Framing, Base Rates and the Quality of Public Decisions: A Reading for Policy Audiences
# Framing, Base Rates and the Quality of Public Decisions: A Reading for Policy Audiences
There is a sentence in Die Architektur des Denkens that has followed me into every committee room and regulatory workshop I have entered since. A surgery with a ninety percent survival rate and a surgery with a ten percent mortality rate are mathematically identical. Psychologically, they are not. The first produces assent, the second produces hesitation. This small asymmetry, repeated across millions of public decisions, forms the invisible grammar of political consent. Regulatory instruments, investor letters, ministerial communiqués: they are all written in one frame or another. None of them is neutral. And yet we pretend, in Brussels as in Frankfurt, that the substance of a decision can be cleanly separated from its presentation. It cannot. The frame is part of the substance. To write this essay, as Dr. Raphael Nagel (LL.M.), is to ask what it would mean if public institutions took that fact seriously in the daily exercise of their legitimacy.
The Three Insights and Why They Leave the Laboratory
In the fourth chapter of my book I described the intellectual friendship between Daniel Kahneman and Amos Tversky, and the paper they published in Science in 1974 that mapped the systematic errors human judgement makes under uncertainty. Three findings from that body of work travel particularly well into public life. The first is that certainty is not a quality mark. The feeling of being sure reflects the fluency of a thought, not its correctness. The second is that the frame determines the answer. Identical information, presented differently, produces divergent judgements. The third is that base rates defeat vivid stories, but only when they are visible. When they are not, the single dramatic case displaces a hundred statistical studies.
These three insights were developed in psychology laboratories. They have spent the last decades travelling, rather unevenly, into medicine, finance and behavioural economics. What remains strangely uncolonised is the European regulatory tradition. Impact assessments, consultation papers, transposition notes: the texts that shape the daily life of half a billion citizens are still written as though the human recipient were the rational agent of mid-twentieth-century economics. That reader does not exist. The reader who does exist is the one Kahneman described, whose system one produces and whose system two rationalises, and whose assent to a rule depends at least as much on the rule’s framing as on its content.
The Regulatory Frame and the Illusion of Neutrality
Consider any directive that concerns risk: financial stability, pharmaceutical approval, environmental exposure, data protection. Each such directive can be drafted in the language of loss avoided or in the language of harm risked. A pension rule can promise that ninety-four percent of contributors retain their capital, or it can warn that six percent will experience partial loss. A medication monograph can report that nineteen out of twenty patients tolerate a compound, or that one in twenty experiences a defined adverse event. The regulator who chooses the first version does not lie. The regulator who chooses the second version does not lie either. Yet the consent these two versions produce in parliaments, in the press and in citizens’ associations is radically different.
This is not a marginal observation. It is a central fact of regulatory legitimacy. When a text reaches the European Parliament, its frame has already been chosen, usually without documentation and almost always without contestation. The consultation that follows treats the frame as given and debates the margins. The outcome is ratified as democratic. It is democratic in procedure and, in a quieter sense, pre-decided in presentation.
A regulatory culture that took Kahneman seriously would require what one might call symmetrical framing: every risk quantified in both directions, every benefit expressed alongside its complementary cost, every probability accompanied by its inverse. The additional cognitive labour would be small. The gain in the quality of collective judgement would not be small. It would, over time, change what European publics understand themselves to be consenting to.
Certainty in the Ministerial Voice
The first of Kahneman’s insights, that certainty is not a mark of quality, is the hardest for public institutions to absorb. A finance minister who says without qualification that a reform will work is rewarded by the markets and by the evening news. A minister who says that the reform is the best available instrument given present evidence, and that it may require correction within two years, is punished as weak. The incentive structure of political communication selects for the very cognitive posture that Kahneman identified as the signature of poor judgement.
This is not a problem that rhetorical reform alone can solve. It is structural. But the structure is not untouchable. Central banks, through decades of patient work, have built a communicative culture in which calibrated uncertainty is not a weakness but a professional norm. A forward guidance statement that speaks in ranges and conditional paths is read by markets as more credible, not less. The transformation did not happen by exhortation. It happened by institutional practice, sustained over a generation.
There is no reason the same transformation cannot be extended to regulatory communication more broadly. A directive that states its confidence intervals, that names the evidence it has weighed and the evidence it still lacks, that specifies the conditions under which it should be reviewed, is not a weaker directive. It is an honest one. And honesty, over long stretches of political time, is the only thing that compounds.
Base Rates and the Vivid Case in Public Discourse
The third insight, that base rates lose to vivid stories in the absence of explicit countervailing practice, has become painfully visible in European public discourse over the past decade. A single failure in a pharmaceutical trial, a single incident in a migration corridor, a single fraud at a listed company becomes the entire evidentiary base for a rule that will govern millions of cases statistically unlike it. The story is not wrong; it is only misleadingly representative. But misleading representativeness, at scale and in print, produces misshapen law.
The remedy is not to banish stories. Stories carry moral weight that statistics cannot. The remedy is to insist, as a matter of institutional practice, that every vivid case be accompanied by the base rate against which it should be read. How often does this occur? In what population? Under what prior conditions? A regulation that responds to a one-in-ten-thousand event as though it were a one-in-ten event will misallocate resources by three orders of magnitude. A regulation that does not know the difference cannot correct itself.
Institutional investor communications face the mirror image of this problem. The single quarter of outperformance, the single analyst downgrade, the single anecdote from an earnings call displaces the long-run distribution. Investors who read their own reports under the discipline of base rates make different decisions from those who read under the discipline of narrative. The two disciplines produce different portfolios. Over a decade, they produce different institutions.
Toward an Architecture of Public Decisions
What would it mean, concretely, to write this insight into European practice? A regulatory impact assessment that did cognitive justice to its reader would contain three elements that are currently absent or optional. First, symmetrical framing: every effect stated in both directions, with the two formulations placed side by side rather than alternating between sections. Second, explicit base rates: the prevalence of the phenomenon being regulated, the historical frequency of the harm being prevented, the comparator rates in adjacent jurisdictions. Third, calibrated certainty: a statement of what the drafter knows, what the drafter assumes, and under what evidence the assumption would be revised.
None of this is exotic. Each element is standard in medical evidence reviews, in central bank communication, in serious meteorological forecasting. The obstacle to extending these elements to regulatory and investor practice is not methodological. It is cultural. The culture of public writing still treats confidence as virtue and uncertainty as failure. Dr. Raphael Nagel (LL.M.) has argued, in Die Architektur des Denkens and elsewhere, that the reverse is closer to the truth: calibrated uncertainty is the mark of an institution that knows what it is doing, and unqualified confidence is often the mark of one that does not yet know what it does not know. The practical consequence is a different kind of text, and over time, a different kind of decision.
An editorial plea, then, to close. Base-rate literacy is not a technical skill reserved for statisticians. It is a civic capacity. A public that cannot ask, of any vivid claim, how often this actually occurs, is a public whose consent can be engineered by any competent communicator. A public that can ask that question, and that demands an answer before proceeding, is a public whose institutions must work harder and, in working harder, produce better decisions. The cultivation of this literacy is not the work of a season. It is the work of schools, of newspapers, of parliamentary committees and of the professional bodies that train regulators and investors. It requires that base rates be spoken aloud in places where they are currently whispered in footnotes. It requires that framing be named as framing, and that the act of choosing a frame be treated as a political act, subject to the same scrutiny as any other. It requires that certainty in public voice be read, as central bankers have learned to read it, as a signal to be interpreted rather than a guarantee to be accepted. None of this will abolish the biases Kahneman and Tversky described. Nothing will. But it will make the institutions that govern us a little more honest about the minds they are addressing, and a little more worthy of the consent they ask for.
Claritáte in iudicio · Firmitáte in executione
For weekly analysis on capital, leadership and geopolitics: follow Dr. Raphael Nagel (LL.M.) on LinkedIn →