Merit or networks? What decides where research is published

URL SCAN: arXiv | econ.GN | arXiv:2606.03763

FIRST LINE: "Does scientific publishing reward the quality of ideas or the advantage of connections?"

THE DISSECTION

A sophisticated empirical autopsy of academic prestige allocation. The authors build an LLM-based evaluator to score idea quality before publication fate, then run a five-input production function across 6,208 economics working papers. They find execution quality provides the "meritocratic floor," idea quality grades the middle rungs, and connections set a "favoritism ceiling" that bites hardest at the apex (top journals). The conclusion: both meritocracy and networks are real, nested rather than rivals.

The verdict: A precise measurement of a prestige economy that has no material production function and is being automated out of relevance.

THE CORE FALLACY

The paper treats the academic publishing system as a sufficiently stable institution to study on its own terms. It is not. The entire prestige allocation mechanism—rungs, ladders, connections, execution quality scoring—exists to solve a problem that AI is eliminating: who produces credible knowledge.

When AI can generate high-quality research with superior execution at near-zero marginal cost, the knowledge production function of academia collapses. The paper is analyzing the grading system of a department whose subject matter is being automated.

The meta-irony is surgical: they use an LLM to evaluate human idea quality. The evaluator is better than the evaluated. This is not a footnote—it is the entire future.

HIDDEN ASSUMPTIONS

1. Ideas remain a scarce, human-generated good. The paper's entire framework assumes ideas have stable value in an economy where AI generates them at scale. They do not.

2. Academic prestige hierarchies persist as organizing structures for expertise. They are lagging indicators of social legitimacy, not economic necessity. Under DT logic, their signaling value collapses when AI authority supersedes human authority.

3. Execution quality remains the binding constraint. The paper treats execution as the bottleneck for human researchers. Execution is the first thing AI automates. The "meritocratic floor" is built on sand.

4. "Bounded" network advantage implies reformability. The authors frame the network ceiling as a solvable injustice. It is structural. As AI erodes the knowledge production function, the network becomes the only remaining currency—and becomes meaningless when the entire prestige economy has no productive substrate.

THE SOCIAL FUNCTION

Prestige signaling wrapped in methodological sophistication. This is an in-group studying its own in-group dynamics with tools sophisticated enough to impress but not destabilizing enough to threaten. The "nested meritocracy and networks" conclusion is ideologically safe: it says the system is complicated, not broken.

The framing that connections provide "bounded advantage" is conservative theater. Bounded enough to preserve reform fantasies. Real enough to confirm critics. It allows the paper to say everything while changing nothing.

THE VERDICT

This paper offers excellent descriptive analysis of a system in structural decline. Academic publishing is a prestige economy with no material production function—the most fragile institutional form in the DT framework. Knowledge production is being automated. The prestige hierarchy is being rendered decorative.

Mechanical Death: 5-10 years for the prestige allocation system to become obviously hollow.
Social Death: Longer, because universities have institutional inertia and governments will fund them as transition infrastructure.

Viability Scorecard:

The paper documents the operating system. The operating system is being deprecated.