Krugman Now Disagrees with His Earlier Critique of MMT

By Scott Fullwiler

In a post yesterday, Paul Krugman notes the CBOs long-term projections for federal government deficits and the national debt now show a reduced projection of nominal interest rates:

This markdown has the effect of making the budget outlook — which was already a lot less dire than conventional wisdom has it — look even less dire.

After a bit of discussion of debt-interest rate dynamics—which I earlier discussed in detail here and in my series here (printable version here)—Krugman explains the importance of understanding currency issuers like the US versus currency users like the Eurozone nations for understanding these dynamics:

Now, wait a second, you may say: higher debt will mean higher borrowing rates, because people will fear that we’re about to turn into Greece, Greece I tell you.

And then he hits the key point:

As many of us pointed out, however, such results were driven almost entirely by the euro crisis; high-debt countries that borrow in their own currencies haven’t seemed to face anything like the same costs. And a funny thing has happened to the euro area itself since the ECB started doing its job as lender of last resort: not only have rates come down, but the relationship between debt and borrowing costs has become much flatter.

Interesting that he says “as many of US pointed out,” since back in July 2011 Krugman openly wondered why interest rates on government debt in Italy were rising while remaining very low in Japan”

A question (to which I don’t have the full answer): why are the interest rates on Italian and Japanese debt so different? As of right now, 10-year Japanese bonds are yielding 1.09%; 10-year Italian bonds 5.76%.

I ask this because in a number of ways the two countries look similar. Both have high debt levels, although Japan’s is higher. Both have awful demography. In other respects, the numbers if anything favor Italy, which has a much smaller current deficit as a percentage of GDP.

So what’s going on? I normally argue that members of the euro zone that have excessive costs — which certainly includes Italy — face a straightjacket in the sense that they will be forced to go through a period of grinding deflation to restore competitiveness. But while Japan has its own currency, it’s suffering from its own deflation all the same.

What is true is that the Bank of Japan is keeping rates at zero, while the European Central Bank seems determined to raise rates. Is that enough to explain the difference? Or is it something about the absence of a proper lender-of-last-resort function?

Or, finally, do Japanese politics — for all their disappointments — just look more mature than those of Italy?

I actually don’t have a firm view. But it seems to be an important puzzle to resolve.

In fairness, Krugman did allude to the lender of last resort function of the ECB in both posts to keep interest rates on the national debt down (more on that below).  However, unlike 2011, he does now recognize that being a currency issuer is the key distinction (and the ECB is the currency issuer in the Eurozone system, which explains the importance of its lender of last resort function in keeping rates low for Eurozone countries that otherwise would have to deal with “bond vigilantes”).

A few months earlier in March 2011, Krugman made the same points as he would later in July, this time as the central part of his critique of MMT (here and here):

So suppose that we eventually go back to a situation in which interest rates are positive, so that monetary base and T-bills are once again imperfect substitutes; also, we’re close enough to full employment that rapid economic expansion will once again lead to inflation. The last time we were in that situation, the monetary base was around $800 billion.

Suppose, now, that we were to find ourselves back in that situation with the government still running deficits of more than $1 trillion a year, say around $100 billion a month. And now suppose that for whatever reason, we’re suddenly faced with a strike of bond buyers — nobody is willing to buy U.S. debt except at exorbitant rates.

So then what? The Fed could directly finance the government by buying debt, or it could launder the process by having banks buy debt and then sell that debt via open-market operations; either way, the government would in effect be financing itself through creation of base money. So?

Well, the first month’s financing would increase the monetary base by around 12 percent. And in my hypothesized normal environment, you’d expect the overall price level to rise (with some lag, but that’s not crucial) roughly in proportion to the increase in monetary base. And rising prices would, to a first approximation, raise the deficit in proportion.

So we’re talking about a monetary base that rises 12 percent a month, or about 400 percent a year.

Does this mean 400 percent inflation? No, it means more — because people would find ways to avoid holding green pieces of paper, raising prices still further.

I won’t go through all the problems with Krugman’s argument—I did that in a lengthy response to Krugman (printable version here) a few days later—except to get at the issue at hand here.  To do that, let’s isolate the key sentence in Krugman’s critique:  “And now suppose that for whatever reason, we’re suddenly faced with a strike of bond holders—nobody is willing to buy U.S. debt except at exorbitant rates.”  Note how this is a 180-degree turn from what he wrote yesterday.

In my 2011 response, I wrote in regard to this particular sentence:

Krugman tries to push this heroic assumption [i.e., bond vigilantes] without anyone noticing by inserting “for whatever reason” into the sentence, as if somehow he can just wave his hands and make it so. Indeed, “whatever reason” is the crux of the matter, and in this two word phrase he has assumed MMT’s vast literature on the monetary system away in order to make his argument against MMT. It’s like proving theory X is wrong by simply assuming all of the supporting evidence for theory X—for “whatever reason”—is wrong.

A bit later, I explained the MMT position from a more general perspective:

A currency issuer under flexible exchange rates that allows itself to receive overdrafts in its central bank account will see the interest rate on its debt equal to the central bank’s target rate at the very lowest.  This is because the central bank cannot achieve its target rat in this case unless it pays interest on the reserve balances created by the government’s spending net of balances drained by taxes, and these central bank outlays will reduce the profits it turns over to the treasury (a de facto interest payment by the treasury).  This is what I like to refer to as the strong form of MMT regarding interest on the national debt.  If the treasury instead decides to issue short-term bills, or is required to by self-imposed constraints, these will arbitrage against the central bank’s target rate; if it issues longer-term bonds, these will mostly arbitrage against the current and expected Fed targets.  I call this the semi-strong form, and explained it in more detail here and here.  Randy Wray does, too, here.  Losing access in the semi-strong form is a non-starter—the arbitrage opportunity grows stronger as the non-govt sector can borrow at a lower rate, and there are primary dealers and thousands of hedge funds that would love to take advantage of that trade. 

In neither of these cases should interest rates on the national debt be considered to be set by “market forces” (aside from what Warren Mosler likes to call “technicals”).  Further, the self-imposed political constraint is not a constraint of any economic significance—if one is given the choice between an overdraft at the central bank’s target and issuing debt at roughly the central bank’s target, does it really matter if the overdraft option is then withdrawn?  MMT’ers say “no.”

Failing the strong and semi-strong forms of interest on the national debt—which I would argue would be exceedingly rare, though the probability is probably not 0—a final option would be for the central bank to purchase the government’s debt in order to keep interest rates on the debt from rising.  I call this the weak from of MMT regarding interest rates on the national debt.  Marshall Auerback and Rob Parenteau explained this option in more detail here.  Here again, access to the bond markets isn’t the issue.

The overall point here is that the interest rate on the national debt for a currency issuing government under flexible exchange rates always is, or at the very worst always can be, a monetary policy variable.  Concerns about “bond market vigilantes” are misplaced, as they could apply only to non-currency issuers (e.g., Greece, California) or fixed exchange rate regimes.

Obviously Krugman’s application of the term “lender of last resort” to the ECBs bond purchases to keep bond vigilantes at bay is essentially the same as the “weak form” of MMT.  And, again, to his credit, he did see this as an option for policymakers, albeit one that would be inherently inflationary aside from a ZIRP world (read my response for a critique of that; I also critiqued this view in more detail here).

At any rate, the key point is that (a) Krugman 2011 did not recognize the importance of a currency issuer vs. a currency user in setting interest rates on the national debt, and based his argument against MMT on the potential for bond vigilantes to raise US Treasury rates “for whatever reason,” but (b) Krugman 2014 does recognize this distinction and now appears to view its significance for interest rates on the national debt in much the same way as MMT.

It seems that Krugman’s current view started somewhere in late 2012, such as in posts here and here discussing “invisible bond vigilantes.”  Again, it’s quite clear from that post that there’s been a significant change from Krugman’s 2011 view of bond vigilantes’ powers relative to a currency issuer and his related critique of MMT.

In closing, while there are still significant differences between MMT and Krugman’s neo-liberal-based views of the monetary system—see my links above as well as MMT critiques of his views on banking by myself here and by Randy Wray here—there have been important areas of continuing convergence regarding interest on the national debt and also (interestingly) the government’s consolidated financial statements (see this post from Stephanie Kelton and me here and the NEP version here).  This is a positive development.

30 responses to “Krugman Now Disagrees with His Earlier Critique of MMT

  1. golfer1john

    “This is a positive development.”

    Indeed. Perhaps, when he thinks through the implications, he may even have to abandon his deficit dove stance of “austerity later”.

  2. Jamie Walton

    Good post Scott. Educating Krugman appears to be a slow process, but it’s probably important, so thanks for your good work.

  3. Scott Fullwiler

    Perhaps of interest to readers–Ed Harrison just tweeted the link to this piece he wrote in 2011 just after Krugman wrote his post on Italy/Japan. It’s excellent.

  4. I keep running into all these articles that quote money managers and financial analysts repeatedly and vehemently declaring that the Fed’s interest rate MUST rise. And yet, from the Fed chair’s comments and a glance a credit market graphs there seems little that imply that a rate rise would be anything but a sledgehammer applied broadly to all interest rates when currently credit quality concerns exist in a very limited number of sectors. The money managers are not stupid people. I wonder if their call for higher rates has a lot more to do with a desire to scoop up the bargains, and thereby meet their yearly return targets and collect bonuses, in a market correction caused by a Fed move than by any fundamental belief that a broad increase in rates is called for under current economic conditions.

    • golfer1john

      I think the simple explanation is that they expect the past to repeat itself, as it has always done. The Fed never stands still for very long, and from here there is only one way to go. The only question is when. There has been no indication from the Fed that any of them think a permanent 0% interest rate (or any permanent rate) is a valid policy. Which makes sense, because then they would all be out of a job.

  5. Ryan Bradley

    I have a feeling Mr. Krugman came around to MMT long ago. It hit him like a ton of bricks, just like me, just like many of us. Things were not happening as he was taught they would. And one day he read something from Mr. Fullwiler, or Mr. Mosler critiquing him and WHAM! All the pieces snapped into place and the clouds parted. He could see.
    But, he has years of IS-LM behind him. One cannot reverse course so quickly. So, he trickles it in here and there so when he writes the book about how economics really works (explaining MMT to the masses) he can point back to how he knew it all along. NY Times Best Seller and on all the talk shows. I Hope I see the day. Then I can say “I told you so.” To all my friends and family. 🙂

  6. Krugman is “evolving” on MMT the same way Obama “evolved” on Gay Marriage. This is critical since he must continually battle the right as long as he continues to occupy the role of most mainstream liberal economist. Since MMT is way far out of mainstream consciousness he would be outflanked from the right if he adopts it explicitly.

  7. I often refer to this blog post

    It refers to a Krugman conference paper where he segregates countries into issuer and user and shows that there is a marked difference between the two.

    It dates from Nov 2013.

    • golfer1john

      Kudos to Krugman. Yes, it’s a simple and obvious thing to do, but he’s the only one to do it. If only he would say it on TV rather than in an obscure academic paper.

    • Neil,

      that papersseems to indicate that Krugman agrees with MMT when we’re “at the zero lower bound” or in “a liquidity trap”, but not otherwise.

      • Scott Fullwiler

        actually, his model assumes that the interest on the national debt is determined by the CB whether in a liquidity trap or not. His point is that by following a Taylor Rule outside of the zero bound, a fall in the exchange rate could lead the central bank to push up interest rates to stop the economy from growing too fast (since the fall in the exchange rate stimulates exports). Though I’m not a Taylor Rule fan, if a CB is following a Taylor Rule, then I’ve got no problem with his analysis there (though his use of and explanation of “liquidity trap” is like fingernails on a chalkboard for me). The point is interest rates are rising because the CB is trying to slow the economy, not because of fixed exchange rates or bond vigilantes. That’s all consistent with MMT.

    • Neil,

      In that paper he keeps pushing the lie that it was Paul de Grauwe who noted the sovereign-nonsovereign distinction. MMT related folks had done it years earlier. All insights have to come from the neoclassical clique in his world.

  8. On some level question is not what government CAN do, but what it SHOULD do. If Krugman fears that reserve buildup could lead to unsustainable explosion in lending one can easily imagine a world where government regulators set limit how much banks can expand their balance sheets. Say, limit new loans to 5% of the total per year and problems is solved. If you weight that policy option against one that produces dysfunctional economy and wastes trillions on potential output, Krugmans position as a deficit dove becomes indefensible. How can you support austerity that has terrifying consequences for the nations well-being instead of some minor technical tweak in the regulations that 99.9% of the people would never even notice?

    I would love to see deficit doves answer that.

  9. Rob Rawlings

    I have just read this post and some of the links that are provided and found it very interesting. I have a general question on MMT that is probably both off-topic and a bit basic (I’ve never studied MMT till now), so hope its OK to ask here.

    I gleaned that one of the main policy aims of MMT is to use the budget deficit to maintain AD at its optimum (maximum, non-inflationary) level and to use monetary policy (CB target rate , interest on reserves etc) to control interest rates.

    My question is this: Won’t there be multiple combinations of fiscal and monetary policy that could deliver optimum AD ? For example you could have a smaller deficit and lower interest rates (outside of ZLB anyway), or a higher deficit and higher interest rates and achieve the same level of AD. Is my understanding of this correct, and if so – how do you decide what the optimum budget deficit is ?

    • The general feeling amongst MMTers is that monetary policy is uncertain in its distribution and response profile. So you don’t use it at all. You just ‘park it’ where it needs to be to ensure there is sufficient capital development in the economy (probably at zero).

      Then you use fiscal policy to manage the economy. Strong auto-stablisers, like the Job Guarantee, to counter balance the business cycle and maintain the economy on ‘automatic pilot’. Then the government sets discretionary fiscal policy at a level that keeps everybody busy and the Job Guarantee pool numbers low at the height of the business cycle.

      You don’t drive a car by throwing your weight around and hoping you can ‘nudge’ the wheels in the right direction. You get hold of the steering wheel!

    • Rob,

      My layman answer would be that of course there would be multiple combinations.

      1) It is hard to say which monetary policy is expansionary and which is contractionary. If debt/GDP is high there is so much income derived from interest that high interest rates may be expansionary
      2) It is a policy choice. Whatever the level of debt/GDP the govt can set rates to 1.3% or to 0% and that is that, so it is a policy choice, and yes, since a lot of combinations are possible there may be disagreements on what is optimal and what not.

  10. I think Krugman visits this website. If so, that’s awesome. He certainly has visited Multiplier Effect, as he once linked to a post by Michael Stephens.

    Paul, if you’re reading this, please advocate a huge payroll tax holiday! Thanks!

  11. Scott Fullwiler,

    I think the major disagreement between MMT and Krugman is on monetary policy.

    MMT argues that you can permanently keep the CB overnight interest rate at or near zero, and so you can run whatever size deficits are required with no problem.

    Mainstream economists like Krugman think that permanently holding the interest rate at or near zero would in all likelihood result in hyperinflation. Krugman thinks that at some point the interest rate will have to rise, and this fact imposes a limit on the budget deficits that you can run over time. This is all tied up with the natural rate of interest theory.

    As far as I’m aware the mainstream also holds that the interest rate can not permanently be below the growth rate of the economy without resulting in accelerating inflation.

    Until MMTers can convince people like Krugman that these sorts of beliefs are mistaken, they will continue to disagree, except in exceptional circumstances (when we are “at the ZLB” or “in a liquidity trap”).

    • My understanding is that MMT argues that a monetarily sovereign government can run whatever size deficits are required with no problem. If interest rates rise, the government’s interest payments will rise as high as they need to rise.

  12. Pingback: Krugman Gives DeGrauwe 2011 Credit for What MMT Has Argued for 15+ Years | New Economic PerspectivesNew Economic Perspectives

  13. Scott Fullwiler

    On MMT and monetary policy:

    As Neil states, we view monetary policy’s effects as being uncertain, particularly if the CB is uisng the short-term rate only. First, the effect of lowering/raising rates depends on the distriubtion of savers/borrowers among the pvt sector and across pvt sector vs. govt. Second, interest rates are a cost of working capital and directly affect prices. Third, interest rates don’t matter as much in the case of financial innovations, speculative bubbles, etc., unless you change them a LOT (we had a bubble in with high rates in the 80s and another with relatively low rates in the 00s). Fourth, the pass through from the short-term rate to the rates the pvt sector pays is uncertain. Fifth, changes in the short-term rate can be highly destabilizing, particularly as the economy pushes into more financially fragile states. Sixth, whether or not rate changes cause people to spend more/less, can depend a lot on their pre-existing debt and debt service levels. There are more, but those are good for starters.

    So, use monetary policy or some similar authority to regulate the financial system countercyciclically and over the longer-term, too, to avoid as much as possible (it’s impossible to eliminate) moves to fragile macro financial states (lots of ways to do this–credit controls, adjust directly interest rates private sector pays rather than fed funds rate, forbid certain types of behaviors, require banks to keep loans they make rather than selling, etc., etc.). Use fiscal policy to adjust spending countercyclically. But don’t maniupulate short-term rates–keep them low (note from my first sentence in this paragraph that this doesn’t mean don’t use monetary policy at all).

    Finally, it’s always important to keep in mind that traditional monetary policy “works” by encouraging more/less spending out of exisiting income (i.e., cut rates to simulate borrowing), whereas fiscal policy works first by adjusting income, then spending through that. This suggests there would be times that clearly fiscal policy is the best option that are overlooked by most economists, even if you disagree with us that short-term rates aren’t a useful thing to be manipulating.

    • thanks for the explanation, Scott. This aspect of MMT needs to be put out there much more, I think. Because the first reaction of most mainstream people when they see ‘permanent ZIRP’ is: Hyperinflation!!

  14. Rob Rawlings

    Thanks for the details answers !

    It seems that what you are saying is: Monetary policy is nowhere near as effective as its supporters think it is and its effects can be indeterminate. This makes it a bad tool (according to MMTers) to manage AD.

    Beyond the need to control the financial system (which could in theory be done via regulations) does the MMT framework even need monetary policy and a CB ? It could optimize AD purely by controlling the size of the deficit and simply allow the banking system to determine the interest rate. No need for bond sales, target rates, or IORs. Is that theoretically possible within the MMT world, or am I missing something? What breaks with no CB ?

    • golfer1john

      With no bond sales, the natural risk-free overnight interest rate – insured banks lending to each other – is 0%. So there could be no need for government (and no ability for the banking system) to “manage” that rate.

      One function commonly performed by central banks is clearing, and that function needs to be done somewhere. But as for government economic policy, I don’t see why the Treasury couldn’t do it all without a separate CB policy-maker, if the government wished to allow the risk-free rate to stay at zero.

      • Rob Rawlings

        Question: Why would the overnight interest rate be zero without a CB ? Wouldn’t it gravitate to a market clearing rate based on lending opportunities and available funds ?

        • golfer1john

 is a definitive analysis, but basically it hinges on your premise of no bond sales. If Treasury didn’t sell bonds, there would be none for the CB to buy, and the result would be (as it is now, due to QE) excess reserves in the banking system – due to the deficit – which would drive the overnight rate to zero (in the absence of interest on reserves). Yes, there’s a market, but why would a bank hold sterile cash if they could lend it out, risk-free, at 0.001%?