A lot of contemporary political wrangling revolves around whether the free market should be unfettered and allowed to determine what lives and dies in the economy. The alternative is typically posed as "regulation," whether in the extreme form of soviet-style state control of industry, Japanese-style industry-government cooperation, or the European style of state-supported industries that are occasionally nationalized. The argument centers around the power of the invisible hand of mass action to make better choices than government bureaucrats, who are inevitably portrayed as less competent.
The power of the free market methodology in turn relies not on the supposed omniscience or wisdom of the individuals participating in the marketplace, but rather in the notion that masses of individuals, each acting rationally in their own best interest will add up to the whole acting in the best interest of the greater society that the market is embedded in. There turn out to be two significant problems with this whole idea. The second issue is that there is no proof of the belief that the mass of individual actions does actually add up to the best interest of the society as a whole. This is a concept that needs to be taken on faith, if it is taken at all.
More interesting, to me, and more important, is that work in the new field of neuroeconomics, and its companion field, behavioral economics, is undermining the other underlying premise of free market capitalism. This new breed of economists is demonstrating, definitively in many cases, that individuals almost never act rationally in their own best interest. A very simple case in point is that for almost everyone, now is a good time to buy stocks. It has been shown time and again that over the long term, dollar-cost averaging (buying stocks low and high, on a timed program) provides results better than all but the very best stock picker, and that the typical stock picker (i.e., 95%+) performs worse than random chance by trying to time market ups and downs. Unless you are 60+ and planning to retire in five years to ten years, you should be staying in the market, but that's not what's happening.
Another interesting result from this new field are some studies recently reported in the NY Times regarding performance when bonuses were at risk. Contrary to conventional wisdom, performance of cognitive (that is, non-mechanical) tasks degraded with higher levels of bonus pay, and were at their worst when the bonus pay was at around 30% of annual compensation - the highest level tested in the study.
These two things, taken together, put a significant stick through the spokes of the free market wheel. Not only do market actors fail to act rationally, increasing their risk-based (bonus) compensation worsens performance. This explains an awful lot about Wall Street. In effect, the entire basis of the free market system is a myth, and as a society we need to do something about it. I'm not personally in favor of centralization and five-year plans, but I think it's pretty fair to say that risk-based compensation needs to be capped, and the behavior of actors in the free market needs to be regulated so that these irrational excursions are kept under control.