Wednesday, October 5, 2016

Jason Smith — Keen, chaos, and equilibrium


Physicist Jason Smith critiques a debate among Steve Keen, Roger Farmer, Noah Smith, and David Andolfatto over Steve' recent Forbes post asserting that the economy is best modeled as a complex non-linear system instead of using the conventional linear stochastic models (DSGE) based on assuming general equilibrium.
Actually, as a physicist, I would say that even if the economy was a complex nonlinear chaotic system, linear stochastic models would still be its effective theory description. Regardless of what the quantum theory of gravity is, general relativity -- and even Newton's universal law of gravitation -- is still its long-distance effective theory.

Anyway, this prompted me to write something about Steve Keen's article in Forbes. Keen suffers from a problem that all public economists seem to suffer: asserting matters of opinion as matters of fact, and ongoing research programs as well-established frameworks. This will be made clear as we progress. Let's begin, shall we?
Information Transfer Economics
Keen, chaos, and equilibrium
Jason Smith

17 comments:

GLH said...

"two hundred years of economic wisdom" Is this guy pulling my leg? Economics is soooo scientific.

Ryan Harris said...

Finding 'the one', the elusive beast, that one right way to model 'The Economy', made this article seem appropriate:

What a Georgetown Professor Got Wrong When He Argued That Maybe Dumb People Shouldn’t Be Allowed to Vote.

Now I know that not all economists are dumb, but think of how much better democracy and the economy would be without their policies over the last 50, 100 years. It's tempting, if impossible, to marginalize them.

Matt Franko said...

Yeah but Ryan from there:

"Democracies tend to pass laws and policies that appeal to the median voter, yet the median voter would fail Econ, History, Sociology, and Poli Sci 101."

This is in the median voter's favor if you think about it... and then:

"Brennan, who is also the author of Libertarianism: What Everyone Needs to Know, ..."

Yikes! (promoting the 'mark of the beast' here... and he's at a Jesuit school no less....) He probably thinks the big mistake that the "stupid" people are making is electing officials who keep running up "the debt!" and wont go back to the gold or silver standards...

The main problem is most people have been trained to look at the world as stochastic rather than deterministic this comes via all the Darwin that is pounded into peoples heads all the time...

First thing to ask somebody is whether we are "out of money!"... if they say yes then bingo THAT is a stupid person you are dealing with.... maybe all of those people shouldnt be allowed to vote...

Matt Franko said...

Yes here is from his book:

"As people (regardless of their race, income, gender, or other demographic factors) become more informed, they favor overall less government intervention and control of the economy. (That’s not [to] say they become libertarians.) They are more in favor of free trade and less in favor of protectionism. They are more pro-choice. They favor using tax increases to offset the deficit and debt."

LOL!

"For it is written, I shall be destroying the wisdom of the wise, and the understanding of the intelligent shall I be repudiating.... Does not God make stupid the wisdom of this world?" 1 Cor

THEY are the truly stupid ones.....

MRW said...

Matt, whose book are you quoting from?

Ryan Harris said...

Yep.

Keen was constructive in talking about the how and why.

I like the discussion about better ways to model the economy. But it descended rapidly into the one, the right way and went off the tracks.

If they think they have one right way, then they are snipe hunting, creating utopian societies and occupying the impenetrable world of the imagination.

To these folks dumb means the people with whom they disagree on measures, values and ideology.

Which then becomes a game of purging the dumb who disagree.

Palm to forehead.

Purge the economists!

Drives me batty.

MRW said...

Ah, Jason Brennan's book. There is an incredible review of Brennan's book on Amazon, called Well-Reasoned And Clever—But Fatally Shallow.

NeilW said...

Jason's problem with his physics analogies is that he forgets that atoms don't have learning facilities and rarely watch the six o'clock news.

Macro models without individual actors fail the Lucas critique. Searching for the 'deep parameters' is like looking for the end of the rainbow. It doesn't exist.

Policies affect individual entities and their reaction via the learning process changes the aggregation outcome. So you really don't know how things will change *other than to try it out in a simulated analogue* where you can get a feel for how thing will change *if you include actual humans in that analogue*.



Ryan Harris said...
This comment has been removed by the author.
MRW said...

It's too bad that Brennan didn't bother to watch Paul Davidson here for an explanation why the physics model doesn't work. Maybe because the title doesn't give a clue. Also, it's an hour long.
The Keynes Solution: The Path to Global Economic Prosperity Via a Serious Monetary Theory
https://www.youtube.com/watch?v=31wjPE-mUb4

At a minimum this, 7 minutes:
Paul Davidson - The Trouble With the Ergodic Axiom 2/4
https://www.youtube.com/watch?v=YAbnuwsid4Q

Tom Hickey said...

We have to keep in view what we are after epistemology (how we know) and ontologically (what there is). To do this we use models. Models show how things stand based on general description. Here generality is assume to correspond with regularity of behavior and observation of it. This is the basis of scientific method.

The more comprehensive (general) and the greater the formalization (math). the more preferred the model as the best explanation. But the ultimate criterion in science is fitting observation.

The quest is for the "best explanation." Normal science is occupied with articulating a model and applied science about employing the model technologically (pragmatically) Theoretical science is about pushing out the envelope.

The scientific method involves competition in developing the best explanation. As Paul Feyerabend argues, ruling anything out undermines competition. Best explanation in science at any point in time is a matter of evidence-based debate. Science is always tentative wrt to discovery, invention and fresh data or a new way of configuring existing data. Science is never static.

Where there is not yet evidence, it's not yet science (evidence-based) but philosophy (assumption-based). Philosophy precedes science, which fills in the gap between assumptions and evidence.

Model creation is philosophical (epistemological). Models are representational ("scientific") in so far as they are ontological (fit experience and behavior).

There are known knowns — models that fit the data and are accepted as settled.

There are unknown knowns — how what is already known can be configured to give new knowledge, e.g., not only new models but different ways of modeling. (This, I think, is what Keen is talking about with complexity.)

There are known unknowns — these are matters for research.

There are unknown unknowns — these are reveal by "surprise" aka shocks. These are matters whee there is no applicable method currently. This gap is sometimes closed by fortuitous discovery.

Tom Hickey said...

Models are elaborate metaphors. Even a simple descriptive sense is a model of a putative fact. A descriptive proposition is a sort of map that enables checking the assertion against observation for truth or falsity. There are not facts in the world. Models carve the world up into "facts," a "fact" being an observation using the model (proposition) as a lens or map.

A descriptive proposition is elementary if one and only if one an only one fact confirms it or the absence of the fact disconfirms it. The propositional calculus is based on this as the fundamental building block. As an ideal science provides a general description of the world.

But many models are based on assumptions of generality that are "stylized facts." They are restrictive for convenience in making a model tractable.

In ordinary language we say, "The exception proves the rule." There just should not be more exceptions that generality. Knowledge of any level of generality is better than either randomness or magical thinking.

This, I think, is what Jason Smith is saying about, e.g., applying models based on physics to econ and social science. We need to find the regularly anyway we can and in as many ways as we can.

The ideal is to discover a general framework in which different theories can be developed. Without such a framework, the subject is not yet properly as science. Physics has developed a framework but econ psych and soc have not yet done so. "Let a hundred flowers bloom" until one emerges.

Jason thinks that information transfer may provide such a framework. Steve thinks it is complexity. Paul thinks rational optimization and general equilibrium. We'll see based on outcomes. Everyone agrees that some kind of systems approach is needed. Now the arguments are over the kinds.

We think using metaphors and metaphors are built on each other. There is no problem using the metaphor of models from physics in economics or social science to the degree that they reveal behavioral regularity (generality, invariance), which is what science is ultimately looking for.

All models are limited by their assumptions and construction. The problem is not with the model itself or even the use to which it is put but whether the conclusions dream from it conform to the model or exceed its boundaries.

Similarly with the biological model of evolution. It might be useful but it can also be used beyond its limited. Societies are not organism for example, even though they may act like organisms in certain respects, owing to the their elements being organisms.

It's a matter of exploiting strengths and reducing the effect of weaknesses.

The temptation to exceed boundaries or overemphasize a particular model or way of modeling presumably results from either ignorance, which is uglily among well-trained professionals, or cognitive bias (including affective and volitional). The obvious suspicion is normative and prescriptive creeping into description or being intentionally imported.

I don't have any problem with amy kind of modeling as such. It not whether complex non-linear models are better than linear stochastic ones, for example. Or even gadgets in teaching.

continued

Tom Hickey said...

continuation

Econ 101 has to be simplified if students are going to get the gist of it. That's true of most other subjects as well. 101 whatever just provides basic concepts and tools.There is nothing wrong with starting Econ 101 with a barter system. Student are sharp enough to realize that we are operating under a monetary system, but they need to know the basic idea of production, distribution (exchange) and consumption as the economic cycle involving real resources. This is required in order to realize that the fundamental issue is availability of real resources and to see how demand draws forth supply

It's the us to which a model is put and the conclusions drawn from it, as well as the result of applying it.

Most complicated situations cannot be reduced a a propositional calculus that allows for confirmation and disconfirmation in terms of elementary propositions that are true or false as indicated by observation. Certainly not macro.

The challenge is to squeeze out as much regularity as possible using whatever approaches are relevant. The search for a single "best explanation" can be no more than an ideal in this. The subject matter is too vast and the current state of knowledge too undeveloped to be able to discover a Newtonian-like system and the desire to do so should not be permitted to warp the developmental process.

MRW said...

Tom, all of this economic modeling talk doesn't take into account how federal accounting actually works in a given economy. In real life.

Tom Hickey said...

om, all of this economic modeling talk doesn't take into account how federal accounting actually works in a given economy. In real life.

That's not the only way to model economic behavior and one type of economy. I don't think that MMT economists are saying it is the only way or that noting else matters. The foundation of econ, MMT economists recognize, is the availability and use of real resources — production, distribution and consumption. In a primitive economy there aren't even any commodities produced for exchange, only for use.

"Money" is just bookkeeping. In fact, money is indeed a veil over barter as the exchange of commodities, that is real goods produced for exchange. What Keynes showed is that money is not a neutral veil but affects circular flow.
And as Marx had already observed, consumption goods are not produced for money chiefly to produce more consumption goods but intrinsic to the process is the desire of the owners of the means of production to accumulate financial wealth.
A monetary economy in which distribution is allocated based on price rationing is one way to do this. There are different monetary systems both historically and potentially. In addition, consumers don't spend all they earn since they also desire to save, at a minimum for security against uncertainty.

MRW said...

That's not the only way to model economic behavior and one type of economy.

I wasn't saying that, neither was I criticizing you. It was more of an aside.

Of course I recognize this: availability and use of real resources — production, distribution and consumption, which is why it pisses me off--well, dumbfounds me--that these economic models don't include how federal banking/accounting/transactions work so that they can address the real resources and constraints that a sovereign federal government is charged with managing for the good of all. I mean, how are these people content to live with these theoretical model bubbles? Because they created them? Doesn't Krugman still sneer at Wray's insistence that he include banking?

Tom Hickey said...

In my view there is a difference between macroeconomics as a theoretical attempt to explain macro economic phenomena and political economy as an aspect of policy science. The problem most arise out of conflating them.

Macro economics is about developing explanatory models. It's not really about developing predictive models because it is extremely difficult to test them for a variety of reasons having to do with both model construction and data.

Policy science is about developing data-based information regarding policy formulation that complements political theory, which is ideological and normative. Different parties and interest groups put forward different policy proposals and attempt to ground them in ideological preference, reasoning, and fact.

Macroeconomics should aim at being scientific. Policy cannot be scientific as long as there is more than one party, that is, in a liberal environment. Even in a one party system there often disagreements over how to proceed based on different assumptions and different data sets.

The reality si that there is no general theory of macroeconomics that passes the smell test. Nor is there in psychology or social science. Moreover, biology is significantly different from physics and chemistry even though it is informed by physics and chemistry.

I am not all that much interested in macroeconomics other than its usefulness in policy formulation. Then the context becomes paramount and in a monetary production economy the operation of the monetary and financial systems become highly significant. But even this is not enough. As Bill Black's analysis shows, understanding the monetary and financial systems is insufficient. More gradual analysis of actual practices including malfeasance must be considered. This is not way to model this formally at this point.

However, it was obvious to experts that the institutional arrangements were highly likely to produce the results they did. Shriller and Akerlof received a "Nobel" for work in this area but it was ignored not only by economists but also regulators, who refused to see it even when the FBI called attention to it.

Modeling only goes so far, but that is not reason not to do it. Just "handle with care."