Wednesday, February 8, 2017

Jason Smith — Qualitative economics done right, part 2

Jason Smith looks at Steve Keen's model and finds it lacking.

Information Transfer Economic
Qualitative economics done right, part 2
Jason Smith


AXEC / E.K-H said...

Like Walrasian, Keynesian, Marxian, Austrian models, Steve Keen’s model is provable false. For details see

Where advanced Heterodoxy — represented by Steve Keen — took the wrong turn

Keenonomics, aggregate demand/change of debt, and some misleading critique

Putting the production function back on its feet

The key relationship between employment and growing/shrinking debt

Debunking Squared

Heterodox economists are scientifically just as incompetent as orthodox economists.

Egmont Kakarot-Handtke

Dan Lynch said...

Blogger ate my comment, and I am not inclined to retype it.

The shorter version is that the author does not understand Keen's model nor modeling in general.

The author never addresses Keen's basic concept -- that private debt is the main driver of capitalist economies, as illustrated by this Keen chart.

Keen's model is based on Minsky view of macro. To reject Keen is to reject Minsky, a strange position for MMT to take.

Tom Hickey said...

It's not in spam, Dan. Something malfunctioned in Blogger apparently.

Dan Lynch said...

@Tom, the problem seems to happen when I first "preview" a comment, then attempt to "publish" it -- at that point Blogger gives an error. Sometimes my comment is still there if I go back to the page, but today it was not there!

It seems to work as long as I publish without bothering to preview.

Prolly some incompatibility with my browser or OS. It's happened to me several times recently on MNE.

Ignacio said...

Dan the author is not an MMT guy, he is attacking SFC models too (generally).

His point is that models with insights w/o connection to data do not advance theory. He is right scientifically speaking.

Tom Hickey said...

@ Dan

That happens with me too. I gave up on previewing which unfortunately shows in the posting.

The safest thing to do is to copy the comment to the clipboard before posting and if there is an error, just close the page and open it again, choose "post comment \." and then paste in the comment. That always works for me.

Tom Hickey said...

The problem is that there are an infinite # of explanations that can be given for anything. Most will likely be absurd or highly implausible, and so can be rejected out of hand. But there may be several plausible explanations. Science is about distinguishing the best explanation available from contenders.

Conceptual explanations are qualitative and mathematical ones are quantitative. In the case of both, there has to be a clear connection between the explanation and the data that is being explained that supports the explanation.

I read Jason Smith as saying that Keen has a qualitative explanation that may have value but he has not established the connection between the analysis and the data sufficiently.

This is an issue with heterodox economics as such as conventional. Pointing out problems with conventional economics calls for an alternative solution that meets the basic criteria for scientific explanation that distinguishes it from storytelling.

Ignacio said...

Adding to Tom post, is important to read previous post from the author in this topic to see where he comes from re. description of quantitative vs. qualitative models and strengths and weakness of each model.

AXEC / E.K-H said...


Jason Smith

You say: “Therefore we don’t really know what the key to macro is.”

YOU don’t know, WE know. This is the elementary core of foundational macro propositions, a.k.a. axioms: (A0) The objectively given and most elementary systemic configuration of the economy consists of the household and the business sector which in turn consists initially of one giant fully integrated firm. (A1) Yw=WL wage income Yw is equal to wage rate W times working hours. L, (A2) O=RL output O is equal to productivity R times working hours L, (A3) C=PX consumption expenditure C is equal to price P times quantity bought/sold X.

These macro axioms are certain, true, and primary, and therefore satisfy all methodological requirements. The set of premises is minimal, that is, it cannot be reduced further, only expanded. The set is behavior-free, contains no nonentities like constrained maximization or equilibrium and no normative assertions. All variables are measurable.

For more details see: How to restart economics

Egmont Kakarot-Handtke

Dan Lynch said...

There seems to be confusion about the purpose of Keen's modeling, and Keen himself hasn't said much about his purpose, other than he obviously enjoys doing it for the sake of doing it.

Keen is professor. Professors publish, and professors teach. Besides the cynical observation that the model gives Keen something to publish, I can see how it would be a useful teaching tool, to illustrate basic concepts. Engineering professors and engineerings students do that sort of mathematical modeling endlessly, so I see nothing unusual about it as far as the academic world goes.

It's fair game to ask if Keen's model is useful outside of the classroom, and it's fair game to point out any flaws or failures in the model.

As an engineering student, I spent something like 90% of my time doing mathematical models. In the real world, engineers spend 90% of their time collecting and analyzing empirical data (when they are not attending meetings or writing reports), rather than relying on the theoretical models they were taught in school. But that does not mean it was a waste of time to learn the theoretical models. The models illustrate the basic concepts, and that's important.

Example: my specialty was controls & manufacturing. As a student, I was forced to spend many hours doing fancy mathematical models of control systems. As it happened, on one of my first real world jobs I was asked to see if I could make the control for one of their manufacturing processes work better. It so happened to use a pneumatic PID control. The system had been in use for years and the company had never been able to get the process control to work satisfactorily. There were no mathematical numbers on the pneumatic controller, only some screws that could be adjusted in or out, so even if I had been able to develop an accurate mathematical model of the physical system, it would not have helped me adjust the controller screws! But .... all those hours of modeling PID control systems had left me with an intuitive understanding of how PID control works, so I was able to tweak the adjustment screws by trial and error, and get the system running quite nicely.

At Ignacio's suggestion, I went back and read the author's Part I. Meh, I was not impressed. That is not how things are done in engineering, so I guess engineers have been doing it wrong all these years? Engineering models can be as complicated or as simple as you like, as general or as specialized as you like. It all depends on what you are trying to accomplish.

The author seems to emphasize the ability of a model to predict the future. That is one possible use for a model, but it is not the only use. Predicting the economic future is tough because it depends on politics, external events, and timing. Lots of economists get predictions wrong, including MMTers. If I knew how to predict the economic future then I could make millions playing the markets and then retire.

Tom Hickey said...

Dan, what you are arguing for is that economic be treated like engineering (applied science) rather than as "science" (theoretical science). I agree, but there are still going to be a lot of economists that want to be considered "real scientists" that build theoretical models that the experimental folks test down the line.

So be it. Let there be separate departments of theoretical economic for the math types and practical or applied economics for the folks that want to apply economic analysis to real world systems and policy issues to produce useful results.

An overarching problem in the later field of applied economics is integrating finance and economics.

Ignacio said...

Dan, as Tom says, you are coming from a different angle (engineering) than from the author angle (a theoretical physicist and scientist).

BTW I'm way closer to your pov and engineering in general, probably because I'm more of a systems oriented person who works at design and practical problem solving. And I find Keen models useful too even if they do not necessarily meet the expectations from a strictly scientific pov (regarding data). This is also closer to how models are done and used in practical finance and trading and, in general, systems modelling.

But I don't think this is where Jason wants to get to, as he wants as a closest approximation to data as possible which describes real dynamics the best it can (and hence has certain prediction power). From a scientific point of view of rejecting a null hypothesis I can see some value on what he is saying.

Still all the focus on financial economics and modern capitalism hurts my head a bit, as that's not what I consider economics to be about at core. We are describing a functioning of human made and evolved 'artificial' systems, instead of talking about core material production, consumption and distribution and underlying connection to physical dynamics.

I would expect a physicist would focus more on the REAL part of economics instead of on the finance, in that regard I wonder how applicable is the information transfer framework he is developing applicable to those dynamics (and is what i'm really interested about).

AXEC / E.K-H said...

Tom Hickey

You say: “The problem is that there are an infinite # of explanations that can be given for anything. Most will likely be absurd or highly implausible, and so can be rejected out of hand. But there may be several plausible explanations. Science is about distinguishing the best explanation available from contenders.”

Confronted with the vastness and complexity of reality every branch of the sciences is confronted with the problem of where to start. John Stuart Mill put it thus: “What are the propositions which may reasonably be received without proof? That there must be some such propositions all are agreed, since there cannot be an infinite series of proof, a chain suspended from nothing. But to determine what these propositions are, is the opus magnum of the more recondite mental philosophy.”

Krugman, for example, is quite explicit about how he has solved the starting problem: “most of what I and many others do is sorta-kinda neoclassical because it takes the maximization-and-equilibrium world as a starting point.”

The neoclassical world is given with these hard core propositions, a.k.a axioms: “HC1 economic agents have preferences over outcomes; HC2 agents individually optimize subject to constraints; HC3 agent choice is manifest in interrelated markets; HC4 agents have full relevant knowledge; HC5 observable outcomes are coordinated, and must be discussed with reference to equilibrium states. (Weintraub)

These microfoundations are the wrong starting point as everybody knows by now. So they have to be replaced by the correct macrofoundations. This is the actual challenge.

For details see: ‘Economists’ three-layered scientific incompetence’

and ‘Macro poultry entrails reading’

Egmont Kakarot-Handtke

Matthew Franko said...

"The author seems to emphasize the ability of a model to predict the future."

that can only be done using derivative action of a functional equation...

even using integral action is not predictive... it uses a summation of previous time intervals and compares it to a target summation...

So you have to start to put together a functional equation in time domain and then look at instantaneous change...

But as Dan references above in a PID controller you can regulate without using Derivative action at all you can just use Integral action or Proportional action or a combination of those 2... iow we dont have to necessarily predict the future to regulate anything successfully....

AXEC / E.K-H said...

Matthew Franko

You say: ‘The author seems to emphasize the ability of a model to predict the future.’

Predicting the future is the business of charlatans/prophets/agenda pushers/morons. See

Science does NOT predict the future

ICYMI Prediction/Forecasting

Prediction does not work? Try retrodiction first

Egmont Kakarot-Handtke

Matt Franko said...


People make successful predictions every day...

AXEC / E.K-H said...

Matthew Franko

You say: “People make successful predictions every day...”

Scientists know better: “The future is unpredictable.” (Feynman)

See also ‘Predictably confused’

Egmont Kakarot-Handtke

Matt Franko said...

Ok AX you can go on trying to solve Pi...

Meanwhile Goodyear will make $1.5B this year selling round tires that transport everything...

AXEC / E.K-H said...

For the parallel discussion see on the Information Transfer Economics blog ‘Qualitative economics done right, part 2’

or ‘The key to macro and Keen’s debt-employment model’ on the AXEC blog

AXEC / E.K-H said...


If an economic equation ever had real content then the structural employment equation. And if you were a scientist instead of a cargo cult scientist you would hurry to test it instead of spreading methodological crap. Your blathering about Einstein and Poincaré symmetry is a blatant distraction.#2

Poincaré, by the way, debunked economic cargo cult science long ago: “Walras approached Poincaré for his approval. ... But Poincaré was devoutly committed to applied mathematics and did not fail to notice that utility is a nonmeasurable magnitude. ... He also wondered about the premises of Walras’s mathematics: It might be reasonable, as a first approximation, to regard men as completely self-interested, but the assumption of perfect foreknowledge ‘perhaps requires a certain reserve’.” (Porter, 1994)

Walras did not get the point and neither did you. So, the real take-away for economists from Einstein is the famous dictum: “Only two things are infinite, the universe and economists’ stupidity, and I’m not sure about the former.”#4

If you want to refute the structural employment equation, test it and do not tell that it is not testable. It is a simple question to decide empirically: is the macroeconomic relationship between wage rate and employment negative as standard economics claims or positive as the structural employment curve claims? This goes in one with determining the relationship between employment and debt.

Stop dropping methodological crap on people.

Egmont Kakarot-Handtke

#1 See the elementary version for the investment economy (ex government, ex foreign trade)

#2 For details go to my blog and enter Einstein in the search field

#3 See ‘The one stone that kills orthodox and heterodox employment theory’

#4 This is what Einstein said about the LOGICAL PRIORITY of theory over testing:
“Whether you can observe a thing or not depends on the theory which you use. It is theory which decides what can be observed.”
“... the axioms Science is the attempt to make the chaotic diversity of our sense-experience correspond to a logically uniform system of thought”
“If then it is the case that the axiomatic basis of theoretical physics cannot be an inference from experience, but must be free invention, have we any right to hope that we shall find the correct way?”
“... any attempt logically to derive the basic concepts and laws of mechanics from the ultimate data of experience is doomed to failure.”

AXEC / E.K-H said...

Part 1 somehow got lost.


AXEC / E.K-H said...

For the full text see
post of Feb 14.