A dated account of risk exposures, changes, limitations and decisions: how to turn a VaR increase and limit breach into accountable follow-up.
Risk report
A risk report is a dated communication that brings together risk measurements, explanations, limitations and required decisions for a defined audience. It tells a portfolio manager, risk manager or committee what the exposures mean for the business and what needs attention. A correct number can still support a poor decision if the report hides its scope, its uncertainty or an unresolved exception.
A risk dashboard usually offers a compact monitoring view: levels, trends and alerts. A report develops the interpretation and preserves the evidence, decision and follow-up. The distinction is functional, rather than a file-format rule: an interactive dashboard can contain a report, and a PDF full of unexplained tiles can fail to provide one.
Read this fictional report excerpt
Consider Portfolio P, reported in USD. The previous and current snapshots are taken on successive reporting dates at 16:00 UTC, with the current report issued at 17:00 UTC. Both use the same portfolio inclusion rules, one-day horizon, 95% confidence level and approved VaR methodology. Positions and market observations may change between snapshots; their separate effects have not yet been established. All figures below are stipulated examples, not institutional data.
Report field
Entry
Headline
One-day 95% value-at-risk rose from USD 0.9 million to USD 1.2 million.
Change
Increase of USD 0.3 million on the stated comparison basis.
Exception
Current risk limit: USD 1.0 million on the same basis. The current limit breach is USD 0.2 million.
Explanation status
Position changes and market-input changes require controlled comparison; no quantified driver attribution is established by these two totals.
Decision required
The authorised risk manager must determine the response under Portfolio P's escalation policy, including whether restrictions, mitigation or a time-bounded exception are permitted.
Accountable follow-up
Fictional portfolio risk owner Maya Chen coordinates the response. Risk analyst Alex Reed must provide the comparison and data checks before the next scheduled review. The authorised risk manager records the decision, action owner and deadline.
The arithmetic answers two different questions, in USD millions:
Change=1.2−0.9=0.3,Breach=1.2−1.0=0.2.
The first compares measurements across reporting dates. The second compares the current measurement with its threshold. Neither amount is realised trading loss. VaR summarises a modelled loss distribution; it does not state the maximum possible loss. The change also does not identify its own cause: “VaR rose because volatility rose” needs evidence beyond the headline totals.
Previous VaR is USD 0.9 million and current VaR is USD 1.2 million against a current limit of USD 1.0 million. The increase is USD 0.3 million; the current excess above the limit is USD 0.2 million. A report distinguishes these measured facts from unresolved drivers and the policy decision.
Original fictional comparison in USD millions. Both snapshots use a one-day horizon and 95% confidence level. The dashed line is the current limit; the chart is a measurement comparison, not a P&L attribution.
A missing driver explanation is a reason to label the analysis incomplete, not to suppress the breach until the explanation is ready. Equally, reporting the breach does not itself authorise a trade or an exception. The risk owner, analyst and decision-maker have related but distinct responsibilities.
Make the comparison reproducible
A report needs enough context for another reader to reconstruct its meaning. Portfolio scope includes the books, legal entities, products and off-balance-sheet positions covered, along with exclusions and aggregation rules. Time context includes valuation time and time zone, data cut-offs, issue time, comparison period and any stale inputs. Currency, valuation basis, horizon and confidence level belong beside a metric that depends on them.
An unchanged methodology does not mean unchanged inputs. To investigate the example, an analyst could rerun the measure with selected positions or market inputs held fixed. Such a decomposition must declare its order, reference snapshot and treatment of interactions. With nonlinear portfolios, one ordering can allocate a joint effect differently from another. Label unexplained residuals rather than assigning the whole increase to a plausible market story. Methodology changes, corrected data and scope changes need their own explanation; a restated comparable history can help distinguish a new measurement basis from a change in the portfolio's risk.
Record the data snapshot, model version, reconciliation results and any manual adjustments behind the report. Data lineage develops the connection between reporting and the run records that produced it, including the danger of different jobs reading different snapshots. Its full lesson requires membership; a public preview is available.
What the headline leaves out
A market VaR excerpt is not a complete report of the portfolio's risks. Stress testing, concentration, liquidity risk, model weaknesses and control incidents may alter the decision even when the VaR number is unchanged. A model's historical inputs may poorly represent a new market regime; a valuation may omit execution costs or assume positions can be held fixed. The report should identify the limitations that matter for its particular decision rather than attach a generic disclaimer.
For systemically important banks, the Basel Committee's risk-reporting principles call for accurate, reconciled reports covering material risks, with analysis suited to recipients, suitable frequency and confidential distribution. These are a useful benchmark for reporting quality; the bank-specific framework is not a universal reporting template for every fund. A scheduled report also needs a route for urgent escalation when waiting for the next issue would delay a necessary decision.
Close the loop with a decision record
A useful action entry says what was decided, who had authority, who will implement it, when it is due, and how completion will be verified. “Monitoring continues” leaves those questions unanswered. If mitigation is agreed, the next report should show implementation and residual exposure; if an exception is granted, it should retain the approval, conditions and expiry. The report should carry open actions forward until there is evidence for closure.
Profits do not erase an unauthorised limit violation. Hull's discussion of limit discipline makes that governance point explicitly: judging a breach only by whether the trade later made money encourages the wrong behaviour. The report must preserve the exception and its response alongside any separate performance result.
Audience changes the emphasis, not the facts. Traders need operational restrictions and actions; risk managers need exceptions and measurement caveats; a committee needs the decision, alternatives and remaining exposure. Communicating to traders, risk managers and investment committees works through briefs built from a shared evidence pack. The full lesson requires membership, with a public preview.
John C. Hull, Risk Management and Financial Institutions, 4th edition (Wiley, 2015), Chapter 28, p. 579, “A Difficult Situation” — limit discipline when unauthorised risk-taking produces profits. Portfolio P, its people, reporting convention and amounts are original fictional examples.
