IAI CP3 cheat sheet

Risk = uncertainty of outcome; can be upside or downside. Risk management involves identifying, assessing, controlling, financing, and monitoring risks.

STARR framework: Stakeholders, Terms of reference, Actuarial analysis, Risks, Rewards. Use to structure any risk management problem.

Risk appetite is the amount of risk an organisation is willing to accept; risk tolerance is the acceptable deviation from appetite.

Hazard = source of potential harm; peril = event causing loss; exposure = extent to which an entity is subject to loss.

Risk control techniques: avoid, reduce (frequency/severity), transfer (insurance, hedging), retain (self-insurance).

Diversification reduces unsystematic risk; systematic risk (market, interest rate, inflation) cannot be diversified away.

Capital allocation: risk-based capital = f(risk measure). Common measures: VaR, TVaR, probability of ruin.

Enterprise risk management (ERM) integrates risk across silos; uses risk registers, heat maps, and scenario analysis.

Liquidity risk = risk of not meeting obligations when due; solvency risk = risk of assets < liabilities.

Life products: term assurance, whole life, endowment, annuities, unit-linked, universal life. Key features: death benefit, survival benefit, premium pattern, investment risk.

Pension schemes: DB (defined benefit) promises a benefit formula; DC (defined contribution) promises only contributions; risk allocation differs.

Valuation of life liabilities: use discount rate, mortality/morbidity assumptions, expenses, and persistency.

With-profits funds: use reversionary bonuses, terminal bonuses, and smoothing; estate (free assets) supports smoothing.

Annuity pricing: depends on mortality, interest rates, expenses, and anti-selection. Mortality improvements increase cost.

Pension funding: contribution rate = (benefits + expenses - assets) / (present value of future salaries).

Guaranteed annuity rate (GAR) options create interest rate risk for insurers; hedge via matching assets or reinsurance.

Life insurance regulation: Solvency II (EU) or local RBC; policyholder protection schemes; capital requirements for life risk.

General insurance covers property, liability, motor, marine, etc. Policies are typically annual and repriced at renewal.

Claims reserving: incurred but not reported (IBNR), outstanding claims, and unexpired risk reserve (UPR).

Chain-ladder method: estimate ultimate claims using development factors; assumes stable development pattern.

Bornhuetter-Ferguson method: combines chain-ladder and expected loss ratio; more stable for volatile or new classes.

Premium rating: risk factors (age, location, vehicle type), experience rating, no-claims discounts, and exposure curves.

Reinsurance: proportional (quota share, surplus) and non-proportional (excess of loss, stop loss).

Catastrophe risk: modelled using frequency-severity, event simulation, and exposure data; return period (e.g., 1-in-200 year).

Underwriting cycle: soft market (low premiums, loose terms) vs hard market (high premiums, strict terms).

Premium = risk premium + expense loading + profit margin + contingency; must cover expected claims and costs.

Health insurance products: PMI (private medical insurance), income protection, critical illness, long-term care.

Morbidity = rate of illness/injury; incidence rate = number of new cases per population at risk.

Critical illness: pays lump sum on diagnosis of specified conditions; definitions and exclusions are critical.

Income protection: pays replacement income during disability; benefit period, deferred period, and occupation definitions matter.

Long-term care: covers cost of care; triggers based on ADL (activities of daily living) limitations or cognitive impairment.

Health insurance pricing: use morbidity tables, select period, and loading for adverse selection.

Medical inflation is often higher than general inflation; needs separate assumption in pricing and reserving.

Regulatory issues: community rating (e.g., in some markets), guaranteed renewability, and pre-existing condition exclusions.

Bond price = present value of future coupons and principal; inversely related to yield.

Duration measures price sensitivity to yield changes; modified duration = Macaulay duration / (1 + y/m).

Convexity measures curvature of price-yield relationship; improves duration-based estimates for large yield changes.

Derivatives: forwards, futures, options, swaps. Forwards are OTC; futures are exchange-traded and standardised.

Option pricing: Black-Scholes model assumes lognormal returns, constant volatility, no arbitrage, frictionless markets.

Put-call parity: C - P = S - PV(K) for European options on non-dividend-paying stock.

Interest rate swaps: exchange fixed for floating; used to hedge interest rate risk. Swap rate = fixed rate that makes swap value zero.

Equity indices, commodities, and FX are underlying assets; correlation affects portfolio risk and derivative pricing.

Actuarial communication must be clear, concise, and tailored to the audience (technical vs non-technical).

Structure a report: executive summary, introduction, data, methodology, results, limitations, conclusions, recommendations.

Use plain language; avoid jargon unless defined. Use graphs and tables to illustrate key points.

State assumptions and uncertainty explicitly; distinguish between facts and opinions.

Professional standards (e.g., TAS, APS) require transparency, integrity, and competence.

Peer review and documentation are essential for quality assurance.

Effective communication includes active listening and responding to feedback.

When communicating risk, use absolute and relative measures; avoid misleading scales.

Data visualisation: choose appropriate chart type, label axes, and avoid clutter.

Model purpose: understand, predict, or value. A model is a simplified representation of reality.

Model risk = risk of error due to incorrect model, assumptions, or implementation.

Model validation: independent review, back-testing, sensitivity analysis, and scenario testing.

Professionalism: act with integrity, competence, and objectivity; comply with actuarial standards and code of conduct.

Documentation: record data sources, assumptions, methodology, and results for reproducibility.

Communication of model results: include limitations, uncertainty, and sensitivity to assumptions.

Ethical dilemmas: conflicts of interest, confidentiality, and whistleblowing. Follow professional guidance.

Continuing professional development (CPD) is required to maintain competence.

Judgement: use professional judgement where data is limited; disclose and justify.

Business environment: economic, political, social, technological, legal, environmental (PESTLE) factors.

Stakeholders: shareholders, policyholders, regulators, employees, government. Balance conflicting interests.

Corporate governance: board structure, risk committee, audit committee, internal controls.

Ethics: utilitarianism, deontology, virtue ethics. Actuarial code emphasises integrity, objectivity, and confidentiality.

Conflicts of interest: identify, disclose, manage, or avoid.

Corporate social responsibility (CSR) and sustainability affect reputation and long-term value.

Financial reporting: IFRS, local GAAP; actuarial involvement in valuation and disclosure.

Regulatory environment: prudential regulation (solvency) and conduct regulation (fair treatment of customers).

Professional ethics: honesty, fairness, transparency, and accountability.

Data quality: accuracy, completeness, consistency, timeliness, validity, uniqueness.

Machine learning: supervised (regression, classification), unsupervised (clustering, PCA), reinforcement learning.

Overfitting: model too complex; use cross-validation, regularisation, and hold-out samples.

Big data: volume, velocity, variety, veracity, value. Tools: SQL, Python, R, cloud platforms.

Generalised linear models (GLMs): extend linear regression to non-normal distributions; link function connects mean to linear predictor.

Decision trees and random forests: interpretable but prone to overfitting; ensemble methods improve stability.

Neural networks: flexible but black-box; require large data and careful validation.

Data ethics: privacy, consent, bias, and fairness. Comply with GDPR and other regulations.

Actuarial applications: pricing, reserving, fraud detection, customer segmentation, and risk management.

Solvency II: three pillars (quantitative, governance, disclosure). SCR (Solvency Capital Requirement) and MCR (Minimum Capital Requirement).

SCR calibrated to 99.5% VaR over one year; can use standard formula or internal model.

Own Risk and Solvency Assessment (ORSA) is a forward-looking assessment of risk and capital.

Capital resources: Tier 1 (highest quality), Tier 2, Tier 3. Eligible capital must cover SCR.

Regulatory capital vs economic capital: economic capital is based on internal risk assessment; regulatory is prescribed.

Capital management: dividends, reinsurance, subordinated debt, and equity issuance.

Stress testing and scenario analysis are required for capital adequacy.

Conduct regulation: fair treatment of customers, transparency, and complaints handling.

International standards: IAIS (insurance), IORP (pensions), Basel (banking).

ALM: match assets and liabilities by duration, currency, and nature. Immunisation when duration of assets = duration of liabilities.

Liability-driven investment (LDI): use bonds and swaps to hedge interest rate and inflation risk.

Asset allocation: strategic (long-term) and tactical (short-term). Consider risk tolerance, return targets, and constraints.

Mean-variance optimisation (Markowitz): maximise return for given risk; efficient frontier.

Diversification: reduce unsystematic risk by investing across asset classes, sectors, and geographies.

Liquidity management: hold sufficient liquid assets to meet obligations; consider cash flow matching.

Investment risk: market risk, credit risk, liquidity risk, concentration risk.

Performance measurement: time-weighted vs money-weighted returns; benchmarks and attribution analysis.

Regulatory constraints: quantitative limits on asset classes, currency matching, and custody requirements.