Risk Mitigation for Linked Product Guarantees

Why Mitigation Must Be Evaluated on More Than One Measure

A design change can reduce an insurer’s exposure to an equity-linked guarantee while also reducing policyholder benefits or changing the timing of cash flows. A charge may provide revenue but leave a smaller account value at the guarantee date. An asset-allocation limit may reduce severe downside outcomes while also limiting upside. The effect therefore cannot be assessed from a single figure in isolation.

This tutorial uses simple hypothetical models to compare guarantee shortfalls, policyholder payments, discounted reserves, and explicitly defined stress measures. The calculations are aligned with the SOA Exam ALTAM (Advanced Long-Term Actuarial Mathematics) syllabus. Any stress-capital measure used below is a stated illustrative proxy, not a complete regulatory capital formula.

Definition: Let $G$ be a guaranteed maturity amount and let $F$ be the linked account value immediately before the maturity payment. The guarantee top-up is $T=\max(G-F,0)$, and the total maturity benefit is $B=F+T=\max(F,G)$. For a one-payment model with payment at time $1$, annual effective interest rate $i$, and an in-force policy, the time-0 reserve for the guarantee payment is $V_0=E[T]/(1+i)$. This reserve excludes expenses, taxes, other benefits, and future premiums unless those cash flows are explicitly included.

A Consistent Comparison Framework

For each design, first identify the account value at the payment date and the guarantee that applies. Calculate the top-up in each outcome, then apply the outcome probabilities. Keep conditional quantities distinct: an expected top-up for a policy still in force at maturity is not automatically the expected top-up per policy originally issued. If the probability of remaining in force is $r$, and the top-up is zero when the policy is not in force, expected top-up per policy issued is $rE[T\mid\text{in force}]$.

To compare designs fairly, use the same scenario probabilities, timing, and discount rate unless the problem specifies otherwise. Also report policyholder benefits separately from insurer costs: the guarantee top-up is only one part of the total maturity benefit.

Specify the contract.
State the guarantee, the account value used to determine the benefit, and whether any charge is deducted before the guarantee test.

Calculate by outcome.
For every scenario, calculate $T=\max(G-F,0)$ and $B=\max(G,F)$. Do not average account values before applying the maximum.

Value and stress the results.
Weight top-ups by their probabilities, discount payments to the valuation date, and calculate the specified stress measure using its stated stress scenario.

Guarantee Redesign: Lowering the Floor

Reducing $G$ directly limits the amount that must be paid when the account value is low. It does not necessarily reduce every maturity benefit: if $F$ exceeds both the old and new guarantees, the benefit remains $F$. The policyholder impact is therefore outcome-dependent.

Worked Example: Reducing a Maturity Guarantee

A hypothetical linked contract has a maturity guarantee of $\$1{,}000$. At time 1, the account value is $\$800$ with probability 0.20, $\$950$ with probability 0.30, or $\$1{,}200$ with probability 0.50. A redesign reduces the guarantee to $\$900$, without changing the account-value distribution. The annual effective interest rate is 5%, and the policy is in force at maturity. Calculate the expected top-up, reserve, expected maturity benefit, and the top-up under a separately specified stress account value of $\$700$, under each design.

Method. Apply the maximum separately in each outcome. The reserve is the expected top-up discounted one year. The stress measure here is the top-up under the stated $\$700$ stress account value; it is not presented as a full capital requirement.

Original guarantee. With $G=\$1{,}000$, the top-ups are $\max(1{,}000-800,0)=\$200$, $\max(1{,}000-950,0)=\$50$, and $\max(1{,}000-1{,}200,0)=\$0$. Thus

E[T]=0.20(200)+0.30(50)+0.50(0)=40+15+0=\$55.

The reserve is $55/1.05=\$52.38095$, or $\$52.38$ to the nearest cent. The maturity benefits are $\$1{,}000$, $\$1{,}000$, and $\$1{,}200$, so the expected benefit is $0.20(1{,}000)+0.30(1{,}000)+0.50(1{,}200)=\$1{,}100$.

Redesigned guarantee. With $G=\$900$, the top-ups are $\$100$, $\$0$, and $\$0$. Therefore $E[T]=0.20(100)=\$20$, and the reserve is $20/1.05=\$19.04762$, or $\$19.05$. The maturity benefits are $\$900$, $\$950$, and $\$1{,}200$; the expected benefit is $0.20(900)+0.30(950)+0.50(1{,}200)=180+285+600=\$1{,}065$.

Under the stress account value of $\$700$, the top-up falls from $1{,}000-700=\$300$ to $900-700=\$200$, a $\$100$ reduction. In the specified probability model, expected top-up falls by $\$35$, and the reserve falls by $\$52.38-\$19.05=\$33.33$. The expected maturity benefit falls by $\$35$ as well. These differences follow from the outcomes: in the $\$950$ state the old guarantee pays $\$1{,}000$, but the redesigned benefit is the account value, $\$950$.

Check. Each expected benefit equals the expected account value plus the expected top-up: the account-value expectation is $0.20(800)+0.30(950)+0.50(1{,}200)=\$1{,}045$. Adding $\$55$ gives $\$1{,}100$; adding $\$20$ gives $\$1{,}065$. A common mistake is to assume that reducing the guarantee leaves the benefit unchanged in every state.

Why the Expected Shortfall Formula Works

For a finite set of mutually exclusive account-value outcomes $F_j$, with probabilities $p_j$ summing to 1, the top-up in outcome $j$ is $\max(G-F_j,0)$. The expectation of a discrete random variable is the probability-weighted sum of its values. Therefore the expected guarantee payment is

E[T]=\sum_j p_j\max(G-F_j,0).

If each top-up is paid at time 1 and the effective annual interest rate is $i$, one dollar payable then has time-0 value $1/(1+i)$. Linearity of expectation gives

V_0=E[vT]=\frac{E[T]}{1+i}, \qquad v=\frac{1}{1+i}.

This result assumes that the payment is exactly one year from valuation and that the specified interest rate applies to discount it. For a payment at a different time, the discount factor must reflect that timing. It also values only the stated guarantee cash flow; it is not a complete reserve for a contract with additional benefits or expenses.

Charges: Revenue and a Smaller Guarantee Base

A charge deducted from an account before maturity can reduce the insurer’s net exposure if the charge is retained as revenue. However, the deduction also reduces the fund available to the policyholder and can increase the guarantee top-up. The charge, benefit, and guarantee payment must be placed on the same timeline before comparing them.

Worked Example: Adding a Maturity-Date Charge

A hypothetical one-year contract guarantees $\$1{,}000$. Before a proposed charge, the maturity account value is $\$900$ with probability 0.40 and $\$1{,}100$ with probability 0.60. A new $\$20$ charge is deducted at maturity from the account and retained by the insurer. The policy is in force, the annual effective interest rate is 4%, and the charge is collected in either outcome. Compare expected top-ups, expected benefits, and the time-0 reserve for net guarantee outgo, defined as top-up less charge revenue.

Without the charge. The top-ups are $\$100$ and $\$0$, so expected top-up is $0.40(100)+0.60(0)=\$40$. Benefits are $\$1{,}000$ and $\$1{,}100$, giving expected benefit $0.40(1{,}000)+0.60(1{,}100)=\$1{,}060$. The reserve for the guarantee outgo is $40/1.04=\$38.46154$, or $\$38.46$.

With the charge. The account values used in the guarantee test are $\$900-\$20=\$880$ and $\$1{,}100-\$20=\$1{,}080$. The resulting top-ups are $\$120$ and $\$0$, so expected top-up is $0.40(120)=\$48$. The benefits paid to the policyholder are $\$1{,}000$ and $\$1{,}080$, giving expected benefit $0.40(1{,}000)+0.60(1{,}080)=\$1{,}048$. Expected net outgo at time 1 is $\$48-\$20=\$28$, and its time-0 value is $28/1.04=\$26.92308$, or $\$26.92$.

The charge increases expected guarantee payments by $\$8$, but expected net outgo falls by $\$12$, because the $\$20$ charge is collected in both outcomes. Expected policyholder benefit falls by $\$12$. The net reserve comparison is valid here because both the charge and top-up occur at time 1 and the charge is certain. If the charge were conditional, paid at a different time, or not retained by the insurer, its expected discounted value would need to be calculated separately.

Check. With the charge, the expected benefit plus the expected charge revenue equals $1{,}048+20=\$1{,}068$. The expected account value after charge is $0.40(880)+0.60(1{,}080)=\$1{,}000$, and expected top-up is $\$48$, so benefit is $\$1{,}048$; adding charge revenue gives $\$1{,}068$. The $\$8$ difference from the original expected account-plus-top-up total reflects that the charge is deducted from the account.

Asset-Allocation Limits and Downside Risk

An asset-allocation limit may change the distribution of account values rather than the guarantee formula. Its effect should be measured against a clearly specified baseline using comparable probabilities and timing. In particular, a lower expected top-up does not by itself establish that every risk measure or every policyholder outcome improves.

Worked Example: Limiting Risky-Asset Exposure

A hypothetical linked contract has a $\$1{,}000$ maturity guarantee. Under the current allocation, the maturity account is $\$700$, $\$1{,}000$, or $\$1{,}400$, with probabilities 0.25, 0.25, and 0.50. A proposed allocation limit changes these outcomes to $\$850$, $\$1{,}020$, or $\$1{,}250$, with the same respective probabilities. The annual effective interest rate is 3%, and the policy remains in force. Compare expected top-ups, reserves, expected benefits, and the top-up in a stress scenario using the lowest account value in each design.

Current allocation. Top-ups are $\$300$, $\$0$, and $\$0$, so expected top-up is $0.25(300)=\$75$. The reserve is $75/1.03=\$72.81553$, or $\$72.82$. Expected benefit is $0.25(1{,}000)+0.25(1{,}000)+0.50(1{,}400)=\$1{,}200$. The specified downside stress top-up, at account value $\$700$, is $\$300$.

Allocation limit. Top-ups are $\$150$, $\$0$, and $\$0$, so expected top-up is $0.25(150)=\$37.50$. The reserve is $37.50/1.03=\$36.40777$, or $\$36.41$. Expected benefit is $0.25(1{,}000)+0.25(1{,}020)+0.50(1{,}250)=250+255+625=\$1{,}130$. The downside stress top-up, at account value $\$850$, is $\$150$.

The allocation limit reduces expected top-up by $\$37.50$, reserve by $\$36.41$, and the stated stress top-up by $\$150$. Expected maturity benefit also falls, by $\$70$. The largest account value falls from $\$1{,}400$ to $\$1{,}250$, so the reduction in guarantee exposure is accompanied by less upside in this model.

Check. Under the current allocation, expected account value is $0.25(700)+0.25(1{,}000)+0.50(1{,}400)=\$1{,}125$; adding expected top-up gives $\$1{,}200$. Under the limit, expected account value is $0.25(850)+0.25(1{,}020)+0.50(1{,}250)=\$1{,}092.50$; adding $\$37.50$ gives $\$1{,}130$. A common mistake is to compare only the guaranteed portions and ignore the account values paid when they exceed the guarantee.

Interpreting the Measures and Avoiding Common Traps

The examples use separate metrics because they answer different questions. Expected top-up describes average guarantee claims under the specified scenarios. A discounted reserve expresses the value at the valuation date of those expected claims under the stated timing and interest assumptions. A stress top-up measures the claim in a particular adverse scenario. None should be called “capital” without a specified capital framework. A stress top-up can be reported as an illustrative capital proxy only when that convention is explicitly stated.

A comparison should also identify the policyholder effect. Reducing a guarantee can lower payments in outcomes where the account value lies between the old and new guarantees. A charge reduces the account value even if the guarantee keeps the payment above that account value. An allocation limit can lower downside exposure while reducing upside. These effects are not interchangeable with the insurer’s net cost or reserve.

Do not average first and apply the guarantee second. In general, $E[\max(G-F,0)]$ is not $\max(G-E[F],0)$. Apply the guarantee formula in each outcome before taking the expectation.
Apply charges at the specified time. If a charge is deducted before the guarantee test, use the post-charge account value in $T=\max(G-F,0)$.
Keep benefit and top-up distinct. The benefit is $\max(G,F)$; the top-up is only the amount needed to raise a below-guarantee account to $G$.
Match probabilities to the population being valued. Conditional top-ups for policies in force at maturity require an in-force probability if the requested result is per policy issued.
State what the stress figure means. A single stress scenario is not automatically a full solvency capital calculation.

Key takeaway: Evaluate a proposed guarantee mitigation by recalculating scenario-level benefits and top-ups, then weighting and discounting cash flows consistently. Report policyholder benefits, guarantee reserves, and any specified stress measure separately so the trade-off is visible.

Check Your Understanding

For each question, use the stated assumptions and calculate scenario-level top-ups before averaging.

A one-year contract guarantees $\$1{,}200$. The account value is $\$900$, $\$1{,}150$, or $\$1{,}400$ with probabilities 0.30, 0.40, and 0.30. The annual effective interest rate is 5%, and the policy is in force. Find expected top-up and reserve. (Answer: expected top-up $\$110$; reserve $\$104.76$)
A one-year contract guarantees $\$800$. Before a maturity-date charge, the account is $\$760$ or $\$900$, each with probability 0.50. A $\$15$ charge is deducted at maturity in both outcomes and retained by the insurer. At 4% effective annual interest, find expected top-up after the charge, expected net outgo, and its time-0 value. (Answer: expected top-up $\$27.50$; expected net outgo $\$12.50$; reserve $\$12.02$)
A $\$1{,}000$ guarantee has account outcomes $\$600$, $\$1{,}000$, and $\$1{,}300$ with probabilities 0.20, 0.50, and 0.30. An allocation limit changes them to $\$800$, $\$1{,}000$, and $\$1{,}200$, with probabilities unchanged. Compare expected top-ups, expected benefits, and lowest-outcome stress top-ups. (Answer: expected top-ups $\$80$ and $\$40$, a $\$40$ reduction; expected benefits $\$1{,}090$ and $\$1{,}060$; stress top-ups $\$400$ and $\$200$)
A one-year guarantee is reduced from $\$950$ to $\$900$. The account is $\$850$, $\$925$, or $\$1{,}100$ with probabilities 0.40, 0.40, and 0.20. At 5% effective annual interest, find the expected top-up and reserve under each guarantee, and the expected benefit under each. (Answer: expected top-ups $\$50$ and $\$20$; reserves $\$47.62$ and $\$19.05$; expected benefits $\$980$ and $\$950$)
Of 100 policies issued, 80% are expected to remain in force at the guarantee date. Conditional on being in force, account values are $\$700$, $\$900$, and $\$1{,}200$, with probabilities 0.25, 0.50, and 0.25. The guarantee is $\$1{,}000$, and policies not in force receive no guarantee payment. Find expected top-up conditional on being in force, expected top-up per policy issued, and aggregate expected top-ups for the 100 policies. (Answer: conditional expected top-up $\$125$; expected top-up per policy issued $\$100$; aggregate expected top-ups $\$10{,}000$)

What You'll Learn