    Here you go:

    C -$10 (E)
    I -$40
    G +$100 (A)
    T -$100 (B)

    Y +$50 (C)
    Y-T -$50 (D)
    Sp -$40
    Sg +$0
    S -$40

    Satisfied?

  77. Gravatar of D R D R
    12. January 2012 at 07:15

    Ach, sorry. Sign on T is positive, obviously.

  78. Gravatar of 123 123
    12. January 2012 at 07:43

    Scott, You said:
    “I agree with Nick about the burden of the debt.”

    The source of the burden in the OLG models is the wrong pattern of consumption in the period 1 when the debt is created. You can design fiscal policy that leaves no OLG burden for the future generations by focusing the payroll tax cut on the young workers only. Or you can also argue that the private debt that is induced by the monetary stimulus has the same OLG defect too. So in my opinion OLG criticism is not valid against the fiscal stimulus.

    The choice between the fiscal deficit stimulus vs. monetary stimulus depends on the relative costs of private vs. public debt (debt collection vs. tax collection, economies of scale, financial stability etc.) In 1998 Krugman was a market monetarist as he favoured private sector debt and argued against the fiscal stimulus. My guess is that he changed his mind because he has reconsidered question of the relative costs of private vs. public debt finance.

  79. Gravatar of Morgan Warstler Morgan Warstler
    12. January 2012 at 08:13

    Scott, it only follows logically.

    Inventions happen at rates based upon conducive market situations, the best situation is an unregulated market where governments never have any meaningful power past protecting the rate of invention.

    Those people (DeKrugman) who argue for more government, are simply asking we slow down the rate of invention.

    As such, there are inventions we could have today we do not have.

    And there are inventions in the 1970’s we should have had but did not have.

  80. Gravatar of UnlearningEcon UnlearningEcon
    12. January 2012 at 08:40

    Scott:

    Is that true? Do you have any graphs/tables?

  81. Gravatar of Ted Durant Ted Durant
    12. January 2012 at 09:13

    The primary mistake is the arrogant assumption of equality a) between physics and economics as “science”, and b) between a Nobel Prize and a Nobel Memorial Prize.

    Of course, there’s also the arrogant assumption that either prize is an indication that the recipient has achieved god-like omniscience.

  82. Gravatar of Andrew C. Andrew C.
    12. January 2012 at 11:26

    It often seems like Scott’s way of thinking and Keynesian way of thinking are the same, just draped in different language. But other times it seems like there are real differences. For example when Scott says this about how a monetarist might justify a balanced budget multiplier greater than zero:

    Now that doesn’t mean the balanced budget multiplier is necessarily zero. Here’s the criticism that Wren-Lewis should have made:

    Cochrane ignores the fact that tax-financed bridge building will reduce private saving and hence boost interest rates. This will increase the velocity of circulation, which will boost AD.

    This doesn’t just seem like keynesianism dressed in monetarist language, it seems wrong. Or at least wrong in the keynesian model. If a tax-financed increase in gov spending manages to increase interest rates, it reduces, not increases the balanced budget multiplier by crowding out some investment (or consumption too if C is dependent on interest rates). If, on the other hand, there is no increase in the interest rate, the keynesian model predicts no crowding out, and the balanced budget multiplier is equal to 1.

    There is a kind of causality-reversal between Keynesian and monetarist thinking. Keynesians don’t talk about V, but if they did, they would say the increase in NGDP from the government spending caused V to go up, not the other way around. And if interest rates go up due to an increase in liquidity preference, both V and NGDP rise by less.

    Scott thinks Keynesian language is confusing, but I find MV=PY a lot more confusing.

  83. Gravatar of ssumner ssumner
    12. January 2012 at 12:37

    DR, Yes, I’m satisfied, but that doesn’t help poor Wren-Lewis. He acted as if the consumption smoothing is the only decline in “spending,” yet your example has I falling 4 times more. In that case what’s to stop it from falling 9 times more, which was Cochrane’s implicit assumption?

    123, I have no problem with that, as his point is a sort of theoretical curiosity. I rely on the tax burden (deadweight cost of taxes) and the redcution in investment. (Crowding out.)

    Or perhaps he changed his mind because Bush is no longer President. 🙂

    Morgan, Didn’t the government invent the internet?

    Unlearning Econ, I don’t have graphs, but it’s true for most business cycles. Think of this recession and the 2001 recession and the 1930s, as examples of when rates were low. Rates were high in 1999-2000.

    Andrew, if rates don’t rise then money demand is perfectly elastic, hence a liquidity trap. So a miniscule rise in interest rates cause an infinitely large rise in V.

    And you can’t assume that more G means more NGDP, which is the whole point of the new Keynesian revolution. So which Keynesian model do you like?

  84. Gravatar of Floccina Floccina
    12. January 2012 at 13:05

    And
    Watson and Crick have made some wacky statements. Linus Pauling have made some wacky statements about vitamin c. I am sure that there are many other examples of noble prize winners getting off.

    The other sciences are not that far from economics.

  85. Gravatar of D R D R
    12. January 2012 at 13:26

    Are you kidding me?

    The whole point is that is a difference in model which must be defended. You cannot simply dismiss Wren-Lewis on the basis of accounting.

    Indeed! What is to stop consumption from *increasing*?
    I can still hold assumptions A-C and revise D, E such that

    D) after-tax income to rise
    E) consumption to rise by one-tenth amount than after-tax income (one-ninth the increase in private savings)

    C +10
    I +90
    G +100
    T +100

    Y +200
    Y-T +100
    Sp +90
    Sg +0
    S +90

    Tada! Same degree of consumption smoothing, no additional government borrowing, savings *rise* and stimulus works.

  86. Gravatar of Andrew C. Andrew C.
    12. January 2012 at 14:12

    Scott, you said:

    And you can’t assume that more G means more NGDP, which is the whole point of the new Keynesian revolution. So which Keynesian model do you like?

    I’m working from the plain old Keynesian model. I haven’t had a chance to study New Keynesian models (my macro classes didn’t get that far unfortunately), but I don’t quite follow you when you say “that’s the whole point of new keynesianism.” If Y=C+I+G, then if G goes up, either Y goes up or C+I goes down. My point is if interest rates don’t rise, we have no reason to expect C+I to go down. This is the best case scenario for the balanced budget multiplier. Whereas in the story you tell, it sounds like interest rates have to go up for G to affect Y, which doesn’t make sense to me.

    The model you seem to be working from is governed by:

    1. MV(r)=PY (the quantity equation)
    2. S(Y-T,r)=I(r)+G-T (the loanable funds market)

    Whereas the keynesian model is governed by
    1. M/P=L(Y,r) (LM equation)
    2. Y = C(Y-T,r) + I(r) + G (IS equation)

    These equations are basically equivalent, but the story you tell in your head is very different. The story I tell goes like this:

    1. G goes up, putting upward pressure on Y.
    2. Higher Y leads to rise in liquidity preference, putting upward pressure on r.
    3. Higher r crowds out some investment and consumption.
    4. Overall balanced budget multiplier is less than 1, but probably greater than zero, depending on parameters.

    Whereas you story goes like this:

    1. Higher T makes disposable income Y-T go down, reducing saving. This pushes up interest rates in the loanable funds market.
    2. Higher r causes velocity to pick up, causing NGDP to go up.

    This second story might mathematically be equivalent to the first story, but if we’re arguing about which story is a clearer or more intuitive window to understanding of what goes on in macro, the first story makes a lot more sense to me. But I suppose it’s just a stylistic preference.

  87. Gravatar of Morgan Warstler Morgan Warstler
    12. January 2012 at 14:59

    Yes Scott, they also invented Tang and Velcro.

    One more time: the only reason we don’t hunt down DeKrugman is because we do not know yet what awesome future stuff we don’t get to have today.

    (I love the mental image of a crazed mob chasing DeKurugman)

    Having just returned from CES, I’m even more jaded with this view.

    Also check this out:

    http://www.makerbot.com/

  88. Gravatar of D R D R
    12. January 2012 at 16:21

    OK. Maybe I can bridge this gap.

    If I understand Scott’s argument it is as follows:

    Suppose G and T each rise by $100 and this induces a change in consumption of $z. If I does not change, then Y rises by $100+z and disposable income rises by $z. If an increase in disposable income of $z induces an increase in consumption of $z, then either z=0 or consumers are not smoothing. (Or… I does change.)

    Now, let’s perform a charitable reading of Wren-Lewis. If the government spends $100, aggregate demand increases by $100. Now, one charitable reading is that this is the initial shock, and Wren-Lewis is not suggesting in this that spending multiplier is 1.0. But let us set that aside for the moment.

    If we do read Wren-Lewis as asserting a multiplier of 1.0, then this means a $c change in consumption is offset exactly by a $c reduction in investment. c may be anything, but both income and disposable income must rise by $100.

    If, just for kicks, we assert on Wren-Lewis’ behalf a smoothing factor of 10, then c=10. This results in a rise in private savings of $90, and a fall in government savings of $100. Both national savings and investment fall by $10.

    Now we impose a $100 tax on the consumer. In order to smooth this away, the consumer reduces consumption by $n for a total of increase of $10-n. As for investment, Wren-Lewis did not count any fall in investment here, and his argument does not obviously change if investment increases, but let us first assume that there is no investment effect.

    Now, if investment does not rise as a result of the tax, then income has on net increased by $100-n, and disposable income has fallen by $n. Thus, -n=10*(10-n), or n=$11.11.

    Overall, then, consumption has fallen by $1.11, investment has fallen by $10, and government has grown by $100. Income is up $88.89 although disposable income has fallen $11.11. (Again, we see that 10:1 ratio to consumption)

    Private savings, however, has fallen by $10, and government savings have not changed. Thus, national savings and investment have fallen by $10. (Checks out with investment)

    So. Wren-Lewis is arguing that with a smoothing factor of 10 and no investment effect as a result of the tax hike (neither assumption was stated explicitly) we wind up with a net increase in aggregate demand of $88.89. I don’t see where this falls apart. Did I make a math error?

  89. Gravatar of UnlearningEcon UnlearningEcon
    13. January 2012 at 02:05

    Scott,

    Right but isn’t it true that the fed acts to reduce interest rates in bust times and has a ‘let’s get real’ mentality in boom ones (something I detest, incidentally). Surely we’d have to adjust for fed operations?

  90. Gravatar of ssumner ssumner
    13. January 2012 at 06:48

    DR, The burden of proof in on Wren-Lewis, who said Cochrane was wrong be cause he ignored consumption smoothing. But Wren-Lewis (and Krugman) were wrong, consumption smoothing has nothing to do with it. Cochrane could say that if you assume a $100 fall in private income leads to only a $10 fall in consumption, then BY DEFINITION you are assuming that investment falls by $90, which disproves your point. Wren-Lewis acts like the fall in consumption is equal to the fall in “spending” but he ignores investment spending!

    Andrew, The new Keynesians would argue that interest rates will rise if G rises, that’s why they lost interest in fiscal stabilization after 1985.

    You said;

    “But I suppose it’s just a stylistic preference.”

    No it’s not. Because the Keynesian story leaves out money, and hence is a special theory that will only be right in certain cases, under certain monetary regimes. The monetary approach is always valid, and hence is less likely to lead to error. With all due respect your sense that the Keynesian model is more “intuitive” reflects that fact that you don’t fully understand what’s going on here. It’s much more complicated than those equations make it out to be, which is why new Keynesians have moved on from that model. The focus now is all on future expected monetary policy. I recommend my “special theory” post from a couple days ago.

    UnlearningEcon. No, interest rates were also procyclical before the Fed existed, there is less demand for credit during recessions.

  91. Gravatar of D R D R
    13. January 2012 at 07:32

    Scott,

    As I laid out in my previous comment, I do NOT believe that Wren-Lewis ignores investment at all. I presented numbers which work from an accounting standpoint and are perfectly consistent with Wren-Lewis’ description of the economy.

    You are correct that he doesn’t explicitly mention investment. But he sure does imply it when he says the spending multiplier is 1.0 and there is consumption smoothing. I think that’s about all you need to imply that a change in government spending induces a consumption effect AND an investment effect.

    (Note that offsetting the spending with a tax need not simply reverse all the effects of the spending, because the tax multiplier may be different. I believe that the effect of the $100 tax in Wren-Lewis’ model is to reduce income by $11.11– from +$100 to +$88.89–… resulting in a tax multiplier of only 0.11.)

    If you want to say that Wren-Lewis *makes different modeling assumptions* than Cochrane, then fine. Point out where that difference lies. That is not the same thing as saying he made an accounting error or that his model is inconsistent.

  92. Gravatar of Andrew C. Andrew C.
    13. January 2012 at 09:38

    Scott, you said:

    The new Keynesians would argue that interest rates will rise if G rises, that’s why they lost interest in fiscal stabilization after 1985.

    Ok, but the old Keynesian model also predicts interest rates will rise, as I laid out in my comment. Do New Keynesian models somehow predict interest rates will rise even more?

    Because the Keynesian story leaves out money, and hence is a special theory

    How is M/P=L(Y,r) leaving out money? In fact, how is it different from MV(r)=PY?

    With all due respect your sense that the Keynesian model is more “intuitive” reflects that fact that you don’t fully understand what’s going on here.

    Maybe. But what am I missing? I’m trying really hard to understand your point of view. Was my “second story” an accurate portrayal of your way of thinking?

    It’s much more complicated than those equations make it out to be, which is why new Keynesians have moved on from that model. The focus now is all on future expected monetary policy. I recommend my “special theory” post from a couple days ago.

    I’ve read that post, but all the points you make in it are perfectly comprehensible with an old keynesian IS-LM model. I’m not sure what I’m supposed to have learned that would lead me to reject it. Yes, if the fed targets inflation (and assuming it can credibly do so) then the AD curve is flat and AS determines output. And yes, if the fed wants to counter fiscal stimulus, it can. I don’t see how any of this contradicts anything I’ve said.

  93. Gravatar of TheMoneyIllusion » It’s tough to argue against an identity TheMoneyIllusion » It’s tough to argue against an identity
    13. January 2012 at 12:17

    […] to rehash the Wren-Lewis/Krugman mistake, because I don’t think I explained it effectively in this post.  Here’s how Simon Wren-Lewis criticized John Cochrane’s claim that when the […]

  94. Gravatar of D R D R
    13. January 2012 at 12:54

    Please excuse the length of this comment.

    I suppose it is worth looking at what Wren-Lewis quoted:

    “Before we spend a trillion dollars or so, it’s important to understand how it’s supposed to work. Spending supported by taxes pretty obviously won’t work: If the government taxes A by $1 and gives the money to B, B can spend $1 more. But A spends $1 less and we are not collectively any better off.”

    It is also clear what Cochrane means by “spending”– is this merely consumption, or consumption plus investment? It seems the safest reading is that he means the latter.

    Let cA be the change in A’s consumption and cB be the change in B’s consumption. Similarly, we have changes in investment iA and iB and changes in private spending xA and xB. (G does not change)

    Now, Cochrane states that xB=$1 is in the realm of possibility. This is not a terribly informative statement. Cochrane does not disallow xB$1.

    Wait, what? How is it that xB can exceed $1 unless B dissaves? Well, no. If B invests an additional $1 as well as consumes an additional $1 , then total income rises by $1, allowing savings to rise. That is, it may be possible that…

    EXAMPLE 1
    ———
    cA = -$1
    cB = +$1
    iA = $0
    iB = +$1

    G does not change, so total income (and therefore disposable income) changes by (cA+cB)+(iA+iB)+G = $1. Total consumption is unchanged (cA+cB = $0) so private savings rises by $1. Fun, huh?

    Next, Cochrane states that xA=-$1. Why? Is this an assertion? If not, what leads Cochrane to this conclusion? Does this necessarily follow from previous assumptions?

    Well, we can disprove by counterexample that it does not *necessarily* follow:

    Suppose that A’s spending is unchanged and B’s rises by $1. Then total income rises by $1. Disposable income rises by $1, and private savings… who knows? It may be possible that…

    EXAMPLE 2
    ———
    cA = $0
    cB = +$1
    iA = $0
    iB = +$1

    G does not change, so total income (and therefore disposable income) changes by (cA+cB)+(iA+iB)+G = $2. Total consumption rises by $1 so private savings rises by $1.

    We are now faced with two choices. Either Cochrane has unstated and nonobvious assumptions which directly bear on his reasoning, or Cochrane is *assuming* that xA=-$1.

    So let us proceed on the *assumption* that xA=-$1. Cochrane next states that “we are not collectively any better off.” Based on his next sentence, this means “income does not rise” Since G is unchanged, this implies that xA+xB<=0. Why? Is this an assumption? If not, what leads Cochrane to this conclusion? Does this necessarily follow from previous assumptions?

    Again, we can disprove by counterexample that it does not *necessarily* follow. In fact, Example 1 above does the job.

    As it does not necessarily follow, perhaps, then, it is an assumption. We have assumed from the start that G does not change. If we now assume that private spending does not rise, then sure, it follows that total spending does not change.

    So Cochrane's argument is "A nonpositive number added to zero yields a nonpositive number, and therefore the sum is nonpositive." There is no economics at all in this argument.

    We are forced to conclude that Cochrane has left out at least one nonobvious assumption.

  95. Gravatar of Andrew C. Andrew C.
    13. January 2012 at 13:39

    I’m probably going to regret making this comment (I’m moving into broken-record territory), but this whole issue becomes extremely lucid when you use IS-LM. What Krugman/Wren-Lewis are saying is that people’s propensity to save is less than 1 so on net the government’s policy shifts the IS curve to the right. What happens to output, as Scott has said, depends on monetary policy. If the fed is targeting interest rates, then the LM curve is flat – so the balanced budget multiplier is greater than 0. If the fed targets the money supply then the LM curve is upward sloping and interest rates rise a little crowding out some investment, but the policy is still modestly expansionary. If the fed targets NGDP, then LM is vertical, and all this is a moot point.

  96. Gravatar of Andrew C. Andrew C.
    13. January 2012 at 13:40

    Oops. I meant to post that in the new thread.

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

    Scott, you wrote:

    “Cochrane could say that if you assume a $100 fall in private income leads to only a $10 fall in consumption, then BY DEFINITION you are assuming that investment falls by $90, which disproves your point.”

    Well, yes, if Y-T=C+I falls by 100 and C falls by 10, then I falls by 90. But then I would have to ask what point it disproves.

    Look, Cochrane claimed he proved something, which, if anything, was a trivial result. It most certainly did not prove that transfers from A to B cannot increase aggregate income. It was even less of a proof (if such is possible) that any stimulus is so doomed to failure.

    Either it was a failed proof or it was an example of how stimulus *might* fail.

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

    As I keep arguing, the difference between the two is modeling. Maybe Scott finds neither convincing. That’s his choice. But then Scott argued:

    “Either way Wren-Lewis’s example is wrong. If viewed as accounting it’s wrong because he ignores saving and investment. If viewed as a behavioral explanation it’s wrong because he assumes consumption will fall, but that’s only true if the fiscal stimulus failed.”

    None of these statements by Scott are defensible.

    First of all, Wren-Lewis only wrote that consumption fell in response to the tax hike– not that it fell in response to both actions on the part of the government. Furthermore, Scott’s argument is false, as may be demonstrated by any number of examples (see earlier comments.)

    Second of all, Wren-Lewis cannot have ignored investment– for it follows directly from his assumption that the government spending multiplier is 1.0 that any G-induced consumption crowds out any G-induced investment dollar-for-dollar. It’s too bad for Scott that this was merely implicit.

    Note that I had assumed that this applied only to the spending multiplier. But carrying it over to the tax multiplier only strengthens the case… the total change in disposable income goes to zero, with zero change in consumption, and with zero change in investment… but the $100 increase in G still remains. $100 tax-financed expenditure increases income by $100, rather than $88.89. Private savings, by the way, doesn’t change either. Is it a free ride? Not exactly. Someone had to produce $100 worth of whatever the government bought. So there was some sort of labor involved, and natural resources consumed. But that doesn’t make the argument wrong!

  98. Gravatar of ssumner ssumner
    14. January 2012 at 07:53

    Andrew, New Keynesian models generally assume the central bank is inflation targeting, therefore interest rates rise even more with fiscal stimulus.

    DR, I’ll pass over these because I already responded to numerous comments of yours that are similar in more recent posts. If I missed anything important feel free to bring it up in more recent posts.

  99. Gravatar of 123 123
    15. January 2012 at 05:30

    Scott, You said:
    “I have no problem with that, as his point is a sort of theoretical curiosity. I rely on the tax burden (deadweight cost of taxes) and the redcution in investment. (Crowding out.)

    Or perhaps he changed his mind because Bush is no longer President. :)”

    Yes, OLG is a theoretical curiosity, but is a very fascinating – Nick and Andy can’t wait to receive Krugman’s reply.

    I guess during Krugman still supports the fiscal discipline during normal times because of the tax burden and crowding out arguments. But these days the deadweight loss of private debt collection is high too, as collateral values have declined. Instead of crowding out we could have crowding in, if you throw in some explanation about the Fed not reducing AD in response to a fiscal stimulus. The case of Japan was important for Krugman too – recently he said Japan did it better than Bernanke (the same comparison was recently made by the governor of BOJ), and fiscal deficits were important in softening the damage in Japan.

  100. Gravatar of ssumner ssumner
    15. January 2012 at 07:42

    123, Did Krugman really say that? It would be the second time he changed his mind:

    1. First he said the big Japanese deficits were not helpful, monetary stimulus was needed.

    2. Then when he became reborn as an old-style Keynesian he claimed there were no big deficits after all, it was false accounting.

    3. Are you saying he now says big deficits in Japan made the crisis smaller? I’d love a quotation.

  101. Gravatar of D R D R
    15. January 2012 at 13:11

    Scott,

    Can we get links?

    I’m just guessing here, but I wonder

    1) if he didn’t say that deficits were *insufficient* rather than “not helpful”
    2) if he really called it “false accounting” or if he had some other reason for saying whatever he said

  102. Gravatar of 123 123
    16. January 2012 at 10:32

    Scott,

    What I wrote was a mix of real Krugman and my interpretation of him.

    Here is what he wrote about Japan recently:
    “But Fingleton is right in this: the data don’t match the picture of relentless decline that is so widely held.

    And Japan did go through all this period without anything like the suffering, the human disaster, that America is experiencing.

    I’ve been saying for a while that when people ask whether we might respond to our crisis as badly as Japan did, they’re way behind the curve. We are, in fact, doing worse than Japan ever did.”
    http://krugman.blogs.nytimes.com/2012/01/09/japan-reconsidered-2/

    Here is how Krugman describes his (and Eggertsson’s) debt deflation model:
    “In the model, deficit-financed government spending can, at least in principle, allow the economy to avoid unemployment and deflation while highly indebted private-sector agents repair their balance sheets, and the government can pay down its debts once the deleveraging crisis is past.”
    If we apply his model to Japan, this means that yes, big deficits in Japan made the crisis smaller. I am not sure what is more important for Krugman – the debt part or the government spending part in his deficit-financed government spending proposal, however, both parts are important: ” Also note the middle term: in this model tax cuts and transfer payments are effective in raising aggregate demand, as long as they fall on debt-constrained agents. In practice, of course, it’s presumably impossible to target such cuts entirely on the debt-constrained, so the oldfashioned notion that government spending gets more bang for the buck than taxes or transfers survives. And the model also suggests that if tax cuts are the tool chosen, it matters greatly who receives them”

    However, in my view, the logic is sound. When the monetary stimulus and the associated (net) creation of private debt is too costly (e.g. after a shock to collateral values) or impossible for some reason (central bank time inconsistency), it makes more sense to switch to fiscal stimulus. For me the shock to the collateral values argument is more important, however, Krugman only mentions the central bank time inconsistency problem in his paper.

    Here is more:
    “What does modeling the liquidity trap as the result of a deleveraging shock add? First, it gives us a reason to view the liquidity trap as temporary, with normal conditions returning once debt has been paid down to the new maximum. This in turn explains why more (public) debt can be a solution to a problem caused by too much (private) debt. The purpose of fiscal expansion is to sustain output and employment while private balance sheets are repaired, and the government can pay down its own debt after the deleveraging period has come to an end.
    Beyond this, viewing the shock as a case of forced deleveraging suggests that fiscal policy will, in fact, be more effective than standard models suggest – because Ricardian equivalence will not, in fact, hold. The essence of the problem is that debtors are liquidity-constrained, forced
    to pay down debt; this means, as we have already seen, that their spending depends at the margin on current income, not expected future income, and this means that something resembling oldfashioned Keynesian multiplier analysis reemerges even in the face of forward-looking behavior on the part of consumers.”

    Here is a quotation on Ricardian equivalence from the same paper:
    “It is a familiar proposition, albeit one that is strangely controversial even within the macroeconomics community, that a temporary rise in government purchases of goods and services will increase output when the economy is up against the zero lower bound; Woodford (2010) offers a comprehensive account of what representative-agent models have to say on the subject. Contrary to widely held belief, Ricardian equivalence, in which consumers take into account the future tax liabilities created by current spending, does not undermine this proposition. In fact, if the spending rise is limited to the period when the zero lower bound is binding, the rise in income created by that spending fully offsets the rise in future taxes; the multiplier on government spending in a simple one period liquidity trap consumption-only model
    like the one considered here, but without debt, ends up being exactly one (once multiple periods are studied, and expectations taken into account, this number can be much larger, especially at the zero bound as for example shown in Christiano et al (2009) and Eggertsson (2010a))”

  103. Gravatar of ssumner ssumner
    16. January 2012 at 12:40

    DR, I was basing that on 123’s comment, but his new comment doesn’t read the way he suggested, so I withdraw my suggestion that Krugman flipped twice on Japan. I have some old posts that show him warning about big deficits in Japan. Later he said he been convinced by someone (Posen?, Koo?) that the deficits in net terms weren’t that big in Japan, and that’s why the deficit spending didn’t work. I presume he’s still saying that, but don’t have time to research right now. It’s no big deal, I change my mind all the time.

  104. Gravatar of 123 123
    16. January 2012 at 13:08

    Scott,

    Here is Krugman this summer:
    http://krugman.blogs.nytimes.com/2011/07/16/italy-versus-japan/
    “Both have high debt levels, although Japan’s is higher.”
    So #3 is true.

    On the other hand, #3 is not a flip, as both #2 and #3 are right. #2 refers to net debt which is reasonable in Japan. It is the net debt that creates the burden.

    #3 refers to gross debt. Gross debt provides the stimulus during the debt cycle.

  105. Gravatar of 123 123
    16. January 2012 at 13:11

    Oh, and what are we doing here – spending lots of time analyzing and doublechecking Krugman. If only he did the same before writing his very very wonkish comparative statics nonsense.

  106. Gravatar of I Figured Out What Scott Sumner Is Talking About « Uneasy Money I Figured Out What Scott Sumner Is Talking About « Uneasy Money
    16. January 2012 at 18:00

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

  107. Gravatar of ssumner ssumner
    17. January 2012 at 16:58

    123, I thought net debt proved the stimulus, as it measures the budget deficit. Are you sure it’s gross debt?

  108. Gravatar of 貯蓄は「お金をためておく」ことじゃなくて「資本財形成」だって! by Scott Sumner – 道草 貯蓄は「お金をためておく」ことじゃなくて「資本財形成」だって! by Scott Sumner – 道草
    17. January 2012 at 21:07

    […] 貯蓄は「お金をためておく」ことじゃなくて「資本財形成」だって! by Scott Sumner // サムナーのブログから”Saving isn’t “setting money aside,” it’s BUILDING CAPITAL GOODS“(12. January 2012)。クルーグマンばかり読んでいると、サムナーが財政乗数ゼロの件でボロクソ言われているようですが、訳者が読むに、サムナーがレン-ルイス(とそれを引用称賛したクルーグマン。その邦訳)を批判したのはそういう話ではないと思います(もっと言うとそのことがケインジアン的分析批判にもつながるっているかと)。クルーグマンが引用した部分でレン-ルイスが「spending(支出)」と「consumption(消費)」を混同している(また「貯蓄」についてもだと別エントリで)、というのがサムナーの主眼。よって、先にこちらのエントリを紹介しておきます。  […]

  109. Gravatar of 123 123
    18. January 2012 at 14:01

    Scott,

    Gross debt provides stimulus by repairing balance sheets of liquidity-constrained debtors, although the net debt is in most cases more powerful.

  110. Gravatar of ノーベル錬金術賞? by Scott Sumner – 道草 ノーベル錬金術賞? by Scott Sumner – 道草
    19. January 2012 at 05:31

    […] by Scott Sumner // サムナーのブログから、”Nobel Prizes for alchemy?“(11. January […]

  111. Gravatar of ssumner ssumner
    19. January 2012 at 14:08

    123, I thought the difference between the two was money that went into federal accounts like Social Security, Medicare and pensions. Am I wrong?

  112. Gravatar of Free Banking » Missing from the debate on multipliers Free Banking » Missing from the debate on multipliers
    19. January 2012 at 19:43

    […] back and forth about fiscal multipliers, in a debate with many other participants. (Here are the first post and the latest post by Sumner on the issue.) For those of you who have not followed the debate, the […]

  113. Gravatar of 123 123
    21. January 2012 at 09:05

    Scott,
    in the case of Japan the difference mostly refers to various assets, such as shares, forex reserves, etc. So the expansion of gross debt represents a sort of quasi-QE.

  114. Gravatar of ssumner ssumner
    22. January 2012 at 09:30

    123, Thanks for the info, I’m tempted to quip that “quasi-QE” is like being “quasi-pregnant.” You are or you aren’t. Given that Japan’s had no NGDP growth since 1992, I’d say it isn’t.

  115. Gravatar of Time to Move On – But Not Before I Explain (Definitively) What it all Means « Uneasy Money Time to Move On – But Not Before I Explain (Definitively) What it all Means « Uneasy Money
    24. January 2012 at 12:50

    […] principals were accompanied by various interventions on either (or neither) side by Brad DeLong, Scott Sumner, Nick Rowe, Karl Smith (to name just a few) and by responses and rejoinders by Cochrane, Wren-Lewis […]

  116. Gravatar of Krugman on Mistaken Identities « Uneasy Money Krugman on Mistaken Identities « Uneasy Money
    31. January 2012 at 13:18

    […] economist, into serious trouble. That, I suggested, is what happened to Scott Sumner when, in a post about whether a temporary increase in government spending and taxes would increase GDP, he relied […]

  117. Gravatar of Beware of the Investment=Savings ‘Identity’ « Decisions, Decisions, Decisions Beware of the Investment=Savings ‘Identity’ « Decisions, Decisions, Decisions
    10. August 2012 at 21:08

    […] authors.  The treatment of the ‘identity’ is debated by authors such as Graber and Summers but to my mind the foundational problem has not been […]

  118. Gravatar of Missing from the debate on multipliers – Alt-M Missing from the debate on multipliers - Alt-M
    29. March 2015 at 13:31

    […] back and forth about fiscal multipliers, in a debate with many other participants. (Here are the first post and the latest post by Sumner on the issue.) For those of you who have not followed the debate, the […]

Leave a Reply