Public Choice Links, 5/9/2026

Alice Evans on sub-Saharan Africa; Daniel Yu on profit-seeking vs. non-profits; JFV on bureaucracy; Ilana Redstone on AI and public discourse

Alice Evans writes,

Why did Sub-Saharan Africa’s earliest sites of political centralisation become so fragile? I suggest that the region’s embrace of Islam has been distinctive.

In Eurasia and North Africa, Islamic armies burst out of tribal Arabia and conquered preexisting, often ancient, imperial structures built on extracting agricultural surpluses from fertile lands. In Sub-Saharan Africa, however, the methods and ideology of Islamic governance could not overcome Sub-Saharan Africa’s obstacles to constructing enduring territorial states. Islam could not make African soils richer nor its terrain easier to control.

But Islam was nevertheless a force for state-building. The distinctively African-Islamic solution was to build states by selling salt, gold and slaves to the Middle East in exchange for warhorses (and other goods). Since these states were largely organised around man-hunting rather than steady control of territory, they did little to build law and order, public goods, or national cohesion.

I am reminded of the distinction between stationary bandits and roving bandits.

Daniel Yu writes,

A successful commercial firm does something no NGO can: it issues “cash transfers” to a large group of people every month, indefinitely, funded by the market rather than donor whims. Indeed, private sector growth is the key to the structural transformation required to create hundreds of millions of jobs. No rich country today has become wealthy through the intervention of NGOs.

Pointer from Tyler Cowen. Non-profits have been getting some bad press lately. Allegations of widespread fraud. Then the Southern Poverty Law Center accused of providing funding to racist groups in order to maintain the need to fight such groups. As you know, I have been wanting for a long time to knock non-profit organizations off of their pedestals.

Jesús Fernández-Villaverde writes,

The real issue is that the [hypothetical socialist commonwealth] would need to organize large bureaucratic organizations. Without them, it cannot develop and deploy vaccines, MRIs, scanners, and the rest. The need to scale is the key mechanism at play, not who owns the property.

And, because of their scale, these large bureaucratic organizations will suffer the type of problems that critics of “capitalism” attribute to “capitalism.” The organization will be impersonal and alienating, and inefficient due to career concerns, asymmetric information, conformity effects, and internal politics.

…Those who deny the need for large bureaucratic organizations are living in a fantasy world. I am pretty sure the day they are told they have prostate cancer, they will run to their closest large bureaucratic organization for treatment.

Those who deny the problems of large bureaucratic organizations, and how deeply irresoluble those problems are, have not seen how not-for-profits work. I have never seen more acrimonious fights than within not-for-profit organizations, where some shared sense of the common good unites members. The fights are fierce precisely because profits play no role.

Pointer from Alex Tabarrok. Socialism works below the Dunbar number, because people can police one another informally. But if you want nice things, you need a society that is larger than the Dunbar number, and at that point you need written rules, hierarchy, and organized division of labor.

Ilana Redstone writes,

Imagine an AI that, when asked about a politically charged topic, surfaced its assumptions the way a good academic paper is supposed to surface its priors. This response treats disparate outcomes as evidence of discrimination — a contested causal claim. This response assumes that historical injustice creates ongoing institutional obligations — a contested normative position. This response frames the question in terms of group identity rather than individual circumstance — a contested explanatory choice.

It would improve public discourse if AI did this. It would improve public discourse if humans did this. But I fear that when it comes to dysfunctional public discourse what appears to be a problem on the supply side is more of a demand-side problem.

substacks referenced above: @

Many of the Problems AI Creates for Democracy Can Be Solved With This Basic Change

@

Why Are Sub-Saharan Africa's Earliest States now Epicenters of Terrorism?