Economics and similar, for the sleep-deprived

A subtle change has been made to the comments links, so they no longer pop up. Does this in any way help with the problem about comments not appearing on permalinked posts, readers?

Update: seemingly not

Update: Oh yeah!


Wednesday, November 30, 2005

 
The long awaited Freakonomics post

This is the first in a series of posts with which I hope to present a critical assessment of "Freakonomics", a book which I have now read (I didn't take part in the original Crooked Timber seminar due to lack of time)[1] and which I humbly assess to be a really bad book. It's reasonably topical apparently because there is a bit of a kerfuffle about his work on abortion and crime, but I won't be addressing that here. In this first post (and readers beware; I have a lot of previous form when it comes to starting these series and never completing them), I'm mainly taking a look at the chapter entitled "Why Do Drug Dealers Still Live With Their Moms?", which largely summarises Levitt's work with Sudhir Venkatesh on a crack-dealing gang in Chicago. I'll note at this stage that I have not read the underlying research papers on which Freakonomics is based; normally I would regard this as necessary, but in this case I don't think I should have to bother. I'm not, for the most part, reviewing his academic work or trying to establish the truth or falsity of the claims he makes; I'm trying to establish a particular point about the book Freakonomics, which is that it's not a good book. Levitt and Dubner placed this thing into the public sphere, they were not shy with their activity in publicizing it and they have not, at any point which I'm aware of, tried to make it clear that the book cannot possibly stand alone or that its conclusions are tentative and subject to very many or important caveats discussed elsewhere in the literature. So as far as I'm concerned, the book is an entity in and of itself, and can be discussed, debated and reviewed as such, just as if it were written by Thomas Friedman, James Surowiecki or any of the other occupants of the neighbouring bookshelves to the ones from which Freakonomics is sold. So with those caveats out of the way, I'll begin with a few motivating remarks about why I don't think much to the methodology of this book (a subject to which I'll return in part 3), followed by a close look at one of its chapters to try and convince you guys that these shortcomings lead to real and practical problems. So here goes:




Chapter One: Introduction and Crack Gangs

Malcolm Gladwell is a seriously clever bastard.

I mean, one of the indications of Noel Coward's genius is that he came up with a way of being honest about a friend's play to their face when you didn't like it (he'd say "darling, good was notthe word!" or "well, you did it again!"). But Gladwell has taken it to a new level he's managed to pull the same trick, but so subtly that the authors of Freakonomics have actually put it on the back cover blurb of their book. Gladwell is quoted as saying:

"Steve Levitt has the most interesting mind in America, and reading Freakonomics is like going for a leisurely walk with him on a sunny summer day as he waves his fingers in the air and turns everything you once thought to be true inside out. Prepare to be dazzled"

And Gladwell is right; the waving of the fingers in the air is indeed the main argumentative and rhetorical technique by which this book attempts to establish its conclusions.

There are a number of things I don't like about Levitt's approach, but the biggest problem I have with him is his[2] habit of saying (in various forms of words) "whichever way you look at the numbers, XYZ" when he means "whichever way I look at the numbers, XYZ". On a lot of these subjects (by far the most obvious example is abortion/crime, but it is an issue in all of them to a greater or lesser degree), Levitt is looking at quite large, clearly multicausal issues where any model is likely to be partial and all manner of conflicting theories can claim support from the data. "Freakonomics" absolutely does not recognize this fundamental truth of econometrics; it might be because the authors don't have the statistical chops to understand it but I think it is much more likely that they are just trying to copy the monolithic tone of voice adopted by social reformers and similar blowhards who hand out their assertions with no data at all. In all honesty, I think that theL&D approach is a retrograde step; it's easy for the untrained reader to spot when someone has no empirical support at all for his position, but much more difficult to deal with someone who systematically overstates the empirical support that he does have. This is at least 90% of what makes John Lott so pernicious, and it seems to me that L&D are involved at least partly in the same sort of game.

It's the game of pretending that difficult social questions have easy non-sociological answers. There are lots of people in this space, and not all of it is by any means bad. Any look at a difficult question is going to be either hopelessly oversimplified or hopelessly unreadable, and I would certainly prefer it if people erred in the first direction. There's also a decent Hayekian (or indeed Bayesian) point to be made here that if you're entering into the marketplace of ideas to try and extract the truth from a number of differing viewpoints, then you want everyone to give you their idea, not to caveat it all about with bits and pieces of other people's ideas. That's why I'm prepared to give a (limited) free pass to Malcolm Gladwell or James Surowiecki when they write books like this which, in my view, present absurdly oversimplified views of the world, because I understand what they're trying to do; to present their view on a question, not to give the final indisputable answer.

The problem comes in when someone attempts to present their view of a question as if it is the final indisputable answer. A lot of the things in Freakonomics are things that I wouldn't make too much of a fuss about if the authors were just advancing them as their view of one way of explaining the facts. But they don't do that; at key points in the book, they keep claiming that they're reporting the facts when they're clearly (to me at least) reporting a particular spin on the facts. This is the pop-science approach to social questions, because it's trying to combine the authority of a scientific investigation with the unequivocal certainty of a theoretical pronouncement. What Levitt and Dubner are doing is exactly the same thing that Thomas Friedman does; telling a bunch of stories and then explaining how these stories fit into their view of the world. However, in the case of Friedman it's always obvious that someone else could tell entirely different stories about the same kinds of people and events and fit them into an entirely different worldview. Because of the way that Freakonomics has pitched itself at the pop-science crowd (constantly banging on about Levitt's John Bates Clark medal and referring to all the statistical analyses; for fans of cringeworthy exegesis, page 161 of the American edition contains what I strongly believe to be the worst description of the linear regression model ever committed to print), however, they are always either implying or outright saying that their stories are the only ones consistent with the facts, so we can either fit their stylized facts into our own worldview or (preferably) drop ours and buy theirs. As you can tell, I don't like this.

The pop-science approach to economics is dangerous and irritating in itself (Krugman's "Pop Internationalism" refers). But when combined with the panache of a seasoned magazine journalist, it becomes downright sloppy. What actually set me off on this trail � the initial clue that there was something very wrong about Freakonomics was a throwaway remark, presumably inserted by Dubner and certainly unsupported by any of Levitt's work, to the effect that "the typical prostitute earns more than the typical architect". This remark is asinine. What on earth are they talking about? There is probably a reasonable working definition of a "typical architect" (though I can think of about five different types of architect off the top of my head), but what is a "typical prostitute"? Do they mean per hour or on an average annual earnings basis? Is there any data to back this up (the only study I could find put average earnings for street prostitutes in Los Angeles, who are about as "typical" as any other prostitutes at $23485 in 1991, which seems low for an architect)? Fair enough, this is really just a throwaway remark aimed at illustrating a point about labour market theory, but surely the whole freakonomicsing selling point of this book was meant to be that the authors didn't make lazy assumptions and throwaway remarks but checked things against the data. I'm sorry, but if a bloke says "of course, prostitutes make a mint, they do, they earn much more than you or I", then in my estimation it is going to count very much against his subsequent claim to never take things on trust or to tirelessly question conventional wisdom. And once you start looking at Freakonomics with a critical eye and the view that some of the facts in it might not have been checked all that well, you find a lot of other things start shimmering and vibrating with the temper of a fact at bay.

The factoid about swimming pools and guns on page 150, for example, is really troubling to me. As presented in the book, the argument is obviously wrong. Levitt divides the number of child deaths caused by guns by the number of guns, then divides the number of child deaths caused by swimming pools by the number of swimming pools, compares the two numbers and says "if you have a gun in your house and a swimming pool, the pool is more likely to kill your child than the gun". Which might or might not be true, but this calculation can't possibly be the right way to prove it. Riddle me this; what proportion of the guns in the USA are held in households with no children in them? What proportion of the swimming pools in the US are owned by households with no children in them? Is there perhaps a pretty good reason to believe that households which differ in their gun ownership and swimming-pool ownership will also differ in other potentially relevant ways? Is there a good reason to believe that the fact that a house has a child in it will be informative about the relative likelihood of gun ownership and pool ownership? Now, Levitt might, for all I know, have actually done the more rigorous analytic work which would support his claim here. But if he did, I bet he did it in a proper journal where he stated the claim with the proper caveats and was totally clear about the degree of confidence that could be placed in it. But that's not what he does in "Freakonomics". He just a) puts the factoid straight in front of the reader with no qualifications at all and b) backs it up with a calculation that is clearly flat out wrong. He's simultaneously teaching the lay reader to make definitive statements without acknowledging estimation problems, and to ignore correlations between explanatory variables. How on earth can this not be worsening the overall level of debate?

So we've got two central problems here; first, the admixture of empirical evidence and extrapolated conclusions, and second the curious mixture of carefully checked scientific work and completely unrigorous off the cuff assertions. We can find both of these problems hard at work in the chapter of the book where it is discussed what insights neoclassical economics can bring to understanding criminal gangs. From now on in, I'm going to assume that you have read the relevant chapter of Freakonomics; to be honest if you haven't you can probably work out what I'm talking about and it will help you to later pretend to have read the book at dinner parties[3].

The first thing I have to note in expounding an alternative theory of the criminal firm is that there is only a loose requirement on me to be consistent with the contents of the spiral notebooks which contained the "accounts" of the gang and which form the centerpiece of the chapter. With due respect to the risks Venkatesh took in the name of science in getting them, I think that by treating these notebooks as "data" on a par with the output of the BLS, Levitt is behaving in an extremely na�ve fashion. In particular, one of the things that (presumably) Dubner's attempts to make the overall narrative swing along doesn't fully conceal is that in Chapter 1 of the book, Levitt and Dubner are telling us that high school teachers cheat and misreport numbers all the time because there is an incentive for them to do so. However, when it comes to looking at the unaudited self-reported incomes of street crack dealers to a nominal boss who exercises next to no direct supervision over them, this cynicism is gone and we are meant to assume that what the numbers say is what happened. I simply don't buy this. Because of police busts if nothing else (and the street dealer's policy of swallowing the drugs he has on him if he even suspects a bust, after which operation the crack is no longer resaleable), the shrinkage in crack dealing must be at least as bad as in other forms of retailing. The opportunity for the street dealer to under-report sales, over-report losses and skim profits is very obvious and in the absence of either audit or supervision I would expect that this would happen all the time. If JT were to attempt to have an IPO of the Disciples, I would certainly not buy shares in it on the basis of the spiral notebooks[4].

They do present some anecdotal evidence that the gangsters were not well paid that doesn't depend on the notebooks, but it's if anything even weaker. The simple fact that someone lives with his mother is not actually knockdown proof that he is strapped for cash; something like thirty per cent of young Italian men do it for the simple reason that it's better than cooking and cleaning for yourself. I also think it's quite na�ve to assume that when the gang members (who were, we shall remember, full-time drug dealers) asked Venkatesh to try and get them a janitorial job at the university, this showed that anything, even cleaning toilets on minimum wage, was a better life than the Gangster Disciples. I am hardly the most streetwise guy around, but even I can work out a couple of other possible reasons why a full time drug dealer might want a job which allowed him to wander round a university campus more or less at will. Students buy drugs[5].

Furthermore, even if we take the numbers in the notebooks as reliable, we are faced with the observable fact that crack dealers (even street soldiers) have expensive tastes and hobbies. Even leaving aside the question of trainers and jewellery (on which I have no hard data about ownership to argue against Levitt), it is an undeniable fact that even the most junior members of the Gangster Disciples were able to engage in the hobby of pistol shooting, a popular but expensive middle-class pastime which I would consider to normally be beyond the means of a burger-flipper at McDonalds. The non-salary compensation of JT's street dealers might be really quite high; access to guns, free admission to nightclubs, favourable deals on stolen goods and clothing, regular social events with local rappers, it all adds up and compares really quite well to the fast food trade, and as far as I can see the informal healthcare plan was also quite generous compared to most mainstream employers in that it covered family members and had substantial death-in-service benefits which would have been worth quite a lot in a neighbourhood that was not exactly Hampstead even for non-gang members. I find the seeming absence of any analysis of the non-salary component of compensation quite strange, particularly since the underlying work was done working with a sociologist who would at least have some analytical framework which one might use to measure the value of the benefit to the gang member of being in a gang and thus having some degree of status in a community where status mattered.

Once more, I'm not saying here that my own interpretation of the evidence is the right one, just that Levitt and Dubner are in the business of telling stories here (as opposed to adding up numbers), that other stories can be told, and that in my opinion the authors of Freakonomics try much too hard to give the impression that they are reporting facts when they are actually just telling stories. So anyway, on to the two central claims of the Freakonomics crack gang chapter. These are 1) that a criminal gang is reasonably similar to a capitalist business and 2) that a criminal gang is usefully modeled as a "tournament" reward structure where people are prepared to accept a very low initial reward and/or very unsatisfactory conditions because there is a small chance of rising to a level where there is a very high reward.

I'll deal with the second claim first; Levitt & Dubner describe the economics of a tournament as I have described above. It is about at this point that they make the silly claim about prostitutes and architects, and then go on to say that "cheerleaders from Nebraska" are prepared to move to Hollywood and accept low wages as waitresses (unless, I suppose, they choose to become prostitutes, given that this is apparently Adam Smith rewritten as Letters to Penthouse) because they dream of getting a big break as movie starlets.

Fair enough, but this implies a whole host of rather shaky looking empirically testable propositions. In general, do organized crime gangs recruit youngsters by promising them that if they knuckle down and perform well, they might end up being the top man? In existing organized crime gangs, is the whole thing set up so as to reward young chaps who want to rise to the very top and take over (was The Godfather completely deceitful on this point?) Subjectively as reported, do street-level crack dealers aspire to being gang bosses themselves? There is no evidence in the book at all from the gangsters themselves that they do (as opposed to Levitt and Dubner attributing this motivation to them in circumstances where other motivations are entirely possible).

There are a few organizational predictions that would normally drop out of this theoretical framework too, none of which look all that attractive. In most companies which operate "tournament" reward structures (paradigmatically, law firms structured as partnerships), there is a clear culture of "up or out". Employees are encouraged to leave if they are not making progress up the hierarchy. This doesn't seem particularly like the culture of a criminal gang at all; gangs don't like people leaving them. In fact, if you read between the lines of the description of the gang in Freakonomics, it seems pretty clear that JT spent a lot of time and effort on keeping his street-level soldiers contented with their lot. There's really no sense in there that the Gangster Disciples were all aware that life as a low-level gangster was meant to be rough and that it wouldn't be worth living unless they could get to the top. Quite the opposite, in fact.

In any case, there's a more fundamental oddity in the reward structure that Levitt & Dubner describe, and that's that according to Freakonomics, the street dealers of the Black Disciples were paid a wage by their leader (which went up or down depending on how dangerous the streets were), and didn't make any material part of their income from sales commissionThis is an extremely weird way to structure the remuneration of a sales force (actually, if I had my own Freakonomics hat on, I would almost say that it's so weird that it calls into question the credibility of all of the other claims made about the economic organization of the crack gang). It might be that Levitt & Dubner are speaking loosely and when they say "wages" they mean "total compensation including sales-related commission", but if they are, it's unforgivably sloppy; if there's one distinction that an economist ought to be persnickety about, this is it, so I think they aren't. If it really is true that the Disciples paid their gang members purely on the basis of hard work and/or bravery rather than revenue productivity, then this means that the "tournament" if tournament it is, is a pretty strange one. It's made clear at numerous points during the chapter that sales performance is almost irrelevant to this "tournament"; not only is there no commission, but promotion to the upper ranks is dependent on intangibles like "personality" and "leadership", coupled with actively counterproductive behaviour like starting turf wars and killing customers. Even Tyco didn't have an executive incentive scheme this warped.

But given that the organization doesn't reward behaviours which contributed to the economic success of the business, what's to make anyone think that the individuals in the organization valued economic success above anything else? If the Black Gangster Disciple Nation was prepared to reward and promote people who systematically undermined its main business activity by starting turf wars and killing for honour, then might this not be taken as good evidence that the Black Gangster Disciples cared more about turf and honour than they did about money? And if that's the case, then why do we assume that people joined the gang with the main objective of enjoying the lifestyle of its top members?

I am now shading into my critique of the other main assertion of the chapter on the Disciples, however; the claim that there is any meaningful analogy between a criminal gang and the neoclassical theory of the firm. I think that there isn't, but this argument merits a separate post (coming soon), in which I will attempt to see if other economic traditions can do any better and develop "The Heterodox Economic Theory Of The Criminal Firm". So long for now.


[1] Just to make it clear, the reason that I'm posting this here is that the quality of content on D2D has been right down the toilet recently, rather than anything more sinister with respect to Crooked Timber.

[2] (possibly Dubner's, but it is a figure of speech that I think sounds more like an economist than a journalist). In general I'm going to say "Levitt" as if he was the only author of the book; I will explain why I am following this policy in a future post.

[3] I of course mean by "you can pretend to have read the book at dinner parties" that when you are at parties you can pretend to have read the book, not that you can pretend to be the kind of person who goes to a party and sits in the corner reading Freakonomics.

[4]Actually, the fact that they are specifically referred to as "spiral notebooks" is not just a minor detail of stationary which gives descriptive colour to the story, and if Venkatesh had gone to the business school rather than the economics department to find a co-author someone might have tipped him off. If you get a job doing audit work at Deloittes, try going round writing things on a spiral pad and see how long it takes for someone to give you a bollocking. Spiral pads are not acceptable stationary for scientific lab-books or for auditing, precisely because it is easy to rip pages out of them. I don't have any evidence that this happened with respect to the notebooks in question, but it's the sort of thing that needs to be taken into consideration.

[5] As well as the fact that students buy drugs, there are other good reasons why drug dealers would want to have minimum wage jobs other than to supplement their meager incomes. The police are in general much more difficult to deal with if they believe that you are a "career criminal", so it makes sense to have at least a token stab at gainful employment; obviously you would out of preference want something with little or no supervision and hours that did not conflict with your criminal business. In his autobiography, the British celebrity villain "Dodgy Dave" Courtney mentions having been given this advice in the 1960s and getting a job as a binman as a result, and I suspect that gangsters in Chicago would have worked it out too.
1 comments this item posted by the management 11/30/2005 08:16:00 AM


This page is powered by Blogger. Isn't yours?




Links:

Bitch : Lab
Aaronovitch Watch
Balkanalysis
Perfect.co.uk
Maxspeak
Brad Delong
The Robert Vienneau blog

Political and philosophical heroes

Subcomandante Marcos
Will Rogers
Boris Vian
The English Svejk

RSS Feed:
This seems to matter to a lot of people

If you liked this "Daniel Davies" website, you might be interested in

"Danux", the web developer
The martial artist (and fan of extremely annoying Flash intros) from Blackburn
The Welsh political journalist
A Scouse computer programmer who collects Soviet cameras
"Danimal", the heavy metal drummer
Canada's finest recorder of radio jingles
More of the same, at the Guardian
A tailor's in Lampeter where Jimmy Carter once bought a hat
An advertising man who has written a novel about dogging (I think we sometimes get each other's email)
An award-winning facilities manager in Dubai
The son of the guitarist from the Kinks Update: he is apparently "balls-out motherfucking shit-dicked exxxstatic" to be included on a Kerrang magazine giveaway CD of Iron Maiden covers, which is nice.
"Fritz Gretel" from the Ramones film "Rock 'n' Roll High School"
The former presenter of the leading politics talk radio show on the Isle of Man, now a business change manager in the Manx government secretary's office
An aquarium curator in Sussex who keeps on scoring home runs like this (this is the first stable link I've found, but he is constantly kicking ass in acquarial terms)

If you didn't like this "Daniel Davies" website, then don't give up on the Daniel Davies industry completely!

An American "Christian Political Analyst" who has the same name as me
A student at Patrick Henry College
these two might be the same guy ...
"Scatter", the deceased Liberian gangster
A naked man stuck in a chimney in Wigan
A thug in Barrow



This blog has been going downhill since ...

August 2002
September 2002
October 2002
November 2002
December 2002
January 2003
February 2003
March 2003
April 2003
May 2003
June 2003
July 2003
August 2003
September 2003
November 2003
December 2003
March 2004
April 2004
May 2004
May 2005
June 2005
July 2005
August 2005
September 2005
October 2005
November 2005
December 2005
January 2006
February 2006
March 2006
April 2006
May 2006
June 2006
July 2006
August 2006
September 2006
October 2006
November 2006
December 2006
January 2007
February 2007
March 2007
April 2007
May 2007
June 2007
July 2007
August 2007
September 2007
October 2007
November 2007
December 2007
January 2008
February 2008
March 2008
April 2008
May 2008
June 2008
July 2008
August 2008
September 2008
October 2008
November 2008
December 2008
January 2009
February 2009
March 2009
April 2009
May 2009
June 2009
July 2009
August 2009
September 2009
October 2009
November 2009
December 2009
January 2010
February 2010
March 2010
April 2010
May 2010
June 2010
July 2010
August 2010
September 2010
October 2010
November 2010
December 2010
January 2011
February 2011
March 2011
April 2011
May 2011
June 2011
July 2011
August 2011
September 2011
October 2011
November 2011
December 2011
January 2012
February 2012
March 2012
April 2012
May 2012
June 2012
July 2012
August 2012
September 2012
October 2012
December 2012
February 2013
April 2013
June 2013
July 2013
August 2013
March 2014
April 2014