Choosing Rate and Capital Actions Under Risk Limits

Introduction

A rate action can improve expected underwriting profit without making a portfolio acceptable under a capital limit. Reinsurance can reduce the tail of retained losses, but its price may absorb part of the expected margin. An additional capital contribution can address a capital shortfall without changing the underlying loss distribution, but it uses a limited resource. A sound decision therefore tests profitability and risk feasibility separately.

This tutorial uses a one-period framework to compare these actions under a premium cap and a tail-risk limit. All scenarios and figures are hypothetical. The calculations are aligned with the SOA Exam ASTAM (Advanced Short-Term Actuarial Mathematics) syllabus.

Model and conventions: Let $X$ be the insurer’s one-period gross aggregate claim amount, $P$ its gross premium, and $R$ a deterministic reinsurance premium. Let $Z$ be the insurer’s retained aggregate claims after reinsurance. Underwriting profit is $P-R-Z$. For confidence level $q$, define $\operatorname{VaR}_q(Z)=\inf\{z:F_Z(z)\ge q\}$. The required capital measure used here is $K=\max(0,\operatorname{VaR}_q(Z)+R-P)$. Available capital $C$ makes the action feasible only if $K\le C$. A premium cap additionally requires $P\le P_{\max}$.

The capital measure is the $q$-quantile of the one-period underwriting loss $Z+R-P$, floored at zero. It is a specified decision metric, not a universal definition of regulatory capital. If a problem defines its capital measure differently, use that definition instead. Unless a case explicitly gives a charge for supplied capital, the charge is not deducted from underwriting profit.

\begin{aligned} \text{Expected underwriting profit} &= P-R-E[Z],\\ \text{Expected margin above retained cost} &= P-R-E[Z],\\ \text{Risk constraint} &: \max(0,\operatorname{VaR}_q(Z)+R-P)\le C,\\ \text{Premium constraint} &: P\le P_{\max}. \end{aligned}

For this framework, the expected margin above retained cost is the expected underwriting profit; it is not a guarantee of a positive outcome in every period. If capital has an explicit annual charge, show it separately from underwriting profit and subtract it only when comparing net economic results.

Why the Capital Formula Works

The formula follows from the underwriting loss for the period. Since $R$ and $P$ are fixed amounts, the random underwriting loss is $L=Z+R-P$. Adding a constant shifts every possible value of a random variable by that constant, so it shifts its quantile by the same amount.

\begin{aligned} F_{Z+R-P}(l) &=P(Z+R-P\le l)\\ &=P(Z\le l-R+P)\\ &=F_Z(l-R+P). \end{aligned}

Thus the smallest $l$ at which this distribution function reaches $q$ is $\operatorname{VaR}_q(Z)+R-P$. Taking the maximum with zero gives $K$. This derivation assumes deterministic premium and reinsurance cost for the period. If either is random, the constant-shift argument does not apply as written; the quantile must be calculated for the joint underwriting-loss variable.

For a fixed reinsurance arrangement and available capital $C$, the risk constraint can also be written as a minimum premium condition:

P\ge \operatorname{VaR}_q(Z)+R-C.

This inequality is useful, but it does not replace the premium cap or profitability test. If the required premium exceeds $P_{\max}$, that action cannot meet the capital limit through pricing alone. Conversely, a premium within the cap can still leave expected profit too small or required capital too high.

Worked Decision Method

Specify each action.
Record the gross premium, reinsurance terms and cost, retained claim amount, and any additional capital.

Calculate expected profit.
Find $E[Z]$, then calculate $P-R-E[Z]$. Keep any stated cost of capital separate until making a net economic comparison.

Calculate the tail measure.
Find $\operatorname{VaR}_q(Z)$ from the retained-loss distribution, then calculate $K$.

Apply both limits.
Check $P\le P_{\max}$ and $K\le C$, where $C$ includes only capital actually available under the proposed action.

Compare feasible actions.
Compare expected profit, required capital, and any explicitly stated capital charge. Do not select an action merely because one measure improves.

Rate Increases Under a Binding Capital Limit

A premium increase shifts the underwriting loss downward dollar for dollar in this model, so it reduces the capital measure when the measure is positive. It does not change the distribution of claims. A premium cap limits how far this lever can be used. Discrete loss distributions also require care: a quantile is determined by the first cumulative probability that reaches $q$, not by interpolating between loss values.

Worked Example: Finding a Feasible Rate Within the Cap

A hypothetical insurer has one-period claims of $500 with probability 0.90 and $1,500 with probability 0.10. It charges $900 currently, may charge no more than $1,100, and has $450 available for the capital measure. Use $q=0.95$, with no reinsurance. Calculate current expected profit and required capital, then test a premium increase to the cap.

Method. First calculate the expected claim and its $0.95$-quantile. With no reinsurance, $Z=X$ and $R=0$.

E[X]=0.90(500)+0.10(1{,}500)=450+150=\$600.

The cumulative probability at $500 is 0.90, which is below 0.95. At $1,500 it is 1.00. Therefore $\operatorname{VaR}_{0.95}(X)=\$1,500$.

At the current premium of $900, expected profit is $900-600=\$300$. Required capital is $\max(0,1{,}500-900)=\$600$, which exceeds the $450 available. The current action is profitable in expectation but fails the capital constraint.

At the capped premium of $1,100, expected profit is $1{,}100-600=\$500$, and required capital is $\max(0,1{,}500-1{,}100)=\$400$. The premium is at the cap, and $400 is no greater than the $450 available. This action is feasible and has a positive expected margin.

Check. The rate increase adds $200 to premium, so expected profit rises by $200, from $300 to $500, and required capital falls by $200, from $600 to $400. The common mistake is to calculate the expected claim as $650; the correct weighted sum is $450+150=\$600$.

Reinsurance, Rate, and Capital Together

Reinsurance changes the retained-loss distribution and adds a reinsurance cost. The tail effect depends on both the contract and the selected confidence level. A layer may reduce a very high quantile substantially while leaving a lower quantile unchanged. The premium for reinsurance also enters the underwriting-loss quantile in this framework, because $R$ is part of the period’s underwriting cost.

Worked Example: A Tail Limit That Needs Both Reinsurance and Capital

A hypothetical portfolio has gross claims $X$ of $0 with probability 0.80, $2,000 with probability 0.15, and $6,000 with probability 0.05. The insurer may charge at most $1,000, has $500 of available capital, and uses $q=0.975$. It can buy excess-of-loss reinsurance attaching at $2,000 for a fixed premium of $250. An additional capital contribution of at most $800 is available at an annual charge of 4% of the contribution. Compare no reinsurance with the reinsurance option at the premium cap, and determine whether either can meet the risk limit without or with an allowed capital contribution.

Method. Compare each action’s retained claims, expected profit, and capital requirement. The reinsurance recovery is $(X-2{,}000)_+$, so retained claims are $Z=\min(X,2{,}000)$.

Without reinsurance, expected claims are $0.80(0)+0.15(2{,}000)+0.05(6{,}000)=300+300=\$600$. Since the cumulative probability at $2,000 is 0.95, below 0.975, the gross-claim quantile is $\operatorname{VaR}_{0.975}(X)=\$6,000$. At premium $1,000, expected profit is $1{,}000-600=\$400$, and required capital is $6{,}000-1{,}000=\$5,000$. It exceeds the $500 available by $4,500. An additional contribution of $4,500 would be needed, which is above the allowed $800; therefore this action is infeasible under the stated capital limit.

With reinsurance, retained claims are $0 with probability 0.80 and $2,000 with probability $0.15+0.05=0.20$. Thus $E[Z]=0.20(2{,}000)=\$400$. At $q=0.975$, the retained-loss quantile is $2,000. At the capped premium, expected underwriting profit is $1{,}000-250-400=\$350$, while required capital is $\max(0,2{,}000+250-1{,}000)=\$1,250$. Reinsurance alone is not enough: it leaves a $750 capital shortfall.

An additional $750 of capital meets the risk limit exactly: available capital becomes $500+750=\$1,250$. The contribution is within the $800 maximum. Its stated charge is $0.04(750)=\$30$, so net expected economic profit after this charge is $350-30=\$320$. The action meets the premium and capital limits and remains positive after the capital charge.

Check. Reinsurance reduces the $0.975$-quantile from $6,000 to $2,000, but the reinsurance premium still contributes $250 to the capital measure. The common mistake is to call reinsurance alone feasible because the retained quantile is only $2,000; the risk calculation must also include $R$ and subtract $P$.

Comparing Capital and Reinsurance Actions

A capital contribution raises available capital but does not lower required capital $K$. Reinsurance can lower $K$, but it also changes expected retained losses and incurs a premium. A direct comparison should keep these effects visible. If the decision framework assigns a cost to capital, include that charge in a separate net economic measure rather than disguising it as a change in expected claims.

Worked Example: Choosing Between Reinsurance and More Capital

A hypothetical portfolio has claims of $0 with probability 0.90, $1,000 with probability 0.08, and $5,000 with probability 0.02. The current premium is $1,100, the premium cap is $1,400, available capital is $1,500, and $q=0.99$. The insurer can either operate without reinsurance and contribute up to $2,200 of additional capital at an annual charge of 5%, or buy excess-of-loss reinsurance attaching at $1,000 for $250. Compare both actions at the premium cap.

Method. Find the gross and retained-loss quantiles, calculate profit and capital needs at $P=\$1,400$, and then include the capital charge only for the capital-contribution option.

The expected gross claim is $0.90(0)+0.08(1{,}000)+0.02(5{,}000)=80+100=\$180$. The cumulative probability at $1,000 is 0.98, below 0.99, so $\operatorname{VaR}_{0.99}(X)=\$5,000$.

Without reinsurance, expected underwriting profit at the cap is $1{,}400-180=\$1,220$. Required capital is $5{,}000-1{,}400=\$3,600$, so the additional contribution needed beyond the $1,500 already available is $3{,}600-1{,}500=\$2,100$. This is within the $2,200 maximum. Its charge is $0.05(2{,}100)=\$105$, giving net expected economic profit of $1{,}220-105=\$1,115$.

With reinsurance, retained claims are $\min(X,1{,}000)$. They are $0 with probability 0.90 and $1,000 with probability 0.10, so $E[Z]=\$100$ and $\operatorname{VaR}_{0.99}(Z)=\$1,000$. Expected underwriting profit is $1{,}400-250-100=\$1,050$. Required capital is $\max(0,1{,}000+250-1{,}400)=\$0$, so no additional contribution is needed. Both actions meet the premium and capital constraints, but the capital-contribution option has higher net expected economic profit: $1,115 versus $1,050.

Check. The no-reinsurance option needs $2,100 of additional capital, not $3,600: $3,600 is total required capital, of which $1,500 is already available. The comparison also depends on the stated 5% capital charge; without that charge, the two options’ expected underwriting profits would still differ, but the decision criterion would be different.

Common Traps and Decision Checks

Stopping at expected profit. A positive expected margin does not prove the capital constraint is met. Calculate the specified tail measure separately.
Forgetting the premium cap. A premium that would meet the capital inequality is not an available action if it exceeds $P_{\max}$.
Ignoring reinsurance cost. In this framework, $R$ is both deducted in expected underwriting profit and included in the underwriting-loss quantile.
Using the wrong quantile boundary. For discrete losses, use the smallest loss value whose cumulative probability is at least $q$. Do not interpolate unless a different method is explicitly prescribed.
Confusing required and available capital. A contribution increases available capital; it does not reduce $K$. Report both quantities and their difference.
Counting a capital charge twice. The capital charge is a separate economic cost when specified. It is not part of $K$ in the stated formula.

Key takeaway: A defensible decision must pass the premium cap and capital limit, while also reporting expected underwriting profit. Rate increases, reinsurance, and additional capital act on different parts of the calculation; compare them using the same loss model and the same confidence level.

Check Your Understanding

For each hypothetical case, calculate the requested quantities before checking the answer. Use the stated VaR definition and capital measure.

A portfolio has claims of $300 with probability 0.80 and $1,800 with probability 0.20. Premium is $950, $q=0.90$, and available capital is $700. Find expected profit and required capital, then state whether the risk limit is met. (Answer: Expected profit $350; required capital $850; not feasible, short by $150.)
Claims are $0 with probability 0.85, $1,000 with probability 0.10, and $4,000 with probability 0.05. With no reinsurance, premium is $700, $q=0.95$, and available capital is $250. Find expected profit and required capital. (Answer: Expected profit $400; $\operatorname{VaR}_{0.95}=\$1,000$; required capital $300; not feasible, short by $50.)
Claims are $0 with probability 0.80, $1,500 with probability 0.15, and $4,500 with probability 0.05. Excess-of-loss reinsurance attaches at $1,500 and costs $180. At $q=0.95$, premium is $800 and available capital is $900. Find expected retained claims, expected profit, and required capital. (Answer: Expected retained claims $300; expected profit $320; required capital $880; feasible with $20 of capital remaining.)
Claims are $0 with probability 0.90, $1,000 with probability 0.08, and $5,000 with probability 0.02. Excess-of-loss reinsurance attaches at $1,000 and costs $200. At $q=0.99$, premium is $1,250. Find the retained-loss VaR and required capital. (Answer: $\operatorname{VaR}_{0.99}(Z)=\$1,000$; required capital $0.)
Claims are $0 with probability 0.90, $2,000 with probability 0.08, and $8,000 with probability 0.02. Premium is capped at $1,800, $q=0.99$, and existing capital is $2,500. Without reinsurance, additional capital is limited to $3,000. Alternatively, excess-of-loss reinsurance attaches at $2,000 and costs $350. At the premium cap, find the additional capital needed without reinsurance and determine whether the reinsurance option meets the capital limit without an additional contribution. (Answer: Without reinsurance, required capital is $6,200 and additional capital needed is $3,700, exceeding the $3,000 limit; with reinsurance, retained-loss VaR is $2,000 and required capital is $550, which is covered by existing capital.)

What You'll Learn