Wednesday, May 28, 2014

Bad Data Science of Love

I was sorta shocked when I received an invitation to take a course on Data Science from the "data scientist" behind the OK Cupid blog. While the blog is whimsical and fun and was largely responsible it seems for Ok Cupid's popularity, the statistics were largely un-grounded and amateurish, which is fine for a blog, but now he's going to teach these methods.. as "science?!"

A good example of this junk data science was on this recent ted talk featured on npr. It starts with a common Fermi problem of estimating how many suitable partners there are in your city (so common, I stumbled into this conversation the one time I visited the Stanford physics lounge). Usually by adding more and more criteria (your age, right gender, right education, food tastes, physical attractiveness, height), you find there is only 1. But usually those estimates are exaggerations because they assume independence which is not true.

She then uses an arbitrary formula for optimizing her ideal mate.  She made uparbitrary weight to arbitrarily selected characteristics.  Much like the simple scoring systems my mom used to use to pick jobs or things growing up. A useful tip, but its not science.

She then collected data from OKCupid profiles to figure out which are most attractive, but again, did nothing to separate causality from correlation, and most of her conclusions were based on qualitative assessments. Again, interesting, but not science.

There was a recent Wired story of a recent math phd who at least used machine learning algorithms to construct the optimal profile. There was at least some science there. But still, mostly ad hoc assumptions not grounded by theory.

I guess all this reminds me of the difference between econometrics and statistics. I was recently interviewed for a consulting opportunity and was asked why should we hire you as opposed to a statistician.

I think the answer is that to really do "science" about behavior, you can't look at data alone, you need theory. You need theory to separate causation from correlation. And on questions of choice, economics have 100 years of thinking hard and formally on the axiomatic primitives behind how choices are made.

Tuesday, May 27, 2014

Cultural Replication of Food Trends: Next Big Thing

My institutional theorist professors would appreciate the global cultural replication of culture as evidenced by trends in fine dinnig. For better or worse, food trends in fine dining restaurants around the world seem to follow the same playbook, (perhaps just delayed by 1-20 years depending on restaurant and location). For better or worse, I've sat down in restaurants ranging from Vietnam to Budapest to Istanbul that all served menus that could easily have come off a Jean-George spinoff from the late 90's.

In my brief experience 15 years eating at fancy places, it started with the new international paradigm, (e.g. rare wasabi encrusted ahi tuna steak), that gradually shifted to farm to table (as pioneered by Chez Panisse), had a brief fling with molecular gastronomy (as pioneered by El Bulli) with such extravagant gestures like Moto's edible menu and laser vaporized orange peel, before those foams and liquid nitrogen shattered edible soils, got synthesized into the mainstream farm to table style that prizes locavore sourcing and parochial cultural authenticity which seems to be the existing dominant fine dining paradigm.

There has been rumblings of a new Kuhnian shift recently in the food world though, as heralded by Noma in Copenhagen ascension to the top of world league charts, that I hadn't quite been able to put a name to, until now, with all the buzz around Dan Barber's new Third Plate book. Barber, whose Blue Hill at Stone Barns is perhaps the apotheosis of farm to table, by literally being on a farm, calls farm to table 2nd plate, the new thing is 3rd plate.

I think of this as an off shoot of the Farm to Table idea of Nose to Tail, Snout to Hoof (e.g. fergus henderson). But instead of limited by all parts of an animal, instead we eat all parts of an eco-system (Fungus to Flora perhaps you could call it). Noma Redzepi's favors local fungi, wild roots, multi-year cured meats. Barber has been advocating eating what's good for the ecosystem, eating what is traditionally seen as weeds or animal feed, rather than eating foods based on habit or tradition, that developed at a time with different technologies and agricultural circumstances. Both hearken a bit to even earlier traditions, of hunting and gathering and somehow being in harmony with the environment.

I'm sort of excited. Farm to table has gotten a bit old. It's also neat to see paradigm shifts and cultural replication in action. As an economist, I'm a little bit skeptical of Barber's efficiency logic behind the third plate, but as an eater, I'm excited at the new prospects.

Thursday, May 15, 2014

Varys' Riddle - the game theory of game of thrones

In George RR Martin’s book of political intrigue one character poses the following riddle:

In a room sit three great men, a king, a priest, and a rich man with his gold. Between them stands a sellsword, a little man of common birth and no great mind. Each of the great ones bids him slay the other two. 'Do it,' says the king, 'for I am your lawful ruler.' 'Do it,' says the priest, 'for I command you in the names of the gods.' 'Do it,' says the rich man, 'and all this gold shall be yours.' So tell me- who lives and who dies?

Above is a link to the scene from the HBO tv show.

Game Theory, it turns out, has an answer. Consider what some call the repeated game of daily life. In each period, every member of society is randomly matched with someone else whereupon they play a Prisoner’s Dilemma. In each period, each player has two choices, either to be violent and slay the other, or to be peaceful, and cooperate. It is individually rational to be violent in any given period, but mutual cooperation is better for both than mutual violence.
In analyses of two player versions of this game, we see a wide variety of strategies can maintain mutual cooperation. For example grim-trigger or famously tit-for-tat. The common take away, people play nice in any given round for fear of being punished in the future.

In the large N versions, the problem becomes more difficult because you interact with somebody new each period. Thus in order to maintain cooperation you need some kind of social order. If somebody defects and decides to slay, we need to rely on somebody else to carry out the punishment.

Thus the answer to Varys' riddle depends on which equilibrium we are in, or perhaps if we are out of equilibrium, which sub-game of the current equilibrium.
Calvert (1992) finds an equilibrium of this game where players maintain equilibrium by choosing one member (a king) and using the king to collect information on who has defected, and thus giving the king power to decide who should be killed. Like in Hobbes’ Leviathan, we give the king power because he maintains order by punishing anyone who defects, thereby keeping everybody in line. Assuming this equilibrium is in play, the sell sword will listen to the king or risk being punished by whoever he encounters next.

Milgrom, North and Weingast (1990), find that merchants actually can maintain order without a government or king. By studying the “law merchants” of the Champagne trade fairs of medieval Europe (if you ever visit Paris, take a train to the town of Provins, today it’s basically a medieval times ren faire for French tourists, but you can still find the animatronic law merchants). Here, contracts were enforced by judges who kept a board that recorded the names of anyone who breaks a contract. In equilibrium, all merchants had to defect against anyone who ever violated a contract, or have their name recoded on the board as well (making sure the equilibrium is sub-game perfect, the judges themselves also would be punished if they were dishonest, ensuring sub-game perfection in their play as well). So long as this equilibrium of contract enforcement is in place, the merchant can credibly offer future money to the sell sword in exchange for his life.

Kandori (1993) gives conditions on when we can sustain cooperation based on some sense of moral authority alone that has a Kantian flavor. Consider a moral norm based on a grim trigger strategy where I cooperate in every period, unless I am ever wronged. However, if I am ever wronged, I will bring violence to everyone I encounter from here on out. Even though I may never see the person who wronged me ever again, if I start bringing violence to others, that should trigger a chain reaction where they start bringing violence to others, which brings down the entire system. Thus, people cooperate, to avoid setting off this chain reaction that brings down the system as a whole. A priest could appeal to the sellsword’s rational morality, that every act of violence has a chance of bringing down civilization as a whole.

Of course, all of these above institutions only work when we are on the equilibrium path. But when we are in an era when Winter is Coming, then we very well may fall out of equilibrium. In such conditions, Acemoglu and Robinson (2001) model the process where Latin American countries transition between cooperative behavior and a state of mutual violence, what Hobbes’ referred to as the “nasty, brutish and short” state of nature. In such a world, all that is left is violence, and whoever is best at violence, the sellsword, prevails. 

Tuesday, May 13, 2014

Economics of Apologies on Vox

More Shameless Self Promotion, my Op-Ed about apologies on Vox.

Tuesday, May 06, 2014

Thank you Professor Becker

Gary Becker was the biggest influence on my advisor, Eddie Lazear, and by transitivity, me. In terms of applying simple mathematical models based on the assumption that people take purposeful actions, motivated not just by money, but on things like values, emotion, and culture. I agree that he may have had the most profound influences on the social sciences in the past half century, where his work is routinely cited across the social science disciplines, and was fundamental to the economic imperialism into other disciplines that has let me study the economics of apologies and guilt and social identity. I was fortunate enough to get to have lunch with him just a few months ago so it is shocking to hear of his passing. Thank you, Professor Becker, and may he rest in peace.