Disclosure futility

Walk into any chain restaurant and the calorie counts are printed beside the prices. Factories file emissions reports; suppliers publish audits of their supply chains; cigarette packs carry warnings. Every one of these rests on a single idea: that the root cause of a bad market outcome is information asymmetry (one party to a transaction knowing more than the other), and that the cure is to make the hidden information public. If people knew the full consequences of their choices, the reasoning goes, they would choose differently. For many markets that reasoning holds. It works for search goods, whose quality a buyer can verify before purchase, and for simple experience goods, whose quality becomes clear after one use. Tell a shopper the thread count or the crash-test rating, and the market corrects itself. It breaks down precisely where the stakes are highest. In Private-Systemic Tension domains — the domains where a private gain and the welfare of the shared system pull in opposite directions — disclosure alone cannot close the gap. Three independent results from behavioral and information economics, taken up in the sections that follow, show why: when the people receiving the information face temptation or neurochemical dependency, telling them more changes nothing. A further possibility, which this chapter treats as an open empirical question rather than a settled charge, is that disclosure survives politically because it is easier to enact than structural reform.

A smoker who reads that cigarettes cause lung cancer already knows, and smokes anyway. The warning is not disbelieved; it is simply overpowered. Once a substance has rewired the brain's reward chemistry, the craving speaks louder than the fact. Formally, when an agent's utility function carries an addictive state variable, the response to new information is non-monotonic — it does not shift steadily in one direction as the agent learns more. The Bernheim-Rangel framework¹, which models addiction as cue-triggered, state-dependent choice, gives this its structure: the decision to use is set off by an environmental cue, not by a deficit of knowledge. That locates the effective lever, and it is not disclosure. Disclosure supplies information; the cue remains. The intervention that bites is removal of the cue that triggers the addictive state. For any substance with neurochemical lock-in — brain chemistry that sustains dependence regardless of what the user knows — a disclosure mandate is structurally incapable of changing behavior at scale.

Addiction is the hard case. Temptation is the quieter one, and disclosure fares no better against it. Picture a dieter who asks the waiter to take the dessert menu away. The dieter is not short of information about the cake; the point is to remove the option, because its mere presence on the table exacts a cost — the effort of resisting — whether or not the cake is eaten. That is the shape of a temptation preference: the agent prefers a restricted menu to a full one. Full disclosure does the opposite of what such an agent needs. It expands the information set while leaving the choice set untouched. The tempted agent now knows the full cost of the Hollow Win and still cannot resist it. The Gul-Pesendorfer framework² draws out the sting: disclosure can raise the welfare cost rather than lower it, because the agent bears the disutility of knowing on top of the disutility of yielding. In Private-Systemic Tension domains, then, a disclosure mandate can leave the tempted agent worse off than silence would.

The sharpest version of the point comes from Lipnowski and Mathevet³. Consider an information designer — a party who gets to choose how much to reveal — facing a tempted audience. Their result is that revealing everything is welfare-inferior to an optimal partial disclosure: the designer who strategically withholds produces better outcomes than the one who tells all. For policy this is an uncomfortable finding. Full mandatory disclosure, the standard regulatory instrument, is formally worse than a well-designed partial-disclosure regime in exactly the domains under discussion. And it points ahead to a political puzzle: if full disclosure is the weaker tool, why is it the one governments reach for? The answer, taken up below, is that it satisfies the appearance of action while remaining suboptimal in substance.

Set neurochemistry and temptation aside, and disclosure still fails — this time for a reason that owes nothing to psychology. The Missing System Theory holds that system welfare (W) is not a function of the two parties' payoffs⁴. The coordinate that would register the health of the shared system is simply not in the payoff space. Disclose more about those payoffs — revenues, costs, margins, supply chains — and you add data in the wrong dimension. However complete the bilateral record becomes, it never contains the system-welfare state, because that state was never one of its entries. Disclosure operates inside the payoff space; the problem it is meant to solve lies outside it. This is not a behavioral limitation that better information could overcome. It is a structural feature of the information environment itself.

Put those results together and an uncomfortable corollary follows: governments may prefer disclosure precisely because it is ineffective. A disclosure mandate looks like action. It generates compliance costs that can be held up as proof of seriousness. And it leaves the underlying game intact — it does not alter the payoff matrix, does not break any Private-Systemic Tension axiom, does not change what either party stands to gain. This is not an accusation of cynicism against any particular official. It is a structural prediction. The political equilibrium selects for interventions that are visible but weak, because the interventions that would actually work — production bans, fiscal restructuring, functional regulation — impose real costs on Party B, and Party B is frequently the government itself or the constituencies it depends on.

None of this makes disclosure useless everywhere. It works at the edges — for products with low neurochemical lock-in, high visual salience, and a pathway for producers to reformulate. A front-of-pack warning label on an ultra-processed food changes purchases; a sugar tax nudges manufacturers to cut the sugar. But look at how those cases actually work, and the mechanism is not the classical one. The label succeeds through salience, catching the eye at the moment of choice, not through reasoned information transmission. The tax succeeds by restructuring producer incentives, not by educating the shopper. These are behavioral-design and fiscal measures wearing disclosure's clothing. Which sharpens the political point rather than softening it. When disclosure is chosen over structural reform in the domains where the tension is most severe, it is a revealed preference for the weaker instrument exactly where the stronger one is needed.

Disclosure arrives too late. By the time the information is public, the decision has been made and the payoff structure has already rewarded the Hollow Win. Decision Accounting intervenes earlier, at the one moment when the outcome is still open: the record is created before approval, while the decision can still be revised⁵. At that point a material decision has to carry a contemporaneous basis — who decided, what, when, where, and why; the named authority; the evidence; the uncertainty; the Field 16 prediction; the communication position; and the Field 17 system-welfare impact, among others of its seventeen fields. The first readers of that record sit inside the lawful governance chain — management, counsel, compliance, audit, the board, or supervisors, as the setting requires. External Conflictoring agents reach it only unevenly⁶, through examination, discovery, or the public record, as the Conflictoring chapter details. What changes behavior is the prospect of that review: several independent lanes with different powers, each able to reconstruct the governance state at the moment of decision, grade the prediction, test the record for gaming, and feed an enterprise learning loop that checks what was written against what later happened.

Two very different critics — the FASB (the board that writes US financial-accounting rules) and the short-seller (an investor who profits when a security's price falls) — arrive at the same rule: welfare exposure disclosed without a plan to reduce it is bad disclosure design. A number that only frightens is a liability, not a contribution. One can imagine the disclosure such a rule would call for. A company might report the industry's β_W, the estimated annual system-welfare loss, its own share of industry revenue, and, from that share, an estimate of its implied contribution to the industry's loss. No such architecture exists today — not as a FASB or IASB (its international counterpart) requirement, not as an accepted accounting standard, not as a peer-reviewed protocol. Were it ever adopted, it should never travel alone. It should carry the game-change plan beside it: what the company is changing inside its own operations, what industry-wide change would cut the loss, which lawful collaboration or policy channels that would take, which Conflictoring agents the board and CEO intend to engage, and what milestones would show progress. The plan is the answer to the short-seller's framing. Disclose the exposure, disclose the route out of it, and make the process credible enough to believe.