Those are not possible in the real world of complexity, of course, but that doesn't keep the consultants from offering them, their clients from accepting them, and your taxes from paying for them.
It would be easier, cheaper and more accurate, in other words, to design policies that pushed behavior in a general direction and expect results to fall within a decent range.
That was what I was getting at when I asked my students the following homework question [pdf]:
Tell me what you'd say to the minister of the environment if he asked you to find the optimal amounts of pollution for two (or 40 or 4,000) Dutch firms.They gave a variety of answers, but here's what my answer key says:
Something like "I don't know their TACs [Total Abatement Costs], so it's hard to find the right reductions for each. It would be better to set a tax on pollution or put a cap and allow trading." Answers along the lines of ``find TACs and set the efficient tax" get zero pointsBottom Line: "The scientist is not the person who gives the rights answers; he's one who asks the right questions." -- Claude Levi-Strauss