Joseph Stiglitz contributed the "Information" entry to The Concise Encyclopedia of Economics. He wrote, "Since about 1970, an important strand of economic research, sometimes referred to as information economics, has explored the extent to which markets and other institutions process and convey information. Many of the problems of markets and other institutions result from costly information and many of their features are responses to costly information."
Later in the essay, he notes that "The standard theorems that underlie the presumption that markets are efficient are no longer valid once we take into account that information is costly and imperfect. To some, this has suggested a switch to the Austrian approach, most forcefully developed during the 1940s and later by Friedrich Hayek and his followers. They have not attempted to 'defend' the market by the use of theorems. Instead, they see markets as institutions that have evolved to solve information problems."
And (disagreeing with Hayek) "The fact that markets with imperfect information do not work perfectly provides a rationale for government actions."
Some of us would add that government actions are also imperfect. Many years ago, Harold Demsetz warned about "nirvana economics."
Last night, on 60 Minutes coverage of the bailout, former Treasury official Roger Altman mentioned AAA ratings given to novel credit instruments that no one really understood. The market vetting institutions had fallen behind the innovations.
At econtalk.org, Pete Beottke talks about the economics of disasters. He mentions that mainstream economics in the 20th century went the way of physics (theorems) when it should have gone the way of biology (evolution).
Now we hear that the credit instruments that no one adequately understood were designed by physicists that the investment banks had hired.