Deriving Capital Sensitivity to Balance-Sheet Changes

Why Balance-Sheet Changes Affect a Capital Ratio

A capital adequacy ratio can change even when only one balance-sheet amount moves, and a single transaction can change both its numerator and denominator. To assess the effect, distinguish available capital from required capital, state how the proposed change affects each, and then calculate the ratio again. A first-order sensitivity is useful for a small change; an exact calculation is needed when the change is material or a target must be met precisely.

Definition: Let $A$ be available capital and $K$ be required capital, with $K>0$. The capital adequacy ratio is $R=A/K$. A target ratio $q$ is met when $R\ge q$. Unless otherwise stated, a loss or dividend reduces available capital dollar for dollar. Any change to required capital must be specified separately.

This definition alone does not guarantee that increasing required capital lowers the ratio: the effect depends on the sign of $A$. The usual interpretation that a higher denominator reduces adequacy applies when available capital is positive. If $A=0$, changing $K$ leaves the ratio at zero; if $A<0$, increasing $K$ makes $A/K$ less negative and therefore raises the ratio. In the examples below, available capital is positive.

Exact Changes and First-Order Sensitivity

Write the changes in available and required capital as $\Delta A$ and $\Delta K$. The starting amounts are $A$ and $K$, while the amounts after the change are $A+\Delta A$ and $K+\Delta K$. The post-change denominator must remain positive: $K+\Delta K>0$.

\Delta R =\frac{A+\Delta A}{K+\Delta K}-\frac{A}{K} =\frac{K\Delta A-A\Delta K}{K(K+\Delta K)}.

The exact formula follows by putting both ratios over the common denominator $K(K+\Delta K)$. Its numerator is $K(A+\Delta A)-A(K+\Delta K)=K\Delta A-A\Delta K$. This form shows why the changes cannot generally be treated as separate percentage-point adjustments to the ratio.

For small changes, the first-order approximation is the differential of $R=A/K$. It is a local approximation, not an exact identity for finite changes.

dR \approx \frac{\Delta A}{K}-\frac{A}{K^2}\Delta K.

The partial derivative with respect to available capital is $1/K>0$: raising $A$ raises the ratio. The partial derivative with respect to required capital is $-A/K^2$. It is negative if $A>0$, zero if $A=0$, and positive if $A<0$. These sign conditions matter whenever a sensitivity is interpreted.

Worked Example: Exact Change Versus a First-Order Estimate

A hypothetical insurer has available capital of $180 million and required capital of $120 million. A market shock reduces available capital by $12 million and increases required capital by $8 million. Calculate the starting and ending ratios, the exact change, and the first-order estimate. State one common mistake avoided by the exact calculation.

Method. First calculate the starting ratio. Then update both amounts and use the exact finite-change formula. Use the differential only as a comparison.

Starting ratio. $R=180/120=1.5$, or 150%.

Updated amounts. $A+\Delta A=180-12=168$ million and $K+\Delta K=120+8=128$ million. The post-change ratio is $168/128=1.3125$, or 131.25%.

Exact change. The ratio changes by $1.3125-1.5=-0.1875$, a decrease of 18.75 percentage points. The exact formula gives the same result:

\Delta R =\frac{120(-12)-180(8)}{120(128)} =\frac{-2880}{15360} =-0.1875.

First-order estimate.

dR \approx \frac{-12}{120}-\frac{180}{120^2}(8) =-0.1-0.1=-0.2.

The linear estimate predicts a decrease of 20 percentage points, compared with the exact decrease of 18.75 percentage points. The approximation differs because it omits finite-change effects. The check is that the calculated post-change ratio, $168/128$, is 1.3125 and its difference from 1.5 is exactly $-0.1875$.

Common mistake avoided: treating the denominator increase as a direct subtraction of $8/120$ from the ratio. The denominator sensitivity includes the factor $A/K=1.5$; its first-order contribution is $-1.5(8/120)=-0.1$, not $-8/120$.

Solving for a Maximum Loss

A maximum-loss calculation turns a target ratio into an inequality. The answer depends on how required capital responds to the loss. Suppose, for a stated scenario, a loss $L\ge0$ reduces available capital dollar for dollar and increases required capital by $\beta L$, where $\beta\ge0$ is a specified constant. With starting amounts $A_0$ and $K_0$, the target condition is

\frac{A_0-L}{K_0+\beta L}\ge q, \qquad K_0+\beta L>0.

For a positive target $q$, the denominator is positive, so multiplying both sides by it preserves the inequality. Rearranging gives

A_0-L\ge qK_0+q\beta L \quad\Longleftrightarrow\quad L\le \frac{A_0-qK_0}{1+q\beta}.

This expression is a usable nonnegative loss allowance when $A_0\ge qK_0$. If that condition fails, the starting position is already below the target, so there is no nonnegative loss allowance under this model. The formula is specific to the stated linear response of required capital; it should not be used if the scenario specifies a different relationship.

Worked Example: Maximum Loss With Higher Required Capital

A hypothetical insurer has available capital of $210 million and required capital of $160 million. Management sets a minimum ratio of 1.25. For a specified stress, each $1 million of loss reduces available capital by $1 million and increases required capital by $0.20 million. Find the largest loss consistent with the target and verify it.

Method. Let $L$ be the loss in millions of dollars. Use the scenario’s stated capital response in the target inequality; do not hold required capital fixed.

The starting ratio is $210/160=1.3125$, above the target. After a loss $L$, the ratio is $(210-L)/(160+0.20L)$. The denominator is positive for $L\ge0$, so the target inequality can be rearranged without changing its direction:

\frac{210-L}{160+0.20L}\ge1.25 \quad\Longleftrightarrow\quad 210-L\ge200+0.25L \quad\Longleftrightarrow\quad 10\ge1.25L \quad\Longleftrightarrow\quad L\le8.

The maximum loss is therefore $8 million. At that loss, available capital is $210-8=202$ million and required capital is $160+0.20(8)=161.6$ million. The ratio is $202/161.6=1.25$, exactly the target. Any larger loss violates the rearranged inequality.

Common mistake avoided: using the fixed-denominator allowance $210-1.25(160)=10$ million. That ignores the $0.20 million increase in required capital per $1 million of loss; using the correct denominator response reduces the allowance to $8 million.

Solving for a Maximum Dividend

A dividend is a transfer out of the insurer, so in this simplified setting it reduces available capital. Required capital may also change, but that effect must be supplied by the scenario rather than assumed. If a dividend $D\ge0$ reduces required capital by $\gamma D$, where $\gamma\ge0$, then the target condition is

\frac{A_0-D}{K_0-\gamma D}\ge q, \qquad K_0-\gamma D>0.

For $q>0$, multiplying by the positive denominator and collecting terms yields

A_0-D\ge qK_0-q\gamma D \quad\Longleftrightarrow\quad A_0-qK_0\ge(1-q\gamma)D.

When $1-q\gamma>0$, the inequality gives an upper bound on the dividend. If $1-q\gamma=0$, the target condition is independent of $D$ after rearrangement; if $1-q\gamma<0$, its direction when solving for $D$ reverses. In every case, also enforce $D\ge0$, the positive-denominator condition, and any other transaction limit in the scenario. Do not assume that a falling denominator automatically offsets a dividend: the effect depends on its size relative to the target.

Worked Example: Dividend Limit When Required Capital Falls

A hypothetical insurer has available capital of $300 million and required capital of $220 million. It must maintain a ratio of at least 1.25. A proposed dividend $D$, measured in millions of dollars, reduces available capital by $D$ and reduces required capital by $0.10D$. Find the largest permissible dividend, and state the largest dividend if payments must be made in whole cents.

Method. Substitute both effects into the target inequality. Check the denominator remains positive and verify the boundary value.

For the relevant range, required capital is $220-0.10D$. It remains positive at the boundary found below. The ratio condition becomes

\frac{300-D}{220-0.10D}\ge1.25 \quad\Longleftrightarrow\quad 300-D\ge275-0.125D \quad\Longleftrightarrow\quad 25\ge0.875D.

Thus the continuous maximum is $D=25/0.875=200/7=28.571428\ldots$ million dollars. At this boundary, available capital is $300-200/7=1900/7$ million and required capital is $220-(0.10)(200/7)=1520/7$ million. Their ratio is $(1900/7)/(1520/7)=1900/1520=1.25$. The denominator is positive.

The continuous maximum is $28,571,428.571… when expressed in dollars. If the dividend must be stated in whole cents, the largest permissible amount is $28,571,428.57: rounding down to the nearest cent keeps it below the continuous boundary. Rounding to $28,571,428.58 would exceed that boundary.

Common mistake avoided: subtracting the dividend from available capital but leaving required capital fixed. Here, the required-capital reduction is part of the stated scenario and must appear in the inequality.

Using Sensitivities Responsibly

A local sensitivity helps explain direction and approximate magnitude; it is not a substitute for updating the balance sheet. For example, when $A>0$, a loss that lowers $A$ and raises $K$ pushes the ratio down through both channels. A dividend that lowers both amounts has competing effects: the numerator reduction lowers the ratio, while the denominator reduction raises it. The exact target inequality determines which effect dominates for the scenario.

A useful calculation sequence is to record the starting amounts, write the post-change values as functions of the proposed transaction, state the domain in which the denominator stays positive, and then solve the target inequality. Check the boundary by substituting it into the original ratio, not just the rearranged equation. If an operational amount must be rounded, round in the direction that preserves the constraint.

Set the starting position.
Record $A_0$, $K_0$, the starting ratio, and the target $q$.

Model the balance-sheet response.
Write post-change available capital and required capital using the scenario’s specified effects.

Solve and check.
Apply the target inequality with a positive denominator, solve for the transaction amount, and verify the boundary in the original ratio.

Key takeaway: For $R=A/K$, the exact change is $\{K\Delta A-A\Delta K\}/\{K(K+\Delta K)\}$, while the first-order change is $\Delta A/K-A\Delta K/K^2$. A maximum loss or dividend is found by modeling both balance-sheet effects and solving the target inequality under its stated assumptions.

Check Your Understanding

For each problem, state the relevant assumption and check the result against the stated target or updated ratio.

A hypothetical insurer starts with available capital of $275 million and required capital of $220 million. A stress reduces available capital by $11 million and increases required capital by $5.5 million. Calculate the post-stress ratio to four decimal places. (Answer: $264/225.5=1.1707$)
A hypothetical insurer has available capital of $190 million and required capital of $160 million. A dividend of $12 million reduces available capital dollar for dollar; required capital is unchanged. Calculate the resulting ratio and say whether a target of 1.10 is met. (Answer: $178/160=1.1125$; yes)
A hypothetical insurer has $A_0=260$ million and $K_0=200$ million. A loss $L$ reduces available capital by $L$ and increases required capital by $0.10L$. What is the maximum loss that preserves a ratio of at least 1.20? (Answer: $L\le20/1.12=17.8571$ million)
A hypothetical insurer has $A_0=350$ million and $K_0=250$ million. A dividend $D$ reduces available capital by $D$ and required capital by $0.20D$. Find the continuous maximum dividend consistent with a target ratio of 1.30. (Answer: $D\le25/0.74=33.7838$ million)
At $A=180$ million and $K=120$ million, a small change has $\Delta A=-6$ million and $\Delta K=3$ million. Calculate the first-order change and the exact post-change ratio. (Answer: first-order change $=-0.0875$; exact post-change ratio $=174/123=1.4146$)

What You'll Learn