A response to Paul Krugman

Paul Krugman has a lot in common with quasi-monetarists like me. [Update: I'm going to re-emphasise this. On reading Paul's post, it is now much clearer to me than it was in the past that there is a helluva lot in common between Paul Krugman and quasi-monetarism. So many economists just don't get the central importance of monetary exchange and excess demand for money in understanding recessions. Paul does. Take as read a general rant about such economists.] To oversimplify just a little, we agree on the diagnosis, but disagree on the proposed cure. Which is a bit strange, though not impossible.

Unlike many macroeconomists,  Paul Krugman sees that the fact we live in a monetary exchange economy, and not a barter economy, is absolutely central in understanding recessions. A recession is not just a fall in output and employment. A bad harvest or earthquake could do that, even in a barter economy. It is a fall in output and employment accompanied by a generalised excess supply of goods in terms of money. A recession is an excess demand for money. If barter were costless (it obviously isn't) it would be very easy to escape a recession. Unemployed workers would simply barter the labour they cannot sell for the extra goods they could produce but which their potential employers cannot sell. And then the newly-employed bakers could barter their bread-wages for the beer-wages of newly-employed brewers.

Ultimately, I think, Paul Krugman believes that the Paradox of Thrift is really a Paradox of Hoarding. (I reckon I might have misinterpreted him in that old post).

His babysitting model is a Paradox of Hoarding model. That model wouldn't work if they could barter babysitting services, and "money" were just a savings vehicle.

And if you do think of recessions as an essentially monetary phenomena, it's also a bit strange not to propose a monetary cure. Strange, but not impossible.

But what precisely is the underlying difference between the cure Paul Krguman would propose and a monetary policy cure?

Paul Krugman would, I think, want US Nominal GDP to be (say) 5%  higher than it is right now, and then grow at (say) 5% per year thereafter. I don't think there would be much disagreement between Paul Krugman and (say) Scott Sumner over the precise numbers. Or even over the target variable itself. In any case, both would agree that something even vaguely like that would be much better than what the US has right now.

And if that's what you want monetary and/or fiscal policymakers to do, why not have them announce it, make it a commitment, that they will do as much as is necessary for as long as is necessary to hit that target? Paul Krugman would agree that expectations of higher NGDP would help the US escape recession. His model says so.

OK. The difference is on the means used to hit that target and fulfill that commitment. Paul would I think want a policy like:

1. The Fed prints money and Treasury uses that money to build a new bridge.

OK. Let's start there.

2. Any objection if they charge a toll on the bridge (subject to demand and collection costs)? The revenue from tolls could always be spent on other worthwhile government things, if you think that's a good use of the funds, from either micro or macro grounds.

3. Any objection if Treasury sells that bridge at a later date, and returns the money to the Fed, once the NGDP target has been hit, and there's a danger of overshooting the target? It would be good to make this policy reversible.

4. Any objection if the management of the bridge and collecting the tolls were turned over to the private sector? Or rather, any objection on purely macroeconomic grounds?

5. Any objection if the bridge is built, owned, and operated by the private sector, and the government just buys all the shares and collects the dividends?

6. Any objection if the government instead buys 50% of the shares in two new bridges, rather than 100% of the shares in one new bridge?

7. Any objection if the government buys commercial bonds in new bridges, rather than shares? Make the private owners the residual claimant, so they have the incentive to do their job right?

8. Any objection if we extend this to all new investment projects, not just bridges?

9. Any objection if we allow the Fed/Treasury to buy some old investment projects instead of just new ones? There's presumably high substitutibility between old and new investment projects, so the previous owners of the old investment projects will go looking for new ones with their new cash?

If there's no objection so far, then the cure for the recession is for the Fed and Treasury to announce a joint commitment to a level-path for NGDP, and to implement that policy by buying commercial paper with freshly-printed money.

Want to go the final step and have the Fed alone buy government bonds, on the grounds that government bonds are close substitutes for commercial bonds?

Or, maybe back up a little, and get Treasury to allow the Fed to buy the S&P 500 index? That's definitely OK by me.

Where exactly does fiscal policy end and monetary policy begin? (Not sure I know the answer to that one).

What is the real underlying difference between Paul Krugman and quasi-monetarists?

(It can't be just pessimism/optimism on the ability of monetary authorities to make credible commitments, can it?.)

Comments

[rabbit]

I suspect Krugman's political viewpoints are in part guiding his economic ones. In particular, he appears to desire a far larger role for government in the economy, and is becoming increasingly hostile to "unfettered capitalism".

[Nick Rowe]

rabbit: maybe. But perhaps the same thing could also be said about Scott Sumner, or me.

[brian]

fed buy S&P, not sure who really benefits on that front.. we have already had paper/fake asset inflation and it has only widened wealth gaps and hurt small businesses even more. I fail to see where inflating paper assets and closing output gaps have anything to do with one another. That type of thinking is just another fake monetary policy tool, haven't we established this is a fiscal policy problem not a monetary one?

[Luis H Arroyo]

Very good, indeed.
I´m jus have read this paper of Friedman I suppose you know (http://nb.vse.cz/~BARTONP/mae911/friedman.pdf), wich I see as a perfect complement for your text.
The problem is that the paper is from 1948, when Milton was for the 100% reserve banks. But his propposal of a financial debt without interest is brilliant, I think. It is the best "authomatic stabiliser" I never see. Do you think that his condition of 100% is absolutely necessary?

[Paul]

Could their be something to do with the certainty of governement spending on such projects?

With the construction labor market facing so much uncertainty, the bridge project alone may be more efficiently built/run privately, but knowing that regardless of external economic conditions or internal company conditions, this bridge will be built, might that not also have indirect effects to those benefiting from its construction?

I guess given curent conditions that would lead me to think before agreeing to point #5.

[Alex Godofsky]

Paul: are you really that certain that a government project won't have its funding cancelled? See also: tea party.

[Ian Lippert]

"Paul Krugman would, I think, want US Nominal GDP to be (say) 5% higher than it is right now, and then grow at (say) 5% per year thereafter. I don't think there would be much disagreement between Paul Krugman and (say) Scott Sumner over the precise numbers. Or even over the target variable itself. In any case, both would agree that something even vaguely like that would be much better than what the US has right now."

I didnt have much attention to pay attention to Scott Sumner's blog when he was proposing this method but are there any good resources for understanding this view? Is there any evidence that it will work? What happens if it only works temorarily and wears off once investors inflation expectations adjust? Are we then going to argue that we should try a 10%/yr target? I just dont really understand how an economist like Scott Sumner can propose something like this without having at least some fear that it will set off inflation at some point (or stagflation). Can somone explain this, I would really be interested in understanding their point of view.

[Patrick]

How about this: Let's do lots of #1 (doesn't have to just be bridges), and in the meantime you guys can go argue about 2 thru 9. We'll stop when you get it figured out. Deal?

:)

[Kevin Donoghue]

"It can't be just pessimism/optimism on the ability of monetary authorities to make credible commitments, can it?"

Why can't it? Decades ago Krugman was writing about the role of credibility in exchange-rate policy. And he never had a problem with letting the Fed take care of demand management as long as the ZLB was far away.

[marcus nunes]

Nick In his next post PK shows that the IMF indicates the "Confidence Fairy is dead", i.e. that Fiscal Consolidation is contractionary.
But then, I think:
Fiscal Consolidation is contractionary, especially if monetary policy is also \"non stimulative\". But note that fiscal stimulus was also \"non expansionary\" because monetary policy was, in fact, contractionary!
So, instead of giving up on monetary policy back in early 2009 to go \"all in\" for fiscal stimulus, mostly to \"peddle\" your liquidity trap view, you should have used your colums to \"cry out\" for more aggressive monetary action, especially since at the time inflation was completely absent!

[Martin]

Nick:

"What is the real underlying difference between Paul Krugman and quasi-monetarists?"

It is the distribution of debt. When people focusing on rebuilding their balance sheet the best monetary policy can do is 'reduce' that debt through inflation.

You [the QM's or MM's) never talk about a balance sheet recession (I nearly said BS recession), Krugman following Koo does. How does QM look at a balance sheet recession: a reduction in real spending to reduce debt? Because of the balance sheet recession, the demand for money is a symptom, just look at the kantooseconomics site http://kantooseconomics.com/2011/09/12/mein-unbehagen-mit-quasi-monetarismus-my-discomfort-with-quasi-monetarism/#English%20version

he basically agrees with that.

[Jacques René Giguère]

Nick: 1 ( yes Patrick lots of 1!) through 8 is almost non-controversial, But at 9, you're back at possible cash hoarding. Especially if the former bridge owner thinks like the owner of the London-to-York railway.

[Mike Smitka, Washington and Lee University]

The CP purchase option mentioned above depends as well on businesses being credit constrained: they see lots of profitable opportunities, they'd employ more if they could, they just can't get financing. Instead they're acting like households -- paying down debt, cash-rich.

The bridge example would work if infrastructure was largely private. But the network externality means that on the margin building new infrastructure isn't profitable, most of the gains accrue to holders of existing infrastructure (we, the people). Of course the Fed could purchase debt in projects that are not (at least by them, in private) expected to be profitable. But legal and political strictures are formidable, and in practice it may create jobs but not the benefits of infrastructure. I for example lived along an old national road in rural Japan that paralleled a new limited-access highway. But the tolls were so high on the latter that few used it. So the bonds used to finance the construction will (eventually) prove to be worth a fraction of their face value (they were purchased by the postal savings system) and the economic benefits are nil, once the job creation of the construction period passes. Furthermore, that sort of system is begging for abuse, and was abused, so it wasn't even very good at creating jobs.
This digresses from the formal theory, but it does suggest that (indirect) monetary policy [critical proviso] in the face of a zero lower bound is ineffective, given even a modicum of controls that keep it from being a near-transparent version of fiscal policy.

[Mattay]

"Any objection if they charge a toll on the bridge?"

Yes, assuming the toll "proceeds" go to the government. A toll has the same effect as a tax of transfering financial wealth from the private sector to the government sector. Our goal is directly the opposite - to increase private sector wealth.

"Where exactly does fiscal policy end and monetary policy begin?"

The Fed engages in contractionary fiscal policy any time it makes a "profit."

The Fed engages in expansionary fiscal policy any time it makes a "loss."

The Fed engages in monetary policy any time it breaks even (or if you like, any time it comes reasonably close to breaking even).

Given the current state of the economy, any time the Fed makes a "profit" it is a national tragedy, while any time the Fed makes a "loss" that is cause for tremendous celebration.

Now, let's take your last suggestion of the Fed buying the S&P 500 Index - SPY is currently at 118.

Supposing that the Fed posted a bunch of bids to buy SPY at $150, that would be great - especially if they then followed up by selling their SPF shares back to the market at $118. Why would that be great? Because the Fed would be making a "loss."

Alternatively, suppose that the Fed posted a bunch of bids to buy SPY at $118, that would be ok. It would become fantastic if the next day then the next day they sold all their SPYs at $1. Why would that be great? Because the Fed would be making a "loss."

But I would suggest that buying SPYs is not the most equitable, fair, or effective way to do this - people who own stocks tend to be much wealthier than the average citizen and have lower than average propensities to consume or pay down debt. It would be much better to just mail all citizens a check, and dispense with the charade of taking an asset in return.

[Max]

The main objection to each progression is: yes, that might work, but it's less likely to work.

Also, if you really want to use expectation effects, what you want is a stimulus program which is obviously too large and effective. "If this doesn't work, we will do more of the same" is not very credible.

[Paul]

alex - I'm more defending the possibility that this type of government expenditure could be more useful now, then predicting that it would be useful. The gov't at least in theory should be able to to take a longer term and large scale approach to spending and infrastructure investment then private buisnesses.

While most of the time these processess may be more efficient by channeling much of the operation through private channels, agreeing with what Nick stated above. I think in the current situation, having a gov't that can do the process start to finish (and do it fairly well), certainly until number 3, would be best. That isn't to say it could do it fairly well.

[JKH]

He'd object at number 9 and everything following, unless the first 8 were exhausted. The key qualifier is "new". That's fiscal. "Old" is monetary.

[foosion]

I believe Krugman has said that in order to move the needle on NGDP or inflation a meaningful amount under current conditions would involve such a vast issuance of money as to be impossible as a practical matter.

[Chris J]

Amateur question: Don't we object to number 2? If the goal is to add money, create a bit of inflation because it is too low, and build something useful, then charging a toll before the economy improves seems counter productive.

[Greg Ransom]

No, as Kling & Hayek & Mankiw point out, ia tradype cycke is a dis-coordinated economy, most especially in the areas of invedtment / capital and labor -- items which to a limited degree are substitutes, but which are specialize and _don't_ fit when the economy is in severe discoordination -- some items, like lumber hauling rail cars & CDSs & houses -- lose their prior value status, which was dependent on an unsustainable equilibrium.

Insisting on ignoring an empirical pattern staring you in the face, because "theory" won't allow to to think it, is not a competent strategy for a scientist.

Nick writes,

"It is a fall in output and employment accompanied by a generalised excess supply of goods in terms of money."

[Greg Ransom]

Sorry for the Ipad typos ...

[Luis Enrique]

Nick,

I think it's quite sensible to believe that government command spending can do things (build bridges) that the private sector would not do, if monetary policy conducted by Fed buying corporate equity and commercial debt. You don't have to think government run economies are good ideas to think that in some circumstance, a bit of command and control can be useful.

Here's something that has been puzzling me, would be grateful for your thoughts.

If we did want to monetize government debt to enable some fiscal stimulus, the Fed could buy debt and announce it plans to roll that debt over in perpetuity so everybody can regard it as written off, and the real government debt (that will ever need repaying) has shrunk. So the only difference between that and QE is the intention to roll over, rather than sell back to the private sector in due course. Its puzzling because the only difference between supposedly non-inflation stocking QE and supposedly inflation exploding Zimbabwe style monetizing debt concerns future intentions.

Why are discussions about doing more monetary stimulus always about buying more, as opposed to changing what the Fed plans to do with what it has already bought? If it wants to create a bit of NGDP growth, and thinks some fiscal elbow room would be useful, why isn't there explicit talk of monetizing (writing-off) some portion of what's already on the balance sheet? And ... what do we know, and how do we know it, about the relationship between funding government spending via the printing presses, and inflation, in the current circumstances? How much could we get away with, how responsive is inflation to it?

[Greg Ransom]

This is a lingustic stipulation, it's not a fact of reality:

"A recession is an excess demand for money."

It's you attempting to legislate a use of an essentially contested word.

Everyone else is free to say, "Bunk".

Money, credit, leverage etc. allow an economy to discoordinate the structure of production across time -- e.g housing finance & housing construction & housing manufacturing support capital, etc. -- this massive discoordination of the interconnection of production and consumption will but specialized goods out of work, looking for a valued place in the re-eqypuilivrating system.

Some of these will never find a new place of values use -- value greater than cost of their productive exploitation.

This is Econ 101, but econ 101 is banned from macroeconomics due to the stipulated dogma of your definition of
"recession" and due to the bogus causal entites of aggregated GDP, aggregated labor, aggregated capital, etc. which are pretended to directly interact the "tool kit" everdya macro models stuck in the head of macroeconomic "scientists".

[Nick Rowe]

Patrick: we can do 9 much quicker than anyone can do 1.

JKH: suppose the government buys a brand new office chair. That's fiscal. Suppose it buys a chair that was produced 5 minutes ago? 5 hours ago? 5 days? 5 weeks? 5 months? 5 years? 50 years? Someone's got the cash, and will want to replace the chair.

Mattay: So if the Fed bought the S&P500, and this caused the economy to recover, and S&P profits to increase, and the value of the S&P500 to increase, so the Fed sold them back during a boom at a profit, that would be a failure? There's more to economic policy than income (wealth) effects.

[Determinant]

A balance-sheet recession is state-space stable. That means that once the state condition is in place (businesses want to pay down debt and hoard cash rather than invest) it is stable, sticky, persistent and won't change unless we change the state-space parameters.

It means he economy is money-constrained instead of capacity constrained. It also means additional investment to expand capacity is useless.

In order to change the state-space condition you need to inject money into the economy. We need money supply inflation, which because of the balance-sheet recession state will not lead to price level inflation, the kind of inflation everybody hates.

Once we are back to a capacity-constrained economy which needs investment to expand to meet demand and can use that investment, then we can worry about price inflation.

Additional investment in a balance-sheet recession is only useful as a method to deliver money. It works because it's an acceptable method to reach households. Remember in our economy we distribute wealth through wages and employment. The investment is a delivery vehicle for added money, not an end in itself.

Balance-sheet recessions turn some fundamental conditions of a normal capacity-constrained economy upside down. The fact that it does and these are not conditions we normally encounter is what makes the solution so difficult because the methods are atypical, counter-intuitive and are inappropriate for a normal capacity-constrained economy.

[Nick Rowe]

Martin: David Beckworth has done a number of posts on the balance sheet recession theory. I have addressed it too, obliquely. Whenever I do weird posts on antique furniture and the Pro-usury party I talking about whether an excess of desired saving can cause a recession. And Richard Koo's theory is that a balance sheet recession is when creditors stop lending, and borrowers stop borrowing, and this (somehow) causes an increase in aggregate desired saving and this (somehow) causes a recession.

[Jeff Fisher]

"There's presumably high substitutibility between old and new investment projects, so the previous owners of the old investment projects will go looking for new ones with their new cash?"

Sounds like an assertion that investors are just as likely to use cash to fund new projects during bad times as during good. I don't buy it.

[K]

Luis Enrique: "If we did want to monetize government debt to enable some fiscal stimulus, the Fed could buy debt and announce it plans to roll that debt over in perpetuity so everybody can regard it as written off"

You need to be very clear on exactly what you mean.  If you are paying interest on excess reserves then you are just replacing long term debt with short term debt - Operation Twist. I don't think that's what you mean.

If you are not paying IOR, then you *cannot* raise interbank lending rates above zero until you have removed the excess reserves. Otherwise there would be no bid for settlement balances, and the CB would lose the ability to control the marginal cost of funds for commercial banks. So a commitment to permanent QE is a commitment eternal ZIRP. It's not credible.

The basic problem is that there is a very limited demand for non-interest bearing money when interest rates are not at zero - just basic liquidity requirements. You can't will there to be arbitrarily large amounts of it (or the risk free rate will collapse).

[rabbit]

Rpwe:

"But perhaps the same thing could also be said about Scott Sumner, or me."

Or anyone.

But unlikely to the degree that Krugman has. He is showing increasing signs of bitterness, polarization, and contempt for those who disagree with. It's not a pretty sight.

[Andy Harless]

"Any objection if the government buys commercial bonds in new bridges, rather than shares?"

Yes, I object. When interest rates are very low, bonds are a fairly close substitute for money, while real investment is not. Buying the bonds therefore won't necessarily result in more bridges being built; it might just result in more household assets being held as money rather than bonds. (I'll grant you it would probably work if the government buys enough bonds, but the number required might be quite huge relative to the amount of new investment created.)

Of course we've had this argument before, and your answer is something like:

"maybe back up a little, and get Treasury to allow the Fed to buy the S&P 500 index"

which is fine with me and probably with Paul Krugman as well, though many (perhaps including him) will object that this is now fiscal policy rather than monetary policy. His larger point, though, would be

"that the quasi-monetarists are trying too hard to find a deep essence when what’s really needed is just a model."

And here I'm going to switch and put my quasi-monetarist hat on. I think there's a good reason for trying to find a deep essence. It's a political issue: Paul Krugman is saying, "Hey, let us technocrats take care of it, since we have a model that works well enough." That's not going to fly, because people don't trust technocrats. If you can find an essence, you can perhaps get people to see that it really is the essence, and there is some hope of getting them to realize that there is a monetary problem that requires a monetary solution (although the monetary solution could take the form of a fiscal policy designed to reduce the demand for money).

[Nick Rowe]

Chris J: there is nothing to prevent the government taking that toll money and putting it right back into the economy, either to finance more bridges, or handouts, or tax cuts, or whatever. If it's needed. The Bank of Canada's income from past money creation gets handed right back to the government in any case. Or, set the toll at $0 till the recession is over, if you like. Or hand it back to the drivers as they exit the bridge! It doesn't matter. But it's important that the government can collect the toll revenues and use them to retire debt and/or money when the recession is over. And those *potential future* toll revenues are what allow the government to sell the bridge and get its money back when the recession is over and it needs to reverse policy.

[Determinant]

@Greg Ransom:

That's sounds a bit Austrian. I derived my balance-sheet recession position from Roger Garrison's Time and Money model. If you treat the Loanable Funds market as the money supply and draw a circle around it to show the size of the money supply and assume a monetary economy where all trades are for money, then you can easily get monetary disequilibrium and make the model demonstrate a balance-sheet recession with an output gap.

Sure it is degenerate behaviour but it is plausible and self-consistent. As an engineer by training I have learned to accept suboptimal and degenerate cases.

Nick's definition is simply declaring a frame-of-reference. In physics there is no preferred frame of reference so you always have to define your own. In monetary economics that preferred frame of reference is money. A shortage of money is a demand deficiency which is also the same as glut of good relative to money. It all depends on which frame of reference you use, goods or money.

Goods to goods is not a valid frame of reference in this debate.

[W. Peden]

Martin,

What is the difference between a balance sheet recession and a rise in the demand for base money? Here's what I think: a balance sheet recession is when the public try to rebalance their debts. This drives up the demand to hold money i.e. people want to save more and borrow less. However, unless the money supply is increases to match demand, they will be frustrated in their efforts through debt deflation. A balance sheet recession is just a way that you can get a disequilibrium between the supply & the demand for money. It is wrong to separate the former from the latter.

The monetary disequilibrium IS the cold. It is not a symptom. The balance sheet problems are just how the cold is caused, but there are many ways of getting a cold. It just happens to be the case that, while there is no cure for the cold, there is a cure for monetary disequilibrium.

(A different case of demand-side monetary disequilibrium might be something like interest rates on excess reserves, which increase the demand for to hold base money.)

[Martin]

@Nick, I believe Richard Koo focuses mostly on that borrowers do not want to borrow because they want to minimize debt and that this behavior - debt minimization - depresses spending and this depresses AD. The creditors in this instance are the banks that accumulate cash on their balance sheets: I believe Nomura had about 17% Equity. The job of the government is to mop up those savings and to spend them, because banks cannot find anyone else to borrow. That is how this is supposed to cause a recession.

I believe he also argued that the spending gap was quite large 10-20% of GDP that was filled with a modest amount of stimulus in Japan.

As I understand MM, the BOJ should have announced an NGDP target and done everything to reach it. If there however is a spending gap of 10-20%, how much inflation is that going to generate from a year to year basis? I know it must be less than 10-20% as it also makes taking out credit more attractive, but how much can you expect with such a gap?

[Mattay]

Nick said:

"So if the Fed bought the S&P500, and this caused the economy to recover, and S&P profits to increase, and the value of the S&P500 to increase, so the Fed sold them back during a boom at a profit, that would be a failure? There's more to economic policy than income (wealth) effects."

Your hypothetical scenario sounds great to me. If the outcome that you describe did indeed occur, that would be wonderful. To be clear, I have no real objection to trying it. I simply don't understand why it would work, beyond the fiscal effects that could stem from fed "losses."

What mechanism would cause this hypothetical fed fueled recovery to actually occur?

[Martin]

@W Peden:

"What is the difference between a balance sheet recession and a rise in the demand for base money? Here's what I think: a balance sheet recession is when the public try to rebalance their debts. This drives up the demand to hold money i.e. people want to save more and borrow less. However, unless the money supply is increases to match demand, they will be frustrated in their efforts through debt deflation. A balance sheet recession is just a way that you can get a disequilibrium between the supply & the demand for money. It is wrong to separate the former from the latter."

Let's say it does, off-setting that demand, however does not mean you have solved the balance sheet recession, you have merely prevented a secular depression. It just hasn't gotten worse, but the problem is still there.

[W. Peden]

Martin,

Assume money isn't in disequilibrium and the public as a whole try save more (for whatever reason). Two things can happen: (1) the cost of financing government expenditure falls as the demand to borrow from the government rises; or (2) the public adjust their desire to save and borrow in line with their new money balances (Fisher's dance of the dollar). Either way, the debt deflation cycle has ended with the increase in the money supply, because debt deflation only holds if people can only reduce their balances with existing money.

Hence there have been recoveries from severe deflationary recessions (like Britain after September 1931) with tight fiscal policies with compensating monetary policies. Monetary disequilibrium models have a very good explanation of why something like that would happen.

[Luis Enrique]

K

You are talking about things I'm not sure I understand. I had in mind something simple. If I lend you $10 and then every time you repay it, I lend you another $10, that debt is effectively written off. The Fed can do that*. You say it's "not credible" ... I don't think by that you mean we could go ahead and use the printing presses to finance government debt with no consequences for NGDP.

* indeed it doest that. Over the long run, that's how it realises net new money into the economy, by bringing govt debt onto it's balance sheet and keeping it there.

[Henry Kaspar]

I obviously can't speak for Krugman, but more me:

(It can't be just pessimism/optimism on the ability of monetary authorities to make credible commitments, can it?.)

... is indeed critical, especially at or near the zero bound.

Further, the key step in your example is 6. From step 6 one needs a private agent willing to spend cash on stuff rather than hoarding it - and in a liquidity trap one may not find that agent. By contrast, pure government spending translates directly into demand and creates shortages in goods markets. Thus it always works - as long as the government has capacity to indebt itself - while monetary policy may not.

[Martin]

W. Peden,

I disagree, your argument seems to be premised on the idea that you can only have a debt-problem when there is debt-deflation. When debt is outpacing NGDP growth, then you will have a debt problem in the future without it being caused by debt-deflation. Worsened, yes, caused no.

[Bill Woolsey]

Did anyone read the comments on Krugman in the NYT?

Yikes!

[W. Peden]

Martin,

Fair point. I should have distinguished more clearly between debt deflation (which is sustaining disequilibrium within a recession) and a balance sheet recession (which can cause a recession). The way that monetary equilibrium solves them both is the same though: the increase in the money supply solves the monetary disequilibrium, and in monetary equilibrium one person's rebalancing is another person's profits.

I certainly don't want to say that debt is only a problem when there is debt deflation (or any sort of deflation). I do think that debt-rebalancing is turned into a benign and stabilising process by monetary equilbrium. The first half of Great Moderation in the UK was basically a story of people rebalancing their portfolios after the early 1990s housing bust. It was a period of astonishing stability in output, falling unemployment, and inflation between about 1.5% and 3.5%. The rebalancing of portfolios through paying off debt was totally commensurate with this stability. (Incidentally, fiscal policy slowly tightened over this period, leading to surpluses in the late 1990s.)

[K]

Luis Enrique: I'm going to (slightly) take back what I said (but not materially :-).  A commitment to perpetual QE is not a commitment to perpetual ZIRB. Reasonably there is some equilibrium ratio of quantity of money, M, to NGDP (It's one over the velocity of money). So a commitment to maintaining a particular value of M, is a commitment to maintaining ZIRP until inflation/growth has rendered NGDP up to the equilibrium level implied the that value of M. So lets say we doubled M since last time the economy was at a reasonable equilibrium (2007ish) and commit to keeping it there. Then we are saying we are going to keep rates at zero until inflation/growth has doubled NGDP. *That* is not credible. Maybe if we had increased the money supply by, e.g., 20% then that would be a reasonable commitment. Maybe they should increase reserves by 5% each year and *promise* never to decrease them.

But then are we really not just back in the contingent path of rates game? It's just a commitment not to increase rates above zero until NGDP, inflation, whatever,  is at some target. M has nothing to do with it cause it's not doing anything anyways while we are at the ZIRB. The only relevant action is the announcement of the conditions under which we exit the ZIRB (and the contingent path of rates thereafter). The market doesn't care one iota about M in the meanwhile.

[Lars Christensen]

Bill, as always - you are so right! Yikes!

[Nick Rowe]

Henry: "From step 6 one needs a private agent willing to spend cash on stuff rather than hoarding it - and in a liquidity trap one may not find that agent. By contrast, pure government spending translates directly into demand and creates shortages in goods markets."

Suppose we had a Keynesian and a Monetarist arguing like this:

Keynesian: "Y=C+I+G+NX. Therefore if G goes up, Y directly goes up. (Unless there's some weird offsetting effect through C, I, or NX)."

Monetarist: "MV=PY. Therefore if M goes up, Y directly goes up. (Unless there's some weird offsetting effect through V or P.)"

Why should it matter whether the government buys 100% of 1 new bridge or 50% of 2 new bridges, in consortium with private partners? You *might* argue that 2 extra bridges get built in the second case. It all depends what those partners would have done otherwise. But you also wonder what everybody else would have done otherwise in the first case.

[Nick Rowe]

Bill, Lars: which comments? As usual there were a lot of comments from people who don't know much economics, criticising both quasi-monetarism and Paul Krugman from all sorts of directions, right and left.

[Lars Christensen]

Nick, its Market Monetarism;-)

[Lars Christensen]

But we are talking about the comments you did your blog on...so I guess you are covered.

[Martin]

@W. Peden,

I agree with you on the monetary equilibrium, the monetary authorities should do their utmost to maintain this. However how benign re-balancing is going to be seems to depend on the amount of debt and how it is distributed in the system.

I also do not think that you need NGDP to be on target for there to be a monetary equilibrium and the desirability of NGDP being on target when there is a debt re-balancing depends on the amount of inflation that is going to accompany this. This seems to depend on what the stance of fiscal policy is and how much debt is depressing spending.

In case of the UK, I think that improved macro-economic conditions in the world made it possible for the UK to export its way out of that mess: http://research.stlouisfed.org/fredgraph.png?g=2bo This, is what probably allowed them to tighten.

[W. Peden]

Martin,

I don't know about that. I suspect that world conditions were more favourable in the 1960s, on balance, but we managed to get ourselves into accelerating unemployment and inflation during that period. What was really important was (a) better supply-side policy and (b) better demand-side policy. The former made sustainable low unemployment possible, the latter kept inflation in check. Also, there's no theoretical reason (that I know of) which would preclude doing what the UK did in an autarkic economy.

And while macro-economic conditions improved in the world during that period, few countries improved as much as the UK. Some of the best were those engaging in fiscal tightening e.g. the US during the late 1990s.

As for NGDP being on target requiring inflation, what fundamental difference is there between paying off debt out of increasing real incomes and paying off debt due to inflation?

[Lord]

I think the entire means of the Fed to create money under current circumstances is suspect and only leads to exchange of instruments of one kind with instruments of another. If we legalized the money drop, the distribution of sums equally to every adult citizen unbacked by debt, I have no doubt it would work. Money must get into the hands that most need and want it and allow them to reduce debt or spend as they see fit. That is a little more likely with fiscal than monetary policy, but only a little.

[K]

Luis Enrique: My last comment could have been more concise. To summarize: What you are saying is "I will not decrease the money supply below M" which is equivalent to "I wont raise rates until NGDP/V > M, where M is a certain amount of QE. Why don't you just say I wont raise rates until NGDP is some value, or inflation is some value, or whatever. Why involve the quantity and velocity of money (which nobody cares about) in your commitment.

[And now I'll stop since we are obviously OT. Sorry Nick.]

[jean]

Paul Krugman would certainly object on the reversibility of the policy.
Krugman has repeatedly claimed that the liquidity trap problem was a commitment problem. The central bank would like to have a bit of inflation now but not too much inflation later. So once the recovery has occurred, the central bank would tighten its policy. Anticipating this tightening, money market actors don't react to the announcements of the central bank.
Budget policy has the same time-inconsistency problem (called Ricardian equivalence), but neither the government nor the central bank would not let happen something like a sovereign default.
I agree that such a commitment device is rather strange. I think that letting the central bank have negative equity would be quite efficient too.

[Nick Rowe]

jean: the reversibility issue *might* be a problem. But if there is a commitment not to reverse the project unless the NGDP target looks like being overshot, and if you have set that future NGDP target high enough, that should be OK. I.e. you promise not to reverse it at the expense of your future NGDP target.

[rsj]

Monetary policy is a balance sheet operation that leaves your net-worth unchanged. It may, however, have some fiscal side-effects.

Fiscal policy is an income operation that increases your net worth (in the case of a payment received from the government) or decreases your net worth (in the case of taxes paid to the government).

Anything that increases your net worth is fiscal. Anything that changes the composition of your assets is monetary.

To the degree that there is an overlap between monetary fiscal policy, in the sense that purchasing assets can cause fiscal policy to occur as a side effect, the Federal Reserve is generally forbidden from engaging in it.

There are only two exceptions: setting the OIR, which obviously can cause net-worth to go up or down based on changes to the discount rate, and lending to banks, which if they default will create a fiscal transfer ex-post. The Fed is supposed to regulate banks to ensure that lending to them does not cause this problem, and it is also supposed to only lend against collateral to further limit any fiscal spillovers.

So interest rate management is the only thing the CB can do to try to increase the net-worth of the private sector.

That is why the CB is only allowed to purchase assets guaranteed by the U.S. government. Not SP 500 In that case, the fiscal operation happens when the guarantee is made, and then the asset swap is monetary policy should the CB purchase the asset.

The central bank is also only allowed to purchase obligations -- e.g. financial assets -- not real assets. Not bridges.

Fiscal operations must be passed by the legislature, which doesn't like for the government to purchase private capital. Monetary operations can be done by committee, with the bounds set for them by congress (e.g. to limit spill-overs in which monetary policy creates fiscal side-effects).

[Andre]

(It can't be just pessimism/optimism on the ability of monetary authorities to make credible commitments, can it?.)

I am not an economist, but my layperson reading of Krugman is that for him it is largely about expectations and the central bank's ability to credibly commit to higher than normal target inflation. QE can prevent a deflationary spiral but won't lead to much growth if expectations are that the central bank will keep a lid on inflation (e.g. 1.5-2%).

http://krugman.blogs.nytimes.com/2011/08/26/academic-debate-real-consequences-wonkish/

http://krugman.blogs.nytimes.com/2010/08/31/japan-1998/

But perhaps I've misunderstood...

[Nick Rowe]

rsj: "Monetary policy is a balance sheet operation that leaves your net-worth unchanged. It may, however, have some fiscal side-effects.

Fiscal policy is an income operation that increases your net worth (in the case of a payment received from the government) or decreases your net worth (in the case of taxes paid to the government)."

OK, So say the Fed prints $1 billion, lends it to treasury, and Treasury buys a new bridge worth exactly $1 billion, and hands the $1 billion cash over to the people who built the bridge.

Did the government's net worth change? Is this fiscal or monetary?

Andre: or maybe we've misunderstood. Dunno. But if QE can prevent a deflationary spiral, why can't a little bit more or better QE get the economy out of the recession?

[Steve Waldman]

Reading one through nine, the voices in my head began to shout. "There's many a slip 'twixt the cup and the lip!"

Points 2, 7, 9 and the conjectured substitutability of Treasuries for risky commercial paper are all places that seem likely to slip. Politically, the user fees collected might not in fact be spent. If we are buying private commercial paper, we might simply bid down the yield on existing projects rather than inspire new real expenditure. Do we know the elasticity of real investment to reductions in the commercial paper rate? Does this change during periods where "animal spirits" are low, and the many front-loaded costs and coordination problems associated with planning and executing new real investment must be overcome by human beings who do not have an investment opportunity set all planned and ranked, just waiting for a reduction in their cost of capital? How substitutable is government paper for private paper? Isn't part of the problem that during difficult times, investors view these as categorically rather than incrementally different?

This is all a very internecine conversation. I would love, and I'll bet Paul Krugman would love, for Scott Sumner to be elected pooh-bah and announce a target NGDP or a price level path. I'd be hopeful, but ultimately skeptical that it would work, as long as the central bank is restricted to swapping money for government paper or well-collateralized claims on banks. I like all of the quasimonetarists very much, but the expectations fairy is invoked a bit too often for my comfort. It's not that I doubt the power of expectations, but I do doubt the power of the central bank to set expectations if its tool set is very limited.

If the Fed can buy commercial paper directly (as it did during the crisis, but would have a harder time doing again, post-Dodd-Frank), I am a bit less skeptical. If the Fed could buy paper issued in primary markets, and restrict its purchases to entrepreneurs seeking funding for new projects, I'd be less skeptical still. If the Fed could fund new projects directly, and set the level of incremental taxes or fees at a level it deems appropriate on macroeconomic grounds, all skepticism has vanished.

But that's a different Fed than the one we have in the United States, and as we eliminate potential slips, traveling down your list from 9+ back to 1, the name for the kind of policy we are pursuing slowly shades from "monetary" to "fiscal".

[Nick Rowe]

Hi Steve! Ah, I'm slowly leading you down the slippery slope or garden path to quasi-monetarism!

OK, let's suppose that there is a slow slippage from 1 to 9. And that $1 billion spent on 1 would have more effect than $1 billion spent on 2, on 3, etc...If so, we would have to increase it from $1 billion, to $2 billion, etc...to get the same effect.

But:

1. What's the opportunity cost of that extra $1 billion, $2 billion, etc.? Isn't it $0, if all these projects are reversible? And actually, as we move on down the list, it gets easier to reverse them and recoup the government's investment.

2. The credibility of the government's commitment is far more important than what it actually does. It's that increased expected future NGDP that's going to do all the heavy lifting. The only purpose of 1 through 9 is to put some real wings on the poor old confidence fairy. It's the credible threat to do 1 through 9 that matters, not actually doing 1 through 9. Simply being willing and able to do it if people don't believe you about NGDP. And if they do believe you are willing and able to do it, they will believe in NGDP. And as we slide down the scale from 1 through 9, it gets easier to do it, politically, so the threat to do it is more and more credible.

Hell, if that threat was perfectly credible, so expected future NGDP took off, the Fed and government would have to do 1 through 9 in reverse, to stop the economy overheating. The existing stock of cash plus the existing deficit are far too big to be compatible with full employment equilibrium and NGDP growing at some sensible rate like 5%.

[Scott Sumner]

Nick, You said;

"(It can't be just pessimism/optimism on the ability of monetary authorities to make credible commitments, can it?.)"

I think it is. There are minor differences such as my view that it's wrong to think of monetary stimulus in terms of lowering the real interest rate, and instead it's about lowering the nominal interest rate relative to NGDP growth. But I think the sentence you provided is the key difference.

[Nick Rowe]

You might be right Scott.

But damn, with his latest post in particular, the differences between Paul Krugman and quasi-monetarism (market monetarism) look so very small. Until now, I had thought he might be (like many New Keynesians) an excess saving theorist. But it's quite clear here that he recognises that excess saving only causes a recession if it spillsover into excess hoarding of money.

[Michael Simms]

When I took economics in university I never really though there was much of a difference between Keynesians and monetarists. They have the same goals, they just prefer different tools.

[Nick Rowe]

Michael: the joke is that the battle between Keynesians and Monetarists was resolved when the Keynesians agreed to (almost) all the monetarist policy proposals and the Monetarists agreed to call themselves New Keynesians. But the current crisis has re-opened some of the debate.

[Leigh Caldwell]

Your 9-point scale is a good outline of the spectrum from fiscal to monetary intervention.

There are certainly adjacent steps on the scale which are not equivalent. Your example in a later comment is a good one:

"Suppose the government buys a brand new office chair. That's fiscal. Suppose it buys a chair that was produced 5 minutes ago? 5 hours ago? 5 days? 5 weeks? 5 months? 5 years? 50 years? Someone's got the cash, and will want to replace the chair."

In fact, this example shows why the two policies are different. There are lots of old office chairs which would not be replaced if the government bought them. In fact, there are lots of chairs in abandoned office buildings which would be brought out and dusted off to sell to the government if Congress passed a cash-for-chairs scheme. This is an argument both for and against fiscal policy. For, because buying newly produced chairs is better than buying old ones (antiques?). Against, because it shows how easy fiscal policy is to game.

Behaviourally, this is explained by some kind of status quo bias/risk aversion - as a business I might hang onto equipment just in case it comes in useful - its net present value is slightly greater than zero. But it doesn't mean that I would replace it if someone bought it.

Now that should not be true of government or corporate bonds because they are liquid, with a clear tradeable value. But it will definitely be true of investment in, or lending to, small businesses, which is not fungible. It will be true of many fiscal interventions. There are a few reasons why intervention at step 1 will not be equivalent to intervention at step 9 - most of which I think arise from the leap from step 8 to 9, because old and new projects are only equivalent in a limited subset of the economy (liquid corporate bonds and some equities).

Which is not an argument for or against step 1 relative to step 9. Either might be right. Each has different positive and negative impacts. Each is reversible to a different extent, has different political acceptability, raises different credibility questions, and creates different kinds of incentives. You've demonstrated quite well why they are not as different as they might instinctively feel. But they are different, and perhaps these essentially political questions of credibility and incentives are the main reason they are different.

Is there really a difference between an "excess saving theorist" and someone who believes in "excess hoarding of money"? I think I have been persuaded by you that these two are the same, but that's from a starting point of some macroeconomic naivety. What is the alternative theory?

[JW Mason]

As others have said: At some point between 6 and 9, it becomes doubtful that the bridge will in fact be built.

At some high level of abstraction, the question "is this bridge profitable enough to be worth building" should be equivalent to the question, "will someone pay enough for a security representing the stream of future profits from operating bridge to cover the cost of building it," which in turn should be equivalent to, "if we convert the expected stream of returns to the bridge to an equivalent fixed stream of annual payments, is the ratio of one payment to the cost of the bridge above or below the prevailing rate of interest."

But in practice, these questions don't seem to be equivalent at all. There's a big empirical literature on the sensitivity of business investment to interest rates, and the findings more or less across the board are that it's minimal or nonexistent. As Alan Blinder says, "you have to torture the data pretty ruthlessly before they confess to an interest elasticity of investment." So in practice, buying the security that funds the bridge is very different from building the bridge directly. And buying some other, notionally equivalent, securities is more different still.

It's an important theoretical question why this is so. But from a policy perspective, the important point is, it is. The normal channel for monetary policy isn't business investment, but housing. As for why that's not likely to work right now ... well, it's not hard to think of reasons, and they have nothing to do with the ZLB.

[JW Mason]

I should add that Keynes' argument in chapter 11 of the GT makes exactly the equivalences I criticize in the second paragraph there. In some sense the convergence between Keynesianism and quasi-monetarism that Nick talks about, goes back right to the beginning.

[Lord]

The opportunity cost is infinite if it is undoable, which with the Tea Party, it is.

[JW Mason]

It's the credible threat to do 1 through 9 that matters, not actually doing 1 through 9. Simply being willing and able to do it if people don't believe you about NGDP. And if they do believe you are willing and able to do it, they will believe in NGDP.

It's not obvious to me that this is the case. Or rather, it's not obvious what additional assumptions are required for it to be the case.

Certainly there are many actors in the economy who face binding liquidity constraints. For them, a change in the expected path of NGDP will have no direct effects on their behavior, while actually doing some of 1 through 9 might.

[Min]

Nick Rowe: "Did the government's net worth change?"

Does it make any sense to talk about the gov't's net worth?

[rsj]

"rsj: "Monetary policy is a balance sheet operation that leaves your net-worth unchanged. It may, however, have some fiscal side-effects.

Fiscal policy is an income operation that increases your net worth (in the case of a payment received from the government) or decreases your net worth (in the case of taxes paid to the government)."

OK, So say the Fed prints $1 billion, lends it to treasury, and Treasury buys a new bridge worth exactly $1 billion, and hands the $1 billion cash over to the people who built the bridge.

Did the government's net worth change? Is this fiscal or monetary?"

When the CB "lends" to Treasury, why wouldn't Treasury just sell bonds?

That's how we do it -- the CB sets the borrowing costs for Treasury with its reaction function, and then Treasury sells bonds to the private sector. It then builds a bridge, paying the workers the amount it borrowed. The difference between the CB lending directly to the treasury and the CB setting the overall risk-free rate, is that the government is not a recipient of a special interest rate that is different from the overall interest rate in the private sector. The CB sets the rates for all, which is how we should do it.

Now the question is, if the CB wants to keep borrowing costs from rising too much when the government announces it will build the bridge, does it need to print $1 Trillion of money or not?

It doesn't, it is able to control interest rates via OMO with minuscule injections of new reserves. You don't need $1 Trillion of base money to fund a $1 Trillion project without raising borrowing rates for the private sector. How much do you need?

Here is a simple example of how I think about it.

Let's suppose that we are in output gap, that inflation is below target, but that the private sector wants only some of those assets -- it wants $500 billion in additional market assets to restore itself to its optimal consumption plan, because there was an asset re-valuation that cut private sector market assets by $500 billion.

Then treasury supplies 1 Trillion of market assets to the private sector, and households turn around and consume more, to get themselves back to their ideal wealth/income ratio. In the course of that spending, prices rise somewhat. As prices rise, banks require more reserves, and then the CB, as it accommodates, supplies those additional reserves.

So in this case, the CB will need to supply *some* additional base money in order for the Treasury to sell $1 Trillion of bonds at the current interest rate.

Now suppose that prices have risen back to the inflation target and the output gap is closed.

Treasury then wants to build another bridge.

Again, it sells the bonds to the private sector, but in this case, as the private sector does not wish to hold these market assets, if the interest rate remained the same, households sell assets to each other to buy consumption. But the (nominal) price of assets does not fall, as this is set by arbitrage with the short rate, so it is the nominal price of consumption that goes up. To prevent this, the CB raises the borrowing rate, and as it does, some private sector investment is no longer viable, and the existing market wealth of households is also reduced. In a perfect world, this happens up until

the reduced market value of pre-existing assets + the federal bonds from the new bridge - the private investment crowded out = desired market wealth

Then there is price stability and ex-post, households are back on their consumption plans, and public borrowing has crowded out private borrowing while the ratio of reserves/GDP remain unchanged.

So this story is perfectly compatible with the notion that the base money supply rises together with prices, but cause and effect are different.

Moreover, it is consistent with the notion of crowding out if there is no output gap, but again via a different casual channel.

Which casual channel is in operation is important for the policy decisions. The role played by each institution is also important.

So it's good to think of the CB as controlling the costs of borrowing for the government, rather than controlling the quantity of money. The central bank can, if it wishes, negative the stimulative effects of fiscal policy by increasing interest rates. But this doesn't happen automatically in a monetary economy. The private sector is not able to shift in such a way so that government borrowing crowds out private borrowing, as real rates fall when households rush to sell their wealth for present consumption. Rather, the CB needs to perform the crowding out as part of its inflation fighting efforts. Crowding out of private investment by government investment is the result of conscious government policy at attempting to maintain price stability. Without this policy, government investment would crowd out private consumption, not private investment, as real rates would be too low, which is effectively a tax on consumption and a subsidy for investment. You see investment bubbles in nations that keep interest rates too low, and the consumption share of GDP declines, while the investment share goes up.

In terms of the net worth of the government changing, that also depends on your model. On the one hand the liabilities of the government have increased, as it has the $1 Trillion debt obligation. On the other hand, output will also go up, which means tax revenues will also increase without the need to raise tax rates. I.e. disposable income can rise even if the government's borrowing rate is not lower than the growth rate of the economy. Whether or not this happens depends on whether the fiscal policy succeeds in increasing national income or not.

If you *assume* that fiscal policy cannot increase national income, then you can make a ricardian argument. But you can't use that assumption to then prove that fiscal policy is ineffective, as this is a tautology.

[Nick Rowe]

Leigh: "Is there really a difference between an "excess saving theorist" and someone who believes in "excess hoarding of money"? I think I have been persuaded by you that these two are the same,..."

Hang on. I'm saying they are different. I believe in the Paradox of hoarding, but i don't believe in the Paradox of thrift. Or did I misunderstand you?

JW: well, actually, thinking of interest rates as the (sole, or main) transmission mechanism isn't really quasi-monetarist. Indeed, in full equilibrium, a doubling of the money supply should double all prices and leave interest rates the same. A credible commitment to loosen future monetary policy should probably raise interest rates, as investment demand responds to higher expected future demand.

(But I still find it puzzling why econometricians finds to little empirical evidence. Not that I know that literature. Probably reverse causality? Plus the main channel through which lowered interest rates would increase Id is through increasing stock prices?)

"Certainly there are many actors in the economy who face binding liquidity constraints. For them, a change in the expected path of NGDP will have no direct effects on their behavior, while actually doing some of 1 through 9 might."

Wouldn't an increase in expected future NGDP relax those liquidity constraints? I'm more likely to be willing to lend to someone if I think he's less likely to be unemployed in future.

Min: "Does it make any sense to talk about the gov't's net worth?"

Well, not really. Only if we hold a lot of things constant.

[Nick Rowe]

rsj: "If you *assume* that fiscal policy cannot increase national income, then you can make a ricardian argument. But you can't use that assumption to then prove that fiscal policy is ineffective, as this is a tautology."

hey, I was just going to make the same argument against you! I was waiting for you to say that if policy increased private sector net wealth, then it was fiscal policy. And if it didn't, it was monetary policy. Which means fiscal policy is by definition effective and monetary policy is by definition ineffective!

[rsj]

Yes -- and I am not saying that fiscal policy must always increase private sector wealth.

Sometimes it does and sometimes it doesn't. I assumed that it did in my example. But I can make a case -- a plausible case -- for why it would in the current environment.

I can also make a case the revenue-neutral re-distribution will increase output as well.

[rsj]

"Wouldn't an increase in expected future NGDP relax those liquidity constraints? I'm more likely to be willing to lend to someone if I think he's less likely to be unemployed in future."

No. The credit constraints faced by households have two sources. The first is prudence -- households will choose *not* to borrow if their market wealth is zero, because in that case there is a risk that next period they will have no income.

Imagine that there is a p% chance that you will receive zero income in each period. Then you will never borrow if you have zero market wealth and live for only a finite number of periods.

now take the limit as p --> 0. Still, you will not borrow any amount. I think this is similar to your robustness argument, in the sense that a perfect foresight model, in which households have zero income uncertainty, is not the limiting case of a ratex model, in which households face a positive probability of income uncertainty, even if you take the limit as the standard deviation of income --> 0.

Similarly, an infinitely lived household's consumption problem does not have the same solution as the limiting case of a finitely lived consumption problem if there is non-trivial additive income risk. The solutions are the same if you take the limit of wealth --> infinity.

This is similar to your robustness argument. The notion that consumption is set by preferences, rather than constrained by income is not a robust belief. Add just a little bit of income uncertainty and finite lived agents, and the resulting consumption solution looks a lot more Keynesian. Throw in large levels of inequality, so that the majority of households are on the steep portion of their consumption function each period, and you get something extremely Keynesian.

I think, as an aside, that one reason why the Phillips curve relation began to break down was that we had low levels of wealth inequality in the 1970s, and most people were on the portion of their consumption function that looked closer to the perfect foresight model (as their market wealth/income ratio was much higher).

The second source is that human capital is valued differently, in terms of collateral quality, than market wealth. The difference in interest rates between a collateralized loan and a credit card loan dwarfs any differences in "general economic outlook".

[JW Mason]

thinking of interest rates as the (sole, or main) transmission mechanism isn't really quasi-monetarist. Indeed, in full equilibrium, a doubling of the money supply should double all prices and leave interest rates the same. A credible commitment to loosen future monetary policy should probably raise interest rates, as investment demand responds to higher expected future demand.

Here we may have a continued difference between quasi-monetarism and Keynsianism, then? On the Keynesian side, we think of the interest rate as equating the supply of money with liquidity demand. So in general we would expect part of any increase in the money stock to be reflected in higher prices, and part to be reflected in lower interest rates as money holdings increase relative to income. But haven't you been known to use the IS-LM model yourself?

I still find it puzzling why econometricians finds to little empirical evidence. Not that I know that literature. Probably reverse causality? Plus the main channel through which lowered interest rates would increase Id is through increasing stock prices?

The Minskyan answer, which I personally find persuasive, is that external financing involves substantial additional costs to both the lender and the borrower. Firms are also typically rationed in their borrowing. So the variation that matters, in terms of financing investment, is in cashflow or internal funds. Steve Fazzari and various coauthors have a bunch of papers exploring this argument empirically. In any case, firms only finance a vanishingly small portion of investment through stock issues.

Wouldn't an increase in expected future NGDP relax those liquidity constraints?

Well, maybe. But maybe not, or not very much. Here we're getting into all sorts of behavioral and institutional questions. It's a much bigger step than I think you're making it out to be.

[Nick Rowe]

JW: "But haven't you been known to use the IS-LM model yourself?"

Yep. But ISLM is open to more than one interpretation of the transmission mechanism. Maybe we reverse the normal story. IS determines r, and LM determines Y! So, an increase in (say) G shifts the IS curve, and increases r at a given level of Y, which increases V, which creates an excess supply of money, which causes Y to increase.

[Nick Rowe]

rsj: people get offered loans, and accept those loans, even when both lender and borrower know there's a positive probability the loan will not be repaid.

[JW Mason]

rsj-

That's a really interesting comment (9:57 PM). Can you point to any good articles/authors that develop the argument?

[rsj]

"people get offered loans, and accept those loans, even when both lender and borrower know there's a positive probability the loan will not be repaid."

Of course. But households do not generally borrow to fund consumption in excess of their market wealth. They get loans to buy a house or a car or some other asset that is bought with the loan, and there needs to be some form of downpayment, which means positive market wealth is needed. When the buy an asset with a loan, they are not reaching a state of insolvency, where their market wealth is negative.

The loans to fund consumption are credit card loans. Credit card interest rates come with hard limits and are substantially higher than interest rates for mortgages.

The only difference in allowing credit cards is that without them, the consumption function in terms of wealth/permanent income each period is a curve that starts at the origin and is very steep, eventually flattening out as wealth/income --> infinity. Add some credit cards with hard limits and the whole curve shifts a bit to the left. It is still highly concave.

Effectively households are still credit constrained, whether by their own prudence, or by the underwriting prudence of their lenders, because their ability to borrow unsecured against uncertain future labor income is much lower than their ability to borrow to purchase a house with a large downpayment.

It also means that the distribution of wealth is an important variable for determining the effects of fiscal policy.

Ironically, it is the redistribution and income guarantees of the Keynesian era that ultimately made fiscal policy much less effective, because the effect of these policies to make the consumption function less credit constrained and more dependent on preferences.

But by the same token, the ratex models that assume each agent is not credit constrained and is wealthy also become less effective at describing the world when their prescriptions are adopted and income guarantees and redistributive policies are rolled back.

[rsj]

JW Mason,

Look at "A theory of the consumption function, with and without liquidity constraints", JEP 2001

available here

Note that he says a lot of nice things about Freidman, but not so many nice things about how Friedman is currently interpreted. In Friedman's original formulation, permanent income was to be discounted at a rate of about 30%, and was defined as the average income over the next 3 years or so. That has been blown away by infinitely lived agents discounting at the risk free rate.

See also “Consumption Growth Parallels Income Growth: Some New Evidence.” Christopher D. Carroll and Lawrence H. Summers

[Steve Waldman]

Nick — You sure do draw a shady crowd here. (And by shady, I mean "excellent"!)

A few things:

1) With respect to frictions, matters of degree can become matters of kind. There is a sense in which, if I blow to the north-east, after a few hours you should feel a breeze. But as a practical matter, that simply won't happen. The frictions simply overwhelm the force applied.

2) Of course, there is some energy of huffing and puffing that would, in fact, ruffle your hair. If is far beyond the capacity of my lungs, but if we posit that the cost of incremental wind energy is literally zero, then I can make any desired degree of mess of your coiffure at zero cost. Knowing this (and how much I hate a tidy dew), you won't even bother to comb your hair and I won't have to go through the trouble.

3) I don't think it's at all realistic to say that the cost of monetary interventions is zero. (Even though, in purely fiscal terms, the cost is often negative!) If there is any cost at all, then severe frictions may render the operation impractical (which is why your hair is safe, despite the availability of cheap electric fans!) As a practical matter, I think diverging estimates of both the costs of monetary policy and the degree of friction in transmission largely determine which side of the modern demand-side family one ends up on, quasi-monetary or fiscal-ish. If the transmission mechanism is sticky, so that it takes a very great deal of intervention in the Treasury market to get just a bit of output gap undone, that has effects on the distribution of wealth and on patterns of investment that may not be desirable. To the degree the supply of real activity is inelastic to interventions, their effect falls on the price of assets, creating "regressive" transfers of wealth and income for incumbent holders of financial assets. It is a matter of controversy, both on normative and positive economic grounds, whether such transfers are harmful, but if you think that they are (I do), then exortations to apply sufficient force, never mind the side-effects, won't be very persuasive. Of course, to the degree the interventions will be reversed (which is usually in practice incomplete, especially if one includes the spreads paid on the round-trip of transactions), you might argue that these wealth transfers will be reversed. But that's not quite right, because some incumbent holders of assets are cleverer about tracking the asset-price implications of monetary policy than others, and the losses will be distributed to different people than the gains. You might argue that's what the investment casino is all about, distributing profits from the more clever to the less clever in the service of the general welfare, but if investors profit more reliably by tracking macro stabilization rather than by crafting and evaluating valuable real economic projects, that creates a severe opportunity cost in terms of information work not done. When clever central bank Kremlinologists profit from expansions and pass losses onto passive holders of index funds and the beneficiaries of agency-cost-riddled pension funds, that strikes me as costly. Further, the frictions are not themselves uniformly distributed. Expansionary monetary policy has an easier time inspiring some sorts of real activity than others. In the US at least, the channel through which monetary expansions most reliably provokes real expansion is famously housing. Whatever the particulars, if sectoral biases are strong, then distortion of activity adds to the cost of monetary policy. (See also this piece by Ashwin @ macroresilience on how monetary stabilization affects the incentives of banks.)

Of course there are lots of costs to expansionary fiscal policy as well, in terms of distortion and undesirable transfers. I don't think we have any good means of quantifying and comparing the costs of fiscal vs monetary interventions, and I'm not optimistic that further research will satisfactorily resolve the issue in such a cloudy and ideologically charged space. I think the right approach going forward is to deemphasize the monetary/fiscal feud and to try to design specific interventions and institutional contexts that maximize benefits and minimize costs regardless of their conventional categories. My preferred option is to have something institutionally very much like a central bank use variations in a rate of flat transfers to target NGDP, because I think that minimizes normative "fairness" problems and allows widely dispersed demand to determine the shape of new investment. But that is certainly not a perfect proposal, as absent tax authority there is a zero lower bound on transfers that may prevent sufficient contraction. I'm open to lots of things, regardless of whether you call it fiscal or monetary. But I remain skeptical that swapping money for government securities in any quantity will do the job at tolerable costs. If my skepticism is well-founded (or at least widely shared), we can't rely on expectations to do the job, at least not initially. Central banks will have to put up or shut up and prove that they can meet their targets before market participants like me back down from testing their resolve.

[Min]

Nick Rowe: "Richard Koo's theory is that a balance sheet recession is when creditors stop lending, and borrowers stop borrowing, and this (somehow) causes an increase in aggregate desired saving and this (somehow) causes a recession."

Gee, if lenders stop lending and borrowers stop borrowing, won't the supply of private money drop? (Forget about desired saving.) And in general, won't that in itself tend to a recession? Thanks. :)

[Luis Enrique]

Thanks K, I understand. At then same time, many people are worried the government is carrying too much debt, so I don't understand why the potential ability to write off some of it, is given no weight. If you combine that with the idea that if building bridges is what the economy needs, and the government is better able to build them than a pessimistic private sector, no matter how much money is shoved down their throats, that could add up to an argument for conducting monetary policy the Krugman way ,i.e. via the government.

[Kien]

Dear Nick, I haven't gone through the comments and while I've been doing my best to follow your posts, I may have missed some of your writing and just plain misunderstand you. With that qualifier and apology, let me make the following points ...

I don't think Prof Krugman is saying fiscal expenditure is the solution. As I understand his writings (e, fiscal expenditure will help ameliorate the downturn and lessen unemployment. It will lessen the short term adverse impact (with longer-term consequences) on individuals, particularly those without work.

The cure, as I understand Prof Krugman's 1998 paper on the liquidity trap, lies with getting people to believe that there will be inflation in future, and therefore enabling the negative real intest rate that the economy "wants".

On your very nice argument about the substitutability of government building a bridge and the private sector building the bridge, isn't the problem that the private sector doesn't want to take the risk, especially with deflationary expectations in mind. If the private sector expects deflation, it won't invest. But the government can invest and take that risk. Also, by investing, the government could help signal its commitment to inflation, and therefore help address the problem of deflationary expectations.

Nick, you are much smarter than me. So tell me why I am wrong. But please don't forget the short-term costs, especially the impact on the unemployed. I fear some economists ignore the immediate human impact in their pursuit of long-term outcomes.

Best wishes, Kien

[acarraro]

I think that the difference between Quasi(Market)-Monetarist and Krugman is Japan.

Japan formed a lot of Krugman's view on this issue. He believes they tried (and failed) monetary policy in a similar situation. He repeated that more than once. The BOJ has been doing QE for 10 years and still we have no sign of inflation or higher NGDP growth.

I think Scott Sumner argued in the past that BOJ never really tried enough (they did withdraw some reserves at some point) and that the market still believes it's all temporary (even if it has been going on for 10 years now). I have some sympathy for that view, but I can't shake the feeling that there is a problem we still don't understand there...

In fairness Japan seems to prove that wasteful fiscal policy has no much hope either (given 220% debt-to-gdp). Not sure where that leaves us.

[Nick Rowe]

Kien:

1. Paul Krugman wants to do more than just make life a little easier for the unemployed. He wants to increase Aggregate demand to escape the recession and get unemployment back down to normal. He's right.

2. Paul focusses more on increasing expected inflation to increase AD. Quasi Monetarists say OK, but increases in expected real GDP growth can also increase AD. Lets roll those two effects together and say that increasing expected inflation+expected real growth=expected nominal GDP would increase AD.

3. Paul draws a downward sloping IS curve. This means that equilibrium real interest rates would need to be lower in a recovery. I draw an upward-sloping IS, so that real interest rates can actually be higher in a recovery.

[Nick Rowe]

Steve: It's a rough bunch here, at times. They give me a good workout. I can't keep up with them all!

In principle, you could give the central bank special tax/transfer authority, and run macroeconomic stabilisation by helicopter/vacuum cleaner operations. My mind boggles at the practical and political difficulties though. The helicopter is not so bad. The Bank of Canada just mails everyone a cheque for $1,000 in base money. Sure, it might miss a few. But how could the Bank of Canada operate a vacuum cleaner, and make everyone mail it a cheque for $1,000?

[Nemi]

I don’t understand. How is it not obvious that Krugman would object?

What I assume Krugman would think:

1: Fine – that’s expansionary.

2-5: No, it does not matter whether the construction is done by the state or a private firm.

6-8: WTF?

9: ahh – you assume that the amount of personal recourses that a person will put into real investments is fixed. That is nonsense. If they get money and don´t see any profitable investment opportunities they will simply save the cash (or, at best, increase the prices of financial assets).

Or is your point that these actions somehow would create inflation (why?) and thus change the payoff of real investments (relative to holding cash)?

[Nick Rowe]

Nemi: what's the difference between the government buying 100% of 1 new bridge and buying 50% of 2 new bridges (5 and 6)? "WTF?" is not really an answer.

[Nemi]

Because, at least as I read it, up until point 5 it was implicitly assumed that the state initiated the project and paid someone for it.

After 5, at least as I read it, the state would co finance a project that the private sector had initiated (why else would they want to put up with 50 %).

The “WTF” was a shorthand for “what does co financing of privately initiated (and thus profitable without considerations to multipliers etc.) investments has to do with stimulation of AD”.

At 9 he would understand why you think so. But, as I read Krugman, he would not agree.

As I read Krugman – he do not think there exist a lot of private investment opportunities. If the state paid half of the investment, the investor would simply end up with more cash than otherwise. It wouldn’t lower interest rates (already zero) – and thus there is not a mechanism by which this extra cash in the hands of the private sector would increase investments.

Maybe I misunderstand what you wrote (or what Krugman writes, or both of you)

[Winslow R.]

I hit a mental roadblock when ever you mention it would be a good idea for the Fed to buy the S&P 500.

I was permanently banned from Brad Delong's site for, what I thought was a perfectly reasonable response to a 'Nick post' on this topic. What are you two thinking, the S&P is unscammable due to the forceful oversight of the SEC?

Perhaps the banning was also due, in part, to my constant alignment of Brad with the mistaken policies of Larry Summers and Bob Rubin. Nevertheless, the whole idea stinks of a scam, and an endorsement of the folly of recent history (last 30 or so years).

Though I find Nick's position exasperating, the post comments were well worth the read.

[Greg Ransom]

There is a singularly significant set of FALSE assumptions snuck in here:

"If barter were costless (it obviously isn't) it would be very easy to escape a recession. Unemployed workers would simply barter the labour they cannot sell for the extra goods they could produce but which their potential employers cannot sell. And then the newly-employed bakers could barter their bread-wages for the beer-wages of newly-employed brewers."

The evenly rotating perfect knowledge / perfect equilibrium construction of the "barter" economy rules out by assumption the distortions in the structure of production (e.g. housing production generated by false interest rates, leverage, bandwagon optimism & non-transparency in financial markets & across the veil of time) and consumption (e.g. luxury goods financed by 2nd mortgages) which make up the core of the boom and bust cycle.

Recessions are escaped in this world because by assumption they have been ruled out at the start by assumption.

Assume a DISCOORDINATION barter environment where prices are massively out of line -- and many goods have completely lost their economic status (e.g. hundreds of miles lumber hauling rail cars once used to build un-economic housing).

THEN WHAT?!?

In this world you simple "instant" barter solution in non-sense. Some types of labor are now NON-ECONOMIC GOODS. Some production goods are now NON-ECONOMIC GOODS, or goods which cost more to use than the value of what they produce (e.g. the houses and casinos that have been bulldozed in America, post 2007).

Because prior exchange ratios are now meaningless -- they were revealed to be FALSE signals creating a system of totally incompatible economic plans -- no one will know what to barter for what.

The "simple" solution (just barter) is actually not even available -- no one know what to trade for what or with whom.

In other words, the socialist calculation problem / knowledge problem is directly relevant to the problem of thinking about the discoordination problem of a "recession" in the so-called "barter economy".

It shows us that the "barter economy" doesn't give us a benchmark at all of how re-coordination works in a recessionary disequilibrium situation -- i.e the false idea of just plugging in the "correct" dose of new money to re-equilibrate the system.

[Jon]

Nick: one of the best phrases I've heard to describe the current situation is "capital strike". Private bridges are not being built. Now Krugman thinks this means we need to build public bridges, but thats because he is not considering why there is a general capital strike going on.

Second, Sumners has been partial right about tight money problems but treats it as the whole story. I don't agree. There is an identification problem here: a capital strike and cash hoarding are similar ideas. The latter is a specific case if the former. Both manifest in is-lm in a similar way but whereas monetary policy expectations can drive us back to equilibrium during a hoarding situation, there is no transmission mechanism in the general case.

I agree that monetary policy had made matters worse at times. So we need to ask ourselves why there is a general capital strike. One factor is precisely the idea that the government will be building bridges or directing large asset purchases. Krugman's remedy rubs salt in the underlying wound.

The US congress has been a serial policy disaster machine since 2006. Many people have forgotten now but one of the early triggers of rumbling in the subprime market came when congress obliterated the student loan industry. This was unintentional as ostensibly the bill was only going after profits but the entire industry basically collapsed overnight. This was right before the school year began... Only a hastily constructed new federal program plugged the gap just in time. So the man on the street forgets this happened but lenders and businesses do not.

The message was clear: if you make money we (the government) will take it.

This keeps happening over and over now. Just recently there was that story about the govt manipulating the oil lease permitting process so that it could renegotiate royalty terms on a recent large find.

[Winslow R.]

Jon wrote: "The message was clear: if you make money we (the government) will take it."

My goodness, how does that idea jive with the reality expressed by government's move 'to privatize gains and socialize losses'?

Your read on the student loan program could use some additional analysis. Start with bankruptcy law, for profit education institutions, tuition costs, and job prospects for graduating students. Perhaps the collapse had something to do with falling returns to private investment?

[Ashwin]

Just chiming in to say that I completely endorse Steve's idea of simple direct transfers as primary monetary stimulus tool. And the vacuum cleaner can be some simple consumption tax like the VAT in Europe. Not saying this is perfect but it is much more neutral than current monetary policy practise which by propping up existing assets ends up being blatantly regressive. The market today rewards an ability to anticipate CB action - an activity that frequently implies unethical insider knowledge and is completely restricted to the elite in the financial community.

[K]

Steve, Ashwin: yup. Citizen's dividend + VAT. Direct. Equal.

[Too Much Fed]

Jon said: "Nick: one of the best phrases I've heard to describe the current situation is "capital strike". Private bridges are not being built. Now Krugman thinks this means we need to build public bridges, but thats because he is not considering why there is a general capital strike going on."

What if there are enough bridges and no more are needed?

[Too Much Fed]

Min said: "Gee, if lenders stop lending and borrowers stop borrowing, won't the supply of private money drop? (Forget about desired saving.) And in general, won't that in itself tend to a recession? Thanks. :)"

Good question. I'd rather phrase it like this. Won't the amount of medium of exchange go down? I believe you need to consider debt defaults too.

[Too Much Fed]

Nick said: "The Bank of Canada just mails everyone a cheque for $1,000 in base money."

Will the $1,000 check have a bond attached (making it debt) or not?

[nemi]

Jon said: "Nick: one of the best phrases I've heard to describe the current situation is "capital strike". Private bridges are not being built. Now Krugman thinks this means we need to build public bridges, but thats because he is not considering why there is a general capital strike going on."
????
I guess that this was ment to be a retoric question but I do not get it. Why is there a capital strike and how does that affect whether the public should build bridges? I hope that you don´t think a lack of private return in a investment somehow mean that it can´t be a good investment for the public - even the old greeks knew better than that.

The private cost of having someone digging a hole is that persons wage. Societys cost, if the person otherwise would have been unemployed, is as a first approximation zero (probably negative if you belive in multipliers, positive if you don´t and belive taxation implies great DWL - but still not as big as the private cost)