Thursday, July 25, 2013

Learn the basics of New Keynesian models without hard math

This post by Miles Kimball is a master class in how to explain ideas effectively. In one post, he explains the core ideas of what he calls "Neomonetarism", which I tend to just call "monetarism", and which most economists would refer to as "New Keynesian models with investment". This sort of model is squarely in the mainstream of what modern business cycle theory. Miles' academic work on Neomonetarism in the 1990s is heavily cited by Frank Smets and Raf Wouters, creators of the workhorse New Keynesian model used by central banks.

Miles manages to explain the ideas of the model without any reference to difficult math, while providing a clear and lucid description of most of the "microfoundations". He also explains what the "natural rate of interest" means in the context of these models (hint: there are actually two of them). Note that the policy recommendations that pop out of Miles' model are only slightly different from what Scott Sumner, David Beckworth, and other "market monetarists" have been advocating through their blogs.

So if you want to understand the way that mainstream business cycle theorists think about the macroeconomy, but you don't have the time or expertise to pick through a DSGE model, Miles Kimball's explanation is your constrained-optimal strategy for achieving understanding.

OK, that was a short post.

What, you wanted me to argue with the model? OK, OK, I guess I can play Devil's Advocate.

What are the potential shortcomings of this sort of model? In a 2012 speech (flagged helpfully by Steve Williamson), Charles Plosser lays out a number of them. Note that Plosser is a pioneer of the rival Real Business Cycle type of modeling, which generally lost out to New Keynesian models as the dominant form of business cycle theory, though it remains a substantial minority view. 

A...friction often assumed in Keynesian DSGE models is that firms and households have to wait a fixed interval of time before they can reset their prices and wages...A rule of the game in these models is that the interactions of these nominal frictions with real frictions give rise to persistent monetary nonneutralities over the business cycle... 
[Realistic price adjusting strategies] are ruled out by the “rules of the game” of New Keynesian DSGE modeling. 
Another important rule of the game prescribes that monetary policy is represented by [a rule] that policymakers are committed to follow and that everyone believes will, in fact, be followed...[T]his commitment is assumed to be fully credible according to the rules of the game of New Keynesian DSGE models. Policy changes are then evaluated as deviations from the invariant policy rule to which policymakers are credibly committed. 
I have always been uncomfortable with the New Keynesian model’s assumption that wage and price setters have market power but, at the same time, are unable or unwilling to change prices in response to anticipated and systematic shifts in monetary policy. This...raises questions about the mechanism by which monetary policy shocks are transmitted to the real economy in these models... 
From a policy perspective, the assumption that a central bank can always and everywhere credibly commit to its policy rule is, I believe, also questionable...[C]ommitment is a luxury few central bankers ever actually have, and fewer still faithfully follow...Policy actions have become increasingly discretionary... 
Given that central bankers are, in fact, acting in a discretionary are we to interpret policy advice coming from models that assume full commitment to a systematic rule? [A] number of central banks have been openly discussing different regimes, from price-level targeting to nominal GDP targeting. In such an environment where policymakers actively debate alternative regimes, how confident can we be about the policy advice that follows from models in which that is never contemplated?
What Plosser is arguing is that New Keynesian models have unrealistic microfoundations. Specifically, the "sticky price" mechanism and the "monetary policy rule" mechanism, both highly important for New Keynesian models, don't seem realistic to Plosser.

On one hand, this sort of criticism might sound a bit rich coming from a pioneer of RBC theory, whose microfoundations - perfectly flexible prices, technology that vanishes from human knowledge, etc. - are even less realistic. But I think the criticism itself is worthwhile.

The "monetary policy rule" is not as big of a problem as Plosser thinks, because more recent New Keynesian models, like the aforementioned Smets-Wouters model, allow the monetary policy "rule" to fluctuate randomly, meaning that they don't assume that the central bank can perfectly commit to future policy.

But the "sticky price" microfoundation remains troubling. You really do need sticky prices for these models to work. More specifically, as Miles points out in another paper with Bob Barsky and Chris House, you need durable goods prices - houses, cars, TVs, etc. - to be sticky. But why these prices would be sticky, and when they would be more sticky or less sticky, is not well understood.

Yes, sticky-price models do appear to fit the macroeconomic facts much better than do "flexible-price" models like Plosser's RBC. But sticky prices could just be a stand-in for something else. As Greg Mankiw and Ricardo Reis point out, "sticky information" - i.e., people simply not paying enough attention to price changes -  gets you similar results.

So I'm not saying I think we should chuck the models of Miles, Smets/Wouters, etc. out the window. Far from it. First of all, they may allow slight improvements in forecasting (and I am not as quick as most macroeconomists to dismiss the importance of forecasting!). And on the policy advice front, well, the evidence is strong that monetary policy sometimes really does work in the way we think it should (a clear and simple example being the Volcker disinflations; more complicated examples involving impulse responses and event studies and moments of aggregate joint time series are also available).

I just think we don't yet really know why monetary policy sometimes works. And because we don't know why it works, there's the possibility that monetary policy will do screwy things, or fail to work the way we think it should, in some situations. A liquidity trap situation, with prolonged deflation, may be one of those situations. I think Miles' "electronic money" idea should be tried (especially in Japan), but I'm not as confident as Miles that it will work as advertised!


  1. Unanimous6:37 AM

    I don't know about the Volcker disinflation. It happened to coincide with demographic changes that also appeared to be a disinflationary force. You need to check other country's records to try to get a feel for the demographics and monetary policy effects to see which is more of an explanation.

  2. Noah,

    As long as investment/consumption decisions depend on the interest rates, doesn't that provide a ready answer to why monetary policy works? Or do you mean it in some different sense?

  3. Let me offer the reverse Lucas Critique: sticky prices can be a convenient way to model non-price related frictions in markets. That is, if you ignore inaccurate price predictions (as RBC and New Keynesian practitioners often do), basically any failure of a market to clear can be modeled as a price discrepancy in an otherwise friction-less market. That is, using parameter values derived from empirical micro is not guaranteed to produce the best prediction when you have the wrong model. Hence, I tend to favor macro models with somewhat more price stickiness than what the micro data would suggest.

    And for the record, New Keynesian models have been much more successful in predicting prices than have RBC models.

    1. Anonymous11:54 AM

      Markets clear in these models.

    2. Ok, I used imprecise language. They clear in the sense that firms always adjust supply to match realized demand. Two points: 1) in the new keynesian model markets do not clear in the sense that firms and households make plans that are not realized--ie the firm plans to sell goods at its optimal price, but sticky prices prevent it from realizing that price, and thus the firm does not actually sell the planed amount. 2) markets only clear by assumption. The New Keynesian model assumes that firms are obligated to supply whatever amount consumers want at their posted price. This simplifies the math a lot, but is definitely not profit maximizing--if sticky prices prevented firms from raising prices, for example, the profit maximizing behavior would involve restricting supply, causing excess demand and failure of the market to clear.

  4. DSGE models always seemed to me like they are more work than they are worth. But I suppose if you have a large research team constantly using them they become fairly routine.

  5. Williamson would probably also argue that these models lack "deep money". Notes and bills appear in utility functions without any effort to explain how and why they gets there in the first place. Frictions related to monetary exchange get ignored.

  6. I agree that the detailed nature of the underlying inefficiencies that support nominal rigidities are not totally clear. But I think you are really missing the big picture here: *RATES MATTER*.

    If the fed hiked rates to 10% right now, inflation would plummet, the real rate would sky rocket and the economy would plummet. In rbc models, inflation would just rise by 10% because the real rate is always at the natural rate (how wrong is that!?!?).

    If there is one thing anybody should understand about good macro models it's that 1) the real rate matters and 2) the cb sets the real rate.

    If there is a second thing you anyone should understand about macro models it's this: because of the above dynamic the liquidity trap is an inescapable fact. If the nominal short rate is above equilibrium, the real rate will tend to go *up*, not down with time.

    I there is a third thing it's that QE is snake oil (long story - told it many times).

    None of these facts depend in any way *WHATSOEVER* on the exact pricing dynamics, or details of the competitive inefficiencies. They are 1) an absolutely necessary feature of any non-useless macro model and 2) all that is required to show that the liquidity trap is a problem. NK is just happens to be an analytically tractable framework that includes these required features. The rest is details.

  7. Do these models contain sticky nominal debt obligations as the way households and firms finance spending?

    1. Krugman's debt deleveraging paper is a New Keynesian model where some households finance spending through nominal debt obligations. Typically though, debt is not explicitly modeled in most New Keynesian models (since aggregate net debt is zero, models with a single representative agent simply can't model debt).

  8. AlanInAZ11:35 AM

    My first impression after reading the post is that the model does not conform to my real world experience. I worked for a multi-national manufacturing company for more than 3 decades and was involved in many investment decisions. Every decision was based on global demand and global manufacturing alternatives. It is not clear how this global element fits into the model.

  9. The one stickiness that is clearly empirically correct is the downward stickiness of nominal wages. While it is convenient for the dominant Smets-Wouters model (I'll accept that it is derived from the work of your mentor, the admirable Miles Kimball), everybody has always known that the Calvo pricing mechanism is a goofball as silly as rational expectations.

    However, it should be kept in mind that while these models are not ratex or a variation on DSGE, there are plenty of models around, mostly in the Post Keynesian tradition, that emphasize that one can have downturns and business cycles even with perfectly flexible prices and wages. Indeed, Keynes himself emphasized this point in the GT, although many either never knew this or have conveniently forgotten it. One obvious mechanism is when one can have prices that "overshoot," as with bubbles, although, of course, we know that either those do not happen or they can be handled by introducing "financial frictions" (yaaaaay!!!) into DSGE models of the NK or some other variety (hack, couhg).

    Barkley Rosser

    1. However, it should be kept in mind that while these models are not ratex or a variation on DSGE, there are plenty of models around, mostly in the Post Keynesian tradition, that emphasize that one can have downturns and business cycles even with perfectly flexible prices and wages.

      So far the models of that type that I have seen model macroeconomic aggregates using systems of differential equations. In other words, no microfoundations at all. While I am not averse to non-microfounded models, why use a complex one instead of a simple one?

    2. You are thinking of Keen's model. I am thinking agent-based ones as done by people like Gallegati and coauthors (including me), Chiarella and coauthors, Hommes and coauthors, Westerhoff and others. Westerhoff's stuff has been used by the EU to study effects of a Tobin tax (answer, depends on micro market structures). You may dislike the microfoundations various agent-based models have, not ratex and thus "ad hoc," according to many, but they are microfoundations. And those models most definitely can give you bubbles and all that.

    3. Sounds excellent! Link me to a couple papers!

    4. There's a paper by Westerhoff and Franke here which has the very basic model(s)

      Basically, what you wind up with is an AR(2) process with time varying coefficients - time varying, because they depend on the relative share of agents using a particular strategy which also evolves over time.

      The "ad hoc"/microfoundations, or in other words, the point at which the rabbit gets stuffed into the hat (which is fine, all models have to put that bunny in there at some point) are essentially the "attractiveness" equations, though I understand that in some more sophisticated papers they examine the robustness etc. of various potential rodents.

      Personally I was wondering to what extent the fluctuations the models generate reflect true endogenous dynamics (which is why you need the 2 in the AR(2)) rather than the outside shocks that are assumed. That was/is a big criticism of the RBC, and many of the NK models; that they generate "realistic" cycles by assuming just the right exogenous shocks. Unfortunately when I try to replicate the dynamics for the second model (the goods mkt one) I get explosive fluctuations. Which means either I've made a typo somewhere or there's a typo somewhere in the paper (perhaps one of the values of the coefficients from the first model got accidentally imported into the second one?)

    5. Nathanael11:02 PM

      It's worth noting that every single result in psychology says that *people* are pretty much ad hoc!

      A mishmash of different behaviors triggered by different specific situations, with people showing a great lack of consistency from one situation to another.

      Accordingly, assumptions which appear "ad hoc" are very likely to be the correct assumptions, if they are based on real psychology results.

  10. JohnR4:38 PM

    Perhaps I'm missing something elementary (as I often do - I am but a man), but "I don't understand how it works, so it can't work" doesn't seem to be a particularly strong criticism of anything, model or not. Perhaps a bit of creative relabeling of variables would help? Instead of 'the "sticky price" mechanism and the "monetary policy rule" mechanism', why not just call them "Thing A" and "Thing B"? That way there need be no agitation that the label leads one down a garden path into the Slough of Confusion. Maybe this could be the start of a New School of Economics - the Humpty-Dumpty School, where any variable label means only what the modeler intends it to mean, no more and no less. It might be confusing at first, but think of the efficiency - one model could do for almost anything, if the meaning of the factors involved simply arbitrarily changed for each modeler. It could be called the Universal Schoen Model (after Jan Hendrik Schoen of happy memory). I leave the details (if any) to bigger brains than mine; I'm more of a big-picture kind of guy. Naturally, if all this is based off of a silly misunderstanding, I invoke the Emily Litella Defense.

  11. Noah,

    I just read Roman P's comment above, ie the short version of what I said: monetary policy works because real rates control investment and consumption decisions. If, following Miles' proposal, the BoJ drops the short rate to -5%, the 2s/10s curve will steepen about 10% which will absolutely set both consumption and investment on fire for a few years, just what the doctor ordered. This just isn't mysterious. It's basic micro.

    1. Agreed with K. I don't know why every blogger I read these days is suddenly skeptical of the effects of lower interest rates. Drop rates and either prices or quantities or some combination of the two will rise. I'm not sure why anyone would be skeptical of Kimball's negative rate plan, except for the fact that getting it implemented seems like an impossible task.

    2. JP Koning, K,

      Well, it sure might be not that straightforward. For example, for some the interest on loans is income, so the impact of lowering interest rates on AD is likely to be mixed. Next is the question of how sensitive investment is to the interest rates: some propose that it is relatively insensitive. After all, interest rates are costs of investment, but ultimately the decision to invest depends on the expected profits, so there might be a lot of investment with high interest rates (prospects are good) or stagnated investment with low interest rates (prospects are bad). The last, even if monetary policy of a central bank is able to set a nominal interest rate, how a real interest rate is formed is a question that I don't think is decisively solved. If low nominal interest rates cause low inflation (per, say, mark-up pricing process) than setting interest rates lower is kinda self-defeating: real interest rates won't fall so much.

      Though, all in all, I think that higher interest rates do lower aggregate demand. Besides, maybe some of the objections I put above are bunk.

  12. Or to paraphrase an old post of Noah's... Interest rates... they work, bitches!

  13. What does mr. Kimball has against workhorses?

    I didn't understand why the IS curve was so quickly dismissed. 'No micro-foundations' and its gone; the next paragraphs criticize alternative interpretations of the IS. Other attempts at simple NK models for teaching and policy usually have it, even if they discard the LM curve: Carlin and Solstice, Romer.

    David Romer has a micro-founded IS curve, as noted by the awesome Brad DeLong

    To be honest I don't know enough about the consumption Euler equation controversy, but it still might be a useful link.

    But the important thing to me is that the IS curve still "comes from somewhere" even if it doesn't have micro-foundations and you get a pretty useful model - a workhorse. That is why IS-LM still refuses to die when you force economists to make judgements (outside of academia).

  14. "So far the models of that type that I have seen model macroeconomic aggregates using systems of differential equations. In other words, no microfoundations at all. While I am not averse to non-microfounded models, why use a complex one instead of a simple one?"

    I find this remark incomprehensible.

    Models with one good or with one (representative) agent do not have microfoundations. I do not know of in what sense microfounded models are simple.

    Part of the point of the study of (low-dimensional) systems of differential and difference equations is that models with a simple structure can can complex structural and temporal dynamics. This point has been made in economics, including with Post Keynesian economics.

    I was particularly surprised to find that this point can be illustrated with a discrete time version of Kaldor's 1940 non-linear multiplier-accelerator model. A formalization yields a two-dimensional system of difference equations with interesting bifurcations and, thus, varying dynamics, depending on the location of the model parameters in phase space.

    If you want to confine your attention to neoclassical models, a similar point can be made in OverLapping Generations (OLG) models. (I have somewhat explored bifurcations of equilibria in such models in Cambridge Capital Controversy technologies. The CCC also shows that the concept of the natural rate of interest is incoherent and should be abandoned - see Coling Rogers work.)

    Anyways, one who understands complex dynamics might realize that regularities in prediction errors needed to ground the Lucas critique of previous macroeconomic models need not arise in those models. Thus, the Lucas critique may need to be rethought or rejected for now. Agliari, Chiarella, and Gardini make this point in a 2006 JEBO article.

    1. I find this remark incomprehensible.

      Well whose fault is that? ;-)

      Models with one good or with one (representative) agent do not have microfoundations. I do not know of in what sense microfounded models are simple.

      In general, they are not. Note that I never said they were.


      Part of the point of the study of (low-dimensional) systems of differential and difference equations is that models with a simple structure can can complex structural and temporal dynamics. This point has been made in economics, including with Post Keynesian economics.

      Naah, as soon as you take them to data, it'll be curtains.

  15. It seems to me that there is an important difference between prices being sticky in general, and prices being sticky in the face of monetary policy innovations. The latter is required* for monetary policy to work in these models, but a lot of the discussion around price stickiness seems to be about why firms might be reluctant to change prices in general, and the empirical work likewise.

    Can anybody point me to any discussion of this distinction?

    * is required too strong a word?

    1. It seems to me that there is an important difference between prices being sticky in general, and prices being sticky in the face of monetary policy innovations.

      Yep, that's the good old Lucas Critique!

      Can anybody point me to any discussion of this distinction?

      Well, see Plosser's speech itself. There is a literature on this but I forget the names of any papers or authors. Try Google!

    2. I think I see what you mean, although if I understand you correctly I think we are coming at this from different directions.

      I think you are suggesting that even if prices look sticky empirically, behaviour might change if the policy regime changes. That sounds like the Lucas Critique to me.

      I was rather thinking the opposite - prices might look pretty flexible, empirically, but that doesn't mean price changes neutralize monetary policy. Firms might know how to changes prices for reasons of competition or in response to changing costs or patterns of demand, but need not know how to change prices in response to monetary policy.

      Quoting Gali's textbook "monetary non-neutralities are, at least in theory, a natural consequence of the presence of nominal rigidities ... if expected inflation does not move one for one with the nominal interest rate when [it] is changed. "

      Even in the absence of menu costs or whatever, I'd be surprised if firms adjust prices in that fashion.

  16. Roman P,

    "Though, all in all, I think that higher interest rates do lower aggregate demand."

    If the Fed raises the short rate to 1%/week and says it will keep it there for a year, do you *think* demand will drop, or do you know it?

    1. K,

      Uh, I'm sort of rusty with financial mathematics, but effective interest rate for a period of a year with 1%/week ought to be ~73%. If prime interest rate is a federal funds rate + markup (which as I gather is non-controversial in finance) then interest rates on loans will also rise by about 73%. I think this will by itself prevent a lot of investment projects and will significantly lower AD. I am not sure what do you mean by *knowing* it - I am not a professional macroeconomist, but even those guys could only be reasonably certain of those non-self evident propositions. But this goes too close into uncomfortably philosophical territory...

      I actually think that this hypothetical scenario will lead to disruption of production and investment on a very wide scale. A lot of firms will have to go bankrupt, they will lay off workers, recession will deepen even more. High interest rates will translate into inflation which will disrupt everything even more, despite money being very tight and credit unavailable for most. Economy of USA will collapse and dollar will be substituted by bottles of whiskey, cigarettes and euro. Something like that happened with Russia in the early 90's, actually...

  17. Anonymous12:53 PM

    A footnote regarding economic history. Smets and Wouters is a wonderful econometric paper. But it implements the model developed in Christiano, Eichenbaum and Evans, JPE, 2005. The latter used partial information statistical techniques. Smets and Wouters estimated the same model (except that they had more shocks) using full information maximum likelihood techniques.

    1. Yup!

      I just wanted to point out that Miles' work and Smets-Wouters are of the same family.

      Note that Miles' work on NK models with investment predates Christiano, Eichenbaum, and Evans by about a decade. CEE include a lot more real frictions, like habit formation and investment adjustment costs.

      The added shocks of Smets-Wouters make it a lot better at fitting aggregate time-series. So it's not just a new estimation technique, though that is what is emphasized by the name of the paper.

  18. Roman P,

    I agree with the first part of your answer but I think it gets muddled right around high rates leading to high inflation. You need to distinguish short and long run effects. My point was that the short run dynamics are unequivocal in terms of direction and critical to get right in a macro model. In the short run higher rates lead to decreases in consumption and investment (and therefore disinflation). In the long run (eg Russian default/inflation) everything depends on the details of the microstructure and the central bank's response function. The Russian inflation, in fact, occurred because the monetary authority was caught between keeping rates high enough to control inflation and low enough to prevent default. In the end they got both. RBC models hold that (short run) inflation rises pari passu with short rate increases. This is patently false both from a theoretical and empirical perspective. The successful use of Taylor rule - type inflation targeting the world over, during the past 20 years or so, as well as the volcker disinflation (and Swedish crisis with the short rate as high as 500%) provide plenty of evidence that inflation varies negatively with nominal rates and that central banks therefore control the real rate (and therefore liquidity traps are a mandatory feature in any macro model if there are possible states of negative nominal natural rates).

  19. Honestly, we would never have been having this debate before 2008. The theory and evidence was and is overwhelming. And how, on earth, evidence gathered from the past 5 years of no nominal rate changes can lead anybody to see new evidence for the impact of nominal rate changes on inflation and output is beyond me.


    Obviously I've adopted a bad tone again in these comments. Maybe I shouldn't. The reason is that I find this kind of wishy-washy post to be just a more sophisticated take on opinions-on-the-shape-of-the-earth-differ type journalism. Important proposals, such as Miles', deserve better and will require (for starters) a high level of public support from intelligent public voices. "Who knows, we should just try it!" isn't a vote of confidence under the circumstances with powerful right wing organizations (BIS, OECD) clamouring for rate hikes. Miles is a passionate advocate of policies he believes can lead to the improvement of the human condition. Sometimes I don't feel the urgency in your writing.

  20. Part 1 of 2

    Interesting post! I am not sure whether this post is intended for non-economists. However, I think it is important so I have four comments.

    First, before you can communicate with ordinary people about your models you need to understand the questions that ordinary people ask about these models.

    Think about something important which you don’t normally think about e.g. the sewage system. If there were a problem with the sewage system which threatened you, your family and your property, you would want to know the answers to certain questions:

    What is the cause of the problem?
    What are you going to do to fix it (including discussing options, and their pros and cons, if appropriate)?
    When will it be fixed?
    How do we know it won’t happen again?

    For ordinary people, economic models are like the sewage system. “The economy is broken. Economists failed to predict it. There must be a problem with their models. It is very important that economists fix their models”. We then ask the sewage system questions about economic models.

    Noah’s post doesn’t address these questions. I am interested in economics enough to search actively for answers to the sewage system questions. However, it feels like the sewage system manager is offering me a tour of the sewage treatment plant to show off its latest technology, and is not even aware of my questions.

    Second, when I think about economics, the interesting bits are the things that scare me. I want to understand them because they scare me: stock market bubbles, depressions, bank runs, hyper-inflation etc. All of these things are examples of herding behaviour. The issue is not whether people are rational or how often businesses change their prices. It is that most of the people behave in the same way at the same time. It seems to me that the models discussed in this post are unlikely to detect herding behaviour as they seem to be focused on the behaviour of an individual. In combination with my first point, and the fact that these models also don’t include the banks that were at the centre of the current crisis, I get a sinking feeling in my stomach.

    Third, although he would use different words, Joseph Stiglitz has spoken on several occasions about the many flaws in existing models and that they need to change. Why do other economists think that Joseph Stiglitz is wrong about this and that it is sufficient to carry on with the same models as though nothing is wrong? I trust Stiglitz because he understands my questions EVEN IF HE DOESN’T HAVE THE ANSWERS. I don’t trust economists who are oblivious to my questions.

  21. Part 2 of 2

    Fourth, I only skim-read Miles’ post as it didn’t seem to address my earlier points. However, I picked up enough to understand that one of the main reasons it is “a master class in how to explain ideas effectively” is that Miles takes great care to define and explain his terminology.

    I note that economists like to give their ideas silly names: Austrianism, Keynesianism etc. For my own amusement, I also give economists silly names. For example, a common characteristic in economics is ‘Greenism’. Greenism is best explained by an example.

    Greenist economist: The sky is green.
    Ordinary person: No the sky is blue.
    Greenist economist: No, you don’t understand how economists use the word ’green’.

    There are other sub-schools of Greenism with different behaviours. For example, using one word to mean different things without making clear which meaning is appropriate in a given sentence; using many different words to mean the same thing (see first paragraph of Noah’s post when, apparently, I am supposed to be impressed that Miles, Noah and other economists have three different names for the same thing); using metaphors to describe the world, oblivious to the fact that the whole point of science is to move from a metaphorical understanding of the world to a more realistic one; giving each other silly names which aren’t well-defined e.g. neoclassical, and then arguing about it incessantly; defining each other by the names of specific policy biases and then arguing that economists are unbiased scientists.

    Economics looks different when viewed through an understanding of Greenism. For example, you begin to notice that, on a fairly regular basis, Paul Krugman will intervene in a blogosphere debate and say something like “Guys, you realise that these things you are talking about are just metaphors. They aren’t real”. These interventions always make me smile. At least Krugman understands Greenism and that it is one of the biggest sources of economic derp. I’d also point out that the people with the worst track record in the current crisis are the ones who think that ‘money printing’ (metaphor) will cause the ‘money multiplier’ (metaphor) to do its thing and that hyper-inflation is the only possible outcome. Derp derp derp.

    Miles should be applauded for defining his terms clearly. However, economists should realise that defining terms clearly is a fairly low hurdle to pass. Economics needs an anti-Greenist coalition. I think you would be surprised at the members of this coalition. (I have a theory that anti-Greenism is one of the reasons that the Post Keynesians split from the mainstream, although the Post Keynesians are themselves the authors of ‘neoclassical Greenism’ where all problems can be resolved by calling someone else a neoclassical economist).

    1. Nathanael10:59 PM

      FWIW, regarding Greenism, I've been complaining that genuine money printing -- mailing newly printed $100 Greenbacks to everyone -- would actually work, but that nobody has been willing to try it.

  22. Two further points on this.

    First, economists have a very narrow view of the word ‘model’. I came across some YouTube videos recently which use a very different type of model to tell economic stories. I think that these models are very effective communication tools for difficult economic ideas. I have attached a link below.

    The key points to note are not related to any specific policy message, or to promoting these specific videos or their producer. They are:

    These models are of are a very different type to the ones discussed in Noah’s post. For example, they are visual rather than mathematical. We understand most things using visual models but economists almost never use such models
    They mostly just tell stories about how stuff works
    The symbols used in the diagrams are consistent and represent real world objects such as banks and households
    The stories are mostly about the relationships and interactions between the objects, and how they can cause crises
    There is virtually no made-up econo-derp about understanding the world through metaphor etc.

    The type of credible economics I am looking for would make use of this type of visual model (or any available variant) to explain how the economy works in easy to understand chunks. It might also use this type of model to discuss the implications of different policy options in an unbiased way. I understand that older economists such as Paul Krugman are too set in their ways to use novel techniques. However, I don’t understand why younger economists don’t employ this type of technique. Lack of imagination? Drowning in econo-derp? Too much focus on fixing broken mathematical models?

    Effective economics is as much about the communication of the message as it is about the message itself.

    Second, in my attempt to work out which economists, if any, to trust, I have been trying to work out why the Post Keynesians split from the New Keynesians and why they hate the mainstream so much. The New Keynesians don’t discuss this as they barely acknowledge the Post Keynesians at all. Meanwhile the Post Keynesians discuss this subject in very technical terms which require economic knowledge beyond my abilities.

    Yesterday, I found the following Post Keynesian post.

    It is taking issue with the definitions in the post by Miles Kimball recommended in Noah’s post. The Post Keynesian response is critical of Miles’ definitions. However, that is not my point. Note that:

    The dispute is about the definition of basic terms
    The Post Keynesian response is only possible because Miles was very clear in his definitions
    If there are significant debates around such terminology then I’d suggest that there is no chance of meaningful debate on ideas and policies based on the disputed terminology.

    One of the crucial events in the Post Keynesian break-away seems to have been something called the Cambridge Capital Controversies (CCC). These appear to have been about the precise meaning of the term ‘capital’ although I don’t understand the detail. Note that the criticism of Miles’ definitions ends with a link to a post about the CCC.

    Definitions of words matter, particularly if economists refuse to use visual models. Without effective words and pictures all that is left is derp.

  23. Nathanael10:57 PM

    Prices are sticky downward because people don't like to sell at a loss. Which is a known result in psychology.

    That is all. We don't need any more microfoundations than that.