2021-03-09

On Radical Reforms, Technocracy and Seeing Like a State

In his latest post, Scott writes on the success of radical top-down reforms, contrary to his previous writing on the dangers of radical top-down reforms.

This leads to confusion when he approaches the Acemoglu et al. paper The Consequences of Radical Reform:

I think my real concern here is that someone might use this paper to support some sort of far-left reform, saying “come on, this shows that reforms work better than leaving institutions in place”, when an alternate lesson is “capitalism works better than not-capitalism”.

But how do we know which lesson is appropriate here? Scott concludes:

maybe the moral of the story is something like - replacing stagnation and entrenched interests with good reform is good, and with bad reform is bad. Which sounds obvious, but I do think that considerations of “is this potentially challenging a carefully evolved system of traditions?” is less important than I originally believed.

This is a deeply unsatisfying conclusion to an otherwise excellent series of posts. As a low-hanging counter example: The Vietnam War was at least ostensibly about fighting off the expansion of communism. That sounds like a good reform, but it went horribly wrong.

Is the lesson that good reform is good, but war is so bad that it’s altogether a net negative? But the original paper is about countries subjected to Napoleon’s conquest, so this same lesson ought to hold. Maybe war just looks better in retrospect.

The more fundamental confusion here is about trying to draw a clear distinction between top-down and bottom-up systems, when in fact the two work together and no clear line can be drawn.

Consider the example Scott gives where France invades top-down and establishes bottom-up free markets in a country that goes on to experience outsized economic growth. Who exactly is this a win for? There are endless examples with arbitrarily convoluted dependencies:

A small group of Founding Fathers top-down determine the governance system for the American colonies, and they choose democracy, the most bottom-up system to date.
A committee top-down designs the LSAT as an entrance exam, creating a bottom-up process for any applicant who wants to be judged fairly. Those students go on to become technocratic lawyers who pass top-down judgements.
Bottom-up competition allows one CEO to emerge as the dominant titan of industry, she then uses her power to top-down determine future product lines

This last example is a riff on Coase’s classic The Nature of the Firm, which asks why all of our free-market companies are run in a totally centralized fashion by an executive leader or small governance board. At the other extreme, one might ask today why today’s communist youth movements seem to favor decentralized governance. There ends up being a good answer in both cases, but the question remains: is this a win for top-down or bottom-up systems?

Or if these abstract cases are boring, consider some more realistic examples closer to home:

Elon Musk is the Ayn Randian hero who privatized space travel… funded by contracts from NASA, a federal agency.
Open Philanthropy is a technocratic organization that unilaterally determines the allocation of millions of dollars… due to the patronage of Dustin Moskovitz who got rich starting a company and competing in free markets.
Donald Trump’s 2016 victory was a populist movement… funded by corporations and hedge funds.

Who’s to praise (or blame) here? Which system deserves the credit? Is the whole debate pointless?

Tyler Cowen would say no, we just have to get more specific: “Earlier in history, a strong state was necessary to back the formation of capitalism and also to protect individual rights… Strong states remain necessary to maintain and extend capitalism and markets.”

In this view, it’s not about reforms vs tradition or mechanism vs judgement or anything else. It’s just that very specifically capitalism, markets and individual rights are good, and states are justified in using just enough coercion to ensure those systems remain healthy. Cowen’s own thoughts are a bit broader, suggesting that state capacity is also critical for “infrastructure, science subsidies, nuclear power… and space programs”.

Does this mean that top-down rule is always justified in the service of future bottom-up freedoms? Perhaps, but I would be shocked to find anyone still willing to justify the US invasion of Iraq, although it was ostensibly in the name of promoting democracy. So perhaps it really just is that war is very bad, so much so that it offsets potential gains from imposing improved institutions.

Still, I don’t think any of this can be taken as a totalizing framework.

Consider the entire field of mechanism design, of which Weyl is a prominent member. While “free markets” and “democracy” might just feel obvious, the specific mechanisms involved are subject to judgement. There are many ways to aggregate popular preferences into a collective decision, and voting theory remains an active area of research. If Weyl designs a voting mechanism, is that judgement? Is it mechanism? Bottom-up or top-down? The whole dichotomy falls apart.

Also consider the not-so-distant future where many mechanisms for bottom-up aggregation do not even have the flavor of democracy. Google Search can be thought of as an aggregation mechanism. It takes user-generated data, and synthesizes it into a centralized model which makes decisions. How is that different than voting? Is it less democratic? Less populist?

To sum up my views:

Democracy and free markets are generally worth promoting.
Some costs, such as the horrors of war and other violations of individual liberty, may be too high a price to pay.
Even if the ostensible aims [1] are good, top-down enforcement is simply ineffective in some scenarios (Vietnam War, Invasion of Iraq).
In other cases, the distinction between “mechanism” and “judgement” is simply unclear, and there is no guarantee that all forms of bottom-up evolution are as effective as democracy + free markets.

At this point, you might ask: why even make sweeping statements? Why not just get specific and leave it at that?

In the rationalist tradition, Bayes’ rule dictates the consideration of an outside-view. The point is not to say “X is good because of reason Y”, but to say “Things in the class of X have historically gone well, and this is our prior. The specific reason Y counts as evidence, and can be used to update that prior”. This sometimes gets you into trouble when it comes to establishing an appropriate reference class, but it’s still a useful technique.

So if this whole debate sounds silly, that’s fine so long as you don’t take it as a serious and literal attempt to figure out if one style of governance is always right. James C. Scott’s book had the wonderful and appropriate subtitle “How Certain Schemes to Improve the Human Condition Have Failed”. The point isn’t to hit you over the head with one case after another of high modernist failure, it’s to understand the failure modes, understand the prior, and try to do less poorly in the future.

The Scholar’s Stage post is particularly underappreciated. His argument is basically: in the past, tradition was better than technocracy because we had time for slow cultural evolution. Post-industrial revolution, the world is moving too quickly for useful traditions to establish themselves: “The traditions are gone; custom is dying. In the search for happiness, rationalism is the only tool we have left.”

Footnotes
[1] As I understand it, the Invasion of Iraq also had several not-so-good actual aims.