« How can we know if government spending is money-financed? | Main | The great tuition debate: A rejoinder »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Nick,

"If real interest rates fall, government spending G should very probably increase, especially government investment spending, on purely micro-theoretic cost-benefit grounds. Because the Net Present Value of a government spending project will probably be higher if real interest rates fall."

What does this even mean? Taking that the non-discounted value of all future tax revenue is infinite and that the government can pull forward tax revenue from any future point in time to fund an investment project, what "cost" is to be considered in a cost-benefit analysis?

Frank: I'm not going to spend my time in these comments teaching micro-theoretic cost benefit analysis. This post isn't even about it. Off-topic. Google "cost benefit analysis", and teach yourself the basics.

Just establishing that the "natural rate" of interest depends on G (and on lots of other things, like the trade balance) is a big step forward.

I think one of the things that handicaps these discussions is the idea -- going back at least to Wicksell -- that the natural rate of interest in the sense of the monetary interest rate consistent with price stability/full employment, is somehow the same as the natural rate of interest in the sense of the intertemporal rate of substitution that would prevail in a Walrasian economy somehow corresponding to the actual economy. In fact the two things have nothing to do with each other (and the latter is, in my opinion, not really a well defined concept at all.) But the confusion of the two notions of "natural" leads people to think that the rate targeted by monetary policy should be determined, at least on average, by the economy's fundamentals, ant by things like the government fiscal position.

should be: .. the economy's fundamentals, not by things like the government fiscal position.

JW: glad to see you correct your typo in the last sentence. Because I didn't understand the original version at all!

I would say that fiscal policy *is* one of the fundamentals. There's government investment, and government saving, too.

In simple models, it is usually dG/dt, rather than G, that affects the natural rate. But yes.

Yep. IIRC, Wicksell at one point defined the natural rate as the rate that would prevail in a barter economy(?). But a real barter economy might still be back in the stone age, where everything is different. Nevertheless, a model of a hypothetical frictionless barter economy might be of *some* use in helping us understand the determinants of the natural rate in a real-world monetary economy. I wouldn't go so far as to say that the two things have *nothing* in common. The frictionless barter (or Walrasian) model's predictions would probably get the sign right.

Well, I don't know. In the discussion of policy, we are talking about the rate of interest as the yield on money loans. And the "natural" rate is the money-loan yield that maintains stable prices and full employment by keeping the demand for current output equal to the economy's productive potential.

In the Walrasian barter world (which Milton Friedman also used in his definition of the natural rate), there is no money, hence no loans of money, and no reason anyone should want it. There is no unemployment or inflation -- in fact the term "inflation" is meaningless in a world without money -- and no possibility of a divergence between aggregate demand and aggregate supply, since there is no difference between them. Furthermore, such a world will have one intertemporal price for every commodity, so there will be a single "interest rate" only if we further assume that relative prices are the same in all periods. I can see how calculating an intertemporal price in such a hypothetical world could be an interesting exercise, but I don't see how it has anything to do with the problems of monetary policy.

now, you can construct a model in which there is only a single commodity, where loans are contracted only to shift consumption between periods, where no other asset except central bank money is available for that purpose, and where everyone knows their lifetime income but where there can be excess demand for output in particular periods. Woodford's models, for example, are carefully designed to have those properties. But these are very special assumptions.

JW: In a monetary exchange economy, there are many real interest rates. The apple real rate is the nominal rate minus the inflation rate on apples. The banana real rate is the nominal rate minus the inflation rate on bananas. Etc.

And in a barter/Walrasian economy there will be the same set of real interest rates. one for apples, one for bananas, etc. People will make loans that are promises to pay future apples, or future bananas, etc.

In a model where money is superneutral, the whole vector of real rates in a monetary exchange economy *may* (or may not) be the same as the vector of real rates in a model of a frictionless barter/Walrasian economy.

There certainly are special assumptions involved, but I don't think the single commodity assumption is one of them. I think that one can be relaxed.

If there are no futures markets in apples and bananas in a monetary economy, then that may change the vector of real interest rates relative to a barter economy. Because we won't be able to hedge risk for relative price changes.

I very much like and agree with this list.

I have a question. You say one option is "Temporarily increase government spending above the micro-theoretic optimum, to raise the natural rate of interest.". What is the mechanism that would cause an increased govt deficit to increase the natural rate ? Is it simply that the govt would compete for funds and drive the natural rates up, or something else ? (I was thinking the govt might engage in unfunded deficit spending and that would tend to push rates down, not up).

MF: Thanks.

"Is it simply that the govt would compete for funds and drive the natural rates up,..."

That's the simplest way to think about it, yes. But it depends on how private consumption/saving and investment choices respond to that increase in G. If it's a purely transitory increase in G, that has no/negligible effect on permanent disposable income at potential output, and if it's not expenditure on goods that are close substitutes for private consumption and investment, desired C, S, and I will be unaffected, so the real interest rate compatible with potential output will rise.

Excellent, I agree 100%.

Thanks Scott. I figured you probably would agree. You have influenced my views on this a lot, of course.

Nick,

I have been enjoying your blog for a while but I haven't commented before. So please bear with me.

I have a stupid question. As I recall the Bank of Canada has two main ways of controlling the money supply: they can change the bank rate or they can buy or sell government bonds. Both of these methods force the BoC to rely on financial markets and institutions to "transmit" its monetary policy. If there is no friction in the system then money created by the BoC flows immediately out into the economy at large and everybody is happy. However, if the economy is in a liquidity trap then this transmission mechanism breaks down. As a result lower interest rates and bond purchases will be ineffective at increasing aggregate demand. Monetary policy becomes ineffective. (There is an income distribution effect here that never gets discussed but should be kept in mind - people make a living off of that friction.)

Can we cut out the middleman here? Can we have the BoC implement monetary policy by making direct deposits into the bank accounts of every Canadian? For example, if we wanted an inflation rate of 2% and the money multiplier were 5 (I have no idea what it actually is) then I figure each Canadian would get around $13 a month.

The system would be structured so that the BoC would be able to quickly turn the taps on and off as the economy needed. Is the economy slowing? Then let's boost the payment to $25 next month. Is it getting too warm? Let's cut next month's payment to $8. Is the economy going into a depression? Let's announce we will pay out $100 for the next 6 months. (Of course, there is a zero lower bound problem here too.)

One benefit of this would be that stimulus would more counter-cyclical than it is now. As we have seen, governments find it very difficult to spend counter-cyclically. They like to cut spending during recessions and the like to increase spending during periods of full employment. (It isn't too hard to understand once one realizes that consumers are also voters.)

So, is this even technically feasible? Of course there are problems. There always are. But would it even work?

Brad: yes it is technically feasible. It's called "helicopter money". It can work. But whether it works or not depends so much on whether it is seen as a permanent or a temporary increase in the money supply. See Scott Sumner's latest post on this.

Thanks Nick. Scott's post was interesting, and worrying.

Nick,

If durable goods exist, then isn't the micro-theoretic optimum of government spending something that looks very closely like active fiscal policy when at/near the ZLB (by conventional policy restrictions)?

In the US, the supposed real interest rate available to the government is negative: it can issue short-term bonds with interest rates lower than both current and targeted inflation. In that case, the micro-theoretic optimum suggests that even if it runs out of other ideas, the government should be purchasing and storing durable goods (such as Twinkies) for later resale and consumption.

From a "Government and CB are one" perspective, this looks something like QE since the net-government is offering to purchase goods in limitless quantity until the real interest rate becomes positive through demand ("crowding out") effects. From the perspective of the real split between the government and CB, each silo is handling its own affairs along existing legal divisions of responsibility.

From the money market perspective, this looks like credibly-permanent QE because new money is issued into the market until demand picks up. From the "real economy" perspective, these actions also directly stimulate demand (or at least bring it forward), which overcomes concerns about transmission effects of currently-implemented, limited-duration QE.

At the same time, this looks like a micro-theoretic optimum because the government is using its fiscal position in a way that will take a profit from an "irrational" market.

Majro: "If durable goods exist, then isn't the micro-theoretic optimum of government spending something that looks very closely like active fiscal policy when at/near the ZLB (by conventional policy restrictions)?"

Yes. It would be hard to tell them apart in practice, unless you did a very careful Cost Benefit Analysis.

You have to be a little bit careful with the stored goods at negative real interest rates example. Because if relative prices are expected to change, you need to subtract the expected inflation rate **on those stored goods** from the nominal rate, to see if you make a profit. (And we need to ask why the private sector wouldn't be doing this already?)

Majormax,

This is baffling to me:

"In the US, the supposed real interest rate available to the government is negative: it can issue short-term bonds with interest rates lower than both current and targeted inflation. In that case, the micro-theoretic optimum suggests that even if it runs out of other ideas, the government should be purchasing and storing durable goods (such as Twinkies) for later resale and consumption."

I would counter that even without a unified central bank / government, the government is not constrained by a micro-theoretic optimum.

See Fisher's Separation Theorem

@Nick:

> Because if relative prices are expected to change, you need to subtract the expected inflation rate **on those stored goods** from the nominal rate, to see if you make a profit.

True, I'm making the unrealistic assumption that we have a single-good (Twinkie) economy. It also works best if we assume price level targeting rather than inflation targeting, since a deflationary period can result in an arbitrary loss position over a finite span of time.

> And we need to ask why the private sector wouldn't be doing this already?

That's back to allegedly Keynes but really Shilling: markets can remain irrational longer than you can remain solvent.

Stockpiling durable goods (or taking a long position in the futures market) increases leverage and risk of ruin if deflationary/ZLB conditions continue. That alone means that private sector actors who carry credit risk may not be able to exploit the opportunity.

@Frank:

> I would counter that even without a unified central bank / government, the government is not constrained by a micro-theoretic optimum.

I don't think anyone has argued that the optimum is a constraint. It's just that acting along with that optimum is both (on average) profitable for the government and monetarily expansionary.

Frank: You are making no sense at all, and just throwing out random stuff to disrupt the comment thread. This has happened too many times.

Stop commenting.

Thanks for an interesting post Nick. These kinds of posts are very helpful to have to refer back to.

"4. The ZLB is only a strictly binding constraint if the central bank runs out of things to buy. "

I was wondering the other day if anyone the largest proportion of total assets in an economy any CB has ever owned? My WAG would be around 10%. Paging Mark Sadowski...

http://blogs.wsj.com/economics/2014/05/28/niall-fergusons-mostly-comforting-history-of-central-bank-balance-sheet-expansions/
https://www.ecbforum.eu/up/artigos-bin_paper_pdf_0551614001400679837-360.pdf

The graph on page 12 is probably surprising to a lot of people.

IIRC total wealth usually around 10x GDP, so 50% of GDP would be about 5% of total wealth.

TallDave: Thanks. Yep, I decided to write down this summary to clear my own head, and so I wouldn't forget what conclusions I had come to!

Good links there. Interesting to see it all in historical perspective.

My guess (just a guess) is that the Bank of England might have owned a largish amount of assets after the Napoleonic wars? But I guess still a very small share of total assets.

"The ZLB is only a strictly binding constraint if the central bank runs out of things to buy."

What about negative interest rates on reserves? One thing that the public (including banks) still own, if the central bank has bought everything, is the monetary base. So provided a central bank can charge banks for holding reserves at its account, it can perform stimulatory concrete steppes even if it has bought everything.

> What about negative interest rates on reserves?

That's tough because banks and citizens can simply stockpile paper money instead. Reserves and cash are supposed to be equivalents, both being part of the money base, but it's impractical to apply a negative interest rate to paper money.

W.Peden: Yep. That's why I put in that bit about 0% interest currency being available and cheap to store. Banks could hold wads of notes in their safes.

Possibly a stupid question, but is it really true that currency and reserves are strict equivalents? If one was going to hold one of the two as an asset, I don't see how it's hard to choose.

Well, only banks have the choice, because normal people can't hold a chequing account at the central bank.

For banks, reserves would dominate cash, because it's easier to make interbank payment with reserves than with cash, and it's easier for robbers to steal cash. But if the quantity of reserves gets very large, and if reserves pay a small negative interest rate, currency might dominate.

So insofar as reserves dominate currency, and the demand for currency vs. goods is less than the demand for reserves vs. goods, this is concrete steppes stimulation, however meagre?

(I'm not trying to press the point vigorously. It just seems to be a naively intuitive point. And obviously a very pedantic one.)

Greg Mankiw wrote a while back about a negative interest rate here:

http://www.nytimes.com/2009/04/19/business/economy/19view.html?_r=0

"Imagine that the Fed were to announce that, a year from today, it would pick a digit from zero to 9 out of a hat. All currency with a serial number ending in that digit would no longer be legal tender. Suddenly, the expected return to holding currency would become negative 10 percent."

W. Peden: I agree with you.

Nick:"Well, only banks have the choice, because normal people can't hold a chequing account at the central bank."
With a debit card, money is transferred automatically between the customer and the store. If they are not using the same bank, the clearing involves tranfering reserves. In the modern world, the deposit side of the banks is only the public brand name of a central bank de facto subsidiary. (If this humble IO guy is not mistaken...)

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad