NBER study finds zero fiscal multiplier . . .

. . .  in countries with flexible exchange rates (i.e. autonomous monetary authorities.)

In How Big (Small?) Are Fiscal Multipliers? (NBER Working Paper No. 16479), co-authors Ethan Ilzetzki, Enrique Mendoza, and Carlos Vegh show that the impact of government fiscal stimulus depends on key country characteristics, including the level of development, the exchange rate regime, openness to trade, and public indebtedness. . . .

Exchange rate flexibility is critical: economies operating under predetermined exchange rate regimes have long-run multipliers greater than one in some specifications, while economies with flexible exchange rate regimes have multipliers that are essentially zero. The differences in the responses to increases in government consumption in countries with fixed and flexible exchange rate regimes are largely attributable to differences in the degree of monetary accommodation to fiscal shocks in these nations. The results imply that the central banks’ response to fiscal shocks is crucial in assessing the size of fiscal multipliers.   (Italics added.)

I wonder how many multiplier studies modelled the central banks’ response in their estimates of the size of “the” multiplier.

PS.  I haven’t had time to do any blogging, and am somewhat overwhelmed with work.  If you email me don’t expect an immediate response.  I’ll try to do something on the current world situation when I have time.  This post is just a stop-gap to show I am still around.


Tags:

 
 
 

16 Responses to “NBER study finds zero fiscal multiplier . . .”

  1. Gravatar of Full Employment Hawk Full Employment Hawk
    16. March 2011 at 16:57

    “while economies with flexible exchange rate regimes have multipliers that are essentially zero.”

    What assumptions are they making with respect to the small, versus large economy and the degree of capital mobility?

    There are theoretical reasons for believing that this will affect the results. For example if the Mundell-Fleming model is modified to apply to a large country with less than perfect capital mobility and flexible exchange rates, it implies that expansionary fiscal policy (if not offset by contractionary monetary policy) is effective, though less so than in a closed economy.

  2. Gravatar of Andy Harless Andy Harless
    16. March 2011 at 17:51

    IIRC the standard Mundell-Fleming result for a small open economy with flexible exchange rates is that the fiscal multiplier is zero. But IIRC the M-F model doesn’t deal with (and can’t handle) the case where the interest rate is bound by a zero constraint, so it would be interesting to see if the empirical result holds up for a sample with that characteristic.

    Of course, the US today doesn’t have a flexible exchange rate regime because one of our largest trading partners is controlling the bilateral exchange rate. So we’re somewhere between the flexible case and the fixed case. And the US is also an extreme case on the axis between “high-income” and “developing” countries, so even if the exchange rate were flexible, one would have to consider the impact of the observation that “the output effect of an increase in government consumption is larger in industrial than in developing countries” (from the linked summary).

    Obviously, though, when one speaks of a fiscal multiplier, one should specify whether it is ceteris paribus for monetary policy. Both concepts are useful for different purposes, but it is bothersome that casual discourse (and even, I think, a lot of academic work) often ignores that critical distinction.

  3. Gravatar of Mark A. Sadowski Mark A. Sadowski
    16. March 2011 at 21:01

    Krugman’s oft repeated argument is that the zero lower bound negates the classic Mundell-Flemming model argument that fiscal stimulus is impotent in a flexible exchange rate environment (with perfect capital mobility). But I have yet to see a convincing proof of this.

    Have we forgotten how to debase a currency? If so I suggest we send the FOMC to Zimbabwe to take notes. As I have often said before, hand me the printing press and by the end of the day….

  4. Gravatar of e e
    16. March 2011 at 21:04

    Scott,

    I know you’re pretty busy but I wanted to get your reaction to a specific PK article from when you were out.

    http://krugman.blogs.nytimes.com/2011/03/11/ricardian-confusions-continued-seriously-wonkish/

    I’ve bugged you about a similar idea DeLong posted but he basically assumes Ricardian Equivalence and then shows that Fiscal Stimulus will still raise NGDP in the period it is spent in since rational people won’t cut back on private spending all at once. Makes sense, but I am surprised b/c it sounds a lot like the idea of the Fed doubling the money supply but promising to halve it over the course of a few years. Krugman specifically states that monetary policy which is known to be temporary will have no effect.

    It also seems to me like Krugman’s stimulus could be offset by a known temporary monetary contraction. Maybe a known temporary contraction would be damaging but intuitively it seems like a temporary contraction should also have no effect, and if fiscal stimulus can be offset by a policy that has so effect, it seems unlikely to have one itself.

    Anyways like I said I know youre busy and you probably want to respond to his PKs liquidity trap post so I thought I would try and get your thoughts on this before it gets buried.

  5. Gravatar of e e
    16. March 2011 at 21:09

    Actually I guess level targeting is based on the idea that if the fed can be counted on to make up for a negative NGDP shock, making them all temporary, then those shocks will have little or no effect. So it seems like you probably have to think that a temporary monetary contraction will have no effect. I am not sure if follows from PKs thoughts on monetary policy though.

  6. Gravatar of How Might Japan Affect the Global Economy? « Economic Sophisms How Might Japan Affect the Global Economy? « Economic Sophisms
    16. March 2011 at 23:50

    […] Japan where RGDP shocks were due to transient phenomena like non nuclear disaster earth quakes, wasteful stimulus spending and deflationary monetary […]

  7. Gravatar of Richard Allan Richard Allan
    17. March 2011 at 04:28

    Hey Scott, maybe a little off-topic but have you seen this paper? I think it’s amazing :3

    http://ideas.repec.org/p/fip/fedlwp/2010-009.html

  8. Gravatar of Full Employment Hawk Full Employment Hawk
    17. March 2011 at 05:48

    “If so I suggest we send the FOMC to Zimbabwe to take notes”

    The hyperinflation boogyman again! This is designed to undercut rational discussions of the effects of a moderately more expansionary policy. The effect of printing huge amounts of money in causing hyperinflation is well understood and no notes need to be taken.

    Controlling NGDP growth at a rate that will produce a moderate rate of inflation is in no way comparable. And the fact that this will cause a moderate depreciation in the exchange rate is well understood and no notes need to be taken here either.

    With the economy being depressed and with the Unites States having a serious structural deficit in the trade balance, the effect of expansionary monetary policy in depreciating the dollar is beneficial in the short run because it will stimulate consumption and investment spending and therefore Aggregate Demand. And in the long run it is needed to reduce the deficit in the trade balance.

  9. Gravatar of Full Employment Hawk Full Employment Hawk
    17. March 2011 at 05:58

    “Krugman’s oft repeated argument is that the zero lower bound negates the classic Mundell-Flemming model argument that fiscal stimulus is impotent in a flexible exchange rate environment (with perfect capital mobility).”

    But the problem here is that while short-term interest rates are at the lower bound, longer and long-term rates are not and have even recently been rising. So the question is, which interest rate is THE interest rate in the Mundell-Fleming model?

    Clearly analysis of the effects of expansionary fiscal policy in a closed economy will give badly misleading results for an open economy like the U.S. But analyzing them with the small country perfect capital mobility Mundell-Fleming model will also do that.

  10. Gravatar of Scott Sumner Scott Sumner
    17. March 2011 at 06:30

    FEH and Andy, I agree that with different assumptions one can get different results. The title of my post was misleading–my real interest was in the sentence I put in italics.

    Mark, Krugman’s theory implies that rumors of QE2 would have had no impact on stock prices, the dollar, and TIPS spreads. That theory is now in the dustbin of history. There’s no point in even discussing it any longer.

    e, As I recall, the conditions are rather complex. If fiscal stimulus goes into worthless output, then measured GDP rises but actual GDP does not rise.

    If fiscal stimulus goes for things that people already tended to purchases privately, then Ricardian equivalence holds. It crowds out an equivalent amount of private spending.

    The best case is where fiscal stimulus goes into goods that are useful, but that would not be purchases privately. Even then there are all sorts of other hurdles than must be considered, but it is possible that it might work. (I’m ignoring monetary offsets, deadweight losses from future taxes, etc.)

    Richard, I read a similar article by a different Chinese author. I can’t comment, as I find those highly abstract models to be of limited interest. My hunch is that the deadweight loss from inflation is fairly low, as people now tend to shop with credit cards, and the actual money payments (paying the credit card bills) are predictable.

  11. Gravatar of Ralph Musgrave Ralph Musgrave
    17. March 2011 at 11:55

    Since no one is sure what the fiscal multiplier is, it is silly to rely on it. And since no one is 100% sure what the effect of QE or changing interest rates is, it is equally silly to rely on monetary policy alone.

    So instead of separating fiscal policy from monetary, we should combine the two. For example if stimulus is needed, then have the government / central bank machine raise its net spending: e.g. have a payroll tax reduction which is not covered by extra borrowing.

    That’s what Abba Lerner advocated, and Keynes agreed, I think. The only reason fiscal policy ever got separated from monetary, as far as I can see, is that we insist on having two institutions responsible for influencing aggregate demand: central banks do monetary and elected politicians do fiscal. You might as well have a car with two steering wheels operated by two different people.

  12. Gravatar of Mark A. Sadowski Mark A. Sadowski
    17. March 2011 at 12:15

    Full Employment Hawk,
    Don’t ever confuse me with an Austerian inflationphobe.

    I merely invoked Zimbabwe because I find it incredible that anyone can think it is ever impossible to debase a currency. If a total basketcase like Zimbabwe can do it so can the US. The liquidity trap is a total myth.

    IMO your question about which interest rate is the one that matters most in the MF model underscores the problem with thinking about the MTM in terms of interest rates. As Scott has argued ad nauseum interest rates are a very poor indicator of monetary policy stance be they nominal, real, short, long or whatever.

  13. Gravatar of Bababooey Bababooey
    17. March 2011 at 15:07

    The original notion was that national governments would use tax-and-spending policies to compensate when euro-zone monetary policy was too loose or too tight for their economies. But it hasn’t worked out. Spain and Ireland ran surpluses, but not big enough to offset easy money, misguided bankers and borrowers, and inadequate regulations. Deficit limits were evaded by nearly all””hence current attempts to fashion stronger fiscal constraints.

  14. Gravatar of ssumner ssumner
    18. March 2011 at 10:49

    Ralph, You said;

    “You might as well have a car with two steering wheels operated by two different people.”

    This is exactly why we need only one policy. Set monetary policy in a position expected to produce on-target NGDP growth, and you’ll find there is no need for fiscal policy.

    (It would help if they would create a NGDP futures market.)

    Mark, Don’t worry, I’ll never call you a Austerian inflationphobe. :)

    Bababooey, Reading that really drives home my point about monetary policy being waaaaaay stronger than fiscal policy. Even with surpluses, they weren’t able to hold back the tide. Would it be tasteless to use the analogy that for Spain and Greece to try to prevent overheating with fiscal surpluses–when money is too easy for their needs–is about as likely to be effective as holding back a tsunami with a picket fence?

  15. Gravatar of Full Employment Hawk Full Employment Hawk
    19. March 2011 at 16:45

    “The best case is where fiscal stimulus goes into goods that are useful, but that would not be purchases privately.”

    Actually the best case is where fiscal stimulus goes into goods that are complements to goods produced in private markets. For example, if the government paves the roads, the demand for cars increases over what it was before the roads were paved.

  16. Gravatar of ssumner ssumner
    20. March 2011 at 04:55

    FEH, That sounds right. I wish more of the stimulus had gone into removing potholes–that would have been useful.

Leave a Reply