Sunday, July 08, 2007

Some Principles Of Heterodox Economics

  1. Methodology (rather than just method) is important to understanding economics.
  2. Human actors are social and less than perfectly rational, driven by habits, routines, culture and tradition
  3. Economic systems are complex, evolving and unpredictable - and consequently equilibrium models should be viewed sceptically.
  4. While theories of the individual are useful, so are theories of aggregate or collective outcomes. Further, neither the individual nor the aggregate can be understood in isolation from the other.
  5. History and time are important (reflecting (3)).
  6. All economic theories are fallible and, reflecting (4), there is contemporary relevance of the history of thought to understanding economics.
  7. Pluralism, i.e. multiple perspectives, is advocated (following on from (3) and (6)).
  8. Formal mathematical and statistical methods should be removed from their perceived position as the supreme method - but not abandoned - and supplemented by other methods and data types.
  9. Facts and values are inseperable.
  10. Power is an important determinant of economic outcomes.
-- Andrew Mearman (2007). "Teaching Heterodox Economics Concepts", The Economics Network (June)
(I previously posted a quotation of ten principles of institutional economics.)

33 comments:

James said...

I wonder who would admit to disagreeing with any of this.

walt said...

I disagree with 6 and maybe part of 8.

YouNotSneaky! said...

I'm with Walt on 8 and also

3 can be easily overstated

for 7, just because something is a "good" it doesn't mean that there is no budget constraint. The question is "how much pluralism?". There are clearly perspectives NOT worth pursuing.

I disagree with 9, mostly. Just because it's not easy doesn't mean it can't be done, up to a point. And if facts and values really are inseparable, what are we arguing about? Let's just agree that we have different values and that my values are better.

walt said...

I think equilibrium is in the eye of the beholder, so it's really a matter of taste.

9 is too vague for me to be sure if I agree or disagree.

Anonymous said...

I'm still waiting to find out why math models based on unrealistic (indeed false) assumptions give more truth about the world then does, say, poetry.

All defenders of math models (as the highest form of thinking) can do is mutter, "science, physics, science, physics,.." which ain't an argument at all.

The fact/value dichotomy has been discussed for a long time. One thing is clear: the 1920s version of "values" accepted by the orthodox is hard to maintain. In this perspective "values" are merely subjective statements about likes/feelings/etc and absolutely nothing permits us to choose between them.

This perspective leads to statements such as "well, it was right for Hitler."

Assuming facts exist and assuming non-factual statements are made, no reason exists to suppose that all non-factual statements are equally wrong or right. We can engage in reasoned argument to narrow down what subjective statements are "acceptable" and those that we reject. A selection process among non-factual statements can be found that does not involve invoking facts.

But this is all old stuff to philosophers. Most economists, however, think that 1920s philosophers were the end of the line of thinking in the west.

Plus as I endlessly say, orthodox economics is infused with values from top to bottom: it is good to let people get what they want! That is a value statement. Just because "we" all agree with it does not mean it is not loaded with value. That is, we might all agree that "murder is wrong" but this does not mean that such a statement is not a value-statement.

Eric

YouNotSneaky! said...

I'm still waiting to find out why math models based on unrealistic (indeed false) assumptions give more truth about the world then does, say, poetry.

Because all models, mathematical or otherwise, are going to have some unrealistic assumptions. Show me a model with "realistic assumptions" only. Those assumptions are there for a reason. Modeling means simplifying things, i.e. making unrealistic assumptions.
As I've said before, it's not about realistic assumptions vs. unrealistic assumptions, it's about which set of unrealistic assumptions you're okay with.
So who's muttering what friend?

Plus as I endlessly say, orthodox economics is infused with values from top to bottom: it is good to let people get what they want! That is a value statement.

Well, that's certainly assumed in normative economics. But that's why it's called "normative". Otherwise it's factual claims - price goes up, quantity demanded goes down, etc. You can still work "values" into those kind of statements I guess (what is "price", what is "quantity demanded" etc.) and there's SOME value in thinking about it that way, but mostly it's a empirically verifiable objective statement.

I have no idea what Hitler has to do with any of this.

YouNotSneaky! said...

And I agree with Walt on the equilibrium thing. It has some specific meanings (a price vector such that blah blah blah...) in specific context but more generally it just means that there's a particular outcome(s) that you expect from a particular set of conditions. It's interesting to ask how a social system moves from one such outcome (equilibrium) to another when parameters change but I wouldn't get too worked up about the word "equilibrium" itself.

YouNotSneaky! said...

More to the point, if economic systems really are "unpredictable", why bother? And if they are "unpredictable" then any model/school that actually says something is just as bad as any other.

Should we cut taxes? Who knows, it's "unpredictable"!

Should we implement labor standards in international trade agreements?
Who knows, it's "unpredictable"!

Are the stagnant wages and rising profits a sign that labor's bargaining power is being eroded? Who knows, it's "unpredictable"!

Is the wage gap between male and female earnings due to discrimination? Who knows, it maybe due to some butterfly that flapped its patriarchal wings over in China!

Of course some things really are unpredictable. But to consign all economic phenomenon - the "economic system" - to that dustbin is plain ol' wrongheadedness. And it's usually done because someone somewhere doesn't like some conclusion for ideological reasons so they invoke "complexity" (usually misunderstanding the term) for a lack of better argument. And part of the job of economic theory is to separate that which is unpredictable from that about which we can actually say something.

Anonymous said...

>Because all models, mathematical or otherwise, are going to have some unrealistic assumptions.

True but not relevant.

This is like an answer: "saying, yes no car today can get 1,000 miles per gallon so we must be satisfied with one that gets 50 miles per gallon." The question? " Can you drive a car to mars?"

The answer is true, but not relevant. Yes math models often make simplifying assumptions. But this does not justify any claim that math models tell us something about the real world.

You might be okay with some set of assumptions but what is the relationship between a math model and reality? A math model tells you the implications of certain assumptions but this does not carry over to it saying anything about reality.

I still here muttering but not answer. I've yet to find an orthodox economist who can justify the faith they have in math models.

>Well, that's certainly assumed in normative economics. But that's why it's called "normative".

Many avenues to take here. I'll just note that the concept of pareto optimality is 100% ethical. Behind PO is the notion that it is good to let people have what they chose. That is an ethical statement.

A concern with prices, in particular with prices being "right" only makes sense if something like PO is assumed. "Right" prices are "efficient prices" and efficient prices are what you'd get if people were able to make their own choices and had the most they could. That is ethical content. So even a concern with prices is directly an ethical concern.

In orthodox economics is also an explicit notion that all that matters when you look at the economy is the consequences. And the only consequences that matter are those related to choice/consumption/etc.

The claim that only consequences matter is modern utilitarianism. This perspective is challenged by many. The book by Smart and Williams (Utilitarianism: For and Against) covers important material related to this.

But I'm sure you've read it. ;>

Eric

h.e. said...

I've yet to find an orthodox economist who can justify the faith they have in math models.

I'm not sure if "faith" is the right word. In any case, it requires some qualification of what exactly it *is* the mathematical models *do* that orthodox economists have faith in. I tend to think of the role of formalism from a different perspective: economists require that ideas be communicated in a certain way. Mathematical formalism is a disciplinary norm. The line that I generally hear is that formal (mathematical) modeling is important because it makes the logic underlying an argument transparent. That the correctness of this logic applies to a purely imaginary world doesn't seem to matter. Economists are also generally nonplussed when one points out that certain operational relationships with their models, upon which results turn, are ad hoc or fraught with conceptual difficulties. I include Robert's favorite hobby horse, capital theory, into the latter category.

YouNotSneaky! said...

Eric,

First, I said "mathematical or OTHERWISE". Whenever somebody makes some statement about a cause and effect phenomenon they have, whether they know it or not, a model in their mind. Just because you don't do math doesn't mean you're not making assumptions. Worse, you're making assumptions and you're probably not even aware that you're making them or what they are. Again, I challenge you, or anyone else to show me a model/theory/statement about social phenomenon with only realistic assumptions, whatever these may be.

Second, mathematical models are not just the province of orthodox economists. Robert posts mathematical models all the time and obviously he's not orthodox.

Third please make your analogy with the car more precise and explicit because I, in my limited intellectual capacity, fail to see its applicability to the issue at hand. All else being equal, of course we want more realistic assumptions (again, whatever these are and if we can agree on what constitutes "realism" in this instance). But that ain't how it works. If it was that'd be a free intellectual lunch sitting right there on the table and we all know those don't come around very often. Making assumptions more realistic (i.e. relaxing them) is a difficult endeavor (and a pointless one oftentimes)

Fourth,

You might be okay with some set of assumptions but what is the relationship between a math model and reality? A math model tells you the implications of certain assumptions but this does not carry over to it saying anything about reality.

Why not? Because the model is not reality itself? I hate to belabor an overused example, but a road map is based on some unrealistic assumptions (the earth is flat like a piece of paper, my house is usually not marked upon it, and highways are actually yellow in color), but nonetheless it tells you A LOT of useful information about the real world.
Same with economic models (and let me emphasize - you're not objecting to any PARTICULAR model but to math modeling as a whole).

Is that enough justification for you? I think I'm pretty clear on my position here and why I think mathematical models are useful. On the other hand your qualms seem to be pretty imprecise and vague - and saying "but it makes unrealistic assumptions" is exactly that - so again, who's the one muttering friend?

I don't feel like getting into Utilitarianism and all that right now. But do note that PO is not Utilitarian, and Utilitarianism is explicitly labeled as, again, NORMATIVE, economics.

YouNotSneaky! said...

Or to put the end to all the muttering, please outline your most favored methodology or approach to understanding social phenomenon. If I wanna know what is the most likely outcome of a cut in tariffs on the US current account deficit how I should go about thinking of it? Feel free to recite poetry at me.

Anonymous said...

>If I wanna know what is the most likely outcome of a cut in tariffs on the US current account deficit how I should go about thinking of it?

If you wanted to help people flourish through their interaction in the economy how would YOU do it?

My questions are different from your questions. What is the effect of tariffs on human flourishing? Can your model tell me?

Orthodox economics talks about certain things not because those things MUST be part of an economic theory but because these are the thing it thinks are important. But it might be wrong. And often is wrong.

Formal models used by the orthodox merely discover the implications of the set of assumptions they have made. But whether this model says anything about the real world is not clear. Let's say that you could generate 1,000 different models of "the economy." Do they ALL tell us something about the economy or do they just merely spell out the implications of their own assumptions?

It is merely a custom within orthodox economics to do math modeling. Why is it done? Unclear except that it is done.

FYI, I'm not opposed to math modeling and can model with the best. But I never thought that just because I wrote stuff on a piece of paper that I was discovering stuff about the real world.

Orthodox economists never stop to think about WHY they do models except "that is what economics is." So they have been taught. And never question what they have been taught because that's the sort of people they are.

Mutter...science...physics...

PO is very much part of the utilitarian package and also part of the western liberal package (such as Mills). It is ethical from beginning to end.
Eric

h.e. said...

Just because you don't do math doesn't mean you're not making assumptions. Worse, you're making assumptions and you're probably not even aware that you're making them or what they are.

But this is as true of mathematical models as it is of verbal models. What assumption about the properties of human capital permit an ordinal measure of the total quantity? Is such a concept sensical or even analytically useful. And so on.

Robert introduced me to the capital controversies, and subsequent conversations with colleagues convinced me that there is no reason to believe the use of math neccessarily leads to greater conceptual clarity. If anything, the use of math in economics has led to a kind of laziness, where researchers write down models with abstract variables, and then look for variables from the real world to use as "proxies".

I'm not hostile to mathematical models either, because I think it is possibly to illustrate some interesting ideas that might be difficult to convey otherwise. But I have no doubt that the reliance many economists have on mathematical reasoning has lead to other deficiencies within the discipline.

YouNotSneaky! said...

Uh, what does it mean for people to "flourish"?

It sounds like you're trying to weasel out of my question through the use of important sounding but ambiguous words.

YouNotSneaky! said...

Sure, not every mathematical model tells us anything useful about the real world. No one said that every one of'em did. No one's muttered anything about physics or science either.

And there are obvious reasons why mathematical modeling is the practice within orthodox, and many non orthodox, economics. Clarity, precision, conciseness and the fact that it's easier for everyone to know what it is you're talking about. Unlike the word "flourishing".

Orthodox economists never stop to think about WHY they do models except "that is what economics is."

How do you know what orthodox economists think about and what they don't? You're making stuff up here, plain and simple. As an orthodox economist I'd like to say that I think all the time about WHY I do models. Because I think it's the best approach to understanding social phenomenon. God, by now I'm sounding like a broken record in my effort to not mutter and give precise, clear answers to the question you posed. I'm sure it won't count for much though.

PO is not utilitarian. Where's the "good", where's the "greatest number", where's "diminishing marginal utility"??? Hell, you don't even need utility functions for it. I don't know, it may be part of the "Western liberal package". Whatever. You can do orthodox economics without stopping to think about PO even once.

YouNotSneaky! said...

But this is as true of mathematical models as it is of verbal models.

Sure, it can be true if you got a lazy economist. But because every thing's clearly written out for everyone to see there'll almost always be someone to point out the assumptions to you, even if they're not immediately obvious. Just think of all the sweat and tears that've been poured out over every aspect of the assumptions of so called "rationality" in economics. Why? Because the definition of economic rationality - maximization of complete and transitive preferences over a choice set - has been precisely and clearly formulated.

Whereas verbal exposition leads to a bunch of people endlessly arguing about "What did X really means about Y when he said Z?". I guess it does provide employment opportunities for HET people.

YouNotSneaky! said...

My questions are different from your questions.

My questions are the questions that most people, including those who set policy that affects us all, are more interested in. And you haven't given me an answer. Perhaps you should dump all of that foolosophy and re-read the last few pages of Candide.

walt said...

PO is clearly a utilitarian criterion, one that sidesteps the obvious objections to utilitarism, while sharing the same basic world-view. Someone who thinks the correct social welfare function is the total devotion of humanity to the exaltation of the Lord is not going to put much stock in the Pareto criterion.

The fact/value or positive/normative distinctions are not sufficient to really classify the behavior of economists. What is the invocation of complete and transitive preferences? What is the independence axiom in vN-M utility? They're not "facts", since facts are generally not falsified by famous experiments. They're not really "values" either, since economists use it positively and not normally. Usually, when economists invoke "economic theory", they're working in this third category of things that are not facts or values. And it's this third category where most of the contention lies.

h.e. said...

And there are obvious reasons why mathematical modeling is the practice within orthodox, and many non orthodox, economics. Clarity, precision, conciseness and the fact that it's easier for everyone to know what it is you're talking about.

Except that episodes like the capital controversies show that everyone doesn't know what they're talking about. "I'm going to introduce a variable Z into the production function that is a measure of the sum of human knowledge." "Umm, what do you mean by knowledge, and how would you produce such a measure." "I don't really know, but I'll just add up the number of patents and call that a proxy." In fact, the economic literature seems to encourage exactly this degree of abstraction. So if you want to say that by writing down a model, anyone can check to see if your math is correct, I agree completely. If you want to make the case that everyone knows exactly what you are talking about when you introduce symbol x and call it whatever, I disagree. The concepts are only well defined as mathematical entities.

Part of the problem is that language produces categories, but not well defined metrics for summing up within categories. It's one thing to talk about pollution. Clearly there will already be problems around agreeing to what substances can be correctly categorized as pollution. Economists tend to sidestep these issues entirely, leaping to discussions about how ill-defined "quantities" of pollution affect social welfare, etc. How one can possibly declare that this practice leads to precision of argument is completely beyond me.

Robert Vienneau said...

gjbuwThanks for all the comments. Some of these paths are well-trodden. I'd rather read than comment, though.

By the way, Eric has referenced Hilary Putnam on the fact/value distinction. I haven't read that particular book, but I think I've seen an earlier argument along the same lines in Putnam's Resaon, Truth, and History. (I've also read Putnam on Pragmatisim and on Artifical Intelligence in his Representation and Reality. He's a very good philosopher.)

Gabriel said...

>> "They're not "facts", since facts are generally not falsified by famous experiments."

They're "false facts". Facts of a world that's not this one, if you will.

I don't know how to study this world other than by studying other, more mathematically tractable worlds. A.k.a. models.

If you can show, for example, that the time path of an observable variable from this world gravitates around a smooth path from an artificial world calibrated in a certain way, then you're ready for crude predictions.

The difference between this and usual, mental models is the clarity of exposition and the fact that you get to use cookbook calculus, I guess.

Gabriel said...

P.S. Philosophers like to argue (understatement!) but the is-ought distinction is still mainsteram (Take, for example, the recent-ish literature on moral skepticism as proof of interest).

It's a problem of (formal) inference. How can you reach an "ought" conclusion without having an "ought" in the premises?

That being said, of course that the Pareto criterion is normative/ethical/etc. and that it's at best uninteresting.

Talk of human flourishing all you want, but prince controls with still f* up the economy.

YouNotSneaky! said...

Except that episodes like the capital controversies show that everyone doesn't know what they're talking about.

Funny, I was actually thinking of the CCC as a perfect illustration of my argument. I think (of course it's just what if speculation, but take it for what it's worth) that if Samuelson and co. had been arguing verbally without having specific mathematical models in mind, Robinson would have never gotten him to admit he was wrong.

Part of the problem is that language produces categories, but not well defined metrics for summing up within categories.

Sure, but I think here we're arguing about the matter of degree. While mathematical modeling can still be unclear and ambiguous, whatever unclarity and ambiguity it contains will also be present in a verbal exposition... plus some more! That's the point.

YouNotSneaky! said...

I don't particularly feel like a have a dog in this fight but

PO is clearly a utilitarian criterion

PO may be the bastard son of utilitarianism, but it is not utilitarianism itself. IT IS an obvious, perhaps unsuccessful, attempt to get the ethics out of something like utilitarianism, but still retain something useful - a criteria. But like I said, you can do all economic theory, with the exception of Social Choice, without bothering about the Pareto Criteria once. Of course then you have a lot less to say.

The fact that most economists do consider it a useful criteria doesn't change anything. Just like if most economists thought that doing the Lord's work was the supreme criteria this would in no way affect the accuracy of the theory - whether demand curves slope down, whether prices are equal to rates of marginal substitution or production coefficients, whether tariffs lower imports, etc.

h.e. said...

Sure, but I think here we're arguing about the matter of degree. While mathematical modeling can still be unclear and ambiguous, whatever unclarity and ambiguity it contains will also be present in a verbal exposition... plus some more! That's the point.

Perhaps. But why didn't the CCC lead to at least a continued acknowledgement in at least the textbooks that there are deep conceptual problems? Instrumentalism alone can't explain it, because not every person that studies economics is interested in prediction. Nor is a science simply about prediction; the theory of evolution explains observed phenomena without being able to provide precise predicitons.

I think the real reason the CCC had no effect is because current practice does permit a kind of "free lunch". Current practice permits construction of tractible models with clean predictions. Again, though, what these predictions mean or how they are to be "tested" is unclear, even avoiding discussions about ceteris paribus and so forth, when researchers have given so little thought to what the variables are supposed to represent, how (and if) they can be meaningfully measured, and whether ad hoc relationships between these variables have any justification.

I suspect part of the problem is that economists generally rely on secondary sources for their data, and hence give little thought to measurement. There's something painful about talking to someone with a Ph.D in economics and asking them what the unit of measurement is for capital and getting a blank look followed by incoherent mumbling. And as I've stated before, this is not an issue that applies only to macro-aggregates.

h.e. said...

Of course then you have a lot less to say.

To other economists, perhaps. Again, though, there is something suspicious about a field of study that emphasizes a rather dubious moral criteria above all else. And again, whether or not economic research relies heavily on this criteria (and I would argue it does in a number of fields, including trade, growth and environmental economics), it forms the basis of undergraduate teaching. The claim that markets produce too much pollution is generally based on a very specific criteria. Unfortunately economists generally seem to receive very little instruction in moral philosophy.

So you end up with a discipline making policy prescriptions that (i) has limited knowledge of moral philosophy, (ii) a dated view of science based on Popperian falsification, and (iii) little knowledge of social measurement. A very sad state of affairs indeed.

Peter H said...

The criticism of formalism is not that it's worthless, but rather that formal mathematical models have costs as well as benefits. To quote the Post-Autustic Economics Review:

"The Post Keynesians Victoria Chick and Sheila Dow make an equally powerful, if largely implicit, case for pluralism in their penetrating analysis of what is implied by mathematical modelling in economics. Formalising an argument is not, they suggest, an unambiguous improvement, as neoclassicals believe. On the contrary, it is a matter of costs and benefits. Formalism entails a particular view of the world, namely that it displays event regularities strong enough for it to approximate to a closed system. It also requires that the meaning of economic terms be fixed rather than context-specific, and that these terms are separable rather than internally related. If these assumptions are rejected, classical or formal logic is inapplicable and Keynes’s ‘ordinary logic’ may be needed in its place. Ordinary, common-sense or human logic ‘generates knowledge which is imperfect, partial or vague’, and provides ‘reasoned grounds for belief which are nevertheless not conclusively demonstrable’ (Chick and Dow, 2001:711, 714). Economic statements may therefore be true in some historical and institutional circumstances, but false in others."

See here.

YouNotSneaky! said...

Like I said before, I'm slow when it comes to philosophy (again, last few pages of Candide works for me). Gimme an illuminating example of this "ordinary logic" that you speak of.

Gabriel said...

Allow me to interject and translate: "Talking your way through problems."

Lucas, in his Nobel address I think, makes a point of explaining that this is exactly the reason why early monetary theorist couldn't make any progress.

I must confess that I have no idea who Chick and Dow are, but I'm sure that they never tried to talk their way through a model with serious interactions between expectations and random shocks or game theoretic setups.

Alternatively, they might mean "weekend psychologizing". The behavioral assumption in the General Theory are presented as psychological/common sense facts of sorts. Bad idea.

Robert Vienneau said...

Gabriel is, of course, mistaken about what he says he is "sure" of. Chick has given a mostly verbal exposition of a model with those characteristics. And I believe I have previously mentioned Tony Lawson on this blog.

Gabriel said...

I stand corrected. They (she) did. And what do you know, it's a "What did the Master really meant?" text. My favorite!

walt said...

So Gabriel, you think what, exactly? If you can't put something in a mathematical model, then it doesn't exist?