Gelman's Law

April 27, 2018 11:31 AM

A quick rule of thumb is that when someone seems to be acting like a jerk, an economist will defend the behavior as being the essence of morality, but when someone seems to be doing something nice, an economist will raise the bar and argue that he’s not being nice at all.

From Andrew Gelman's blog. Elevated to the status of Natural Law by Henry over at Crooked Timber.

A favorite bit from the comments so far, regarding the purported penis envy by lay-people of economists', um, penetrating quantitative skills: "I would suppose that the number of people envious of the ability to comprehend the entire universe as a vast linear regression is likely to be small."

posted by mondo dentro (30 comments total)

What did the economist say when he walked into a room and discovered a man trying to cut off his own fingers with a pair of scissors?

...

"Can I interest you in a sharper pair of scissors?"

(Not the fairest joke, admittedly, but so long as I'm maximizing my subjective utility in telling it, I don't see what grounds an economist would have for objecting to it.)
posted by belarius at 11:43 AM on April 27, 2018 [16 favorites]

There's that old joke about physicists and spherical cows, but I feel like the joke might be more accurately attached to economists. And also the economists are in denial about the fact that their model has made any simplifications at all.
posted by tobascodagama at 11:52 AM on April 27, 2018 [4 favorites]

(Rationally, I'm aware that there are definitely good economists -- probably most of them, in fact -- who not only appreciate the differences between their models and reality but also avoid falling into the pattern Gelman is talking about here. But there's a sort of publicity bias here, where the sad state of journalism means that hackish contrarians are almost the only academic voices that ever get amplified.)
posted by tobascodagama at 11:55 AM on April 27, 2018 [4 favorites]

Economists: the first #slatepitch-ers
posted by leotrotsky at 11:57 AM on April 27, 2018 [1 favorite]

Freakonomics: it’s counterintuitive!
posted by sjswitzer at 12:10 PM on April 27, 2018 [1 favorite]

I think it's not described as the essence of morality, it's described as the essence of practicality.
posted by Nancy Lebovitz at 12:11 PM on April 27, 2018

Economists/statisticians have their own jokes about the inaccuracy of their models:

Three econometricians went out hunting, and came across a large deer. The first econometrician fired, but missed, by a meter to the left. The second econometrician fired, but also missed, by a meter to the right. The third econometrician didn't fire, but shouted in triumph, "We got it! We got it!"
posted by FJT at 12:14 PM on April 27, 2018 [15 favorites]

The smartest economists I know got an advanced degree in the subject and thereafter refused to take a job in the field.
posted by tilde at 12:20 PM on April 27, 2018 [8 favorites]

My economist acquaintance was so frustrated with the field, she finished her phd and went directly for an mdiv.
posted by aniola at 12:54 PM on April 27, 2018 [5 favorites]

The issue of American Niceness. P.S. In Middle English, the word "nice" meant "stupid/foolish". It then changed throughout the 17th-18th centuries to mean, "exact"; hence, "niceties". Today, of course, it means, "pleasant/kind", but one always wonders if words can be haunted by their etymologies.
posted by Tarn at 12:57 PM on April 27, 2018 [8 favorites]

Well you know, according to the labor theory of value ... oh the heck with this. I fork and quit.

Stab a donut once, why stab it again?
posted by Beginner's Mind at 12:58 PM on April 27, 2018 [1 favorite]

an economist will raise the bar and argue that he’s not being nice at all.

Also known as 'No Good Deed Goes Unpunished'.
posted by eclectist at 1:11 PM on April 27, 2018

I thought Gelman's Law was that the 1000th cat is you.
posted by grumpybear69 at 1:28 PM on April 27, 2018 [1 favorite]

Economists are the grease monkeys for the machine of Capitalism -- or put slightly differently, economists apply the lube for capitalists.
posted by jamjam at 2:17 PM on April 27, 2018 [3 favorites]

There's that old joke about physicists and spherical cows, but I feel like the joke might be more accurately attached to economists. And also the economists are in denial about the fact that their model has made any simplifications at all.

When a physicist starts talking about spherical cows, at the end there's almost certainly going to be actual math that can be evaluated. Whereas when an economist (especially the popular book/blog writing kind) starts talking about a 'model', only rarely do you get a concrete set of equations, far more frequently you get napkin sketches of curves, an outline of what those equations might vaguely look like, and then a great deal of confident speculation about what those unwritten equations are going to do. It's like a spherical model of a model.
posted by Pyry at 2:24 PM on April 27, 2018 [18 favorites]

I think it's not described as the essence of morality, it's described as the essence of practicality.

It's described as normative - what should be done.
posted by justsomebodythatyouusedtoknow at 2:35 PM on April 27, 2018 [1 favorite]

Pyry: When a physicist starts talking about spherical cows, at the end there's almost certainly going to be actual math that can be evaluated. Whereas when an economist (especially the popular book/blog writing kind) starts talking about a 'model', only rarely do you get a concrete set of equations, far more frequently you get napkin sketches of curves, an outline of what those equations might vaguely look like, and then a great deal of confident speculation about what those unwritten equations are going to do. It's like a spherical model of a model.

That's not my experience of economics, or at least not my experience of economics from the physics-envy era. Most of what I've seen is towers of equations built on shaky assumptions, with the answer to any objection being that you're not allowed to criticize any of it until you understand all of it.

"Why are y'all building this giant inverted Eiffel Tower on top of a child's sandcastle?" is not considered a polite or reasonable question.
posted by clawsoon at 2:44 PM on April 27, 2018 [12 favorites]

There are two kinds of Economists: those who buy into the system, and those who don't.
posted by ovvl at 3:51 PM on April 27, 2018

The tent of economics is large enough to fit both those who would treat options pricing as a quantum spin glass and those who would decide tax policy based on a curve drawn on a cocktail napkin.
posted by Pyry at 4:02 PM on April 27, 2018 [3 favorites]

Two economists are walking together. One of them sees a new idea on the sidewalk and says, "Hey look, a new idea! I should pick it up!"

The other one doesn't even bother to look down. "If there really was a new idea, another economist would've picked it up already."
posted by clawsoon at 4:04 PM on April 27, 2018 [13 favorites]

There are two kinds of Economists: those who buy into the system, and those who don't.

I'd say rather that there are two types of economists (and two types of social scientists in general) those who devote their careers to describing the system as they find it and those who ask further questions about how things might work otherwise. You don't need to buy into the system (i.e. think of it as normative or necessary) to find describing its inner workings useful. Marx didn't "buy into" capitalism, but he certainly studied it in great detail.
posted by AdamCSnider at 10:30 PM on April 27, 2018 [2 favorites]

And also the economists are in denial about the fact that their model has made any simplifications at all.

For the life of me I can't remember if it was a statistician comparing his field to economics, or the other way around, but whoever it was complained that the first thing every student learns in chapter one about (utility functions / gaussian distributions) is what a huge simplification of the universe it is, followed by 4 years of training working out conclusions as if they were literally true.

I'm pretty sure it was Nobel winner Kenneth Arrow who pointed out that there are lots of terms in econ (like "efficiency" and especially "optimum") that have technical definitions that do not match their English language meaning, but get trotted out to justify all sorts of crap as if they did.
posted by mark k at 11:18 PM on April 27, 2018 [6 favorites]

I wonder if some economists got into the neoliberal program of transforming all human relationships into market relationships and all incentives into monetary incentives not because they are monsters who wanted to enjoy the resulting suffering, but because they wanted their equations to work.
posted by clawsoon at 4:29 AM on April 28, 2018 [4 favorites]

This burn from the blog post, though:

Similarly, in the world of pop economics, or neoclassical economics

posted by clawsoon at 4:34 AM on April 28, 2018 [4 favorites]

I work alongside a person with an economics degree, whenever human behaviour disagrees with his model it's nearly always the person acting stupid, I'll informed or irrational. Not that the model isn't correct.
It's a most peculiar science.
posted by 92_elements at 4:51 AM on April 28, 2018

I'd be interested in hearing from a primatologist about this, from the comments:

I’ve heard a related point regarding primatology. Supposedly, primatologists are willing to describe violent primate actions in the vernacular—”fight, hit, kill,” etc. On the other hand, affectionate gestures are described in a latinate and removed sense—”affectionate gestures”, “mutual grooming”, etc.

posted by clawsoon at 5:07 AM on April 28, 2018 [7 favorites]

"Your faves are inefficient."
posted by pykrete jungle at 6:25 AM on April 28, 2018 [2 favorites]

Hm. I don't think that's accurate - if I'm getting my primatological jargon on, I'll talk about affiliative behaviors - grooming, proximity, tolerance - and agonistic behaviors - supplant, displace, threaten, display, chase, bite. I'm not really sure how to describe mutual grooming in a more vernacular way - someone suggests "nuzzle" or "stroke," but they're not nuzzling, they're grooming each other. It's describing the behavior, just like "bite" or "hit" does. I wouldn't say that two individuals "fought" without then further describing exactly what that fighting encompassed. Vocalizations? Approach-retreat? Threat yawns? Lunges? Chases? Biting or hitting? I have notes like "Norm grabbed Helga's tail and bit it repeatedly." How else would I describe that event?

Ideally, when you're describing behaviors, you're not ascribing any sort of intent to them. Some things are unambiguous. If you bite someone while they scream and try to get away, I can pretty much guarantee that that's an agonistic interaction. If you are grooming someone and they get so relaxed that they fall asleep while you're picking through their fur, that's pretty obviously affiliative. But what makes, say, proximity affiliative? When is displacing someone else an agonistic encounter and when is it just that one monkey got bored and moved and another monkey came in to take its place? Part of the remove in describing behaviors - both agonistic and affiliative - is that you run the risk of ascribing intent incorrectly if you're more casual in your language. It's pretty obvious when someone kills someone else. It's less obvious what an "affectionate gesture" is (and I've never seen "affectionate gesture" as a category on an ethogram). Describing behaviors (mutual grooming) rather than implicitly giving them a value (stroking eachother) helps maintain that objectivity when interpreting what you've observed.

I went through multiple rounds of revisions on a paper where I was reporting an infanticide attempt because I had originally failed to acknowledge that the infant could have died as the result of generalized aggression rather than assuming a targeted infanticide, although this male attacked a female with an infant repeatedly, didn't attack any females without infants, and then wouldn't let any females approach the screaming infant who had fallen 20m to the forest floor. We're generally pretty careful about our language and our methodology because people are so alert for anthropomorphism from sloppy primatologists who are swayed by the charismatic fuzziness of monkeys.
posted by ChuraChura at 8:26 AM on April 28, 2018 [7 favorites]

Another economists' two-step I've noticed is that, if the capital reacts poorly to a regulation or legal change, we'll hear how the market dislikes uncertainty. The response applies even if they grant the change itself is laudable, like an attempt to reduce CO2 emissions. But you need to take the losses caused by disruption and uncertainty into account!

On the other hand if labor reacts poorly to a disruptive change--say, lower trade barriers, more outsourcing, less job security--they are just uneducated and stupid.

(Also both Henry and Gelman are great.)
posted by mark k at 10:03 AM on April 28, 2018 [4 favorites]

Economists tend to be contrarians. That’s because economics teaches one to look for hidden patterns, unrealized motivations, and surprising mechanisms — for alternative explanations to commonly held beliefs.

Economics: "Well, actually" as a profession.
posted by clawsoon at 11:29 AM on April 28, 2018 [6 favorites]