Radar Perene / Archive / science

Conflict of interest in reverse: what changes when whoever calculates the index sells no product on it

◦ Index methodology v2.2 (working papers with DOI). See the methodology.

Science

The most useful question to ask of any market analysis is not "is this right?" — it is "what happens to whoever wrote this if I believe it?". Across most of the industry, the answer carries a currency sign: the analysis lives in the neighborhood of a sale, and the reading that favors the product meets less internal resistance than the reading that gets in its way. No one needs to lie for this to work; it is enough that one of the two readings has the easier path.

This house's model inverts the position of the currency sign: the indices are published with open DOIs and no commercial financial product exists attached to any of them — the house manages no money, distributes no funds, sells no trading signals. The structural consequence is that there is no "convenient" reading: if a house index errs, there is no product to protect; if it is right, there is no product to promote. What remains at stake is the operation's only asset — the credibility of the record.

The classic conflict, described without villains

The industry's dominant design is well known and has its own treatment in this library, in conflict of interest in bank research: the institution producing the analysis is paid through the sale of the products the analysis opines on. It is not a scandal — it is an architecture, declared in disclaimers no one reads, managed by internal barriers of variable effectiveness, and rational for everyone involved, except perhaps the reader who takes the material as disinterested opinion.

The point of this piece is not to accuse that design — names, in fact, do not matter, because the problem is structural, not moral. The point is to note that the reader rarely asks what incentive architecture sits behind the ruler they are using. And different architectures produce different materials, even with equally honest authors.

What the inversion buys — and what it does not

A provider that sells no product on its own indices gains a specific freedom: that of recording what the ruler shows when what it shows is uncomfortable. The house's series carries documentary proof of that freedom — a published working paper whose conclusion dismantles the very thesis that motivated it, open revisions of findings that fell, and the policy of publishing the entire hit rate. None of this would be impossible under the classic design; it would merely be more expensive, and avoidable expenses tend to be avoided.

Honesty demands completing the account: commercial independence does not buy correctness. An index free of conflicts can be wrong through incompetence, bad data, or aged methodology — the mechanisms of this trail (retesting, self-assessment, open revision) exist because the absence of conflict replaces none of them. What the inversion buys is something else: the guarantee that, when the error appears, there will be no commercial reason to hide it. Is that little? It is the difference between an error that becomes a public revision and an error that becomes an ever-smaller footnote.

How the reader uses this in practice

The incentive architecture is verifiable by anyone, with no privileged access: it is enough to ask what the operation sells and to whom. When the answer is "products the analysis opines on", the material deserves the corresponding reading — useful, possibly excellent, and structurally partial. When the answer is "nothing beyond the record itself", the material deserves another ruler — and another suspicion, in fact: the operation must live on something, and the reader does well to understand on what. In this house's case, the answer is public: the open archive is financed by readership, not by allocation — and that distinction belongs on the checklist any market material should cross, next to the flags of financial pseudoscience.

The contrast, for that matter, is not binary: between the two extremes live mixed architectures — subscription-paid research inside institutions that also sell, indices licensed to product issuers, analyses sponsored case by case. The reader need not classify every variant; they need only locate where the revenue lives before calibrating the reading.

Frequently asked questions

Does commercial independence guarantee correct analysis?

No. It only guarantees that the error, when it exists, will have no internal sponsor. Correctness depends on method — retesting, revision, self-assessment — and those mechanisms must exist and work under any architecture.

Does the house have no interest at stake at all?

It has one, and it is declared: credibility. An operation whose only asset is the reliability of its record has a strong incentive to protect it — which includes admitting errors before others find them. It is an interest, but it points toward the reader, not against them.

How can one verify whether a provider sells products on its own indices?

By looking at what the operation offers commercially: funds, portfolios, signals, structured products referenced to its own rulers. The information is usually public, if not highlighted — its absence from the materials is, in itself, a datum.

Can't third parties use open-DOI indices to sell products?

They can — an open DOI does not control others' use. What the model guarantees is the house's position: none of the house's revenue depends on any particular reading of its own indices.

---

Continue the trail: The checklist the house applies before citing its own source →

House reading: today's reading is in the Diário; the archived episodes, in the Atlas.

Rereading a third-party report in light of the incentive architecture that produced it is an exercise the house does in conversation, over concrete material.