Why Earnings Quality Matters to Capital Resilience
An insurer’s capital position can absorb losses, but a strong position today does not by itself show how readily the insurer could recover from a future shock. An assessment of capital resilience also asks whether the insurer can generate earnings consistently and retain enough of them to rebuild capital. The quality of earnings matters because reported profit may include temporary items, volatile sources, or gains that are not expected to recur.
A rating assessment may consider the level of earnings, their variability over time, the diversity of their sources, and the portion likely to remain in the business. These are related but distinct questions. High earnings in one year do not establish a durable earnings base; several sources of profit do not guarantee that they will remain independent under stress; and accounting income does not automatically become available capital.
Recurring Earnings and Normalization
Reported earnings are the starting point, not necessarily the best estimate of a continuing earnings run rate. A useful analysis identifies items that are unusual, nonrecurring, or not representative of ordinary operations. Examples might include a one-time gain on a hypothetical asset sale, a temporary tax benefit, or a discrete restructuring charge. Whether an item is truly nonrecurring requires judgment and evidence; an insurer should not label a recurring cost “one-time” merely because it is inconvenient.
A simple normalization exercise adjusts reported earnings for specified items. If a reported figure includes an after-tax one-time gain, remove it to estimate recurring earnings. If it includes an after-tax one-time expense that is genuinely nonrecurring, add it back. The result is an analytical estimate, not a replacement for reported results. The adjustment should be transparent, consistently applied, and checked against the circumstances that produced the item.
The formula is only as sound as its classifications. A series of annual “one-time” charges may actually indicate a continuing cost of doing business. Similarly, an investment gain may be repeatable in some form but still too volatile to treat as dependable operating earnings. The analyst should distinguish “could happen again” from “reasonably expected to recur at a stable level.”
Worked Example: Normalizing Earnings Before Estimating Retention
A hypothetical insurer reports after-tax earnings of $42 million. The figure includes a $12 million after-tax gain from a sale stipulated to be nonrecurring and a $5 million after-tax restructuring expense stipulated to be nonrecurring. For a planning illustration, assume dividends equal 30% of either reported or normalized earnings. Estimate normalized earnings and the retained amount under both measures. State the method and identify the mistake it avoids.
Method. Adjust reported earnings by removing the one-time gain and adding back the one-time expense:
The assumed dividend payout is 30%, so the retained proportion is \(1-0.30=0.70\), or 70%. Retention based on normalized earnings is:
For comparison, retention using the unadjusted reported earnings is:
The difference is \(\$29.4-\$24.5=\$4.9\) million. This difference equals 70% of the net \(\$7\) million by which the specified one-time items raised reported earnings: \((\$12-\$5)\times0.70=\$4.9\) million. Thus, the arithmetic reconciles. The common mistake is to project all $42 million as a continuing source of capital. Under the stated assumptions, the normalized estimate supports $24.5 million of retention, not $29.4 million.
Stability and Diversity of Earnings Sources
Earnings stability concerns how much earnings vary and why. A short history may be informative, but it cannot establish that future earnings will be stable. The analyst should consider whether results are driven by recurring operations, volatile investment outcomes, unusually favorable claims experience, or other factors that could reverse. Stability is a judgment about the underlying drivers as well as a description of past numbers.
Diversification of earnings sources can support resilience when different sources contribute to earnings and their adverse movements do not reliably occur together. For example, underwriting earnings and fee income may arise from different activities. But a list of business lines is not proof of effective diversification: sources can still depend on common economic conditions, counterparties, customers, or management processes. A small number of years with offsetting results is evidence to examine, not proof that the offset will persist.
The following comparison uses a simple range rather than a dependency model. It illustrates how total earnings may look steadier than either source in a particular observed period; it does not establish a future relationship between the sources.
Worked Example: Interpreting Offsets Between Earnings Sources
A hypothetical insurer reports two stipulated recurring after-tax earnings sources, in millions of dollars, over three years. Source A is underwriting earnings; Source B is fee income. No nonrecurring items are included. Assess the observed stability of total earnings and explain what the short record does and does not establish.
| Year | Source A | Source B | Total earnings |
|---|---|---|---|
| 1 | $20 million | $10 million | $30 million |
| 2 | $24 million | $6 million | $30 million |
| 3 | $16 million | $14 million | $30 million |
Method. Calculate each source’s observed range by subtracting its minimum from its maximum. Source A ranges from $16 million to $24 million:
Source B ranges from $6 million to $14 million:
For each year, add the two sources. The totals are \(\$20+\$10=\$30\) million, \(\$24+\$6=\$30\) million, and \(\$16+\$14=\$30\) million. The observed total range is therefore \(\$30-\$30=\$0\) million. In these three observations, movements in the two sources offset exactly.
The result supports a limited statement: total earnings were stable across the three supplied years despite variation in each source. It does not show that future earnings will remain at $30 million, that the sources will continue to offset, or that their underlying risks are independent. A common mistake is to treat this exact historical offset as a guaranteed diversification benefit. The correction is to describe the observed pattern, then investigate the sources’ drivers and the possibility that they could weaken together.
From Earnings to Capital Replenishment
Earnings can replenish capital only to the extent that they are retained and available to support the capital measure being assessed. Dividends and other distributions reduce the amount retained. In a simplified illustration that excludes all other changes, the capital balance after a period equals the starting balance plus earnings less distributions. This relationship follows by separating the period’s increase in resources into the part kept in the business and the part distributed.
If distributions are a fixed fraction \(p\) of net earnings, and earnings are positive, then retained earnings are \((1-p)\) times net earnings. This is an assumption about the payout policy, not a universal rule. Actual capital changes may also reflect other items, and a rating analysis may use its own definitions of capital and earnings. The simplified relationship is useful for understanding the mechanism, but it should not be mistaken for a complete balance-sheet calculation.
Use a stated recurring earnings estimate or an explicitly defined stress scenario; do not silently substitute a favorable reported year.
Determine the amount or proportion of earnings assumed to be paid out rather than retained.
Subtract distributions from earnings, keeping the time period and units consistent.
Under a constant-earnings, no-other-changes assumption, accumulate retained earnings until the shortfall is recovered.
Worked Example: Comparing Capital Recovery Under Two Earnings Levels
A hypothetical insurer begins with $300 million of capital and experiences an immediate $60 million capital loss. After the loss, assume annual after-tax earnings are either $25 million in a baseline case or $15 million in a stress case. In both cases, dividends equal 20% of earnings. Assume earnings and the payout proportion remain constant each year, and assume no other capital changes, interest, new capital contributions, or further losses. Estimate the years required to recover the $60 million loss in each case.
Method. First calculate the post-loss capital:
The retained proportion is \(1-0.20=0.80\), or 80%. In the baseline case, annual retained earnings are:
The number of years required under the stated constant-earnings assumptions is the shortfall divided by annual retention:
Checking the accumulated amounts, after one year capital is \(\$240+\$20=\$260\) million; after two years it is \(\$280\) million; after three years it is \(\$300\) million. The original capital level is recovered exactly.
In the stress case, annual retained earnings are:
The recovery time is:
Checking, post-loss capital rises by $12 million per year: \(\$240\), \(\$252\), \(\$264\), \(\$276\), \(\$288\), then \(\$300\) million after five years. Thus the calculations reconcile in both cases. The common mistake is to divide the shortfall by gross earnings, which would give \(60/25=2.4\) years or \(60/15=4\) years. That ignores dividends. Under the assumptions given, the correct recovery estimates use retained earnings and are 3 years and 5 years.
Interpreting the Evidence Carefully
A strong earnings assessment combines arithmetic with a reasoned account of what could sustain or weaken the earnings base. It does not treat a single ratio or historical average as a complete conclusion. In particular, an analyst should distinguish the following:
- Level and stability: a high average can coexist with large year-to-year variation. Examine both the amount earned and the pattern that produced it.
- Recurring and nonrecurring items: adjustments should be specific and supported. Repeated exceptional charges may be evidence of an ongoing burden rather than a reason to keep adding them back.
- Source diversity and offsets: multiple sources may reduce reliance on any one activity, but the observed mix may not persist. Ask whether adverse conditions could affect several sources at once.
- Retention and distributions: earnings paid out cannot simultaneously be counted as retained capital. Any recovery estimate depends on its payout and earnings assumptions.
- Scenario dependence: recovery time calculated from constant annual earnings is a simplified illustration. If earnings fall after a shock, distributions change, or further losses occur, the actual path will differ.
Rating agencies may use different measures, evidence, and judgment when considering earnings and capital resilience. The analysis here is conceptual and does not reproduce a particular agency’s formula. It is aligned with the CAS Exam 9 (Risk Management for Actuaries) syllabus: the essential discipline is to state the earnings basis and assumptions clearly, then distinguish what the calculations show from what remains uncertain.
Check Your Understanding
For each hypothetical problem, state the calculation or judgment that supports your answer.
- An insurer reports after-tax earnings of $28 million, including a stipulated nonrecurring $6 million gain and a stipulated nonrecurring $2 million expense. What are normalized earnings? (Answer: $24 million.)
- A hypothetical insurer’s two recurring earnings sources are $14 million and $9 million in Year 1, $11 million and $12 million in Year 2, and $17 million and $6 million in Year 3. What is the total earnings range across the three years, and what limitation should be noted? (Answer: Totals are $23 million, $23 million, and $23 million, so the range is $0 million; three observations do not establish that the offset will continue.)
- A company has normalized after-tax earnings of $32 million and assumes a 25% dividend payout. How much is retained, and what common error would overstate replenishment? (Answer: $24 million is retained; using all $32 million as retained earnings ignores dividends.)
- A hypothetical insurer has a $45 million capital shortfall, annual recurring earnings of $18 million, and a 25% payout ratio. With constant earnings and no other capital changes, how long is the estimated recovery period? (Answer: Retention is $18 million × 0.75 = $13.5 million per year; $45 million ÷ $13.5 million per year = 3⅓ years.)
- Two earnings sources each vary substantially over a short record, while their combined total is unchanged. What can be concluded, and what cannot be concluded from that observation alone? (Answer: The combined earnings were stable over the observed period; the observation alone does not establish that future offsets will persist or that the sources are independent.)