Fiscal expectations traps: Reply to Thoma

Mark Thoma has a response to my recent follow-up posts on the fiscal expectations trap.  Before getting into the details, let me just say that the “Paging Brad DeLong” post was meant to be sort of half-joking.  I would add that I don’t think anyone would deny that I have gotten beaten up for drawing what I thought was a clear inference in a Krugman post, which apparently wasn’t there.  So there’s some history there.  But I am thankful that Mark is taking my idea seriously enough to devote quite of bit of space to it in his very popular blog.

Here’s how Mark Thoma concludes his post (he is discussing Woodford’s model):

Note that the source of the increase in expected inflation is the expected increase in government spending in the next time period. All that’s required for expected inflation to rise is that fiscal policy is expected to persist another period. However, the Fed won’t do anything in response to the rise in inflation expectations because under the assumptions of the model the target interest rate remains negative (and to go back to an earlier point in the discussion, the expected inflation is credible due to observable changes in fiscal policy in the present time period).

The bottom line is that despite recent claims to the contrary, when the economy is at the zero bound fiscal policy is still effective, i.e. it has a multiplier greater than one, even under strict inflation targeting.

If I’ve mischaracterized anyone, I’m sure I’ll hear about it.

I think Mark Thoma understands these models pretty well, probably on some levels better than I do.  But there are many levels to these models.  There are formal mathematical models, and then there are policy implications that require some sort of game theory guesswork.  The policy implications that come out of the models depend in part on the assumptions you make about how one policymaker responds to another.

I have been aggressive in pushing this idea because I am pretty sure that I will be shown correct in the end, it’s just a question of how correct.  If the Congress passes a $700 billion stimulus and the next day the Fed becomes terrified of inflation that they raise interest rates to 20% and put us in a depression, I think everyone agrees that there will be no “fiscal multiplier.”  And I agree that under some of the assumptions in the Woodford model you can get a much bigger fiscal multiplier at the zero bound than otherwise (assuming an inflation-targeting central bank.)  I would never have the courage to attack the nuts and bolts of the Woodford model–he knows much more about this than I do.  What I question is the sort of policy options that he considers.

Let’s return to the monetary expectations trap for a moment.  Woodford argues that if you fall into a deflationary liquidity trap, the right policy is to do price level targeting, not inflation targeting.  As long as the policy is credible, you can still lower real interest rates even after nominal rates hit zero.  I agree (although I’d prefer NGDP targeting to price level targeting, but put that aside.)

So I consider price level targeting to be the benchmark for a central bank that is serious about escaping a liquidity trap.  Bernanke also suggested price level targeting for Japan (around 2003, I believe.)  How could price level targeting fail?  Well suppose Bernanke said we are going to target a price level path that rises 2% per annum from September 2008.  This means prices will be roughly 20% higher in Sept. 2018 (plus compounding.)  The policy would fail if people thought Bernanke would go back on his word after we got out of the recession.  Why might they think that?  The argument is that central bankers don’t like inflation rates over 2%.  And if you fall below target (as we are doing now) the public won’t find promises to get back on target to be credible.   BTW, I don’t agree with this view, I think an explicit promise would be believed.  But put that aside and let’s assume the pessimistic case when the public didn’t think the Fed would allow prices to be 20% higher in 10 years.  That’s a monetary expectations trap.

What can fiscal policy do in a monetary expectations trap?  Fiscal policy is supposed to shift the AD curve to the right, and thus raise inflation.  Mark suggests that the Fed can’t easily stop this, as the fed funds rate is stuck at zero.  He might concede that they could raise rates to 20%, but how likely is that?  The point is that if the Fed keeps rates at zero, fiscal policy has a clear field.  But that’s where I disagree.  From a price level targeting perspective the Fed can tighten monetary policy without any change in short term nominal rates.  How?  By suggesting a lower future price level target.  Or even suggesting that they won’t allow inflation to rise, as it is already below the 2% target path.  Indeed I think the Fed is doing a bit of this with all its exit strategy talk.  They are sending out signals that they won’t allow the price level to climb back to that 2% trajectory.  They are suggesting that they’ll raise rates if necessary to prevent this from happening.  That’s contractionary right now, even if current interest rates stay at zero.  Don’t believe me?  Read Woodford.  He has made a career of emphasizing that changes in future expected monetary policy are far more important than changes in the current setting of monetary policy.  That’s why I’m so confident that I’ll be shown to be correct in the long run.  Woodford is on my side regarding the fundamental assumption required to drive my result; changes in future expected monetary policy are more important than changes in the current policy setting, fiscal or monetary.

The rest is game theory.  And I’m confident there as well.  It seems far more likely that the future Ben Bernanke would try to sabotage fiscal stimulus he thought excessive, than that the future Ben Bernanke would try to undercut an explicit price level target path objective set out by the current Ben Bernanke.  (Not that either is particularly likely.)

One final point.  I am not saying the fiscal multiplier is low at the zero bound, I am saying that the fiscal multiplier is low at the zero bound if you assume the sort of perverse central bank that would create an expectations trap for monetary policy.



7 Responses to “Fiscal expectations traps: Reply to Thoma”

  1. Gravatar of Indy Indy
    5. June 2010 at 16:36

    I like the game theory implications, for example:

    What if Fiscal Stimulus is trying to goose AD expectations, but is also a signal to the Fed not to stand in the way and counter-react, because if they respond with contractionary monetary policy, there will be both *more* deficit-spending Fiscal stimulus (which eventually threatens national solvency) and also calls to remove the Fed’s political independence because it is “thwarting the will of the people” or something.

    If Fiscal Stimulus can intimidate the current and future Fed to lay down their anti-inflation quivers, then Krugman et al would have a good case.

    Otherwise, not so much – and the lack of clearly convincing rebuttals makes me think you’re most likely correct in this matter.

  2. Gravatar of TravisA TravisA
    5. June 2010 at 18:11

    Nice post, Scott. You and Thoma are iterating toward clarity on your point of disagreement.

    At first I thought that you were dead right, but I think that you may be missing a part of Thoma’s argument. If we have deflationary expectations (-3% inflation) and suppose that government expenditures increase inflationary expectations to -1%. Also, suppose that the Fed has a strict zero inflationary target. The market has nothing to fear from any future Fed actions. So under those assumptions, fiscal policy has a current multiplier that won’t be reduced by fear of future Fed actions.

    BUT, you are right if the goal of fiscal policy is to increase inflationary expectations above the Fed’s target — e.g. +3%. The market knows that the Fed won’t allow fiscal policy to have this impact, so the impact won’t occur.

    I think that the Woodward footnote at the bottom of page 25 clearly agrees with you (“…it is assumed that the central bank pursues a strict zero inflation target as long as this is consistent with the zero lower bound…”).

  3. Gravatar of e e
    5. June 2010 at 18:51


    I’m probably not qualified to be in this argument, but I dont see why the “Tinkerbell Effect” if it exists at the zero bound doesn’t exist everywhere for monetary policy. If nominal rates were 2% and the world predicted an upward shock in demand for money so the fed eased to 1% but people thought that the fed would ultimately raise too quickly in order to prevent inflation, why wouldnt you see the same effect at 1% nominal rates?

    I can see an argument that the Fed has really bad credibility at the zero bound b/c they keep messing up there but no sane person would bet that experience with the BoJ in the 90s will outweigh experience with the current fed in predicting the current fed’s actions for long if the current fed stopped acting like the BoJ, would they? Is there some fundamental reason why the “Tinkerbell Effect” cant exist everywhere?

  4. Gravatar of scott sumner scott sumner
    6. June 2010 at 05:25

    Indy, That’s a very good point. We should not expect stable parameter values for the multiplier–it will be highly sensitive to the reaction of the Fed, which will depend on all sorts of subtle factors, including who is in charge of the Fed.

    Travis, I completely agree. But my point is that this is also true of monetary policy. There is no monetary policy expectations trap in that case either. If we have -3% inflation, and Bernanke tries to bump that up to -1% inflation, no future Fed will try to sabotage that action.

    Does that make our positions even closer? Remember, I am not arguing a fiscal expectations trap is likely, I am saying it is more likely that the exceedingly unlikely monetary expectations trap.

    e, You said;

    “I’m probably not qualified to be in this argument, but I dont see why the “Tinkerbell Effect” if it exists at the zero bound doesn’t exist everywhere for monetary policy.”

    You are not only qualified, but perhaps more qualified that Nobel Prize-winning Paul Krugman. You are right and he is wrong on that point.

  5. Gravatar of travisA travisA
    6. June 2010 at 10:27

    Scott, I really don’t think that there is any disagreement between Thoma and you. Thoma is saying that fiscal policy is always effective at the zero bound because he *assumes* that fiscal policy doesn’t boost inflationary expectations above the Fed’s target for inflationary expectations.

    Here’s the question to ask him: can fiscal policy boost inflationary expectations above the Fed’s target for inflationary expectations at the zero bound? It think the answer is pretty clearly no for the reasons you specified.

  6. Gravatar of e e
    6. June 2010 at 11:26


    I think if the fed could credibly allow expectations to be boosted up to their target, they would be credibly inflation targeting which everyone in this argument seems to think would better than fiscal stimulus. Thoma has a game theory argument saying that the fed isnt able to do so if the inflation is caused by monetary stimulus.


    Thanks for the reply.

  7. Gravatar of scott sumner scott sumner
    7. June 2010 at 06:57

    Travis, I think people misunderstand what this is all about. I’m not arguing that fiscal traps are likely, but rather that they are more likely than monetary expectations traps. Thoma had argued that monetary traps were likely–that is what I responded to.

    You said;

    “Scott, I really don’t think that there is any disagreement between Thoma and you. Thoma is saying that fiscal policy is always effective at the zero bound because he *assumes* that fiscal policy doesn’t boost inflationary expectations above the Fed’s target for inflationary expectations.”

    But if this were true, it would also suggest that monetary traps are not a problem, in which case Thoma would be wrong.

    e, You’re welcome.

Leave a Reply