Introduction to Actuarial Mathematics

Chapter 7
Premium Principles and Product Pricing

Post Graduate Program in Insurance Management
Introduction to Actuarial Mathematics   •   Post Graduate

Premium Principles and Product Pricing

Introduction

Insurance pricing converts uncertain future cash flows into a present premium that is adequate, fair, competitive, and commercially sustainable. Actuarial pricing must balance several interests: policyholders seek value and affordability; insurers require solvency and profitability; regulators require fairness and transparency; distributors require appropriate remuneration; and shareholders expect a risk-adjusted return.

A premium is therefore more than an estimate of claims. It is the price of a bundle of contingent benefits, expenses, capital requirements, options, guarantees, and services. This chapter develops the principal approaches to premium calculation and shows how net premiums become gross premiums through the addition of expenses, commissions, risk margins, profit provisions, and allowance for behavioural effects such as lapses.

Learning objectives

By the end of this chapter, the reader should be able to:

  1. Explain the equivalence, expected-value, standard-deviation, variance, and percentile premium principles.
  2. Calculate single and periodic net premiums using present values and expected values.
  3. Construct gross premiums from benefits, expenses, commissions, profit margins, and capital costs.
  4. Incorporate lapses, guarantees, options, and adverse selection into product pricing.
  5. Perform sensitivity testing and interpret pricing results for management decisions.
  6. Identify governance, fairness, regulatory, and model-risk issues in actuarial pricing.

1. The pricing problem

Let a policy generate random future benefits and expenses. Denote the present value of benefits by ZB and the present value of expenses by ZE. If the premium is paid at inception, the insurer’s present-value outgo is

Z = ZB + ZE

The random variable Z is measured in currency units at the valuation date. Pricing seeks a premium P such that the insurer receives adequate compensation for expected outgo, uncertainty, capital usage, and the required return. In a simple expected-value framework, P is related to E[Z], but a prudent commercial price generally also reflects variability and non-hedgeable risk.

Important distinctions are:

  • Net premium: the amount required for expected insured benefits, excluding most expenses and profit allowances.
  • Gross or office premium: the amount charged to the customer, including expenses, commissions, taxes where applicable, risk margins, and profit provisions.
  • Single premium: one payment at the beginning of the contract.
  • Level periodic premium: equal payments made at regular intervals.
  • Premium rate: premium per unit of sum insured, exposure, or benefit.
  • Present value: a future amount discounted to the pricing date.

Pricing is conducted on a specified basis: mortality or morbidity, interest, expenses, lapses, claim incidence, inflation, taxes, and contract terms. A premium is meaningful only together with its basis.

2. Present values and expected loss

If a cash flow of amount Ct is payable at time t, and the annual effective discount rate is i, its present value is

PV(Ct) = Ct(1+i)-t

For several deterministic cash flows,

PV = ∑t=0n Ct(1+i)-t

The formula assumes that cash-flow timing and the effective annual interest rate are known and that t is measured in years. For a random cash flow Ct, the expected present value is

E[PV] = E[∑t=0n Ct(1+i)-t]
= ∑t=0n E[Ct](1+i)-t

when the discount rate is deterministic and the expectation is finite. If interest rates are stochastic, the discount factor and cash flow may be jointly random, so the expected product must be modelled rather than separated automatically.

For an annual life-insurance benefit of amount S payable at the end of the year of death, with death probability qx in year one, the expected present value is

E[ZB] = S qx v

where v = (1+i)-1. For a multi-year policy, probabilities must be conditional on survival to the relevant duration. If {}kpx is the probability that a life aged x survives k years and qx+k is the following-year death probability, then a benefit payable at the end of year k+1 has expected present value

E[ZB] = S ∑k=0n-1 {}kpx qx+k vk+1

This expression assumes benefit payment only on death during the term, a fixed benefit, deterministic interest, and no benefit reductions or selection effects beyond those represented in the mortality probabilities.

3. Premium principles

Different premium principles reflect different attitudes toward uncertainty. No principle is universally correct; the choice depends on the risk, available capital, market constraints, and governance requirements.

3.1 Equivalence principle

The equivalence principle sets premium equal to expected present-value outgo:

PEP = E[Z]

For a pure benefit, this is the net single premium. It produces zero expected underwriting profit before expenses and risk compensation. It is attractive for its transparency but does not compensate for parameter uncertainty, random fluctuation, capital costs, or model error.

For a level annual premium payable at the beginning of each year while the policy remains in force, let Y be the present value of one unit of premium. Then the net annual premium is

P = E[ZB] / E[Y]

For a policy with payments at times 0,1,...,n−1, conditional on survival and in-force status, a common form is

E[Y] = ∑k=0n-1 Pr(policy in force at time k)vk

The denominator is an annuity value. It must use the same interest, survival, lapse, and payment-timing assumptions as the numerator.

3.2 Expected-value principle

The expected-value principle adds a loading proportional to expected loss:

PEV = (1+θ)E[Z], θ ≥ 0

Here θ is the loading factor. The loading may represent expenses, uncertainty, capital, profit, or a combination, but these components should normally be separated for transparency. This principle is simple and useful when the coefficient of variation is reasonably stable across risks.

3.3 Variance and standard-deviation principles

The variance principle is

PVar = E[Z] + α Var(Z), α ≥ 0

The parameter α has units inverse to currency because Var(Z) is measured in currency squared. Thus, it requires careful calibration and is sensitive to the monetary scale of the risk.

The standard-deviation principle is

PSD = E[Z] + β√Var(Z), β ≥ 0

Here β is dimensionless because the standard deviation has currency units. This principle increases the premium with dispersion and is often more stable under changes in monetary scale than the variance principle.

3.4 Percentile and risk-measure principles

A percentile principle may charge a high quantile of loss:

Pq = VaRq(Z)

where VaRq is the smallest amount p such that Pr(Z ≤ p) ≥ q. For example, q=0.995 targets the 99.5th percentile. Value at Risk does not describe losses beyond the percentile. An alternative is expected shortfall:

ESq(Z) = E[Z | Z > VaRq(Z)]

subject to the chosen definition and continuity assumptions. Risk-measure pricing is particularly relevant for catastrophe, solvency, and economic-capital frameworks, but it requires a credible loss distribution and robust tail modelling.

3.5 Comparison of principles

PrincipleFormulaMain strengthMain limitation
EquivalenceE[Z]Simple and transparentIgnores volatility and capital
Expected value(1+θ)E[Z]Easy to implementLoading may be arbitrary
VarianceE[Z]+αVar(Z)Rewards diversification assumptionsScale-sensitive; units issue
Standard deviationE[Z]+βSD(Z)Reflects volatilityDoes not fully describe tail risk
PercentileVaRq(Z)Explicit confidence targetIgnores severity beyond quantile
Expected shortfallESq(Z)Captures tail severityData- and model-intensive

4. Net premiums: worked example

Illustrative example. A one-year term policy pays 100,000 currency units at the end of the year if death occurs. The mortality probability is 0.002, and the annual effective interest rate is 4%. There are no expenses, lapses, or other benefits.

The expected benefit is 200 currency units. The discount factor is

v = 1/1.04 = 0.961538

Therefore,

PEP = 100,000 × 0.002 × 0.961538
= 192.31

The net single premium is 192.31 currency units. If instead the insurer applies a 20% expected-value loading,

PEV = 1.20 × 192.31 = 230.77

The second amount is not a net premium in the strict benefit-only sense; it is a risk-loaded premium before explicit expense construction.

For a two-year policy with death probabilities 0.002 in year one and 0.0025 in year two conditional on survival, the expected benefit is

E[ZB] = 100,000(0.002)v + 100,000(0.998)(0.0025)v²
= 192.31 + 230.93
= 423.24 approximately

The survival factor in year two is essential. Omitting it would slightly overstate the expected claim cost.

5. From net premium to gross premium

A gross premium should be constructed from identifiable cash-flow components rather than by applying an unexplained percentage. Let:

  • P = gross single premium;
  • B = expected present value of benefits;
  • Ef = fixed acquisition and issue expenses;
  • eP = premium-proportional expenses;
  • cP = premium-based commission;
  • Et = maintenance expenses;
  • R = risk or capital margin;
  • Π = target profit provision.

Then the prospective pricing equation is

P = B + Ef + eP + cP + Et + R + Π

Solving for the premium gives

P = (B + Ef + Et + R + Π)/(1 − e − c)

This formula assumes that all listed amounts are present values, and that premium-proportional expenses and commissions are linear in P. The denominator must be positive. If e+c ≥ 1, the structure is economically infeasible because variable deductions consume the entire premium.

For periodic premiums, the correct form is usually an equation of present values:

P × E[Y] = E[ZB] + E[ZE] + E[ZR] + E[Z_Π]

Thus,

P = (E[ZB] + E[ZE] + E[ZR] + E[Z_Π]) / E[Y]

The numerator and denominator must use consistent timing, discounting, mortality, lapse, and expense assumptions.

Expense categories

  1. Acquisition expenses: underwriting, medical evidence, commission, policy issue, and distribution costs.
  2. Maintenance expenses: billing, customer service, administration, and periodic statements.
  3. Claims expenses: investigation, settlement, legal review, and fraud management.
  4. Premium-related expenses: percentages of premium, payment processing, and taxes.
  5. Inflation-sensitive expenses: costs that rise with wages or consumer prices.
  6. Capital costs: remuneration for capital held against adverse outcomes.

Expense assumptions should be based on credible unit costs, credible allocation methods, and expected productivity changes. Arbitrary allocation can distort product profitability.

Worked gross-premium example

Suppose the present value of benefits is 192.31, fixed expenses are 25, maintenance expenses are 10, and the required risk/profit provision is 20. Commission is 10% of premium and other premium-proportional expense is 5%.

P = (192.31 + 25 + 10 + 20)/(1 − 0.10 − 0.05)
= 247.31/0.85
= 290.95 approximately

The gross premium is approximately 290.95. Commission is 29.10 and premium-proportional expense is 14.55. Their sum, together with the fixed components, reconciles to the premium.

6. Lapses, selection, and policyholder behaviour

Lapse assumptions affect both sides of pricing. Lapses may reduce future claims and maintenance expenses, but they also reduce future premiums, cause acquisition expenses to be recovered over fewer payments, and may create anti-selection. A lapse model should distinguish:

  • voluntary lapse and involuntary termination;
  • duration, age, premium size, payment mode, and distribution channel;
  • shock lapses after premium increases;
  • partial withdrawals or reductions;
  • reinstatement and persistency after repricing.

If {}kpx represents survival and {}klx represents survival in force after lapses, then an expected premium annuity may be written as

E[Y] = ∑k=0n-1 {}kpx {}klx vk

The benefit numerator must use the corresponding in-force probabilities. Lapse rates are not merely a denominator adjustment; they can change the composition of remaining policyholders and therefore claim rates.

A key pricing issue is recapture of acquisition costs. If a policy lapses early, the insurer may not recover initial underwriting and commission costs. Products with high first-year expenses are particularly sensitive to early lapse.

7. Guarantees, options, and embedded value

Guarantees and options have economic value even when they do not generate an immediate expected claim. Examples include guaranteed renewal, minimum crediting rates, conversion privileges, premium holidays, surrender values, and guaranteed annuity rates.

Their pricing requires scenario-based or stochastic analysis when outcomes depend on future interest rates, equity returns, inflation, or policyholder decisions. A simplified decomposition is

Gross price = expected benefits + expenses + option cost + capital/risk margin + profit provision

The option cost should be estimated consistently with the contract’s exercise behaviour. Assuming that all policyholders exercise an option, or that none do, may materially misprice a product. Dynamic lapse models can link exercise rates to economic incentives, such as surrender when market guarantees become valuable.

Guarantees also create basis risk: assets used to support the guarantee may not move perfectly with liabilities. Asset-liability management, hedging cost, liquidity, and reinvestment risk should therefore be considered alongside actuarial claim assumptions.

8. Sensitivity testing and scenario analysis

Pricing should not rely on a single best estimate. Sensitivity testing changes one assumption at a time to identify material drivers, whereas scenario analysis changes several related assumptions to represent a coherent future environment.

For a premium P depending on assumption a, a local sensitivity can be measured by

Sensitivitya ≈ [P(a+∆a) − P(a)]/∆a

The result has units of premium per unit change in a. A relative sensitivity is

Elasticitya ≈ [∆P/P]/[∆a/a]

Useful tests include:

  • mortality or morbidity increases;
  • lower interest rates;
  • higher inflation and expense growth;
  • higher early-duration lapses;
  • lower persistency after repricing;
  • adverse claims severity;
  • commission changes;
  • delayed investment income;
  • guarantee take-up and option exercise;
  • catastrophe or pandemic scenarios where relevant.

A pricing committee should examine not only the revised premium but also profitability, break-even duration, capital strain, solvency impact, and customer affordability. A product can be profitable on an expected basis but unacceptable under a plausible stress scenario.

9. Product pricing process and governance

A disciplined pricing process normally follows these stages:

  1. Define benefits, exclusions, guarantees, policy terms, and target market.
  2. Establish exposure, claims, lapse, expense, inflation, and investment assumptions.
  3. Select a premium principle and risk-margin methodology.
  4. Project cash flows by duration and calculate present values.
  5. Determine gross premiums and test affordability and competitiveness.
  6. Perform sensitivity, scenario, and profitability analysis.
  7. Review capital, reinsurance, tax, accounting, and regulatory effects.
  8. Obtain independent actuarial and governance review.
  9. Document assumptions, limitations, controls, and approval decisions.
  10. Monitor actual-to-expected experience and revise assumptions when justified.

Governance should include version control, data validation, peer review, model-change procedures, and clear ownership of assumptions. Pricing models should be independently challenged, particularly where they use complex lapse, catastrophe, machine-learning, or option-exercise components.

10. Case study: illustrative term assurance pricing

Background. An insurer is designing a one-year renewable term assurance product with a benefit of 100,000 currency units. The expected first-year death probability is 0.002. The discount rate is 4%. Fixed acquisition expenses are 25 currency units per policy; maintenance expenses are 10; the target risk and profit provision is 20. Commission is 10% of premium and other premium-proportional expense is 5%.

Problem. Management wants to know the indicated premium and whether the result is sensitive to mortality and lapses.

Analysis. The benefit present value is 192.31, as calculated earlier. The indicated premium is

P = (192.31 + 25 + 10 + 20)/0.85 = 290.95

Suppose a stress increases mortality by 25%. The benefit component becomes

192.31 × 1.25 = 240.39

The stressed premium is

Pstress = (240.39 + 25 + 10 + 20)/0.85 = 347.52 approximately

The increase is 56.57, or approximately 19.4% of the base premium. Although mortality increased by 25%, the total premium increased by less because fixed expenses and the risk/profit provision did not change proportionally.

A lapse stress should be modelled through projected in-force policies rather than simply adding a percentage to claims. Higher lapses may reduce future premium income, reduce later claims, increase the proportion of unrecovered acquisition costs, and alter selection. For a renewable product, management should therefore examine profitability by duration and renewal cohort.

Approach and solution. The insurer should validate mortality against the target market, test distribution-channel selection, project renewal behaviour, and assess whether the 5% proportional expense remains realistic at the proposed price. It should also compare the indicated premium with competitor prices without allowing market comparison to override adequacy requirements.

Outcome. The base indicated premium is 290.95, while the mortality-stress indication is approximately 347.52. The product should not be approved solely on the base result; underwriting controls, renewal pricing rules, reinsurance, and monitoring triggers should be specified.

Learning points. Fixed and variable expense structures produce different sensitivities. Behavioural assumptions affect both numerator and denominator. A transparent pricing model supports management decisions better than an unexplained aggregate loading.

11. Limitations and contemporary developments

Premium models are limited by sparse data, changing medical practice, inflation, legal change, climate risk, and behavioural uncertainty. Historical averages may not represent a new distribution channel or underwriting process. Small portfolios may require external data, credibility methods, or Bayesian updating. Data must be relevant, representative, legally obtained, and sufficiently granular to avoid unfair discrimination.

Technology can improve pricing through automated data pipelines, experience monitoring, and scenario engines. Machine-learning models may detect nonlinear relationships, but predictive accuracy does not guarantee actuarial fairness, stability, explainability, or regulatory acceptability. Sensitive variables and proxy discrimination require governance. Human oversight remains necessary for assumptions, overrides, validation, and customer impact.

Contemporary pricing practice increasingly integrates actuarial pricing with enterprise risk management, economic capital, climate scenarios, real-time experience monitoring, and fair-value assessment. The central professional principle remains unchanged: a price should be adequate for the risks undertaken, reasonable in relation to benefits and services, and supported by evidence and governance.

Chapter Summary

Premium pricing begins with the present value of uncertain benefits and expenses. The equivalence principle charges expected present-value outgo; expected-value, variance, standard-deviation, percentile, and expected-shortfall principles add compensation for uncertainty in different ways. Net premiums cover expected benefits, while gross premiums also include acquisition and maintenance expenses, commissions, risk or capital margins, and profit provisions. Premium-proportional costs must be solved algebraically rather than simply added to the numerator.

Lapses influence premiums, claims, expense recovery, and selection, so they must be modelled consistently. Guarantees and options may require stochastic valuation and behavioural assumptions. Sensitivity and scenario testing reveal the drivers of profitability and resilience. Effective governance requires validated data, documented assumptions, independent review, monitoring, and attention to fairness, consumer value, model risk, and regulation.

Key Terms

Premium principle: Rule used to convert a random loss into a premium.

Equivalence principle: Premium equal to expected present-value outgo.

Net premium: Premium for expected insured benefits, generally excluding expenses and profit.

Gross premium: Customer premium including benefits, expenses, risk margins, and profit provisions.

Single premium: One payment made for the contract.

Periodic premium: Premium paid at regular intervals.

Expected value: Probability-weighted average of a random variable.

Present value: Value of a future cash flow discounted to the pricing date.

Expected-value loading: Proportional addition to expected loss.

Variance principle: Premium based on expected loss plus a multiple of variance.

Standard-deviation principle: Premium based on expected loss plus a multiple of standard deviation.

Value at Risk: A selected quantile of a loss distribution.

Expected shortfall: Average loss beyond a selected quantile.

Lapse: Termination of a policy before scheduled expiry, other than by insured event.

Persistency: Proportion of policies remaining in force over time.

Acquisition expense: Cost of obtaining and issuing new business.

Maintenance expense: Ongoing cost of administering policies.

Risk margin: Amount for uncertainty, capital usage, or adverse deviation.

Embedded option: Contractual right whose value depends on future economic or policyholder conditions.

Sensitivity analysis: Assessment of output changes when one assumption changes.

Review Questions

Short Answer Questions

  1. Distinguish a net premium from a gross premium.
  2. State the equivalence principle and explain its main limitation.
  3. Why must lapse assumptions affect both premium and benefit projections?
  4. What is the difference between Value at Risk and expected shortfall?
  5. Identify three categories of insurance expense.

Long Answer Questions

  1. Explain and compare the equivalence, expected-value, variance, and standard-deviation premium principles.
  2. Derive the gross-premium formula when fixed costs, premium-proportional expenses, and commission are present.
  3. Discuss how guarantees and embedded options affect insurance product pricing.
  4. Explain the role of sensitivity testing, scenario analysis, and profitability monitoring in pricing governance.
  5. Critically evaluate the use of machine-learning models in insurance pricing.

Analytical / Application Questions

  1. A benefit has an expected present value of 500 currency units. Fixed expenses are 60, maintenance expenses are 20, risk and profit provisions total 40, commission is 12% of premium, and other premium-proportional expense is 8%. Calculate the gross premium.
  2. Explain how a 30% increase in early lapses could affect a product with high first-year commission and a benefit payable at maturity. Identify at least four channels of impact.
  3. Design a sensitivity-testing programme for a renewable term assurance product. Specify assumptions, stress levels, outputs, and management actions.

Multiple Choice Questions with Answers

Question 1

Under the equivalence principle, the premium is generally set equal to:

A. The maximum possible claim

B. The expected present value of outgo

C. The standard deviation of claims only

D. The policyholder’s income

Correct Answer: B. The expected present value of outgo

Question 2

In the formula P = (B + Ef + Et + R + Π)/(1 − e − c), what do e and c represent?

A. Interest and inflation rates

B. Mortality and lapse rates

C. Premium-proportional expense and commission rates

D. Expected and actual claims

Correct Answer: C. Premium-proportional expense and commission rates

Question 3

A one-year benefit of 100,000 has death probability 0.002 and discount factor 0.961538. What is its expected present value?

A. 192.31

B. 200.00

C. 2,000.00

D. 96.15

Correct Answer: A. 192.31

Question 4

Which premium principle adds a multiple of the standard deviation to expected loss?

A. Equivalence principle

B. Variance principle

C. Standard-deviation principle

D. Percentile principle

Correct Answer: C. Standard-deviation principle

Question 5

Why is the survival probability included in a multi-year death-benefit calculation?

A. To increase the discount rate

B. To represent survival to the later year before death can occur then

C. To eliminate expenses

D. To convert a gross premium into a net premium

Correct Answer: B. To represent survival to the later year before death can occur then

Question 6

If premium-proportional expenses are 10% and commission is 15%, the denominator in the gross-premium formula is:

A. 0.25

B. 0.75

C. 0.85

D. 1.25

Correct Answer: C. 0.85

Question 7

Which statement best describes expected shortfall?

A. It is always equal to the mean loss

B. It measures the average loss beyond a selected quantile

C. It ignores tail losses

D. It is the same as a lapse rate

Correct Answer: B. It measures the average loss beyond a selected quantile

Question 8

A lapse assumption in a periodic-premium product can affect:

A. Only mortality

B. Only investment income

C. Premium income, claims, expense recovery, and selection

D. None of the projected cash flows

Correct Answer: C. Premium income, claims, expense recovery, and selection

Question 9

A sensitivity test normally changes:

A. One assumption or driver to assess its effect

B. The legal contract without analysis

C. Only the accounting date

D. The answer choices in a model

Correct Answer: A. One assumption or driver to assess its effect

Question 10

Using the chapter’s illustrative case, the approximate premium under a 25% mortality increase is:

A. 240.39

B. 290.95

C. 347.52

D. 425.00

Correct Answer: C. 347.52

Introduction to Actuarial Mathematics
Chapter 7 of 14