SuperAPI AI Directory

actuarial-analyst

Coding & development

View repository

Updated 10-16 04:56

About this skill

Actuarial Analyst

When to Use

  • Build or refresh loss development triangles and support IBNR / ultimate loss estimates
  • Prepare pricing support: experience summaries, trend, credibility-weighted indications
  • Validate, reconcile, and document actuarial data feeds and model inputs/outputs
  • Produce reporting packs: exhibits, bridges, roll-forwards, and memo-ready tables
  • Tie statutory or management reporting schedules to actuarial workpapers
  • Apply approved assumptions in models under actuary direction (not set enterprise policy)
  • Run reasonability checks and peer-review prep before actuary sign-off
  • Support pension or group benefit valuation data prep and schedule tie-outs (analyst depth)

When NOT to Use

  • Appointed actuary opinions, statement of actuarial opinion, or final sign-off → actuary
  • Engagement scoping, SOW, client decks, M&A consulting framing → actuarial-consulting
  • Enterprise assumption governance, assumption papers, and change control → assumption-setting
  • ALM strategy, duration matching, and investment policy → asset-liability-management
  • P&C coverage legal wording, claims handling, or underwriting authority → property-casualty-insurance
  • Life/health product design and reserving sign-off → life-health-insurance (line context); technical judgment → actuary
  • Pension plan design, funding policy, or trustee advice → pension-retirement-funds (strategy); analyst prep stays here
  • Chart design, dashboards, and visualization craft only → data-visualization
  • Source-system ETL, deduplication pipelines without actuarial judgment → data-scrubbing
  • Capital markets quant research and backtesting → quantitative-researcher

Related skills

Need Skill
Actuarial judgment, sign-off, capital overview, governance memos actuary
Consulting engagement, SOW, client communication, DD framing actuarial-consulting
Assumption governance, papers, and enterprise change control assumption-setting
ALM, duration, and asset–liability strategy asset-liability-management
P&C products, claims, and underwriting context property-casualty-insurance
Life/health benefits and product mechanics life-health-insurance
Pension funding policy and plan design (overview) pension-retirement-funds
Data cleaning pipelines and source normalization data-scrubbing
Charts, dashboards, and visual design data-visualization
Statistical/ML modeling beyond standard actuarial methods quantitative-researcher

Core Workflows

1. Task intake and scope (analyst)

Before opening triangles or models:

  1. Assignment type — Reserving, pricing support, reporting tie-out, model run, data prep
  2. Basis — Statutory, GAAP, IFRS 17, management; valuation date and cutoff
  3. LOB and segment — Direct vs assumed; reinsurance netting rules
  4. Deliverable — Exhibit list, workpaper IDs, deadline, reviewer (actuary name if known)
  5. Source of truth — Prior memo, approved assumptions version, model build ID
  6. Out of scope — Flag sign-off, legal interpretation, or assumption changes needing actuary / assumption-setting

See references/actuarial_analyst_scope.md.

2. Reserving and loss development

  1. Pull paid, incurred, case, and count triangles by origin and development age
  2. Run diagnostics (factor stability, calendar-year check, paid/incurred divergence)
  3. Apply selected methods (chain ladder, BF, Cape Cod) per actuary instruction—document picks and alternatives
  4. Build IBNR and ultimate tables; reconcile to prior evaluation
  5. Prepare PYD bridge and reserve roll-forward exhibits
  6. Escalate tail, cat, or large-loss treatment questions to actuary

See references/reserving_and_loss_development.md.

3. Pricing and rate indication support

  1. Aggregate experience by segment (exposure, frequency, severity, pure premium)
  2. Compute trend and on-level adjustments per documented methodology
  3. Apply credibility weights as specified; show gross vs credibility-weighted indications
  4. Summarize indicated rate change vs current; tie to loss ratio targets if provided
  5. Document data cuts, outliers, and one-time events in a support schedule
  6. Outline GLM or multivariate specs for actuary or modeler—do not present as final model without review

See references/pricing_and_rate_indication.md.

4. Data validation and model I/O

  1. Reconcile premium, exposure, and claim feeds to control totals
  2. Validate triangle dimensions, accident/valuation dates, and duplicate keys
  3. Compare model inputs to source exhibits; log exceptions with $ impact where possible
  4. Review model outputs for reasonability vs prior run and plan metrics
  5. Archive run logs, parameter files, and data cut manifests for audit trail
  6. Route assumption changes to actuary; do not alter approved assumption tables without approval

See references/data_validation_and_model_io.md.

5. Reporting and workpapers

  1. Build exhibit index aligned to filing or management pack structure
  2. Produce bridges (reserve change, premium earned, claim counts) with footnoted definitions
  3. Tie external schedules (statutory blanks, management reports) to internal workpapers
  4. Draft memo support tables (facts and exhibits)—not full executive narrative unless asked
  5. Version files: YYYYMMDD_basis_LOB_vNN with change log tab or README row

See references/reporting_and_workpapers.md.

6. Quality review and documentation

  1. Complete analyst checklist before actuary review (below)
  2. Ensure formulas trace and no hard-coded plugs without comment
  3. Cross-foot subtotals and tie to GL or bordereaux where required
  4. List open items with owner and target date
  5. Package review folder: inputs, outputs, diagnostics, reconciliation, change log

See references/quality_review_and_documentation.md.

Analyst checklist (before actuary review)

Analyst QA:
- [ ] Valuation date, basis, and segment definitions match memo / prior period
- [ ] Triangles tie to source data; paid + case = incurred (where applicable)
- [ ] Development factors documented with $ impact of alternative picks (if material)
- [ ] Reserve roll-forward reconciles opening + movement = closing
- [ ] Pricing experience A/E and exposure bases foot to bordereaux
- [ ] Credibility and trend applied per approved assumption memo
- [ ] Model run ID and assumption version recorded
- [ ] Exhibits numbered; cross-references in memo outline updated
- [ ] Open questions listed for actuary (not buried in cells)

Key metrics (analyst must define each time)

Metric Analyst responsibility
Earned premium Match earning pattern to loss basis in exhibit footnote
Loss ratio State numerator (incurred vs paid) and denominator
Development factor Age-to-age vs cumulative; calendar vs accident year
IBNR Definition (incurred vs paid basis); split case/IBNR if required
Pure premium Exposure unit documented (car-year, member month, payroll)
Indicated change Gross vs credibility-weighted; constraint notes from underwriting
PYD Bridge to prior quarter/year with volume vs rate vs development split

Tool patterns (agnostic)

Task Typical approach
Triangles and factors Spreadsheet triangles; SQL/R aggregation; actuarial reserving tools
Reconciliation VLOOKUP/XLOOKUP keys; SQL joins; Python pandas groupbys
Version control Dated folders; change log sheet; avoid "final_final" without version ID
Model runs Parameter export; diff prior vs current outputs; checksum on input files
Documentation Exhibit index; definition tab; color only for review flags, not logic

Prefer reproducible steps actuaries can re-run; avoid opaque manual cell edits without audit trail.

Line-of-business notes (analyst)

LOB Analyst focus
P&C Class/territory splits; cat and large-loss flags; long-tail tail factors
Life In-force census reconciliations; experience study tables—not product filing
Health Seasonality in monthly data; benefit change flags in A/E
Reinsurance Ceded vs net triangles; treaty mapping; clash aggregates if instructed
Pension Census and liability roll-forward data; tie to actuary valuation file

Deliverable standards (analyst)

Deliverable Minimum content
Triangle workbook Data tab, triangle, factors, ultimates, diagnostics, reconciliation
IBNR support Method, selections, ultimates, PYD bridge, open questions
Pricing support Segments, experience, trend, credibility, indicated change table
Model run pack Inputs manifest, parameters, outputs, reasonability vs prior
Reporting tie-out Schedule line ↔ workpaper cell map; unexplained differences listed
Review package QA checklist, change log, exhibit index

State limitations (data gaps, one-time events, immature years). Do not present analyst work as opinion, legal advice, or filed regulatory submission without qualified actuary and counsel review.

Assignment type matrix

Trigger phrase Primary workflow Lead reference
triangle analysis / loss development Reserving support reserving_and_loss_development.md
IBNR / reserve analysis / reserve roll-forward Ultimates and bridges reserving_and_loss_development.md
rate indication / pricing support Experience and indication pricing_and_rate_indication.md
experience study (tables) A/E prep; assumption change → actuary pricing_and_rate_indication.md
actuarial workpaper / actuarial reporting Exhibits and tie-outs reporting_and_workpapers.md
model run / model output review I/O validation data_validation_and_model_io.md
credibility (application) Weighting per memo pricing_and_rate_indication.md

When tasks span reserving and pricing on the same block, separate workpapers and flag independence concerns to the actuary.

Ethics and reliance (analyst)

  • Label work as draft or analyst support until actuary review
  • Cite data limitations, restatements, and one-time events in open items
  • Do not change approved assumptions locally—request update through actuary and assumption-setting
  • Distinguish technical indication from implemented rate or posted reserve
  • Refer legal or filing questions to counsel; refer sign-off to actuary

When to load references

  • Scope, boundaries, ethics → references/actuarial_analyst_scope.md
  • Reserving and triangles → references/reserving_and_loss_development.md
  • Pricing and indication → references/pricing_and_rate_indication.md
  • Data and model I/O → references/data_validation_and_model_io.md
  • Reporting and workpapers → references/reporting_and_workpapers.md
  • QA and documentation → references/quality_review_and_documentation.md