Actuarial Communication for Risk Committees

By Jonas Osman Abdelghafour.

The technical quality of an actuarial or quantitative risk analysis is a necessary condition for its usefulness, not a sufficient one. The analysis has to be understood, trusted, and acted on by a risk committee whose members typically have wide experience but limited time and, in many cases, limited technical background. Communication design — what is shown, in what order, with what caveats — determines whether the analysis changes a decision.

Start with the decision

Every paper to a risk committee should open with the decision being asked. If there is no decision, the paper is for information and should be labelled as such and kept short. If there is a decision, the opening page should state it, the options, and the recommendation, with the analysis behind it in the pages that follow.

This inverts the sequence in which the analysis was performed but matches how a committee reads.

Ranges, not point estimates

Almost every material actuarial output is a range. Reserves have a best-estimate and a range of reasonable outcomes; capital measures depend on distributional assumptions that are themselves uncertain; scenario losses depend on scenario calibration.

Reporting a single number without a range implies precision the analysis does not have. Reporting a range without saying what drives its width leaves the reader without a way to assess whether it is realistic. The useful format is a central estimate, a range, and one sentence for each of the two or three assumptions that most affect the width of the range.

Surface the assumptions

Every material assumption should be visible: which discount rate, which loss development pattern, which correlation matrix, which management action credit. A risk committee cannot challenge what it cannot see. Where an assumption is contested or evolving, that should be stated explicitly rather than buried in an appendix.

For scenarios and stress tests, the assumption pack should include the narrative, the calibration reference (historical episode, regulatory reference scenario, or portfolio-specific tail analytic), and the management actions credited. See ORSA scenario design and board use for the scenario side of this discipline.

Acknowledge limitations

Limitations are not weaknesses of the paper; they are properties of the analysis. Stating them plainly increases trust rather than reducing it. Useful limitations are specific: "the reserve range assumes claims inflation of X%; a persistent Y% would move the central estimate by approximately Z"; not "the analysis is subject to uncertainty".

Where the analysis is unable to answer a question the committee is likely to ask, saying so — and outlining what would be needed to answer it — is more useful than a synthetic answer with a large error bar.

Visualisation choices

Three principles usually apply. First, one chart per idea; charts with three overlaid metrics on two axes rarely land. Second, direct labels rather than legends; a committee reads a chart in seconds. Third, consistent conventions across the pack — the same colour for the same scenario, the same axis orientation for the same metric — so that comparisons are visual rather than cognitive.

Sensitivity waterfalls, fan charts of reserve ranges over development periods, and simple bar charts of scenario impact versus appetite thresholds are the workhorses. Complex heatmaps are usually a sign that the underlying data has not been aggregated to the level of the decision.

Language

The language should be neutral, specific, and free of marketing register. Prefer "the model estimates" to "our best-in-class model shows"; prefer "the assumption is conservative relative to historical experience" to "we are being prudent". Where terminology is unavoidable — Solvency II modules, IFRS 17 building blocks, capital metrics — a one-line footnote on first use is worth more than a glossary at the back.

Q&A discipline

The Q&A after the paper is where communication most often succeeds or fails. Preparation matters: the top five likely questions, with drafted answers and any additional exhibits, should be reviewed before the meeting. Where a question cannot be answered fully in the room, "I do not have the data to answer that with confidence today; I will come back with X by Y" is a stronger response than a partial answer that later has to be corrected.

Common failure modes

Three patterns recur. Papers that lead with methodology instead of the decision. Point estimates presented without ranges, forcing the committee to either accept them at face value or reject them without a basis. And appendices that grow to the point where the committee stops reading them, hiding the assumptions the analysis actually depends on.

Limitations

Good communication does not fix a weak analysis. It exposes the analysis for what it is, which is exactly the point. Committees that receive well-communicated weak analyses make better decisions than committees that receive well-communicated strong analyses presented as certainties.

Conclusion

Actuarial communication is a professional skill in its own right. Structured around the decision, honest about ranges and assumptions, specific about limitations, and disciplined in language and visualisation, it turns technical work into governance value. Related notes: ORSA scenario design and board use, model risk remediation and finding closure, and enterprise risk management, demystified.

---

Written by Jonas Osman Abdelghafour, actuary and financial risk manager. Background and contact details are on the about page.