July 2014

Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    
Blog powered by Typepad

« Time Selects Lin Ostrom As One of the 100 Most Influential People | Main | Is Debt Forgiveness the Answer to the Sluggish Recovery? »


TrackBack URL for this entry:

Listed below are links to weblogs that reference What Can Economists Know?:


Feed You can follow this conversation by subscribing to the comment feed for this post.

I have a good heuristic that economists can use:

Use "as if" optimization not as a way of explaining thought processes, but as a way of modeling behavior (two very different things), unless you know of a heuristic that is more accurate.

The bias-variance tradeoff point at the end is very important. But notice how he made that point - by showing us a graph of the prediction error as a function of the order of the polynomial, and minimizing it.

In other words, our brains are still optimizing. I guess my point is that I agree on how important heuristics are to understanding human behavior - but we should not think about heuristics as an alternative to optimization. Instead we should think about them as an alternative to calculation. And that trade off can be set up as an optimization problem itself because calculation expends cognitive resources.

I find gigerenzer's point coherent with Vernon Smith + Hayek. My problem with such a kind of optimization is that local optimization sometimes is not the same thing that global optimization. I live in Spain (and I saw human behavior under a bombastic bubble, all of us optimizing).
In this session, the conference that I enjoyed so much was that of Nancy Cartwirght.

Also today, Peter Dorman has this comment at Econospeak (http://feedproxy.google.com/~r/espeak/~3/LQCGxrmI8eE/anonymous-wisdom-on-so-called.html):
"All we can do, at best, is arbitrarily identify a large set of assumptions and forecast conditionally on them. (And one catch-all assumption is that none of the consequential unknown unknowns will materialize.) For this reason, forecasting may be a false goal. A more serviceable one would be to identify processes with known dynamics in as close to real time as possible."

Combining this thought with the speech topic, I think economists can 'know', or at least study, the heuristics individuals use in specific real-time situations, such as catching a baseball. I'm not sure this is still optimization since there is no defined domain. Using the baseball example, among other things, an outfielder must also consider the potential for a change in wind. The best path may therefore be different than any previous path.

Very interesting!

I see Austrian econ as having developed out of the heuristic approach. Beginning with the School of Salamanca, scholars observed what people did in the market and found patterns.

That is similar to the way a person learns to catch a ball: they remember patterns of behavior that worked.

Theory came in to explain the thought processes of the observed behavior.

Mainstream econ, instead, began with abstract theory - equilibrium - and hopes that some day the real world will conform to the theories and equations.

PS, by predicting patterns instead of striving for exactness, Austrian econ avoids over-fitting that mainstream econ suffers from.

Really interesting. I've ordered his book 'Rationality for Morals' to find out more about this.

FWIW, I reviewed Gigerenzer & Selten's edited volume "Bounded Rationality: The Adaptive Toolbox," in JEBO 53(3), http://tinyurl.com/7vpx66f.

Oops. This link should work:



I'd love to read your review -- but at an access cost of $41?

Though I'm sure that each of your words is a "pearl of wisdom" . . . I think I'll wait for the movie version.

Richard Ebeling

Yeah, if you institution does not have a subscription, it's crazy pricy. Sorry about that. More fuel for the Elsevier boycott, I suppose.

I must say that I was fascinated by this video discussing the difference between real decision-making under uncertainty vs. choice-making under knowable and calculable risk.

The discussion of following simple patterns for action rather than presuming complex calculations reminded me of the controversy over marginal decision-making within the firm that Fritz Machlup participated in, in the early 1950s. Some may recall that Machlup used the example of a driver who decides whether or not it is safe and possible to overtake a car ahead of him with another car coming toward him in the opposite lane.

The driver does not and cannot make a precise "equation-like" calculation in his head to decide whether or not it is safe to overtake the car in front of him.

He follows tacit rules based on experience on "measuring" distances and safety based on visual estimate in these matters -- pattern relationships in his head -- with the result the most of the time all of us in such situations make the "right" decision.

Marginal decision-making, Machlup argued, follows this latter method of deliberation within the heads of the decision-makers in the firm, and does not require the mathematical version with full and detailed knowledge of production conditions.

May I suggest that the same applies to our consumer marginal choices. We do not make mathematically precise estimates of the marginal gain vs the marginal cost from this trade-off or that.

We follow "rules" in our head that assist us in these choices when we cannot know ahead of time all the details we would need to know to make that presumed "rational" choice under conditions of complete knowledge or information.

This is, perhaps, one of the reasons that Frank Knight stated in his 1933 article on marginal utility that the theory seems more commonsensical, realistic and persuasive in Carl Menger's formulation of marginal decision-making than that in Jevons' and Walras' more mathematically refined versions of the concept.

Richard Ebeling

Another example: On a home repair show on TV a carpenter was building a railing for a porch in which he wanted the vertical slats evenly spaced. He discussed the math involved, which was complex for non-math people. Then he showed his method.

He bought a roll of elastic such as is used in the waist of pants. Next he drew marks on the elastic spaced an inch apart, one mark for each slat. He placed the elastic on the bottom board of the rail and stretched the elastic the length of the board. Then he transferred the marks from the elastic to the bottom board.

Heuristics are great!

To add to Richard's remarks, we must ask this question:

What the heck is the pure logic of relational valuation for? Why did it have to be "discovered" and articulated, how does this articulation advance our understanding, and what use is it in addressing the problem of design-like / plan like coordination among many via money exchange _without_ any top-down designer or planner?

It seems to me that 99.999% of all economists don't think to ask such questions and can't come close to competently answering them.

We don't need logic to be logical -- and we check formal logic with everyday examples of reasoning with which we have far greater and more confident competence.

But then, once we have the formalisms of logic memorized, they become useful.

But their relation to explanation, argument, and advancing understanding remains obscure, contested ... and inevitably badly botched by most top "logicians" and "philosophers", etc.

Not much separates economics and philosophy at their kissing cousin core in this regard.

They each produce a logical formalism -- and can't make sense of what they are or what their relation is to anything else.

"Heuristics are great!" Here's what I've been going around saying. Philosophy is the study of what men say...while economics is the study of what men do. Actions speak louder than words. Every action you take sacrifices all the other actions that you could have taken. Giving people the freedom to prioritize the use of their limited time/money is the recipe for economic prosperity.

The other day on Netflix I really enjoyed watching Stephen Hicks' documentary..."Nietzsche and the Nazis". Really really great stuff. It's definitely worth your time to watch it. It should go without saying that I can't really know that.

At the end of the video...Hicks says it's really important that we understand the Nazi ideology in order to inoculate ourselves against any future manifestations. He said it a lot better than I have...of course. But I really like that idea of inoculation. You know how I know for a fact that we are not inoculated? We don't give taxpayers the freedom to choose where their taxes go.

While I'm recommending Netflix videos...another one that I'd highly recommend is "The Man from Earth".

Between both those very simple, yet extremely powerful films, you get the sense that we're shooting ourselves in the foot by not doing something similar.

The heuristic that I'm applying to everything is the idea of sacrifice. Sacrifice is something that men do for something that they want. If you want to understand economics then you need to understand sacrifice...from the big stuff...Jesus' sacrifice...all the way to the little stuff...people tossing coins into wishing wells.

Are you willing to sacrifice your time to watch those two Netflix films? Why did I sacrifice my time to tell you about those two films?

When I get a chance I'm going to e-mail Stephen Hicks and see if I can't encourage him to do a documentary on scarcity, sacrifice, tolerance and prosperity.

As a general statement,Gigerenzer is terrific. He is in keeping with Vernon Smith insofar as Gigerenzer provides more "micro foundations" for Smith's ideas on the importance of institutions. Smith has always downnplayed how excactly individuals make or do not make decisions within a framework.

Most importantly, Gigerenzer sounds the death knell to the idea of contentless or axiomastic theories of choice rationality. This is clear from, among other places, his article "I Think, Therefore I Err." in his book RATIONALITY FOR MORTALS.

Of course another advantage of using heuristics under uncertainty is that if others are following the same heuristics you can predict their behaviour and thus, as Hayek says, extend the range of your knowledge.

Richard's Machlup example only makes sense if we know that the other drivers will follow the same rules of the road as us.

ALso well worth reading is Gary Klein on the power of intuition and heuristics. Kahneman in Thinking Fast and Slow describes his debates with Klein.

Btw, Kahneman totally ignores Gigerenzer. My suspicion is this is because he cannot answer him. But hey, Kahneman has a Nobel prize.

I tell myself I must have faith. Rushangshenglongba

This is what AI researchers found out a long time ago, logic is not enougth in a uncertend universe. If you look at how the Google IA car drives around its a lot of formulas, calculation and prediction BUT they are all guesses and get ajusted all the time when diffrent information comes in. As soon as the car next to you speeds up new guesses are mad where the car could be next.

Its was the baseball player does when he catches the ball, the diffrence is that the baseball player has brain that does this calculations in a way we dont understand, we have to make up more or less complex models to do this.

When we build a car that drives by itself the care must learn to take in new date (like when the baseball-player looks at the ball), calculat new guesses based on mathmatical models (we dont know how the baseball-player does this) and then ajust the action to the new guesses, this process is done over and over again multible times per second.

There is no great calculation at the start, its all bad and incomplet data together with rough guesses combined in a feedback loop.

The comments to this entry are closed.