Every significant accountability failure in institutions follows a similar sequence. Warnings surface. Internal reports are filed. Staff raise concerns through the designated channels. The system receives each signal, acknowledges it, and continues without changing direction. Then something external (a regulator, a market, a court, an incident) imposes the cost the institution declined to impose on itself. The correction is always described as sudden. The record that preceded it tells a different story.
Why the systemic accountability crisis keeps repeating
The standard explanation for institutional failure is individual. A negligent executive. A board sleeping at the wheel. A regulator who looked the other way. Any of these explanations is attractive because it implies that better people, placed in the same positions, would produce different outcomes. The actual evidence runs against that explanation.
In October 2024, TD Bank entered a guilty plea to violations of the US Bank Secrecy Act and agreed to pay more than $3 billion in penalties, one of the largest anti-money laundering enforcement actions on record. The conduct the settlement addressed had been running since 2019.
The sequence is worth examining. It is worth it, precisely because it is unremarkable. TD Bank’s anti-money laundering systems generated alerts. Staff escalated concerns about the adequacy of controls. Supervisory findings had already flagged weaknesses before the most damaging period began. The institutional feedback loops were operational in a procedural sense. Warnings went up the chain. They were received, noted, and filed, but none of them carried a cost that matched the risk they were describing.
Business growth continued. Process inertia absorbed the warnings. For four years, the monitoring systems kept flagging the same problems and the bank kept not fixing them. No internal mechanism closed the gap between what the monitoring systems were reporting and what the institution was willing to do about it.
This is the mechanism that produces what looks, from the outside, like negligence. What it is, more precisely, is the rational output of a system in which acting on internal feedback costs more, in the short term, than absorbing it. The staff member who escalates a concern fulfils their procedural obligation. The committee that receives the escalation fulfils its governance role. The report is filed, acknowledged, and placed in the business’s records. At no point in this sequence does anyone bear a personal responsibility or cost proportionate to the risk the report described. The institutional feedback loops operated. They produced documentation but did not produce change.
How accountability failure in institutions invites external correction
The correction, when it arrived, came from outside TD Bank’s governance apparatus. US federal authorities imposed criminal liability, financial penalties exceeding $3 billion, and operational restrictions that the institution’s own systems had never generated. The external cost in monetary and operational terms was orders of magnitude larger than the cost and operational burden of acting on any single, early alert during the years this conduct accumulated.
The asymmetry is systemic. Institutions that absorb internal feedback without acting on it do more than simply deferring correction. They raise the eventual price of that correction. Every year that a warning sits in a filing cabinet without a response, adds to the bill that the external enforcer will eventually present. Each quarter in which the alerts accumulated without corrective action and the conduct continued, added to the exposure. When the bill finally arrives, it arrives at the scale the years of deferral had made inevitable.
This pattern extends well beyond a single bank. Financial regulators globally imposed $4.6 billion in anti-money laundering penalties in 2024, with TD Bank’s settlement alone accounting for the majority. In 2025, global AML penalties reached $3.8 billion, and the first half of the year saw enforcement actions quadruple compared to the same period in 2024. Years of partial supervision and accumulated regulatory enforcement failures across multiple jurisdictions helped produce this correction cycle. The penalties now arriving reflect the compounded cost of feedback that was generated, reported, and systemically ignored for years before the enforcement arrived.
This process functions much the same in organisations that would never describe themselves as non-compliant. A risk register carrying the same amber items for three consecutive quarters. A culture survey whose results are acknowledged in a town hall and filed without response. A performance management system that produces differentiated ratings on paper while every employee receives the same percentage increase. Each is a feedback loop generating signal without bearing cost. The signal exists. The enforcement of that signal has been made procedurally expensive, politically awkward, or disconnected from the decisions it was supposed to inform. Or a combination of them.
The cost of this mechanism often falls on the people who operated in good faith within the system it produced. The staff who escalated concerns, followed procedures and filed accurate reports, discover that the system they trusted to act on their information was built to receive it. The feedback function was procedural. The trajectory continued regardless.
Organisations that have operated this way long enough develop a specific institutional character. The reporting infrastructure grows more sophisticated with each cycle. More dashboards, more committees, more escalation pathways. The gap between what the system knows and what it does about what it knows, keeps widening because the apparatus rewards the production of reports while the correction of what the reports describe remains unaddressed.
The person who recognises this in their own environment has a specific piece of information. Every report that was filed and absorbed without consequence is still accumulating somewhere in the system’s exposure. The correction will arrive. It always does. And when it lands, it does it at a scale that reflects every quarter in which the system knew what it knew and carried on as though knowing were the same as “acting.”
The information was never the problem. The reports existed, the alerts fired, the committees convened. What was absent was cost.
Pressure produces reports. Cost produces reform. A system that has been swallowing its own warnings for long enough, eventually forgets they were warnings at all. It looks solvent. The bill just hasn’t arrived yet.