Fish don't feel the water they swim in. (I've heard that's false, but it's too good to check). Paul Krugman does not feel the Monetarist hegemony he swims in. The only part of Monetarism that he does feel, and that he now identifies with Monetarism, is that one small part of Monetarism that failed to achieve hegemonic status. Milton Friedman in 1970's believed that M2 growth ought to be kept small and constant – the k% rule. Milton Friedman lost that battle. But he won every other battle. He won the war.
Here's Paul:
"I’ve always considered monetarism to be, in effect, an attempt to assuage conservative political prejudices without denying macroeconomic realities. What Friedman was saying was, in effect, yes, we need policy to stabilize the economy – but we can make that policy technical and largely mechanical, we can cordon it off from everything else. Just tell the central bank to stabilize M2, and aside from that, let freedom ring!
When monetarism failed – fighting words, but you know, it really did — it was replaced by the cult of the independent central bank. Put a bunch of bankerly men in charge of the monetary base, insulate them from political pressure, and let them deal with the business cycle; meanwhile, everything else can be conducted on free-market principles." (my bold).
(I'm picking on one small part of his post, by the way, and the rest of it is well worth reading.)
There's a lot more to Monetarism than the k% rule.
Paul Krugman, like nearly all of us, forgets all the battles that Milton Friedman won. "That's not Monetarism, that's just normal economics; everyone believes that!" We don't feel the Monetarist water we swim in. Only the Post Keynesians, who took a different path in the 1970's, can feel the Monetarist water, because they don't spend their whole lives swimming in it.
As a Post Keynesian blogger (help me out someone, because my memory has failed) recently said, the canonical New Keynesian model is Monetarist. Well, it isn't exactly Monetarist, but it's very close. A Keynesian Rip Van Winkle who fell asleep in 1970 would certainly see it as much more Monetarist than Keynesian. And if he woke up and saw independent central banks targeting inflation, with flexible exchange rates, no price and wage controls, and fiscal policy relegated to the sidelines for emergency use only, he would know that Milton Friedman had won his war.
The canonical New Keynesian macro model has three equations:
1. An expectations-augmented Phillips Curve with long-run neutrality at the natural rate of unemployment/output and short-run non-neutrality. Pure Friedman. The vertical (or near vertical) long-run Phillips Curve was once controversial. Now it's the standard benchmark. It's part of the water.
2. An Euler-equation IS curve. This is nothing more than a way to translate Milton Friedman's Permanent Income Hypothesis into math. Unlike the Old Keynesian IS curve, current income does not appear as a constraint determining current consumption demand. Current consumption is part of an intertemporal optimisation plan against an infinite-horizon budget constraint. Again, it's the standard benchmark. It's part of the water. When you depart from this benchmark, as Paul Krugman and Gauti Eggertsson did in their recent paper, by introducing borrowing constrained consumers whose consumption depends on current income, that very feature is what makes your model stand out from the benchmark of Milton Friedman's hegemonic Permanent Income Hypothesis.
So far, it's pure Milton Friedman all the way.
3. Some sort of Taylor Rule reaction function for monetary policy. OK. Here's it's not pure Milton Friedman. It's not a k% rule for the money supply. But it's still very Monetarist in spirit. First because it is a monetary policy reaction function that closes the model, and not a fiscal policy reaction function. The model assumes that Milton Friedman has won the war: that monetary policy is in charge of aggregate demand, and that fiscal policy is relegated to handling micro stuff and so can be ignored. Which is what Milton Friedman wanted. Second, because that reaction function is built around a target level of inflation. That is one way to give expression to Milton Friedman's dictum "Inflation is always and everywhere a monetary phenomenon". It means that inflation is the responsibility of the monetary authority. It's not something you try to control with fiscal policy or wage/price controls, while monetary policy targets something else.
Looking at those three equations, I call it 2.5 out of 3 for Monetarism. From the standpoint of the 1970's, the canonical New Keynesian model is five-sixths Monetarist.
(Of course, it all looks different from the standpoint of today, and quasi-monetarists like me notice other ways in which the canonical New Keynesian model fails to do explicit justice to the use of money as a medium of exchange and nominal anchor, but that is by-the-by. And this leaves out the monopolistic competition underpinnings as well. The canonical New Keynesian model is the child of Milton Friedman and Joan Robinson, raised by its paternal grandfather, Knut Wicksell.)
Worthwhile Canadian Initiative 
As I joked in one of my first blog posts, the dispute between the Monetarists and the Keynesians was resolved when the Keynesians conceded all the substantive points and the Monetarists agreed to be called Keynesians.
I wonder what Brad DeLong would say about this. In 1999 (I believe) he published a piece in the JEP than made a number of similar arguments–although I find your explanation to be clearer.
A while ago I argued that the term ‘liberalism’ basically meant “Whatever the intellectual elite thinks about public policy at any given point in time.” Classical liberalism, social democratic liberalism, neoliberalism, all represented a sort of consensus in their day.
I then argued that the same was true of economics. Modern neoliberalism obviously isn’t really “liberal” as the term was defined in 1960. And New Keynesianism isn’t really Keynesianism circa 1960. Most academics are left of center, even if only moderately left of center. Thus there were always far more Keynesians than monetarists. So when the consensus changes, they have the power to call the new consensus whatever they want. And it flatters them to call it “Keynesian” in economics, and “liberal” in politics.
Krugman keeps talking about how Samuelson had all the answers back in 1948. But in those days Samuelson was saying that a Soviet-style centrally planned economy would produce higher living standards than our economy. I wonder how Krugman thinks that prediction is working out. Makes Friedman’s slip up with the k% rule seem rather trivial by comparison.
Andy, Scott: I think you are both spot on.
Scott: to my mind, you agree with Milton Friedman, except you think the money demand function isn’t very stable, so a k% rule won’t work. Fair caricature? I never really saw the k% rule as that central to Friedman’s view of the world.
Andy; You seem to have stopped blogging? I hope not for good. You have posted some really good stuff, over the months and years.
The Keynesian/Monetarist false dichotomy has bothered me for years. Thank you for putting it in perspective.
Scott: “Most academics are left of center, even if only moderately left of center.”
I think that’s been fairly well documented in the US, but I’d be willing to bet it’s not true for Canadian academic economists. A lot of Canadian academic economists are trained in the US, work closely with Americans, etc., so tend not to differ hugely from their US colleagues in terms of outlook. However since the Canadian political spectrum is so much to the left of the US political spectrum, what’s a center-left position in the US context becomes a center- right position here.
I was just chatting today with an American living here in Ottawa – he’s gone from been a staunch life-long Democrat to a conservative Harper supporter, without changing his political views much at all.
Of course, it could be the same story on your side of the border – it’s not that American academics are particularly left of centre, it’s just that their views are closer to those held by intellectuals in Europe, Canada, other parts of the world…
Nick, so are we getting into a competition to figure out who can come up with the best titles? 😉 But seriously – why won’t it speak its name? Why be ashamed?
Sorry, I figured out why it doesn’t speak its name – it need not. It’s hegemonic.
Frances: Yep. I would like to think I came up with the title. Actually, I got it from Dave, in conversation, about a year ago!
British Monetarists were always seen as moderate Keynesians, by US Monetarists.
So Nick, what exactly is the difference between a Kenyesian and a monetarist?
The original difference.
“As a Post Keynesian blogger (help me out someone, because my memory has failed) recently said, the canonical New Keynesian model is Monetarist.”
I’m guessing that you mean Robert Waldmann, although I don’t know whether he is a Post Keynesian. But if he’s the one you have in mind, what he actually said is that “self declared ‘New Keynesians’ would be called ‘classicals’ by Keynes.” It’s not quite the same thing.
Any version of Keynesian economics takes some version of classical economics as its starting point. It has to because the whole idea is to generalise the classical model, in such a way as to permit some form of unemployment which is not voluntary. Keynes’s own theory took Marshall and Pigou as its classical point of departure. Eggertson-Krugman starts from what RW calls the ‘classical’ NK model.
Kevin, in the canonical NK model, and in Eggertson-Krugman, there is no involuntary unemployment.
The labour market is entirely “classical”.
How do you assert “hegemony” in the face of Krugman’s comments:
“When monetarism failed – fighting words, but you know, it really did…”
and:
“Milton Friedman was wrong: in the face of a really big shock, which pushes the economy into a liquidity trap, the central bank can’t prevent a depression.”
Kevin, in the canonical NK model, and in Eggertson-Krugman, there is no involuntary unemployment.
That’s true; my comment was a bit muddled. I agree with Robert Waldmann that such models shouldn’t be called Keynesian. (But I certainly wouldn’t call them near-Monetarist as Nick does; that’s not how I remember Friedman’s views at all.) I mostly just wanted to clarify who his dimly-remembered ‘Post-Keynesian’ might be and what s/he actually said.
Adam P.: “So Nick, what exactly is the difference between a Kenyesian and a monetarist? The original difference.”
Good question. I will try to answer it.
Since “keynesian” has changed over the years, I’m going to talk about the difference in the early to mid 1970’s, because: that’s when Monetarism came into prominence as a serious alternative to Keynesianism; it’s the earliest I can remember.
At the policy level, the difference between M’s and K’s was a question of which policy instrument to assign to which target.
Keynesianism:
Fiscal policy is assigned aggregate demand and unemployment.
Monetary policy is assigned the composition of aggregate demand between consumption and investment (or, in a small open economy, the composition of demand between net exports and domestic absorption).
Which leaves inflation left over. They distinguished between “demand-pull” and “cost-push” inflation. Cost-push inflation (inflation even when the economy is at less than full-employment) was variously assigned to: industrial policy (try to raise productivity and reduce the monopoly power of firms and unions); and various forms of price and wage controls (does anybody remember TIP and MAP, which where taxes on inflation and a cap-and-trade quota on inflation?).
Monetarism:
Monetary policy is assigned aggregate demand and inflation.
Fiscal policy is assigned the composition of demand between consumption and investment, net exports and domestic absorption.
Which leaves unemployment left over. Unemployment was assigned to the sorts of labour market policies that the Keynesians thought of as anti-cost-push policies.
In short, all the targets and instruments swapped partners.
It was never easy to get at the root of the theoretical differences that underlay these policy differences on the assignment question.
The Keynesians really did believe that the Phillips Curve was downward-sloping, even in the long run. The whole talk about “cost-push inflation” (which is how Keynesians interpreted the adverse shift in the Phillips Curve in the 1970’s) shifting the Phillips Curve northwards only makes sense if you believe the PC slopes down.
It was not so many decades ago that every budget speech talked about the effect of G and T on AD, and whether fiscal policy was needed to stimulate or reduce AD.
Canada/US/UK really did impose price and/or wage controls to combat inflation.
Economists really did argue that monetary policy could not control inflation, or, even if it could, the price would be permanently higher unemployment than it otherwise would be.
The recent past really was a whole other world.
To give an example. The mainstream “Keynesian” view was that aggressive unions could cause inflation. Friedman said they could not cause inflation, but could cause unemployment.
Kevin: “I’m guessing that you mean Robert Waldmann, although I don’t know whether he is a Post Keynesian. But if he’s the one you have in mind, what he actually said is that “self declared ‘New Keynesians’ would be called ‘classicals’ by Keynes.” It’s not quite the same thing.”
Yes. I think Robert Waldmann is the blogger I was remembering, or misremembering. “Classical” and “Monetarist” is not the same thing.
And the labour market itself in many NK models is “classical” in the sense of market-clearing wages, but the level of employment is: inefficiently low on average; subject to inefficient fluctuations. Because the goods market is not classical, so the derived demand for labour is “wrong”.
The mainstream “Keynesian” view was that aggressive unions could cause inflation.
That’s a bit of a caricature I think. There was usually an assumption that the central bank was too weak to resist cost-push inflation. We knew only too well that a really severe “credit squeeze” (the term I remember from the late 1960s) could certainly curb inflation at the cost of unemployment. Keynesians of that era were also Kaleckians: recessions were the instruments used by the capitalists to discipline the workforce. The political aspect of monetary policy was stressed by Keynesians and glossed over by Monetarists.
“Economists really did argue that monetary policy could not control inflation, or, even if it could, the price would be permanently higher unemployment than it otherwise would be.”
After that point of view was discredited, the prediction about permanently higher unemployment turned out to be correct for Europe (though I suppose few if any would accept the old Keynesian explanation).
BTW I haven’t (intentionally) stopped blogging. Just haven’t managed to get a post out recently.
Here’s how the Keynesians thought about inflation and unions etc.:
My inflation = some function of: my power (which depends on structural forces and also depends negatively on unemployment); and on your inflation (through catch-up and indexation etc.).
Therefore, an increase in my structural power will cause higher inflation, unless fiscal and/or monetary policy deliberately increases unemployment by enough to offset it.
New Keynesians take the same equation, assert that the equation is HD1 in {my inflation, your inflation} solve it out for the implied natural rate of unemployment when my inflation = your inflation, and assert that structural power increases the natural rate of unemployment.
Ok Nick,
I think that was a pretty good answer (though I’d actually asked for the “original” distinction which presumably would have taken us back well before the 70s).
So, in that distinction how is Friedman’s permanent income hypothesis in any way related to “monetarism”?
To pre-empt you, the answer is that it’s not. Just because it came from Friedman doesn’t mean it has anything to do with monetarism.
Same goes for the Euler equation. RBC models are models that generate recessions even in barter economies (just as NK models can). RBC models all have Euler equations. Since many RBC models don’t even have money (and in the ones that do money has no real effects of any kind) you can’t possibly be correct that the presence of an euler equation in the model is “monetarist” in nature.
So, I’ve only just got started and I’d say we’re already down to 1.5 out 3 equations being “monetarist”. Thus, at best the model is 50% monetarist in nature, that’s already a far cry from 5/6ths.
PS: And furthermore, couldn’t we just as much claim that the Euler equation is the modern incarnation of the Modigliani’s life-cycle model of the consumption function?
Was Modigliani in any way a monaterist?
How would you compare the economic performance of the two regimes? In terms of unemployment, productivity growth, GDP growth, budget deficits, not to mention median wage growth, the monetary policy regime doesn’t look too good. And it’s starting to not look so good on the price stability front, also.
There’s a lot more to Monetarism than the k% rule.
Yeah, but you have to remember Monetarism got its name from the fundamental views about rigid money supply and about how money rules it all (intention does not matter).
Once the k% rule was thrown away as non-working and once you guys started including intention (fiscal …) dynamics you all became keynesians without realizing it – you just refuse to call yourselves that, for emotional, nostalgic or whatever stupid political reasons.
There’s very few original monetarists left.
Friedman was at least intellectually honest: he knew that Monetarism would be dead without the k% rule (and some other details) and he stuck to those details until the end.
That old theory simply does not work in some important circumstances and I have no idea why you continue to call your current set of theories ‘Monetarism’ – it’s modern keynesian economics in all but name.
Nick, My biggest disagreements with Friedman aren’t related to the k% rule, but the identification of monetary shocks. I think he misidentified them in the Monetary History of the US, which led him to assume that policy lags were longer than they really are. In my view a monetary shock is a change in the expected future path of policy, which shows up as a change in expected future NGDP. Also, like most economists Friedman takes a “wait and see” approach toward policy. I believe that we know immediately whether a policy has worked or not, because the goal of policy is to target expectations, and stabilize expectations. Thus Friedman predicted high inflation in 1984, and was wrong. I would never predict inflation except to the extent that I acknowledge market forecasts. Modern monetarists are again predicting high inflation based on rapid growth in the base, despite no evidence in the actual CPI or the TIPS spreads to back up their prediction. I see that as the biggest weakness of monetarism, even bigger than the k% rule.
If the econ profession ever adopts Svensson’s “target the forecast” approach, much of 20th century monetary policy discussion (on both sides) will look like voodoo.
I also hope Andy keeps blogging. In my comment section I recently said that Andy made the best Keynesian arguments in my blog.
Anon, Even Krugman doesn’t agree with the Krugman quotation you provided. Krugman has argued that the central bank can prevent a depression with a higher inflation target, and has strongly criticized the BOJ and Fed and ECB for not doing so. So it isn’t “can’t” prevent a depression, it’s “won’t” prevent a depression. (If central banks could not stimulate the economy at the zero bound, Krugman’s current criticism of the Fed would be nonsensical.) BTW, there were no big shocks in the first two years of the Great Depression in the US. The worst was a banking panic about 16 months in, and that was a fairly small panic. The Fed let the monetary base fall in the first year of the Depression (about 7% between Oct. 1929 and Oct. 1930) so not only was it not trying to prevent the Depression, it was helping cause it.
Scott: I don’t think Krugman ever said the Fed can create inflation at the zero bound. All I’ve ever heard him say is that QE is their best shot at it.
K remarks:
But of course that’s the point. Krugman writes what’s politically expedient for his party today, whereas his academic writings on which his reputation were built say something rather different. His academic work says the Fed CAN create inflation at the zero bound.
Still he is smart and executes this parlor trick subtly and with enough ambiguity to later claim that the rubes just misunderstood his columns.
“His academic work says the Fed CAN create inflation at the zero bound.”
Actually, no that’s not what it says. Not at all.
Adam: A simple definition of “monetarism” is “whatever Friedman believed (about money/macro)”. It’s crude, and it slights other monetarists, but at least it’s got some content.
Your point about Modigliani is a good one. (You could also cite Phelps, on the vertical LR Phillips Curve). In many ways, it was Keynesians themselves who slowly chipped away at the underpinnings of what used to be thought of as Keynesianism. And you might argue (and some Post Keynesians do) that this goes all the way back to Hicks’ ISLM. Lucas himself, IIRC, said that he considered himself a Keynesian in 68, and saw the Lucas/Rapping paper as just filling in the background details in the Keynesian perspective. But when the Keynesians had finished building the microfoundations, and stood back to look at the result, it all looked a lot closer to Friedman’s view of the big picture.
The Lifecycle/Permanent Income Hypothesis was subversive of Keynesian macroeconomics because it massively reduced the marginal propensity to consume out of current income. This in turn reduced the multiplier — understood as the deviation-amplifying positive feedback mechanism that played a large role in earlier Keynesian writings. That multiplier has disappeared totally from New Keynesian macro. (It comes back a bit in the Gauti Eggertsson/PK paper, and I was trying to bring it back in that post a few weeks back).
The multiplier played a big role in Old Keynesianism, none in New Keynesianism, and none in Friedman.
As an aside, I think that Friedman’s PIH is closer in spirit to the Euler equation than is Modigliani’s LIH. That’s because, IIRC, Friedman based his analysis on the Irving Fisher diagram. (present consumption on one axis, future consumption on the other, and an intertemporal budget constraint tangent to an indifference curve). The Irving Fisher diagram is the same as the modern treatment, except it only has 2 periods, and ignores uncertainty.
Part of the problem in distinguishing Keynesians and Monetarists is the Friedman never successfully laid out a formal macro model to show the differences. His one attempt to do so was a hopeless failure. It was only when early New Classical and New Keynesian models appeared that you could look back and say “OK, that’s what Milton Friedman must have meant”.
“A simple definition of “monetarism” is “whatever Friedman believed (about money/macro)”. ”
Seriously? I’d have to say that’s a horrible definition that is entirely without content.
White Rabbit: “Friedman was at least intellectually honest: he knew that Monetarism would be dead without the k% rule (and some other details) and he stuck to those details until the end.”
I don’t think that’s true. IIRC, Friedman abandoned the k% rule some time in the 1990’s.
“Once the k% rule was thrown away as non-working and once you guys started including intention (fiscal …) dynamics you all became keynesians without realizing it – you just refuse to call yourselves that, for emotional, nostalgic or whatever stupid political reasons.”
Certainly not true in my case. I call myself a Keynesian, and a quasi-monetarist. And precisely because I keep on insisting that Keynesian macroeconomics only makes sense in a monetary exchange economy, and not in a barter economy (a controversial position – ask Adam P.) I insist that “quasi-monetarists” like me are in some ways more true to the original Keynesian vision (more “Keynesian”) than are New Keynesians.
Not everything is emotion, nostalgia, or stupid politics. It’s intellectually lazy to dismiss others’ viewpoints like that.
Adam: IIRC, the word “Monetarism” was originally coined by those who opposed Friedman’s views as a name for Friedman’s views (and the views of people like Friedman) on money/macro.
Ironically (and this goes back to White Rabbit’s assertion that some people cling to the label “Monetarism”) it was Paul Krugman I think who coined the name “quasi-Monetarist” to describe people like David Beckworth, me, and I think Scott Sumner. I’m happy to accept the label, because I think it has some descriptive power, as long I’m recognised as a quasi-Keynesian too.
IIRC, Friedman abandoned the k% rule some time in the 1990’s.
Yes, in an interview with the FT. On the general question of what Monetarism was all about, this paper is as good as anything I’ve seen:
“Different exponents of Monetarism stressed different propositions, but it would be fair to say that, from the point of view of the non-academic observer whose main concern was the conduct of economic policy, Monetarism involved first a theory of inflation, second a theory of the cycle, and third, as a corollary of these, a recommendation for the conduct of monetary policy. Specifically, inflation was said to be explicable in terms of the rate of growth of the money supply, and the cycle, or more precisely its turning points, in terms of changes in that rate of growth.” — David Laidler (1990; PDF file).
Adam writes:
One word: expectations.
“One word: expectations.”
indeed.
“New Keynesians take the same equation, assert that the equation is HD1 in {my inflation, your inflation} solve it out for the implied natural rate of unemployment when my inflation = your inflation, and assert that structural power increases the natural rate of unemployment.”
This flew straight over my head. Is there a reference?
RSJ: sorry. Let me be clearer (or try to be)
Suppose we have an equation something like this:
W = axMarketPower – bxUnemployment + cxP
Where W is one type of inflation, in one part of the economy (“My inflation”), and P is inflation in the rest of the economy (“Your inflation”). And a, b, and c are all positive parameters.
If you look at that equation, it seems to tell you that an increase in market power will cause inflation, unless it is offset by a big enough increase in unemployment. (The more market power I have, the more I can raise my price, unless unemployment is high enough to create a countervailing force. And if prices rise in the rest of the economy, that will cause some sort of wage-price, or wage-wage, or price-price spiral, causing me to raise my prices even more.)
Now, assume c=1 (that is what I meant by HD1 or Homogenous of Degree One in nominal variables). And suppose I aggregate up over all sectors of the economy, so that M=P on average (because all the “My’s” are the same as all the “You’s”). Then I get:
P = axMarketPower -bxUnemployment + P
Solving I get:
Unemployment = (a/b)xMarketPower
Market power determines the natural rate of unemployment, and not the equilibrium rate of inflation. And there is no trade-off between unemployment and inflation.
P.s. I know I’m being really sloppy here. especially on the Homogeneity of Degree One thingy.
I think y’all are getting too serious about this question of what constitutes monetarism. When the definition of a word becomes highly controversial, the word loses its communicative value. It appears that the word “monetarist” has crossed that threshold.
One may, however, talk about “what monetarists [typically] believed” without committing to the question of which of those beliefs constitute monetarism per se and which are merely incidentally correlated. To the extent that “what monetarists believed” in the mid 1960’s differed from “what Keynesians believed,” it’s pretty clear that monetarist-associated beliefs are closer to the current new Keynesian orthodoxy.
But if one insists on pushing the definitional issue, I’d ultimately have to say that monetarism, as such, is not hegemonic. In my old blog post (which I cited earlier), I went on to say that, in actuality, the one point the Keynesians never conceded was the raison d’être of monetarism (namely the reasonably stable and predictable relationship between money supply growth and inflation, according to some definition of “money supply”).
Perhaps we can all agree (unless there are any die-hard k%-ers here) that, given his own definition, Krugman is making a true statement when he says that “monetarism failed.” But we have to recognize that his definition is limited and that “monetarism failed” does not mean that “the monetarists failed.”
Ahh. Thanks.
That is clever.
But you are using homogeneous of degree 1 in a different way than I am used to — the “general” case wouldn’t have the P’s cancel, so you would get something like
aP = market power – bunemployment
which we could re-write as:
a(inflation) + b(output gap) = market power = nominal interest rate charged – nominal “natural” rate of interest.
Looks a bit like the Taylor rule. Or perhaps I am totally missing the rules of this game.
Andy: I can live with that. I don’t think there are any k%ers left. I did float a k% trial balloon a few weeks back http://worthwhile.typepad.com/worthwhile_canadian_initi/2010/10/was-milton-friedman-right-after-all.html but it didn’t really fly very well.
RSJ: Yep. I wasn’t really using HD1 correctly. What I really meant is that if you take all the variables measured in $, and multiply them all by the same constant (just changing the units from dollars to cents, or whatever), and leave all the other variables and parameters unchanged, the equation should still hold true. That’s the “absence of money illusion”. (I would have to re-specify my equation in multiplicative form to make it work, then take logs to make it linear).
You lost me on the second part. There’s the Phillips Curve equation, and then there’s the Taylor Rule equation. The Phillips Curve must be HD1. The Taylor Rule must NOT be HD1, otherwise the inflation rate is indeterminate in the system as a whole.
Did you ever see my old post on “Units”? That was about this homogeneity stuff. http://worthwhile.typepad.com/worthwhile_canadian_initi/2010/05/units.html
OK, I just re-read that post. I remember reading it when it came out, and I was a bit confused about short-run versus long run.
It seems to me that if you are going to claim that something is short run not-neutral but long run neutral, then this should fall out of the constraints, rather than having different constraints for the short and long run. And one way to do this would be to have some parameters that are constant over the short run, but variable over the long run, in which case, the long run degree of the polynomial will not be the same as the short run degree, so the same constraint can go from not being HD1 to being HD1, or vice versa. Or am I totally missing how you would model long and short runs? What is the technical definition of “long run”?
Yep. Theory says it should be HD1. And it looks to be roughly HD1 given long enough. But not immediately. So we cook up some reason why it isn’t HD1 immediately. Like lags in adjusting prices, or lags in adjusting expected inflation to actual. The technical definition of long vs short run depends on the theory we cook up.
OK, now I realize I don’t have any idea of what the NK philips curve is supposed to be:
I was looking at this paper
Click to access InflationDynamics.pdf
They provide a formula (equation 4) for inflation in period t, p_t, that is of the form
p_t = lambdasum_k(B^k E_k{mc_(t+k)} )
where k goes form 0 to infinity and mc_t is defined as
“the percent deviation of the firm’s real marginal cost from its steady state value… The benchmark theory thus implies that inflation should equal a discounted stream of expected future marginal costs.”
In that case, mc_t should go to zero as t goes to infinity. Then we can look at the difference between the real and nominal interest rate for an N period zero coupon bond bought in period 0. It will be the nominal yield corrected by the average of the expected future inflation rates. So re-arranging, this becomes for the Nth period correction term:
1/N(E_1{mc_1}+ .. +E_1{mc_(N)}) + a term that goes to zero as N goes to zero.
Here I’m just using the law of iterated expectations plus some fiddling with re-arranging the terms.
But if mc_N goes to zero as N goes to zero, given that over the long term the real marginal cost will converge to its steady state value, then the arithmetic mean will also go to zero.
Does this mean that in the limit as the bond term gets very large, the nominal long term rate converges to the real long term rate? Or does mc_N not go to zero?
“I call myself a Keynesian, and a quasi-monetarist. And precisely because I keep on insisting that Keynesian macroeconomics only makes sense in a monetary exchange economy, and not in a barter economy)”
I think that’s also a Post-Keynesian thing
RSJ: you lost me about halfway down.
A couple of points to note:
That (Calvo) Phillips Curve is used in most NK models because the math is tractable, and there’s a simple story behind it – each firm has a constant probability each period of being allowed to change price.
The deviation of real MC from steady state is a positive function of the deviation of output from steady state (it would be a negative function of unemployment, if there were unemployment in the model).
If Beta is less than one, the Long Run Phillips Curve in that model would not be vertical. That’s widely seen as a “glitch” in that model. You can remove that glitch by assuming growth equal to the rate of time preference, which is similar to setting Beta = 1. The LR Phillips Curve is then vertical. If output is permanently above steady state (so real MC is permanently above steady state) inflation is infinite. Then it’s a true natural rate model. On average, output is expected to equal the steady state value, though it can deviate from the natural rate in the short run.
Rob: “I think that’s also a Post-Keynesian thing”
I think that too. And the Post Keynesians are absolutely right on that. And it’s really important. And it bugs me that some Keynesians don’t get it, and I can’t convince them. (Some New Monetarists don’t get it too, and that bugs me too).
Can I call myself a “Post Keynesian”? Better not. It would only add to the confusion.
It does go to show, though, that labels like “Keynesian”, Monetarist”, “Post Keynesian” etc., have a rather limited usefulness. And we shouldn’t identify too closely with any particular “school”.
I keep on insisting that Keynesian macroeconomics only makes sense in a monetary exchange economy, and not in a barter economy.
Indeed you do. But what you call a barter economy looks suspiciously like a frictionless, Panglossian world in which macroeconomic problems are excluded by assumption. Keynesian macroeconomics typically takes that sort of model as a starting point and throws in some complications. But the complications don’t have to have anything to do with money as a medium of exchange. A rigid wage and/or price level, or Calvo-type sluggish price adjustment, is just as believeable in a (non-Panglossian) barter world as it is in a monetary economy. One of the interesting things about macroeconomics is that all sorts of models which would have looked very weird to Keynes generate somewhat ‘Keynesian’ results — see for example how Eggertson and Krugman get a (sort of) Keynesian consumption function in a paper Keynes would almost certainly have thrown in the bin. He evidently found Ramsey’s paper a bit hard to take.
Kevin: tu quoque!!
“But the complications don’t have to have anything to do with money as a medium of exchange.”
Yes they do! (except, the Keynesians aren’t explicit about it, and may not even realise they are doing it).
OK, that does it. Time to re-do Barro and Grossman 71 for a New Keynesian model. This will not be easy.
But even the Barro-Grossman paper (which I probably would never have read if not for your blog) treats money in a very perfunctory way. It could just as easily be precious stones that people seek to hoard. Clower gets a mention, but B&G don’t actually work through any CIA constraint or the like. The ‘Keynesian’ consumption function emerges simply because the household is rationed in the labour market, so the effective budget set is a sawn-off Walrasian budget set. That’s all you need.
Kevin: “But even the Barro-Grossman paper (which I probably would never have read if not for your blog) treats money in a very perfunctory way. It could just as easily be precious stones that people seek to hoard.”
There’s a Q=QD semi-equilibrium condition for the market in which consumption is traded for stones; a Q=QD semi-equilibrium condition for the market in which labour is traded for stones; but there is no such condition for the market in which consumption is traded for labour. Because that market does not exist. B&G is a model of a monetary exchange economy precisely because they assumed that market did not exist.
Think about the constrained vs notional labour demand curve in the B&G model. The monetary exchange equilibrium at point B, vs the Walrasian equilibrium point where the MPL curve crosses the labour supply curve. It’s exactly the same in the New Keynesian model, except you have to multiply the MPL by a constant (1-(1/e)), where e is elasticity of demand for an individual firm’s output.