Wednesday, April 30, 2008

The minimum wage and dataset bias

Megan McArdle has a good post on ye old Minimum Wage Monopsony debate. The post is a little technical, but it's a technical subject. She concludes that bad data is the best explanation for the Card and Krueger minimum wage results, which showed an increase in employment in New Jersey after the state minimum wage was raised there. This is because all of the other stories--collusion, employment searching costs, heterogenous preferences, etc.--are all extremely unlikely when it comes to the fast food industry, for obvious reasons.

(See also this good 2006 Will Wilkinson post.)

I've always considered the Card and Krueger study a bit overblown. We get to hear the same argument every time minimum wage increases are brought up in the public sphere, and none of the numerous other minimum wage studies showing a negative employment influence are as generously cited. The discussion tends to turn into "theory vs. history," without even well establishing the history.

So what's going on here? David Card is an honest and intelligent economist. He is not a crackpot or an ideological crank. He has some data and he is trying to make sense of the world accordingly--at least, that's how he strikes me in his interviews. I wonder if some sort of general researcher bias has snuck in here, which is common enough. Researchers invest a lot of time and money into their datasets and they don't easily give them up, even when there are lots of complaints about their methodology.

Surely David Card knows much more about labor economics than I do. But researchers of all stripes have an incentive to protect their past works, and the monopsony argument just doesn't make much sense for the fast food industry. Read more.

Limited Liability

Interesting summary from Alex Tabarrok on the arguments for and against limited liability and the evidence concerning its efficiency. Do read Stephen Bainbridge's followup; a lawyer's perspective is always useful for these things. Read more.

Is economics a science?

It's a silly question, so it's no wonder that the people who so frequently ask it write silly articles.

I'd have a tough time trumping Ronald Bailey's take on it, so take a gander at that. But a couple of comments nonetheless. A few excerpts from the article:
But what is not widely known is that these now legendary economists—William Stanley Jevons, Léon Walras, Maria Edgeworth and Vilfredo Pareto—developed their theories by adapting equations from 19th-century physics that eventually became obsolete. Unfortunately, it is clear that neoclassical economics has also become outdated.
. . .
The physical theory that the creators of neoclassical economics used as a template was conceived in response to the inability of Newtonian physics to account for the phenomena of heat, light and electricity. In 1847 German physicist Hermann von Helmholtz formulated the conservation of energy principle and postulated the existence of a field of conserved energy that fills all space and unifies these phenomena. Later in the century James Maxwell, Ludwig Boltzmann and other physicists devised better explanations for electromagnetism and thermodynamics, but in the meantime, the economists had borrowed and altered Helmholtz’s equations.

The equations used to describe marginal economic behavior are not outdated because physicists use different models now. A derivative is still the slope of a line at a point. Nobody ever said economic behavior could be modeled the same as electromagnetism, and frankly, this is one of the stupidest critiques I've ever read. Economists use calculus and algebra. Last time I checked, neither had been invalidated.

If the environmental crisis did not exist, the fact that neoclassical economic theory provides a coherent basis for managing economic activities in market systems could be viewed as sufficient justification for its widespread applications. But because the crisis does exist, this theory can no longer be regarded as useful even in pragmatic or utilitarian terms because it fails to meet what must now be viewed as a fundamental requirement of any economic theory—the extent to which this theory allows economic activities to be coordinated in environmentally responsible ways on a worldwide scale. Because neoclassical economics does not even acknowledge the costs of environmental problems and the limits to economic growth, it constitutes one of the greatest barriers to combating climate change and other threats to the planet.

Want economics to acknowledge environmental costs? It's very easy: start including a base utility function for animals in your public policy analysis. Most economists don't already do this because, well, we can't speak to animals. We don't know that they experience things the same way we do. But if you go ahead and assume that they don't like pain and they don't want to die, and that these facts should have any weight at all in your cost-benefit analysis, things start looking quite a bit different.

It's a bad, bad article, but you can find some more general and somewhat more reasonable questions about the scientific validity of economics elsewhere. I'll say something about them soon. Read more.

Ebert blog

Roger Ebert has a blog. Neat. Read more.

Speaking of Siren

The original English dubbing of Siren is quite bad. All of the distinctly Japanese characters are given awkward British voices, names like "Risa" (Lisa) are left poorly-Romanized, and many of the significant archive items you find throughout the game are left in Japanese. (They are partially translated in subtitles, but on a few of them there are undescribed bits you can make out if you look closely. Or so I'm told.)

However, it seems the game is getting remade and retranslated for the PS3. (Video here.) I'm not confident the dubbing will be much better, but if they rework the myriad frustrating elements of the original--the glitches, the busted checkpoint system, the ridiculously overpowered enemies, the terrible combat, etc.--while keeping the narrative intact, it will be, no joke, the best game ever made. Read more.

Marathon Infinity and Siren

Been playing through the Marathon trilogy, a series of first-generation first person shooters originally Mac-exclusive. (Each game is freely downloadable for various operating systems at the above link.)

Much of the level design and ideas for the games are typical for the genre at the time. Long, dark corridors filled with aliens, invincibility and cloaking powerups locked behind hidden doors that look like the surrounding walls, lots and lots of switch-pressing. The main difference is that Marathon has an actual story. The first two games are a semi-interesting scifi venture, but Marathon Infinity is a shining example of how videogame narrative can be used to bolster a story, while so many games treat the narrative as incosequential.

Why? (Spoilers ahead.) Well, the story is entirely told through monitors--computer terminals you find in the game that display commands, usually from one of the rampant AIs who are pushing you around. You can find the text of every monitor in the game on Bungie's own site and read the story that way, but you simply won't get the context of the levels. The game's puzzles are not "which AI is telling me to do what," but "who is shooting at me and why?" Sometimes the aliens you've been killing for the last two games are assisting you to mass-slaughter human enemies. Sometimes the other way around. And there's another alien race that appears to switch allegiances willy-nilly, no matter which AI is giving you orders; puzzling, considering that their liberation and eventual siding with one of the AIs is the main plot of the previous two games. The monitors are simply insufficient evidence to figure out what's going on; both humans and aliens recognize your character and attack or defend accordingly, so they must be basing their aggression on your past actions.

And then there are stages like "Electric Sheep," which are devoid of any action--or any contact from the AIs who treat you like a pawn. Monitors in these levels tend to display odd poetry or dream-like sentences. But the rest of the level is significantly surreal: Enemies glow and are suspended in mid-air, everything seems to be in stasis. The game leaves entirely open the possibility that you're just crazy or dreaming.

I'll stop talking about Marathon there; I don't want to step on the toes of my buddy who's writing a Marathon article for the next issue of TGQ. But I wanted to note that the only other game I've played with narrative-story conjunction on par with Marathon Infinity is Siren. I could dedicate a long, long essay to the brilliance of Siren (though it might be reduntant), but for now I'll just say that Siren is similar in the way the game is essential to determine context. There is no way you can understand what's going on just by watching all the cutscenes. The town has a sensible layout that's key to understanding what characters are where, and the Shibito zombies are characters with personalities themselves. In the writing world, we would call Marathon Infinity and Siren examples of "showing not telling," even though both games have quite a bit of telling to do. It's also interesting that both games take the approach that much narrative should be left open to player interpretation, which is unusual for most stories. Read more.

Monday, April 28, 2008

Science is awesome

Scientists are experimenting with summoning lightning bolts with lasers.

(HT Megan McArdle) Read more.

Sunday, April 27, 2008

The Limits to Growth

From an entry on Skepchick by Donna:
Over the past few years I’ve spoken to several conservatives about economics and they can’t get it through their frakking heads that growth cannot be sustained infinitely.

And from the comments:
I know there are many problems that would not be solved by reducing the number of people, but all of the problems having to do with the environment, food, water, and so forth would be come much less of an issue if there were less than a billion humans on the planet. Of that, I am sure.

I'm not so sure. It's a question of which billion humans, innit? If we left a billion humans alive in only Africa, they would still have water problems. Most of Africa (especially the poorest countries) is sparsely populated, after all.

Of course, I can imagine other solutions too. A cheaply-produced pill that decontaminates water would solve those pesky water problems. So would a machine that did the same. Etc. As long as ideas like these are not impossible, then our current problems are hardly an absolute limit to growth.

What if we were all computer uploads, living in the same network, with no more need for food, water, air, transportation, or shelter? Would that eliminate the so-called limits to growth? What if a self-improving AI made itself so intelligent that it could create inventions humans could never dream of? Would that put a dink in limits-to-growth theory? If we had robotic bodies? If we developed biogenetic implants that made our bodies more efficient? If we evolved such that our current consumption patterns were no longer desirable?

The socialist skeptic (or unbeliever, as it were) has a problem. I understand why Roman Catholics believe self-improving AIs and future human uploads are impossible. I don't know why a person who doesn't even believe in a soul thinks this. Even if these grand ideas don't come to fruition, that doesn't mean other important changes won't occur. The future tends to be weirder than we expect. If you want to know why people can't get "limits to growth" through their frakking heads, it's because past human experience shows igenuity solving problems. This is no proof that the trend will continue in the future, but it is evidence.

And then there's this:
On a related topic, it depresses me that the American lifestyle requires slave labor to sustain itself. This was true at the time of the American Revolution, and it is basically true today. It’s just that these days we primarily get our goods made by wage slaves who live on other continents so we don’t have to look at them.

This is some extreme economic silliness that would require many entries to dispell--if you think there's anything to it, you might want to listen to this lecture/podcast--but here's some food for thought: Why is it that when we exploit people, they tend to become richer by thousands of dollars per year?

I may write more later with some more conclusive data. Read more.

Saturday, April 26, 2008

Penn and Teller explain why magic works (and voting doesn't always)



This clip is from the first episode of The Unpleasant World of Penn and Teller, a six-episode special that was filmed and aired in the UK. It's a great lesson about magic, but it also works as a good analogy of Bryan Caplan's Myth of the Rational Voter.

The audience member publicly professes an irrational belief (the card is lost in the deck) until Penn asks him to put some money down. Likewise, Caplan theorizes that voters profess (and politicians pander to) irrational beliefs for a host of reasons. However, when serious money is at stake, people are more likely to conform their beliefs to evidence. In other words, people vote for bad policy because it makes them feel good and they, as individuals, don't have much at stake (although the losses for the entire country can be quite high). Read more.

Wal-Mart and evidence

So you know how Wal-Mart is the big evil corporate empire that destroys small businesses? Well. . .
Our research suggests that the popular belief that Wal-Mart has a significant negative effect on the size of the mom-and-pop business sector of the United States economy is statistically unfounded. After examining a plethora of different measures of small business activity and growth, examining both time series and cross-section data, and employing different geographic levels of data and different econometric techniques, it can be firmly concluded that Wal-Mart has had no significant impact on the overall size and growth of U.S. small business activity.

The link is a PDF of an article from Regulation Magazine, the quote is the concluding paragraph. They begin by explaining "creative destruction," a process most economists consider necessary but a counter-intuitive idea to, if not many citizens, at least organizations like Wal-Mart Watch. Just as you can't say that the decline of American manufacturing employment means the decline of the American manufacturing sector, you also can't say that the closure of many retail stores means the death of small business.

The article goes on to examine the statistical claim that the presence of Wal-Mart stores entails the destruction of small business and finds, based on several measures of small business activity, no such correlation.

Thinking like an economist--thinking like any kind of scientist whatsoever--requires thinking like a skeptic. This new analysis reminded me of the Benjamin Barber/Will Wilkinson debate of last year, shortly after the release of Barber's Consumed. Early in the debate (transcript available here), Barber alleged:
Here's the problem. We know that big-box stores have social consequences that we don't think of when we shop there. They make normal mom-and-pop, small retail impossible, but the problem is in the United States and in most cultures, small-town, small-scale retail is the core of our community development whether they're villages or towns or urban neighborhoods, and you put small-scale retail out of business and you destroy communities, you destroy the basis for the community values that come with those communities. There's a cost.

There are a lot of heavy claims in there--all without even establishing a statistical link between Wal-Mart and small business decline. As James Randi, ur-skeptic, might say, this is just like debating the mechanism behind psychic spoon-bending without first even demonstrating that spoon-bending is a real effect rather than a trick. When a questionable idea pops up in the public sphere, like "Wal-Mart destroys jobs" or "trade deficits cause massive recessions," it's probably a good idea to wait until there's some good statistical evidence on the subject before you include it as a major point of your thesis.

There may be many other valid criticisms of Wal-Mart, but critics would do well to discard this one from their repertoires, at least until they have some real evidence.

(HT Don Boudreaux.) Read more.

Astrology fails (again)

A recent scientific study has found astrology to have no predictive power for anyone's lives. Just like every other scientific study of astrology. The methodology of this one is particularly good, and the results very robust:
For several decades, researchers tracked more than 2,000 people - most of them born within minutes of each other. According to astrology, the subject should have had very similar traits.
. . .
Researchers looked at more than 100 different characteristics, including occupation, anxiety levels, marital status, aggressiveness, sociability, IQ levels and ability in art, sport, mathematics and reading - all of which astrologers claim can be gauged from birth charts.

The scientists failed to find any evidence of similarities between the "time twins", however. They reported in the current issue of the Journal of Consciousness Studies: "The test conditions could hardly have been more conducive to success . . . but the results are uniformly negative."

If you know anything about statistics, you know that's a darn good sample size. It's especially useful because of a typical astrologers' copout I noticed in an old video recently. It was an appearance of James Randi on Oprah, along with three absolute nutters, one of which was a professional astrologist. When Randi brings up the most obvious scientific refutation of astrology--twins are not known, statistically, to have identical personalities or futures--she countered that the precise time of birth mattered as well. In fact, each prediction and fortune of astrology comes from a difference of four minutes in time of birth! So if a twin is born within the next minute of his sibling, but that minute falls within the next four-minute block of astrology prediction, you can't expect them to be the same.

So it's nice to have an enormous, decade-long survey that tracks even time of day so closely.

(Randi on Oprah video, in five parts, here. Hat tip Pharyngula. Read more.

Blog change

I've spread myself a bit thin on the internet. I thought that nobody wanted to read about both economics and videogames--and it's true, a lot of people don't--but if Megan McArdle can blog about vegan cooking and subprime mortgages, I can do whatever I want. So I'm consolidating other blogs into this one. Don't worry, I'll use labels so you can skip whatever posts you want without trouble.

The blogging bug has hit me hard again, I don't know why. Maybe it's the approach of grad school, maybe it's meeting new people who are interested in, rather than bored by, things I have to say. So expect a lot of updates from here out. Read more.