For the 247th time, the fiscal multiplier is roughly zero

Keynesian economists have never been able to accept my assertion that the fiscal multiplier is roughly zero because the Fed steers the (nominal) economy.  There’s a mental block on their part (or on my part from their perspective) that prevents us from seeing eye to eye on the issue, even if we agree on the need for monetary stimulus.

Everyone seems to agree that the fiscal multiplier is zero when we aren’t at the zero bound.  The issue being debated is whether this is also true when rates are near zero.  I say yes, most Keynesians say no.  See if you think the Fed continues to steer the nominal economy at the zero bound:

Bernanke and his colleagues may be considering more measures to aid growth and improve public understanding of Fed policy, which could be unveiled as soon as their next meeting taking place Jan. 25-26, said Julia Coronado, chief North America economist at BNP Paribas. The Fed reiterated that it expects joblessness to drop “only gradually.”

“They still see downside risks, so I still think they’re tilted toward easing,” said Coronado, a former Fed researcher who is based in New York. She said she expects a new round of asset purchases in the second quarter, or as soon as the January or March meetings should the economy deteriorate faster.

The “recent strength in data” allows Fed officials to “be a little more patient than they otherwise might be,” Coronado said.

Case closed?  Unfortunately the answer is no.  Both sides are dug in pretty deeply, so it will take more than a snippet from to change minds.

Stay tuned for snippet number 248 in the near future.

PS.  When I say “everyone seems to agree,” I of course mean everyone but Joe Stiglitz.



39 Responses to “For the 247th time, the fiscal multiplier is roughly zero”

  1. Gravatar of Brian Brian
    14. December 2011 at 09:13

    Anybody have the over/under on how much time will elapse before Menzie Chinn attacks Scott’s personally for posting this?

  2. Gravatar of Andy Harless Andy Harless
    14. December 2011 at 09:40

    It seems to me your own posts make the case that the multiplier is not zero at the ZLB. To wit: “We expect to fail, but we’ll keep a close watch on things just to make sure.” Would the Fed ever expect to fail if it were NOT at the ZLB? And if not, then couldn’t fiscal policy, by bringing the economy above the zero bound, cause the Fed to stop failing? My contention, also, is that, the deeper we are into ZLB territory, the more failure the Fed is willing to experience (at least up to a point), because the Fed is averse to taking risky assets onto its own balance sheet.

  3. Gravatar of Ryan Ryan
    14. December 2011 at 09:59

    Why is it zero and not negative?

  4. Gravatar of John Thacker John Thacker
    14. December 2011 at 10:05


    But at the same time, the Fed is reiterating that it is not out of options, and is planning to unveil new monetary measures.

    If the Fed is averse to taking risky assets, then fiscal policy will presumably dissuade the Fed from doing those measures.

    In other words, by your own contention, Andy, the multiplier is close to zero.

  5. Gravatar of anon anon
    14. December 2011 at 10:39

    John, no, Andy seems to be right. If, say, investment subsidies or fiscal policy can bring nominal interest rates up to a policy regime where Fed officials are no longer averse to easing, then fiscal policy is “non-neutral” in that sense. And if the Fed is specifically worried about buying risky assets, there is an even better case that the Treasury dept. should provide a guarantee by issuing riskless debt as needed.

  6. Gravatar of Benjamin Daniels Benjamin Daniels
    14. December 2011 at 11:24

    Saying that the fiscal multiplier is zero is totally misleading. You have to decompose the impact: fiscal stimulus has no impact *ex-post* because it only offsets later Fed actions. Even as a Keynesian economist I can agree with that. But fiscal stimulus has quite a lot of impact in any time period where Fed action is constant — meaning that fiscal stimulus can change net nominal spending sooner.

    In addition, to quote from my comment at Worthwhile Canadian Initiative (

    “[T]he central bank is always doing “something”. But what I understand from right-leaning NGDPT proponents like Sumner is what I think of as “strong” NGDPT – the idea that “we should never use fiscal stimulus because the central bank is implicitly setting NGDP anyway”. Sumner argues this quite a lot, and I think it’s naive monetary neoliberalism.

    I support NGDPT, but a “weak” version: “fiscal policy works, but it has no net impact if the Fed decides to target NGDP anyway.” Note – no “net” impact. What this means is that fiscal policy can and should be used to decide the composition of stimulus, because the mechanism for netting out fiscal stimulus is “implicit offset” (ie the Fed does less stimulus later).

    The impact of implicitly offset fiscal stimulus takes three elements: first, the public/private composition, second, the progressivity of the stimulus, and third, the labor-intensity of the stimulus. None of these things can be controlled through the NGDPT lever, and these are exactly the things that lefty economists tend to worry about.

    To see NGDPT advanced without clearly taking into account the role of fiscal stimulus puts off lefties like me. I think that this adaptation of the theory gives us a lot more room to talk about what we think economics is really about.”

  7. Gravatar of Marcelo Marcelo
    14. December 2011 at 11:36


    I think the biggest counter to this argument is that the Fed itself is asking for more fiscal stimulus. To say that the fiscal multiplier is zero would mean that if fiscal policy became more expansionary at the margin that the Fed would contract at the margin (I believe this is what you are arguing).

    What we have seen from the Fed is that they are (for whatever unwilling, I like to think that the hawks keep policy from actually being useful and Bernanke wants more aggressive policy), but I think there is a strong case to be made that further fiscal easing would NOT be matched by ANY change in monetary policy. I think this is the case because, as you have pointed out many times, that no one operating at the upper levels of policy is really using an AS-AD framework, meaning the Fed can say its doing all it needs to, but fiscal policy must step up.

    I suppose my argument, summed up, is that there is an upper limit to what central banks are willing to do. Once they hit this bound (not necessarily the ZLB as we have seen) they will allow for fiscal policy to push AD instead. I don’t think is optimal policy (obviously), but I do think it is a fair representation of the real world.

  8. Gravatar of Ram Ram
    14. December 2011 at 13:03

    You’re talking about two different things. Ceteris paribus, a temporary increase in the government’s budget deficit raises the natural rate of interest, stimulating aggregate demand. If the central bank is targeting aggregate demand (directly or indirectly), then fiscal stimulus must be offset by monetary contration in order to keep aggregate demand on target. The net effect, then, is no change in aggregate demand, but the effect of the fiscal expansion alone is to stimulate aggregate demand. This matters because the central bank doesn’t have to target aggregate demand (directly or indirectly), and so if the fiscal stimulus is offset by monetary contraction, that is not a flaw with the fiscal stimulus, but rather a flaw with monetary policy. Your ultimate point, that fiscal stimulus only makes a difference in practice if the central bank permits it to, stands, but the blame for the inefficacy of fiscal stimulus may lie with the central bank, and not with the very idea of fiscal stimulus.

  9. Gravatar of David Beckworth David Beckworth
    14. December 2011 at 13:24


    There are recent studies by Eric Leeper, Nora Traum and Todd Walker (2011), Lawrence Christiano, Martin Eichenbaum, and Sergio Rebelo (2010), and Michael Woodford (2010) that acknowledge that fiscal policy is muted by monetary policy with one big exception: the zero bound. The argument is that if the policy rate at 0%, fiscal policy can raise inflation expectations and lower the real interest rate. I think there are a number of problems with this view as outlined here:

  10. Gravatar of Morgan Warstler Morgan Warstler
    14. December 2011 at 13:49

    B Daniels, you analysis on Euro is HORRIBLE.

    The logic is this:

    Greece eats it. They eat it deep and large. The ONLY real outcome from them is long term becoming like Southern US states, bending over for businesses on taxes, labor, and regulations. Anything else is a meaningless chapter in my larger story.

    Forgiving debt is NO BIG DEAL, what matters is a country actually having to live on receipts.

    Deep down they know this, they know everyone else knows this, so they will always blink.


    Look, if the FM is zero (or less), it means the government will not be allowed to “invest” – which correctly orders the world…

    Ubermen run businesses, using only private cash, and the public sector only survives where it plays neutral referee without an opinion of what the market turns us into.

    No one smart bets / gambles against the house.

    Real gambling happens when poker players play EACH OTHER, the guy hosting the party knows his place.

  11. Gravatar of Morgan Warstler Morgan Warstler
    14. December 2011 at 14:02

    Scott, I’ve got a new argument I’m ready to stand on…

    You have to FAVOR small business over big business with tax policy and regulator policy otherwise you get big government.

    Big business = big government. Every time.

    Since that is fact, you can’t claim to favor small government unless you favor small business outright.

    In other news, see Matty proves my point here:

  12. Gravatar of Andreas Reinsch Andreas Reinsch
    14. December 2011 at 14:20

    Hi Scott,

    asking as non-economist:
    Would you say the same thing about the eurozone? If so, would you agree that if the current eurozone austerity measures were implemented by all countries simultaneously, they would have no effect on the economy? Or do you disagree, since you expect the ECB to not loosen monetary policy further even if the threat of disinflation or even deflation materializes?

    Of course is only one small country (like Greece) implements austerity measures, there is probably no or only a very small effect on ECB’s policy, so that the fiscal multiplier for this country would probably be expected to be non-zero.

  13. Gravatar of John John
    14. December 2011 at 15:03

    The multiplier seems like a dubious concept. It is especially misused when economists like Mark Zandi say that this amount of government spending on this or that project will result in a x increase in GDP. In order to predict a multiplier effect in a scientific manner, you would have to completely forecast all individual spending decisions of the people who receive money from the project and then contrast this with the completely forecasted spending decisions of a counterfactual scenario.

    Needless to say this is impossible which is why professional economists should make as many predictions as possible so that eventually they will be right and everyone will hopefully forget all the bad forecasts they had in the past.

  14. Gravatar of dwb dwb
    14. December 2011 at 17:02

    yeah, sorry, not clear why its zero at the ZLB either. The feds current policy does not steer the nominal economy, it is pseudo-inflation targeting, which means nominal income can still (and does fall). I am nearly 100% sure that if the US announced a “stimulus” package tomorrow as part of the end-of-the-year payroll tax cut bonds would drop, inflation expectations would rise and the market (including commodities) would rise. Certainly the market would anticipate that some of the current AD slack would be closed. if we were at full employment the fed would lean against it, but clearly we are not and they wouldn’t, until the output gap was closed, threatening inflation. So its tough to make the case the fiscal multiplier is zero at the ZLB faced with abject policy failure in the first place.

    I have to confess when I go back and read older historical works, the line between fiscal policy and monetary policy is sometimes razor thin or nonexistent and therefore hazy in my mind – the central bank and the govt are the same so the govt can print money and spend it (or bury it for people to dig up). so the lines get blurry. I think I would agree that if the treasury minted 2Tn coins and then spent them, the AD slack would vanish. I think I would agree that it would have more effect than identical Fed money creation (the transmission between reserves and credit creation is diminished). Is that money creation or fiscal policy? both! And I don’t see people making decision about the entire expected future path of taxes, so riccardian equivalence is flawed in my mind.

    And most professional forecasters are Keynsians in my experience, so fiscal stimulus might have an effect merely because professional bank economists expect it to. hmm.

  15. Gravatar of ssumner ssumner
    14. December 2011 at 17:50

    Brian, Does Menzie Chinn do personal attacks against UW grads?

    Andy, That’s possible, but I very much doubt it can bring the economy above the zero bound, as the Fed will sabotage the efforts to do so by doing less monetary stimulus–just as this article suggests.

    Ryan, It might be negative, I said “roughly zero.”

    anon, See my reply to Andy.

    Benjamin, You said;

    “But fiscal stimulus has quite a lot of impact in any time period where Fed action is constant”

    Define “constant.” For me, constant Fed action is constant NGDP. For others it’s constant Mbase, or M2, or interest rates, or Taylor Rule.

    There is nothing “right wing” about my preference for monetary NGDP targeting. It has no bearing on the size of government, tax increases, tax cuts, spending, whatever. Those decisions should be made on their merits. Many people wrongly think there is something “progressive” about Keynesian fiscal stimulus. This is a myth. You could slash government spending during booms and have “normal” government spending during recessions and that would be a countercyclical fiscal regimes that satisfies any definition of Keynesian fiscal stabilization policy. Yet it would be right wing. You are confusing two completely unrelated issues.

    In addition, you seem to assume I want no changes in fiscal policy. Not true, I want to abolish both personal and corporate income taxes. Now that is a neoliberal agenda. But it has nothing to do with optimal monetary policy. I support NGDP targeting whether the government is getting dramatically smaller or bigger.

    The idea that you should push a big government agenda as a countercycical device was discredited in the 1970s. It would eventually bankrupt the government, as spending would ratchet upwards during each recession. And if you don’t your “ideological” argument collapses. G just cycles around trend. The fight is over where the trend should be.

    Marcelo, You said;

    “I think the biggest counter to this argument is that the Fed itself is asking for more fiscal stimulus.”

    No, that’s not even consistent with Andy’s argument. Suppose the Fed wants 4.5% NGDP growth, and will use on and off QE2 to get it. Also suppose they hate taking the heat for monetary stimulus from all the wackos. Obviously they’d then prefer the fiscal authorities deliver the 4.5% NGDP growth, so they get criticized less.

    You said;

    “I think there is a strong case to be made that further fiscal easing would NOT be matched by ANY change in monetary policy.”

    Obviously that’s possible, but it’s completely inconsistent with the article I quoted, which seems quite plausible to me.

    1. The Fed responds to growth.
    2. Fiscal policy affects growth (ceteris paribus)
    3. Ergo, the Fed responds to fiscal policy.

    Where’s the flaw in my logic?

    Ram, You said;

    “This matters because the central bank doesn’t have to target aggregate demand (directly or indirectly), and so if the fiscal stimulus is offset by monetary contraction, that is not a flaw with the fiscal stimulus, but rather a flaw with monetary policy.”

    Congress tells the Fed to control AD (not in those exact words, but the dual mandate amounts to that.) And then if Congress tries to control AD they complain that it’s not working because the Fed is DOING EXACTLY WHAT CONGRESS MANDATED THEM TO DO? Sorry for all the caps, but I really feel strongly that everyone is missing the point. Congress needs to think about what the Fed is doing, that’s part of the policy process. Just like supply-siders need to think about what corporations would do with tax cuts. They need to be pragmatists, how will policy work in the real world, not in some fictional setting where “everything is held constant.”

    David, Thanks for the link. Those models assume monetary policy stops at the zero bound. But we know it doesn’t, hence we know those models are wrong.

    Andreas, Good question. Regarding a single country like Greece, you are right that there’s no monetary offsets. But fiscal stimulus is more iffy for small open economies, especially bankrupt ones where confidence also matters. Not saying Greece can do nothing, just that’s it’s a complicated problem. I’d expect a rough offset at the macro level, as long as the ECB has a 2% inflation target. But I’m open to suggestions that they are less proactive than the Fed, and hence a coordinated fiscal stimulus might have a small short term effect over there. I discussed that possibility earlier in “God of the gaps” posts.

    John, Good point.

  16. Gravatar of ssumner ssumner
    14. December 2011 at 17:56

    dwb, You aren’t responding to my argument at all. I’d recommend you go back and read my “god of the gaps” posts. Or consider this. Suppose the Obama administration does no stimulus in 2009. What rate of NGDP growth would the Fed produce in that case? What would their counterfactual policy be? Now justify your claim by citing Bernanke’s known views on monetary policy. Where would unemployment be today? I think it might be 7.6%, or it might be 9.6%, and I can make very plausible cases for each outcome. It’s actually 8.6%, and I have no idea whether fiscal stimulus helped or hurt.

    The stock market likes tax cuts when we aren’t at the zero bound–I doubt it is for AD reasons.

  17. Gravatar of Shane Shane
    14. December 2011 at 18:27

    The multiplier is zero if the central bank is targeting inflation or NGDP. But it isn’t. That’s the problem in the first place. The fact remains that when deflation occurs, the Fed will slowly act to return inflation to the 1-2% range. But it will never push it above that range because it is afraid that we will become the Republic of Zimbabwe-Weimar for some damn stupid reason.

    You are being too logical, forgetting that institutions that have manifestly failed are like cornered animals incapable of reasoning. Bernanke defines the term “learned helplessness”–he’s waiting for some outside savior to do his job. That’s why he’s begging for more fiscal stimulus–he knows that seeing catch-up growth accompanied by moderate inflation would lift the spirits of the FOMC and ward off tightening. But he also knows that the Fed is too tired, hungry, scared, and cold to take such bold action itself.

  18. Gravatar of John John
    14. December 2011 at 19:32


    1. We don’t know what the natural rate of unemployment is with 99 week UI

    2. The stimulus was and will be an absolute waste of money. No matter what Krugman tries to tell you, being at the zero lower bound and having unused resources does not nullify the LAW of opportunity cost. Government stimulus plans have always been about buying votes.

  19. Gravatar of Russ Anderson Russ Anderson
    14. December 2011 at 20:52


    Really? Congress mandated the Fed to maximize employment and stable prices, not maintain high unemployment. The Fed has behaved more like the ECB where maintaining price stability is the priority, even at the expense of high unemployment. The Fed certainly hasn’t maintained stable AD or a return to AD trend. Quite simply, the Fed has refused to do what they are mandated to do.

  20. Gravatar of Benjamin Daniels Benjamin Daniels
    14. December 2011 at 21:07


    By ‘constant’ I mean ‘unchanged in response to fiscal policy’. Depending in the time between policy changes, and depending on institutional barriers to policy changes, this can make a good deal of room for fiscal stimulus to have power. As we discussed the day before ‘Operation Twist,’ I think the current FOMC is very unresponsive and therefore the implicit tightening could be a long way in the future – meaning that fiscal stimulus can increase NGDP *now*.

    Making fiscal desicions ‘on the merits’ is a red herring. The fact is that government investments become extremely attractive when the interest rate is low, and that stimulus dispersed through that channel disproportionately creates employment and output rather than profits. So it’s always a political decision, not one that can be resolved by some sort of CBA.

    I think ‘optimal monetary policy’ is the same for both of us – the Fed should be stabilizing the rate of NGDP growth. But in my conception, the Fed should ‘fill the gap’ between what fiscal stimulus accomplishes and the target, whereas in yours it should be the sole responsible entity, so low interest rates don’t justify additional federal spending (which they can, and do, by lowering the cost of borrowing).

    I’m not pushing a big- or small-government agenda on the whole; I agree with you that that’s somewhat a different topic. In fact I think there is quite a bit of uncertainty as to how that should even be defined (since you could have a very big government even if it spent a small share of GDP, and a very small one even if it spent a large share). But amping up government spending *during recessions* works, even if some of that spending becomes permanent, because the ensuing growth means that a higher nominal level of spending quickly becomes a lower level as a percent of GDP, and the debts incurred in the recovery effort vanish with growth. Just see postwar US or UK debt patterns for proof.

  21. Gravatar of K K
    14. December 2011 at 22:11

    I can’t figure out what you are saying, Scott.

    “the Fed continues to steer the nominal economy at the zero bound” because “a former Fed researcher” says that the Fed is going to do stuff if things deteriorate??? Well, no kidding they are going to “do stuff.” They are not going to throw up their arms and *admit* total failure. They might as well write Ron Paul a letter agreeing that he’d be right to kick them out on the street. The fact of them (or some former Fed employee) protesting that they are still in charge is utterly devoid of significance.

    “Where’s the flaw in my logic?”

    It’s right in 1) The Fed responds to growth.

    I know you think the fed is omnipotent. We all know that. But I don’t believe that and I think the Fed doesn’t believe that either. But they are never going to come out and tell you that, right? You do see that that wouldn’t help them? So lets stop taking whatever the fed *says* they can do as some kind of evidence for what they can or cannot achieve.

  22. Gravatar of Ryan Ryan
    15. December 2011 at 02:39

    Scott, thanks. I thought you might say that. I guess my question is really: “I would expect the fiscal multiplier to be decidedly negative. Is there any theoretical reason I should expect it to be otherwise?” Just wondering if I’m missing something obvious.

  23. Gravatar of sdfc sdfc
    15. December 2011 at 03:42

    Real interest rates have been negative for an extended period and yet the US economy remains sluggish, providing plenty of evidence that monetary policy is relatively ineffective when the economy is in a debt trap.

    Let’s not forget that the increase in the deficit over the past four years has been overwhelmingly caused by an increase in transfer payments and lower tax revenue. You might want to explain how ripping income from households is pro-growth when non-financial private sector debt is around 160% of GDP.

    Public sector borrowing has monetary effects, government borrowing and spending ends up as deposits in private sector bank accounts. The fiscal multiplier is all about cash flow.

  24. Gravatar of anon anon
    15. December 2011 at 08:33

    “Andy, That’s possible, but I very much doubt it can bring the economy above the zero bound, as the Fed will sabotage the efforts to do so by doing less monetary stimulus-just as this article suggests.”

    In order to “do less monetary stimulus”, the Fed must scale back QE and perhaps raise nominal interest rates. But (by assumption) this would make them more willing to ease policy, cet. par. Regardless, it remains the case that extensive stimulus would push interest rates above the ZLB.

  25. Gravatar of Coray Coray
    15. December 2011 at 10:13

    Hi Scott, I have a question. If the fiscal multiplier is roughly zero (when interest rates are sufficiently above zero) why do so many VAR-style models identify a multiplier of roughly 1?

  26. Gravatar of Floccina Floccina
    15. December 2011 at 11:20

    If the Fed is averse to taking risky assets, then fiscal policy will presumably dissuade the Fed from doing those measures.

    It is interesting to me that the Fed has the ability to make the assets that it buys less risky. A little wage inflation and those mortgages get paid.

  27. Gravatar of Dan K Dan K
    15. December 2011 at 19:50


    It would be helpful to me — who has followed you at least irregularly for a while — and perhaps to newer readers, if you could confirm or correct as necessary, my interpretation of what you are saying:

    It’s not that fiscal policy is by definition ineffective, but given how the Fed currently behaves, the Fed insures that fiscal policy will have (roughly) zero net effect (because as soon as it sees a positive effect, say, it pulls back). This would not be the case, however, if the Fed followed a NGDP targeting rule, or perhaps some other objective target (other than inflation).


  28. Gravatar of ssumner ssumner
    15. December 2011 at 21:11

    Shane, The Fed is quite busy with QE2, operation twist, promises of 2 years of low rates. If the fiscal authorities do more, they’ll do less. Count on it.

    Russ, I’m saying that they are mandated to control AD. If they offset fiscal stimulus (and I agree that’s debatable) they’d be doing what Congress asked them to do. Congress wants them to control AD, that means they should offset fiscal stimulus. The question of whether their target is too low (and obviously I agree it is) is a completely separate issue.

    Benjamin, You said;

    “By ‘constant’ I mean ‘unchanged in response to fiscal policy’.”

    We are going around in circles. Now you’ll need to tell me what you mean by “unchanged.” I have no idea what it means for the Fed to “not change” it’s monetary policy. I know what I mean by the term (constant NGDP growth) but I don’t know what other people mean.

    You said;

    “Making fiscal desicions ‘on the merits’ is a red herring. The fact is that government investments become extremely attractive when the interest rate is low,”

    You criticize me for talking about the “merits” and then you answer back with “merits”. That seems illogical to me, or am I missing something?

    You said;

    “But amping up government spending *during recessions* works”

    It would only work if you did it in response to future expected demand side recessions. And with optimal monetary policy there are no future expected demand side recessions. Which is the problem with Krugman’s “Depression economics.” He doesn’t seem to realize it should be “expected depression economics.”

    The postwar debt payoff was through unexpected inflation, it doesn’t have much bearing on our current situation.

    K, Bernanke believed monetary policy was all powerful when he was an academic. You think he was making it up then, before he had any political reason to lie? He and I did almost the exact same research on the Great Depression, and we saw how powerful monetary policy was. His academic writings make 100% sense to me. It was my conclusion too. Why in the world would you want me to assume he’s suddenly started lying about things he’s believed his entire adult life? I don’t get it.

    Ryan, That’s too complex an issue to discuss here. I’d guess it’s close to zero, but it might be negative if government is very inefficient.

    sdfc, You said;

    “Real interest rates have been negative for an extended period and yet the US economy remains sluggish, providing plenty of evidence that monetary policy is relatively ineffective when the economy is in a debt trap.”

    You are obviously new to my blog, real interest rates tell us the economy is weak, they say nothing about whether money is easy or tight. Interest rates are the price of credit, not the price of money.

    anon, We’ve done “extensive stimulus,” and we are still at the zero bound. Japan did “extensive stimulus” for 20 years, and is still at the zero bound. How much evidence do people need?

    Coray, It’s not just me who says the multiplier is zero, it’s the standard model. I presume they get that estimate because they don’t account for the monetary policy reaction to fiscal stimulus. They do “ceteris paribus” estimates, which of course have no policy implications, because ceteris isn’t paribus. Or they might pick up supply-side effects from cuts in MTRs.

    Dan K, No, I’m afraid it would also be ineffective if they followed a NGDP target. Any fiscal stimulus that attempted to raise NGDP would be met with offsetting Fed tightening.

    Everyone, I’m way behind in grading so the older comments will have to wait until this weekend to be answered.

  29. Gravatar of Andy Harless Andy Harless
    17. December 2011 at 09:58

    Scott, now AFAICT you’re suggesting that there is a “regime discontinuity” at the zero bound, whereby the Fed will “fail” in a consistent way (i.e., have a certain effective NGDP target that is below its unconditional preference) when we are at the zero bound and succeed in a consistent way (i.e., aim for its unconditional preference) when we are above the zero bound. So in that world, the marginal fiscal multiplier is infinite at one point (when we are just barely at the zero bound) and zero everywhere else. I find that theory rather implausible. Can you write down any vaguely reasonable objective function for the Fed that would justify that kind of discontinuity?

    It seems much more plausible to me that, if the Fed totally hates to do that little bit of QE that is necessary to get us to its unconditional preference when we are just barely at the zero bound, then it will even more hate to do the huge amount of QE that gets us to its “barely at the zero bound” preference when we are deep into the zero bound. If hitting the zero bound causes the Fed to deviate from optimal policy, then falling deep into the zero bound should cause it to deviate even more from optimal policy. So the fiscal multiplier is positive whenever we are at the zero bound (although I can imagine it starts to become smaller as we get extremely deep into the zero bound, when the marginal deviations from optimal policy become so costly that they start to outweigh the Fed’s dislike of unconventional stimulus).

  30. Gravatar of sdfc sdfc
    18. December 2011 at 04:22

    Interest rates are the price of credit, not the price of money.

    Considering money supply growth is pretty much driven by banking sector financing decisions that’s a pretty meaningless statement.

  31. Gravatar of Scott Sumner Scott Sumner
    18. December 2011 at 08:38

    Andy, That’s a good argument, and I concede it might be true. But there are other arguments cutting the other way:

    1. Fisher’s recent statement that implies the Fed needs to be tight to offset the damage done by reckless fiscal stimulus.

    2. The Fed probably underestimates the effectiveness of it’s various “nuclear options” such as an explicit target, level targeting. It’s quite possible that the Fed would have exactly that had there been no fiscal stimulus in 2009. In that case maybe we get more NGDP growth than the Fed expected.

    I would add that I think plausible levels of fiscal stimulus are quite weak. They have little impact on the amount of unconventional stimulus required to hit any given NGDP level. So even if my simplistic “normal times/bad times” dichotomy is too simplistic, it may not be far from reality.

    I’d probably be a bit less obsessive about this issue if I saw evidence that the various progressive bloggers even understood the dilemma raised by the article I quote. The problem it raises for estimates of “the multiplier.” But by and large I don’t see evidence of that awareness (with a few exceptions.)

    sdfc, I define money as the base. That’s determined by the Fed. It’s often true that the price of money is soaring even as the price of credit is collapsing. This occured in 1929-1933 for instance. One of the great weaknesses of the Keynesian model is that they overlook that distinction. It helps to consider monetary policy in an economy without banking, if you’d like to better understand the difference. Money’s been around far longer in human history than banking.

  32. Gravatar of dwb dwb
    19. December 2011 at 13:31

    dwb, You aren’t responding to my argument at all. I’m agnostic on the fiscal multiplier at the ZLB because I think good arguments can be made either way.

    IF the fed leaned against it, sure, the multiplier is zero. But if the Fed was doing what they were supposed to do we would not be at the ZLB to begin with. Heck, if they were doing what they were supposed to do, all the time, then this ridiculous scenario they dreamed up would not be keeping me busy (its riducluous mainly because they might let it happen). It’s ridiculous that the ECB warns about the recession risks in Europe.

    Plus, even if its not zero I would prefer not to expand govt spending except as a very very last resort (govt spending no matter how “temporary” seems to generate a life of its own, as I can testify since I live near DC, in a state that actively covets/seeks large multi-year govt spending projects). So, investigating the fiscal multiplier to me is a bit like asking if two wrongs make a right (using fiscal policy is as bad as the poor monetary policy its trying to correct).
    I’d by far prefer we were not in the position to ask the question in the first place.

  33. Gravatar of ssumner ssumner
    19. December 2011 at 14:41

    dwb, Fair enough.

  34. Gravatar of orionorbit orionorbit
    21. December 2011 at 00:30

    There is something I really don’t get, that is, if we look at the Fed and the Government as a consolidated entity, why would the multiplier on 1 dollar that the govt spends at the ZLB be any different than 1 dollar that the fed spends?

    The only plausible story I can think of is that when fed spends it invests it “passively” i.e. it doesn’t have to make any decision on how to spend the stimulative funds, it simply lends to the banking sector which allocates it to various investments. On the other hand when the govt spends it has to decide how to allocate the capital and we can reasonably assume that the govt allocates capital worse than the private sector.

    Is there some other difference I am not getting?

  35. Gravatar of Scott Sumner Scott Sumner
    21. December 2011 at 07:40

    orionorbit, You are missing a huge distinction. The Fed creates new money, whereas the Congress moves around money that already exists (via taxes or borrowing.)

    Money creation is far more expnasionary in normal times.

    At zero rates it’s much harder to evaluate money creation, it depends whether it’s expected to be temporary or permanent. I agree that temporary money creation has little effect (indeed even if you aren’t in a liquidity trap, it has little effect.)

  36. Gravatar of K K
    21. December 2011 at 12:35


    The accused man is protesting that he is innocent. And you write a post saying “Look, I told you he’s innocent. He just said so!!!”  What I’m saying is that he may or may not be innocent, but the fact of him saying so *on the stand* is no evidence for or against. Now you say that he has always been a good man, and he would never, and whatever… And that’s a good argument (or it would be if it weren’t for the fact that he has publicly come out and said that he has changed his mind and expressed far greater sympathy for the predicament of the BOJ than he had as an academic, i.e. he admits to criminal tendencies). The point is, when somebody is *absolutely required* to say something by virtue of their predicament and there is no upside for them in not saying it then the fact of them saying it is evidence of exactly *nothing*. Why do you not see this? You have to distinguish between these two propositions:

    1) Bernanke believes

    2) Bernanke says *means* Bernanke believes

    Logically they are not the same and proposition 2 patently false. I don’t buy the first one either, but at least it’s debatable.

  37. Gravatar of orionorbit orionorbit
    21. December 2011 at 18:17

    Scott, yes I agree about your point on normal times, but let’s just stick to the ZLB times, where your disagreement with Keynesians lies.

    Right now, the banking sector holds a lot of cash in the fed which does not raise the NGDP because it is not lent to the people and businesses that would spend it. That is because the banks think it’s safer to earn a few bp with no credit or interest rate risks, rather than lend it to an investment grade company. They could lend the money they got from QE for 2 years to a private company with no interest rate risk, given the fed’s commitment. They don’t, which means they think the credit risk is too high so they’d rather forego the extra 100bp or so that they would be making.

    Now, when the government issues debt, it does not “shift money around”. As long as the private sector thinks this debt is credit risk free, the government by borrowing simply tells the bank that instead of leaving their proceeds from QE at the fed and earn almost 0, they can lend it to the government for say 5 years and earn say 2%.

    So the way I understand it, you and the Keynesians agree on the problem (too low NGDP) and what to do about it (expansionary monetary policy). Your disagreement is on what to do in order to make QE end up in the real economy.

    Keynesians want that the government issue bonds, so that banks will use their fed balances to buy them (because cash and bonds given fed guarantees to stay on hold are assets identical in credit and interest rate risk; their only difference is in the interest they pay, so it’s reasonable to assume that when the govt borrows in the ZLB banks will jump to buy the assets which are identical to their cash and pay slightly more). So keynesians say “well, let’s have the govenrment take up all the money and spend it on something”

    on the other hand you favor monetary means. Your answer to the “cash sitting idle at the fed” is that the govt says to them “if you don’t lend your money to someone i’m gonna tax you”, which it can do through either introducing negative interest rates on reserves (here the cash flows are identical to a tax on reserves if you see the fed and the govt as a consolidated entity), or through inflation (which is easier to implement but has the drawback that it is a tax on everyone, not just banks).

    So to sum up, you and the Keynesians agree on the problem (too low NGDP) but dissagree on the solution.
    Keynesians: “the consolidated govt should trade all the idle cash the fed created for bonds and spend it on behalf of the banks”
    You: “the consolidated govt should tell the banks to spend the idle cash or else”.

    If I am right up to this point, then the disagreement is merely on who should make the decision on how to spend the cash. The keynesians say it should be the government, you say it should be the banks. And unlike NGDP, where every impartial observer looking at the data would agree that the problem lies in a failure of the consolidated fed/govt to keep it on its steady 5% growth path, there exist no clear evidence that would overwhelmingly convince you that it’s better that the government decides how the money be spent, or convince the keynesians that it should be banks that make this decision.

    And if I am right on this one, there is no “mental block” or any other kind of deep disagreement. The reason you disagree on what to do is simply lack of evidence on who is better suited to decide how the money the fed printed should be spent. And this is a very very small disagreement, that does not warrant posts like “MULTIPLIER IS X!!! I TELL YOU IS XX”, at least in my eyes, given that I don’t understand why you think that a dollar from the fed reserves spent by the government has a multiplier of 0 while the exact same dollar spent by the banks has a multiplier of >0. The only plausible explanation i can think of is if you think that banks will spend the money more efficiently. But even then you would have to do more explaining on why this “higher efficiency” should be interpreted as a 0 multiplier for govt and >0 for banks, because to me it seems more reasonable to assign say a 0.5 multiplier for govt and 1.5 for banks or something like that? Unless there is something that else i am missing?

  38. Gravatar of Scott Sumner Scott Sumner
    22. December 2011 at 08:06

    K, Obviously we can never resolve this, but here’s my botrtom line.

    1. Someone believes X as an academic (presumably his true belief)

    2. He takes a government job and nothing happens that would lead him to revise his belief in X.

    3. He continues to insist he believes in X.

    The most likely interpreation is that HE DOES BELIEVE X.

    orionorbit, I don’t accept your entire framwork, including the way you describe Keynesian views.

    I think the Keynesians now agree with me that a higher NGDP target is the way to go. I don’t favor taxing banks or the public via inflation, I favor encouraging people to reduce real cash balances via faster expected NGDP growth, that’s very different. And Keynesians say they also want faster NGDP growth, so there is no difference there either. The inflation “tax” would be the same either way.

    It’s not about whether the banks or the government spends the money, it’s about whether the government or the private sector allocates resources. I think each should play a role, but that has nothing to do with monetary policy. The proper response to undesirably low NGDP is not to raise the government’s share of the economy, in terms of allocating resources, it’s boosting NGDP back to normal. Government spending should be based on its merits.

    You are wrong about banks allocating resources. An expansionary policy that raised the equilibrium interest rate above zero would quickly lead to almost all excess reserves being pulled out of circulation. You are confusing money and credit, which is a common mistake of Keynesians. Banks allocate credit.

  39. Gravatar of orionorbit orionorbit
    23. December 2011 at 00:04

    Well, as I understand it the difference between money and credit is that the latter comes with an obligation to pay interest. If the consolidated fed/govt issues debt, it has to pay interest on whatever amount it spends. If the consolidated fed/govt issues pieces of green paper it doesn’t have to pay interest on it, but someone else is doing the spending (those that the banking sector decides to allocate the credit to). But with the fed on hold for 2 years, the government can issue 2 year bonds which -at the ZLB- are the same as cash. The distinction between a $1 2y bond that pays 0% and a dollar bill is that if I am holding the first I am also holding interest rate risk, while if I am holding cash I am holding the option that if interest rates rise i can invest it with a yield higher than 0%. But if the fed commits to stay on hold for 2 years, the option component of the dollar bill is worthless and a 2 year bond is for all practical purposes the same as cash, here the distinction between credit and money breaks down. So if the government decided to borrow and spend, the part of the spending that would be financed by <2y paper should have the same multiplier as the fed printing cash (minus whatever loss of efficiency you get by altering the govt purchases as a share of GDP) so it would not be zero, or am I wrong?

Leave a Reply