Escalation has two exits and only one of them leaves a record: a risk can fall below the threshold because it was treated, or because the threshold moved, and no instrument in general use distinguishes them.
A risk has been red on the delivery report for four months. This month it is amber, and nobody raises it, because a risk moving in the right direction is not what a governance meeting is for. The pack shows the change and offers no line explaining it.
Two things could have produced that colour. Somebody could have done something: a supplier qualified, a design frozen, a licence renewed. Or somebody could have decided that an exposure unacceptable in April is acceptable now. The first is delivery. The second is a change in the enterprise's risk position, invisible in exactly the way the first is visible.
This is not an accident of reporting practice; it is written into the doctrine. A level receiving an escalated risk is offered two responses: manage it, or adjust the delegated tolerance for the level below. Both are proper. Only one generates an artefact — a treatment with an owner, a cost, a due date and a closure. The other generates a colour.
Two further instructions, each sound alone, compound this. Risk appetite is temporal: a board may vary how much risk it will accept as circumstances change, and treating appetite as fixed is named an error. A delivery leader inherits the risk-return profile that justified the investment and must deliver within it. So the boundary may move from above at any time, the person held to it may not move it at all, and that leader is asked to stay attuned to shifts in appetite. Attunement is what a system asks for when it has no notification.
Risks reducing, the most-quoted line in delivery reporting, has two causes with opposite meanings, and a board reading improvement may be reading its own decision returned to it as news.
The Strategic Context
Consider a hypothetical facilities services contractor running multi-site cleaning and static guarding on thin margins, with high turnover and regional coverage subcontracted to smaller operators. One risk has sat red for months: currency of guard licensing across that network, which the client contract makes a condition of payment.
There are two routes to amber. The contractor can build an expiry check into rostering, so an uncurrent guard cannot be allocated a shift. Or an executive can conclude that a small proportion of non-compliant shifts is survivable against the margin, and the internal threshold moves from none to some.
The two routes produce an identical pack and are not the same decision. One is an operating improvement that costs money once. The other changes exposure across every site and client, and will be found by an audit or a claim rather than a governance meeting.
What Leaders Commonly Misread
The first misreading is that a falling risk profile is evidence of delivery performance. It is evidence of arithmetic: one side of an inequality moved, and the report does not say which.
The second is that tolerance changes must be visible because senior people make them in scheduled meetings. Seniority does not create a record. Minutes capture what somebody framed as a decision, and moving a threshold is framed as a judgement about a risk, not an amendment to a boundary.
The third is that the register holds the answer. A register records a risk's rating and status, not the standard it was assessed against, so a re-rating and a re-standarding change the same field in the same direction.
An item can also leave the process by a third route, reclassified as something other than a risk; that route and its cost to an enterprise is the subject of [Related article: Where Do Your Near-Certainties Go?] and is not examined here.
Reframing the Issue
Escalation compares a measurement against a boundary, so every reported change in status is a change in one of two operands, and the instruments in general use name neither. The useful question is not whether a risk is still red but which operand moved. The remedy follows from it: the boundary must carry a version history, as a schedule baseline or an approved budget already does. Nobody would accept a favourable cost variance without knowing whether the spend fell or the budget rose, and risk reporting is accepted on exactly those terms every month.
This article concerns the second exit and the record it fails to leave. What never reaches the threshold at all, and why modest exposures accumulate beneath it unseen, belongs to [Related article: Escalation Sized by One Risk] and is not covered here.
The Exit That Leaves No Record
The ambiguity is sanctioned, not accidental
Guidance on monitoring appetite states that a breach may indicate either that behaviour has drifted outside acceptable bounds or that the appetite itself needs recalibrating. No instrument in general use supplies a test for telling the two apart afterwards. The discipline names the fork and leaves the enterprise standing at it.
The authority runs one way
The freedom to vary appetite sits with the level that granted the tolerance. The obligation to deliver within it sits with the level that received it. Between them there is no notification duty, no distribution list, no dated instrument. A delivery leader can be in breach of a boundary they were never told had tightened, or credited with an improvement they did not produce.
The second exit is used because it is cheap
Moving a tolerance is instant, costs nothing in the period, requires no supplier, design change or reserve, and produces the same reported outcome as a treatment that would have taken two quarters and real money. Nothing in the governance system prices it.
A hypothetical operator managing fishing quota across several species shows the cost. Reported by-catch exposure can fall because gear and fishing behaviour changed, or because the internal limit widened after a strong season made the penalty affordable. The stock does not distinguish, and the second route reverses only after the position has moved.
Decision Framework
The tolerance-change minute makes the second exit as visible as the first. Any change to a tolerance, threshold, appetite statement or escalation trigger takes effect only when a minute exists carrying five items.
- The boundary before and after, in the original units: the two numbers, not a description.
- Who moved it, and under what authority. Named individual, named delegation.
- The circumstance that changed. A tolerance moves because the environment moved. If the stated reason is a fact about the risk rather than the enterprise, this is a re-rating and belongs in the register.
- The affected register lines, named, with status before and after. This converts an invisible act into a visible one.
- A lapse date on which the wider boundary reverts unless renewed by the same authority. Widening is temporary by default.
Two rules make it bind. No risk may improve its reported status in the same period as a change to the tolerance governing it unless the minute reference appears on its line. Where no minute exists, the boundary has not moved, whatever anyone believes was agreed.
The minute records who moved a boundary. What that person may commit the enterprise to owe is set by an entirely different mechanism, examined in [Related article: What May They Cost You?].
From Strategy to Execution
Immediately. Take twelve months of risk reporting for one programme and list every item that improved without a completed treatment action against it. Ask, for each, what changed. The count is usually the argument.
Over the next two quarters. Put the tolerance-change minute into the governance calendar, and add a boundary-version field so every rating carries the standard it was assessed against. Require the minute reference on any improvement not backed by a closed action.
Over the longer horizon. Treat appetite as a versioned baseline reviewed on the cadence of the business case, and make a change to it a notified event with a distribution list. The board keeps its freedom; what ends is exercising it unannounced.
Signals to Monitor
Watch for improvements that cluster immediately before a gate, a bid or a reporting cut-off, and for a widening gap between the count of risks improved and the count of treatment actions closed in the same period; that gap is the signature of the second exit.
Watch too for a risk that improves and then reappears at its old rating months later under a slightly different description, which is what a moved boundary looks like when the exposure beneath it was never touched.
Questions for the Leadership Team
- Over the last four quarters, how many risks improved in reported status without a completed treatment action, and which were they?
- Which tolerances or appetite statements changed in that period, who authorised each, and where is the record held?
- When our appetite last changed, which delivery leaders were told, by what means and on what date?
- For our three largest programmes, what risk-return profile was approved at investment, and does the tolerance now in force match it?
- Which of our currently green risks would be red under the tolerances in force twelve months ago?
- Who here can widen a tolerance without producing a document, and was that authority granted or acquired by custom?
Closing Perspective
A board's freedom to vary its appetite is legitimate and worth defending; an enterprise that cannot re-price its own risk position is rigid rather than disciplined. The problem is not the second exit but its silence.
Silence here is not neutrality. An unrecorded widening converts a governance decision into a compliment paid to the delivery organisation, in the one document the board relies on to judge whether its investments are managed. Over enough periods an enterprise can improve its whole reported risk profile without changing anything in the world, everyone acting in good faith.
The choice is narrow and practical. Keep the freedom and pay for it with a dated line saying what moved, who moved it and when it lapses. Or keep the silence, and accept that part of the improvement reported each month is the board's own decision, coming back around.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.