Redefining "inflation": Austrians, NGDP targets, and my mother

Some Austrians (I can't remember which ones) define "inflation" as "a growing supply of base money". [Update: Jonathan Finegold gently corrects me.] This used to strike me, as it strikes most economists, as a bit daft. OK, I would think, you can define "inflation" that way if you want, but  "rising prices" is what the rest of us mean by "inflation", and it just makes conversation harder.

But after spending a fortnight with my dear old mother on the farm in England, I'm begining to think those Austrians sort of have a point. And it's related to Simon Wren-Lewis' post about the political difficulties of increasing the inflation target to reduce the risk of hitting the Zero Lower Bound on nominal interest rates. And it's also related to arguments Scott Sumner has made about abolishing the concept of "inflation" and targeting NGDP.

Let me try to get my head clear on this:

Mother doesn't like inflation, and she's seen a lot of it in 90 years. And she kept telling me about it. I even suspect that if the Conservatives promised higher inflation than Labour she would switch to Labour. I didn't even try to argue with her, because I knew I would fail. Which confirms Simon's point.

Forget macro/money for a minute, and think about partial equilibrium/micro. Suppose you asked a microeconomist: "Is a rise in the price of apples a good thing or a bad thing?". She would look at you funny. She might say it depends on whether you are a buyer or seller of apples. She might say it depends on what causes the price of apples to rise. But basically she doesn't like the question, and doesn't want to answer it. She doesn't like the question because it's not a well-formulated policy question. If instead you asked "Would a government quota on apple production be a good thing or a bad thing?" she would be much more comfortable giving you an answer.

Back to macro/money. If we define "inflation" the way some Austrians define it (as a growing monetary base) then the question "Is inflation a good thing or a bad thing?" is a well-formulated policy question. Any macroeconomist would feel comfortable answering that question, even though (for all except a tiny minority of macroeconomists who want to target the monetary base) the answer would include a lot of "it depends on whether..." that would make for a very ugly answer. But because nearly all macroeconomists don't want to target the monetary base, and because we want to avoid ugly answers, and because everyone else defines "inflation" differently and we want to make conversation easier, we should not follow that Austrian definition of "inflation". But they do have a point, nevertheless.

When a macroeconomist hears the question "Is inflation a good thing or a bad thing" what the macroeconomist hears is: "Would a change in monetary policy that resulted in a higher rate of inflation, holding constant all those other things that would stay constant if monetary policy changed exogenously, be a good thing or a bad thing?"

That is not what my mother hears when she hears the same question. Mother doesn't spend her time thinking about monetary policy, so of course she won't hear the question the same way. What my mother hears is probably something like: "Would a change in any of those things that would result in a higher rate of inflation, holding constant all those other things that would stay constant if any of those first things changed exogenously, be a good thing or a bad thing?"

Admittedly, macroeconomists are usually a little bit more explicit than my mother about what's in the ceteris paribus clause, but you get the point. More importantly, if Mother did include "holding nominal income constant" in her implicit ceteris paribus clause there would be nothing logically wrong with her doing so.

"Would a change in any of those things that would result in a higher rate of inflation, assuming monetary policy holds the path of nominal income constant, be a good thing or a bad thing?" is a coherent question. And while one could dream up exceptions (everyone approaches Nirvana and gets everything they want without working) macroeconomists hearing this question would then agree with my mother that inflation (or the things that cause inflation) are a bad thing because they reduce our real incomes. Under NGDP targeting, the inflation fallacy is not a fallacy.

There are two ways to change monetary policy to reduce the risk that the Zero Lower Bound on nominal interest rates would be a binding constraint: raise the inflation target; switch from an inflation target to an NGDP level-path target. Draw a graph with "average inflation rate" on one axis and "risk of hitting the ZLB" on the other axis. Both are bads. Draw in your own indifference curves between those two bads. They slope down. Curves to the South-West are better. Now draw the curve showing the feasible trade-off under inflation targeting. It slopes down. The higher the target rate of inflation, the lower the risk of hitting the ZLB. Now draw the curve showing the trade-off under NGDP level-path targeting. It also slopes down. The higher the target rate of NGDP growth, the higher the average rate of inflation, and the lower the risk of hitting the ZLB. But the curve for inflation targeting lies everywhere North-East of the curve for NGDP targeting. That's because: NGDP level-path targeting has stronger automatic stabiliser properties than inflation targeting so nominal interest rates won't need to change as much in response to shocks; and because NGDP targeting will generate higher inflation when real growth is low, which is when the risk of hitting the ZLB is highest, and lower inflation when real growth is high, which is when the risk of hitting the ZLB is lowest. Therefore NGDP level-path targeting is a more efficient monetary framework that every macroeconomist who fears the ZLB but dislikes high inflation otherwise should choose.

Simon could try to convince my mother to move along the inflation targeting curve to a higher inflation target to reduce the risk of hitting the ZLB. He could try to explain to her that higher inflation caused by looser monetary policy (and not by all those other things that Mother thinks may cause inflation) would cause nominal income growth to rise at least in proportion to higher inflation (though in Mother's case it probably wouldn't) and would even rise by more if this reduces the ZLB risk. But Simon knows he would have a difficult job ahead, even though he has never met my mother.

Or he could try to persuade Mother that it would be a good thing if the Bank of England made people's incomes grow at (say) 5% per year, acknowledge that making people's incomes grow any faster might simply cause inflation to rise equally faster too, and happily agree with Mother that inflation really is a bad thing that makes people worse off. And if he added that he wants to do this because it would reduce the risk of her getting 0% interest on her savings account, I think he would get her vote (but maybe he shouldn't mention badgers).

Sometimes, just sometimes, policies that are economically right are also easier to sell politically. This is one of those times.

Comments

[Yancey Ward]

Sounds like the same thing as asking a citizen "Do you think it is good thing to raise your taxes, and give the proceeds to someone else?"

[Frank Restly]

Nick,

"There are two ways to change monetary policy to reduce the risk that the Zero Lower Bound on nominal interest rates would be a binding constraint.."

The zero lower bound on money interest rates is a fallacy of its own. It rests on two absurd assumptions:

Assumption #1: The federal government bears solvency risk (they don't).
Assumption #2: The realized rate of interest paid by the private sector is the same as the stated interest rate on a loan (it isn't).

The simplest way to explain things to dear old mother is to say: Mother you would like to receive a positive rate of interest. For that to happen, someone else must borrow at that positive rate of interest. The federal government would like to satisfy the concerns of both borrower and lender, and so the government through tax policy can either:

1. Collect enough tax revenue to pay the positive rate of interest on it's debt and / or
2. Lower the after tax cost of servicing private debt

On an after tax basis, the cost of servicing private debt can be below zero.

[Ralph Musgrave]

The word inflation 30 years ago or so was actually defined as “an increase in the money supply”. At that time I spotted the change in meaning of the word to something like “rising prices”. But the Oxford Dictionary hadn’t spotted the change. So I wrote to them and told them to change their definition.

When Austrians use the word in the old sense, I suspect they’re into linguistic subterfuge. For the intellectually dishonest, ambiguity offers endless opportunities: you ever known a politician say anything that didn’t have three different possible meanings?

[RPLong]

What does it mean when NGDP is constant and inflation changes? Doesn't that mean that what really happened is RGDP changed? Do proponents of NGDP targeting believe that they are controlling RGDP?

[Alex Godofsky]

@RPLong

What does it mean when NGDP is constant and inflation changes? Doesn't that mean that what really happened is RGDP changed?

Yes.

Do proponents of NGDP targeting believe that they are controlling RGDP?

No, except inasmuch as they are preventing it from from being adversely affected by nominal shocks.

[Nick Rowe]

Ralph: "The word inflation 30 years ago or so was actually defined as “an increase in the money supply”." I didn't know that. Are you sure it wasn't just an Austrian who wrote that bit of the OED?

"When Austrians use the word in the old sense, I suspect they’re into linguistic subterfuge."

Or, they are just being sticklers for tradition, if what you say about the old definition is right.

Or, they are insisting on a definition which makes "Is inflation a good or bad thing?" a well-formulated question about policy.

RPLong: I think Alex's answers are good ones.

Yancey: it's sort of like that, but not exactly like that. "Is increasing taxes a good or bad thing?" does specify a policy variable, so is a well-formulated question in that regard, but it is an incomplete question, because it doesn't say what the government will do with the extra revenue.

Frank: you lost me.

[Jonathan Finegold]

I have to agree with Ralph. If an Austrian switches focus from the CPI (or w/e other index) to base money, it's an attempt to avoid the conclusion that inflation has not been high ("what, look at the amount of base money!"). And yes, I remember three or so years ago priding myself over believing the "original definition of inflation." But, IMO, the original definition (and the "Austrian" [Mises'] definition) is just to encapsulate the theory that inflation is everywhere and always a monetary phenomenon.

[Nick Rowe]

Jonathan: But, IIRC, some Austrians were defining inflation as an increase in money supply (or base money supply) well before the recent recession.

(I updated to add a link to your post BTW.)

[Frank Restly]

Nick,

You borrow $10,0000 and pay 5% annual interest = $500.00 per year (assuming simple interest to keep the math easy). Government refunds you $700.00 per year on your taxes. After tax cost of servicing debt is -$200.00 (negative 2%).

You might counter that it would cost the government $700.00 in lost tax revenue to which I would counter - are you worried about the solvency of the federal government?

[Jonathan Finegold]

You're right, they have. That's how it's defined in Human Action (which means that's how it's defined in Man, Economy, and State), but I think the complete separation of the price level from the money supply is more recent, and is a misunderstanding of the "original" definition. It's no coincidence that the "strict" money supply definition is used by the same Austrians who always and everywhere reject "inflationary" monetary policy.

[Nick Rowe]

Frank: OK, you mean interest subsidies, as a way to get around the ZLB. Suppose I borrow $1 trillion from you at 5%, and you borrow $1 trillion from me at 5%. We both clean up with a very large subsidy, but there's no more effect on aggregate demand than if the government just helicoptered us the same amount of cash. But some economists have proposed a related idea: investment subsidies.

[Frank Restly]

Nick,

"OK, you mean interest subsidies, as a way to get around the ZLB. Suppose I borrow $1 trillion from you at 5%, and you borrow $1 trillion from me at 5%. We both clean up with a very large subsidy, but there's no more effect on aggregate demand than if the government just helicoptered us the same amount of cash."

Not exactly. That is why I brought in tax policy rather than government spending policy. If I borrow $1 trillion from you at 5% and you borrow from me at 5% and the government offers to reduce each of our taxes by 7% x $1 trillion per year, each of us must still have a taxable income to realize that gain. Whereas if the government just gave us an interest subsidy, you would be correct - lots of gain for no pain.

Also there can be a significant effect on aggregate supply. It comes down to the paradox of toil. Suppose I borrow money to fund the production of some good and I decide to produce more than the market is demanding forcing prices down. Normally I would be out of my mind in doing this because I am raising the real rate of interest that I am paying on the loan. However, when the government reduces my after tax cost of servicing that debt I can produce more than is demanded and still remain afloat.

[Kevin Donoghue]

Ralph is correct. From my 1972 Chambers 20th Century Dictionary:
Inflation: undue increase in quantity of money in proportion to buying power, as on an excessive issue of fiduciary money.
Mind you I'm a bit surprised that definition was still in use as late as '72.

[Fran R Woolley]

Nick, enjoyable post. Sounds like you have as many problems selling inflation to your mother as I do selling carbon taxes to my cousin!

[Tom Brown]

In Sumner vs Schiff on CNBC, you saw the Austrian definition on full display:

https://www.youtube.com/watch?v=ZuSIFisFEis

with Schiff saying that high inflation was already here! Just look at the excess reserve levels! Duh! Lol.

[ACEMAXX ANALYTICS]

The inflation rate in Switzerland is for 21 months in a row in the red zone over the previous year: meaning thatt the deflation persists. There are no pricing pressures.

But if the EU had measure price developments in Switzerland with its own method (HICP), the inflation would be higher by 0.5%.

[J.V. Dubois]

Even discussion here is a proof that inflation is highly misleading - even in economic discussion. It is not rare to have discussion with austrian economist who says that inflation is connected to money supply only to hear him later how "deflation in IT" was a good thing. If inflation has anything to do with money supply then one cannot speak of a change in money supply in one particular sector of economy. It does not make sens, right? I suspect that many times conflating different definitions of inflation is done on purpose in a spirit of "end justifies the means". If someone always considers inflation - that part of the price level increase caused by increase in money supply (ignoring velocity) - as bad but not easily explained to public, just scare them by equating inflation with cost of living and call it a day. Hell, even Mario Draghi used it this way with his famous "with low inflation you buy more stuff".

PS: If we are confused even if we didn't touch things like how inflation is actually measured then we are truly lost. I agree with Scott Sumner, inflation is highly misleading, hard to explain and measure. Different definitions of inflation are important for different policies. You want to use core inflation when guiding monetary policy and "inflation in consumptiom basket of seniors" when you want to maintain living standards of pensioners. Just focuse on NGDP.

[Dan Kervick]

What mother is worried about is not the macroeconomic impact, but the effect on Mother. She's probably on a fixed nominal income with few debts but lots of everyday expenses. An inflation will be bad for her.

Others are a more complex situation. They deserve clearer answers from the economists. Are the latter proposing a true general inflation, an increase in the price of everything, including labor? If so, then great. But people then have a right to know how this amazing uniform increase will be engineered.

If on the other hand the economists say that they plan to measure the inflation via the CPI, then there is reason to be wary of the economist's plan, which sounds more like an internal devaluation of labor costs than a general inflation. And working people have been down this road for several decades now.

[Nick Rowe]

So it seems Ralph was indeed right. I'm surprised, because I'm old enough to remember 1972, and I don't remember people defining inflation as anything other than "rising prices". Maybe my memory is bad, or maybe dictionaries were lagging behind popular use of the word (as often happens, given that dictionaries are partly descriptive and partly prescriptive, which is an interesting discussion in its own right).

I do remember people saying "inflation is too much money chasing too few goods", but I always interpreted this as a causal statement, rather than a definition.

Dan: I'm not sure if you understand this or not, so I'm going to assume you don't, just in case:

We aren't talking about what economists are proposing. We are talking about what economists are proposing **for monetary policy**. Monetary policy is just one policy lever (there are others). One lever can only have one target. (If you want two targets you need two levers). You can make that target for the monetary policy lever anything you want (within reason, and it must be a nominal variable measured in $ and not a real variable like employment or real wages, because monetary policy can't hit real targets in the long run as we learned in the 1970's). You could target CPI inflation, OR wage inflation, OR NGDP, OR the price of gold, OR etc, etc. Pick ONE. Or you could pick the AVERAGE of price inflation and wage inflation if you want. But you can't pick BOTH price inflation AND wage inflation.

[Nick Rowe]

Hey, what's happening? My comments aren't going directly into spam any more! Did Typepad just fix itself?? Crosses fingers.

Edit. Damn. This one went straight to spam, but the previous one didn't. How come?

[W. Peden]

Nick Rowe,

My little piece about some pop-Austrians like Peter Schiff is that in 2008 they were saying that the increase in the base was going to lead to inflation (a causal statement) and now they're saying that the increase in the base IS inflation (an identity statement).

Of course, since gold is money, the falling price of gold in dollars recently means that the US is experiencing a severe deflation.

[Frank Restly]

Nick,

If you have two levers available, you can use one lever to hit one target and another lever to hit another. Or you can focus both levers on hitting the same target.

For multiple levers and a single target you would need policy coordination.

[K]

[ OT, sorry...

Nick,

"monetary policy can't hit real targets in the long run as we learned in the 1970's"

Well, it certainly can't hit excessive output targets in the long run. But I think we have an "interesting" experiment going on to see whether if inflation is really downward sticky, it's possible to hit a long run deficient output target. If the medium run is long enough to prevent people from starting businesses, investing in schools and education, and even starting families for maybe a decade or two, I'd think you could do some *serious* long run output damage. Interestingly, this phenomenon doesn't seem symmetric at all on the upside. Even if inflation is super sticky upward, the consequences of over-investment just don't seem that horrible.

/OT]

[Ben Rizzo]

A response - The war of semantics: http://benrizzo.wordpress.com/2013/08/08/the-war-of-semantics/

[Dan Kervick]

Actually Nick, I don't understand that. I don't understand the difference between proposing P and proposing P as the goal of monetary policy. If you tell someone to put on a tie, what does it matter whether you also say that their putting on a tie is an expression of sartorial policy, and so you are only making a sartorial policy proposal? Are you telling them to put on the tie or aren't you?

I don't understand your last point about not targeting wages inflation and consumer price inflation. First of all, if inflation were truly always and everywhere a monetary phenomenon, and if monetary effects can be engineered by monetary policy, then its a bit mysterious why these monetary effects don't cause uniform inflation of all prices concurrently. If there is some change in the ratios of relative prices along with an overall increase in the rate of change of some average of prices, then something else is happening besides monetary phenomena. My understanding is that has always been part of the point, right? Policy economists promoting inflation have sometimes appealed to the idea that inflation combats downward price stickiness and allows relative prices to adjust more rapidly to reestablish some kind of efficiency, no? If that is the ultimate purpose behind a higher inflation or higher NGDP targeting proposal, then people have a right to hear that policy stated in clear terms, without all the central banker gobbledygook and stylized verbal formulas.

Why wouldn't you target increases in the rate of change of CPI and increases in the rate of change of wages at the same time? Assuming such a thing is actually possible, then wouldn't that have the desired effects of reducing the real value of nominally fixed debts and allowing the market to achieve a its negative real natural rate of interest, if such a creature there be?

[Frank Restly]

Dan,

"First of all, if inflation were truly always and everywhere a monetary phenomenon, and if monetary effects can be engineered by monetary policy, then its a bit mysterious why these monetary effects don't cause uniform inflation of all prices concurrently."

It is because monetary policy is not money policy. Monetary policy operates through a credit channel. To get uniform price changes all that you would need to do is redefine the units of currency - $1 old dollar becomes $2 new dollars - and bam!!! All prices change instantaneously.

[Peter K.]

Kervick "Why wouldn't you target increases in the rate of change of CPI and increases in the rate of change of wages at the same time?"

Well Rowe does say "Or you could pick the AVERAGE of price inflation and wage inflation if you want." It's what the Bank of Japan and Abe are currently doing, targeting 2 percent inflation. As Rowe says - and you ignore, you're a champion ignorer and cherry-picker - there are multiple levers. Abe is employing multiple levers, monetary policy included.

[J.V. Dubois]

Dan Kervick: Inflation does not mean "uniform inflation of all prices concurrently". This does not apply to different goods in consumption basket compromising CPI as one the price of one good may increase more than the price of other good. And this surely does not apply to different measures of inflation (like CPI inflation vs GDP deflator). Nick for sure does know that there are real shocks that can affect relative prices of different goods.

Nick's point (I think) is that because central bank has only one tool to steer monetary policy (printing money) there can be only one future path of money printing. Central bank can decide that this money printing path will be steered so that the of gold will remain stable, or that the price of selected goods in some basket will rise 2% a year, or that will make NGDP level growth stable on 5% or it can even be some plain aritmetic expansion (like Friedman rule). But that can always be only one path. There is more on this here: http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/06/no-multiple-own-rates-of-interest-dont-matter-but-the-monetary-target-does-matter.html

Having two targets means that central bank will have to shift from one to another. If this switching has some rule such as if CB is required to follow some weighted average of these targets than we ended up with one target. If central bank has discretionary power to switch between various targets than it stops making sense to talk that it has any target (other than some unknown one that is in the head of Central Bank executives)

So while there seems to be endless possibilities of things that should be used to steer money printing path, which one central bank target? The one that has trouble reaching equilibrium, or in other words the one with more sticky prices.

[Edited to fix typo NR.]

[J.V. Dubois]

Ah, it should be: "If this switching has some rule such as if CB is required to follow some weighted average of these TWO targets than we ended up with ONE target"

[I edited it to fix it. NR]

[Dan Kervick]

Frank,

I have long argued that central bank policy operates almost entirely through the credit channel by means of interest rates. But I was under the impression that Nick disagreed with that.

[Dan Kervick]

Could someone tell me then what the "wage lever" is for policy? Is there a monetary policy operation for that lever? Is it a monetary policy that is under the control of the central bank?

And could I get this straight: Are we to understand the term "inflation" to be a synonym for "rate of increase in the CPI?"

And J. V. Dubois, if it is true that central bank policy works entirely by printing money, then how can the effects of that policy differentiate between wages and consumer goods. We use money to pay for both labor and consumer goods. So how can money printing target the rate of change in the weighted average price of a basket of commodities, without at the same time targeting the rate of change in every other price.

This was Friedman in his Nobel prize lecture:

Only surprises matter. If everyone anticipated that prices would rise at, say, 20 percent a year, then this anticipation would be embodied in future wage (and other) contracts, real wages would then behave precisely as they would if everyone anticipated no price rise, and there would be no reason for the 20 percent rate of inflation to be associated with a different level of unemployment than a zero rate.

[J.V. Dubois]

Dan: How can money printing change the price of CPI goods and not impact other prices (like prices of investment goods) with the same magnitude? In the same way that money printing can be consistent with the price of one good in the CPI basket rising and another one lowering - by relative price changes. Relative price changes are in turn caused by real shocks, be it positive (like new technology) or negative (like bad harvest).

If FED would promise to stabilize the the price of floating point operations per second on computers we would have different monetary policy (money printing path) compared to the one where FED promises to stabilize the price of gold.

So while (permanent) money printing affects all prices uniformly, using different targets to define how much money is to be printed obviously has different effect on how much money gets printed and therefore on inflation.

PS: I also thing that you are using the sentence "inflation is always and everywhere monetary phenomenon" in a strange way - which by the way stems from different definitions of what inflation is. For instance the Warsaw ghetto economy during WWII went through "hyperinflation" - if by hyperinglation one means that the prices experienced 3 digit increases - purely for supply side reasons. So if inflation is "too much money chasing too few goods" than the quantity of goods has some say in it too.

[Dan Kervick]

J.V., I understand that relative prices can change in conjunction with money printing, but then I'm wondering what sense it can make to target consumer prices and not wages, or wages and not consumer prices. Where are the alleged levers? I suppose the Fed can say "we would like (or expect, or whatever) the CPI to be such and such." They could make similar pronouncements about wages. But, you know, big deal.

Of course the Fed doesn't just print money in some generic sense. It emits dollars in electronic form to buy financial assets, mainly from financial institutions, which affects interest rates in the market for those assets but causes little change in the overall private sector holdings of financial assets. So the whole model of the Fed releasing money into "the economy" do its textbook business of chasing up prices seems bogus.

[Nick Rowe]

Basic points:

1. If the only thing that mattered was inflation, it wouldn't matter whether we used monetary policy, fiscal policy, Diocletian-style price controls, or whatever, to hit an inflation target. But other things do matter. If all I care about is that you wear a tie, it doesn't matter who knots your tie. But if everyone else is busy doing other things that matter too, I want you to be the one who knots your tie.

2. If you have one instrument you can only hit (at most) one target. Unless by sheer fluke, two targets line up exactly. It would only be by sheer fluke that monetary policy could target (say) 2% price inflation and (say) 3% wage inflation at the same time. Because there are lots of other shocks hitting the economy that will affect price inflation differently to how they affect wage inflation. For example, if a shock affected equilibrium real wages, so that wages started to rise at 4%, but prices only rose at 1%, should the central bank tighten or loosen monetary policy? It can't do both.

3. Most economists believe that in the short run monetary policy will have real effects. We don't believe money is "short run neutral". That's the most important reason why monetary policy matters. That's why we argue about whether inflation targeting is better than NGDP targeting or money supply targeting or gold price targeting or exchange rate targeting etc. It is possible that in the short run monetary policy will affect price inflation differently from wage inflation, which is a real effect, because it affects real wages. (This is quite distinct from all the other shocks that are hitting the economy that may affect price inflation differently from wage inflation.) For example, if wages are stickier than prices, monetary policy may have a bigger effect on price inflation than on wage inflation in the short run.

4. Most economists believe that monetary policy may have some long run real effects too, in the sense that targeting 4% inflation may be better or worse than targeting 2% inflation (or targeting 7% NGDP growth may be better or worse than targeting 5% NGDP growth). We don't believe that money is strictly "long run superneutral". Simon Wren-Lewis believes targeting (I think) 4% inflation would be better than targeting 2% inflation. He's probably right. But I think that targeting 5% NGDP would be better still. And that's what we are discussing here, in this blog post.

[Frank Restly]

Nick,

"If the only thing that mattered was inflation, it wouldn't matter whether we used monetary policy, fiscal policy, Diocletian-style price controls, or whatever, to hit an inflation target."

But success in hitting that target would depend a lot on which is used. Monetary policy is not truly independent of fiscal policy. In as much as a central bank is lender of last resort, a government is a borrower of last resort. Monetary policy gets no traction at any interest rate if no one borrows money.

"We don't believe that money is strictly "long run super neutral". Simon Wren-Lewis believes targeting (I think) 4% inflation would be better than targeting 2% inflation. He's probably right. But I think that targeting 5% NGDP would be better still.

The first question is why is 4% inflation better than 2% inflation? Picture a nation redefining the units on its currency. One old dollar is worth $1.04 new dollars (or $1.08 or $1.20 or $2.00). Picture also that all existing contracts (labor, debt, etc.) are rewritten based upon the new currency denomination. Wages go up 4%, prices go up 4%, debt levels go up 4% - what has really changed? In this case money really is long run super neutral, changes in denomination will have absolutely no affect on real economic growth.

[Dan Kervick]

If you have one instrument you can only hit (at most) one target.

This seems like a very vague statement. Is the CPI one target or an artificial jerry-rigged combination of multiple possible price targets on multiple distinct goods, slammed together with artificial weighting formulas to create the fiction of a single thing?

I think its a little bit tendentious to use the term "prices" in a generic sense as opposed to "wages", when what is actually meant is "prices of consumer goods." After all, a wage is a price too. Why is it treated as as something distinct from other prices.

But anyway, I think I have the answer to my question. And that is that when economists use the term "inflation" in 2013, they almost always mean the rate of change in the consumer price index, and so when they argue for targeting higher inflation, what they are talking about is targeting a higher rate of change in the prices of consumer goods.

And in that case, I think mother is right to be suspicious.

[Nick Rowe]

Dan:

1. Even if the CPI were (it isn't)"...an artificial jerry-rigged combination of multiple possible price targets on multiple distinct goods, slammed together with artificial weighting formulas to create the fiction of a single thing..." that would indeed be *one* target. It's a scalar, not a vector. It's one number, not several numbers.

2. "And that is that when economists use the term "inflation" in 2013, they almost always mean the rate of change in the consumer price index, and so when they argue for targeting higher inflation, what they are talking about is targeting a higher rate of change in the prices of consumer goods."

Yep. Sometimes we do say "CPI inflation", just to be explicit. But when we say "the Bank of Canada targets 2% inflation" we mean CPI inflation, because that's what the Bank of Canada does in fact target, and everyone knows this.

3. "Could someone tell me then what the "wage lever" is for policy? Is there a monetary policy operation for that lever? Is it a monetary policy that is under the control of the central bank?"

I'm not quite sure what you mean by this, but let me try this: If the Bank of Canada wanted to, it could stop targeting 2% CPI inflation and instead target (say) 3% nominal wage inflation in (almost) exactly the same way that it currently targets 2% CPI inflation. No problems (in principle) doing that at all. Slight change in the models they use, and maybe (or maybe not) some political pushback, but if you told the nerdy guys at the Bank to do this instead of targeting 2% CPI inflation they would shrug their shoulders and start doing it. Or 1% or 2% or 4%, or 5%, or whatever. (They might worry about the ZLB problem if you made it too low a number, is all.) Some (e.g. Scott Sumner) might prefer this to targeting CPI inflation. Maybe you would too?

And yes, some economists (I am one) think that central banks should target something other than 2% CPI inflation. My preference would be 5% NGDP level-path growth. That is (partly) what this blog post is about. What would you prefer? 2% CPI inflation? 0% CPI inflation? 3% wage inflation? There are loads of other possibilities.

[Peter N]

Dan:

Imagine a collection of compound pendulums to each of which is attached a solenoid. All the solenoids are wired to fire at once. You can target the motion of a particular pendulum by timing the firing. The behavior of the other pendulums under this regime is undefined, absent equations of motion for them (perhaps other than statistical properties).

So you can target motion by when and how hard without specifically targeting where force is applied. If this sounds rather crude, it is.

Of course a central bank has more than one lever. Unfortunately, there is no agreed upon theory of how to use multiple levers in combination. The hope is that having a single target will create such a strong expectations channel, that the details won't matter that much.

A more complicated model might damage the expectations channel and do more harm than good. I don't believe this, myself, but it's a respectable argument.

[Andy Harless]

I love this post, but I want to point out one thing:

"There are two ways to change monetary policy to reduce the risk that the Zero Lower Bound on nominal interest rates would be a binding constraint: raise the inflation target; switch from an inflation target to an NGDP level-path target. "

There is (some, such as Miles Kimball, argue) a third way, which is to remove the constraint altogether. Basically, this means tricking people into using a medium of account that can't be physically hoarded, so that there would be nothing to prevent people from paying for the privilege of holding it by accepting a negative interest rate. Miles thinks we can do this by breaking the fixed exchange rate between paper money and electronic money, in such a way that the medium of account status would fall on electronic money, and paper money would depreciate sufficiently quickly that the return to holding it would always be lower than the return to holding electronic money.

[Nick Rowe]

Thanks Andy! You are right of course. I only considered two of the ways.