May 28, 2003

Daniel Davies Recommends Donald Mackenzie

Daniel Davies recommends the work of Edinburgh sociologist Donald Mackenzie...

Posted by DeLong at May 28, 2003 08:30 AM | TrackBack

Comments

OK, so I just read the paper Davies recommended:

"An Equation and its Worlds: Bricolage, Exemplars, Disunity and Performativity in Financial Economics" linked to from:

http://www.ed.ac.uk/sociol/Research/Staff/mcknz.htm

I found it to be an interesting (hey, even entertaining) history of the history and pre-history of the Black-Scholes equation.

I guess to make it sociology, there were words added on top of about how it was cool that the equation was essentially derived in several different ways by people with different assumptions and backgrounds. And the author apparently thinks we should all be astounded by the fact that it was not derived in one sitting by somebody who just sat down with a blank piece of paper and some mythical equations of economic force or something. Rather, people played around with bits of calculus that they didn't know quite how to deal with, did some curve fitting, made lucky guesses...all of the stuff I would have *thought* you would expect to happen. But I guess sociologists just live in a different world. :-)

Posted by: Jonathan King on May 28, 2003 10:43 AM

> But I guess sociologists
> just live in a different world. :-)

Yeah, the real one. >smack!<

;-)

Posted by: Kieran Healy on May 28, 2003 11:20 AM

Brad: thanks very much for picking up on my incredibly broad hints and giving this guy the plug he deserves in the economics community.

Jonathan: What's your point?

Posted by: dsquared on May 28, 2003 11:32 AM

dsquared writes:
>
>Jonathan: What's your point?
>

I thought I already said this, but I guess I had two points, and the second one wasn't clear.

a) I liked the paper, whose complete URL is:

http://www.ed.ac.uk/sociol/Research/Staff/optionpricing.pdf

I found it very easy reading compared to many papers I try to read that are well outside my field; that's obviously a strength.

b) Despite the fact that I liked the paper, it looked to me like a pretty straight-forward narrative of the history of the Black-Scholes formula for option-pricing. *My attempt* to understand this a work of sociology, or social history, as MacKenzie describes his work on his home page, was apparently a failure. In his abstract Mackenzie tells us that "several themes of potentially general importance are revealed". I assumed he used to word "revealed" to imply "these truths are not obvious from the surface". But then here are the deep points:

1) The formula was found through "creative tinkering" rather than some (rote?) rule-following strategy (of derivation, I suppose).

2) The tinkering was, in at least one instance, based on a previous model or example from a similar problem domain.

3) While everybody who worked on the problem was a fairly orthodox member of his profession, not everybody agreed about some of the theoretical underpinnings for the formula. As a sub-point, the fact that people who were not in agreement on these things *could* agree that the formula they came up with matched reality is a strong point.

4) The discovery of Black-Scholes has led to changes in the world, and especially in our perception of the truth of Black-Scholes. Specifically, people now use Black-Scholes to price options, so to some extent the fact that Black-Scholes describes option prices well is...well, not unexpected.

OK, so points (1) and (2) were really not news to me at all. This is just how stuff gets done in science, math, or any other creative discipline. Heck, the approach was even canonized by the mathematician Polya himself in "How to Solve It" decades ago. Point (3) is another basic observation about how science usually works. The deepest work often does turn out to be derivable from many different kinds of assumptions. Having thought about it again, I guess it's probably true that some people *would* be suprirsed by these points. That's fine, too, I guess.

Now that leaves point (4). Personally, I think that point was really the interesting one, but I'm not sure how strong the case is for it. One striking thing Mackenzie points out is that Black-Scholes was derived before there were really options markets where the technical conditions needed were especially close to being true. Then when the CBOT started options markets, people did have Black-Scholes in hand, and, presumably because people who priced options that way did well, the formula's use became wide-spread. I have a feeling, though that Mackenzie is trying to go for something deeper here. If so, I don't really understand it.

So there you have it. I should confess that I have *never* really understood sociology well at all, but I do keep on trying, from time to time. I liked this paper, but not, apparently, because it was a work of sociology. Or maybe I really don't understand what sociology is about.

Posted by: Jonathan King on May 28, 2003 12:51 PM

(THIS just MIGHT be the "magic bullet" boys and girls ;!)

Progress Marches On...

"Psychological Test Probes Minds of Psychopaths

By Patricia Reaney

LONDON (Reuters) - Scientists have adapted a standard psychological test that detects underlying prejudices to delve into the minds of psychopathic murderers....

...'But this is the first data and we need to do a lot more work before we can make almost life or death decisions about whether people should come out of prison or not,' she added.

In research reported in the science journal Nature, Gray and her colleagues evaluated the test on more than 100 violent criminals.

The researchers have also adapted the classic Implicit Association Test (IAT) to use on arsonists, pedophiles and rapists...."
http://story.news.yahoo.com/news?tmpl=story&cid=585&ncid=585&e=2&u=/nm/20030528/sc_nm/science_psychopaths_dc

(Throw some money--AT LEAST as much as they say they need, if not MORE--at those folks: FAST! We DESPERATELY need, immediately if not sooner, to "adapt" AND apply that sucker to ALL politicians, generals, intelligence "operatives", police officers, CEO's and executives in EVERY field, lawyers, economists, preachers, teachers, undertakers, auto mechanics...

Hell, JUST listing "the usual suspects" is wearing ME out ;-)

Posted by: Mike on May 28, 2003 01:01 PM

Interesting that the comments by people about sociological findings assert how "common sense" they are. However, if they think about it, the models in econ are far from common sense and make assumptions about human and firm behavior that are often, to say it nicely, quite over rational.

Also people seemed to have miss the bombshell at the end of the paper where Mackenzie says that Black-Scholes was right because people believed it would be right. Practice met theory - for a while any way.

Posted by: Karim Lakhani on May 28, 2003 02:31 PM

Interesting that the comments by people about sociological findings assert how "common sense" they are. However, if they think about it, the models in econ are far from common sense and make assumptions about human and firm behavior that are often, to say it nicely, quite over rational.

Also people seemed to have miss the bombshell at the end of the paper where Mackenzie says that Black-Scholes was right because people believed it would be right. Practice met theory - for a while any way.

Posted by: Karim Lakhani on May 28, 2003 02:36 PM

Hmm... mandatory screening of everyone to see if they have violent impulses? Ever seen "Minority Report," Mike? }:^) (Yeah, yeah... I know, crappy movie, but still!)

Maybe I need to read the Philip K. Dick version sometime...

Anyway, this Mackenzie fellow looks rather interesting. I'll have to look around his site some more.

It is interesting, though, how assets are priced the way the are because of the Black-Scholes model, though the Black-Scholes model was supposed to just describe how assets are priced. My econ prof was talking about it just a few days ago. (Maybe I should give him a link to Mackenzie!)

Posted by: Julian Elson on May 28, 2003 03:40 PM

Damn. I was under the impression that they were on track to give the rest of a reliable way to IDENTIFY the "psychopaths" among us, Julian, BEFORE they had a chance to 'hang out a shingle', 'earn their stars, badges, sprurs OR stripes', get themselves elected OR 'claw their way to the top of some other heap':

"...'this test may become an important tool for distinguishing psychopaths who are likely to commit extremely violent offences from those who are not," Gray said.'"

If I read more into the story than was actually there, I guess I'll have to ask you to chalk it up to my innately optimistic nature, Julian.

Now I'm depressed again. I hope YOU'RE happy....

Posted by: Mike on May 28, 2003 04:14 PM

I see, fair enough. Thanks for the clarification.

1) and 2) might be a reasonable description of science in general, but they're not really consistent with the self-image of a lot of economists. Economics is meant to be an axiomatic system where everything is deduced from expected utility maximisation. The fact that the Black/Scholes model didn't use such a framework led to it being rejected by a number of journals as being mere engineering. (This is actually a weakness of Mackenzie's article; the Black/Scholes paper is, IMO, very untypical of the way in which economics is already done). It's also worth mentioning that although the CAPM acts as an "exemplar" for the B-S model, option pricing theory is actually very different from CAPM type approaches and Black's attempts to derive one from the other haven't really gone anywhere in the literature.

3) is a point that needs to be made again and again with respect to economists; that there can be fundamental differences of opinion in the interpretation of agreed formulae. The theoretical gulf between Black and Merton (whether continuous-time arbitrage pricing reflects simple application of optimal control or whether it has an interpretation in terms of expected utility maximisation) goes to the heart of some of the deepest issues in economics, and the fact that it takes someone like Mackenzie to emphasise this point suggests that the profession is trying to paper over cracks.

4) didn't impress me either, and I didn't think that the evidence for "convergence" to Black-Scholes prices was all that strong. I thought that the idea that economic theories could be "performative" (for example, by being enacted into the "genuine hedge" rules) was very interesting, but knowing a bit about the subject myself, I wasn't terribly keen on the idea that Black-Scholes had nothing too it beyond providing a rationale for economic actors to converge in expectations. Basic Black-Scholes is a very powerful model indeed (for example, Mackenzie is repeating an urban myth of the finance profession when he claims that it can't handle transactions costs).

Posted by: dsquared on May 28, 2003 11:49 PM
Post a comment