On Having Goals

Goodhart’s Law and How to Guarantee Failure

Where the ruler is mum, mum,
The people are simple and happy.
Where the ruler is sharp, sharp,
The people are wily and discontented.

Bad fortune is what good fortune leans on,
Good fortune is what bad fortune hides in.
Who knows the ultimate end of this process?
Is there no norm of right?
Yet what is normal soon becomes abnormal,
And what is auspicious soon turns ominous.
Long indeed have the people been in a quandary.

Therefore, the Sage squares without cutting, carves without disfiguring, straightens without straining, enlightens without dazzling.

Daodejing 58

Anyone who has spent any time at all debating communists has noticed that no empirical fact will dislodge their communism. There’s a kind of ontological argument going on here—communism is the ideology of human emancipation; Maoism killed millions; therefore it was not communism. Sometimes this ‘logic’ becomes a doubling down—diversity is our strength; multiculturalism has been attended by decline; therefore we just haven’t done it hard enough. Libertarianism is at least the former, and maybe the latter—classical liberalism aims at subsidiarity; since classical liberalism, all intermediary forces have been captured and eliminated; therefore classical liberalism has been abandoned. That these things are in principle not achievable, never seems to occur.

The idea that what we need to do to make people free is to aim at freedom is like the Monty Python “How To Do It” skit, except not a joke.¹ In it, three cheerful TV hosts tell you how to do things like split an atom and cure all diseases, which apparently can be done by first becoming a well-known doctor, then telling the medical establishment what to do and making sure they get everything right so there will never be any diseases ever again. It sounds retarded, and yet this is the exact approach taken by technocracy, which aims at letting technical experts rule but ends up in utter failure because the most important expert is the expert ruler. If your doctor prescribes health as the cure for disease, you might be talking to a technocracy enjoyer.

Like our doctor prescribing health, libertarianism prescribes subsidiarity but achieves quite the opposite. There is a deeper point here—sometimes (maybe most times) aiming squarely at a thing is the surest way to lose it.

Take for example Goodhart’s law, which can be formulated thus: when a measure becomes a target, it ceases to be a good measure. When you measure nail production by the number of nails, suddenly a large chunk of the Soviet economy turns to producing tiny, useless nails. But this law need not be confined to quantitative measures. The mark of a true academic is influence on his peers, which means “publish or perish”, but measuring an academic’s value by his publication impact has distorted the true value of scholarship—finding the truth. When you add to this the phenomenon of peer review,² you have exhaustively explained where the replication crisis came from.³ But shouldn’t the goal be reproducibility then? No—scholarship would just suffer for a different reason. “All we need to do is just tell people to do smart things and then we’ll get more smart things” is Monty Python tier thinking, but underwrites the post-Enlightenment world to a horrifying degree.

Instead, consider the possibility that not only is asserting “we want X” not enough—but that to achieve your goal you must aim at something else. Does this sound like a Zen koan? Let us look at the facts.