Anything IS-LM can do, Fisher did better

Here’s what economists knew before the General Theory:

1.  Monetary policy and velocity determine NGDP growth.

2.  Velocity is positively related to interest rates (and hence investment booms and deficit spending may raise velocity.)

3.  Wages and prices are sticky in the short run.

4.  Because of point 3 a monetary shock may produce a liquidity effect for short term rates.

5.  Because of point 3, money and velocity shocks can destabilize output in the short run.

Here’s what we have learned since:

1.  Macro models need to incorporate expectations that are consistent with the prediction of the model.

Here’s what IS-LM added to our knowledge:

1.  Conventional monetary policy may be ineffective in a liquidity trap.

Here’s what IS-LM subtracted from our knowledge:

1.  Monetary policy may be ineffective in a liquidity trap.

Net gain from IS-LM:   Zero.

What got me thinking about interwar macroeconomics was a recent post by Brad DeLong, which painted a very bleak picture of pre-Keynesian macro, and then argued that a great leap forward occurred in 1937.  It wasn’t clear whether he thought much progress had been made since.  (I certainly don’t think so—but then I also view 1937 as a step backward.)  Here’s the crux of DeLong’s argument:

If you try to read pre-Keynesian monetary theory, or for that matter talk about such matters either with modern laymen or with modern graduate students who haven’t seen this… you quickly realize that this seemingly trivial formulation is actually a powerful tool for clarifying thought, precisely because it is a general-equilibrium framework…. Here are some of the things it suddenly makes clear:


What determines interest rates? Before Keynes-Hicks… there has seemed to be a conflict between the idea that the interest rate adjusts to make savings and investment equal [in financial markets, “loanable funds”], and that it is determined by the choice between bonds and money[, “liquidity preference”]. Which is it? The answer, of course–but it is only “of course” once you’ve approached the issue the right way–is both: we’re talking general equilibrium here, and the interest rate and price level are jointly determined in both markets.

How can an investment boom cause inflation (and an investment slump cause deflation)? Before Keynes this was a subject of vast confusion, with all sorts of murky stuff about “lengthening periods of production”, “forced saving”, and so on. But once you are thinking three-good general equilibrium, it becomes a simple matter. When investment (or consumer) demand is high… [people] are… trying to shift from bonds to goods…. both the bond-market and goods-market equilibrium schedules… shift; and the result is both inflation and a rise in the interest rate.

How can we distinguish between monetary and fiscal policy? Well, in a fiscal expansion the government sells bonds and buys goods…. In a monetary expansion it buys bonds and “sells” newly printed money, shifting the bonds and money (but not goods) schedules….

Of course, this is all still a theory of “money, interest, and prices” (Patinkin’s title), not “employment, interest, and money” (Keynes’). To make the transition we must introduce some kind of price-stickiness, so that incipient deflation is at least partly translated into output decline…. But the basic form of the analysis still comes from the idea of a three-good general-equilibrium model in which the three goods are “goods in general”, bonds, and money….

But step back from the controversies, and put yourself in the position of someone who must reach a judgement about the likely impact of a change in monetary policy, or an investment slump, or a fiscal expansion. It would be cumbersome to try, every time, to write out an intertemporal-maximization framework, with microfoundations for money and price behavior, and try to map that into the limited data available…. [T]hat is why old-fashioned macro, which is basically about that model, remains so useful a tool for practical policy analysis.

First let me talk about where I agree with DeLong.  The last paragraph is exactly my view—I like simple models best.  And I like old-fashioned models best.  The only exception is that I think rational expectations are crucial to any macro model, especially under a fiat money regime.

Here’s where I disagree.  If you read David Laidler you will realize that pre-Keynesian models were actually much more sophisticated than they appear.  Indeed I have a much higher opinion of those models than even Laidler had, because I believe (with Milton Friedman) that for most purposes a partial equilibrium approach is to be preferred to a GE model.  Model the supply and demand for the medium of account to get nominal shocks, and use an SRAS to get business cycles.  Assume money is neutral in the long run.  Interest rates are hard to model, but I’ve never seen any evidence that they play an important role in business cycles.  (Yes, they affect velocity somewhat, but I see interest rates as mostly reflecting expected changes in NGDP growth, which are primarily driven by monetary policy.)

DeLong is not right about the (interwar) classical economists lacking a good framework for evaluating the effects of investment booms and fiscal stimulus.  Indeed he cites John Hicks very favorably, but doesn’t mention that in his famous 1937 paper Hicks claimed that the liquidity trap was the only real innovation in the entire General Theory.  If that is true (and I think it is), then it’s pretty hard to claim that classical economists had no coherent macro model.

People like Irving Fisher had a perfectly good macro model.  Indeed, except for Ratex it’s basically the model that I use in all my research.  But the problem is that these pre-1936 models didn’t use Keynesian language.  And they didn’t obsess about trying to develop a general equilibrium framework. A GE framework is not able to predict any better than Fisher’s models, and is not able to offer more cogent policy advice than Fisher’s model.  Indeed in many ways Fisher’s “compensated dollar plan” was far superior to the monetary policy the Fed actually implemented last October.  (Although I would prefer CPI futures target to a flexible gold price, at least Fisher’s plan had a nominal anchor.)

I’m not just skeptical of IS-LM; I’m fed up with almost all of modern macro.  If it’s 1995 and GDP growth is 3%, I have no doubt that structural models can predict 1996 GDP growth will also be 3%.  And they will usually be pretty close.  Of course when we really need advice from macroeconomists, at turning points like last fall, their models are basically useless.  When I saw the markets crash in October I knew that a few months later the Blue Chip forecasters would be downgrading their “consensus forecast.”  Now I just watch the markets, and hardly even waste time with econometric models and their forecasts.  My prediction is that within a few decades all these economic alchemists will given up on their fruitless attempts to create a golden structural model, and simply infer market forecasts of the policy goal.

The sophisticated reader may be saying I am unfair, models aren’t supposed to offer unconditional forecasts of turning points, they are supposed to offer conditional forecasts of how monetary and fiscal policy will affect the economy.  But can they even do that?  Nobody seems to believe in conventional monetary policy at this point, but what about fiscal policy?  It’s too soon to say, but in 1998 a famous economist expressed skepticism about the rationale for fiscal stimulus in a liquidity trap:

As a starting point for this discussion, we should notice that Japan has unwittingly managed to turn itself into an old-fashioned Keynesian economy, the sort of economy envisioned in the original, 1948 edition of Samuelson’s textbook. With large excess capacity, Japan is unmistakably constrained by demand rather than supply; with call money rates hard up against zero, Japan has the fixed interest rate assumed in naive multiplier analysis. . . .

What continues to amaze me is this: Japan’s current strategy of massive, unsustainable deficit spending in the hopes that this will somehow generate a self-sustained recovery is currently regarded as the orthodox, sensible thing to do – even though it can be justified only by exotic stories about multiple equilibria, the sort of thing you would imagine only a professor could believe. Meanwhile further steps on monetary policy – the sort of thing you would advocate if you believed in a more conventional, boring model, one in which the problem is simply a question of the savings-investment balance – are rejected as dangerously radical and unbecoming of a dignified economy.

Will somebody please explain this to me?

Note the exasperation in the last sentence.  That’s how I have felt for 8 months.  Even today there are professors who think massive government spending can overcome the BOJ’s deliberate policy of targeting inflation at negative 1%.  But we know better.  Who are we?  Who rejects the “orthodox, sensible thing to do?”  Paul Krugman and I.

[Read the entire piece—it’s about Japan, but it’s the best critique of current U.S. fiscal policy that I have ever read.  Much better than I could do.  Yes, one could argue that it’s not applicable to the U.S., where a recovery would be self-sustaining.  But only because our central bank is not likely to pursue deflationary policies indefinitely.  It all comes back to monetary policy.  Fiscal policy won’t work with a bad central bank, and isn’t needed with a good one.]

Bonus points: Can anyone explain this sentence from DeLong’s post?

It is remarkable: if you ask, every macroeconomist says they believe in the quantity theory of money: MV = PY.



29 Responses to “Anything IS-LM can do, Fisher did better”

  1. Gravatar of Lord Lord
    19. May 2009 at 17:55

    Why weren’t the classical models used in 1929? Were they unknown? Were they unbelieved? A lack of imagination at the Fed? A lack of will? Had they already limited themselves to conventional monetary policy? The straight jacket of the gold standard?

  2. Gravatar of Nick Rowe Nick Rowe
    19. May 2009 at 18:16

    I think there are two more (closely-related) things we have learned since:

    1. If one market is in excess supply or demand, buyers or sellers will be quantity-constrained, and that will spillover and affect their demands and supplies in other markets.

    2. Those spillovers will have very different effects in a monetary exchange economy than in a barter economy.

    Perhaps implicit in the General Theory and ISLM, but explicit in Clower etc.. Was it implicit in Fisher?

    (This is why I keep harping on about the medium of exchange vs the medium of account function of money.)

  3. Gravatar of Winton Bates Winton Bates
    19. May 2009 at 20:28

    You write: “My prediction is that within a few decades all these economic alchemists will given up on their fruitless attempts to create a golden structural model, and simply infer market forecasts of the policy goal.”

    I don’t understand. Market forecasts reflect the (thousand ?)models of market participants. Where do market participants get their models from, if not from economic alchemists of various kinds(including defunct macro economists)? This seems to me to imply that there will continue to be plenty of demand for economic alchemists competing to create the golden structural model, even after the central bank ceases to be interested in such modelling.

  4. Gravatar of TGGP TGGP
    19. May 2009 at 21:36

    Lord, according to Lawrence White the Fed was dominated by the Real Bills Doctrine back then. Meltzer agrees, and says that’s why Fisher’s theories had no influence on the Fed.

  5. Gravatar of Kevin Donoghue Kevin Donoghue
    20. May 2009 at 01:00

    “…in his famous 1937 paper Hicks claimed that the liquidity trap was the only real innovation in the entire General Theory. If that is true (and I think it is), then it’s pretty hard to claim that classical economists had no coherent macro model.”

    I don’t think Hicks would have regarded that as an accurate paraphrase. So far as the development of economics as a science (of sorts) is concerned, the real importance of the General Theory was that it provided a way to integrate monetary theory with price theory. Keynes himself says that one of his objectives is “to bring the theory of prices as a whole back to close contact with the theory of value.”

    In that regard he accomplished quite a bit (although he also left lots to be done). I call three expert witnesses:

    Hicks, writing in the very article where you claim he found nothing new in the General Theory apart form the possibility of a liquidity trap:

    “Income and the rate of interest are now determined together at P, the point of intersection of the curves LL and IS. They are determined together; just as price and output are determined together in the modern theory of demand and supply. Indeed, Mr. Keynes’ innovation is closely parallel, in this respect, to the innovation of the marginalists. The quantity theory tried to determine income without interest, just as the labour theory of value tried to determine price without output; each has to give place to a theory recognising a higher degree of interdependence.”

    Samuelson in What Classical and Neoclassical Monetary Theory Really Was:

    “In a real sense there was a dichotomy in our minds; we were schizophrenics. From 9 to 9:50am we presented a simple quantity theory of neutral money. There were then barely ten minutes to clear our palates for the 10 to 10:50 discussion of how an engineered increase in M would help the economy.”

    Pigou in Keynes’s General Theory (1950):

    “Nobody before [Keynes], so far as I know, had brought all the relevant factors, real and monetary at once, together in a single formal scheme, through which their interplay could be coherently investigated. His doing this does not, to my mind, constitute a revolution. Only if we accepted the myth as I regard it that earlier economists ignored the part played by money, and, even when discussing fluctuations in employment, tacitly assumed that there weren’t any, would that word be appropriate.”

    Of course Pigou is right to defend the classics against “the myth” he implicitly accuses Keynes of propagating, but he also acknowledges that they didn’t have a coherent theory and that Keynes provided one. That’s important; predictive power isn’t the only thing that matters in a model, indeed it’s far from being the most important thing if your goal is to understand how the system works.

    In your previous post you employ a military metaphor to explain why you take so many pot-shots at Krugman, the commander of the Keynesian forces. Even as a military doctrine I don’t think this is sound. General Wolfe and Admiral Nelson both won major victories despite being mortally wounded. But in any case Krugman isn’t the guy you need to take out. It’s Keynes himself. In the world of ideas a man doesn’t have to be breathing to have great power. Contrary to what Lucas would have you believe, Keynes isn’t dead in the relevant sense. In fact I don’t think I’ve ever seen him looking so fit.

  6. Gravatar of Adam P Adam P
    20. May 2009 at 02:13

    Well said Kevin, very well said.

  7. Gravatar of Matt C. Matt C.
    20. May 2009 at 03:22

    Call me crazy or nut job, but I believe there was a coherent macroeconomic model that was proposed much sooner than the 1930s. It was considered at the time a classical model, I believe this is a correct statement. Mises’ Theory of Money and Credit was originally published in 1912 and it combined micro and macro theory. Hayek then expands the argument further later. Here is the important element of the Austrian school and Roger Garrison points this out in a reply to Brad DeLong. You can find the critique here:

    “The Austrian theory is not a theory of recessions per se; it is a theory of the unsustainable boom. As such, it has a much stronger link to the underlying microeconomics than does much of today’s mainstream theorizing. The Austrians focus broadly on credit markets and ask what happens when the price of credit, i.e., the interest rate, is held below its market-clearing level. If interest rates were held too low by legislation (interest-rate caps), there would follow an immediate credit crunch. This “if-then” proposition is a direct analogue to the proposition that rent control causes a housing shortage and, more generally, that price ceilings cause quantities demanded to exceed quantities supplied.”

    The problem, I would guess, is that maybe the Austrian theory concentrates too much on the Micro and so it easily dismissed as not being a purely Macro theory at all.

  8. Gravatar of ssumner ssumner
    20. May 2009 at 04:31

    Lord, The problem was partly the gold standard, but policy was bad even given the constraints of the gold standard.

    Nick, I don’t know if the interwar economists thought in terms of excess demand in one market causing excess supply in another. I’m not even sure whether I implicitly think that way. When I teach S&D I usually just say that if D shifts right, P and Q rise at the new equilibrium. I gather that some instructors say “at the old price there is excess demand, hence a shortage, which causes prices to be bid up to the new equilibrium.” Maybe I should teach things that way, but I always assumed that with Ratex prices will immediately adjust in a commodity market. Now how about macro models with sticky prices? Then I simply assume prices are flexible in the long run. If you have more of the medium of account, then its expected future “price” will fall. This causes higher expected future NGDP. Once you get higher expected future NGDP, you discard monetary theory and simply ask how that would affect current AD, current NGDP.

    I don’t have any objection to the excess supply/demand approach, it simply seems unnecessary to me. But I have an open mind on the subject, because I believe there are many transmission mechanisms. And I wouldn’t discount the possibility that it is implicit somewhere in my thought experiment.

    In any case, I was looking at the question pragmatically. I don’t see any flaws in the interwar model that inhibit their ability to explain macro events.

    Winton, I don’t see any evidence that the public now predicts business cycles better than in the 1920s, when the public did not have structural models to rely on. In any case, how do modern markets predict turning points before structural model builders have even noticed the shock—as in early October 2008?

    TGGP, I would agree and add that Fisher’s CDP would have meant the end of the gold standard, so it was not politically acceptable in 1929. When it was implemented by FDR in 1933, it immediately produced rapidly rising NGDP. If not for the NIRA, the Depression would have ended quickly.

    Kevin, Regarding Hicks and IS-LM, I think you are confusing new theoretical constructs, with new ideas. Yes, the IS-LM model with is implicit in the General Theory was new, but Hicks argued that the liquidity trap was the only important implication of that model that was new. The other ideas in the General Theory were concepts that interwar economists knew about but explained in a different language. In addition, I am afraid the the IS-LM model is far less useful than supply and demand. To understand why, we need to start with the understanding that all Keynes/Hicks really did was to envision of general equilibrium model in r/Y space. The problem is that unlike S&D, they didn’t provide a good explanation for how and why r and Y evolve over time. Any economic shock will produce a new r and Y combination, which can be shown in r/Y space, but IS-LM does a very poor job or both predicting and explaining the new combination.

    People are always telling me that I should use IS-LM, because there is some way to express my ideas in that format. Yes, but at what cost? In my view monetary policy has its biggest impact on IS, not LM. But most people (including Keynes) seem to assume that monetary policy generally shifts the LM curve.

    In contrast, S&D is very useful. If OPEC sharply cuts back on oil production, we can anticipate rising prices and falling Q. But monetary policy is different. A really tight money policy (like 1929-30) produces falling r and Y, thus it looks like an IS shock.

    So yes, you are right that IS-LM tries to integrate the real and monetary sectors, but I would argue that it adds no useful information to previous models. What it does do, it does very poorly.

    Samuelson’s argument is just silly. The interwar economists knew that money is non-neutral in the short run and neutral in the long run. That’s what he is referring to. Almost every modern textbook does the same thing. Neutrality in one chapter, non-neutrality in another.

    Pigou is wrong in my view. Don’t confuse a “coherent model” with a “GE model.” One does not imply the other. Our textbook models of the supply and demand for apples are coherent, in my view, but they are not GE models. The only question of interest, and the only question I addressed is USEFULNESS. I suppose smart grad students get psychic satisfaction from doing GE models. Too bad those models (with a few obvious exceptions like models that integrate the current and capital accounts) rarely tell us much of anything of interest beyond partial equilibrium models. In my view GE models are best at explaining not movements in total GDP, but relative changes among its components. But the two big questions of macroeconomics are inflation and business cycle theory.

    The military analogy was half-joking. I don’t really expect Keynesians to run for the hills. I was serious about Krugman being the smartest Keynesian blogger, however. That’s why I focus on him. DeLong shares my interest in economic history, so I also find his blog interesting.

    Adam, I do agree, Kevin’s question was one of the best ever placed on this blog. I just have a different point of view.

  9. Gravatar of ssumner ssumner
    20. May 2009 at 04:54

    Matt, You may be interested in looking at comments in my Austrian post from about three weeks back. There are more than 180 comments (a record). The comments are longer than most books.

    I like some aspects of Austrian economics, but have two problems with Austrian business cycle theory. I still don’t understand what the whole “misallocation of capital” story tells us that is useful. Sticky-wage theory already explains why nominal shocks lead to booms and busts. And I also don’t like the Austrian focus on interest rates, which I believe are a very poor indicator on monetary policy. For instance, I believe monetary policy was too tight in 2002, whereas most Austrians seem to think it was too loose (because rates were very low.) BTW, not just Austrians, but John Taylor also has that view.

  10. Gravatar of James Hasik James Hasik
    20. May 2009 at 05:50

    I’m delighted to read that you’re “not just skeptical of IS-LM,” but “fed up with almost all of modern macro.” If models like IS-LM have so little predictive power, why do people cling to them?

  11. Gravatar of Alex Alex
    20. May 2009 at 06:25

    “It is remarkable: if you ask, every macroeconomist says they believe in the quantity theory of money: MV = PY.”

    I thought MV = PY was just an equation, not a theory. In fact it is an equation that cannot be wrong because the quantity equation is the definition of the velocity of money*, i.e. V = PY/M. It is not a theory until you start making assumptions about how the variables in the equation move. Such as if Y and V are fixed then dM/M = dP/P or if P and V are fixed then dM/M = dY/Y or if V is fixed then dM/M = dNGDP/NGDP.

    *The “right” definition of the velocity of money actually uses transactions and not output. What we usually call velocity is the output velocity of money.

  12. Gravatar of Lord Lord
    20. May 2009 at 09:11

    The reason is I am not sure IS-LM subtracted
    . Monetary policy may be ineffective in a liquidity trap.
    Monetary policy had already been circumscribed into conventional routes through long standing practice. I doubt they would have been comfortable extending it even if they felt they had the authority and the means, and even if they felt unconstrained by the gold standard.

  13. Gravatar of Tom Tom
    20. May 2009 at 09:12

    DeLong may be wrong. Not every macroeconomist is a quantity theorist. There are endogenous-money/reverse-causationists who think, a’ la Joan Robinson, that economic activity determines the quantity of money. And there are still a few who believe, like John Law did in 1705, that permanent changes in the quantity of money determine permanent changes, not in the price level, but in real output. All, however, accept the equation of exchange MV=PQ. As Alex says,it is a truism, a tautology, an identity that cannot be wrong by definition. It becomes a theory when one posits that all the elements are independent of each other, or when one fixes, by assumption, the behavior of one or more of its elements (as when one assumes that the V/Q ratio is a fixed constant so that there is a constant proportional relationship between M and P).

  14. Gravatar of Thruth Thruth
    20. May 2009 at 11:21

    >I still don’t understand what the whole “misallocation of
    >capital” story tells us that is useful. Sticky-wage theory
    >already explains why nominal shocks lead to booms and busts.

    Doesn’t the misallocation of capital story get at the idea that small persistent errors over a period of time can add up to a big “nominal shock”. Whether the post shock effects are driven is sticky prices or imperfect capital mobility or something else is probably academic.

    I’m with you that interpreting the recent “bubble” as a Fed induced misallocation of capital seems like a tough row to hoe, and in particularly pointing at short-term interest rates in isolation is ridiculous. Inflationary expectations by the official statistics were never really too far out of line. Some people suggest that house price increases should somehow have been included in the CPI stats. Perhaps the bigger issue is that the Fed/bank regulators let banks push velocity (or subsitute money?) to unsustainable heights?

  15. Gravatar of Nick Rowe Nick Rowe
    20. May 2009 at 11:23

    Kevin: Wasn’t Samuelson in that 1968 quote praising Don Patinkin’s 1965 “Money Interest and Prices”, rather than the ISLM? Hence the reference to the (valid and invalid) classical dichotomies. My memory could be false.

  16. Gravatar of gnat gnat
    20. May 2009 at 12:06

    You may want to read Keynes’ Treatise on Money published in 1930. Its the GT in QT clothes. Its a tough read and Keynes says after laying out his model in “book 3”: “This line of thought anticipates much that is to come. If it is not fully clear, the reader will nevertheless pick up the drift easily enough if he returns to it on a second reading.”

    Keynes realized that money was endogenous and the QT framework was clunky in breaking out the savings-investment disequilibrium and its effects. Also, both the Treatise and GT are about disequilibrium as Hicks has conceded, and the ISLM does not capture what Keynes was modling. As Keynes said (in Tract on Monetary Reform):

    “The long run is a misleading guide to current affairs. In the long run we are all dead. Economists set themselves too easy, too useless a task if in tempestuous seasons they can only tell us that when the storm is past the ocean is flat again.”

    I took undergraduate Macro in the 60’s and what was emphasized was the sequence of adjustments in each sector that led to a new equilibrium. ISLM is just a short hand that tells us what the equilibrium is.

  17. Gravatar of Jon Jon
    20. May 2009 at 13:16

    “(This is why I keep harping on about the medium of exchange vs the medium of account function of money.)”

    This is one element that reads clearly in “The Theory of Money and Credit”. i.e., gold coin served as the medium of account but exchange was done using bank notes and cheques.

  18. Gravatar of happyjuggler0 happyjuggler0
    20. May 2009 at 15:05

    I have issues with Austrians claiming that monetary feast and famine is responsible for all or the bulk of economic boom/bust scenarios in the real world, or bubbles if you will.

    The majority of the US, geographically speaking, didn’t have a housing bubble. hence, they also aren’t experiencing a pop of that housing bubble, nor are they experiencing a general housing price decline for any reason.

    So clearly something besides money expansion was at least in part responsible for the numerous localized housing bubbles in the US.

    My explanation is pretty simple, albeit hopefully not in the obtuse sense of the word “simple”. 😮

    We all know that if demand exceeds supply, prices go up. In turn, this induces some combination of: new supply to come forth; existing demand to “drop out”; and substitutes to come into being or greater use. This drops the price back down again, either to the old equilibrium, or close enough to it for government work, or even past it (overshooting) for a period of time.

    With regards to housing, in Texas there has been a large influx of migrants from other states in recent decades. But since the mid 80’s anyway, there hasn’t been a general housing bubble (although certain choice luxury neighborhoods “suffer” from quasi-permanent lack of supply), the question is why not?

    The answer is simple. In Texas, there are minimal frictions to building new houses, so the tend to be priced as cost+.

    In a state like California however, it is notoriously difficult to build new houses. Environmental regulations, and just plain old bone-headed zoning regulations are the reason.

    Now go back to econ 101 price rule. Demand exceeds supply, price goes up. But what happens when there is a a significant time delay in new supply coming online? Prices go up still more. Eventually new houses come online, but they are responding to the old signal from perhaps two years ago. So they aren’t enough to measurably reduce prices, if at all, let alone to the old “equilibrium level”. Rinse and repeat and you get rising prices again and again, which eventually sparks rear view mirror speculative buying (gee, it’s going up, I think I’ll buy, after all they aren’t building land anymore) in addition to “organic” demand.

    Eventually new demand starts fading away thanks to higher prices, while the higher prices continue to spark new home building. Eventually the bubble bursts, just like all bubbles do, even if you don’t believe in bubbles, they exist folks.

    So how much of this was explained by Fed easing? The Austrians seem to be saying all of it, or the bulk of it. I argue that it was land use restrictions, coupled with governmental and private “enthusiasm” for low income buyers to buy, because after all, “they aren’t making land anymore, right?”. And so prices have to keep going up right? And so if the (literally) poor borrowers can’t finance their mortgage out of personal cash flow, they can pay the loan back anyway out of capital appreciation.

    Anyway, that’s my operating hypothesis, and I haven’t seen anything to contradict it. By the way, I’m not saying that the Austrian money boom/bust hypothesis is entirely inapplicable here, I’m just saying that it seems a stretch to say that it is the bulk of the explanation. Not all bubbles come from bad monetary policy, they sometimes come from bad fiscal or regulatory policy too, coupled with lack of critical thinking in the private sector which happens from time to time.

  19. Gravatar of happyjuggler0 happyjuggler0
    20. May 2009 at 15:10

    By the way, in my scenario of the CA housing bubble, builders keep building so long as they think they can get “cost plus” enough of a profit to justify building. They don’t stop until they get a “light bulb moment” where they realize that they have overbuilt and that enough demand doesn’t exist. But by that point there is “too much” supply, at least at the market price, which means that market prices have to crash down back to a more sustainable level.

  20. Gravatar of ssumner ssumner
    20. May 2009 at 17:38

    James, I suppose my “fed up” comment sounded petty and self-indulgent to some. I should do a post explaining my frustration. It was the way the profession responded to sharply falling NGDP, as if monetary policy had nothing to do with it. Which goes against the last 20 years of research in inflation and NGDP targeting rules.

    Alex and Tom, You are both right. Two things puzzled me: V is usually defined as P*Y/M, so it isn’t “remarkable” that economists believe MV=PY. And if you use the non-tautolgy version (assume V is stable), which is the “theory” part of the quantity theory, then not that many economists believe in it.) I think Brad meant that the QT does tend to hold in the long run, in the sense that changes in M cause proportional changes in P in the long run, and just got a bit sloppy. Many textbooks are vague about the distinction between the QT of M and the quantity equation.

    Lord, I think you need to distinguish between theory and practice. I agree that under the gold standard they weren’t comfortable doing what is necessary to inflate the price level. But they never really doubted that a sufficiently expansionary (unconventional) policy would do the job. If the Fed had set the nominal price of gold at $350 an oz, not $35 an ounce in 1933, nobody would have doubted the inflationary impact of such a policy. I had thought that under a fiat money regime the Fed would not feel restrained about doing what is necessary. I’ll give Krugman credit–he was right and I was wrong. The Fed is still reluctant to do what it takes once rates hit zero.

    Tom, Your comment is exactly right.

    Thruth, There was a lot of discussion of this a few weeks back, but I also had trouble seeing how the theory added much. Yes, it may have described the discomfort we experienced switching resources from housing to other sectors from mid-2006 to mid-2008. But I don’t see how it adds to our understanding of the recent big drop in NGDP.

    Nick, You may be right. But as a former Laidler student I presume you agree that Samuelson was being unfair to classical economists. I cannot recall a single example of classical economists claiming short run money neutrality.

    New classicals are another story . . .

    gnat, I like the Tract quite a bit, the Treatise a bit less, and the GT still less. I don’t remember the Treatise very well, but I do recall a distinction between income and profit inflation, which seems useful to me.

    I’m not a fan of 1960s Keynesian macro–it was the dark ages as far as i am concerned.

    Jon, I talked about the gold issue earlier—perhaps in my money theory post. The idea was that under wildcat banking the medium of exchange did not serve a dual role as a medium of account. I claimed that under that system what mattered was the supply and demand for the medium of account (gold or silver) not the medium of exchange (banknotes that traded at a fluctuating discount.)

    Happyjuggler0, I agree with most of your comments, especially about the role of the fed being overrated, and the Texas vs. california housing markets. But I aslo think people throw around the term “bubble” too loosely. I moved out here in 1982, before the bubble, and I don’t see the bubble as having ever burst. Sure there was a pullback in 1990, and again in 2007. All markets have ups and downs. But the bubble never really burst, and I don’t think it will. When I moved here in 1982 houses were not expensive in the western suburbs of Boston. (Relative to other cities.) Now they are, and I expect them to stay expensive the rest of my life. I don’t think this “bubble” will burst.

  21. Gravatar of happyjuggler0 happyjuggler0
    20. May 2009 at 18:22


    Thanks for your reply.

    In your recent bubble thread (posted after this one), I mentioned that I lived in Waltham. In fact I grew up in Needham, and lived in about a half dozen towns in MA (east, west, and the Cape) as an adult before leaving for good about 5 years ago.

    I left in small part because I thought housing prices were ridiculously high, and I also thought that they weren’t going down any time soon due to severe (compared to the bulk of the rest of the US) population density and the fact that Boston is a coastal city (hence half of its “natural” suburbs are in Boston harbor). Mostly however I left for professional reasons outside the scope of this post.

    I actually thought that housing prices in the Boston area might go down significantly recently, but I seem to be wrong. Oops. Perhaps the fact that Boston is “college central” with a new and ready supply of high productivity earners has something to do with it. Or perhaps there really are limitations to how densely you can build houses “on top of” one another. I now have old friends I grew up with who live in 495 towns, or near Rhode Island, who commute to the 128 region. Living in 128 towns is out of their reach financially.

    Anyway, I’ve long suspected (I’m not alone, this is hardly original), correctly or not only time will tell, that there will be a significant price decline in housing once the baby boomers start retiring, at least in areas they “used to” live, perhaps in the Boston area. Housing should either get more expensive (e.g. Cape Cod) in areas where there is a natural limit on new supply, or stay roughly the same in retirement areas though. Something to stew over if you are of a mind that your house (assuming you own) won’t go down in price in the next decade or so.

    Or maybe prices still won’t go down in the Boston area. I;ve been fooled before.

  22. Gravatar of Kevin Donoghue Kevin Donoghue
    21. May 2009 at 01:42

    Adam, Scott: thanks for the kind words. But it’s not hard to make a case when you have witnesses like Hicks, Samuelson and Pigou.

    Nick Rowe: Wasn’t Samuelson in that 1968 quote praising Don Patinkin’s 1965 “Money Interest and Prices”, rather than the ISLM?

    No, I’m sure Samuelson has great regard for Patinkin, but in that paper he is telling him how it was in the old days: “From 2 January 1932 until an indeterminate date in 1937, I was a classical monetary theorist.”

    Scott: I cannot recall a single example of classical economists claiming short run money neutrality.

    Samuelson isn’t saying that they did. He is saying that their short-run theory (such as it was) was divorced from their long-run theory. Keynes provided a coherent short-run theory which enabled Samuelson, for one, to see more clearly what the classical theory ought to be: “It is as if to understand Gary, Indiana, I had to travel to Paris. I began to understand neoclassical economics only after Keynes’ General Theory shook me up.”

    Obviously it’s possible to say that the classical economists had a fine short-run theory, if by theory you mean treating velocity as a function of something-or-other rather than a constant. To me that’s just waffle. But you’re quite right in saying that I don’t distinguish between new theoretical constructs and new ideas. If a theoretical construct isn’t an idea, what the hell is it?

    I can’t make sense of what you say (at 20 May 04:31) about supply and demand – the point Hicks is making in the passage I quoted is that, thanks to Keynes, we have a coherent theory of supply and demand at the macroeconomic level. Of course Keynes didn’t need to do much with the supply side; he just extended what Marshall had taught him to get his Z (aggregate supply) function. The demand side was trickier, since a Marshallian industry-level demand curve is drawn with national income taken as given. He needed a theory of income determination which didn’t simply assume full employment. That is what he developed. If you think saying “MV=PY where V=f(w,x,y,z) hence Y=MV/P” is just as good, well, I grant you it’s a lot less trouble.

  23. Gravatar of ssumner ssumner
    21. May 2009 at 05:46

    happyjuggler0, Years ago Mankiw made a housing prediction in the WSJ based on baby boomer demographics, and promptly fell on his face. Don’t assume that less baby boomers will lower property prices. Places like Boston, Manhattan and SF have low birth rates, but they have a steady inflow of yuppies.

    The high prices in Boston have nothing to do with availability of land, it’s all about zoning, about land with permission to build. We could build as many houses as Houston or Dallas if we had Texas-style zoning. There is plenty of land around here.

    Kevin, Maybe I should have been clearer about the theoretical construct point. To the extent that Keynes was correct, he restated classical ideas (like the short run non-neutrality of money) in a different language. His important new ideas (monetary policy ineffectiveness in a depression, lack of self-correcting mechanism to return economy back to the LRAS curve) all happened to be wrong. I don’t understand what Keynes added to the short run/long run distinction made by classical theorists that has any value.

    Why is fluctuating velocity a waffle? It’s like saying C+I+G=Y is the the Keynesian tautology and allow the MPS to change is a waffle. All theories start with tautologies, and then attempt to model each variable. I don’t see the difference. Does Keynesian theory allow us to predict business cycles better than the QTM? If so, where is the evidence? Does it give good fiscal policy advice?

    I thought you were arguing that IS-LM was analogous to S&D. If you are arguing that Keynes came up with AS/AD, that is flat out wrong. AS/AD was basically the model of most good interwar macroeconomists (Fisher, Hawtrey, etc) Maybe they never actually drew the graph, but that was clearly their model. Fisher’s macro model was basically the AD/SRAS/LRAS model we all teach in our intro texts. They simply defined the AD curve differently. They implicitly thought of AD as a rectangular hyperbola, which is still the way I teach AD. Yes, you can derive an AD curve than is not a hyperbola from IS-LM, but what is the point? The original purpose of IS-LM was to explain how monetary and fiscal policy shocks (as well as investment shocks) impact interest rates and output. But we know know it does that job very poorly. It is easier to think of AD shocks as simply fluctuations in M and V. The policy implications are just as clear, and they both predict equally poorly. So again, what’s the advantage of the far more complicated theory? Perhaps the answer is that some economists still believe that the Keynesian model produces useful estimates of “multipliers” and other such artifacts. After the past 15 years of Japanese history, I can’t see why people would still believe in that sort of stuff, but if they do, more power to them. I simply haven’t seem any output from the entire IS-LM approach to macro that makes me want to give up my much simpler analytical framework. Tell a specific world event that IS-LM explains better than MV=PY.

  24. Gravatar of Bill Woolsey Bill Woolsey
    21. May 2009 at 07:32

    I don’t like “V.” I like money demand. So, we define V = 1/k where Md = kPy. Well, that only is useful if people choose to hold money as a fraction of income. As Marshal said, people choose to keep a certain fraction of their income in the form of ready money balances. But, suppose they don’t?
    In fact, it sees that if the real income elasticity of money demand is different than one, then they don’t. And so, “k” is just a complication. Md = k(y)Py.

    Money demand is the sum of individual money demands (like the market demand for a product is the sum of individual product demands.) Money demand depends on opportunity cost, income, wealth… whatever. It is a good, with special characterstics. The quantity of money is created by profit maximzing banks, and of course, the Fed is involved.

    Velocity? Uneeded macroeconomic variable.

    Aggregate expenditure needs to be generated out of the supply and demand for money– not from — the average number of times each dollar is spent on final goods and services.

    IS-LM has plenty of problems, but it is at least based on the supply and demand for money.

  25. Gravatar of Devin Finbarr Devin Finbarr
    21. May 2009 at 07:36

    I agree entirely with Bill.

  26. Gravatar of Scott Sumner Scott Sumner
    21. May 2009 at 09:57

    Bill, I much prefer k to V, partly for the reasons you cite. I merely mention V because I think most people are less familiar with k. If you are interested in modeling the price level the most convenient way to is use nominal money supply and real money demand. If you want to model NGDP the most convenient way is to use k, which is nothing more than the demand for money as a fraction of nominal income (NGDP.) It is very similar to real money demand. You divide M by P*Y, whereas you divide M by P in the calculation of real money demand.

    Devin, I halfway agree with Bill (or is it 3/4?) I especially agree with his statement “IS-LM has plenty of problems”

  27. Gravatar of Kevin Donoghue Kevin Donoghue
    22. May 2009 at 03:31

    I thought you were arguing that IS-LM was analogous to S&D

    No, that would be like saying London is analogous to a city. IS-LM is a supply-and-demand model. Goods-market equilibrium and asset-market equilibrium are S&D concepts. Whatever merit pre-Keynesian arguments based on velocity may have had, they certainly weren’t any better grounded in S&D analysis than IS-LM was. Of course an IS-LM analysis, like any S&D analysis, is only as good as the underlying equations for demands and supplies. Hicks knew that. Your response to his comment doesn’t affect his claim that “Mr. Keynes’ innovation is closely parallel, in this respect, to the innovation of the marginalists.”

    As to what Keynes’s “important new ideas” were – all of them wrong, you say – don’t they include the principle of effective demand, whch Tobin describes as Keynes’s central idea? Is that wrong too, or right but well-known to Fisher under some other name? I’m quite willing to believe that the neoclassicals thought about what happens to the budget constraints of workers when they are fired. It certainly seems like something they should have thought about and incorporated into their analysis. But I’m not aware that they did.

  28. Gravatar of ssumner ssumner
    22. May 2009 at 04:05

    Kevin, I think you misunderstood my comment about S&D. I don’t doubt that IS-LM is “correct” in some sense, I question whether it is useful. S&D are useful because we can identify supply shocks. Thus we know a crop failure causes higher prices and lower consumption. We don’t know the impact of a monetary shock on interest rates–and that is perhaps the single most important shock that IS-LM was set up to explain. Monetary shocks can affect both the IS and LM curve—the stronger the monetary shock, the more powerful the relative impact on IS. I see that as a huge problem. I prefer to study interest rates in a much more disaggregated way, short vs long term, safe vs risky.

    Regarding income constraints, I suppose they could affect velocity, although it’s not clear to me that they do. Remember that if workers get laid-off because wages are sticky and prices are falling, then those workers not laid off (the vast majority) see their real income rise. So I don’t know the net effect on “effective demand.” I don’t use the term effective demand, I simply try to explain nominal spending (which of course equals nominal income=nominal GDP.) I think it can best be explained by monetary policy. If monetary policy is constrained (unable or unwilling to offset velocity changes, as in the gold standard) then you also have to model velocity, or the real demand for money, or some analogous concept.

    Keynes tries to explain nominal shocks by independently modeling C, I, G, etc. But that seems to get things exactly backward. Why did Zimbabwean NGDP soar last year? Was it C? Was it I? Was it G? Or was it too much M? I don’t see “M” in the tautology that Keynesians use (C+I+G+NX=Y) But it does appear in another famous tautology.

  29. Gravatar of The Fisher-Friedman-Sumner-Svensson axis « The Market Monetarist The Fisher-Friedman-Sumner-Svensson axis « The Market Monetarist
    14. November 2011 at 11:45

    […] is Scott Sumner in […]

Leave a Reply