Investment Management and Asset–Liability Management

Introduction and context

An insurer is both a risk-transfer institution and a financial intermediary. It receives premiums today, invests them, and uses the resulting resources to meet claims, benefits, expenses, and other contractual obligations in the future. Investment management therefore cannot be assessed solely by return. The relevant question is whether the portfolio produces adequate, risk-adjusted returns while remaining capable of meeting liabilities when they fall due.

Asset–liability management (ALM) is the coordinated management of assets and liabilities so that an insurer remains solvent, liquid, profitable, and able to honour policyholder promises under changing economic conditions. It links underwriting, product design, reserving, reinsurance, investment, capital management, and financial reporting.

The central principle is liability-driven investing (LDI): the characteristics of liabilities should determine the risk capacity, time horizon, liquidity profile, and interest-rate exposure of the investment portfolio. A portfolio that earns a high average return but is unable to fund claims during a market stress event is unsuitable for insurance purposes.

Learning objectives

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

1. Explain the purpose and structure of insurer investment management and ALM.
2. Distinguish duration matching, cash-flow matching, immunization, and dynamic hedging.
3. Construct a strategic asset allocation subject to solvency, liquidity, and policyholder constraints.
4. Calculate present value, duration, convexity, and basic funding gaps.
5. Evaluate policyholder participation requirements in life and participating business.
6. Analyse an insurer’s ALM problem and recommend governance and risk controls.

1. The insurer’s balance sheet and the ALM problem

The simplified economic balance sheet is:

where A is the market value of assets, L is the market-consistent value of liabilities, and E is economic own funds or surplus. The identity is measured at a particular valuation date and is not a statement that assets and liabilities have identical risk.

Solvency depends on the level and volatility of surplus. If asset value falls by more than liability value, surplus declines. A useful change relationship is:

where ∆ denotes change over a specified period. This relation is an accounting approximation for economic values; it does not by itself measure regulatory capital.

Insurer liabilities differ by line of business:

ALM must consider at least five dimensions: amount, timing, currency, interest-rate sensitivity, and optionality. Policyholder options such as surrender, withdrawal, conversion, renewal, or loan features can make liability cash flows dependent on economic conditions.

2. Liability-driven investing

LDI begins with a liability benchmark. The benchmark may be a projected schedule of claim payments or a replicating portfolio whose duration, convexity, currency, and liquidity resemble the liabilities. Assets are then divided into two broad functions:

• Hedging or matching assets: designed to fund contractual liability cash flows and reduce market risk.
• Return-seeking assets: designed to improve expected return, diversify risks, or support surplus growth.

LDI is not limited to government bonds. Depending on credit quality, liquidity, regulation, and cash-flow reliability, it may include high-grade corporate bonds, inflation-linked securities, private credit, infrastructure debt, swaps, or other instruments. Derivatives may hedge interest-rate or currency exposure, but they introduce collateral, basis, counterparty, model, and operational risks.

A liability benchmark should be documented in terms of:

• projected nominal and real cash flows;
• expected claims and benefit timing;
• discount-rate methodology;
• duration and convexity;
• currency exposure;
• liquidity needs and stress outflows;
• embedded options and policyholder behaviour; and
• acceptable tracking error.

3. Present value and cash-flow matching

The present value of a set of deterministic cash flows is:

where CF_t is the cash flow at time t, y_t is the applicable annual effective discount rate for that maturity, and n is the final payment period. If a single flat annual effective rate y is assumed, the expression becomes PV = ∑ CF_t ÷ (1+y)^t. Cash flows and present value must use the same currency and time unit.

Cash-flow matching attempts to select assets so that asset cash flows cover liability cash flows in each relevant period:

where A_t and L_t are asset and liability cash flows in period t. In practice, exact matching may be impossible because liabilities are uncertain, assets are indivisible, and transaction costs exist. A surplus or liquidity buffer is therefore normally maintained.

Worked example: basic cash-flow matching

Suppose an insurer expects liability payments of 40, 60, and 100 monetary units at the ends of years 1, 2, and 3. It purchases bonds with cash flows of 45, 65, and 110 in those years. Before expenses, defaults, and reinvestment effects, the annual cash-flow surplus is 5, 5, and 10. The cumulative surplus is 20. This is a cash-flow match only under the assumptions that payments occur on schedule, bond cash flows are received in full, and no earlier liquidity requirement exists.

If the bond portfolio instead pays 30, 80, and 110, it has a shortfall of 10 in year 1 and a surplus later. It may still be economically acceptable if liquid assets cover the first-year shortfall, but it is not a strict cash-flow match.

4. Duration, convexity, and immunization

Duration measures interest-rate sensitivity. Macaulay duration is the cash-flow-weighted average time to receipt:

Here D_M is Macaulay duration in years, t is the payment time in years, CF_t is the cash flow, y is a constant annual effective yield, and PV is present value. This formula assumes a single yield and discrete annual cash flows.

For a bond or liability valued using an annual effective yield, modified duration is:

It approximates the proportional price change for a small parallel yield change:

The yield change must be expressed as a decimal. For example, 50 basis points equals 0.005. The approximation is most reliable for small changes and parallel shifts.

Convexity improves the approximation for larger movements:

where C is convexity, measured in years squared under the chosen convention. Convexity is not a complete description of risk when the yield curve twists, spreads move, or liabilities contain options.

Immunization

Classical immunization seeks to protect surplus against small interest-rate changes by matching the present value and duration of assets and liabilities:

where subscripts A and L denote assets and liabilities. Under a parallel yield shift, similar cash-flow conventions, and sufficient convexity, the first-order effect of the shift on surplus is reduced. A stronger practical condition is that asset convexity should be at least as large as liability convexity when the present values and durations are matched.

Immunization is not a guarantee. It can fail because of non-parallel yield movements, credit spread changes, defaults, prepayments, policyholder options, liquidity needs, and changes in liability assumptions. It also requires rebalancing because duration changes as time passes and as interest rates move.

Duration example

Consider a liability with present value 1,000 and modified duration 6 years. If rates rise by 1 percentage point, the first-order estimated change in value is:

∆PV ÷ PV ≈ −6 × 0.01 = −0.06

Thus the liability value is estimated to fall by 6%, or 60, to approximately 940, ignoring convexity and other effects. An asset portfolio with the same duration would experience a similar first-order percentage change, reducing the interest-rate effect on surplus.

5. Strategic asset allocation

Strategic asset allocation (SAA) establishes long-term exposures consistent with liabilities, risk appetite, capital, and return objectives. It differs from tactical asset allocation, which makes shorter-term deviations from strategic weights.

A typical SAA process is:

1. Project liability cash flows under base and stress assumptions.
2. Define solvency, liquidity, currency, duration, concentration, and participation constraints.
3. Estimate asset returns, volatilities, correlations, spreads, default losses, and liquidity characteristics.
4. Construct candidate portfolios and assess them using scenario analysis and stochastic models.
5. Select strategic ranges rather than relying exclusively on point estimates.
6. Approve the policy through governance committees and board oversight.
7. Monitor performance, risk, and deviations; rebalance according to documented rules.

Mean–variance optimisation can be expressed as:

subject to ∑w_i = 1 and other constraints. Here w is the vector of portfolio weights, μ is the vector of expected returns, Σ is the covariance matrix, and λ represents risk aversion. The objective is dimensionally meaningful when returns and variances are measured over the same period and the risk-aversion parameter is appropriately scaled.

For insurers, unconstrained mean–variance optimisation is inadequate. It may recommend illiquid or highly volatile assets that perform poorly under regulatory or liability stress. An insurer should augment optimisation with:

• liability-relative risk measures;
• solvency capital consumption;
• liquidity-at-risk and cash-flow-at-risk;
• stress and reverse-stress testing;
• credit migration and default analysis;
• ESG and stewardship considerations;
• concentration and connected-counterparty limits; and
• model uncertainty and parameter ranges.

6. Liquidity management

Solvency and liquidity are related but distinct. An insurer may have assets exceeding liabilities but still be unable to make immediate payments without selling assets at distressed prices.

Liquidity planning should classify cash outflows into:

• expected claims and benefits;
• operating expenses and commissions;
• collateral and margin calls;
• policyholder surrender or withdrawal payments;
• reinsurance recoverables timing;
• tax and debt-service obligations; and
• extraordinary catastrophe or legal payments.

A liquidity ladder compares expected inflows and outflows by day, week, month, and year. Liquid assets should be assessed after haircuts, not at an assumed full market value. Stress tests should include claim surges, delayed reinsurance recoveries, market closures, spread widening, surrender spikes, and collateral calls.

Liquidity buffers may include cash, central-bank eligible securities, short-term government instruments, committed credit facilities, and pre-arranged collateral capacity. Reliance on selling illiquid assets during stress should be explicitly quantified.

7. Policyholder participation and surplus distribution

Participating life policies may provide guaranteed benefits plus bonuses or dividends influenced by investment performance, mortality, expenses, and management policy. The insurer must balance:

• fairness between generations of policyholders;
• guarantees and solvency protection;
• smoothing of bonuses over time;
• shareholder expectations;
• reserving and capital requirements; and
• transparency of credited returns.

A smoothing account can reduce year-to-year volatility, but it cannot permanently distribute more than the underlying economic resources support. Excessive smoothing may transfer value unfairly between cohorts or conceal emerging losses. Asset allocation should therefore reflect the liability’s participation mechanism, not merely the average historical return.

For unit-linked products, policyholders generally bear investment risk, but the insurer may retain guarantees, minimum death benefits, operational obligations, and liquidity responsibilities. Product terms must be analysed before assigning risk to policyholders.

8. Governance, regulation, and ethical responsibilities

An effective ALM framework normally includes a board-approved investment and ALM policy, clear delegated authorities, independent risk oversight, documented limits, valuation controls, and escalation procedures. Responsibilities should be separated among portfolio management, actuarial valuation, risk management, finance, compliance, and internal audit.

Regulatory regimes commonly require attention to admissible assets, diversification, prudent-person principles, capital charges, liquidity, related-party exposures, derivatives, and stress testing. Exact rules vary by jurisdiction; therefore, this chapter provides principles rather than legal advice.

Ethically, managers must not select assets solely to maximise reported yield while obscuring liquidity or default risk. Fair treatment requires that investment decisions, bonus declarations, valuation assumptions, and policyholder communications are consistent and supportable. Conflicts involving affiliated investments, remuneration, external managers, and sustainability claims must be controlled.

9. Technology, data, and model risk

Modern ALM uses policy administration data, actuarial models, market data, scenario generators, optimisation tools, and dashboards. Artificial intelligence may assist with cash-flow forecasting, anomaly detection, document extraction, and early-warning indicators. However, AI outputs require validation because historical claims and surrender data may contain bias, regime changes, missing values, or inappropriate proxies.

Model governance should address data lineage, validation, explainability, version control, override authority, cyber risk, and outcome monitoring. A complex model is not automatically superior: transparency and robustness are particularly important when models influence investment limits or policyholder outcomes.

10. Illustrative case: Meridian Life Insurance Company

Background. Meridian Life, a fictional insurer, sells participating endowment policies and immediate annuities. Its liabilities have an economic value of 500 million monetary units and an effective duration of 9.2 years. Its assets are worth 525 million, but the asset duration is 5.8 years. Forty per cent of assets are in equities and private credit, while near-term benefit payments are concentrated in government and high-grade bond cash flows.

Problem. A parallel rise in interest rates would reduce asset values less than liability values because liabilities have longer duration. At the same time, a recession could reduce equity values, increase surrenders, and lower the value of private credit. The insurer also has a policyholder participation policy that requires a competitive credited return.

Analysis. The initial surplus is 25 million. The duration gap is 3.4 years, calculated as 9.2 − 5.8. Using a simple duration approximation, a 1 percentage-point increase in rates produces an estimated asset change of −5.8% × 525 = −30.45 million and a liability change of −9.2% × 500 = −46.00 million. The estimated surplus change is therefore approximately 15.55 million favourable, because liabilities fall more than assets. Conversely, a 1 percentage-point fall in rates would produce an estimated adverse surplus change of approximately 15.55 million. This calculation ignores convexity, spread movements, options, tax, and policyholder behaviour.

The analysis also identifies a liquidity weakness: private credit cannot necessarily be sold quickly, while surrender payments may rise when policyholders need cash.

Approach. Meridian adopts a liability-relative SAA. It increases the matching portfolio of high-quality bonds, uses a controlled interest-rate hedge to reduce the duration mismatch, establishes a twelve-month stressed liquidity buffer, and limits private credit concentrations. It separates the participating fund’s assets from shareholder assets for performance and fairness analysis. Bonus declarations are reviewed against long-term distributable surplus rather than one-year investment returns.

Outcome. The revised strategy reduces interest-rate sensitivity and improves liquidity resilience, although expected long-term return falls modestly. Management accepts the lower expected return because the probability and cost of forced asset sales and unfair bonus support have declined.

Learning points. First, an asset portfolio must be evaluated against liabilities, not in isolation. Second, surplus risk can be material even when an insurer appears strongly capitalised. Third, participation rules make asset allocation a policyholder-governance issue as well as an investment issue. Finally, duration analysis is useful but incomplete without liquidity, spread, option, and behavioural analysis.

11. Limitations and best practices

ALM models depend on assumptions about mortality, morbidity, claims inflation, surrender behaviour, interest rates, defaults, expenses, and future management actions. Long-term forecasts are especially sensitive to small changes in assumptions. A robust framework therefore uses ranges, stochastic scenarios, reverse stress tests, and model overlays rather than a single deterministic forecast.

Best practices include:

• maintain a documented liability cash-flow projection;
• define risk appetite in liability-relative terms;
• match major currencies and material cash-flow dates;
• monitor duration, convexity, key-rate duration, spread duration, and basis risk;
• maintain a realistic liquidity buffer;
• test collateral and derivative arrangements under stress;
• incorporate policyholder options and management actions;
• assess capital consumption alongside expected return;
• validate models independently; and
• report exceptions promptly to senior management and the board.

Chapter Summary

Investment management for insurers is an exercise in meeting uncertain liabilities while preserving solvency, liquidity, fairness, and sustainable returns. ALM integrates asset and liability cash flows, interest-rate sensitivity, credit risk, market risk, policyholder behaviour, capital, and governance. Liability-driven investing places liability characteristics at the centre of portfolio design. Cash-flow matching seeks asset receipts that cover payments, while duration matching and immunization reduce sensitivity to interest-rate movements under stated assumptions. Convexity, key-rate exposures, basis risk, options, and non-parallel yield changes limit the protection provided by simple duration matching.

Strategic asset allocation must incorporate solvency and liquidity constraints rather than rely only on expected return and volatility. Participating products require careful treatment of bonus smoothing, intergenerational equity, guarantees, and distributable surplus. Technology and AI can improve forecasting and monitoring, but model governance remains essential. Effective ALM is therefore a continuous governance process, not a one-time portfolio decision.

Key Terms

Asset–liability management: Coordinated management of assets and liabilities to control financial risk and meet obligations.

Liability-driven investing: Investment approach in which liability characteristics guide portfolio construction.

Cash-flow matching: Structuring asset receipts to cover liability payments by time period.

Duration: Measure of the sensitivity of present value to interest-rate changes.

Macaulay duration: Present-value-weighted average time to receive cash flows.

Modified duration: Duration measure used to approximate proportional value changes for yield movements.

Convexity: Measure that improves duration-based estimates for larger yield changes.

Immunization: Strategy intended to reduce the effect of interest-rate changes on surplus.

Strategic asset allocation: Long-term portfolio policy consistent with objectives and constraints.

Tactical asset allocation: Short-term deviation from strategic portfolio weights.

Liquidity risk: Risk of being unable to meet obligations when due without excessive loss.

Duration gap: Difference between asset and liability duration.

Basis risk: Risk that the hedge and the exposure do not move identically.

Policyholder participation: Allocation of investment or other surplus benefits to policyholders.

Surrender risk: Risk arising when policyholders terminate or withdraw from contracts.

Solvency capital: Financial resources available to absorb adverse outcomes.

Stress testing: Assessment of financial effects under adverse scenarios.

Reverse stress testing: Analysis of scenarios capable of causing a specified failure or threshold breach.

Prudent-person principle: Requirement that investments be made with appropriate care, expertise, diversification, and regard for beneficiaries.

Model risk: Risk of loss or poor decisions caused by incorrect, misused, or misunderstood models.

Review Questions

Short Answer Questions

1. What is the principal objective of liability-driven investing?
2. Define modified duration and state its approximate price-change relationship.
3. Why are solvency and liquidity different risks?
4. What is the purpose of a stressed liquidity buffer?
5. Name two reasons why classical immunization may fail.

Long Answer Questions

1. Explain how an insurer should develop a liability-relative strategic asset allocation.
2. Compare cash-flow matching, duration matching, and immunization, including their assumptions and limitations.
3. Discuss the effect of policyholder participation on investment governance and bonus decisions.
4. Explain how credit risk, liquidity risk, and interest-rate risk interact in an insurer’s portfolio.
5. Evaluate the role of governance, model validation, and regulatory limits in ALM.

Analytical / Application Questions

1. An insurer has assets worth 800 million with modified duration 4.5 years and liabilities worth 760 million with modified duration 7 years. Estimate the first-order change in surplus following a 1 percentage-point fall in rates. State the limitations of the calculation.
2. Design a liquidity stress test for a health insurer facing a pandemic claim surge, delayed reinsurance recoveries, and falling asset prices. Identify assumptions, metrics, and management actions.
3. A participating life insurer is considering replacing government bonds with illiquid private credit. Analyse the expected benefits, risks, policyholder-fairness implications, and governance approvals required.

Multiple Choice Questions with Answers

Question 1

What is the central principle of liability-driven investing?

A. Maximise equity exposure regardless of liabilities

B. Use liability characteristics to guide investment design

C. Avoid all fixed-income investments

D. Match accounting profit in every quarter

Correct Answer: B. Use liability characteristics to guide investment design

Question 2

Which expression represents the basic economic balance-sheet identity?

A. A = L − E

B. E = A + L

C. A = L + E

D. L = A + E

Correct Answer: C. A = L + E

Question 3

If modified duration is 6 and yields rise by 0.005, the approximate proportional price change is:

A. −0.3%

B. −3.0%

C. +3.0%

D. −30%

Correct Answer: B. −3.0%

Question 4

Cash-flow matching primarily seeks to ensure that:

A. Asset cash flows cover liability cash flows by period

B. All assets have identical maturities

C. Portfolio volatility is zero

D. Equity returns exceed bond returns

Correct Answer: A. Asset cash flows cover liability cash flows by period

Question 5

Classical immunization generally requires matching asset and liability:

A. Book values and equity weights

B. Present values and durations

C. Currencies and commissions only

D. Default rates and tax rates

Correct Answer: B. Present values and durations

Question 6

A 50-basis-point yield change expressed as a decimal is:

A. 0.5

B. 0.05

C. 0.005

D. 0.0005

Correct Answer: C. 0.005

Question 7

Which risk can exist even when an insurer has assets exceeding liabilities?

A. Liquidity risk

B. No risk at all

C. Only mortality risk

D. Only accounting risk

Correct Answer: A. Liquidity risk

Question 8

Why is unconstrained mean–variance optimisation insufficient for insurers?

A. It cannot calculate any expected returns

B. It ignores the existence of liabilities by design

C. It may recommend portfolios inconsistent with liquidity and solvency constraints

D. It always produces only government bonds

Correct Answer: C. It may recommend portfolios inconsistent with liquidity and solvency constraints

Question 9

In the Meridian case, the asset duration was shorter than the liability duration. A fall in rates would, ignoring other effects, tend to:

A. Increase liabilities more than assets and reduce surplus

B. Reduce liabilities more than assets and reduce surplus

C. Have no effect on either balance-sheet item

D. Increase surplus only through equity dividends

Correct Answer: A. Increase liabilities more than assets and reduce surplus

Question 10

Which practice best supports responsible ALM model governance?

A. Using a model without validation if it is complex

B. Removing all management review of model outputs

C. Maintaining data lineage, validation, version control, and monitoring

D. Reporting only favourable stress-test results

Correct Answer: C. Maintaining data lineage, validation, version control, and monitoring