BEHAVIORAL ECONOMICS... What are your thoughts?

[tm_member]

Just to add to the debate, I'll share a fantastic quote that got me into behavioral economics in the first place... I can't remember where it's from, but it goes something like this...

"The goal of human nature is reproduction, not maximizing happiness"

That made me question everything and realize that there is another huge force of nature acting on us.

[chateauheart]

Too much of behavioral economics is motivated by a naturalistic rejection of "old" neoclassical economics without really understanding why the neoclassical assumptions are made, or when they are useful. Take the example of self-interest, which is really a tautological construction designed to highlight the fact that individuals have their own preferences which can be quite separate from social preferences (in contrast to the traditional social scientific view that people, in particular political actors, will naively follow what the "social planner" expects), and that these distinct incentives will have huge consequences for the applicability of some public policies. Of course, that rational choice paradigm has been turned into this huge strawman that all individuals are just giant calculators trying to maximize their pecuniary payoffs and hedonic satisfaction. Which is, of course, nonsense.

Believe me, the founders of neoclassical economics were not fools; they had to fight tooth and claw to get the rest of the discipline (then dominated by Keynesian style handwaving about human nature, which quite closely resembles some parts of modern behavioral economics) to follow their approach. And the rest of the discipline eventually were persuaded. The neoclassicists included people like Friedman, Buchanan and Lucas, who were all socialists in their youth. Lucas was actually a Marxist all the way until graduate school. Surely they didn't just become leading neoclassicists because they were naive. Even radical Marxian economists like Bowles and Gintis were later convinced of the usefulness of neoclassical economics (they are now working on evolutionary game theory), after Stiglitz managed to incorporate asymmetric information into the neoclassical framework. How easy it is to forget the knowledge that we've gained.

This is not to say that behavioral economics is not useful or that there aren't a lot of extremely capable behavioral economists; I'm only ranting about a small subset of the field. But this subset of the field is enough to make all of it quite meaningless; they produce strawmen to knock down; then others trying to point out that they are really strawmen, and in the end we've learned nothing. Or, in other cases, people find something that doesn't hold true in controlled experimental settings where the costs involved are insignificant; but some other people find out that it's a completely different story when you move to actual situations in the market. The work on risk aversion and prospect theory is the one field where I really think they added some value to "traditional" economics; but even then the significance is hugely overblown. I think there's a huge adverse selection effect here that dilutes or negates much of the productive ideas in behavioral economics and make people turn to polemical, poorly done research. (The fact that many of the more polemical behavioral economists gain tons of popularity from some political ideologues hellbent on proving neoclassical economics is evil is quite convenient.)

[jaimelannister]

Just to add to the debate, I'll share a fantastic quote that got me into behavioral economics in the first place... I can't remember where it's from, but it goes something like this...

"The goal of human nature is reproduction, not maximizing happiness"

That made me question everything and realize that there is another huge force of nature acting on us.

Yes, I think economists can gain a lot from learning a bit about evolution theory. People might be surprised in how both fields have in common.

[tm_member]

Too much of behavioral economics is motivated by a naturalistic rejection of "old" neoclassical economics without really understanding why the neoclassical assumptions are made, or when they are useful. Take the example of self-interest, which is really a tautological construction designed to highlight the fact that individuals have their own preferences which can be quite separate from social preferences (in contrast to the traditional social scientific view that people, in particular political actors, will naively follow what the "social planner" expects), and that these distinct incentives will have huge consequences for the applicability of some public policies. Of course, that rational choice paradigm has been turned into this huge strawman that all individuals are just giant calculators trying to maximize their pecuniary payoffs and hedonic satisfaction. Which is, of course, nonsense.

Believe me, the founders of neoclassical economics were not fools; they had to fight tooth and claw to get the rest of the discipline (then dominated by Keynesian style handwaving about human nature, which quite closely resembles some parts of modern behavioral economics) to follow their approach. And the rest of the discipline eventually were persuaded. The neoclassicists included people like Friedman, Buchanan and Lucas, who were all socialists in their youth. Lucas was actually a Marxist all the way until graduate school. Surely they didn't just become leading neoclassicists because they were naive. Even radical Marxian economists like Bowles and Gintis were later convinced of the usefulness of neoclassical economics (they are now working on evolutionary game theory), after Stiglitz managed to incorporate asymmetric information into the neoclassical framework. How easy it is to forget the knowledge that we've gained.

This is not to say that behavioral economics is not useful or that there aren't a lot of extremely capable behavioral economists; I'm only ranting about a small subset of the field. But this subset of the field is enough to make all of it quite meaningless; they produce strawmen to knock down; then others trying to point out that they are really strawmen, and in the end we've learned nothing. Or, in other cases, people find something that doesn't hold true in controlled experimental settings where the costs involved are insignificant; but some other people find out that it's a completely different story when you move to actual situations in the market. The work on risk aversion and prospect theory is the one field where I really think they added some value to "traditional" economics; but even then the significance is hugely overblown. I think there's a huge adverse selection effect here that dilutes or negates much of the productive ideas in behavioral economics and make people turn to polemical, poorly done research. (The fact that many of the more polemical behavioral economists gain tons of popularity from some political cranks hellbent on proving neoclassical economics is evil is quite convenient.)

Again, can we please avoid referring to anyone as a "crank"...

[Humanomics]

Just to add to the debate, I'll share a fantastic quote that got me into behavioral economics in the first place... I can't remember where it's from, but it goes something like this...

"The goal of human nature is reproduction, not maximizing happiness"

That made me question everything and realize that there is another huge force of nature acting on us.

You will enjoy Joan Roughgarden's response to Dawkins: The Genial Gene. She's a wicked smart sociobiologist, and finds virtually no evidence to support the coy female / aggressive male version of sexual selection implied by economistic "cheap sperm" and "expensive eggs" etc. She concludes that what's being maximized, globally, is the rate of successfully reared births -- not the rate of fertilizations. What you see thus in most animals is striking cooperation (the whole "it takes a village to raise a child" idea), as against the Spencerian notion of "nature red in tooth and claw."

"Reproduction" is also used (IMO annoyingly) in other social sciences in a path-dependency sense. I asked a very good political theorist at my school to get me clear on it and he reported thus:

I'm surprised that you hear "reproduction" in pol sci; soc seems more likely. I'm guessing that used in this way there's some sort of connection back to or through Marx. Chapter 23 of Capital, vol is titled "Simple Reproduction" . . . . So from Marx, among others, you have along with the idea of society the idea of social reproduction--that any society needs not only to produce but to reproduce.

This reproduction refers not only narrowly to human reproduction, but to the reproduction of conditions of production, relations of production, etc. This is conceived very much as a material process assigning no mysterious agencies etc . . . that we live under conditions where things call the shots, where dead labor directs living labor rather than the other way around, etc. under capitalist social conditions capital as social relation and as self-expanding value has a kind of prosthetic agency.

If the talk is about ideology or something like that there might be a link of some sort to Althusser and other strains of twentieth-century structuralism, many of them traceable in different ways back to Marx. Here again the ideas themselves aren't supposed to be at work in any independent way.In some ways, this talk of reproduction isn't all that far removed from what the 18th century called "custom" and "habit"--it's just got a bit more of an industrial machine-like and quasi-deterministic cast to it. A lot of it is just recognizing that the way we live at any time or place has its conditions of possibility, and that many of those conditions of possibility (the ones that depend on human agency) have to be repeatedly serviced or replaced, consciously or not.

Whether materially-based or not, I don't agree that any kind of independent industrial, social, or ideological machinery reproduces itself with agents just kind of bumping around inside that structure. So I really don't like the term "reproduction," but stuff like the above really helps inform and know your opponent.

[Humanomics]

Believe me, the founders of neoclassical economics were not fools; they had to fight tooth and claw to get the rest of the discipline (then dominated by Keynesian style handwaving about human nature, which quite closely resembles some parts of modern behavioral economics)

Keyne's animal spirits were a small part of his thesis -- as I understand it most of his thesis turned on wages lagging behind prices (a pretty loony assumption in a market where workers are even relatively alert or remotely prudent). And I don't want to presume to much in your meaning, but "hand waving" does not mean "social science done without Greek notation" it means "logical statements averred without sufficiently discussing their axiomatic assumptions." You will find, for instance, a great deal of hand-waving in Rhetoric or Philosophy departments, where lots of proofs are written without epsilons and subscripts.

The neoclassicists included people like Friedman, Buchanan and Lucas, who were all socialists in their youth.

This is largely true of the vast majority of economists of that generation. Just about everyone who has been convinced of Price Theoretic results got to their believing, motivated by an enormous concern about exploitation etc. The debate on positivism in fact, can be read as a sort of born-again fanaticism about not socially engineering, which most of these people were completely intent on doing when they came to economics at first.

"Neoclassical" anyway is usually used historically incorrectly, a fancied up version of "mainstream economics," which is an even dumber way to divide up schools of thought, by popularity. McCloskey writes,

"I think it's about time to stop calling them the "mainstream," or even the anachronistic "neoclassicals," . . . "Neoclassical" economics includes in historical precision the Austrians and the Marshallians, who find the contents of most economics journals nowadays to be unscientific rubbish. The Samuelsonians are best thought of as the children and grandchildren of Paul Anthony Samuelson and his brother-in-law Kenneth Arrow. Larry Summers is their mutual nephew---honest, he is . . . The family bible of the Samuelsonians is Paul's modestly entitled Ph.D. dissertation, The Foundations of Economic Analysis, which proposed in 1947 that all of economics should produce qualitative existence theorems out of a method of constrained maximization under the sign of logical positivism.

What most people have in mind when they talk about "neoclassical" economics is Samuelsonian economics.

And this is the real problem with people's thinking on economics outside Samuelsonian economics:

I think there's a huge adverse selection effect here that dilutes or negates much of the productive ideas in behavioral economics and make people turn to polemical, poorly done research. (The fact that many of the more polemical behavioral economists gain tons of popularity from some political ideologues hellbent on proving neoclassical economics is evil is quite convenient.)

A majority of justification for discounting work that's not strictly reducible to a game theoretic or constrained maximization choice problem is not a conscientious reading of, e.g. institutional, evolutionary, or network theory but rather an ad hominem sneer about how the fringes attract a bunch of rejects who weren't good enough at math or responsible enough researchers to make it in the smoking club. It's a penny-ante way to conduct disciplinary politics. Or wait, we don't have politics in economics -- we have price signals successfully cordoning all the bad researchers into tidy categories like "Behavioral," which might be neatly discounted and ignored toward the betterment of the science.

Lobbing stones at subdisciplines who aren't based solely in Samuelsonian micro foundations is not justified simply by asserting that they're a bunch of spanghewing meshuggeners who threw the first stones at us "neoclassicals."

[Humanomics]

Think about a plant. It cannot learn calculus, or even do it subconsciously. Yet it reaches for the sun as if finding the minimum distance between it and the sun. Friedman (53) argued that the fact it acted as if optimizing was sufficient ground to continue the Samuelsonian program -- who cares how the plant gets to its optimum; it gets to its optimum. I think he was only half way there, and this is where Behavioral is now coming in: we cannot afford to ignore the heuristic the plant uses to actually find where the sun is brightest (how people actually do march up and down supply and demand curves).

Globally, sure, something like a stable general equilibrium emerges from billions of market transactions. But that "second tier" analysis diminishes in returns offered when trying to understand the things specific agents say and do in an economy, on the first tier. How does the plant actually search for the sun? How does a firm get information about demand? Talking to its customers? But I thought the only way to acquire information on preferences was by revealing them through price-voting. Wait.

We have treated (with great success) the origins of property rights themselves, and tastes themselves, merely as givens. But purposive agents construct these givens, and they are scientifically important variables. There are empirical margins on which the Samuelsonian program is failing miserably -- technological innovation, for instance. That doesn't mean the Samuelsonian program is garbage, it means if we are to understand economies more thoroughly we need to have a mature conversation about contested perspectives.

[aecon641]

This conversation always leads to confusion when people fail to define rationality and thereby neglect the proper target of behavioral theory. You learn in micro 101 that rationality entails i) transitive and ii) complete preferences. All behavioral theories essentially challenge either i), ii) or both.

Consider the theory of hyperbolic discounting (Frederick, Loewenstein,O’Donoghue) which posits that agents have time-inconsistent preferences. At its heart, this is a challenge to the claim that preferences are transitive. To illustrate, under neoclassical assumptions of rationality, if on Monday I choose to do my taxes on Tuesday, I will still prefer to do my taxes on Tuesday when Tuesday arrives (this much will be true when I exhibit exponential discounting). On the other hand, if I exhibit hyperbolic discounting, it is possible for me on Monday to prefer doing my taxes on Tuesday while when Tuesday arrives I suddenly (ceteris paribus) prefer to do my taxes Wednesday. In other words, my utility function exhibits a present-bias in each period.

While this model captures human propensity towards procrastination (among other things), it violates the assumption of preference transitivity. Hence, relaxing the assumption of rationality in certain instances (e.g. savings behavior) allows one to more closely approximate or explain human activity.

[jaimelannister]

Not necessarily. The behaviour can be still be 'transitive' (trivially at least) by specifying the choice as a tuple of the original choice and time. But this just illustrates the point RCT is just maths. You can add bells and whistles to the underlying model but there must be a purpose to it otherwise it can add a cost without much gain. Modelling behaviour with hyperbolic discounting may be pertinent to some applications but for a lot of cases I think making the standard assumptions is fine.

[Humanomics]

Gintis assumes transitivity and completeness at the beginning of Bounds of Reason and proceeds. That work would definitely fall under Behavioral. And I don't agree that rationality is properly defined as only those two preference axioms. Utility maximization goes in there, along with (depending on your method), the equating of cost and benefit at the margin.

The work on preferences is very useful though.