Last night I finished The (Mis)behavior of Markets by Benoit Mandelbrot and Richard L. Hudson.

Though it was published in 2006, this book provides an interesting background to the current financial mess. The main thrust of the book is that the assumptions underlying much of modern finance are invalid, and have been known to be invalid for many years.

In particular, Mandelbrot argues that the distribution of prices in the market follows more closely a fractal/power law rather than the standard assumed Gaussian/Normal distribution, and that there is a dependence between current prices and past prices, though not in a way that is useful for so-called "chartists" who try to predict today's price based on previous data.

The first point doesn't seem to be very controversial anymore. Many studies have shown that markets do not match the Gaussian/Normal distribution very well, particularly in the "fatness" of the tails - the markets show much higher proportion of large price shifts than would be expected under Gaussian conditions - even though most finance courses still teach the CAPM/Black-Sholes types models that are based on this assumption. The second point is more controversial, and the book does not fully present that argument in enough detail for me to judge it, nor does there seem to be sufficient agreement in other studies.

I was hoping this book was more of a middle ground between detailed (and usually near incomprehensible to non-specialist) academic papers and pop-science type descriptions, but the authors chose to go the pop-science route and drop almost all mathematical detail from the book.