First all, I agree with you that the author's recommendations on what should be done are pretty ridiculous--it seemed like he was just looking for an excuse to push his outdated Keynesian economic theories... that having been said, I agree with him that things people do are often irrational, not just 'less coherent.'

To me a rational act is one that provides the actor the greatest net benefit.

Most of the examples given in the book seemed to focus around one principal premise--people were primed with something seemingly unrelated to a decision that influenced them imperceptibly before they made the decision, causing their choices in the same decision at different times to vary widely depending on the priming influence.

Every decision doesn't have exactly equal benefits. This means means that some decisions are better than other and, by extension, that for each person there is an optimal choice for decision. The first question is how do we know what is optimal for each person. The second question is whether the priming influence is enough to change what the optimal decision actually is.

I believe the optimal decision comes down to a persons values. If I value economic benefits more than morality it's always in my best interest to do whatever will put more money in my pocket. This is an unrealistic example of an extreme, but it works for less extreme situations too--most people who value economic benefits have a limit on what they're willing to do to get money--at a certain point morality kicks in. That's not the point though, the point is that people have values. I think most people's values are fairly established - at least enough to prevent them from being changed from one moment to the next by influences as slight the priming examples mentioned in most of the experiments in the book. It seems to me that what happens is not that the priming influence actually change someone's values, but instead it temporarily blinds the person to their established values causing them to make a decision that they would later see as inconsistent with their values and less than optimal in providing them with a net benefit.

If someone can be temporarily persuaded to make a decision that goes against their values and doesn't provide them the greatest net benefit, I think the decision can safely be considered irrational. If we can predict when those types of decisions can be made, and I agree with the author, we can, there is a a level of predictable irrationality that comes with being a human that is more than just incoherence.