Paul Krugman on fiscal and monetary policy

Here’s a puzzling passage from a recent Paul Krugman post:

One question I’ve been asked a lot is why I spent 2009 campaigning for fiscal expansion rather than monetary expansion. Well, at the Keynes conference this morning Mike Woodford gave an overview of policy options when you’re up against the zero lower bound that in some ways expressed better than I’ve managed to what I was thinking at the time.

First, Mike argued that monetary expansion once you’re at the ZLB mainly works, if it does, through affecting expectations. If people don’t perceive the expansion as representing a change in policy that will persist even after the economy has recovered, even big changes in the monetary base have hardly any effect. Mike had a chart of Japan 2000-2008 that I’ve crudely reproduced:

Note both that Japan reversed much of the initial expansion in the monetary base, confirming the expectations of those who might have regarded that expansion as temporary – and Japan did this even though deflation continued! Note also that nominal GDP never moved at all despite the huge amount of money “printed”.

Krugman and Woodford are both extremely smart guys, so I can’t understand why they continue to misinterpret this graph.  The Bank of Japan clearly wants to avoid inflation.  That’s why they sharply tightened monetary policy in 2006 “even though deflation continued!”  Indeed that’s the only plausible explanation for this tightening.  But Krugman and Woodford continue to insist that this shows the impotence of monetary stimulus, i.e. the difficulty of changing expectations.  They continue to assume that the BOJ was trying to inflate, but failed.

In fact, this shows that expectations in Japan were rational.  The public correctly predicted what the BOJ would do.  The problem isn’t convincing that public that the BOJ wants to inflate; the real problem is convincing the BOJ that they should try to inflate.  Krugman and Woodford are confusing cause and effect.  They see low interest rates as inhibiting monetary stimulus, whereas the low interest rates are actually a consequence of tight money, of central banks aiming too low.  It’s not that the BOJ and Fed don’t know how to raise NGDP growth, it’s that they don’t want to.  Don’t believe me?  Then read any recent Fed statement explaining why they decided against continuing QE2.

In 2009 Krugman was too pessimistic about monetary stimulus, and thus focused on fiscal stimulus:

That’s about what I was thinking in, say, January 2009. With the severe financial crisis still relatively recent, and many people still expecting a V-shaped recovery, it didn’t seem possible to persuade the Fed to commit to a permanent rise in the monetary base or a rise in the medium-run inflation target, nor did it seem possible to convince markets that there had been a long-run change in policy. The chances for persuading Congress to agree to a large temporary fiscal stimulus seemed much better.

Back in 1998 Krugman knew better than to rely on fiscal stimulus.  From It’s Baaack:

If temporary fiscal stimulus does not jolt the economy out of its doldrums on a sustained basis, however, then a recovery strategy based on fiscal expansion would have to continue the stimulus over an extended period of time. The question then becomes how much stimulus is needed, for how long – and whether the consequences of that stimulus for government debt are acceptable.

.  .  .

The political point is that Japan – like, we might note, the United States during the New Deal – appears to have great difficulty working up its political nerve for a fiscal package anywhere close to what would be required to close the output gap. Exactly why is an interesting question, beyond this paper’s scope.

Does this mean that fiscal policy should be ignored as part of the policy mix? Surely not. On the general Brainard principle – when uncertain about the right model, throw a bit of everything at the problem – one would want to apply fiscal stimulus. (Even I wouldn’t trust myself enough to go for a purely “Krugman” solution). However, it seems unlikely that a mainly fiscal solution will be enough.

That was quite prophetic.  Fiscal stimulus in the US was woefully inadequate.  The US political establishment did find the consequences for government debt to be unacceptable.  And I love the quotation about a “purely ‘Krugman’ solution.”  That would be one that relied solely on monetary stimulus.  By August 2010 Paul Krugman had attached a different name to the money-only approach:

And I don’t understand at all [Koo’s] argument that monetary expansion is positively harmful. He seems to be making up arguments on the fly here; he’s so determined to defend the primacy of fiscal policy that he has to insist that anything else is a very bad thing. (In that sense, I guess, he’s the anti-Scott Sumner).

So now it’s the “purely Sumner solution.”

In the 1990s Krugman assumed that monetary policy steered the nominal economy, and criticized protectionists who thought Asian exporters stole jobs from the US.  He had lots of great arguments.  If he’s done with them, I’m happy to use his discarded ideas.



13 Responses to “Paul Krugman on fiscal and monetary policy”

  1. Gravatar of StatsGuy StatsGuy
    6. July 2011 at 12:27

    First, Krugman is now endorsing the notion that monetary expansion needs to be considered permanent to work – at least that’s cleared up. (aka, CBs really have to mean it, and sometimes they need to prove they mean it)

    However, there are some still unexplained concerns in Japan. It’s possible that even if monetary expansion was intended to be permanent, the markets anticipated these intentions would change when Japan encountered the consequences (for example, banks faced with insolvency as their nominal borrowing costs increased but their long lending returns did not). “Incentives” can be structural, and even cultural – it’s quite possible that threat of election impact is constraining monetary activity right now (rightly or wrongly – you could say this is because those in power misunderstand the means/ends chain, but it’s obviously having an impact on the debate).

    The only way to prove you mean it, in this case, is to create institutional barriers to undoing the policy. Doing this, however, limits the CBs control over inflation, which the CB does not want. This is one reason CBs like monetary expansion via private debt rather than printing currency – because they can reign it in more easily by raising interest rates. Extracting currency out of the economy is a bit more difficult and time consuming than selling digital assets off the virtual balance sheet.

  2. Gravatar of marcus nunes marcus nunes
    6. July 2011 at 13:41

    It will remain a great mystery why he changed his views so drastically (and is battling to reverse them again). Being the “public figure” he is, he did a lot of damage by going “all in” with fiscal stimulus.

  3. Gravatar of Jeremy Goodridge Jeremy Goodridge
    6. July 2011 at 15:07

    You say:

    “It’s not that the BOJ and Fed don’t know how to raise NGDP growth, it’s that they don’t want to. Don’t believe me? Then read any recent Fed statement explaining why they decided against continuing QE2.”

    I would say the BOJ and the Fed are too AFRAID to try to raise NGDP growth. They fear that they will overshoot and will create excessive inflation. But if they KNEW they wouldn’t overshoot, then I would expect them to push for more NGDP. So, in a way, they feel the risk igniting too inflation is greater/worse than the risk of having too little ngdp.

  4. Gravatar of Wimivo Wimivo
    6. July 2011 at 15:51

    Jeremy Goodridge,

    The BOJ isn’t afraid of “excessive inflation”, they seem to be afraid of *any* inflation whatsoever. That’s why they appear to tighten money at such seemingly inappropriate times.

  5. Gravatar of Benjamin Cole Benjamin Cole
    6. July 2011 at 16:01

    The incredible damage that is done by this ongoing misunderstanding of monetary policy by left-wingers and right-winger is…well…enervating.

    Right-wingers think they are campaigning against loose money and morals, and profligate liberal spending programs, validated by greasy greenbacks.

    Left-wingers think you have to spend federal money, as no one else is.

    Koo is koo-koo. Krugman is off base.

    Please, policy-makers, pay attention to Milton Friedman, the John Taylor of 2006, and Scott sumner.

    And pay attention to Japan, if you think tight money and fighting inflation is so nice a thing. 20 years of tight money has been but a noose around the Nipponese economy. Like applying leechs and bloodletting to a guy with anemia.

  6. Gravatar of Lorenzo from Oz Lorenzo from Oz
    6. July 2011 at 20:22

    This is what Koo was arguing in March 2006. But if, as he claims, corporate Japan had repaired its balance sheets by then, surely it is now way past time for the BoJ to shift to money expansion: to make money no longer a perfect substitute with bonds because there is a rational expectation that money will lose value if held on to.

  7. Gravatar of Lorenzo from Oz Lorenzo from Oz
    6. July 2011 at 21:23

    I have become interested in chartalism, because I am interested in the history of money and coins (the unification of medium of exchange with the unit of account) were clearly a creation of rulers. (So the Austrian just-so story of a medium of exchange becoming preferred by a process of exchange evolution then becoming money does not work for me.) I am struck, however, by how unspoken assumptions, which rather more history would undermine, get smuggled into theory. Such as the apparent identification of money with (1) financial assets and (2) use for tax in “modern monetary theory” (MMT), as chartalism has apparently become.

    Money is only a financial asset if you can lend it out at interest: we presume we can in our society but there were long periods where that was difficult or impossible. So it is not a defining feature of money that it be a financial asset, as MMT implies.

    Also, under the Third Reich, the regime actually paid for some purchases in tax certificates precisely because of its shortage of money, showing up how money is not defined (or even necessarily created) by use for tax liability. Coins were a great way for rulers to collect taxes and pay for purchases precisely because they had wider uses than paying for taxes.

    The apparently “obvious” notion of money and bonds being perfect substitutes in a liquidity trap, making monetary policy ineffective, did not make sense to me because I did not get that money then becomes as good a store of value as bonds. So you may as well hang onto it. More M just means more k so no change in Py (NGDP).

    But I suspect part of my problem in grasping what was involved was that I did not see how it became impossible to convince folk that money was going to become less reliable-as-bonds as a store of value. Too much a child of inflationary times and aware of the history of Great Inflations and debasements I suspect.

    I wonder if, with enough fiscal stimulus, you can actually make money seem more reliable as a store of value than bonds! Could that make your fiscal stimulus contractionary, given expectations about future tax liability? After all, there are arguments over (pdf) whether fiscal stimulus can also end up, in effect, “pushing at a string”.

  8. Gravatar of Rien Huizer Rien Huizer
    7. July 2011 at 04:35


    You want central banks to credibly commit to policies that carry the risk of hard to control and too high inflation. Only then everyone will believe that targeting some positive level of inflation will occur fairly late, or not at all, given the natural lags in inflation targeting policy. And even then, there may be doubters.

    That still ties in with the mechanics of targeting future levels of NDGP. We have a substantial body of knowledge about the technology of rate of price inflation targeting (separate from the desirability of that target) and in many countries this is something politicians and central bankers have been able to live with, especially since it gives politicians a respectable fall guy, the virtuously independent central banker-technocrat armed with his robotic rules (who wants democracy when it is hard?).

    I keep wondering about what technology (and associated politics plus risk management) you have in mind for NDGP targeting (and I would not be following this blog if I did not agree that in an ideal world we should have a machine keeping NDGP from being too volatile and preferably, growing smoothly). A market in NDGP futures, like any futures market, requires market participants (speculators or hedgers as they used to be called). Apart from the FED subsidizing the market, how are these market participants going to make money? Or what would they be hedging? By guessing the realized NDGP level not being where it should be? By guessing FED policy? By guessing the prices of non-NDGP futures instruments influenced by FED policy? etc. Maybe a dumb question, but I have not been able to find an answer on your site. Have the exchanges ever shown interest in such a contract?

  9. Gravatar of Scott Sumner Scott Sumner
    7. July 2011 at 07:46

    Statsguy, Krugman’s held that view of permanent monetary injections since 1998–so he’s been consistent.

    Your hypothetical for Japan is possible, but it didn’t happen there, nor has it happened anywhere else in all of world history. The BOJ tightened in 2000 and 2006 when there was no inflation. That’s all that needs to be said. They showed their true colors, and the Japanese public had an accurate understanding of their ultra-conservatism. The problem isn’t credibility, it’s ultra-conservatism.

    Marcus, Perhaps there were two factors:

    1. Krugman’s realization that central banks were too conservative to do the right thing. (On this point he was ahead of me.)

    2. His growing support for big government, combined with a perhaps naive belief that fiscal authorities were not too conservative to act decisively.

    Jeremy, That may be part of the problem, but in a time when TIPS spreads are observable minute by minute, it’s a really bad reason for inaction.

    Wimivo, Yes, for Japan you are certainly right–their GDP deflator is down about 15%. There’s no danger of it rising back to 1993 levels. Jeremy’s point might be valid for the US.

    Benjamin, Good points.

    Lorenzo, I don’t like the arguments about money and bonds being perfect substitutes at zero rates. There are two possibilities. Either zero rates are expected to last forever, in which case bonds ARE money. Or zero rates are not expected to last, in which case swapping money for bonds will raise the future expected price level and reduce current real interest rates.

    I agree with your point on Japan.

    Rien, You said;

    “You want central banks to credibly commit to policies that carry the risk of hard to control and too high inflation.”

    No I don’t. I favor targeting the forecast. I favor having a much smaller monetary base in the US (which would have occurred if NGDP expectations were on target in 2008-09.)

    You asked:

    “Have the exchanges ever shown interest in such a contract?”

    Thank God they haven’t. If they had shown an interest, then NGDP hedging might distort the market price. Because there is little interest, the market will be dominated by pure speculators, which is what you want.

    If there is little trading in the market, that’s fine. It means the Fed is doing a good job. Imagine a novice steering a boat. Then imagine a skilled sailor standing right behind the novice, pushing on the tiller once and a while if the ship gets off course. If the sailor never once has to nudge the tiller, that’s fine. It means the policy is working. The Fed is the novice at the tiller, and the NGDP futures market is the experienced sailor.

    In October 2008 I would have been selling NGDP futures like crazy. But I also think it’s hard to get rich. Hence others would have also sold NGDP futures, pushing NGDP expectations back on target. If they didn’t, then I’d get very rich. Either way I like NGDP targeting–either it works, or I get very rich.

  10. Gravatar of Rien Huizer Rien Huizer
    7. July 2011 at 17:24


    I cannot resist the temptation to associate your market with that for sports bets (which are good places for people with the right business model, maybe too good). Sports betting is not even supposed to influence the outcome of games, but it may influence behaviors. NDGP targeting using futures contracts, in which experienced market operators have no interest, should do little else than influence outcomes (and make the lucky ones rich). How? Betting markets need lack of precise predictability to function and in your ideal world, a very large elephant will teach that market that what it predicts, happens.. I am getting more and more curious about this pse help!

  11. Gravatar of Krugman and Sumner on the Zero-Interest Lower Bound: Some History of Thought « Uneasy Money Krugman and Sumner on the Zero-Interest Lower Bound: Some History of Thought « Uneasy Money
    8. July 2011 at 08:58

    […] discussion a while longer.  As a diversion, I will follow up on a point that ScottSumner made in discussing Paul Krugman’s reasoning for having favored fiscal policy over monetary policy to lead us out of […]

  12. Gravatar of Scott Sumner Scott Sumner
    8. July 2011 at 17:26

    Rien, It would be no more likely to happen than in any other market. I’m afraid you still don’t understand my proposal. Think of it as an implied forecast of velocity (actually future NGDP over current M). Just because we set this market up, doesn’t make velocity easier to forecast.

  13. Gravatar of What I’ve Been Reading: “A-Long-Time-Ago”-Edition « Seeing Beyond the Absurd What I’ve Been Reading: “A-Long-Time-Ago”-Edition « Seeing Beyond the Absurd
    15. July 2011 at 12:35

    […] Japan never tried monetary expansion, but somehow people think that. As always Scott Summner does have a good point: “The problem […]

Leave a Reply