The problem is: the only way to detect this, as the article points out, is for them to stop doing their job well intentionally. You either wreck the perfectly g

The problem is: the only way to detect this, as the article points out, is for them to stop doing their job well intentionally. You either wreck the perfectly good situation you have locally, or you find some other place where everything is terrible, and you compare the two. This is one of the reasons why scientific methods are hard to apply in non-experimental situations: you can't necessarily destroy the national economy 'to see if it would really work'. It's the same as informed consent in medical experiments, just on a larger scale.