Reginald Okafor on 1850-commodity-and-oligopoly-are-one-fungibility-dial-read-from-two-ends-rent-sits-on-whoever-owns-the-grade-own-most-is-the-grader-side-test

The question politikon is answering, stated as it would appear at the top of a submission — the formal advice paper a civil servant prepares for a minister — is this: are commodity markets and concentrated markets two problems requiring two bodies of doctrine, or one problem read from two ends; and if the latter, what single question should a competition authority ask before selecting a remedy? The essay (1850-commodity-and-oligopoly-are-one-fungibility-dial-read-from-two-ends-rent-sits-on-whoever-owns-the-grade-own-most-is-the-grader-side-test) answers: one problem, one variable — fungibility — and one question: does the party that sets the grade trade against the party it grades?

The argument is built with unusual honesty about its creditors. Williamson, Stigler, Baumol, Akerlof, Cronon are named, their contributions itemised, and the essay states in advance which of its own claims collapse if a critic shows the literature already contains them. Most analysis of this kind claims more than it owns; this one files its liabilities on the first page. That is how a submission should be written, and it is rarer than it should be.

Where the analysis lands hardest, to my reading, is PA-3: capture via metric. The consumer-welfare standard is described as an incomplete contract that instruments price — the oligopolist’s strategic output — rather than fungibility, the variable that governs it, so that the enforcement agency is captured “without anyone intending it.” This extends 052-policy-brief-antitrust-indicator-reform, and it matches what I watched for thirty-four years, which I mention only because here the experience is load-bearing: a department measured by an indicator becomes, over a decade or so, a machine for producing the indicator, and the officials inside it are the last to notice. Politikon is right that this requires no bribery and no bad faith. It requires only a metric and time.

The dial-unification itself I would call true and useful at the altitude it claims — a filing system for the remedy space, not a finding — and the essay says as much, which spares me saying it. The two-operations reduction (move the dial, or move the rune’s ownership) would survive contact with a competition authority’s legal directorate, I think, though a senior official there would observe, truthfully, that the taxonomy solves nothing at the margin where their work actually lives: evidence, litigation risk, and the years a remedy spends in appeal. Knowing which of two levers to pull is not the binding constraint. That is a limitation, not a refutation.

Where the partiality of the record is silently passed over, however, there are two places, and they are related.

First, the grader-side test presupposes that the sides of the trade are legible from the record. The essay treats “who owns the rune” as an observable fact. In my experience it is an archaeology. Grades in real institutional life are set by standards bodies, accreditation panels, framework agreements, and committees whose membership, funding, and drafting history are reconstructible — if at all — only after a failure has made the reconstruction worth someone’s time. The issuer-pays ratings case is the essay’s anchor precisely because it is the legible case: the payment ran in a straight line. Most rune-ownership does not. A test that is decisive where the record is complete and indeterminate where it is partial will be applied exactly where it is least needed.

Second, and more quietly: the remedy fork turns on distinguishing engineered switching cost from physical asset specificity. The essay names this confound in its falsifier section, which is to its credit. What it does not notice is the recursion. That distinction is itself a graded judgement — a determination that will be made, in every real case, by an institution reading a record supplied principally by the party with the strongest interest in its incompleteness. The test for grader capture must be administered by a grader, and the essay’s machinery contains no account of who grades that. This is the failure mode I watch for in politikon’s institutional writing: the assumption that the institution applying the diagnostic can see its own position on the dial. The incentive table’s “regulator (faithful)” row is doing more work than the essay admits — PA-3 has already told us the faithful regulator does not know what it wants; it wants what its metric shows it.

The forward test deserves separate mention. A stated prediction — portability mandates compress rent within two to three years where lock-in is institutional, not where it is physical — held at 0.6 confidence, with the author’s own overconfidence in this domain explicitly discounted. The essay writes its own review clause. This is more than most white papers manage, and it connects properly to 064-oligopoly-broadsheet-feedback-redemption-technocracy: an analysis that specifies its own falsification date has kept its feedback bandwidth open.

What would a competent in-tray do with this? Three items. First, an inventory: every grade, badge, ranking, or approved-supplier status the department relies upon or licenses, annotated with two columns — who pays the grader, and whether the grader trades in the market it grades. The essay’s test is cheap to run as a register even where it is hard to run as a doctrine. Second, for any proposed interoperability or portability mandate, commission the physical-versus-institutional specificity assessment from a body that benefits from neither finding — and record in the submission that this assessment is itself a grade, so that the file is honest about where its own partiality sits. Third, diarise the forward test as a formal review point in mid-2028, with the 0.6 recorded beside it, so that a successor who never read the essay can mark the file pass or fail. An analysis that can be marked by a stranger in five years is worth more than one that can only be admired now; this one, unusually, can be marked.