Archive for the Category Liquidity trap


About that eurozone “liquidity trap”

Just a year ago, Keynesians were telling us that the eurozone was stuck in a “liquidity trap” and that the ECB was “out of ammo”.  Instead, Europe needed fiscal stimulus.  Now markets are predicting that the ECB will raise rates within the next 12 months:

Screen Shot 2017-03-07 at 9.12.20 PMObviously if the eurozone actually were stuck in a “liquidity trap” then it would be absolutely insane to raise interest rates this year.

For years I’ve been arguing that the sluggish NGDP growth we see in many developed countries is due to contractionary monetary policies.  Central banks are perfectly capable of delivering faster NGDP growth, they simply don’t want to.

Prediction:  Even as the ECB raises rates, we’ll still hear from the usual suspects that “fiscal austerity” is the problem, even though the US has done just as much austerity over the past 5 years, if not more.

Suggestion:  Those who don’t think the supply side of the economy is important should take a look at Germany and Greece, both operating under the exact same monetary policy.


Krugman on Sims

Bob Murphy directed me to a very interesting post where Paul Krugman discusses a recent paper by Christopher Sims (on the fiscal theory of the price level.)

Here’s Sims on fiscal policy:

Fiscal expansion can replace ineffective monetary policy at the zero lower bound, but fiscal expansion is not the same thing as deficit finance. It requires deficits aimed at, and conditioned on, generating inflation. The deficits must be seen as financed by future inflation, not future taxes or spending cuts.

I think he’s saying that fiscal expansion works only if it leads to a rise in expected inflation. Or maybe not – the truth is that I’m not sure, which is one problem with too purely verbal an argument. But it’s certainly something I’ve heard from helicopter money types, who warn that something like Ricardian equivalence will undermine fiscal expansion unless it’s money-financed.

But this is a misunderstanding of Ricardian equivalence, on two levels. First, as I’ve tried repeatedly to explain, a TEMPORARY increase in government purchases of goods and services will NOT be offset by expectations of future taxes even if full Ricardian equivalence holds. The kind of argument people like Robert Lucas made sounded Ricardian, but wasn’t – it was Ricardianoid.

Second, less relevant to Sims but very relevant to other helicopter people, a deficit ultimately financed by inflation is just as much of a burden on households as one ultimately financed by ordinary taxes, because inflation is a kind of tax on money holders. From a Ricardian point of view, there’s no difference.

So I’m trying to figure out exactly what Sims is saying. What, ahem, is his model? The little liquidity-trap model I devised way back in 1998 is forward-looking, does implicitly incorporate the government budget constraint, but doesn’t tell anything like Sims’s story. What is he doing differently, exactly? I’m confused – and I hope it’s not because I’m stupid.

Krugman’s obviously not stupid, and for the fiscal models of the sort discussed here he has better intuition than I do.  So this post will probably be wrong.  But I’ll give it my best shot.

Let’s start here:

I think he’s saying that fiscal expansion works only if it leads to a rise in expected inflation.

That seems like an odd way of putting it.  Any (demand-side) policy that is expected to be successful should lead to a rise in expected inflation, as long as the SRAS curve is not 100% flat (and it is not.)  And in any rational expectations model, a policy will only be successful if it’s expected to be successful. I think Krugman might have gotten on the wrong track by trying to frame this issue in Ricardian equivalence terms.

The quoted statement by Sims actually reminds me of Krugman’s 1998 paper, which says that currency injections are only effective if expected to be permanent—where permanent is defined as at least long enough to hold up until you are back in a monetarist world with positive interest rates.  If the currency injections are expected to be permanent, then they are expected to lead to future inflation.  That lowers real interest rates today (in the NK model) and boosts AD (NGDP) today.  Working backwards, a policy not expected to lead to future inflation is not expected to be effective.

Krugman’s right that inflation is a tax, but I think he relies too much on Keynesian reasoning in discussing the implication of that fact.  Consider Zimbabwe, which was hit back in 2008 by a massive inflation tax.  Certainly that hurt consumers badly, but nonetheless nominal consumer spending soared under the Zimbabwe government’s reckless policies.  Yes, it was all nominal growth, but that reflected the poor supply side characteristics of the Zimbabwe economy.  The AD curve was shifting rapidly to the right, so demand stimulus was “working” in that sense.  But it had nothing to do with consumers spending more because they felt richer.  They spent more because money was rapidly losing value

From a (market) monetarist perspective, the effectiveness of any demand-side policy depends heavily on the future expected path of policy.  Krugman and I agree that this is true of money, and everyone agrees that it is true of tax-oriented fiscal policy if you assume Ricardian equivalence.  The interesting question is what if you do not assume Ricardian equivalence?  Or what if the fiscal policy is more G, where G is output that is not a perfect substitute for C or I?

And that’s where my intuition tells me that Krugman is right, although personally I doubt the effect is very large.  In terms of the monetarist’s equation of exchange, I doubt whether G alone has much impact on V.  However if G is financed by a permanent increase in the money supply, it has a huge impact on M*V.  Even a credible future increase in M would by itself boost V far more than a sizable (current or future) increase in G.

Krugman’s 1998 paper can be generalized to show that any temporary change in M or V is ineffective at the zero bound.  I.e. that current M*V is strongly impacted by changes in expected future M*V.  In that case a temporary fiscal expansion probably has very little impact on demand.  On the other hand, in standard NK models it can still boost output, even if NGDP does not rise, as an increase in G will reduce C, and the newly impoverished workers will offer more labor, increasing aggregate supply—or something like that.

But Keynesians usually focus on the impact on fiscal policy on AD, and in that case I am inclined to support Sims’s claim that a policy not expected to boost inflation is also not likely to be effective.

That’s not to say I’m a fan of the fiscal theory of the price level; I don’t think it’s a useful theory for the US, although it obviously is for countries like Zimbabwe.

PS.  Krugman also says this:

Just to be clear, I’m all for fiscal expansion under whatever excuse. I’m even reluctant to question arguments for helicopter money, lest my intellectual skepticism give ammunition to those still possessed by austerian instincts.

I’m not at all reluctant to question arguments for helicopter money, as I do not worry at all about empowering people with austerian instincts.  So if you want to know what Krugman privately believes about helicopter money, read my posts showing why it did not work in Japan.

PPS.  On Wednesday September 7th, I will be presenting a paper at a joint Mercatus-Cato conference on monetary policy rules.  The conference will be livestreamed at

Got a blister on your pinky? Let’s amputate your right arm.

There is a rising chorus in the economics community calling for the abolition of cash.  The argument is that cash is the cause of the zero bound problem—the fact that nominal interest rates cannot be cut (very far) below zero.  And, so it is claimed, this causes weak growth in aggregate demand.

Actually, the severe recession of 2008 had nothing to do with the zero bound, as interest rates were still above zero.  It was caused by tight money. And after 2008, Bernanke always insisted that the Fed could do more and that it simply chose not to do more.

I suppose one could argue that during 2009-15 the real problem was that the zero bound led to a need for unconventional monetary stimulus, and the major central banks are reluctant to do enough unconventional stimulus, even if in theory they could buy up the entire planet.  But in that case the solution would presumably be psychological counseling for central bankers, not abolishing cash.  Ironically, one of the few central bankers who did recently call for more aggressive monetary stimulus, Narayana Kocherlakota, has now joined the call for abolishing cash.

I believe that this is a bad idea on several different levels:

1. There is no good theoretical justification for abolishing cash.  That’s because abolition of cash is strictly dominated by an alternative policy option—a higher inflation target.  The usual argument against a higher inflation target is the so-called “shoe leather” cost of inflation, the fact that people will go to ATMs more often with high inflation, and this makes our monetary system slightly less efficient.  But this argument makes no sense if the inflation target is being raised due to a fall in the Wicksellian equilibrium real interest rate.  The whole point of a higher inflation target would be to simply keep nominal interest rates above the zero bound.  And since the nominal interest rate is the opportunity cost of holding cash, a higher inflation target would not hurt cash holders any more than they were hurt during the 1990s, when nominal interest rates were well above the zero bound.

But let’s say I am wrong and inflation hurts cash holders more than I assume.  I still say there is no justification for abolishing cash.  Try explaining this to the average America:  “We are concerned that if we raise inflation from 2% to 4%, you will have to go to ATMs slightly more often, and that will be annoying.  So our currency system would be slightly less efficient.  And so to spare you from this slightly less efficient currency system, we’ve decided to abolish all currency.  And by the way, those blisters you keep getting on your pinky finger—the doctor suggests amputating your right arm.”

2.  Now you might argue that the shoe leather cost is not in fact the major cost of inflation.  I agree, the biggest cost is the excess taxation of nominal investment income.  But the exact same argument applies there as well.  If the inflation target is increased merely to offset a fall in the equilibrium interest rate, then there will be no problem of excessive taxation of nominal investment income, at least relative to the 1990s, when most economists thought the inflation rate (2% at the time) was perfectly fine.

3.  If there is an argument for abolishing cash, it is to reduce tax evasion.  But I believe that intellectuals (who mostly live in a near cashless economy) underestimate the utility of cash.  Go to an antique show at Brimfield, Massachusetts in the summer, and you’ll see an entire economy of 5000 small time “antique” (i.e. junk) dealers, all operating in a cash intensive economy.  The poor often don’t have much access to banking facilities, and use cash for many transactions.  If you are an upper middle class professional, it’s easy to imagine operating without cash.  But for many people it is not.

4.  In a cashless economy with a ubiquitous internet, the government will know everything about you that it wants to know.  It will know where you drive your car, and what you purchase.  We will be living in a giant panopticon.

I’m not so paranoid that I think the government would actually pay attention to most of our transactions, there aren’t enough bureaucrats.  But the information will always be available, if they want to go after someone.  Fortunately, Hillary and Trump would never even think of using this information to go after their enemies.  They are not vindictive people, or so I’m told by their supporters.  But maybe in the future a “bad guy” will be elected President.

Most importantly, there are other much better solutions that are not susceptible to the zero bound problem.  Replace inflation targeting with NGDP level targeting, as distinguished monetary economists like Michael Woodford, Christina Romer, and Jeffrey Frankel have suggested.  Even some Fed officials have recently pointed to NGDP targeting as an option—it’s no longer a pie in the sky idea.  In contrast, taking away cash would be almost as controversial as taking away guns.  This country still has a strong libertarian streak, and the total confiscation of cash is not likely to occur for many decades, by which time we’ll have much better options available for the zero bound.

HT:  Stephen Kirchner

Fix monetary policy; there are no alternatives

The annual Fed meeting at Jackson Hole will soon begin.  I hope they discuss how to fix monetary policy.  I fear they will be distracted by impractical fiscal/monetary schemes.

The recent discussion about monetary policy has been horribly confused.  Here are a few examples:

1. Prior to adopting a 2% inflation target, there was lots of discussion about where to set the target.  I recall almost universal agreement that the target had to be at least high enough to prevent a zero bound problem.  Now that we have experienced a zero bound problem, has that view changed?  If so, why?

2.  I see a lot of discussion to the effect that the Fed is unable to hit its inflation target, because it’s out of ammunition.  That’s not just wrong, it’s a freshman economics level error.  The Fed raised rates in December (and refused to cut them a few weeks ago), precisely because they are worried about overshooting their inflation target.  Indeed at least since the taper tantrum of 2013, the Fed hasn’t believed the US economy needed any more monetary stimulus than they are already providing.  This discussion of the Fed being out of ammo shows a profession that is deeply confused on some of the most basic ideas in monetary economics.  It’s not a pretty sight.

3.  When I discuss the possibility of reforms such as level targeting, or NGDP targeting, the response is often that the proposal is “politically unrealistic”.  Then the naysayers turn around and recommend fiscal/monetary coordination, which shows an almost laughable naïveté about the actual way that fiscal policy is implemented in the US.  And even if by some miracle the GOP Congress would agree to countercyclical policy, the proponents don’t even seem to know what it is.  They propose fiscal stimulus right now, even though doing so at a time of 4.9% unemployment would constitute a procyclical policy, and hence would make the US economy even more unstable.

I hope the economists at Jackson Hole discuss possible ways to fix monetary policy, and don’t get distracted by hopeless fiscal/monetary chimeras, but what I read in the opinion sections of elite media and blogs makes me very pessimistic.

Anand sent me to a post by Paul Krugman:

And I realized something not too flattering about myself: I’m feeling nostalgic for 2011 or so.

Liquidity-trap macroeconomics — which I didn’t invent, but did play a role in bringing back into the mainstream — had become the story of the day. And the basic message of the models — that everything changes when you hit the zero lower bound — was being overwhelmingly confirmed by experience.

The thing is, it was all beautifully hard-edged: a crisp boundary at zero, a sharp change in the impact of monetary and fiscal policy when you hit that boundary. And the predictions we made came out consistently right.

Yes, except for all the things they got wrong, which the market monetarists got right.  Like the 2013 austerity.  Or the removal of extended UI in 2014. Or whether the BOJ would be able to devalue the yen.

But now things have gotten a bit, well, murky.

Now Krugman is being much kinder than I was at the top of this post.  Since we exited the “liquidity trap” the discussion has been borderline incoherent.

The zero lower bound is not, it turns out, quite as hard a boundary as we thought.

Does “we” include the guy who twice recommend negative IOR in published papers in early 2009, and was scoffed at?

More important, probably, is the fact that two of the major advanced economies — the US and, believe it or not, Japan — are arguably quite close to full employment. We don’t know how close, because we don’t know how much pent-up labor supply is still waiting on the sidelines. But you can no longer argue that supply limits are no longer relevant.

Correspondingly, you can also no longer argue with confidence that there can be no crowding out, because the Fed won’t raise rates.

Yes, before you had to argue they’d do less QE to crowd out the fiscal stimulus.

We are, if you like, half-out of the liquidity trap, with one foot on dry land — but the other foot is still hanging over the edge, and it wouldn’t take much to topple us right back in.

What I would argue is that in this murky, fragile situation we should be conducting policy largely as if we were still in the trap — because we badly need to get both feet firmly on dry land with some distance between us and the quicksand.

(The link advocated fiscal stimulus.)  He seems to be arguing that we need fiscal stimulus because we need to raise rates so they we can cut them again if we get into trouble in a few years. That’s 100 times better than arguing the Fed should raise rates so that it can cut them again (an insane idea you sometimes see in the business press), but I still don’t quite buy it.

Fiscal stimulus is a demand-side policy.  It’s not going to have a significant impact on trend RGDP or trend NGDP.  So if we do more fiscal stimulus when unemployment is 4.9%, we probably won’t see dramatically higher nominal interest rates.  I will concede that it’s possible the Fed would have a bit more “conventional” ammo the next time a recession hit (from rates being a bit higher), but that additional monetary ammo would come at the expense of less fiscal ammo.  In Krugman’s view, fiscal stimulus doesn’t come from high G, it comes from rising G.  If we already have high G when we go into the next recession, it’s going to be that much harder to get rising G.  If fiscal stabilization policy is to work it must be countercyclical, that means don’t do it now.  This new “fiscal stimulus forever to put sand under the tires of monetary policy” (my words, not his) seems more than a bit ad hoc to me, a proposal from people who want shiny new airports and high speed rail, and will grasp any half way plausible model to justify that preference.  I put Larry Summers in that camp, maybe even more so that Krugman.

So I don’t think more fiscal stimulus today would make the US economy any more stable in the long run. Instead we’d just be:

Another day older and deeper in debt


Nonmarketable perpetual bond bleg

James Alexander directed me to a very interesting Bloomberg article:

Ben S. Bernanke, who met Japanese leaders in Tokyo this week, had floated the idea of perpetual bonds during earlier discussions in Washington with one of Prime Minister Shinzo Abe’s key advisers. . . .

He noted that helicopter money — in which the government issues non-marketable perpetual bonds with no maturity date and the Bank of Japan directly buys them — could work as the strongest tool to overcome deflation, according to Honda. Bernanke noted it was an option, he said.

Though Honda said he thought Japan was already engaged in a strategy that involved helicopter money, he wanted to convey the idea to Abe and asked Bernanke to meet with the premier in Japan. While this didn’t happen in the spring, Bernanke joined central bank chief Haruhiko Kuroda over lunch this Monday and on Tuesday he attended a gathering with Abe and key officials, including Koichi Hamada, another influential economic adviser.

Bernanke at the Tuesday meeting said Japan should carry on with Abenomics policies by supplementing monetary policy with fiscal stimulus, according to Hamada. Bernanke told Abe that the BOJ still has instruments to further ease monetary policy, said Yoshihide Suga, Japan’s top government spokesman. The central bank didn’t reveal what Kuroda and Bernanke discussed.

I understand perpetual bonds (aka consols), but I don’t get the “non-marketable” part.  If the Japanese fiscal authorities financed their deficit with marketable consols, and the BOJ bought them in the free market, you’d have an ordinary open market purchase.  It would not be a helicopter drop unless tied to a simultaneous fiscal expansion.  But if tied to fiscal stimulus it would be a helicopter drop even if the bonds were not perpetual.  So is it the “non-marketable” aspect that makes it a helicopter drop?

If these bonds became a large share of the BOJ balance sheet, and if Japan ever exited the liquidity trap and rates rose above zero, then the BOJ might have to sell off the perpetual bonds to prevent hyperinflation.  But you can’t sell “non-marketable” securities—is that the idea?

Bond traders, stock investors and economists have been mulling the possible implications of Bernanke’s visit and the next steps to come in Abenomics. Amid intense speculation about the chances of helicopter money, and the certainty of further fiscal stimulus ordered by the prime minister, Japanese shares have rallied for four consecutive days while the yen has weakened.

Fiscal stimulus makes a currency appreciate, so the recent depreciation is more likely due to the anticipated monetary expansion.

PS.  A few months back Bryan Caplan suggested that governments issue marketable consols as a way of out the liquidity trap.  It’s impossible for the yield on consols to fall to zero:

Step 2: The central bank uses standard open market operations to bid up the price of consols until nominal GDP starts rising at the desired rate.

Notice: With regular bonds, the difference between 1% interest and .1% interest seems trivial.  With consols, it’s massive.  A fall from 1% to .1% multiplies the sale price of a consol by a factor of ten!  There is an even bigger difference between a 1% interest rate and a .01% interest rate.  That multiplies the sale price a hundred-fold.  Can we really imagine that this massive increase in the public’s net worth won’t translate into higher consumption and investment?  And if not .01%, how about .00001%?

The only limit, as far as I can tell, is that the central bank might inadvertently retire its national debt.  When the bond price gets high enough, everyone sells.  But this seems like a remote possibility.

I like this idea even better than the non-marketable approach.  I’m not certain that retiring the entire debt is a “remote” possibility, but then I don’t feel I have good intuition in this area.  If it is a remote possibility, then Bryan’s idea would seem to eliminate the zero bound problem.