Dividend Cut Probability

• dividends
  Full dividend history including ex-dates, declaration amounts (cents per share), and event type (interim / final / special). Used to compute the issuer's historical cut rate and the trailing payout pattern.
• income_statements + cash_flow_statements
  Net profit, operating cash flow, and free cash flow for the trailing 4 reporting periods. Used to compute payout-ratio stretch (dividend / FCF) and coverage trends.
• analyst_forecasts (where available)
  Sell-side EPS forecasts for the next reporting period. Used for earnings-revision momentum input. Optional — issuers without analyst coverage receive a neutral input score on this dimension.
• announcements (continuous-disclosure flags)
  Profit warnings, dividend-policy statements, and capital-raise announcements within the last 6 months. Triggers an additive boost to the cut probability when material warnings are on file.

Source code: lib/dividend-cut-probability.ts

• Low (<25%)
  No material stress signals; payout coverage healthy.
• Medium (25–60%)
  Some stretch in payout coverage or analyst revisions; warrants monitoring.
• High (60%+)
  Multiple stress signals firing; payout coverage thin or deteriorating.

• The score is reported as a probability, not a binary forecast. UI never says “will cut”; it says “X% probability of cut at the next declaration”.
• The score is not an investment recommendation. We do not say “sell” or “hold” in connection with the score, and the FMA fair-dealing language rules apply to all surfaces.
• Per the IR Memos load-bearing rule, this score is a prompt INPUT to AI-generated memo content, never cited in memo prose. The probability appears in clearly-labelled internal-review sidebars only.

These are the known limits of the methodology. We disclose them publicly so readers can weigh the observation in context, and so that any qualified-privilege defence under the Defamation Act 1992 §16(2) rests on demonstrably fair and accurate reporting.

• The score is based on backward-looking financial data. A board with strong forward visibility (signed contracts, completed asset sales, refinancing arrangements) may comfortably sustain a dividend that the score flags as at risk.
• Analyst-forecast input is optional and is set to a neutral score for issuers without sell-side coverage. Smaller NZX issuers may have less reliable forecast input.
• The base-rate floor means issuers that have never cut their dividend cannot score below ~5%, even when stress signals are absent. Conversely, issuers with frequent historical cuts will score higher than newer payers in similar financial shape.
• We do not model issuer-specific dividend-policy frameworks (e.g. payout-ratio targets, progressive dividend commitments) explicitly. A board with a stated 50% payout ratio target is treated the same as a discretionary payer.
• Special dividends are excluded from the base-rate calculation but contribute to the trailing payout numerator. This can briefly inflate the score around the months following a special.
• The score does not predict the magnitude of a cut, only the likelihood of one occurring at the next declaration.