Saturday, May 16, 2015

Paul Romer on mathiness


Top growth theorist Paul Romer has an essay in the AER Papers & Proceedings, in which he comes down harshly "mathiness" in growth theory. "Mathiness" is his term for when people (allegedly) use math in a sloppy way, to support their preferred theories. Romer warns direly that the culture of econ theory has become a lot more tolerant of mathiness:
If mathiness were used infrequently,...it would do localized, temporary damage. Unfortunately...as the quantity increases, mathiness could do permanent damage because it takes costly effort to distinguish mathiness from mathematical theory. 
The market for mathematical theory can survive a few...articles filled with mathiness. Readers will put a small discount on any article with mathematical symbols, but will still find it worth their while to work through and verify that the formal arguments are correct, that the connection between the symbols and the words is tight, and that the theoretical concepts have implications for measurement and observation. But after readers have been disappointed too often by mathiness that wastes their time, they will stop taking seriously any paper that contains mathematical symbols. In response, authors will stop doing the hard work that it takes to supply real mathematical theory. If no one is putting in the work to distinguish between mathiness and mathematical theory, why not cut a few corners and take advantage of the slippage that mathiness allows? The market for mathematical theory will collapse. Only mathiness will be left. It will be worth little, but cheap to produce, so it might survive as entertainment. 
[I]n the new equilibrium: empirical work is science; theory is entertainment. Presenting a model is like doing a card trick. Everybody knows that there will be some sleight of hand. There is no intent to deceive because no one takes it seriously. Perhaps our norms will soon be like those in professional magic; it will be impolite, perhaps even an ethical breach, to reveal how someone’s trick works. 
When I learned mathematical economics, a different equilibrium prevailed. Not universally, but much more so than today, when economic theorists used math to explore abstractions, it was a point of pride to do so with clarity, precision, and rigor...If we have already reached the lemons market equilibrium where only mathiness is on offer, future generations of economists will suffer.
Romer has now joined the chorus of old famous guys - Krugman, Solow, Stiglitz, Farmer - who are very vocally mad about the way mainstream economics theory is done.

Romer is not afraid to name names. Interestingly, although he's talking only about growth theory and not about business cycle theory, most of the people he's mad at are the same guys that the Keynesians are mad at - Robert Lucas, Ed Prescott, and David K. Levine. He also calls out Thomas Piketty.

Romer gives specific examples of what he calls mathiness (links are to working-paper versions):

1. Prescott and McGrattan (2010): Romer says that this paper includes a term that the authors label "location," but that doesn't correspond to any real measure of location.

2. Boldrin and Levine (2008): Romer criticizes this paper for assuming that a monopolist would also be a price-taker, and for making various hand-wavey arguments.

3. Lucas (2009): Romer criticizes this paper for making a hand-wavey argument to dismiss the idea that investment in embodied technology (books, blueprints, etc.) can be a source of sustained growth, when there are well-known models in which it can. Romer also points out a random math error in the paper, and uses this to argue that reviewers don't pay close attention to math.

4. Lucas and Moll (2014): Romer criticizes this paper especially harshly. Lucas and Moll claim that their model, in which there is no creation of new knowledge, is "observationally equivalent" to models in which new knowledge arrives very slowly. Romer shows that the truth of this claim depends on which order you use when taking a double limit. He reveals that he told the authors about the problem, but that they ignored him and left it in the paper.

5. Piketty and Zucman (2014): Romer points out the by now well-known "gross vs. net" problem in Piketty and Zucman's definition of savings.


All in all, this seems like a pretty loose collection of criticisms. Hand-wavey arguments, dubious definitions, bad assumptions, and math errors are all very different things. So this essay at first can seem like a grab-bag of gripes that Romer has with individual rivals' papers.

But I think Romer is on to something about the culture of econ theory, at least in the "macro"-ish realms of growth, business cycle, macro-labor, macro-trade, and macro-tax theory (I don't know nearly as much about the culture of the "micro"-ish fields like game theory, decision theory, I/O, etc.; and I know that finance theory has a very different culture). In these "macro"-ish fields, people seem to view math more as a tool for stylized description of ideas than as a tool for quantitative prediction of observables. 

Romer's examples of "mathiness" are all very recent examples. But going back to earlier models, I don't really see much more tight connection of variables to observables. Yes, in a Solow model you can tie capital K to observable things like structures and machines and vehicles. But you'll be left with a big residual, A. Then you can break A down and extract another term for human capital, H. Can you really measure human capital? Human capital can't be bought and sold on a market; you have to bundle it with other goods. So it's very difficult to get a clean measurement of the value of the existing stock of human capital, the way you could get a clean measurement of the existing stock of delivery trucks. Romer cites human capital as a good example of non-"mathiness", but I don't really see a huge difference between that and the "location" used by Prescott and McGrattan (2010). Maybe a minor difference, but not a huge one. As for the remaining A, there's not really any quantitative way to measure the stock of ideas except as a residual. And as for goofy assumptions, well, any growth model is going to have at least one or two assumptions that would make a newcomer to the econ field throw up her hands in disbelief.

Mathiness isn't anything new, it's just the way these econ fields work. The math is there as a storytelling aid (and possibly as a signal of intellectual ability). I think Karthik Athreya said it best:
My view is that a part of what we do is "organized storytelling, in which we use extremely systematic tools of data analysis and reasoning, sometimes along with more extra-economic means, to persuade others of the usefulness of our assumptions and, hence, of our conclusions...This is perhaps not how one might describe "hard sciences".
Do the guys Romer calls out play a little faster and looser with definitions and rely more on hand-wavey arguments? Oh, I'm sure they do - but that's because they're famous old guys. Writing down hand-wavey stuff is a privilege afforded to famous old guys in every academic discipline I know of. In econ, it gets politely published in top journals, but all the hotshot young people just sort of shake their heads anyway, and the only net effect is to pad out the length of the journals. Are the guys Romer calls out more political than the average economist? Maybe.

But in general, the whole discipline of macro theory - in the general sense, including growth and parts of labor, trade, and tax theory - is chock full of mathiness. Even most of the best models ("best" being a highly relative term, of course). The original Solow model seems to me like a rare exception, not a typical example of the Good Old Stuff.

But in any case, I highly recommend the Romer piece, which is a master class in catching errors in models, as well as a fascinating window into the Byzantine world of academic politics.


Update: Brad DeLong has a follow-up post explaining his view of some of the history behind the argument in the field of growth economics. Basically, the idea is that George Stigler didn't like people using models with imperfect competition, since this might open up a window for government intervention. DeLong thinks that Lucas and other "freshwater" types inherited this anti-imperfect-competition bias, causing them to be too down on Romer's models. This is interesting history that I didn't really know about before.

Update 2: Paul Romer has a response on his blog. Excerpts:
Noah Smith asks for more evidence that the theory in the McGrattan-Prescott paper that I cite is any worse than the theory I compare it to by Robert Solow and Gary Becker... 
There is no such thing as the perfect map. This does not mean that the incoherent scribbling of McGrattan and Prescott are on a par with the coherent, low-resolution Solow map that is so simple that all economists have memorized it. Nor with the Becker map that has become part of the everyday mental model of people inside and outside of economics... 
Noah’s jaded question–Is the theory of McGrattan-Prescott really any worse than the theory of Solow and Becker?–may be indicative of what many economists feel after years of being bullied by bad theory. And as I note in the paper, this resignation may be why empirically minded economists like Piketty and Zucman stay as far away from theory as possible... 
For specific purposes, some maps are better than others. Sometimes a subway map is better than a topographical map. Sometimes it is the other way around. Starting with any good map, we can always increase the resolution and add detail. 
No map is perfect, but this does not mean that all maps are equal. It certainly does not mean that an internally consistent map that with so little detail that you can memorize it is on a par with incoherent scribbling.
In the rest of the post, he goes into depth about why he thinks the McGrattan and Prescott paper constitutes "incoherent scribbling." But he also notes that the other papers he goes after in his "mathiness" piece should not be let off the hook:
Noah also notes that I go into more detail about the problems in the Lucas and Moll (2014) paper. Just to be clear, this is not because it is worse than the papers by McGrattan and Prescott or Boldrin and Levine. Honestly, I’d be hard pressed to say which is the worst. They all display the sloppy mixture of words and symbols that I’m calling mathiness. Each is awful in its own special way.
He also has another short post about a Lucas paper.

It appears that Prescott, Lucas, Levine, and others of the unofficial "freshwater" club have annoyed more high-level colleagues than just Paul Krugman. Only a few months ago, Roger Farmer took to his blog to unleash an anti-Prescott blast. Romer and Farmer are not politically-minded media-engaged types like Krugman and DeLong, but their aggravation with the Lucas/Prescott school is, if anything, even more intense.

And yes, Romer is right that I'm jaded. Is it so obvious? *takes swig from hip flask, rubs beard stubble*

Tuesday, May 12, 2015

Department of "Huh yourself!": British demand edition



Brad DeLong thinks I'm nutty when I say that Britain isn't obviously suffering from a persistent shortfall of aggregate demand:
Graph Gross Domestic Product by Expenditure in Constant Prices Total Gross Domestic Product for the United Kingdom© FRED St Louis Fed
There are no signs looking at wages and prices that this is due to any adverse supply shock...there seem to be no reasons looking at wages, prices, and output to believe that the British economy right now is a full-employment at-capacity-utilization economy. And only in such an economy would monetary and fiscal policies that boost spending simply boost prices and not production...
Graph Employment Rate Aged 15 64 All Persons for the United Kingdom© FRED St Louis Fed
Does the fact that the employment share of British adults is actually high mean that the British economy is, in fact, at potential output? That stimulative monetary and fiscal policies risk rising inflation for no gain? And that it is time to normalize? Certainly Mark Carney at the Bank of England does not believe that is so:
Graph Interest Rates Government Securities Treasury Bills for United Kingdom© FRED St Louis Fed
From my point of view, why so many Britons have taken so many low-pay low-productivity jobs in the past three years is a mystery. But that they have gives us little reason to think that the British economy is now a full-employment at-capacity-utilization economy in which aggregate demand is now equal to potential output.

First of all, let's get a couple things straight. I don't think British stimulus would be particularly counterproductive, I just don't think austerity would be either. If there's not a big demand gap, multipliers shouldn't be large, so stimulus just won't do a heck of a lot (unless the UK has falling-apart infrastructure like we do), but neither will austerity. Also I doubt AS-AD is even always the right model for the macroeconomy.

But with that said...

With a demand shortfall, we ought to see high unemployment. We also ought to see low inflation.

In the U.S. we saw both of these in 2009-2014. In Britain we saw the former in 2009-11, and we never really saw the latter at all. Here is British core inflation:


The Bank of England's inflation target is 2% (whether that's a ceiling or a target is not certain). But in 2010-2013 - four years!! - UK core inflation was above that target. 

So if we stick to the good old Econ 102 AS-AD model, and we look at both prices and quantities, we can come up with the following simple story for the UK:

1. In late 2008 the UK suffered a negative AD shock.

2. Around the same time, the UK suffered a negative AS shock.

3. In early 2010 the negative AD shock began to abate.

4. In 2012 the negative supply shock began to abate.

5. The differences between the UK and the U.S. in GDP, employment, and inflation can be explained by the fact that the UK's AD shock wasn't quite as big or long-lasting, while the U.S. didn't experience a supply shock.

This story also fits what we know about British TFP, which declined from 2007-2009, then flatlined through 2011:


This story is incredibly simplified, and uses a model that probably isn't the best. A real, careful, academic analysis of the situation is certainly warranted. But is there some obvious reason we need to go looking for a story where AD is the only thing moving around here? That story is going to have a lot more moving parts, and it's going to go a lot deeper into micro stuff. 

And it will also look like reaching. Why should we demand a story where demand is the whole story? Is this about stabilization policy, or about the size of the British state?

Economists as all-purpose sages: The case of Freakonomics


The original book Freakonomics, by Steve Levitt and Stephen Dubner, was a very fun read. But it also slightly annoyed me. Why? Because there's very little actual economics in it! The quantitative empirical work is mostly reduced-form regressions with natural experiments. That's a fine and good research technique, but it's not really special to econ - it doesn't include anything about market design, structural estimation of supply and demand, game theory, search, prices, general equilibrium...nada!

That covers three of the six chapters. The other three include 1) an ethnography of drug dealer culture by a sociologist, the excellent Sudhir Venkatesh, 2) a quick gloss of statistics techniques that applied mathematicians use to catch cheaters, and 3) a historical story about the decline of the KKK.

So this book has sociology, history, stats, and some general empirical techniques that could be used by any social scientist. That doesn't make it bad - most of the research the book showcases is really cool (though Levitt's own study, on abortion and crime, ended up having some serious problems). But it means that an empirical sociologist could easily taken Levitt's place as the technical co-author of the book, alongside journalist Dubner.

But it was an economist Dubner got, and Freakonomics was billed as a pop econ book, not a pop sociology book. Why? It seems to me that it's because economists are respected as all-purpose sages. Like I said in my previous post, economists get taken seriously on any topic imaginable.

To use an even more stark example, take the sequel, Superfreakonomics. There's a chapter in that book that's all about geoengineering. That's an engineering topic. A physics topic. A climate science topic. And yet an economist is put forth as an authority on whether it will work. And this is accepted by the book's legions of fans.

People trust economists on any topic.

Why? I don't know, to be honest. Maybe it's because economists are thought to have high IQ and know a lot of math relative to other social science disciplines. Maybe it's because economists are confident in their ability to model any social phenomenon, like Gary Becker and others of the "imperialist econ" school. Or maybe it's because economists are just more willing to engage with the public and hold forth on any topic. After all, op-ed writers are the other group who are treated as all-purpose experts; maybe economists just act like really hi-tech, super-smart op-ed writers.

Or maybe it's because of economists' reputation as being clear-eyed and impartial observers of society. For example, in the intro to Freakonomics, Dubner describes a scene with Steve Levitt:
An elderly homeless man approaches [Levitt's car]. It says he is homeless right on his sign, which also asks for money. He wears a torn jacket, too heavy for the warm day, and a grimy red baseball cap. 
[Levitt] doesn’t lock his doors or inch the car forward. Nor does he go scrounging for spare change. He just watches, as if through one-way glass. After a while, the homeless man moves along. 
“He had nice headphones,” says [Levitt], still watching in the rearview mirror. “Well, nicer than the ones I have. Otherwise, it doesn’t look like he has many assets.”
If you didn't know Levitt was an economist, this scene would just make him sound like a rich insensitive jerk. But the fact that he's an economist imbues the scene with a different meaning altogether - suddenly, Levitt's clinical detachment seems like a sign of impartiality and rationality. Exactly the qualities we'd want in an all-purpose sage.

Anyway, I don't know the answer. But the observation that people trust economists on any social topic - and even some engineering topics - seems pretty obvious. And pretty freaky.


P.S. - If you want a book that is nothing but hardcore economics, and explains everyday economic phenomena in a way that is humble, entertaining, and useful all at the same time, I recommend Tim Harford's The Undercover Economist. The best version is the audio version. 

Saturday, May 09, 2015

Economists don't have "physics envy"



I hear all the time that economists have "physics envy". This doesn't seem even remotely true. I'm not sure whether "physics envy" means that economists envy physicists, or that economists want to make physics-style theories, or that economists wish their theories worked as well as those of physicists. But none of these are true.

Reasons why economists don't have physics envy include:

1. Economists make a lot more money than physicists.

2. Economists are treated as experts on practically anything. An economist can talk about why hipsters have moustaches, and get taken seriously. An economist can talk about which restaurants are the best, and get taken seriously (Update: NO, I'm not saying Tyler's book is bad, I haven't even read it, so HUSH). An economist can talk about politics, marriage, popular music, sex, race, or sports and get taken as seriously as any expert in those fields. An economist can talk about how much progress physicists are likely to make, and get taken seriously. Physicists get taken seriously when they invent quantitative rules for things, but otherwise are treated as just one more tribe of crazy nerds with their heads in the aether.

3. Economic theorists, traditionally, have been free from the constraints of empirical validation. Econ didn't start out with any data to speak of, so it developed a culture where data wasn't the measure of a good theory. That culture has been slowly changing, as IT and statistics allow us to do much more empirics. The wild econ theorist is being slowly tamed, though they occasionally buck against their new constraints. But economics is still nowhere near as enslaved to empirics as physics is. Ed Witten, a brilliant physicist who invented superstring theory and won a Fields Medal, will probably never get a Nobel prize, since no one has yet figured out a way to test superstring theory. In econ, though, un-validated theories - even empirically unsuccessful theories - do sometimes get Nobel prizes, especially in macro.

4. Economists and physicists have totally different reasons for thinking their theories are beautiful. Physicists tend to appreciate the symmetry of their theories, or their connection to geometry. Economists tend to appreciate the fact that their theories can be derived from axioms of human behavior. Economists don't usually see themselves as a high-up floor on the tower of science that begins with math and progresses through physics to chem, then bio, then psych. They tend to see their own field as part of an entirely different tower, unsupported by the other sciences, built on the foundation of axioms - not empirical laws or derived properties - of human behavior.

(Note for physics/math people: what economists call "axioms" are what physicists call "postulates".)

5. Economics theories aren't really much like physics theories in the first place. In particular, the word "equilibrium" means totally different things in the two fields. In econ, the word "equilibrium" is incredibly general - it just means the solution of any system of equations in an economics model, really. A few economic equilibria are similar to equilibria in physics models, but not most. A lot of people outside econ don't seem to understand how economists use the word.


So economists don't have physics envy. But there is a related field that absolutely does have physics envy: Financial engineering.

Financial engineering doesn't use the axiomatic stuff econ uses - instead it fits curves, like applied math. Often, the math is very similar to that used by physicists, and this is why physicists often go into financial engineering. But financial engineering doesn't work nearly as well as physics, which definitely leads financial engineers to wish that it did work as well.

So if you want to tease anyone for having "physics envy," tease financial engineers, not economists. The only way economists are ever going to envy physicists is if they find out how much physicists...well, never mind.

Friday, May 08, 2015

"Signaling" isn't about signaling


Robin Hanson responds to my Bloomberg View post about signaling.
More generally I call a message “signaling” if it has these features:
  1. It is not sent mainly via the literal meanings of words said.
  2. It is not easily or soon verifiable.
  3. It is mainly about the senders’ personal features, perhaps via association with groups.
  4. It is about sender “quality” dimensions where more is better, so senders want others to believe quality is as high as possible, while others want to assess more accurately. Such qualities are not just unitary, but can include degrees of loyalty to particular allies.
Cheap talk cannot send a message like this; one cannot just say such a thing, one must show it. And since it cannot be verified, one must show it indirectly, via how such features make one more willing or able to do something. And since willingness and ability track costs, these are “costly” signals.
This seems to add a few arbitrary restrictions to the set of things we might try to describe with a Spence-style signaling model. That's perfectly OK (especially since I was complaining about the overuse of the term, not the underuse!). But I think it also makes things a bit more complex than they need to be.

Let's focus on what I think are the three key characteristics of signaling models:

1. The signal must be costly to send for all types.

2. Different types must have different costs to sending the signal.

3. If you take away the cost differential there will be a pooling equilibrium.

The first condition assures us that the asymmetric information imposes costs on the economy; without (1), the model just becomes a truth-telling mechanism. Without (1), you just say which type you are, pay 0 cost, and society knows you're telling the truth.

The second condition creates the separating equilibrium. It means that signaling "works", in terms of revealing people's true types.

The third condition is why you need signaling in the first place. The separating equilibrium isn't first-best (that's guaranteed by condition 1), but it's constrained-optimal. The pooling equilibrium - the thing you avoid by creating the cost differential in (2) - is even worse than the separating equilibrium.

Robin thinks that my example of the hipster moustache is signaling, and explains why:
Let’s distinguish three different kinds of messages I might send with my waxed moustache:  
1) “I have thick shiney (sic) hair.”... 
2) “Hipster is one of my interest areas.”...Technically, this is a “cheap talk” message. 
3) “I am especially devoted to the hipster ethos” or “I especially embody hipster ideals.” That is, I am especially willing to identify myself as a hipster, and my personal features are an especially high quality match to ideal hipster features, including having a creative and contrarian yet attractive and coherent personal style that fits with current hipster fashions. These messages are hard to verify, and the interests of observers and I conflict. While observers want to accurately rank me relative to others, I may want them to estimate me as having maximal devotion and quality. Since verification and cheap talk won’t work here, I have to show, not just say, my messages. 
To show my hipster devotion, I can choose an appearance that is sufficiently off-putting to many people’s work, home, church, etc. associates. By paying the cost of putting off possible associates, I show my devotion to hipsterism. To show my hipster features, I can pay to track hipster fashions and to continually search in the space of possible appearances for a combination that simultaneously reflects current fashions while being creative, coherent, and showing off my best personal features. Not being a hipster, I don’t know how exactly that works for them. But I do know, for example, that since lipstick and tight clothes make some bodies look better while making other bodies look worse, they are costly signals of the quality of lips and body shape. There must be similar factors for showing off hipster qualities.
#3 is why Robin thinks of hipster moustaches as "signaling."

I think it's obvious that hipster moustaches satisfy my condition 2 from above - hipsters definitely pay a lower cost for having hipster moustaches, for a variety of reasons (most importantly, they like the style more than others do).

But do they satisfy my condition 1? Do hipsters pay a net cost for their moustaches? Robin suggests that they do, because the moustaches are "off-putting" to people at work, church, etc. But I doubt that this is, in fact, true. Do hipsters look in the mirror and think "Dang, I wish I didn't have to grow this stupid moustache just to prove I'm a real hipster"? I doubt it. I bet they intrinsically enjoy having the moustaches. This is very different from the experience of someone who has to work hard to get some pointless credential just so he can get a job - i.e., the situation Spence originally suggested.

I also doubt that the moustaches are particularly off-putting to most of the people hipsters want to associate with. Sure, people in some Baptist church in Alabama would be put off if I walked in with a big ol' moustache. But I bet hipsters have no desire to actually go hang out in that Baptist church. Also, I bet they tend to have jobs with people who aren't offended or repulsed by hipster moustaches, because I bet they would really hate the kind of workplace environment where people are repulsed by moustaches. So the mere fact that some people are repulsed by hipster moustaches doesn't mean that hipsters actually pay a cost from that repulsion. In fact, offending those people may give hipsters pleasure.

Next: do hipster moustaches satisfy my condition 3 from above? Are there a bunch of wannabe hipsters who would love to pass themselves off as hipsters if they just didn't have to grow that damn moustache? I doubt it. What non-hipster wants to be a hipster? Maybe a few. Maybe there's some guy out there who really has a thing for hipster girls, and who thinks that they only date hipster guys (False, btw!). But I bet these wannabes are few enough in number that distinguishing themselves from the wannabes is not a huge concern for the hardcore hipsters. (I could be wrong; perhaps America gazes upon envy at the hipster community, bitterly wishing it could be part of the fun!)

Anyway, it's possible that hipster moustaches satisfy all 3 conditions of a true signaling model, but I doubt it. More likely it's just a truth-telling equilibrium.

It also seems likely that much of the fad for labeling non-signaling mechanisms "signaling" is just a nerdy way of insulting countercultures and subcultures. It's a fancy way of saying "Hey hippie, get a haircut!" - of attempting to enforce social conformity by accusing standouts of inauthenticity. But it doesn't really seem to get the economics right, except maybe in a few scattered situations.


(P.S. - If you want mathematical models of conformity, nonconformity, subcultures, hipsters, etc., I would recommend this or this.)

Saturday, May 02, 2015

A quick history of 4chan and the rightists who killed it (guest post)


A lot of the cool funny internet stuff we know and love - LOLcats, Doge, etc. - came from a site called 4chan. But recently, 4chan has become a gathering place for online rightist movements like GamerGate and various racists. How did this come to pass?

This is a guest post by a 4chan user and Portland resident who goes by @animemoemoney on Twitter. He was also an early GamerGate supporter, and even organized one of the first real-life GamerGate meetups, before later repudiating the movement. Here, he tells the history of 4chan and the invasion of un-funny, uncreative anonymous rightists that eventually changed the nature of the site. 

________

Anonymous Rightists and Their Slow Resurgence on the Internet

At one time, the internet used to be a smaller place. Large social, content hosting websites like Facebook and Twitter now rival large countries in population size. It’s what people do on the internet—browse social media. It’s rare to be out in public and not see a person flip through their mobile device and maybe check their email. The internet is huge now and a ton of people use it.

Before there were advances in consumer electronics, and other forms of utility-increasing technological devices that allowed users easier access to the internet, there were enthusiast forums. Enthusiast forums are not like social media. They behave the same, but there was a far different community dynamic seen in forums of the past. Enthusiast forums are where gamers, anime watchers, or plain jokesters gathered to discuss their hobbies. An enthusiast was more likely to use a slower, now outdated version of the internet as well. 

Enthusiast forums still exist today, and they still have that handmade look of a website built by two-to-three people. The biggest differences between social media websites and enthusiast forums would be the culture and the board software. Instead of putting their real name and uploading a picture of their face, enthusiasts have taken a persona with a custom username and avatar. Enthusiasts loved to take a different persona because it made them anonymous on the internet, and thereby not be judged by their hobbies in real life.  Because of the anonymous nature of enthusiast forums, it meant users’ online behavior wasn’t judged either. 

When Christopher “Moot” Poole created 4chan, he sought to emulate the anonymous behavior of the Japanese website 2channel. Moot was a large browser of the website Something Awful, a web forum dedicated to making jokes and creating original content. Moot would bring Something Awful’s character to his soon to be popular website. He wanted 4chan to be more anonymous than the enthusiast forums of the past, by stripping away usernames and user accounts. 4chan was a great success, and was soon to represent everything Poole loved—anonymity, games, anime, and original content.  

Before 4chan, forums had a lot of rules. The board software would only allow you to do so much, such as make a thread or make a post, but there were also rules made by the moderation team. Users would be expected to obey the rules enforced by the moderation team if they didn't want to be suspended or banned. It was a particularly sad sight when a member of a community was banned forever due to racism, or trolling.  

Unlike forums before it, 4chan didn’t set (many) limits to what was considered acceptable on the website. Forums like Neogaf.com have a very strict expectation for rule following, and will ban you for subtle things like posting in an incorrect tone. The lax rules of 4chan attracted a lot of content producers, and people who loved to post without the fear of gaining a bad reputation, or being banned for behaving a certain way. On 4chan you can be banned for being off-topic, but not for making racist posts. 

There probably was never a moment were racism couldn’t be found on 4chan, but the trend of making racist posts shifted drastically from being an ironic, edgy joke to a very serious habit for a percentage of the posters. It started mostly with the news board “/n/,“ which was deleted for being too packed with race discussion. While it used to be a great place to discuss economics and politics, the board was slowly consumed by people trying to “red-pill” [i.e., convert to rightism] as many people as possible. These “red-pill” posters probably were channers, but they were a vocal minority at that time. After /n/ was deleted, posters didn’t know where to take their race discussion. They went to the firearms board, the anime board, and the video games board but were unwelcome there.
Later on, Moot decided to remake the news board as /new/, and the same process reoccurred. Some people would make threads linking to news articles, and it would be followed by a couple posters presenting the racial truth behind the story; they’d post race statistics, homicide  statistics, and ACT score graphs. Though predictable, these posters where still channers…and therefore jokesters and trolls, so there would be fun discussion even if the content was offensive. 
As time went on, /new/ got more and more radical. The free speech board attracted radicals from all over the internet. But most prominent of the radicals were the hateful plain-text rightists that would soon to dominate. Every thread was filled with racial discussion by that time. The hateful plain-text rightists did not seem to be channers—their posting style was different, they wouldn’t use images, and they wouldn’t recognize the channer language. When a hateful plain-text rightist would be called a “newfag” for example, he would claim he was very against homosexuality. Because of these plain-text hateful rightists, those who didn’t post images or news, the discussion on /new/ became very, very far right. 

Moot would soon delete /new/, probably for the same reason he deleted /n/. An opportunistic man who went by SaveTheInternet would create a website for /new/’s old user base to post at.  That website would be called 4chon, and it was basically just the /new/ forum, with some other boards for community support. People did make news threads at this forum, but it was crowded out with race discussion, conspiracy theory culture, and a very confused love for the Mises Institute. Unfortunately, the chan culture seemed to leave these posters, and they would become hateful plain-text rightists themselves. Anime posting was soon discouraged, and much of what made a 4channer in the past would be considered “degenerate.” It was very sad to watch channers take in the judgmental attitude of a mysterious group of anonymous rightist invaders and use it to dissolve their own culture. . . which is what made their image-board so special.

Moot recreated the news board as the /pol/ “Politically Incorrect” board, making it the third board where rightists gathered on 4chan. In /pol/, you weren’t expected to have to post news, just whatever politically incorrect ideas that you had at the time. This new board would slowly kill alternative boards like 4chon, which relied heavily on defected 4chan posters. Moot would claim in his 7 hour departure Q&A that creating sites like 4chon, based solely on hatred of other people or other internet forums, “didn’t really create a lot of great content.” The activity of /pol/ would soon rival the activity of the anime forum. In the word-of-mouth between posters, there was a joke that /pol/ was a containment board and just a place to keep anonymous plain-text rightist posters off of the other 4chan boards. 

Despite most of the rightist discussion being limited on /pol/, there was a website-wide rightening of 4chan culture. Memes were mostly critical of different races, and even the most progressive parts of 4chan would bemoan the identity politics of Tumblr. When dislike of identity politics reached its height, the video games forum would launch an anti-left politics movement called GamerGate to demand a reform in game journalism ethics. While GamerGate started off as a very diverse, vocal opponent to what they saw was unethical journalism (before it was debunked), many of the anonymous /pol/ rightists would take advantage of its anti-left character by creating sock-puppets. Unfortunately, activities like GamerGate would prove to be too much, and cause Moot to step down from 4chan’s admin position. 
Today it is hard to find a 4chan user that doesn’t have an attachment to far right politics. By being anonymous, rightists took advantage of their lack of identity to spread a hateful world-view. Some 4channers bought the right wing philosophy completely, others accept only some of it. Many only spread the rightist memes as a joke.

Wednesday, April 29, 2015

A follow-up on college and signaling



The other day I wrote a Bloomberg post about the fad of describing all communication and information extraction as "signaling" even when a Spence-style signaling model doesn't apply.

Some people have been telling me that this Garett Jones tweet is a reply to my piece. Garett writes:
Emoteconomist: One who is sure competition would eliminate inefficient diploma discrimination, but not inefficient gender discrimination.
Personally, I would apply the term "emoteconomist" to a much wider array of economists, but that's beside the point. I highly doubt this tweet is aimed at me or my piece. Garett knows that I have often written in support of Gary Becker's idea that economic competition decreases inefficient gender discrimination. He also of course knows the difference between preference-based discrimination (as in a Becker model) and signaling (as in a Spence model). So Garett's tweet is almost certainly not directed at my piece.

However, this Bryan Caplan post most certainly is a reply to my piece. At first I intended to rest my case, but Bryan has decided to delay our Bloggingheads debate two years (to coincide with the planned release of his book on the topic), so I might as well make some short reply.

Bryan does agree with a few of my points. For example, he agrees that college is unlikely to be important as an intelligence signal. But he disagrees strongly with my contention that college is not needed as a signal of working ability and temperament. In my piece I wrote:
So is college a way to signal conscientiousness and willingness to work? Maybe. But an even better way to signal that would be to actually work at a job for four years. One would think that if young people needed to do some hard work to signal their work ethics, some companies would spring up that gave young people real productive work to do, and provided evidence of their performance. Instead of paying through the nose to send a signal of your industriousness, you could get paid. But we don't see this happening.
Bryan replies:
Like most economists, Noah needs to be more sociological.  In a cultural vacuum, working four years might be a great signal of work ethic.  But no human being lives in a cultural vacuum.  We live in societies thick with norms and expectations. And in our society, people with strong work ethics go to college and people with bad work ethics don't.   
Disagree?  Just picture how your parents would react if you told them, "I'm not going to college.  I'm just going to get a job."  In our society, your parents definitely wouldn't respond, "That makes sense, because you're such a hard worker."  Why not?  Because in our society, most hard-workers choose college.  If a hard-working kid refuses to copy their behavior, people - including employers - understandably treat him as if he's lazy.  Because lazy is how he looks.
Actually, when I think about the college-as-signaling hypothesis, I do often think sociologically. But, as so often, the society I think of is not the United States - it's Japan. In Japan, it is taken as a given that college students don't work hard at their studies. College is even nicknamed "moratorium". Japanese college kids are expected to enjoy themselves and not work hard - in fact, when I ask Japanese young people why they don't consider going to America for college, they usually tell me that American students work too hard. And yet, top Japanese employers all require a college education (usually a Japanese college education) as a precondition for hiring.

But to be honest, Bryan is right that I don't think very sociologically. I don't really know much about sociology. Does he? Perhaps we should call in a sociologist. I will do so on Twitter.

Anyway, Bryan then writes:
Noah overlooks another key trait that education signals: sheer conformity to social norms.  In our society, you're supposed to go to college, and you're supposed to finish.  If you don't, the labor market sensibly questions your willingness to be a submissive worker bee. 
I agree that college, in America and also in Japan, is a hallowed cultural institution, and that there is a lot of social pressure on people to do it. But this seems like part of college's consumption value, not its value as a costly, Spence-style signal.

Next, Bryan quotes this part of my post:
There are many other reasons to doubt the signaling theory of college. A more likely explanation for college's enduring importance is that it provides a large number of benefits that are very hard to measure -- building social networks, broadening people's perspective, giving young people practice learning difficult new mental tasks and so forth. 
He replies:
I'm glad to hear this.  Noah inadvertently grants one of my key points: Most of education's labor market payoff is unrelated to the material your professors explicitly teach you.  Once you accept this heresy, you're stuck with some combination of my multidimensional signaling story, and Noah's amorphous, evasive "large number of benefits that are very hard to measure" story.  If that's the choice, my story will end up with the lions' share of the mix.  Noah is welcome to the leftovers.
Ah, but wait! College most certainly does provide some direct and obvious skill-based human capital benefits: reading, writing, working in groups, communicating, arguing, doing math, programming computers, etc. My point about non-obvious forms of human capital - human networks, cognitive broadening, emotional growth, exposure to new career ideas, sexual maturity and marriage - was in addition to the obvious benefits of coursework and instruction. And a third big chunk of college's value is consumption, which Bryan basically ignores.

After those three big bites, it is Bryan's "conformity signaling" that is left to hunt for the table scraps!

Finally, Bryan mentions the "sheepskin effect":
Final challenge for Noah: If education's rewards stem from this "large number of benefits that are very hard to measure," why on earth would the payoff for graduation vastly exceed the payoff for a typical year of education?  My explanation, of course, is that given the vast social pressure to cross educational milestones, failure to graduate sends a very negative signal to the labor market, leading to discontinuous rewards.  What's Noah's alternative?  Do schools really delay "building social networks, broadening people's perspective, giving young people practice learning difficult new mental tasks and so forth" to senior year?
I have no ready explanation of sheepskin effects - perhaps they are used by employers to extract a signal of how well one actually learned things in one's college courses. But signal extraction does not imply Spence-style signaling. Spence-style signaling must be costly, and for students who have done enough to graduate, collecting that sheepskin is simply not costly.

So I don't need to explain the sheepskin effect in order to rule out Bryan's explanation. Bryan views the sheepskin effect as evidence of signaling, but since it implies that much of the college payoff comes without cost, I view it as clear evidence against the signaling model of college.

Anyway, I think that about takes care of Bryan's points. As a final note, Bryan wants me to be more sociological, but I think he should be more psychological! If college really is wasteful, costly signaling, as Bryan posits, then people who complete it should view it as a wasteful, unnecessary chore. It should be something they wish they didn't have to do. But I bet a substantial majority of college graduates, if you ask them, will speak quite highly of their time in college, and will not wish that they had been able to go directly into the workforce instead.


P.S. - If you don't understand that signal extraction does not imply signaling, just contemplate the following sentence: "Fire doesn't emit smoke in order to prove to observers that it's really a fire."

Sunday, April 26, 2015

Guns don't kill people. Labor kills people.



Arthur Chu, Jeopardy champ extraordinaire, tweets:
"Capitalism made your iPhone" 
No, LABOR made your iPhone. Labor makes things under any -ism. The -isms just determine who gets paid
He's right that "-isms", in econ terms, are about distribution of resources (though he should broaden his definition of resources to include control, not just payment).

But is he right that "labor made your iPhone"?

Consider the following two situations:
A) I make fire by rubbing two sticks together.
B) I make fire by using a butane lighter.

In both of these situations, you can say "labor made the fire". But in the first situation, there was a lot more labor for the same amount of fire. Saying only that "labor made the fire" leaves out this important fact.

Now suppose I want to make fire with no tools. No matter how much labor I apply - the labor of millions of people over millions of years - I will not be able to make fire. 

So saying that "labor makes fire" also leaves out this important point - the necessity of having tools.

Labor is a necessary input into producing an iPhone. But there are other necessary inputs - machines, buildings, land, natural resources, vehicles, tools, etc. And labor is not a sufficient input for making an iPhone - without the right tools and the right organizational system, no amount of labor will get the job done.

But didn't labor "make" the machines, buildings, etc.? Since labor is necessary to create any intentionally produced good, you can say "Labor is what makes everything" if you want to. But you know what else is necessary to create those goods? Electromagnetism, gravity, and the strong and weak nuclear forces. So I could reply to Arthur Chu by saying "Labor didn't make your iPhone. Physics made your iPhone." Now who's right?

The basic point here is that our language, and our intuitive way of thinking about causation, views everything as perfect substitutes. A + B = C. If A + B = C, then you can determine how much of C is due to A, and how much is due to B.

But in reality, things are only partial substitutes. You more often have stuff like
(A^a)(B^b) = C. When you have complementarity, it doesn't make sense to ask how much of C is due to A, and how much is due to B. But we always do it anyway. Arthur Chu's tweet is one example. The slogan "Guns don't kill people, people kill people" is another example. A third example is the perennial debate over whether humans have "free will."

Our intuitive concept of causal attribution is simply wrong and useless in most cases.

Wednesday, April 15, 2015

Steve Williamson is right that I am confused


In a post for Bloomberg View, I wrote about the history of the sticky-price revolution of the late 1990s and early 2000s:
In 1994, economists Greg Mankiw and Lawrence Ball wrote an essay for the National Bureau of Economic Research entitled “A Sticky-Price Manifesto.”...[T]he essay heralded the beginning of a macroeconomics mini-revolution. It was a direct threat to the line of research that had been dominant in the 1980s, which tried to explain recessions without sticky prices... 
The economic establishment reacted harshly to the upstarts. “Why do I have to read this?" fumed Robert Lucas, the dean of macroeconomics. "This paper contributes nothing.” He went on to accuse the sticky-pricers of being opposed to science and progress. 
But Lucas fumed in vain. During the following decade, the sticky-price models went from strength to strength. New math was developed to make them easier to use. Possible reasons for price stickiness were investigated -- for example, “menu costs,” in which the seemingly trivial costs of changing prices add up to a big problem across the broader economy. 
Even more telling, sticky-price theorists proved that you didn’t need a lot of price stickiness to mess up the smooth working of the economy. Even the tiniest dash of stickiness would turn all kinds of theories on their heads. Economists Susanto Basu, John Fernald, and my doctoral adviser Miles Kimball, for example, showed that when prices are even a little sticky, bursts of technological progress actually hurt the economy for a short while, by causing a burst of deflation, before eventually boosting growth. Over time, the addition of various other economic mechanisms, like labor search, has further reduced the amount of price stickiness required to cause major recessions. 
Sticky-price models have become the dominant models used at central banks. The smoothly adjusting, flexible-price models of the 1980s are basically not used anywhere, by anyone, for anything. 
Even some of the biggest skeptics of sticky prices are coming around. In 2004, economists Mark Bils and Peter Klenow looked at how businesses changed prices, and found that the changes were too frequent to be consistent with the sticky-price story. But in 2014, they reversed their stance, looking at evidence on the adjustment of markets in recessions and concluding that “sticky prices...deserve a central place in business cycle research.” Meanwhile...Patrick Kehoe...long-time [opponent] of the mainstream sticky-price models, nevertheless wrote a paper in 2010 entitled “Prices are Sticky After All.”... 
The moral of the story is that if you just keep pounding away with theory and evidence, even the toughest orthodoxy in a mean, confrontational field like macroeconomics will eventually have to give you some respect.
Steve Williamson wrote a response to my post, and for the life of me I can't tell what he's trying to say. He calls me "confused". Well, after reading his post, I am confused.

Williamson takes some potshots at Ball and Mankiw:
The "Sticky Price Manifesto" is in part a survey of the menu cost literature, but it reads like a religious polemic... 
Why should we care what Ball and Mankiw think is going on in the minds of their staw-men opponents, or in the classrooms of those straw-men? Why should we care what Ball and Mankiw "believe?"... 
Noah seems to think that Lucas was being unduly harsh [in his response to Ball and Mankiw], and that he was somehow feeling threatened by these "upstarts." It's pretty clear, actually, that Lucas just thinks it's a bad paper - religion, not science - and that Ball and Mankiw could do a lot better...
He then asserts that New Keynesian models don't have anything to do with the stuff Ball and Mankiw were writing about:
Noah is more than a little confused about the genesis of sticky-price New Keynesian (NK) models. In particular, he thinks that Ball and Mankiw's "Sticky Price Manifesto" was a watershed in the NK revolution. Far from it... 
Where did NK come from? Which of the three threads in post-macro revolution Keynesian economics - coordination failures, sunspots, menu costs - morphs into Woodfordian NK models? To a first approximation, none of them. Perhaps NK owes a little to the menu cost approach, but it's really a direct offshoot of real business cycle theory. Take a Kydland and Prescott (1982) RBC model, eliminate some bells and whistles, add Dixit-Stiglitz monopolistic competition, and you have Rotemberg and Woodford's chapter from "Frontiers of Business Cycle Research." Add some price stickiness, and you have NK. So, NK basically leapfrogs most of the "Keynesian" literature from the 1980s. It's much more about RBC than about Ball and Mankiw.
(For a brief intro to Mankiw's contribution to the New Keynesian research program, see the Wikipedia page for New Keynesian economics. See also the Wikipedia page for Steve Williamson.)

Williamson then tries to claim ownership of New Keynesian models for Chicago/Robert Lucas/RBC/His Majesty the King of Spain/I'm not sure:
[I]t's worth noting that Mike Woodford, the key player in NK macro, was at the University of Chicago from 1986 to 1992, the latter 3 years in the Department of Economics with - guess who - Bob Lucas. Indeed, they wrote a paper together. It's about - guess what - a kind of sticky price model with non-neutralities of money. Later on, Lucas wrote about sticky prices with Mike Golosov. So, I think we could make the case that the influence of Lucas on NK is huge, and that of Ball and Mankiw is tiny.
He then goes off on a long tangent about how central bankers might use sticky-price models to think about financial stability (which, apparently, he thinks is now the main priority for central banks).

It's kind of funny to see Williamson trying to wrest historical credit for New Keynesian models from Mankiw & co., since just in his previous post he had this to say:
Mike Woodford can correct me on this, but my impression is that he came out of graduate school with a specific goal in mind, which was creating a version of Keynesian economics that would fit into modern macro. Ed Prescott's project left central bankers scratching their heads about what they were supposed to be doing, and Woodford and others stepped into the void. Interest and Prices is, I think, intended as a handbook for central bankers. There was a lot of effort put into marketing the whole NK project to the world's central banks. This is ongoing, and has been institutionalized[.]
So NK was reverse-engineering of Keynesian ideas. But actually it was just RBC. But it succeeded because it was promoted via a slick marketing campaign. But actually Lucas was one of its founders.

Also, microfoundations are important. But Mankiw's efforts to microfound sticky prices with menu costs was totally unimportant to the creation of sticky-price macro models.

Got that?

Damn, I guess I am confused.

Tuesday, April 14, 2015

Did macro theory fail us in the crisis?



David Andolfatto and Mark Thoma have posts defending macro theory from (some of) the people who say it failed us in the crisis. Both are good posts, and you should read both.

Anyway, here's a quick (and probably incomplete) taxonomy of criticisms people make about macro with regards to the crisis:

1. "Macro models failed to predict the crisis, therefore DSGE sucks."

This is the criticism that Andolfatto and Thoma reject. I basically agree. There are no other models out there that did forecast the crisis. Nor are expert predictions any better.

Personally I think DSGE techniques haven't  reaped dramatic benefits (yet). But what other alternative is better? When I ask angry "heterodox" people "what better alternative models are there?", they usually either mention some models but fail to provide links and then quickly change the subject, or they link me to reports that are basically just chartblogging. Yeah, sure, if you put out hand-wavey reports saying "capitalism sux, there's gonna be a crash!" every year or two, you're eventually going to be able to say "see, I told you so". But that's no replacement for real modeling.


2. "Macroeconomists were too confident before the crisis, and that gave policymakers false confidence."

It is pretty obvious to anyone who has ever interacted with macroeconomists that most of them take their models way too seriously (this is even more true of the "heterodox", to the degree that they even have models). It's a common disease of academics in general - you have to spend so much effort pushing your theories that overconfidence is selected for.

Did this confidence leak over into the policymaking sphere? I don't have evidence here, but I doubt it. Most of the Fed people are a LOT less confident than academics. And they were being advised by a lot more than just the academic crowd - they had a big stable of chartbloggers, hand-wavers, etc. to draw upon. Plus they themselves had Old Keynesian models in their bag of tricks. As for politicians, it's not clear they even know that academic macroeconomics exists.

If the Fed people were overconfident in 2005-6, I suspect it was mostly due to natural cognitive biases - "Everything seems like it's been going OK for a couple decades, I guess we're doing something right" - rather than the overconfidence of the academics they interacted with.

But I could be wrong.


3. "Macroeconomists weren't focusing on finance enough before the crisis."

Thoma says that this is a valid criticism, and I agree. There are a bazillion models out there. But just having models out there isn't enough; if you're going to give policymakers real advice, you're going to have to choose which model - or which basket of models - to base your advice on. Macroeconomists weren't very worried about finance before the crisis - you didn't see a lot of people waving copies of Geanakoplos (2003) and saying we could be courting disaster.

Belief in the Great Moderation, and in the Fed's ability to stabilize the economy, was too strong. The central problem of depression prevention had not, in fact, been solved. But an awful lot of top macroeconomists - not just Lucas, but the New Keynesians too - thought it had. Their favorite models didn't have any finance in them, with the possible lone exception of the Bernanke-Gertler "financial accelerator" models.

That was a big mistake, especially since the Great Depression and crises in other countries (e.g. Japan) should have suggested that financial crashes were a big deal. To their credit, though, mainstream macroeconomists have been hastening to correct the mistake. Certainly they're going to pay more attention to finance for at least a few more decades.


4. "Macroeconomists don't do enough to kill their models off."

This is something I hear surprisingly few people say, given that I think it's the best of the criticisms out there. If you let a million flowers bloom but don't cut any of the flowers, you get a big warehouse full of flowers. OK, so that metaphor went nowhere, but you get the point. Macroeconomists, when they get defensive, tend to say something along the lines of "We got models for everythin'!" But is that a good thing??

I feel like if you have models for everything, you don't actually have any models at all. Without a way of choosing between models, your near-infinite stable of models turns into one big giant mega-model that can give anyone any results he wants. Worried about a financial crisis? Pull out a model that tells you a financial crisis could be looming. Worried about inflation? Pull out a model where inflation is a big danger. And so on.

Now, technically, you could choose between models based on the plausibility of the assumptions. But three things make this impossible in practice. First, the need for tractability means that the assumptions in almost any modern macro model will be utterly implausible to anyone who has not spent decades in a monastery high in the Himalayas training himself in the art of self-deception. Second, the assumptions are so stylized that it takes a huge amount of talent just to figure out what they are - in fact, we're starting to see the emergence of top macro people, like Matt Rognlie, who specialize in figuring out what the heck models are actually saying. And third, with a near-infinite catalog of models to comb through, there's just no way to compare any significant number of them all at once.

If you ever want macro models to actually be useful, it's not enough to just wave your hands and say "all models are wrong". It's not enough to treat models as ways to "organize our thinking". You've got to have a way to take them to data and decide if you should keep them around, send them back to the shop for alterations, or burn them in a fire.


5. "The crisis exposed the fact that macroeconomics doesn't work."

Well, sure. But it also showed that we need to keep trying to make it work. And macroeconomists, as a whole, don't absorb a significant fraction of our GDP, so I'm not incredibly worried.



Updates

Bumped from the comments, by an anonymous commenter:
[D]on't focus on macro *theory*. It's macro empirics I'd worry about. theory ahead of measurement etc. The difference post crisis is you see greater prominence of macro papers using micro data (e.g. mian sufi or autor/dorn/hanson/acemoglu).
Great point.

Thursday, April 09, 2015

Fixed mindsets


Carol Dweck is one of America's most important public intellectuals, and I believe that her idea, the Growth Mindset, is one of the most good and useful in America today. She's also a hedgehog. Meaning, of course, that she applies her One Big Idea to everything and anything, and tends to exaggerate its power and the evidence in favor of it. That's OK. Most promoters of big good ideas are hedgehogs. There are plenty of hedgehogs in the econ blogosphere - Scott Sumner, Paul Krugman, Steve Williamson. It's no big deal. I think that as a society we've gotten good at recognizing hedgehogs and mentally correcting for the hedgehogginess.

So because Dweck is a hedgehog, I won't totally lambaste Scott Alexander for attacking a straw man in his latest 10,000-word essay. Part of Scott's confusion and misdirection comes from Dweck's over-selling of her idea.

But only part. The rest is derp.

Scott gives a pretty good definition of the growth mindset idea:
[G]rowth mindset is the belief that people who believe ability doesn’t matter and only effort determines success are more resilient, skillful, hard-working, perseverant in the face of failure, and better-in-a-bunch-of-other-ways than people who emphasize the importance of ability. Therefore, we can make everyone better off by telling them ability doesn’t matter and only hard work does. 
This is similar to how Dweck herself would describe the idea. However, Scott does not differentiate between the following two statements:

1. Natural ability doesn't matter; only effort matters.

2. Natural ability doesn't matter on the margin, but the marginal effect of effort is large.

I guess economics just makes me automatically think about whether an effect is an average effect o a marginal effect (or an average marginal effect, if we're running regressions!).

Now, in her book, Dweck makes it clear many times that natural ability does, in fact, matter. She states that among people of similar natural ability, having a growth mindset makes a big difference. What she is saying is that the marginal effect of the growth mindset on performance is large for most people, even though natural ability matters a lot in the average.

But what Dweck does not clarify is whether the optimal growth mindset is itself a belief about average or marginal effects of effort on performance! In other words, to get the effect, do I have to believe that natural ability doesn't matter? Or does it work if I believe that although natural ability matters, effort can have a big effect?

If it's the former, we have trouble. If the only way to get the positive effect of the growth mindset is to convince people that natural ability doesn't matter - i.e., a fiction - then eventually they will realize that the fiction is a fiction, and the growth mindset will go away.

But I don't think that's what the growth mindset is. I think that very few if any people, including the people Dweck describes as having strong growth mindsets, actually do believe that effort could make them as good a basketball player as Michael Jordan, as good a mathematician as Terence Tao, or as good a blogger as Noah Smith (cue hearty laughter).

I am betting that the growth mindset is a belief about marginal effects ("if I work hard I will get better") rather than average effects ("anyone can do anything if they work hard"). Scott Alexander, on the other hand, seems convinced that it is the latter.

One reason I would make that bet is the nature of the experiments that have tested the power of the growth mindset. Scott runs through a couple of the experiments in his post. Basically, these are priming experiments. It would be incredibly weird if a priming task could change people's bedrock beliefs.

Scott sees this and concludes (paraphrasing): "Since the growth mindset is a belief about average effects, and a priming task couldn't change people's beliefs about something as big as average effects, there must be something wrong with this research, and therefore with Dweck's thesis."

I look at the same facts and conclude: "Since a priming task can't change beliefs about something as big as average effects, the growth mindset must be a belief in marginal effects."

I find it highly likely that a priming task could change someone's belief about marginal effects. Marginal effects are local, average effects are global. Going from thinking "Trying harder won't help" to "Trying harder will help" is a lot easier than going from thinking "Natural ability matters" to "Natural ability doesn't matter at all."

(I'm going to stop explaining the difference between average and marginal now, since if you haven't gotten it yet, you just need to give up and be a janitor go back and try harder.)

Anyway, Scott goes on to discuss some other studies that might or might not support mindset theory. With a little rewording, most of his post - the substantive part, the part about research and evidence - could actually have been a post in support of Dweck's idea.

But it is not. It is a post criticizing Dweck's idea. And here is where the derp comes in.

At the beginning of his post, Scott - who is rightly renowned as being an intellectually honest person - states his priors:
I’ll admit it: I have a huge bias against growth mindset... 
I can sometimes be contrarian, and growth mindset is pretty much the only idea from social psychology that is universally beloved... 
Which brings me to the second reason I’m biased against it. Good research shows that inborn ability (including but not limited to IQ) matters a lot, and that the popular prejudice that people who fail just weren’t trying hard enough is both wrong and harmful... 
Which brings me to the third reason I’m biased against it. It is right smack in the middle of a bunch of fields that have all started seeming a little dubious recently. Most of the growth mindset experiments have used priming to get people in an effort-focused or an ability-focused state of mind, but recent priming experiments have famously failed to replicate and cast doubt on the entire field. And growth mindset has an obvious relationship to stereotype threat, which has also started seeming very shaky recently.
Scott is a big believer in the importance of inborn ability. He thinks social psych and related fields have been captured by people who want to deny that fact. He thinks these fields have been using junk science to push their nurture-over-nature agenda. And this biases him against Dweck's ideas, and biases him toward skepticism with regards to research results that support Dweck's ideas.

Remember, derp is not confirmation bias. It is more rational than that. Derp is just repeating your strong priors over and over. That's what Scott's post is.

He cites some research in favor of the growth mindset idea, and expresses skepticism of the results. He restates his priors.

He then cites some research that might or might not favor it, and expresses no skepticism of the results. He again restates his priors.

He then cites some research that shows some stuff that might or might not suggest alternative explanations for the results that support the growth mindset, and expresses no skepticism of the results. He again restates his priors.

Then he restates his priors once more in the concluding section. Derp!

Actually, maybe I'm being too generous to Scott. Looking for flaws in studies that contradict your priors, while not looking for flaws in studies that agree with your priors, is confirmation bias. It is not Bayesian-rational. It appears to be what Scott does in this post. So calling Scott's post "derp", rather than "confirmation bias", is giving him the benefit of the doubt! :-)

Anyway, when discussing the growth mindset idea, I think we all need to keep the following things in mind:

1. Hedgehogs gonna hog (and not hedge). We should remember that Carol Dweck, like most promoters of big ideas, overstates her case. That doesn't mean her idea is not a good and useful one.

2. Average is not marginal. Belief in the average effect of natural ability and belief in the marginal effect of effort are two very different things.

3. Science is really hard. Even in physics and chem it's insanely hard to get robust, reliable, definitive conclusions. In psych it's much harder. That doesn't mean we should ignore big ideas in psych.


Update

Someone asked me how my post squares with the following part of Scott's post:
Likewise, mindset theory suggests that believing intelligence to be mostly malleable has lots of useful benefits. That doesn’t mean intelligence really is mostly malleable. Consider, if you will, my horrible graph: 

Suppose this is one of Dweck’s experiments on three children. Each has a different level of innate talent, represented by point 1. After they get a growth mindset and have the right attitude and practice a lot, they make it to point 2. 
Two things are simultaneously true of this model. First, all of Dweck’s experiments will come out exactly as they did in the real world. Children who adopt a growth mindset and try hard and practice will do better than children who don’t. If many of them are aggregated into groups, the growth mindset group will on average do better than the ability-focused group. Intelligence is flexible, and if you don’t bother practicing than you fail to realize your full potential.
Scott seems to think that his graph, if true, contradicts Dweck's idea. But I think that his graph is exactly Dweck's idea.

Scott also seems to think that in order to realize the benefit of a growth mindset, it is necessary to convince children of a fiction (i.e. that the overall effect of natural ability is 0). But Dweck doesn't think that. That seems unlikely to me. My guess is that you just need to convince kids of the truth - i.e., that effort matters a lot on the margin for most people. No fantasies required.


Update 2

Scott has a follow-up post.

When reading this discussion, it is important to remember that the following three things are different:

1. Dweck et al.'s belief about the effect of the growth mindset on performance

2. The belief of a person with the growth mindset about the effects of effort and ability on performance

3. Dweck et al.'s belief about the effects of effort and ability on performance

Tuesday, April 07, 2015

Bond bubbles



Apparently, I convinced Brad DeLong that the idea of a "bond bubble", though a slight misnomer, is not insane. Brad's explanation of the idea is a little convoluted, so let me see if I can boil it down to essentials.

Why can't a government borrow and spend infinite amounts of money? Well, interest payments might get too high, forcing the government to default. But what if interest rates keep getting lower and lower? As nominal interest rates go to 0, interest payments go to 0, so the govt. will always be able to make interest payments no matter how big the debt gets.

So all the govt. has to do in order to be able to borrow and spend infinite amounts of money is to get the central bank to keep interest rates at 0 forever, which the central bank can do by - basically - printing money and buying bonds from people and banks.

But what if people and companies stop buying the government bonds in the first place? Well, the central bank can just print money and buy the bonds directly from the government. (We call this "the full MMT." It is almost certainly the route Japan will have to take.)

So is this a free lunch? What's the danger? The danger is that all that money-printing will eventually result in a big inflation. A big inflation will also raise nominal interest rates, overwhelming the central bank's ability to keep them down. That will cause a government default.

So is this a danger for advanced countries right now? Well, markets have very low expectations for future inflation, and for future interest rates, for most rich countries. If you believe markets are efficient, this means that it's likely that governments can keep borrowing money, and central banks can keep printing money, without causing inflation - at least for now.

BUT, if markets are not efficient, then people could be underestimating the risk of a rise in inflation and interest rates. That is what people are afraid of when they talk about a "bond bubble." No, it's not a speculative bubble, but whatever. This is what I pointed out to Brad DeLong.

Note, btw, that a bond bubble isn't the only reason to take low inflation expectations with a grain of salt. Those low expectations could be based on people's predictions that governments and central banks won't engage in infinite-borrow-and-infinite-print policies. If governments surprise people by trying those policies, the markets could quickly adjust their expectations. You don't need a bond bubble to be afraid of a run on the currency - you just need Goodhart's Law.

The Hugo Award silliness



I will freely admit that I haven't been very enthused by the Hugo Awards in recent years. Looking back through 2007, they were a very good guide to stuff that I would like. For example, through 2007, I liked about 3 out of 4 Hugo nominees and winners for Best Novel. But since then, I've only really liked one of the Hugo winners - Paolo Bacigalupi's The Windup Girl, and fewer of the nominees than before. In fact, many of my favorite SF novels that have come out since 2007 - Ernest Cline's Ready Player One, Hannu Rajaniemi's The Quantum Thief, Ramez Naam's Nexus - didn't even get nominated.

What's going on? I don't know. Maybe nothing. Maybe just statistical noise.

The thought had crossed my mind that maybe sci-fi fandom had shrunk, migrating to video games or anime or YouPorn or whatever, leaving a core of sci-fi writers who care more about literary writing quality than neat ideas. I enjoy literary quality, of course - I love me some Alice Munro or George Saunders or Russell Banks. One of the best literary writers I've ever read, Margaret Atwood, is also one of the best sci-fi writers I've ever read (Oryx and Crake is the best sci-fi book written since the turn of the millennium, dammit; the fact that it also didn't get a Hugo nomination may have been the beginning of the downward trend). But what writers like in sci-fi and what I like in sci-fi tend to be different, which is why I only tend to like about a third of the Nebula Award winners.

But anyway, this same thought appeared to have crossed the mind of sci-fi writer Brad Torgersen, who organized what he thought was a campaign to take back the Hugos. Joining together with another author named Larry Correia, he has been nominating a slate of authors called "Sad Puppies", which he claims represents a return to consumerism and fun.

If Sad Puppies really had been that, it might have been interesting or even a positive force. Unfortunately, Correia and other Sad Puppies followers had their own ideas of where the problem lay. They decided it was all about right-left politics, and that recent Hugos had been chosen as affirmative action picks. The solution, according to these Puppies, was to nominate a bunch of authors with rightist political beliefs.

Things got really bad when a troll named Theodore Beale, who calls himself Vox Day (presumably because if he called himself "Vox Dei" he wouldn't come up first in Google searches), one-upped the Sad Puppies by creating an aptly named "Rabid Puppies" list of even more right-wing authors. By gaming the rules for fan voting and by recruiting outside voters from GamerGate, the various "puppies" managed to grab most of the nominations for this year's Hugos.

By far the biggest beneficiary of the puppies' putsch was a man named John C. Wright, who writes very disrespectful things about gays and other such people who never did him any harm. He also writes science fiction (as an aside, I think his novels are unreadable, though I really liked this short story; but that is irrelevant to the point of this post). There just aren't many rightist sci-fi authors out there, so Wright became an army of one, grabbing six total nominations, including three in the Best Novella category.

This is basically an experiment in politics-based affirmative action, similar to what Jonathan Haidt wants to inflict upon American universities, but more extreme. My instinct says that it will produce a deluge of craptastic crap. It's not that being a right-winger is incompatible with being a good science fiction writer - I like Starship Troopers and Ender's Game. It's that whenever you select people based on their political beliefs, you select people along a dimension that is at an angle to actual quality. Since the pool of rightist sci-fi authors is small and the number of puppies nominations is large, I expect that the effect will be far more severe than for race- and gender-based affirmative action.

Of course, the puppies also did seem to follow Torgersen's wishes to some degree, picking some works based on fun rather than on Vox Day's right-wing politics. For example, I see a Jim Butcher novel in this year's nominations. Jim Butcher is great, and very fun. The rest of the Best Novel nominations look all right. For the other categories, the only right-wing nominee whose works I have actually read is Wright, and he did once write a story that I liked, so maybe some of these other stories by him are good, despite the fact that he himself is a bit of an orc.

So maybe the puppies won't utterly ruin the Hugos. But they can't have helped. And it's another step in the negative trend of the politicization of geekdom that began with GamerGate. I expect to see a counterattack by SJWs at the next Hugos, and further retaliation by the puppies. When rightists and leftists fight, no one wins. That was true in 1930s Europe, and it's equally true in modern geek fandom.


Update: George R.R. Martin has a great post rebutting the arguments of the puppies.

Update 2: John C. Wright appears to have taken down the disrespectful post that I referenced, and I commend him for doing so.