Use this guide to connect financial theory with property and casualty insurance decisions. Each concept explains a specific relationship, works through an original example and identifies a common error. Start with portfolio and valuation foundations, then apply them to underwriting profitability, risk transfer and capital decisions.
Portfolio theory and asset pricing
1. Holding-period return
Holding-period return includes both the change in asset value and distributions received, divided by the initial investment. Use a consistent period and specify whether distributions are reinvested. This simple formula assumes no intervening external contributions or withdrawals; those cash flows require a return measure that handles their timing.
Worked example: An asset bought for 80 pays 3 and is sold for 85. Its holding-period return is (85 − 80 + 3)/80 = 10%.
Mistake to avoid: Calculating only the price change and omitting the distribution.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
2. Compounding and geometric returns
Successive returns multiply wealth factors. The geometric mean is the constant periodic return producing the same ending wealth, whereas the arithmetic mean averages individual returns. Volatility can therefore reduce compound growth even when the arithmetic average is positive. Match annualization to the actual length and number of periods.
Worked example: Returns of 25% and −20% give 1.25 × 0.80 = 1.00. Wealth is unchanged, so the two-period geometric mean is 0%, despite a 2.5% arithmetic mean.
Mistake to avoid: Adding returns across periods to calculate cumulative investment growth.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
3. Expected return and variance
Expected return is the probability-weighted average of possible returns. Variance is the probability-weighted squared deviation from that average, and standard deviation restores the original return units. These quantities summarize location and dispersion, but they do not uniquely describe skewness, extreme losses or the full return distribution.
Worked example: Equally likely returns of 2% and 10% have mean 6%. Variance is 0.5(0.02 − 0.06)² + 0.5(0.10 − 0.06)² = 0.0016; standard deviation is 4%.
Mistake to avoid: Reporting variance as though it were a percentage standard deviation.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
4. Covariance and correlation
Covariance measures how two returns vary together. Correlation divides covariance by their standard deviations, creating a unitless measure between −1 and 1 when both variances are positive. Correlation describes linear association; a low value does not establish independence or exclude simultaneous extreme losses.
Worked example: Assets with standard deviations of 12% and 20% and correlation 0.25 have covariance 0.12 × 0.20 × 0.25 = 0.006.
Mistake to avoid: Substituting correlation directly for covariance in a portfolio variance formula.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
5. Two-asset portfolio risk
Portfolio expected return is the weighted average of component expected returns. Portfolio variance also includes covariance: w₁²σ₁² + w₂²σ₂² + 2w₁w₂Cov(R₁,R₂). Thus, averaging individual standard deviations generally gives the wrong risk. The formula applies to returns measured over the same period and on the same basis.
Worked example: A 50–50 portfolio of uncorrelated assets with standard deviations of 10% and 20% has variance 0.0125 and standard deviation approximately 11.18%.
Mistake to avoid: Calling 15%, the average component standard deviation, the portfolio’s standard deviation.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
6. The efficient frontier
A mean-variance efficient portfolio offers the highest expected return for its variance, or the lowest variance for its expected return, among feasible portfolios. Feasibility depends on constraints such as short-sale restrictions. Efficiency is conditional on estimated inputs and does not establish superiority under every investor objective or loss distribution.
Worked example: Portfolio A has expected return 7% and standard deviation 12%; B has 8% and 12%. If both are feasible, B dominates A in mean-variance terms.
Mistake to avoid: Selecting a dominated portfolio merely because its expected return is positive.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
7. Combining risky and risk-free assets
With a constant risk-free return, a portfolio investing fraction y in one risky portfolio has expected return r_f + y(E[R] − r_f) and standard deviation |y|σ. Borrowing corresponds to y greater than one, but the same straight-line relationship requires borrowing at the assumed rate and ignores transaction restrictions.
Worked example: Investing 60% in a portfolio returning 9% with 15% volatility and 40% at 3% gives expected return 6.6% and volatility 9%.
Mistake to avoid: Applying the lending rate to borrowing without checking the financing assumption.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
8. Sharpe ratios
The Sharpe ratio divides expected excess return over a risk-free benchmark by return standard deviation. It compares reward per unit of total volatility when horizons and conventions match. It can obscure asymmetric losses, and a ratio calculated from historical observations is an estimate rather than a known future performance measure.
Worked example: A portfolio with expected return 8%, risk-free rate 2% and standard deviation 12% has Sharpe ratio (8% − 2%)/12% = 0.50.
Mistake to avoid: Comparing monthly and annual ratios without consistent annualization.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
9. Systematic and diversifiable risk
Firm-specific uncertainty can often be reduced by holding many exposures, while common economic shocks remain. In a single-factor model with uncorrelated residuals, averaging many similarly sized positions reduces residual variance but preserves factor exposure. Diversification depends on dependence and concentration, rather than simply the number of holdings.
Worked example: For 25 equally weighted investments with independent residual standard deviation 20%, portfolio residual standard deviation is 20%/√25 = 4%. Shared market exposure remains.
Mistake to avoid: Assuming additional holdings eliminate losses driven by the same economic shock.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
10. Beta as market sensitivity
An asset’s market beta equals its covariance with the market divided by market variance. It measures linear sensitivity to market movements, rather than total volatility. Estimated beta depends on the market proxy, observation period and sampling uncertainty. A high-volatility asset can have low beta if much of its variation is idiosyncratic.
Worked example: An asset with market covariance 0.018 and market variance 0.015 has beta 1.2. Under the fitted linear model, a 1% market move contributes about 1.2% to its return.
Mistake to avoid: Treating beta as the asset’s standard deviation or a guaranteed response.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
11. The capital asset pricing model
CAPM links expected return to systematic market risk: E[R_i] = r_f + β_i(E[R_m] − r_f). It is an equilibrium model built on restrictive assumptions, including mean-variance investment decisions and a common investment opportunity set. Use it as a structured required-return estimate, with explicit choices for beta and the market premium.
Worked example: With risk-free return 3%, market premium 5% and beta 1.4, the CAPM expected return is 3% + 1.4 × 5% = 10%.
Mistake to avoid: Using the market’s total expected return where the formula requires its excess return.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
12. Multifactor return models
A multifactor model separates returns into an intercept, sensitivities to specified factors and a residual. Factors may represent market movements or other common drivers. A statistical factor regression does not automatically establish an equilibrium pricing relationship; expected premiums, interpretation and residual assumptions need separate justification.
Worked example: In a fitted excess-return model with zero intercept, market loading 0.8 and interest-rate factor loading −0.3, factor returns of 4% and 2% imply fitted excess return 2.6%.
Mistake to avoid: Adding factor returns without multiplying them by the relevant loadings.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
Market efficiency and behavioral finance
13. Forms of market efficiency
Weak-form efficiency concerns information in past prices; semi-strong efficiency concerns publicly available information; strong-form efficiency additionally includes private information. Each is a statement about an information set and prices, rather than a promise that prices never change or equal an observable true value. Stronger forms encompass the weaker information sets.
Worked example: A repeatable strategy using public earnings releases, after realistic costs and risk adjustment, would challenge semi-strong efficiency more directly than a strategy relying on confidential information.
Mistake to avoid: Treating one successful private-information trade as evidence against weak-form efficiency.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
14. Abnormal returns and the joint hypothesis
An abnormal return is the difference between an observed return and the return predicted by a chosen benchmark. Testing efficiency therefore also tests the benchmark model. Apparent outperformance can reflect omitted risk, estimation error or costs rather than mispricing. Specify whether the comparison concerns realized or expected returns.
Worked example: A strategy earns 11% when its risk model predicts 9%, giving a realized abnormal return of 2%. Costs of 1.5% reduce the net difference to 0.5%.
Mistake to avoid: Interpreting a positive abnormal return as conclusive proof of market inefficiency.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
15. Event-study interpretation
An event study compares observed returns around an event with benchmark expected returns. A simple cumulative abnormal return sums abnormal returns over a short event window. Separate the estimation window from the event window, and investigate overlapping announcements, information leakage and correlated observations before attributing the measured effect to one event.
Worked example: Three daily abnormal returns of −0.4%, 1.8% and 0.3% produce a cumulative abnormal return of 1.7%. A simultaneous acquisition announcement complicates attribution.
Mistake to avoid: Assigning every event-window price movement to the event being studied.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
16. Limits to arbitrage
A perceived pricing discrepancy need not offer a riskless, immediately realizable profit. Traders can face funding constraints, short-sale restrictions, uncertain convergence and losses before correction. Distinguish a theoretical arbitrage with locked-in payoffs from a convergence trade whose eventual success depends on future prices or continued access to financing.
Worked example: Two claims appear to have equal economic exposure but trade at 98 and 102. Borrowing the expensive claim costs 5 over the holding period, exceeding the apparent spread of 4.
Mistake to avoid: Calling a price difference exploitable without accounting for implementation costs and risks.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
17. Behavioral biases in financial decisions
Anchoring gives undue weight to an initial reference, while loss aversion assigns disproportionate weight to losses relative to a reference point. These mechanisms differ from rational changes in risk tolerance or information. Diagnose the decision process using relevant alternatives and updated evidence, rather than labeling every unfavorable outcome a behavioral error.
Worked example: A manager retains a bond bought at 100 solely to avoid selling at 94, despite a reassessment valuing it at 90. The purchase price is not a future cash flow.
Mistake to avoid: Using the original purchase price as the decisive estimate of current economic value.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
18. Data snooping and predictive validation
Searching many strategies increases the chance of finding impressive historical performance by accident. A genuine validation sample must remain outside model selection, feature selection and tuning. Include realistic trading costs and information availability dates. Repeatedly examining and adjusting to the holdout makes it part of the development process.
Worked example: Of 200 candidate signals, one performs well in development data but earns −1% after costs in an untouched later period. Its historical ranking alone does not establish predictive value.
Mistake to avoid: Calling a repeatedly consulted holdout an independent test.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
Financial risk and derivatives
19. No-arbitrage valuation
Under frictionless trading assumptions, portfolios with identical future cash flows must have identical prices. Otherwise, buying the cheaper portfolio and selling the more expensive one locks in a gain without future net obligations. Replication requires matching every relevant payment state, date and currency, with financing and trading actually feasible.
Worked example: A risk-free payment of 105 in one year is worth 100 at a 5% annual rate. An otherwise identical claim priced at 103 creates a theoretical arbitrage through sale and replication.
Mistake to avoid: Comparing claims with different credit risk as though their payoffs were identical.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
20. Forward prices and carrying costs
For a non-dividend-paying asset with no storage costs, the forward delivery price is the spot price accumulated at the financing rate. Income received while holding the asset reduces the replication cost. Distinguish the delivery price on a new contract from the current value of an existing contract with a previously fixed price.
Worked example: With spot price 200 and a one-year effective financing rate of 4%, the no-arbitrage forward delivery price is 208 under the stated assumptions.
Mistake to avoid: Calling 208 the upfront value of a newly initiated fair forward contract.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
21. Derivative payoff versus profit
A derivative payoff describes the contractual amount at settlement. Profit additionally recognizes premiums, financing and other relevant cash flows. For a call, terminal payoff is max(S_T − K, 0); for a put, it is max(K − S_T, 0). Keep the perspective of the buyer or seller explicit.
Worked example: A call with strike 50 costs 4 and expires with the asset at 57. Its payoff is 7 and profit is 3, ignoring financing and transaction costs.
Mistake to avoid: Describing a positive option payoff as a profitable trade without subtracting the premium.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
22. European put-call parity
For European options on a non-dividend-paying asset, with identical strike and expiry, put-call parity is C − P = S₀ − PV(K). The relationship follows by matching terminal payoffs. Dividends, exercise rights and market frictions require appropriate adjustments; the simple equality should not be transferred unchanged to every option contract.
Worked example: If S₀ = 60, PV(K) = 57 and the put costs 5, the matching call must cost 8 under the parity assumptions.
Mistake to avoid: Using parity between options with different expiration dates.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
23. Binomial replication
In a one-period binomial model, choose an asset holding and a risk-free position that reproduce the derivative payoff in both states. The replicating portfolio’s current cost gives the derivative value under no arbitrage. The hedge ratio is the difference in derivative payoffs divided by the difference in asset prices.
Worked example: An asset starts at 100 and ends at 120 or 80. A strike-100 call pays 20 or 0. At zero interest, 0.5 asset units and borrowing 40 replicate it, costing 10.
Mistake to avoid: Computing the hedge ratio using initial price rather than the difference between terminal prices.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
24. Risk-neutral probabilities
Risk-neutral probabilities make the discounted asset price a martingale in the valuation model. They are pricing weights, not necessarily forecasts of actual event probabilities. In a one-period model, the up-state weight is (1 + r − d)/(u − d), with d < 1 + r < u required for an interior no-arbitrage weight.
Worked example: With u = 1.3, d = 0.9 and r = 0.02, the up-state pricing weight is 0.30. A claim paying 40 only in that state is worth 0.30 × 40/1.02 ≈ 11.76.
Mistake to avoid: Interpreting the pricing weight as an estimated real-world probability.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
25. Delta hedging and residual risk
Delta measures the local change in derivative value for a small change in the underlying price, holding other inputs fixed. An offsetting underlying position can neutralize first-order price sensitivity. Delta changes as prices and time change, so a hedge requires rebalancing and remains exposed to curvature, discontinuous moves and other risk factors.
Worked example: A short position in 200 calls, each covering one asset unit and having delta 0.35, has delta −70. Holding 70 asset units makes the initial combined delta zero.
Mistake to avoid: Treating initial delta neutrality as protection against every future price movement.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
26. Gamma, vega and theta
Gamma measures how delta changes with the underlying price; vega measures sensitivity to volatility; theta measures sensitivity to time passing under a stated convention. These capture different risks. Local sensitivities approximate small changes and depend on the valuation model, so their units and sign conventions must be specified.
Worked example: If vega is 2.5 per volatility percentage point, a rise from 18% to 20% increases estimated option value by 5, holding other inputs fixed.
Mistake to avoid: Multiplying by 0.02 when vega is quoted per percentage point.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
27. Duration and convexity
For fixed cash flows, modified duration gives the first-order price response to a yield change. Convexity adds curvature: ΔP/P ≈ −D_modΔy + 0.5C(Δy)². Apply compatible yield conventions. Large changes, nonparallel curve shifts and embedded options can weaken an approximation based on one duration and convexity.
Worked example: A bond worth 1,000 with modified duration 5 and convexity 40 faces a yield increase of 0.01. Estimated change is −5% + 0.2% = −4.8%, giving value 952.
Mistake to avoid: Entering a one-percentage-point yield change as 1 rather than 0.01.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
28. Futures settlement and liquidity
Futures generally realize gains and losses through periodic settlement rather than deferring the entire result until maturity. This creates cash-flow needs even when the contract offsets an underlying economic exposure. Hedge effectiveness and funding capacity are separate questions; a later offsetting receipt may arrive after an immediate settlement payment is due.
Worked example: A short futures position covering 500 units loses 1,500 when the futures price rises by 3. The underlying asset gain may offset economically but provide no immediate cash.
Mistake to avoid: Assuming an economically balanced hedge cannot generate a cash shortage.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
29. Interest-rate swap valuation
A plain interest-rate swap exchanges specified fixed and floating payments on a notional amount. For a receive-fixed position, value equals the present value of fixed receipts minus floating payments. Project each floating payment and discount consistently with the valuation framework. The notional usually determines payments without itself being exchanged.
Worked example: On notional 2 million, annual fixed receipts at 4% are 80,000. If the applicable floating rate is 3.2%, the period’s net receipt is 16,000 before discounting.
Mistake to avoid: Including repayment of the notional when the contract exchanges only interest payments.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
30. Counterparty exposure and wrong-way risk
Counterparty exposure concerns amounts lost if a contractual counterparty fails. A simplified expected credit loss calculation multiplies exposure, default probability and loss given default when their treatment is compatible. Wrong-way risk arises when exposure increases as counterparty credit quality deteriorates, making an independence-based calculation potentially misleading.
Worked example: For fixed exposure 5 million, default probability 2% and loss given default 60%, simplified expected loss is 60,000. This ignores exposure changes and dependence.
Mistake to avoid: Assuming collateral or a favorable contract value eliminates all counterparty risk.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
Corporate finance and capital structure
31. Matching cash flows and discount rates
Discount rates must match the ownership claim and risk of the cash flows being valued. Cash available to equity holders is discounted at the required equity return; suitably defined operating cash flows may use a firm-wide rate. Keep nominal versus real amounts, currency and timing consistent throughout the valuation.
Worked example: An equity distribution of 12 in one year has present value 12/1.08 ≈ 11.11 at an 8% required equity return.
Mistake to avoid: Discounting equity distributions at a borrowing rate merely because that rate is lower.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
32. Net present value and internal rate of return
Net present value subtracts investment cost from discounted future cash flows. Internal rate of return is a rate making NPV zero. For mutually exclusive projects, NPV measures value added at the applicable opportunity cost; IRR rankings can conflict because of scale or timing, and unusual cash-flow patterns can produce multiple roots.
Worked example: A project costs 100 and pays 115 after one year. At a 10% discount rate, NPV is 115/1.10 − 100 ≈ 4.55; IRR is 15%.
Mistake to avoid: Selecting mutually exclusive projects solely by the highest IRR.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
33. Debt costs and tax shields
The contractual interest rate is not always the relevant marginal cost of new borrowing. Under a simplified model where interest is fully deductible and the deduction can be used, after-tax debt cost is r_d(1 − T). Deductibility, tax capacity and timing are assumptions to establish, rather than universal features of financing.
Worked example: With assumed borrowing cost 6% and usable tax rate 25%, the simplified after-tax cost is 4.5%. Annual interest of 60 produces an assumed tax saving of 15.
Mistake to avoid: Applying a tax shield without confirming the assumed deduction is usable.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
34. Weighted average cost of capital
In a standard simplified corporate model, WACC combines market-value-weighted equity cost and after-tax debt cost. It suits cash flows consistent with the assumed operating risk and financing policy. An insurer’s policyholder liabilities and capital needs require careful cash-flow definitions; the formula is not automatically transferable to every insurance valuation.
Worked example: With 70% equity costing 10% and 30% debt costing 5%, plus an assumed usable tax rate of 20%, WACC is 0.7 × 10% + 0.3 × 4% = 8.2%.
Mistake to avoid: Using an average financing rate for a project with materially different risk.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
35. Capital structure under idealized assumptions
In the basic Modigliani–Miller setting without taxes, distress costs or financing frictions, changing debt and equity proportions does not change total firm value. Leverage reallocates operating risk between claims. The result is an analytical benchmark: real financing choices can matter when the excluded costs, benefits or constraints are introduced.
Worked example: If operating assets are worth 150 in the idealized setting and debt is worth 60, equity is worth 90. Replacing equity with additional debt does not itself create asset value.
Mistake to avoid: Treating the benchmark as proof that financing never matters in practice.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
36. Leverage and equity beta
When debt beta is zero and taxes are absent, equity beta equals asset beta multiplied by 1 + D/E. Equity becomes more sensitive because fixed debt claims leave shareholders a leveraged residual. Different debt risk or tax assumptions require a different relationship, and inputs should use consistent economic values.
Worked example: With asset beta 0.75 and debt-to-equity ratio 0.60, the simplified equity beta is 0.75 × 1.60 = 1.20.
Mistake to avoid: Comparing equity betas across firms without accounting for different leverage.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
37. Adjusted present value
Adjusted present value separates an investment’s unlevered value from financing effects such as usable tax benefits or financing costs. This can clarify changing debt schedules that a constant WACC obscures. Value each component using assumptions appropriate to its risk and avoid counting a financing benefit in both cash flows and discount rates.
Worked example: Unlevered NPV of 18 plus financing benefits worth 7 minus financing costs worth 3 gives adjusted present value 22.
Mistake to avoid: Adding tax benefits after already incorporating the same benefits through an adjusted discount rate.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
38. Distributions, liquidity and agency costs
A shareholder distribution reduces resources available within the firm. Its suitability depends on investment opportunities, liquidity and capital needs, not just accounting earnings. Agency conflicts can arise when managers, shareholders and creditors prefer different uses of cash. Evaluate the resulting balance sheet rather than treating a distribution as automatically creating value.
Worked example: An insurer has cash 20, near-term obligations 12 and an internal liquidity requirement of 5. A distribution of 6 leaves 14, below the combined need of 17.
Mistake to avoid: Equating positive earnings with cash safely available for distribution.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
Insurance underwriting profit and return
39. Combined ratios on a consistent basis
On an explicitly consistent earned-premium basis, a combined ratio equals claims and underwriting expenses divided by premium. A ratio below 100% indicates underwriting profit on that basis. Published ratios may use different expense denominators, so reconcile definitions before comparison. Investment returns and financing effects are outside this simplified underwriting measure.
Worked example: Earned premium 120, claims 78 and underwriting expenses 30 give a combined ratio of 108/120 = 90% and underwriting profit 12.
Mistake to avoid: Combining ratios calculated with incompatible premium denominators.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
40. Underwriting margin versus equity return
Underwriting margin compares underwriting profit with premium. Return on equity compares a defined shareholder profit with equity employed. Investment results, financing, taxes and capital intensity explain why the measures differ. Use a suitable average equity denominator and align the profit period before comparing businesses or assessing a return target.
Worked example: Premium 200 and underwriting profit 10 give a 5% underwriting margin. If total shareholder profit is 14 and average equity is 70, return on equity is 20%.
Mistake to avoid: Reading a 5% underwriting margin as a 5% return on shareholder capital.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
41. Insurance float and cash-flow timing
Receiving premium before paying claims creates funds temporarily available for investment. The economic benefit depends on payment timing, investment opportunities and the cost of the insurance obligations. Float is accompanied by liabilities; it is not free shareholder capital. A business can generate substantial float while making an underwriting loss.
Worked example: In a simplified one-year transaction, premium 100 invested at 3% becomes 103 before claims of 102 are paid. The net cash gain is 1, before expenses and capital costs.
Mistake to avoid: Counting the premium receipt as profit without recognizing the associated claim obligation.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
42. Discounted underwriting cash flows
An economic premium assessment compares the premium with the present value of expected outgo and separately identified risk or capital compensation. Discounting changes the contribution of payment timing, not the expected nominal claim amount. Use a valuation basis appropriate to the cash flows; a risky asset’s hoped-for return is not automatically suitable.
Worked example: Expected claims of 110 in two years and immediate expenses of 10 have present value 110/1.05² + 10 ≈ 109.77 under an assumed 5% valuation rate, before additional compensation.
Mistake to avoid: Treating discounted expected outgo alone as a complete risk-bearing premium.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
43. Inflation and claim-payment horizons
Claim costs can grow between the exposure period and payment. Project the appropriate cost driver over the applicable horizon, separating frequency changes from severity inflation where possible. Different claim components may have different drivers. Nominal projected costs should be paired with nominal discount assumptions to avoid inconsistent treatment of inflation.
Worked example: A repair cost currently estimated at 5,000 becomes 5,000 × 1.04³ = 5,624.32 after three years under constant 4% annual severity inflation.
Mistake to avoid: Applying one year of inflation to a cost expected to arise three years later.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
44. Frequency-severity expected losses
For a random claim count N and identically distributed severities independent of N, expected aggregate loss is E[N]E[X]. Exposure measures must match the frequency definition, and severity must match the included claims. Dependence between count and severity, changing coverage or shared events can invalidate a simple product of separately estimated averages.
Worked example: For 800 insured units, expected frequency 0.06 per unit and expected severity 2,500 imply 48 expected claims and expected aggregate loss 120,000.
Mistake to avoid: Multiplying a frequency per policy-year by an exposure measured in policy-months.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
45. Aggregate loss variance
For independent severities and a count independent of those severities, aggregate variance is E[N]Var(X) + Var(N)(E[X])². It includes both severity variation and count variation. With a Poisson count, it simplifies to E[N]E[X²]. These formulas require the stated independence and distributional assumptions.
Worked example: If N is Poisson with mean 20, severity mean is 1,000 and severity variance is 250,000, aggregate variance is 20 × 1,250,000 = 25,000,000.
Mistake to avoid: Including severity uncertainty while omitting uncertainty in the number of claims.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
46. Expected cost and risk loading
A risk loading compensates for uncertainty above expected cost under a specified pricing principle. For example, a standard-deviation loading adds a chosen multiple of loss standard deviation to expected loss. This is an analytical convention, not a universal market rule. Add expenses separately unless the loss variable already includes them.
Worked example: Expected loss 80, standard deviation 20 and an illustrative loading coefficient 0.4 give a loaded loss amount of 88. Adding expenses of 12 produces 100 before other adjustments.
Mistake to avoid: Applying the loading twice by including it in both expected losses and an additional margin.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
47. Reinsurance layer recoveries
For a simplified covered loss L, a layer attaching at a with limit m pays min(max(L − a, 0), m). Retained loss is gross loss minus recovery. The calculation assumes the loss meets contract conditions; exclusions, event definitions, aggregate limits and recovery risk must be considered separately.
Worked example: For a covered loss of 900,000 and a 500,000 layer attaching at 300,000, recovery is 500,000 and retained loss is 400,000.
Mistake to avoid: Treating the attachment point as the recovery limit.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
48. Premium targets and capital support
A simplified premium equation can make a target return explicit: premium covers expected claims, expenses and the required profit on supporting capital. If some expenses are a percentage of premium, solve algebraically rather than adding that percentage to costs. State whether investment income, taxes and timing are excluded or modeled.
Worked example: Expected claims are 70, fixed expenses 10, variable expenses 20% of premium and target profit 5. Ignoring other items, 0.8P = 85, so P = 106.25.
Mistake to avoid: Adding 20% to 85 instead of dividing by the remaining premium proportion.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
49. Capital intensity and business selection
Two products with the same premium margin can require different supporting capital because their uncertainty, duration or dependence differs. Compare profit with the relevant capital commitment as well as premium. Incremental business decisions also require marginal capital and operational constraints; standalone ratios may not capture portfolio interactions.
Worked example: Products A and B each earn expected profit 6 on premium 100. If their consistently defined capital amounts are 30 and 60, their profit-to-capital ratios are 20% and 10%.
Mistake to avoid: Assuming identical underwriting margins imply identical capital efficiency.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
50. Risk-adjusted performance and economic profit
Risk-adjusted return on capital divides a specified profit measure by a specified risk-capital amount. Economic profit subtracts the required capital charge from profit. Align loss allowances, expenses, horizon and capital definition before comparison. A positive ratio alone is insufficient when the required return is higher.
Worked example: Expected profit 9 and capital 60 give a 15% return. With a required capital return of 12%, economic profit is 9 − 0.12 × 60 = 1.8.
Mistake to avoid: Accepting business merely because its expected profit is positive, without assessing its capital charge.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
Enterprise risk and economic capital
51. Risk appetite, limits and ownership
Risk capacity describes the ability to absorb adverse outcomes; appetite expresses the risk an organization chooses to accept. Operational limits translate that choice into measurable boundaries. A limit needs a metric, horizon, accountable owner and escalation response. Capacity alone does not determine whether an exposure is desirable.
Worked example: An insurer can absorb a modeled loss of 50 but chooses a concentration limit of 15. A proposed concentration of 18 exceeds appetite despite remaining below capacity.
Mistake to avoid: Treating available capital as permission to accept every exposure it could absorb.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
52. Economic capital and funded expected loss
Economic capital depends on a defined loss measure, horizon and adverse-outcome criterion. Under an unexpected-loss convention, capital may equal a selected loss quantile minus expected loss when that expected loss is already funded. A total-loss convention produces a different amount. Explain the balance-sheet treatment before interpreting either number.
Worked example: If expected loss is funded at 20 and the selected loss quantile is 90, an unexpected-loss capital convention gives 70. Adding the funding and capital produces 90.
Mistake to avoid: Subtracting expected loss when no separate provision or funding covers it.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
53. Value at Risk
Value at Risk at probability p is the smallest loss level whose cumulative probability reaches p. It identifies a quantile, not the average severity of worse outcomes. Interpretation requires the loss definition and time horizon. Discrete distributions can produce abrupt changes in VaR when probabilities or confidence levels shift.
Worked example: Loss is 0 with probability 90%, 40 with 7% and 200 with 3%. VaR at 95% is 40 because cumulative probability reaches 97% there.
Mistake to avoid: Interpreting VaR of 40 as a maximum possible loss.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
54. Tail Value at Risk for discrete losses
Tail Value at Risk averages the worst 1 − p fraction of losses. For a discrete distribution, use only the probability mass needed at the cutoff to complete that fraction. This quantile-average definition avoids ambiguities caused by conditioning strictly above, or at and above, a quantile containing substantial probability mass.
Worked example: Loss is 0 with probability 96%, 100 with 3% and 300 with 1%. The worst 5% includes these positive losses and 1% at zero; TVaR at 95% is 120.
Mistake to avoid: Averaging only positive losses when the required tail also contains mass at zero.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
55. Subadditivity and diversification
A subadditive risk measure assigns combined risk no greater than the sum of component risks. VaR can fail this condition, especially with discrete outcomes. Diversification conclusions therefore depend on the chosen measure and joint distribution. Tail Value at Risk, under its quantile-average definition for integrable losses, has the subadditivity property.
Worked example: Two independent risks each lose 100 with probability 4%, otherwise zero. Each has 95% VaR of zero, but their sum has 95% VaR of 100 since P(no loss) = 92.16%.
Mistake to avoid: Assuming portfolio VaR must be less than the sum of standalone VaRs.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
56. Correlation-based capital aggregation
A specified square-root aggregation model combines two component capital amounts as √(C₁² + C₂² + 2ρC₁C₂). It resembles standard-deviation aggregation, but its capital interpretation depends on calibration and distribution assumptions. Correlation must represent the relevant dependence; ordinary historical correlation may understate dependence during extreme conditions.
Worked example: Standalone capital amounts 30 and 40 with correlation zero aggregate to 50 under this model, a reduction of 20 from their sum of 70.
Mistake to avoid: Using zero ordinary correlation as evidence that extreme losses cannot occur together.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
57. Euler capital allocation
For a differentiable capital function homogeneous of degree one in exposures, Euler allocation assigns each exposure its size times the marginal derivative of total capital. The allocations sum to total capital. They represent contributions under the selected model, rather than each segment’s standalone requirement, and can reflect diversification or hedging effects.
Worked example: For C(x,y) = √(x² + y²), exposures 6 and 8 give total capital 10. Euler allocations are 6²/10 = 3.6 and 8²/10 = 6.4.
Mistake to avoid: Allocating using marginal derivatives alone without multiplying by exposure sizes.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
58. Stress tests and reverse stress tests
A stress test applies a specified adverse scenario and measures its consequences. A reverse stress test starts from an unacceptable outcome and seeks scenarios that could produce it. Both need coherent relationships across assets, liabilities and operations. A scenario illustrates conditional vulnerability; it does not establish its probability without additional analysis.
Worked example: Assets 200 and liabilities 170 give surplus 30. A coherent scenario reducing assets by 12 and increasing liabilities by 8 leaves surplus 10.
Mistake to avoid: Assigning a probability to a scenario solely because its financial impact was calculated.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
59. Asset-liability matching and liquidity
Asset-liability management considers payment dates, currencies and sensitivity to common economic drivers. Adequate total asset value does not guarantee cash availability when obligations fall due. Liquidity analysis should identify reliable cash, stressed sale proceeds and payment timing. Matching expected payments still leaves exposure to unexpected claims and unavailable inflows.
Worked example: An insurer has assets worth 100 against liabilities of 90, but only 4 cash available before claims of 9 fall due. It must identify at least 5 of timely funding.
Mistake to avoid: Using positive economic surplus as proof that all near-term payments can be made.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
60. Model uncertainty and validation
Process uncertainty concerns random outcomes under a model; parameter uncertainty concerns estimated inputs; model uncertainty concerns the structure itself. Validation should assess data quality, assumptions, predictive behavior and sensitivity to plausible alternatives. Increasing simulation volume reduces numerical error but cannot repair a materially wrong dependence structure or omitted risk.
Worked example: A capital estimate falls from 48 to 42 when dependence assumptions change, while repeated runs differ by only 0.2. Structural sensitivity is much larger than simulation noise.
Mistake to avoid: Treating stable simulation output as evidence that the underlying model is correct.
Reference: Exam 9 — Risk Management for Actuaries | Casualty Actuarial Society
Sources
Source check:
All study guides