Do you feel “stimulated” yet?

In a post on the New Deal from last December Paul Krugman presented an interesting graph from a paper by Gauti Eggertsson.  The graph showed investment soaring in the first four months of the Roosevelt administration.  I don’t know the exact numbers, but I do know that industrial production rose by 57% over that four month period.  Krugman then links to an AER paper by Eggertsson with the following abstract:

This paper suggests that the US recovery from the Great Depression was driven by a shift in expectations. This shift was caused by President Franklin Delano Roosevelt’s policy actions. On the monetary policy side, Roosevelt abolished the gold standard and””even more importantly””announced the explicit objective of inflating the price level to pre-Depression levels. On the fiscal policy side, Roosevelt expanded real and deficit spending, which made his policy objective credible. These actions violated prevailing policy dogmas and initiated a policy regime change as in Sargent (1983) and Temin and Wigmore (1990). The economic consequences of Roosevelt are evaluated in a dynamic stochastic general equilibrium model with nominal frictions.

Both Krugman and Eggertsson are Keynesians, whereas I am not.  But we all agree that the key to monetary stimulus is raising inflation expectations.  Increases in the money supply have little or no effect unless they lead to higher expected inflation.  Eggertsson makes the same argument for fiscal policy.  I don’t know whether Krugman agrees, but it would certainly be consistent with his forward-looking approach to macro.  In any model with rational expectations a fiscal stimulus should raise current inflation expectations, which should then raise current output.  Note that just as with monetary policy, there should be some expansionary impact from the mere expectation of fiscal stimulus, even before the actual disbursement of funds.

Of course the US economy in January was nowhere near as depressed as in March 1933, so I certainly wouldn’t expect a 57% increase in industrial production over 4 months.  But surely any effective stimulus should produce some economic growth.  In early January Krugman was somewhat pessimistic about the economy, as the following post shows:

A bit more information to go with my post on stimulus: forecasters surveyed by the WSJ predict, on average, that unemployment will reach 8.1 percent by end 2009 and peak at 8.4% “” that is, it will keep rising into 2010. That’s a forecast of what will happen with the stimulus plan, not of what would happen absent stimulus, which would presumably be considerably worse.

Two points: 1. This suggests that the consensus is at least as grim as the scenario I laid out 2. This looks to me like an economy that could easily be spun by conservatives as a failure of Obama’s policies.

Of course things are already considerably worse—unemployment is 9.4% and may go even higher by the end of 2009.  An 8.1% rate would look pretty good right now.  There are two ways to spin all this:

1.   Krugman warned the fiscal stimulus package was inadequate, and he was right.

2.  The deficit outlook is already horrifying, and there is no realistic chance that the stimulus package could be enlarged enough to make a significant difference.  Fiscal stimulus failed.  Obama failed to turn expectations around rapidly (as FDR did.)

Krugman’s prediction that conservatives would crow about the failure of fiscal stimulus didn’t take long to come true.  In a recent post entitled “They think we’re idiots”, Krugman complains about all the Republican pundits saying “I told you so.”

By and large I’ve long been inured to the deliberate stupidity of much political discourse. But for some reason the vision of Republicans whining, “where are those 3.5 million jobs Obama promised” “” less than four months after the stimulus bill was signed, and with hardly any funds disbursed “” got to me.

Call me an idiot, but I don’t see why it should matter that not much of the fiscal stimulus has been disbursed.  Not much stimulus had been disbursed in July 1933, by which time investment had soared—a point Krugman cited with approval.  The new Keynesian models used by Krugman, Svensson, Woodford and Eggertsson (all of Princeton University) strongly emphasize the fact that current AD is powerfully affected by expected future AD.  Yes, I suppose it is too soon the call the fiscal stimulus a failure, but when you suggest that critics will call the program a failure if unemployment rises to 8.1% by yearend, don’t be outraged when an unemployment rate of 9.4% in May starts people thinking that maybe those clueless “fresh water economists” were right after all.

A few sharp-eyed observers might notice my apparent hypocrisy.  I continue to complain that we are stuck in recession because money is “tight,” whereas by most conventional measures it simply isn’t loose enough.  Isn’t Krugman making the same claim for fiscal policy?  Yes he is.

But here’s the difference; there are real constraints on the possible size of fiscal deficits.  We are already looking at a massive future tax burden on the economy, which will be a drag on growth for years.  In contrast monetary stimulus reduces future fiscal deficits.  And if the monetary stimulus is powerful, the fiscal deficits get much smaller.  Krugman has mentioned some possible capital losses from QE, but under the EMH the expected loss is zero, and in any case even the numbers he cites are small potatoes compared to the prospective fiscal deficits.

Why did FDR succeed where Obama failed?  FDR relied mostly on a highly effective monetary stimulus.  Not merely changing the current money supply with a few desultory OMOs, but changing the entire future path of monetary policy.  It’s still not too late.



18 Responses to “Do you feel “stimulated” yet?”

  1. Gravatar of saifedean saifedean
    9. June 2009 at 13:41

    Here’s an alternative hypothesis that might fit better with your points:

    FDR and Obama both failed, and you and Krugman are both wrong in your recommendation. Fiscal stimulus fails because it causes deficits and inevitably inflation and it substitutes public spending (dictated by bureaucrats) for private spending (dictated by people’s preferences). Monetary expansion fails because it monetizes the government’s debt, paying for the government’s bad decisions from people’s money, debauching the currency and ruining people’s ability to plan.

    The 57% statistic Krugman mentions is no evidence of recovery or improvement. The Soviet economy, before its collapse had massive industrial production, and yet it all crumbled when it turned out that it all came down to painted rust. Real productive activity is only valuable when it gives people things they want””and are willing to pay for. If the government tomorrow decided to hire every man and woman in America to produce corkscrews that no one needs, then surely industrial production would rise, but that would hardly increase real wealth. On the contrary, it would be massive waste, since no one really needs these corkscrews, and a lot of resources were expended on this waste.

    But by obfuscating the real micro picture of the real economy with meaningless aggregates like AD, then this corkscrew production was the best policy imaginable. And the Soviet Union is the most productive economy in history. (This, incidentally, explains Samuelson’s continued infatuation with the Soviets.)

    By using these aggregates, economists can continue to look at the New Deal as if it spurred the recovery, when clearly it did nothing of the sort. Economic engineers sought to increase meaningless aggregates, thinking that that would somehow translate to magically increasing wealth. But all they did was transfer resources from productive activities to make-work activities, reducing real wealth. The aggregate statistics show continued recovery during the Great Depression and WWII. But anyone who lived then knows full well that there was continuous deterioration in the economy from 1929 to 1945.

    It was only after WWII ended, after the powerful FDR was replaced by the meek Truman, after the majority of the new Deal statutes expired, after the Republican Congress and the Democratic President were constantly at loggerheads and unable to do anything, after the return of troops from unproductive wasteful war to productive economic activity, and after the massive DROP in government spending that the economy recovered.

    In short, it was only after all the recommended policies of neoclassical and Keynesian economics could no longer be enacted that the economy recovered. So long as those policies were being enacted, the actual economy was suffering, in spite of what some meaningless aggregate statistic or the other could be tortured to tell you.

    What we’re seeing today is similar. The economy is in a tailspin because of all the government intervention and spending. Therefore, the unemployment numbers are naturally worse than Krugman expected. He thought his remedies are a cure, when in fact they are poison.

  2. Gravatar of David Landry David Landry
    9. June 2009 at 14:33


    You say that “In any model with rational expectations a fiscal stimulus should raise current inflation expectations, which should then raise current output.” I’m not sure this is entirely true, or actually it doesn’t take the story far enough. Say that an announced future fiscal stimulus does indeed create inflation expectations. Then surely it follows that short-term nominal interest rate must also be expected to increase as well. Because the structure of a long-term interest rate is a moving average of the expected path of short-term nominal rates, the current long-term interest rate should increase as well – how much depends on risk premia, etc. This could lead to some crowding out via private investment, a phenomenon that I know Krugman thinks is completely ludicrous to believe in the current situation due to the zero lower bound business (at least that’s what I gather). But if he thinks fiscal stimulus can be effective by raising inflation expectations, then the liquidity trap argument no longer applies since overcoming it is all about making credible commitments to inflation, right?

  3. Gravatar of Leigh Caldwell Leigh Caldwell
    9. June 2009 at 15:23

    Scott, Paul Krugman is currently giving a short series of lectures at LSE which are quite interesting… and because he has four and a half hours to fill up, they shed more light on his views than his blog postings.

    The third lecture is tomorrow at 6.30pm London time and you can watch it live on – videos of the first two will be available there too.

    I’ll be at the lecture so if there are any questions you want me to ask, please suggest…

    Incidentally, he spent some time on discussing the setting of future monetary expectations – e.g. try watching from 23:45 or from 50:45 in the video of the first lecture – at the latter point he throws it in as a “something you can do but [the Fed] haven’t tried very much”. So I don’t think you and he are miles apart. He suggested today that one way to influence expectations would be by setting the Fed a 3-4% inflation target but believes that it would be politically infeasible to do so.

    David – he thinks that the current fiscal stimulus is not enough to create inflation expectations (giving evidence from the yields of both TIPS and ordinary 10-year Treasuries). Perhaps a much bigger stimulus would be sufficient, but I don’t think you can use that as an argument against having one. Your logic seems to be: if X will solve the problem, the problem will no longer exist and therefore we don’t need to do X.

  4. Gravatar of Alan Rai Alan Rai
    9. June 2009 at 17:54


    You say that:
    ” there are real constraints on the possible size of fiscal deficits. We are already looking at a massive future tax burden on the economy, which will be a drag on growth for years”

    But isn’t this only a problem if the government has explicitly declared that it is NOT trying to inflate away the deficit i.e. it intends to repay the deficit via future surpluses, not through inflation. I think Krugman is advocating something similar: a bigger stimulus might be needed to convince people that the government has no intention of repaying the deficit in the future, and in doing so, boost inflation expectations.

    Also, your arguments about using monetary stimulus as a way out of the crisis seems to be based on the (implicit) Phillips-curve assumption that higher inflation is necessarily always associated with higher output. But insights from experience as well as from theory (in particular the fiscal theory of the price level) are not so clear-cut.

    Since all analyses of monetary policy operate against a fiscal backdrop, at some point the fiscal constraints of monetary policy must take hold, shouldn’t they?

  5. Gravatar of David Landry David Landry
    9. June 2009 at 22:37


    My logic isn’t necessarily that big fiscal spending won’t work (crowding out might not be 100% due to uncertainty etc. in my description of events) and hence it shouldn’t be done. Rather, my argument is that the crowding out can happen with rational expectations before the money even gets spent. Moreover, I thought that Krugman thinks crowding out can’t happen as long as the liquidity trap is in effect, and that trap won’t end until credible inflation expectations are established. A bit more directly (this isn’t meant to be a formal logical proof of my conclusion):

    (1) Krugman thinks crowding out can’t happen in a liquidity trap as long as inflation expectations are low (i.e. zero or negative, like the real interest rate) and stagnant.
    (2) Therefore, stimulus that is inadequate to raise inflation expectations will not do anything to interest rates, and still might do some good even if ulimately “inadequate”
    (3) But if we want an (adequate) fiscal stimulus that credibly raises inflation expectations, his reasoning that no crowding out will occur no longer applies (since we are out of the trap via inflation expectations) – we need to worry about it

    So I’d say I’m thinking more along the lines of: X will work unless it is offset by B, which can only happen under condition A (X=fiscal spending that raises inflation expectations, B=crowding out, A=not being in a liquidity trap). But A and X cannot coexist (because the liquidity trap doesn’t hold when X raises inflation expectations). Therefore, B will occur and mitigate the positive effects of X. I don’t think this is the same reasoning with which you credited my argument, but maybe I’m just being thickheaded.

    P.S. I’ll be at tonight’s Krugman lecture as well

  6. Gravatar of Leigh Caldwell Leigh Caldwell
    10. June 2009 at 02:55

    That’s clearer and seems a reasonable point. A friend and I will probably retire to the Old White Horse pub after the lecture. Join us there if you want to!

  7. Gravatar of ssumner ssumner
    10. June 2009 at 05:19

    Saifedean, We are closer than you may assume about the New Deal. I have argued that the New Deal was a failure, but precisely because it failed to raise measured output. In the first 4 months when industrial production rose 57% we were much better off, as we don’t have a Soviet-style economy. When in a deep slump, monetary stimulus leads to production of things we value. After July 1933 the New Deal kicked in (the NIRA) and industrial production went nowhere for 2 years. Only when the NIRA was ruled unconstitutional did the recovery resume. But then FDR made other mistakes, and there were further setbacks. So we aren’t far apart, except on the first 4 months.

    David, Nominal rates don’t have to rise when expected inflation rises, real rates might fall (especially when in a liquidity trap.) And even if nominal rates do rise because of expected inflation, there is no rise in real rates, and hence no crowding out. But these inflation expectations will also raise other assets prices (stocks, commodities, buildings, etc) which raises current output if wages are sticky.

    Krugman does believe that a liquidity trap can be overcome if we can create inflation expectations, and has said so many times.

    Leigh, Ask why the Fed doesn’t set a 6 or 7% NGDP growth target for this year, and 4-5% thereafter. Because the SRAS is fairly flat in a deep slump, that would keep inflation expectations well-anchored.

    Politically it’s an easier sell than inflation, as in Puritan America inflation sounds like a too good to be true cop-out. Something we’d expect from a banana republic. But NGDP includes real output, and sounds more respectable.

    [I know Krugman will understand the “Puritan” point, as he’s made similar arguments. I don’t know if you are allowed to read a question as long as the 5 sentences above, but if you could read it to him I would be eternally grateful.]

    The 3-4% number suggests he’s backing off a bit–I wonder if he has heard about my complaints about his “high inflation” argument. I had thought he was originally in the 5-6% range.

    Alan Rai, Bernanke and other Fed officials made it very clear they don’t intend to inflate away the debt—and I believe them. If that was their strategy, they would be adopting much more inflationary policies right now, as it’s much easy to prevent debt from forming, than to inflate it away.
    I am just talking about numbers like 2% inflation and/or 5% NGDP growth. And you don’t need budget deficits to create that sort of mild inflation, the Fed can do it all by themselves.
    I don’t assume inflation is necessarily associated with higher output, that’s why I prefer the 5% NGDP target. If it’s all output, so much the better. I do assume that higher AD raises inflation expectations, other things equal. And I think all schools of thought agree on that point.
    I do understand the fiscal view, but I think in the US they have the tail wagging the dog. The Fed makes it decision, and Congress salutes and says “yes sir”. Congress must balance its budgets in the long run, the Fed won’t bail them out.

  8. Gravatar of David Pearson David Pearson
    10. June 2009 at 05:50


    You claim monetary stimulus will result in lower fiscal deficits, but do not provide a rationale.

    Maybe you are implying that monetary stimulus raises real output. However, elsewhere (including above) you refrain from making this claim.

    So maybe you are implying that the real cost of deficit financing will decline as inflation rises. Let’s leave aside the question of bracket creep and focus on real rates. Will 2% inflation, which is what you want, raise or lower real rates across the yield curve? You seem to imply lower. A case can be made for (much) higher based on the tail risk of an inflationary overshoot. Surely, there are dozens of examples where an inflationary overshoot stemming from well-intentioned monetary policy results in escalating real debt costs.

    It seems to me that in the end you assume:

    1) the output gap eliminates inflation tail risk
    2) inflation raises real output

    Both are truly Keynesian viewpoints.

  9. Gravatar of Bill Woolsey Bill Woolsey
    10. June 2009 at 08:57

    Scott, so monetary policy can only impact aggregate demand if it creates expectations of inflation?

    If the Fed foolishing was committed to stable prices at all times, and nominal expenditure fell because of an increase in the demand for money/lower velocity, then your position is that no increase in the money supply, no matter how large, could return nominal income to target unless it caused people to expect that the Fed would break its committment to a stable price level and cause inflation.

    Is that it?

    Or, is your position that only after there had been deflation, could the Fed expand the money supply enough to create the expecation that it would reflate back to the target, and that would raise aggregate demand back to the target. Of course, they would no longer be creating an expecation of inflation after the target for the level was hit.

    I think there is a contradiction in this approach.

    Perhaps this support for 2% inflation always is causing the problem. You are always thinking about a target that does include low inflation.

  10. Gravatar of Leigh Caldwell Leigh Caldwell
    10. June 2009 at 12:08

    Scott – he didn’t pick me so no question asked. Never mind, I’m sure you’ll get him to comment somehow eventually. I have an interview request in with his press office but I suspect his time in London is too short to fit me in.

  11. Gravatar of ssumner ssumner
    11. June 2009 at 04:59

    David, Milton Friedman would be rolling over in his grave if he heard you say that inflation only affects output in the Keynesian model. Let’s be clear that inflation doesn’t affect output in the long run in any model. But it does affect output in the short run in almost all models (Keynesian, Austrian, Monetarist) Friedman argued that deflationary policies caused the Great Depression.

    Virtually all economists agree that recessions automatically make the budget deficit bigger, and there is overwhelming empirical evidence to support that notion. If you want to get into inflation and interest on the debt, I would point out that when inflation is lower than expected, the real burden on borrowers with nominal debts rises. And the US government is a net debtor with lots of nominal debts. So going from 0% to 2% inflation will definitely reduce the real burden of that debt, even if, as you say, all new debt incorporates an inflation premium. The bottom line is that more NGDP helps in two ways, more real growth in the short run (boosts tax revenues), and inflation reducing real interest payments on exiting debt. And I haven’t even mentioned that monetary stimulus would reduce the need for fiscal stimulus, which is another huge burden.

    And just so people don’t accuse me of trying to steal from long term T-bond holders through inflation, let me point out that they bought most of those bonds on expectation of 2-3% inflation, and we have just given them a huge windfall.

    Bill, I should have been more specific. Any increase in AD causes inflation to be higher than it would otherwise be. You are quite right that if you had, say, a 3% NGDP target, than using monetary policy to boost NGDP growth from 0% to 3% wouldn’t create inflation, rather it would make for less deflation, as you moved toward price stability.

    Leigh, Thanks for trying. I’ll try to catch the talks on video.

  12. Gravatar of JimP JimP
    13. June 2009 at 08:04

    Why and how Roosevelt was better than Obama now is – and why the WSJ is entirely insane.

    h/t henrey blodget

  13. Gravatar of ssumner ssumner
    14. June 2009 at 05:40

    JimP, Thanks, That may be worth a separate post.

  14. Gravatar of Alex Alex
    14. June 2009 at 06:21


    Have you read this one from Krugman?

    “Let me add, for the 1.6 trillionth time, we are in a liquidity trap. And in such circumstances a rise in the monetary base does not lead to inflation. I had a couple of charts in my lectures this past week.”

    I agree that the increaase in B does not need to generate inflation, but how can he stick to the liquidity trap claim after everything we have seen in the bond markets lately? Maybe the explanation is his next post…

  15. Gravatar of JimP JimP
    14. June 2009 at 07:53


    I agree. It was real interesting. I especially liked the stuff about expectations. Clearly re is not news to everyone.

  16. Gravatar of JimP JimP
    14. June 2009 at 08:02

    I sure do think it is worth a post. This is the early Roosevelt
    – about whom little is known or appreciated. We need to the the later Roosevelt spending policies stopped. This is the Roosevelt everyone knows about – unfortunately including Obama.

  17. Gravatar of JimP JimP
    14. June 2009 at 08:09

    sp on henry

    and “the” should be “get” in the last post.

    I should proof them.

    But I do – sadly.

  18. Gravatar of ssumner ssumner
    15. June 2009 at 05:07

    Alex, I added a post that touches on the issue you raised, and have still another Krugman post planned.

    JimP, Yes, the early FDR was best.

Leave a Reply