Knowledge and the Wealth of Nations by David Warsh
Warsh provides a much-needed follow-on to Heilbroner's Worldly Philosophers (which stops after Schumpeter). Focusing on Technology (innovations and knowledge), Warsh starts with Smith and follows the economic treatment of Technology as it changed from 1776 to 2006.
Although best read after having taken good classes in introductory micro- and macro-economics (as Warsh glosses over many of the ideas not central to the subject of Technology), the book had enough interesting stories to keep its pace. Discussing Marx, he forwards a parable from Lafargue (page 65):
"A talented painter tries again and again to limn the picture which has formed itself in his brain; touches and retouches his canvas incessantly; to produce at last nothing more than a shapeless mass of colors; which nevertheless to his prejudiced eye seems a perfect reproduction of the reality in his own mind."
As indicated by the ascerbic and just treatment of Marx, Economics itself falls victim to the vicissitudes of Technological Change. Pulling a page from Kuhn, Krugman describes what Economics lost as it changed from Literary (19th century) idea transmission to the rigorous math of Modern Economics:
I suddenly realized the remarkable extent to which the methodology of economics creates blind spots. We just don't see what we can't model.
Literary Economics used simple models of the world typically consisting of just two things: capital and labor (money/equipment and people). As the book progresses to the end of 20th century, this dichotomy of Goods becomes a trichotomy: Ideas, Things, and People. Each of these lives in a two dimensional space of Rivalry and Exclusion. Market-based outcomes work well for rival, excludable goods (like my car that only I use (rival) and my insurance company which will sue you were you to steal my car(excludable)). According to Warsh Market-based outcomes work less well for all other points in the two-dimensional space.
Libertarians frown at the thought, but rival/non-excludable (like equal priority processes on a UNIX server), non-rival/excludable (copyrighted software), and non-rival/non-excludable (GPS) all have a higher susceptibility to markets not providing the optimal number of goods. Unfortunately, without the math behind the work, I don't know why rival/non-excludable goods like processor time on a shared box would market fail (more reading required). Microsoft and GPS look like strong market failures because of the vendor-lock for M$ and the high initial deployment cost of GPS satellites.
All in all, very readable and enjoyable.