Actuary

Actuary

When to Use

• Price insurance or reinsurance products (indication, renewal, new business)
• Estimate reserves, IBNR, and loss development (paid/incurred triangles)
• Build or review mortality, morbidity, lapse, and expense assumptions
• Run experience studies and apply credibility weighting
• Explain loss ratio, combined ratio, and underwriting metrics
• Frame capital and risk metrics (RBC, SCR overview, economic capital concepts)
• Compare product design tradeoffs (benefits, riders, deductibles, guarantees)
• Document actuarial assumptions, methods, and model governance for sign-off
• Outline IFRS 17 / Solvency II / NAIC reporting concepts at technical overview level

When NOT to Use

• Corporate FP&A, budgets, variance commentary, or non-insurance KPI dashboards → financial-analyst (if installed)
• Investment banking comps, DCF equity valuation, or M&A pitch models → comps-analysis, dcf-model (if installed)
• Contract legal interpretation, policy wording, or regulatory enforcement advice → commercial-counsel
• IFRS general ledger recognition without insurance contract measurement lens → ifrs
• Clinical trial design, biostatistics, or epidemiology outside insurance morbidity → health/clinical skills
• Building production software, ETL, or model code only without actuarial judgment → data-scientist, senior-software-engineer
• SOC 2 / ISO audit evidence and control mapping without actuarial models → compliance-engineer
• Executive strategy without pricing/reserving/capital math → business-consultant
• Treasury, fundraising, and cash runway → cfo-advisor (if installed)

Related skills

Core Workflows

1. Engagement scoping

Before any calculation:

1. Line of business — Life, health, P&C, annuity, reinsurance; direct vs assumed 2. Decision — Price change, reserve opinion, assumption update, product launch, regulatory filing 3. Data — Policy/claim bordereaux, triangles, census, experience periods, exclusions 4. Materiality — Volume, tail risk, new riders, or cohort changes driving review 5. Governance — Sign-off chain, model version, effective date, peer review requirements

See `references/actuary_scope_and_principles.md`.

2. Pricing and product design

1. Define coverage, limits, deductibles, and rating variables 2. Segment risks (territory, class, age band, hazard) with sufficient volume 3. Estimate frequency × severity or tabular costs; load for expenses, profit, and risk margin 4. Compare indicated vs current rates; document constraints (competitive, regulatory caps) 5. Stress key assumptions (loss trend, inflation, catastrophe load) 6. Summarize rate change and expected loss ratio / combined ratio impact

See `references/pricing_and_product_design.md`.

3. Reserving and loss development

1. Aggregate paid and incurred losses by accident year and development lag 2. Select development factors (chain ladder, BF, Cape Cod) with diagnostics 3. Estimate IBNR and total ultimate; separate case vs bulk reserves if needed 4. Reconcile to prior evaluation; explain changes (development, large losses, methodology) 5. Document tail selection and emergence patterns for long-tail lines

See `references/reserving_and_loss_development.md`.

4. Assumptions and experience studies

1. Define exposure base and study period; handle data quality and outliers 2. Compare actual vs expected (A/E) by key dimensions 3. Propose assumption updates with credibility (limited fluctuation, Bühlmann, or judgment) 4. Separate trend, level, and volatility where material 5. Version assumptions with effective dates and rationale for audit trail

See `references/assumptions_experience_studies.md`.

5. Capital and regulatory overview

1. Identify regime (NAIC RBC, Solvency II SCR, internal economic capital) 2. Map main risk drivers: underwriting, market, credit, operational (high level) 3. Relate reserves and pricing to solvency metrics—not a substitute for appointed actuary sign-off 4. Flag disclosure topics (IFRS 17 CSM, risk adjustment, LIC) for coordination with ifrs 5. Escalate legal or filing interpretation to counsel and appointed actuary

See `references/capital_risk_and_regulatory_overview.md`.

6. Model governance and deliverables

1. Inventory models (pricing, reserving, ALM); owner and validation status 2. Document inputs, methods, outputs, and limitations 3. Run sensitivity and reasonability checks; retain reproducible workpapers 4. Produce actuarial memo or exhibit: executive summary, exhibits, assumptions appendix 5. Archive versioned data cuts and sign-offs per model risk policy

See `references/model_governance_and_deliverables.md`.

Key metrics (insurance)

Clarify numerator/denominator conventions in every exhibit footnote.

Line-of-business notes

P&C

• Ratemaking on exposure (payroll, sales, vehicles) with class plans and territory
• Long-tail lines need tail factors and paid/incurred reconciliation
• Catastrophe loads separate from attritional; event years flagged in triangles

Life

• Mortality, lapse, and expense drive profit testing and reserves
• Cash-value products need interest and crediting assumptions coordinated with ALM
• Select vs ultimate periods for recently underwritten cohorts

Health

• Morbidity trend split utilization vs unit cost when data allows
• Seasonality and benefit design changes (deductible, network) dominate A/E
• Risk adjustment and pooling (ACA, large group) affect segment homogeneity

Annuity

• Longevity and withdrawal drive guarantee costs; scenario sets for sensitivities
• Hedge effectiveness is investment-led; actuary supplies liability cash-flow shapes

Data requests (starter checklist)

When the user has not supplied data, ask for:

1. Valuation date and reporting basis (statutory, GAAP, IFRS 17) 2. Triangle or bordereaux with field definitions (accident date, paid, case, count) 3. Exposure definition and earning pattern 4. Reinsurance structure (quota share, excess, aggregates) and netting rules 5. Large loss and catastrophe treatment in prior studies 6. Prior actuarial memo, approved assumptions, and model version

Reasonability and peer review

Before sign-off, complete:

• [ ] Triangle diagnostics (stability, tail, calendar-year check)
• [ ] Implied ultimate LR vs pricing and plan
• [ ] Bridge of reserve change to prior quarter (volume, PYD, method)
• [ ] Assumption changes tied to experience study with credibility
• [ ] Sensitivities on top 3 drivers documented
• [ ] Independent reviewer for material models (per tiering policy)

Deliverable standards

Always state uncertainty: ranges, sensitivities, and data limitations. Do not present model output as legal, regulatory, or statutory filing without qualified human review.

Ethics and reliance

• Disclose conflicts when advising both pricing and reserving on the same block without review
• Cite data limitations and one-time events affecting experience
• Distinguish indication (technical) from implemented rate (commercial constraints)
• Refer legal interpretations of policy language or filing requirements to commercial-counsel
• Coordinate IFRS 17 measurement presentation with ifrs; actuary owns cash-flow assumptions, not note legal wording

When to load references

• Scope, principles, ethics → references/actuary_scope_and_principles.md
• Pricing and product → references/pricing_and_product_design.md
• Reserving and triangles → references/reserving_and_loss_development.md
• Assumptions and experience → references/assumptions_experience_studies.md
• Capital and regulation (overview) → references/capital_risk_and_regulatory_overview.md
• Governance and memos → references/model_governance_and_deliverables.md

Actuary scope and principles

Table of contents

1. Role boundary 2. Lines of business 3. Actuarial standards of practice 4. Professional judgment 5. Data and quality 6. Communication with stakeholders

Role boundary

Actuarial work supports insurance economics: expected claims, expenses, timing of cash flows, uncertainty, and solvency. It is distinct from:

Lines of business

Scope each engagement to one measurement basis (e.g., statutory vs GAAP vs IFRS 17) and document conversions separately.

Actuarial standards of practice

When jurisdiction-specific standards apply (e.g., US ASOPs, IAA standards), align workpapers to:

• Clear purpose of the assignment (pricing vs reserving vs opinion)
• Documentation sufficient for another qualified actuary to evaluate
• Disclosure of methods, data, and material limitations
• Conflict identification when roles combine (pricing and reserving on same block)

Do not claim appointed actuary or statement of actuarial opinion authority unless the user holds that role.

Professional judgment

Use judgment when:

• Data is thin (new product, new territory)
• Tail behavior dominates (med mal, workers comp)
• External shocks break history (pandemic, inflation step-change)

Document alternative reasonable views and why the selected view was chosen. Quantify impact where possible.

Data and quality

Minimum data checklist:

• Policy/claim record completeness and lag
• Definition alignment (accident vs report vs paid date)
• Large loss handling and catastrophe coding
• Reinsurance recoverables and netting rules
• Restatements between periods

Flag reconciliation breaks before fitting models.

Communication with stakeholders

Use plain language for ranges and sensitivities; reserve technical appendices for exhibits.

Assumptions and experience studies

Table of contents

1. Study design 2. Actual vs expected 3. Mortality and longevity 4. Morbidity and health trend 5. Lapse and persistency 6. Credibility 7. Assumption governance

Study design

1. Objective — Update pricing, valuation, or both 2. Population — In-force, new business, closed block; include/exclude rules 3. Period — Enough exposure for stability; exclude distortion years or footnote 4. Basis — Exposure measure (earned premium, member months, life years) 5. Dimensions — Age, gender, class, duration, geography, product

Data prep checklist:

• Match policy admin to claims keys
• Handle reinstatements, riders, and terminations
• Define claim types consistently (medical vs RX vs dental)

Actual vs expected

A/E ratio = actual experience / expected experience (on same exposure basis)

Present A/E by dimension and in aggregate; investigate interaction effects before changing factors.

Mortality and longevity

Compare to standard tables (e.g., industry tables) with scaling factors. Document selection period and ultimate period separately for some products.

Morbidity and health trend

• Utilization — Visits, admits, RX fills per member
• Unit cost — Allowed amount per service; fee schedule inflation
• Mix — Shift to high-cost sites of care
• Trend — Separate frequency and unit cost trend where possible

Health experience often needs seasonality and network changes called out explicitly.

Lapse and persistency

Model dynamic lapse (overview) when interest rates or guarantees dominate.

Credibility

Blend observed experience with prior (manual, industry, or last approved assumption):

Report full credibility exposure threshold and Z used per segment.

Assumption governance

Assumption change package:

Maintain assumption library linked to models; no ad hoc hardcoding without version bump.

Capital, risk, and regulatory overview

Table of contents

1. Purpose and limits 2. NAIC and US statutory (overview) 3. IFRS 17 (overview) 4. Solvency II (overview) 5. Economic capital concepts 6. Risk metrics map 7. Coordination with finance and counsel

Purpose and limits

This reference provides orientation for actuaries coordinating with finance, risk, and legal teams. It is not:

• Legal advice on filing requirements
• A substitute for appointed actuary sign-off
• Detailed IFRS journal entries (see ifrs skill)
• Investment or ALM strategy execution

Escalate binding regulatory positions to qualified professionals in the relevant jurisdiction.

NAIC and US statutory (overview)

Actuaries supply reserve adequacy analysis supporting opinion; coordinate asset adequacy testing with investments.

IFRS 17 (overview)

Insurance contracts standard (high level):

Coordinate measurement outputs with `ifrs` skill for presentation, disclosure, and reconciliation to GL.

Solvency II (overview)

Technical provisions split best estimate and risk margin; actuaries often own BE assumptions. SCR uses standard formula or internal model (overview)—not full internal model build here.

Economic capital concepts

Internal frameworks often estimate capital as VaR or TVaR of surplus over a horizon:

economic capital ≈ quantile(loss distribution) − expected loss

Link to:

• Underwriting — catastrophe, reserve error, premium risk
• Market — rates, equity, spread (ALM)
• Credit — reinsurance default, bond downgrades

Document diversification assumptions and correlation matrices at high level; refer to enterprise risk for governance.

Risk metrics map

Present sensitivities (interest ±100 bps, mortality ±%, cat event) as scenarios, not single points, when advising committees.

Coordination with finance and counsel

Document open issues list for appointed actuary or auditor review before external distribution.

Model governance and deliverables

Table of contents

1. Model inventory 2. Validation lifecycle 3. Documentation standards 4. Actuarial memo structure 5. Sensitivity and reasonability 6. Version control and audit trail 7. Model risk escalation

Model inventory

Maintain a register for each production model:

Separate pricing, reserving, ALM, and capital models; document dependencies when outputs feed each other.

Validation lifecycle

design → implement → independent review → approval → monitor → change control

Tier 1 models require independent validation (second actuary or model risk function).

Documentation standards

Each model should have:

1. Purpose and scope — Decisions supported, exclusions 2. Theory — Formulas, assumptions, references 3. Data — Sources, transforms, quality checks 4. Parameters — Tables with version and effective date 5. Output mapping — How results post to finance or filings 6. Limitations — Known weaknesses and monitoring

Avoid orphan spreadsheets: named ranges, change log tab, locked production tabs.

Actuarial memo structure

Recommended sections:

1. Executive summary — Conclusion, key numbers, recommended action 2. Scope and data — Period, entities, exclusions 3. Methods — With alternatives considered 4. Results — Exhibits, bridges, segment detail 5. Sensitivities — Key drivers table 6. Assumptions — Cross-reference assumption library 7. Governance — Preparer, reviewer, approver, dates 8. Appendices — Triangles, tables, technical detail

Write so a non-actuary executive understands the decision in page one; specialists use appendices.

Sensitivity and reasonability

Standard sensitivity pack:

Reasonability checks:

• Implied ultimate LR vs pricing indication
• Calendar-year loss vs plan
• Industry benchmarks (with caveats)
• Cash flow emergence vs paid pattern

Version control and audit trail

Store:

• Data cut timestamp and query definition
• Model semver or build ID
• Parameter file hash
• Output file with read-only archive

Retention per company policy (often ≥ 7 years for statutory work).

Model risk escalation

Escalate when:

• Material model change near filing deadline
• Backtest failure persists two quarters
• Data quality break affects key line
• Undocumented manual override in production
• Conflict of interest (same preparer prices and opines reserves without review)

Route to chief actuary, model risk committee, or audit as defined by policy.

Pricing and product design

Table of contents

1. Pricing workflow 2. Segmentation 3. Loss and benefit costs 4. Loadings 5. Product design tradeoffs 6. Metrics and monitoring 7. Common pitfalls

Pricing workflow

define coverage → segment risks → estimate pure premium → add loadings → compare to current → document

Segmentation

Credible segments have enough exposure to estimate frequency and severity (or tabular rates) with acceptable error.

Document minimum volume rules and off-book segments sent to manual review.

Loss and benefit costs

P&C

• Frequency × severity on earned exposure (or policy count for some lines)
• Trend selections: loss trend, exposure trend, social inflation where relevant
• Large loss caps or separate modeling for property cat

Life / health

• Tabular costs from mortality/morbidity tables adjusted for underwriting class
• Benefit duration and utilization for health products
• Lapse impact on persistency and unit costs

Annuity

• Longevity and withdrawal behavior; guarantee cost via stochastic or scenario set (overview)

Loadings

Show indicated loss ratio before and after loadings:

indicated LR = pure premium / gross premium (excl. reinsurance ceded, per convention)

State convention explicitly in the memo.

Product design tradeoffs

Present option sets with expected LR, volume elasticity assumptions (if provided), and capital sensitivity qualitatively.

Metrics and monitoring

Set monitoring triggers post-launch: A/E by cohort, mix shift, large loss emergence.

Common pitfalls

• Using written premium with incurred losses without earning adjustment
• Ignoring claim reporting lag in experience picks for pricing
• Double-counting reinsurance or subrogation
• Stale trend factors after macro shocks
• Pricing to competitor without reconciling own experience

Reserving and loss development

Table of contents

1. Triangle fundamentals 2. Development methods 3. IBNR components 4. Diagnostics 5. Large losses and catastrophes 6. Reserve reconciliation 7. Workpaper exhibits

Triangle fundamentals

Standard loss development triangle axes:

• Origin period (accident year, policy year, underwriting year)
• Development age (months or years since origin)
• Values — paid losses, incurred losses, case reserves, claim counts

Align valuation date and cutoff procedures across periods.

Development methods

Tail factor selection: fit exponential/inverse power, industry benchmarks, or judgment for oldest age.

Document factor selections vs alternative picks with dollar impact.

IBNR components

total reserve ≈ case outstanding + IBNR (+ IBNER if split)
ultimate = reported to date + IBNR (incurred basis)

Explain prior-year development (PYD) drivers in bridge:

opening reserve + incurred losses − paid − closing reserve = PYD

Diagnostics

Run before finalizing:

• Calendar-year vs accident-year implied ultimates
• Factor stability across recent ages
• Paid/incurred divergence
• Claim count vs severity reconciliation
• Outlier accident years (cat, large loss, one-time events)

Use bootstrap or Mack-style uncertainty (overview) when stakeholders need range, not point only.

Large losses and catastrophes

Remove or cap catastrophe years from factor selection with explicit add-back for cat reserve.

Reserve reconciliation

Reconcile to:

• Prior quarter/year reserve
• Pricing expected loss ratio (if linked)
• Reinsurance statements
• Statutory or GAAP reporting lines (coordinate with accounting)

Material changes (> threshold) require narrative: methodology, data, large losses, assumption.

Workpaper exhibits

Minimum exhibits:

1. Paid and incurred triangles (last N origin years) 2. Selected age-to-age factors and tail 3. Ultimate loss table by origin year 4. IBNR summary and roll-forward 5. PYD bridge to prior evaluation 6. Sensitivity: ± one factor or ± tail

Store data cut hash, model version, and preparer/reviewer sign-off.