Elites Must Rule

While I’ve spent much of my life doing institution/mechanism design, I’ve only lately come to see that, at least on prestigious topics, most people want relevant institutions to take the following ideal form:

Masses recognize elites, who oversee experts, who pick details.

While experts are known by other experts to be knowledgeable and skilled regarding particular topic areas, elites must be widely seen by the masses as having connections and features that are admirable and appropriate for leadership and governance roles. (More on experts vs elites.)

For example, with democratic governance the voter masses elect politician elites, who appoint mixed-elite-expert agency heads, who oversee expert agency personnel, who choose details. And debates on who should get to vote can be seen as debates on voter eliteness.

Long ago, it was seen as sufficient for the masses to recognize the natural eliteness of kings and aristocrats, who descended from or choose each other. You might think we are past that now, but academia and medical regulation both work this way. That is, we are all supposed to recognize that top academics and doctors are sufficiently elite to be trusted to run those orgs, and to pick their own successors, all without substantial outside oversight. And actually, this is how regulation works in many prestigious areas in many nations; the elites of each area are given a free hand to run those areas, and choose their successors.

If one sees the typical rich investor as sufficiently elite, then one can accept investors overseeing CEOs who oversee middle managers who oversee line workers, in a progression from relatively elite to relatively expert. The main reason that people object to this arrangement is that they are reluctant to see investors as sufficiently elite; they prefer instead to have government officials decide which firms get funding, and to have workers elect their managers.

Most people don’t like direct democracy, and dislike it more more when ordinary voters have more influence over the proposals on which to vote. And most people aren’t actually that comfortable with legal juries, unless jury choices are greatly limited by elite judges, and advised by expert lawyers. The U.S. uses juries today much less than it used to, and most nations have little use for them. It seems that most people instead want something closer to the ideal form described above.

Back in 2003 my Policy Analysis Market project hit the news, and was immediately killed. While the loudest complaints against it, of sabotage and price manipulation, had little basis in fact, this oft mentioned criticism was solid: such prediction markets would somewhat displace prestigious intelligence elites with more ordinary people. Markets producing better decisions via more accurate estimates was not seen as sufficient justification.

In the CIA today they tolerate internal prediction markets, but only under the rule that market estimates are never to be cited in official reports, which are the coin of prestige in that realm. Having prestigious reports cite prediction markets would let low level CIA folk somewhat displace CIA elites.

Wariness of many of my other institution ideas, such as tax career agents, life maintenance orgs, and crime vouchers, might also be attributed to their more overtly distrusting elites in various ways, and displacing the usual elites somewhat by for-profit firms and financial investors. And these are parties that many see as insufficiently elite. Producing more cost-effective outcomes is not seen as sufficient justification.

While I will continue to try to persuade people to weaken this constraint, and to consider institutions that rely less on our just trusting prestigious elites, I will now also try to take this seriously as a design constraint, and design institutions that conform better to the ideal form described above. I’ll start in my next post.

Added 1Sep2023: In this post, @VitalikButerin tested my "masses recognize elites" & "elites oversee experts" features against 5 other features in capturing what is an institution. They had #1 and #3 strength correlation, out of the 7 total features.