After a rigid new policy launches, the first people to ask for exceptions are investigated and their cases circulated, so everyone else concludes that asking is not worth the risk.
An example
A regional bank's wealth management unit rolled out a new personal-trading policy six weeks after a regulator's inquiry. Within a month the chief compliance officer sent a lessons-learned bulletin to all four hundred advisors. It described, without names but with enough detail that everyone knew who they were, two associates who had asked for carve-outs to keep positions they owned before joining. Both had their full trading history reviewed, and both were placed on a restricted list. An advisor in the next branch had drafted an exception request for a deferred-compensation account left over from a previous employer, and deleted it that afternoon. Over the following quarter compliance received three requests against the forty it had budgeted to handle, and a rule that blocked advisors from rebalancing their own retirement plans went unchallenged for a year.
Why it works
Asking for an exception stops looking like a reasonable request and starts looking like volunteering for an investigation, so its benefit shrinks, and its cost now includes visibility, embarrassment, and a wider review of one's conduct. It depends on the operator controlling the channel that broadcasts the cases, on the first requesters being plausibly framed as deserving what happened, on a workforce with few exits or little appetite to coordinate, and on a policy that is defensible on its face. Once the early cases land, people enforce the rule on themselves.
The thinking behind it
Michel Foucault In Discipline and Punish: The Birth of the Prison (1975), Foucault argued that power which is visible but unverifiable leads people to regulate themselves, so enforcement is rarely needed. A broadcast of the first exception cases works this way: nobody knows which request will trigger a review, so everyone stops asking. The defense follows from the same idea, since published criteria and aggregate numbers make the power verifiable again.
Amos Tversky Tversky and Kahneman's Judgment under Uncertainty: Heuristics and Biases (1974) describes availability: people judge how likely something is by how easily examples come to mind. Two vivid, circulated cases of requesters being investigated make the risk of asking feel far larger than the real rate of denials. Publishing the full numbers, requests and approvals alike, gives people a base rate to set against the story.
Warning signs
A new rule arrives, and the first denied exception requests are described in an organization-wide bulletin, a training deck, or a town-hall aside. The requesters are identifiable even without names, and their broader conduct was reviewed along with the request. Exception volumes fall far below what anyone expected for a rule of that scope, and known flaws in the policy stop being raised.
How to defend yourself
- Keep a dated copy of the bulletin and of your own request, and ask in writing what criteria govern exceptions.
- Talk to colleagues before filing. Requests filed together, on the same grounds, are hard to turn into a spectacle, and a stack of them tells management something about the policy.
- Frame your request around the flaw in the rule, with numbers, so it reads as a correction the organization needs.
- If the first cases look chosen for their teaching value, ask internal audit, the audit committee, or the board's risk function to review how they were selected and publicized.
The tactic is weakest when the first person caught is sympathetic. If that is you, a measured written account to HR and to your manager's manager puts the cost of the spectacle back on the people running it.
How a leader shuts it down
Separate the people who write a rule from the people who rule on exceptions. Publish the criteria and the aggregate numbers (requests, approvals, denials) without the cases. Treat a collapse in exception volume after a launch as a warning about the rule, and never let compliance communications describe identifiable individuals as examples.