Friday, October 2, 2009

Benoit Mandelbrot on the Efficient Market Hypothesis

Below, Jonathan Authers of the Financial Times interviews the eminent mathematician Benoit Mandelbrot1. Sorry about this video automatically launching into play. I didn't see an embed feature so I cut and pasted the source code here. Update: Thanks to Alex Garcia, here's the embedded version, via YouTube:

1I wonder if Dr. Mandelbrot knows that Jonathan Coulton made this song about him, Mandelbrot Set. Here are the lyrics to the chorus:

Take a point called Z in the complex plane
Let Z1 be Z squared plus C
And Z2 is Z1 squared plus C
And Z3 is Z2 squared plus C and so on
If the series of Z’s should always stay
Close to Z and never trend away
That point is in the Mandelbrot Set


JK said...

Mandelbrot is a genius. Looks like I am going to have to buy that book of his sometime soon. His alternative theory supposedly utilizes a form of Pareto's law. Paging Tim, paging Tim.

DaveinHackensack said...

Yeah, that might be a book worth reading. Incidentally, I had an exchange with a rather pompous EMH adherent yesterday on Megan McCardle's blog in this thread if you want to check it out. What a tool.

JK said...

Your perception may have been different, as my eyes are scarred from reading many blog polemics, but John Thacker honestly didn't seem like too much of a tool to me, I just read strong disagreement, however wrong he was. BulgingBracket was a tool though.

I don't think Mandelbrot's theory exactly lined up with what you were saying either, however. Value investing can be little more than a form of EMH because value investing often makes assumptions about the what the intrisic value of a security should be and discounts any number of large variables -sweeping them under the rug, as Mandelbrot would say with Taleb applauding. He did say in Part II of the interview there is nothing that keeps any given stock price from immediately falling to zero (or any other big move, and this is based on the timing of the last big move - patterns that fit into fractal geometry no doubt). The main beef Mandelbrot has with EMH seemed to be the distribution model, as his predicts much more downside to the market. His model doesn't seem much less 'academic' at first glance. I would probably have to get his book to adequately describe it though.

You were hammering home some good points. They say that experimental physicists are there to keep the theoreticians honest. If theoreticians ruled the world then we should have never developed flight, as it was ruled mathematically impossible by some elite minds in the 1800s.

DaveinHackensack said...

You could be right with your characterization. It's also true that value investing does rely, to some extent, on a weak form of the aspect of EMH that states that all information is priced in to stocks. Strong EHM adherents argue that this is always true, so it's pointless to look for bargains; value investors argue that when they buy the stocks, all the information is not priced in, but eventually most of it will be, and the stock's market price will approach their estimate of its intrinsic value. Weak EMH -- markets being mostly efficient, most of the time -- makes sense.

The aspect of EMH/MPT that pretends returns are normally distributed, or that prices always move continuously makes no sense -- unless, as Mandelbrot slyly notes, you ignore the huge, discontinuous moves that have the biggest impacts.

Magic Formula Investing said...

Here is the video on youtube

DaveinHackensack said...

Thanks a lot, Alex. BTW, I sent you an e-mail re a sneak preview of the new financial info site we discussed over the summer.