Atul Gawande says something I believe to be true and something I believe to be false. What I believe to be true:

To figure out how to transform medical communities, with all their diversity and complexity, is going to involve trial and error.

What I believe to be false:

Getting our medical communities, town by town, to improve care and control costs isn’t a task that we’ve asked government to take on before. But we have no choice. At this point, we can’t afford any illusions: the system won’t fix itself,

Gawande has no concept of the relative ability of government and markets to deal with ambiguity. He thinks that markets are unable to adopt new processes without government pressure. He thinks that government is well equipped to experiment.

There indeed are examples of markets that do not evolve effectively. There are examples of government-led experiments that pay off. But mostly it is the other way around. The incentives work much better in markets. In markets, the tendency is to reward success and to punish failure. In government, failed programs persist, and success receives no special reward.

For the pointer, thanks to Peter Orszag</a. But shame on Orszag for taking the line that government-led change is the only way to address health care policy.