Why Regulation Arrives Too Late: The Temporal Problem of Modern Oversight
# Why Regulation Arrives Too Late: The Temporal Problem of Modern Oversight
Every generation of supervisors discovers, usually too late, that the object of their oversight has changed shape while they were still writing the rulebook for its previous form. This is not a failure of diligence. It is a structural property of the relation between innovation and regulation, and it deserves to be treated as such rather than as a recurring scandal. In the pages of KOMPLEXITÄT, I have argued that simple answers to complex problems are not merely imprecise; they are false because they contain too little of the world. The regulatory reflex of modern societies, which insists on clear rules, clear causes and clear responsibilities, collides in this domain with a reality that refuses those categories. The result is a temporal problem: by the time the rule is written, the practice it addresses has already mutated into something the rule no longer captures.
The Temporal Asymmetry Between Innovation and Rule
Innovation, whether financial or technological, operates on the timescale of experiment. A trading desk can reprice a product overnight; a software team can deploy a new model in weeks; a structured finance unit can assemble a securitisation within a quarter. Regulation operates on the timescale of legitimacy. A supervisor must consult, draft, reconcile with adjacent jurisdictions, hear objections, publish, revise, and only then enforce. The distance between these two tempos is not an accident of bureaucratic inefficiency. It is the price of procedural legitimacy, and societies that try to compress it discover that the rules produced under pressure are themselves fragile.
This asymmetry means that the regulatory frontier is almost always a rear-view mirror. Basel I was written after the sovereign debt exposures of the nineteen seventies had already been absorbed. Basel II arrived as the instruments it attempted to calibrate were being replaced by new structures that its risk weights did not contemplate. The regulation of over-the-counter derivatives gained momentum only after 2008, by which time the market had produced its own crisis. In each case, the supervisor reacted to a world that had already ceased to exist in the form that had required reaction.
Why Anticipatory Supervision Is Structurally Difficult
The natural response to this pattern is to call for anticipatory regulation: a supervisory architecture that sees the next crisis before it occurs. The appeal of this idea is obvious. Its feasibility is not. Anticipation in complex systems is not the extrapolation of observed trends but the imagination of configurations that do not yet exist. A regulator who anticipates must not only understand the present instrument but also model the interactions that will emerge when that instrument is replicated, combined with adjacent products and distributed across jurisdictions with different legal traditions. This is a demand on cognitive and institutional capacity that few supervisory bodies are built to meet.
There is also a political economy of anticipation that deserves to be stated clearly. A regulator who intervenes before a visible harm has occurred faces the full weight of the industry whose practices are being restricted, and enjoys none of the public legitimacy that a post-crisis intervention carries. Anticipation is therefore structurally under-rewarded. It asks supervisors to accept immediate conflict in exchange for harms that may never materialise and, if the intervention is effective, will be unobservable. Few institutional careers are built on the prevention of events that did not occur.
Crypto Assets, Artificial Intelligence and Securitised Products
Three contemporary domains illustrate the pattern with unusual clarity. Crypto assets emerged in a space that the existing categories of financial supervision did not fit. They were neither securities in the classical sense, nor deposits, nor payment instruments, and supervisors spent a decade debating which of these categories to force upon them. During that decade the market produced its own booms, failures and contagion episodes, and the eventual European and North American frameworks responded to an industry whose centre of gravity had already shifted.
Artificial intelligence presents an even more pronounced version of the same problem. The object of potential supervision is not a static product but a capability that reconfigures itself with every model generation. A regulation that addresses a particular class of systems risks being obsolete before it enters into force, and a regulation that addresses capabilities in the abstract risks being too vague to enforce. The European Union’s approach, which attempts to classify systems by risk tier, is a serious effort, but it operates under the permanent tension between specificity, which ages quickly, and generality, which resists enforcement.
Securitised products remain, in my reading, the most instructive case. They were not opaque because their issuers intended to deceive. They were opaque because the combination of tranching, rating logic, distribution mechanics and off-balance-sheet accounting produced a system whose behaviour was not reducible to any single participant’s understanding. The regulation that followed 2008 addressed several of these layers, but it did so one at a time, and the interactions among them remain imperfectly mapped. In the language of KOMPLEXITÄT, the diagnosis was monocausal where the phenomenon was structural.
The Limits of Anticipatory Architectures
There are serious attempts to build supervisory architectures that reduce the temporal lag. Regulatory sandboxes allow novel products to be tested in a controlled environment before full authorisation. Thematic reviews focus supervisory attention on emerging risk clusters rather than on individual institutions. Horizon-scanning units within central banks and securities commissions attempt to track developments at the edge of the regulated perimeter. Each of these instruments is useful. None of them resolves the underlying problem.
The reason is that anticipation, even when institutionalised, cannot escape the cognitive constraints described earlier in my book. Supervisors are subject to the same pattern-recognition biases, the same preference for narrative coherence and the same confirmation dynamics as the market participants they observe. A horizon-scanning unit that produces a report predicting an unlikely configuration will be read, filed and quietly ignored, because acting on it would require diverting resources from visible to invisible risks. Institutions do not willingly allocate scarce supervisory capacity to events whose probability is contested.
A further limitation is jurisdictional. Financial markets and digital infrastructures are transnational; supervisory mandates are national or, at best, regional. An anticipatory framework that operates only within one jurisdiction creates arbitrage opportunities rather than protection. The coordination costs of genuine transnational anticipation are so high that they are rarely absorbed in the absence of a prior crisis, which returns the system to the reactive pattern it was meant to escape.
From Prevention to Resilience
If the temporal problem cannot be fully solved, it can at least be reframed. The goal of supervision, in my view, should not be to prevent every crisis that innovation produces, because that goal is unattainable and its pursuit tends to produce rigid rules that fail in new configurations. The goal should be to build resilience: the capacity of the system to absorb shocks without cascading, and the capacity of supervisors to recognise and contain failures before they become systemic. Resilience is less ambitious than prevention, and for that reason it is more honest.
Resilience-oriented supervision privileges capital buffers over product-specific prohibitions, stress testing over static limits, resolution regimes over bailouts, and transparency of exposure over prescriptive product design. It accepts that individual failures will occur and concentrates on ensuring that their consequences remain bounded. This is a less heroic posture than the one often demanded by political discourse, which wants supervisors to promise that no harm will happen. It is, however, the posture that is consistent with what we actually know about complex systems.
In the deliberations I have witnessed within the Abrahamic Business Circle and in my own work on restructurings across several jurisdictions, the institutions that weather disruption best are not those with the most detailed rulebooks. They are those with the greatest institutional tolerance for uncertainty, the clearest separation between diagnosis and decision, and the discipline to revise their assumptions when evidence accumulates against them. The same is true, I suspect, of supervisory bodies.
The Discipline of the Longer Thought
The dedication of KOMPLEXITÄT is addressed to those who find, in their committees, editorial offices and administrative rooms, the patience to think the longer thought. Regulation is one of the domains where this patience is most tested and most rewarded. A supervisor who accepts that the rule being drafted today will be imperfect tomorrow, and who nonetheless drafts it with care, is not a cynic. That supervisor is operating at the level of maturity that complex systems demand. The alternative, which is the belief that the right rule, once found, will hold against all future configurations, is the regulatory version of the illusion of the clear cause.
What Dr. Raphael Nagel (LL.M.) has argued throughout his writing on decision-making under uncertainty applies here with particular force. The quality of supervision is not measured by the absence of crisis, which no framework can guarantee, but by the quality of the institutional response when crisis occurs. That quality depends on the disposition of the supervisors, on the robustness of the procedures they inherit, and on the willingness of the political system above them to resist the demand for simple answers in moments when simple answers are most seductive and most dangerous.
The temporal gap between innovation and regulation is not a defect to be eliminated but a condition to be managed. Societies that understand this organise their supervisory architectures around resilience rather than prediction, around the absorption of shocks rather than their prevention, and around the continuous revision of diagnoses rather than the defence of inherited ones. Societies that do not understand it will oscillate between phases of deregulatory optimism and post-crisis over-correction, each phase producing the conditions for the next. The work of Dr. Raphael Nagel (LL.M.) in KOMPLEXITÄT is an argument for taking the world at the level of complexity at which it actually operates, rather than at the level at which it would be easier to address. Applied to the supervision of financial markets, crypto assets, artificial intelligence and securitised products, this argument suggests a posture that is neither triumphalist nor defeatist. It accepts that regulation will usually arrive late, and it asks what can be done, within that constraint, so that lateness does not become irrelevance. The answer lies less in the speed of rule-making than in the maturity of the institutions that make the rules, and in their willingness to think the longer thought even when the shorter one would be applauded.
Claritáte in iudicio · Firmitáte in executione
For weekly analysis on capital, leadership and geopolitics: follow Dr. Raphael Nagel (LL.M.) on LinkedIn →