Tuesday, December 18, 2012

Another Crank

As critic of modern economics, I am often contacted by other critics, or fans of other critics. Alas, I often find them not even wrong, as in
2 + zebra ÷ glockenspiel = homeopathy works!
 Steve Keen was recommended, and I looked him up. Interestingly he's a huge Minsky fan. I admire Minsky, because he was intellectually honest and curious, and had some really good insights. If you read his books you'll learn a lot about US macro history from a financial perspective. However, he wasn't perfect. He dismissed microeconomics as apologetics, and so basically ceded that whole field as irrelevant. He didn't try to convince his colleagues, rather, he would dismiss them as fools, and so they would dismiss him. He was forever anticipating another 1929 crash, and would get very excited if the Dow was down a lot intraday (alas, back in the 1980's it would always bounce back). His macro model he failed to concisely formulate, not just in a model, but even in words.

 Minsky should have said his theory was based on endogenous instability from excessive leverage, and then one could have tested it, and found it doesn't work: aggregate leverage is not a great leading economic indicator. I too believe the economy is endogenously unstable. That is, unlike Friedman and other free-marketers who believe recessions are caused by government, I think it usually happens simply via systematic errors among private investors. But unlike Minsky, I think it's more micro-based, happening idiosyncratically in a different subsector of the economy. It's more like an eco-system of Batesian mimicry, where over time mimics fester and create a phase shift in the system as predators learn that, eg, all those poisonous snakes are actually non-poisonous, and a havoc occurs until equilibrium is restored.

The key is that the excesses occur in different industries, using different metrics, every cycle. Off the top of my head, here are the focal points for some big recessions; railroads (1893), conglomerates (1969), oil and real estate (1990), internet (2002), residential real estate (2008). What's consistent is that they are all different. Every cycle is predicated on some new 'new thing' that survived the prior two or more recessions, and thus to most participants is their entire working life.

 Like Minsky, Keen was predicting a crash, and so when one happened, he took credit, just like others (Roubini) with a vague, persistent prediction of a crash whose reputations were burnished by the 2008 crisis.  It makes me think that if something unpredictable happens, like the Mayans really do come back this Friday, those who predicted it won't be prescient, but rather, cranks.  In an case, reading Keen I see he picks up Minsky's dismissal of modern economics. However, Keen takes this to the next level by saying that economics is based on a math error. Now, economists I think have made a mistake, but they are pretty good at working from assumptions to conclusions, that's brute force logic, and there are lots of really smart economists who can solve logic puzzles.

He seems fixated on the market power of firms that are small, not infinitesimal  He thinks they act like monopolists, showing a math error at the root of Samuelsonian neo-classical economics. It reminds me of that guy who thinks fat tails or stochastic parameters invalidates financial theory.  This is simply wrong, but as a rhetorical device I think it convinces a lot of people that he's proved the existing theory is wrong, thus in some way proved his alternative is correct (not that this is logical, just it seems to work on many readers).  He then builds up a pretty standard 1970's macro model to show how money is endogenous and whips us through cycles, as if this wasn't tried for a generation and failed. These models have so many equations and parameters one can't prove them wrong, but then, they don't have unambiguous predictions, just fit the past really well. Thousands have wasted their lives on these approaches. A good way to judge a jumble is to look at the results: what to do? His solution to current problems is to simply print money and pay off everyone's debt. Such a solution doesn't deserve much consideration, because when I have time to think of wacky theories, I prefer Ancient Alien Astronauts or stories of the Mayan Apocalypse.

Like other successful cranks, Keen does seem intelligent, and makes some very good points. But net net, his Weltanschauung has more flaws than what he's criticizing. It's easy to note that existing theory is deficient, much harder to present a coherent, more attractive alternative.  It reminds me of a high school class I took where we all got to give a speech criticizing some great thinker of the past, like Socrates, Nietzsche, or Kant. We all did great. Then, we had to present our own new theory on something important, and it was pretty humiliating. Lesson learned.

I hasten to add I  don't hate all macro critics. For example, I really enjoyed Peter Schiff's The Real Crash. And of course, I think I have a rather pointed criticism of finance that doesn't throw the baby out with the bathwater, and has clear empirical implications. But, I realize I am lumped in with the numerous other critics, and that's kind of depressing.


  1. Anonymous12:51 AM

    You prefer Peter Schiff to Steve Keen? Wow. How long as Schiff been dead wrong on hyperinflation? What's his model? Does he offer any real mathematical or scientific theory? Keen's not perfect but at least he has a model that correctly predicted the crisis and the ensuing debt deflation we are experiencing.. Schiff is a blowhard that's been screaming about the "end of the dollar" for over a decade.

  2. I like Steve Keen, I think he fits the bill on intellectually honest and curious, and had some really good insights.

    Unfortunately, Keen uses the Karl Marx theory of surplus value, and also divides the economy along the Marxist "Worker vs Capitalist" split. Then again, at least Keen is honest about his Marxist theoretical underpinnings, because lots of other economists do the same without being as open.

    Keen correctly points out that money is created primarily by banks (not by government) although government of course these days is closely entwined with the banks (only what Rothbard was pointing out decades ago).

    Keen also demonstrates that classical "equilibrium" does not necessarily ever happen because a very simple model using differential equations and a handful of time constants can show chaotic or non-convergent behaviour. Of course, cybernetics has known about the problems of instability, convergence and feedback systems for a very long time, Keen is merely encouraging economists to catch up a bit. Paul Omerod also linked economics to feedback, nonlinearity and chaos (probably better than Keen), I think the message is probably sinking in, but no one has any good answers for it so they stick with Keynes for some strange reason (tradition? convenience?)

    Also, I like to distinguish "Champaign Inflationists" who would print money and give it to the already well off, as against "Sixpack Inflationists" who would print money and spread it widely from a helicopter to make Joe Sixpack better off.

    The Bernank (despite his helicopter moniker) is fundamentally a Champaign Inflationist, while Keen is solidly a Sixpack Inflationist.

  3. Not wanting to blow my own nose too hard, but I did model Steve Keen's theory of the firm (with links back to his page) and found that my model follows the classical prediction, and that Steve is indeed just plain wrong on that one (even for the case of finite sized firms). I will point out that there's a presumption of input cost averaging that I think is unrealistic, but apparently everyone uses it. I don't think it is solvable unless you do it that way... or maybe a monopoly always wins by using market leverage against their suppliers, but then competitors end up buying their supply chain, and I ain't even trying to model that.

    Anyone who likes my free modelling code should think about what they might get in return for gold!

  4. "Unfortunately, Keen uses the Karl Marx theory of surplus value"

    Seriously, I don't want to be rude but do libertarians just live in a world where everyone on the left is an evil communist?

    If you knew about Keen's work, you'd know he explicitly *rejects* Marxist economics in chapter 17 of his book, based on a few criticisms, most notably Ian Steedman's Sraffian dismissal of the LTV.

    As for splitting the economy into workers and capitalists, it makes a lot of sense. The two groups obviously do different things with their relative incomes. It's not about 'class war;' it's simply a case of what is does with wages versus profits.

  5. Sorry for double post (why doesn't blogspot allow editing? It's silly).

    Some of your comments on Keen's work:

    "but they are pretty good at working from assumptions to conclusions, that's brute force logic, and there are lots of really smart economists who can solve logic puzzles"

    Actually, often assumptions are swept under the rug and it is not noticed that they are contradictory. Here is a (non-Keensian) example:

    To have smooth and differentiable cost curves, one must assume that inputs and outputs are perfectly divisible, homogeneous and available at a set price. OK, this is stupid, but whatever...

    Later on on producer theory, the conventional U-ish shaped MC curve will be justified based on economies of scale, followed by diseconomies of scale. But wait - how can we have these if inputs and outputs are perfectly divisible and homogeneous? Managerial effects? Labourers are the same! Bulk buying? The firm is a price taker! Indivisibilities? Well...

    Given an increase in perfectly divisible, homogeneous inputs at a set price, there is no reason to expect anything other than a constant increase in output (perhaps even falling given geometric relationships). But still a standard textbook will discuss EoS and DoS.

    There are *many* more examples of such contradictions. I won't bore you with all of them, but Keen identifies some, one of which you seem to have mixed up:

    "He seems fixated on the market power of firms that are small, not infinitesimal He thinks they act like monopolists, showing a math error at the root of Samuelsonian neo-classical economics."

    The first part is correct. A firm cannot be a price taker if the demand curve slopes downward. They can *act* as if they are a price taker, but this doesn't change that they will have some effect on price, however small. The *cumulative* effect of all of these firms producing slightly too much from not taking the price decrease into account makes the demand and MR curves diverge in the same way as a monopoly's does.

  6. Unlearning: if you Google something like "Cournot model number of firms" you will see lots of economists showing that when the number of firms is finite MR will be less than Price, and that Cournot equilibrium gets closer to monopoly as the number of firms decreases, and gets closer to perfect competition as the number of firms increases.

    I don't know how old that result is, but it was already old when I learned it as a student, over 30 years ago.

  7. Eric: "I too believe the economy is endogenously unstable. That is, unlike Friedman and other free-marketers who believe recessions are caused by government,..."

    I wonder if that's a false dichotomy (is that the right word?).

    Suppose that whether the economy is unstable or not depends, in part, on the monetary policy regime. And our job as economists is to find a monetary policy regime where the degree of instability (somehow defined) is minimised. Yes, bubbles and crashes will happen, but if we got monetary policy right, they didn't ought to cause recessions.

  8. Mercury11:52 AM

    I suppose whaling (and other) ships used to engage in Batesian mimicry when they would paint their sides in a black/white check pattern to look like gun ports on a man-o-war.

    But how is it the case that it is "not a stable equilibrium to have no mimics over long periods of time" ?

    My whiz-bang macro solution is to just get the hell out of the way and let markets (real estate, interest rate, college tuition, medical care) clear for once.

  9. Chris Auld recently updated
    his critique of Keen

    Ten years ago Auld participated in a discussion at the Economic History site that had Keen so tied up in knots he ended up making an argument so pro-market it would have embarrassed Milton Friedman. It was funny, unfortunately it has largely disappeared from the archives there (along with a lot of other debates).

  10. Nick: Well, one might consider monetary policy government, and I think Friedman, especially with his constant money rule, thought the economy would be pretty much recession free (excluding wars, etc) with this kind of thing. I don't remember if he really rethought that after 1982.

  11. Eric: yep, and he was wrong on that. Or, at least, the k% rule was probably better than the Gold standard, but still not good enough. And Friedman acknowledged that. So we (most of us) tried inflation targeting. And that probably worked better than the k% rule, and better than the gold standard, but still not good enough. So we need to keep thinking, and do better. Which is why some of us propose a nominal GDP level target for monetary policy.

    Just because there was a housing crash, and some banks go bust, doesn't mean there has to be a recession too.

    (Enjoyed your post on mimics, by the way. I try to follow this ecology blog, when I can understand it: )

  12. I would not say Minsky was predicting the business cycle though. Instead his focus is on the long cycle which is primarily financial and for which leverage is a significant indicator, so while each cycle may end in a bubble of a different sort, a financial cycle ends with a bubble of debt whatever it is backed by, or it is possible for the central bank to recover from smaller bubbles, financial or otherwise, when it has operating room but its operating room becomes constrained over multiple cycles and the financial bubbles become larger until they can no longer be countered. I would say it is not only leverage though. Wars lead to high leverage but not bubbles, and deleveraging does not require depressions, but they are likely if it is the private sector doing it.

  13. Nick: hey, glad you liked it. Alas, I don't know how you would model it, but perhaps that's why macro is so hard, because the really interesting stuff isn't easy to model, and the stuff you can model isn't really interesting.

  14. Unlearning, you would know more about Keen if you troubled yourself with research. In particular, Keen proposes that Marx successfully refuted Say's Law:

    Beyond that, Keen has a long and technical article called "THE MISINTERPRETATION OF
    MARX'S THEORY OF VALUE" where he explains that Wagner and Bohm-Bawerk had not understood what Marx wrote, so it was necessary for Keen to go back to the original Marx in order to construct his theory.

    If I'm going to disagree with a man, I take the trouble to read his opinion first, I suggest you might do the same. I do disagree with Keen on the matter of Say's Law by the way, I take the Austrian perspective on that one.

  15. Here's another Steve Keen paper, following on from the previous:


    Conclusion: Post Keynesian economics is thus not as eclectic as both its major proponents and opponents believe; it has merely lacked a clearly articulated theory of value, and an axiomatic basis derived from that theory of value. Both of these exist in Marx, and can be adopted by Post Keynesians without fear of contamination by the labor theory of value, and without abandoning any of the valued aspects of Keynes's philosophical approach to economics.

    The benefits to Post Keynesianism from adopting Marx's neglected foundations are many. A sense of methodological consistency can be gained; disparate insights concerning the role of uncertainty, expectations, money, and the labor market, can be collated into a consistent perspective on capitalism; and most importantly, as Caldwell (1989, pp. 56-57), Pasinetti (1986, p. 427) and Backhouse (1988, pp. 39-40) emphasize, a consistent alternative paradigm is needed if Post Keynesianism is ever to develop into a full rival to neoclassical economics.

  16. Hmm well perhaps I overstated my case, but:

    (a) His use of Marx for Say's Law is not really relevant.

    (b) Calling it 'marxist' (even by Keen himself) is a bit of an overstatement; it's more of an adoption of the classical theory of value in general.

  17. Eric, perhaps you already do, but you should follow Nick's posts on blog Worthwhile Canadian Initiative, they're pretty much the best macro blog out there (cobloggers are skippable).

  18. yeah, that looks pretty interesting...though many macro debates I find infertile

  19. I would not call Keen a "marxist", nor have I done on this page. That is too much of a loaded word, and means so many things to so many people.

    That said, Keen does base his theory and inspiration on Karl Marx, in particular this concept of a monetary circuit:

    Money ==> Physical Commodity ==> Money plus Surplus

    In shorthand:

    M => C => M+

    The principle being that as you go around the circuit you get M => M+ => M++ => M+++ forever. Thus, the perpetual rentier who serves no purpose but to collect money is held up as the counter-example to Say's Law. Worse, the existence of a perpetual rentier implies the theft of labour from someone else, via the medium of exchange and usury.

    That concept is Marx, and it is also wrong. People only use money because it is a convenient medium of exchange. A free people would have the option to either use money, or not use money depending on whether what benefits them. They also have the option to switch from one denomination of money to something different, so if one particular currency is hoarded and demanding usurious rates of interest, then people simply switch to a different currency.

  20. Unanimous5:48 AM

    I read the first edition of Steve Keen's book when it came out around 2001 or so, and have been checking in on him occasionally ever since. He has been highlighting the possibility of economic disaster pretty much continuously since at least 2001, so on that count he might qualify as a crank.

    Nevertheless, I find many of his criticisms quite insightful and helpful in understanding many aspects of economics. Most of his criticisms are about unrealistic and contradictory assumptions, and not about the reasoning that follows from the assumptions. He knows full well the limitations of his own efforts to explain and model the economy, and I believe would agree with you Eric on how much more difficult it is to put something constructive in place than to criticise.

    His short bursts of internet writing are often not well constructed and denigrate people, and for that reason also he can come across as a crank, but you get a better view of his thinking from his books.

    He sees himself as pursuing a line of enquiry and having pieces of "the answer" that aren't yet enough to predict much.

    Isn't the run-up in debt that Steve focuses on driven by something like the mimicry that you identify Eric? People see others doing well from borrowing to spend or invest in something, and so tend to borrow and spend more themselves. This makes GDP and/or assets increase, and then plunge at some point. Sure, each cycle the focus of spending and investment is different, and is on something that hasn't been the focus for many cycles, but I don't see that as incompatible with Steve's ideas. And, isn't it also possible that there is an underlying "super-cycle" of debt that peaked sometime around 2008? I'm not sure why there couldn't be more than one cycle of mimicry operating with different time periods at the same time.

    You also like to criticise people in quite an uncompromising manner, so really, now that I think about it, I'm having some difficulty seeing the difference between you and Steve.

    The debt/cycle/mimicry thing really is quite an old idea, and has been shared by many people over the years. Steve is having another shot at modelling it, but with the addition of an MMT understanding of money. Maybe he will get nowhere. I also think that's the most likely outcome because human fashions are too unstable, but I intend to watch and see where he does get to.

  21. Well, I agree that we are headed for a Greek style end-game in the US. Official debt is $17T, but off the books it is more like $60T, and that latter stuff (social security, medicare) is more likely to be paid back than bondholders.

    But, I'm not a fan of multiple equation macro models, because they are unfalsifiable, and only proven lame via lamenting their real-time ineptness looking back. I won't waste time on them, I know too many people who did, and greatly regretted it (I was an economist for an econ department right out of college).

    Keen's model is one of those pre-Lucas models where expectations are basically ignored. Look how few of his equations use expectations, but rather, people act like cogs in a deterministic fashion. Thus, when he concludes the solution is to print money and give it to everyone, ignoring the effect this would have on future interest rates, he's at least being consistent there. But it's a bad model because of this.

  22. Unanimous5:49 PM

    Treating people like cogs is what distinguishes economics from sociology isn't it? I'd say a cog with rational expectations is a more deterministic cog than one with out assumed expectations. Of course not being explicit about expectations might amount to implicitly making silly assumptions about them depending on the model.

    But anyway, surely the only real way to tell if any theory works is by it's predictions. It's very common for any theory to involve assumptions that even the proponents recognise aren't strictly accurate. But you've got to start somewhere, and then you see how things work out. So far, Steve's predictions are on the lame side, and he would probably admit that, but I have trouble thinking of any economic predictions that aren't on the lame side.

    As for printing money, I think Steve places more emphasis on debt cancellation than printing money. Debts have always been cancelled under various conditions throughout all of human history, it's just a question of how strict those conditions are. Steve is definitely towards the less strict end of the spectrum, and I don't know that I agree with his degree of slackness on this. But anyway it's a continuum, not a binary thing.

    With regard to printing money, the effect will surely depend on how much is printed, under what conditions it is printed, and probably also on how it is presented or marketed even. Most people don't realise how money is created, and when it is explained are pretty incredulous about it. Once it is understood, there is a case for regarding debt cancellation as a form of money printing. Steve recognises this, and so doesn't see some degree of money printing under certain conditions as a bad thing, and it seems to me that he sees it as a pragmatic version of his real aim which is debt cancellation. Again, it all comes down to specifics and the degree. If you regard debt cancellation as a form of money printing, then everyone accepts a degree of money printing.

    I suspect Steve would also see a moderate amount of printing money as akin to a negative interest rate, and would expect future interest rates to be lower rather than higher because of the amount of money printing that he would engage in. I doubt that he is ignoring the effect, but rather is hoping for an effect, and one that is different from what you (rationally?) expect.