Reply to Brad DeLong

Brad DeLong recently commented on my IS-LM post.  A few reactions:

1.  Brad is responding to the wrong post.  He should be addressing Nick Rowe’s argument that the IS curve slopes upward.  Nick’s the expert on IS-LM.

2.  Brad says:

My reaction: What’s there to learn? Scott Sumner believes in the quantity theory of money:

PY = MV

No, no, 1000 times no!  THE EQUATION OF EXCHANGE HAS NOTHING TO DO WITH THE QUANTITY THEORY OF MONEY.  It would be like saying the Keynesian theory is; Y=C+I+G+NX.

3.  I agree that liquidity preference theory combined with sticky prices make it likely that expansionary monetary policy will lower the fed funds rate in the short run.  But surely if interest rates are important, it is the medium and longer term rates that matter.  And it is not true (in general) that expansionary monetary shocks lower medium and longer term rates (real or nominal.)  This is because monetary shocks affect the economy is all sorts of ways, not just those assumed by Keynesians.

PS.  I’m not sure whether I “believe in” the quantity theory of money (it depends how you define it.)   But Brad’s right that I do believe in tautologies.

HT:  CA


Tags:

 
 
 

13 Responses to “Reply to Brad DeLong”

  1. Gravatar of Kevin Donoghue Kevin Donoghue
    12. September 2011 at 08:20

    I’m not sure Nick’s post is all that controversial. This is really all that I object to:

    Why does my IS curve slope up? It is not because I think that a higher interest rate would increase demand for newly-produced goods. Rather, it is because I think that a sustained increase in demand for goods would have such a strong self-reinforcing effect on desired investment and consumption that interest rates would need to rise to keep output demanded equal to output.

    An increase in demand for goods is represented by a shift in the IS curve, not a movement along it. Nick seems to be trying to get the Multiplier-Accelerator model without the hassle of using difference equations. I don’t think it works.

    But the possibility of an upward-sloping IS curve is considered in some textbooks. Consider a simple linear case: C=a+bY, I=h+kY-uR, G-T exogenous. Goods-market equilibrium requires Y=C+I+G so substitution gives us the IS equation: R=constant+[(b+k-1)/u]Y. This certainly could slope upwards. But if the LM curve is horizontal (which is a fair description of short-run central bank behaviour) an upward slope means that the equilibrium will be unstable. The traditional reaction AFAIK is to invoke the correspondence principle so as to rule out that possibility.

    You know, if you won’t tell Brad what your model is, he may just conclude that you’re a true economist and of the Keynesian party without knowing it.

  2. Gravatar of Morgan Warstler Morgan Warstler
    12. September 2011 at 08:28

    DeKrugman is not an economist.

    Anyone who’s proveable MO is to get more free shit for the poorest is not an economist.

    Economics is concerned with the interests, actions and attitudes of active market participants.

    People who cannot earn enough to cover their own nut are charity cases, and we can ALWAYS focus on delivering charity, but the goal of economics is not charity, or justice, the goal is business.

  3. Gravatar of Kevin Donoghue Kevin Donoghue
    12. September 2011 at 08:44

    BTW, regarding Nick’s question: is this a liquidity trap? Krugman’s answer (originally directed at you), which he repeats from time to time, is:

    Well, my definition of a liquidity trap is, purely and simply, a situation in which conventional monetary policy “” open-market purchases of short-term government debt “” has lost effectiveness. Period. End of story.

    That’s also ye olde textbook Keynesian answer AFAIAC. So if I understand Nick’s post correctly the situation he describes is not a liquidity trap; it’s an unstable equilibrium where the flutter of a butterfly’s wings can cause a boom or a slump. As Robert Solow said in another connection, if the world actually worked like that we would be looking at “the wreckage of a capitalism that had shaken itself to pieces long ago.”

  4. Gravatar of Scott Sumner Scott Sumner
    12. September 2011 at 08:53

    Kevin, I can see how you’d argue it’s a shift in the IS curve. But I see that as weakening the utiltiy of the IS-LM model. If the model says; “More money will shift LM to the right, and also might shift IS to the right, and thus might raise or lower interest rates” then how useful is it? Aren’t there better ways to make that point.

    If DeLong wants to know my model, he can read my blog. I’ve spelled it out numerous times. But it doesn’t boil down to a few equations, it involves issues like game theory (how does the Fed react to fiscal policy changes) and numerous other factors.

    The short explanation of my model is that money supply increases expected to be permanent cause proportional rises in expected future NGDP (via the hot potato effect.) These expected changes in future NGDP cause current changes in all sorts of asset prices, plus current NGDP. If wages and prices are sticky then current changes in NGDP have real effects on employment. The model doesn’t mention inflation or interest rates. But it’s very much in the spirit of Woodford and Krugman.

    Morgan, How can I put this politely. You’re nuts; they are excellent economists with some erroneous policy views.

  5. Gravatar of Scott Sumner Scott Sumner
    12. September 2011 at 09:01

    kevin#2, Sorry, that’s too simple. Krugman says conventional OMOs that are expected to be permanent do have an expansionary effect. In addition, during normal times when rates are positive temporary OMPs that are expected to be reversed a week later do not have an expansionary effect. So there zero bound has little to do with effectiveness. Krugman will have to do a lot better than that definition. It’s way too simplistic.

    And conventional Keynesian theory never even wrestled with the “expectations” issues Krugman deals with in “It’s Baaaack” So no point in even talking about how they would feel about Krugman’s definition–they’re clueless on these issues.

    Plus there were times when Keynes argued that even unconventional monetary policy like dollar devaluation would fail to raise output (i.e. his 12/31/33 letter to the NYT.) So it’s not at all clear what Keynes thought.

    If all Krugman is saying is that a temporary increase in the money supply, expected to be reversed, won’t boost AD, I’m fine with that. Indeed I published the idea 5 years before he did.

  6. Gravatar of Lars Christensen Lars Christensen
    12. September 2011 at 09:13

    Scott, should it be MV=PY or MV=PT?

  7. Gravatar of Nick Rowe Nick Rowe
    12. September 2011 at 09:28

    Kevin: “An increase in demand for goods is represented by a shift in the IS curve, not a movement along it.”

    I disagree. The very traditional IS curve already has the old Keynesian multiplier built into it. It already contains that positive feedback mechanism whereby an initial increase in Yd induces subsequent rounds of further increases in Yd. That’s a movement along the curve, not a shift in the curve. And if the marginal propensity to spend (consume plus invest) exceeds one, successive rounds of that multiplier process will get bigger, not smaller. So the IS curve slopes up.

    I agree with Scott that the IS curve is *supposed* to be the curve that stays fixed when monetary policy changes. It’s only supposed to shift if something other than monetary policy changes. If the IS and LM curves both shift whenever the other one shifts, it doesn’t work very well.

    Whether the equilibrium is stable or not depends on the whole monetary policy reaction function. And how quickly the central bank reacts relative to how quickly the multiplier process works. Drawing a horizontal LM curve fails to describe that reaction function very well.

    But yes, there’s nothing really very controversial about my post. Or shouldn’t be. The same basic idea is there in Old Keynesian multiplier accelerator models. (Those Old Keynesians were maybe a bit shaky on expectations, and on the distinction between constrained and notional demand functions, but there’s some useful stuff in some of their ideas.)

  8. Gravatar of Paul johnson Paul johnson
    12. September 2011 at 10:48

    MV=PY or MV=PT?

    Depends on your model. It’s confusing to use Y because it is no longer a simple identity.

    The original Irving Fisher idea was “T”:check it out from the start of a 1912 article…

    “THE EQUATION OF EXCHANGE” FOR 1911, AND FORECAST
    The purpose of the present article is to supplement the statistics of ” the equation of exchange ” for the United States published a year ago in this REVIEW by including the figures for 1911, and discussing the indications for the future. The equation of exchange, expressed in algebraic symbols,’ is

    MV +M’ V’== PT.

  9. Gravatar of Kevin Donoghue Kevin Donoghue
    12. September 2011 at 11:19

    Scott: “Krugman says conventional OMOs that are expected to be permanent do have an expansionary effect.”

    Well, conventional OMOs don’t involve a committment that they will be permanent. It’s one thing to buy T-bills, it’s quite another to buy them saying “you’ll never get these bills back from us at any price, no matter what happens.” That’s a kind of expectations-management which goes way beyond conventional OMOs.

    Nick: “if the marginal propensity to spend (consume plus invest) exceeds one, successive rounds of that multiplier process will get bigger, not smaller.”

    But then the sequence doesn’t converge. Y goes to infinity. It was for precisely that reason that Keynes said the mpc must be less than one. And if his investment function had included income as an argument he would have been compelled by the same reasoning to stipulate that the sum of the marginal propensities must be less than one.

  10. Gravatar of Lars Christensen Lars Christensen
    12. September 2011 at 11:19

    Paul, I am well aware of that, but I raise the point because it is actually important. There is a difference – and we know that Scott knows Fisher well. So I look forward to hear what Scott thinks on this issue…

  11. Gravatar of Scott Sumner Scott Sumner
    12. September 2011 at 17:13

    Lars; MV=PY, PT is a meaningless number, as it includes zillions in financial transactions.

    Paul, MV=PY certainly is an identity, as the definition of V is PY/M. Many people think V is the average number of times a dollar is spent each year, but that’s false for all sorts of reasons. Lots of transactions don’t involve money. There are multiple definitions of money in any case. Lots of transactions don’t involve final goods that comprise GDP.

    In any case, DeLong was clearly talking about a tautological definition, otherwise it’d make no sense to say I “believe” in MV=PY.

    Kevin, You said;

    “Well, conventional OMOs don’t involve a committment that they will be permanent. It’s one thing to buy T-bills, it’s quite another to buy them saying “you’ll never get these bills back from us at any price, no matter what happens.” That’s a kind of expectations-management which goes way beyond conventional OMOs.”

    Actually, it’s not at all clear what was being assumed. What is temporary? One day, one week, one month, one year, one decade? Without knowing that it’s impossible to say what the conventional view of an X% increase in the money supply really was. In monetary economics classes the “thought experiments” typically involve permanent changes in M, because temporary changes have little or no effect EVEN IF INTEREST RATES ARE NOT ZERO. What matters is not whether rates are at the zero bound, what matters is whether changes in M are temporary or permanent. And that’s always true.

  12. Gravatar of david stinson david stinson
    13. September 2011 at 08:39

    @Scott:

    ” PT is a meaningless number, as it includes zillions in financial transactions.”

    Not sure I follow that.

    You have chosen the RHS of one formulation of the equation of exchange (PY) as your target for maintaining monetary equilibrium. MV=PT is just as true as MV=PY (albeit with a different definition of V) in equilibrium. Why would it not be OK in theory to use PT as a target for maintaining monetary equilibrium? I can see how PT would be hard to measure though, which is presumably one reason why it tends to ignored.

    It seems to me that PT would be the more general case, with PY being a reasonable target based on assumptions about EMH and the demand for money, i.e., a special case.

  13. Gravatar of Scott Sumner Scott Sumner
    13. September 2011 at 17:44

    David, You can maintain monetary equilibrium with any of the formulas, I’m trying to maintain macroeconomic equilibrium, labor market equilibrium. Stable PT won’t lead to stable PY, unless you are very lucky.

Leave a Reply