Rather, economics attempts to predict behavior in general, with all other factors held equal. Yes, some managers may choose to forego income in exchange for more social prestige. But do all managers, as a general rule, make the same choice? If there are only a few exceptions to a general trend, does it make sense to call for large-scale government intervention to adjust for the effects of a few exceptions?
The kind of critique outlined in the article makes another, more substantial error. While recognizing that certain aspects of the market may not always function with perfect efficiency or consistency, they seem to assume that the government does not suffer the same flaws as the market. Those calling for regulation assume that the government is capable of making perfect, instantaneous, highly detailed decisions to address the shifting and complicated imperfections in a market economy. Anyone whose ever had a run in with any kind of government committee or a few bureaucrats should know better than to make that assumption.
Any market economy in the real world will never perfect, as even Friedman would admit. That doesn't automatically mean government is the best or first solution to turn to. Using the government to correct minor market imperfections is like using a rusty machete to fix the stray whisker you missed during your morning shave.
Though I admire Friedman I have always thought he put too much weight in rational behaviour. I mean, I don't think we're talking about the odd person here and there. I think people behave irrationally often enough to designate people as irrational (and therefore rational behaviour cannot be expected or inferred).
I think you might see rational trends of behaviour that appear to represent reliable norms but I think the explanation for that is more because of cultural influences than it is from the human acting rationally himself. Cultural influences and trends can also be irrational for long periods though (war is an example). I think rational behaviour can be used as an abstract to offer an 'ideal economy index' against which you can measure the current economy but nothing more. People aren't rational thus the 'ideal model' cannot be used to predict perfectly what they as a whole will do.