Challenging Agency Assumptions and Adjustments

Why an Agency Assumption Deserves a Traceable Challenge

A rating analysis can change materially because of an assumption about exposure, loss severity, asset quality, or the treatment of a risk. A challenge is useful only when it identifies the assumption, establishes what evidence bears on it, and shows the effect of a reasonable alternative. Simply preferring a different result is not evidence that the agency’s treatment is wrong.

Rating agencies use their own methodologies, definitions, and judgment. The calculations in this tutorial use a simplified hypothetical framework to isolate the effects of assumptions and adjustments; they are not formulas used by any particular agency. The objective is to evaluate an analytical treatment, not to claim that an agency must adopt a proposed alternative. This approach is aligned with the CAS Exam 9 (Risk Management for Actuaries) syllabus.

Definition: An assumption is an input or judgment used to estimate a risk, resource, or capital requirement, such as an exposure amount or loss factor. An adjustment is a change to an amount used in the analysis, such as an added capital charge, a haircut to an asset, or a reduction in eligible capital. A challenge is an evidence-based evaluation of whether the assumption or adjustment is appropriate for the stated purpose and whether a different treatment is better supported.

Begin by separating three questions that can otherwise become muddled:

What did the analysis do? Identify the input, adjustment, calculation, and amount affected.
Why was it done? Find the stated rationale, relevant risk, and intended prudence or conservatism.
What does the evidence support? Test the data, method, and alternative, including limitations and uncertainty.

The source trail matters. For each material assumption or adjustment, record the source data and date, definitions and exclusions, calculation steps, approvals or judgment applied, and the affected capital measure. If a number cannot be traced from source to result, the first task is to resolve that gap—not to substitute a more favorable number.

Quantifying the Effect

For a simple comparison, let $K$ denote eligible capital resources, $C$ denote required capital, and $q$ denote a target capital multiple. The ratio is $K/C$, and headroom above the target is the resources remaining after allowing for $q$ times required capital.

\text{Capital ratio}=\frac{K}{C}, \qquad H=K-qC

A positive $H$ means capital exceeds the stated target; a negative $H$ means it falls short. These measures are useful comparison tools, but an agency may use different definitions, thresholds, or measures. Always state which quantities and target are being used.

Derivation: Suppose an adjustment changes eligible capital by $\Delta K$ and required capital by $\Delta C$. The adjusted amounts are $K+\Delta K$ and $C+\Delta C$. Substituting them into the definition of headroom gives $$ H_{\text{adjusted}}=(K+\Delta K)-q(C+\Delta C) =K-qC+\Delta K-q\Delta C =H+\Delta K-q\Delta C. $$ Thus a dollar reduction in eligible capital lowers headroom by one dollar, while a dollar increase in required capital lowers it by $q$ dollars. This follows directly from the headroom definition; it does not assume that every agency measures capital in this way.

The ratio and headroom answer different questions. The ratio expresses resources relative to required capital; headroom expresses the dollar distance from a particular target. A change can affect both, and a ratio alone does not reveal the dollar cushion. For a ratio comparison, calculate the adjusted numerator and denominator separately rather than applying a percentage change to the old ratio.

A Disciplined Review Sequence

Pin down the treatment.
Quote or restate the assumption or adjustment precisely. Identify whether it changes an exposure, a capital charge, eligible resources, or more than one item.

Reconstruct the calculation.
Follow the source data through definitions, transformations, factors, and any aggregation. Reconcile the result to the reported amount.

Test the rationale against evidence.
Assess relevance, completeness, timeliness, consistency, and limitations. Determine whether the evidence addresses the risk the treatment is meant to capture.

Quantify a bounded alternative.
Recalculate the affected amounts under a clearly stated alternative. Show the resulting ratio and headroom, and retain any conservatism the evidence does not resolve.

State the conclusion at the right strength.
Distinguish a factual or calculation error from a judgment difference, and state what evidence would change the conclusion.

Evidence is strongest when it is specific to the assumed exposure and period, independently reconcilable, and consistent with the calculation’s purpose. A favorable recent experience may be relevant but does not automatically invalidate a stressed assumption: it may cover a short period, omit severe conditions, or reflect a different portfolio. Conversely, a general statement that an assumption is conservative does not establish that its size is appropriate. A defensible alternative should explain both why its inputs fit and what uncertainty remains.

Worked Examples

Worked Example: Challenging a Risk-Factor Assumption

All amounts are hypothetical and in millions of dollars. An insurer’s simplified rating analysis has eligible capital $K=690$, a required-capital component of 240 apart from a reserve factor, and a reserve exposure of 2,000. The company uses an 8% factor, while an agency analysis uses 11%. Assume the reserve charge is exposure times the factor, the two charges are added without diversification credit, and the target multiple is $q=1.50$. A reconciled review of the relevant exposure supports a 9% factor for this comparison. Quantify the agency treatment and the evidence-supported alternative.

Method. Rebuild the factor-based charge, add the other capital component, then calculate the ratio and headroom for each treatment.

At 8%, the reserve charge is $2{,}000 \times 0.08=160$. Total required capital is $240+160=400$. The ratio is $690/400=1.725$, or 172.5%. Headroom is $690-1.50(400)=690-600=90$.

At 11%, the reserve charge is $2{,}000 \times 0.11=220$, an increase of $220-160=60$. Total required capital is $240+220=460$. The ratio is $690/460=1.50$, or 150.0%. Headroom is $690-1.50(460)=690-690=0$.

At the proposed 9% factor, the charge is $2{,}000 \times 0.09=180$, and total required capital is $240+180=420$. The ratio is $690/420=1.642857$, or 164.3% to one decimal place. Headroom is $690-1.50(420)=690-630=60$.

The alternative lies between the company and agency factors, reducing required capital by $460-420=40$ relative to the agency treatment. To defend it, the company would need to show that the exposure definition and underlying data reconcile to 2,000, that the reviewed experience is relevant to the factor’s intended risk, and that the 9% estimate is supported by a documented method. Common mistake: applying the 3 percentage-point factor difference to total capital of 400 gives 12, which uses the wrong base. The factor applies to exposure of 2,000, so the charge difference is $2{,}000(0.11-0.08)=60$. The alternative is a sensitivity, not proof that the agency must accept it.

Worked Example: Testing a Recoverable Haircut

In a hypothetical analysis, eligible capital before an adjustment is $K=620$, required capital is $C=360$, and the target multiple is $q=1.60$. The analysis reduces eligible capital by 45 for uncertainty in reinsurance recoverables. The company’s review finds that a haircut of 20 is better supported for the recoverables in scope, based on reconciled balances and documented collateral. Assume the adjustment changes eligible capital only. Calculate the reported and alternative results.

Method. Apply each haircut to eligible capital, leave required capital unchanged, and recalculate both measures.

Before the haircut, the ratio is $620/360=1.7222$, or 172.2%. Headroom is $620-1.60(360)=620-576=44$.

With the 45 haircut, adjusted eligible capital is $620-45=575$. The ratio is $575/360=1.5972$, or 159.7%. Headroom is $575-576=-1$, a shortfall of 1 against the specified target.

With a 20 haircut, adjusted eligible capital is $620-20=600$. The ratio is $600/360=1.6667$, or 166.7%. Headroom is $600-576=24$. Relative to the 45 haircut, the alternative increases eligible capital and headroom by $45-20=25$; it does not eliminate the haircut.

A persuasive challenge would reconcile gross and net recoverable balances, identify which balances are collateralized, and demonstrate that the collateral and collection evidence apply to the same recoverables and valuation date. It should also explain risks not removed by collateral or past collections. Common mistake: subtracting the 45 haircut from required capital treats a resource adjustment as a risk-charge adjustment. The stated treatment reduces $K$, so the correct denominator remains 360.

Worked Example: Evaluating a Concentration Add-On

A hypothetical insurer has eligible capital $K=520$, base required capital of 300, and a concentration add-on of 40. The target multiple is $q=1.50$. A review of exposure mapping and the documented basis for the add-on supports an alternative add-on of 15, while not establishing that concentration risk is zero. Assume the add-on increases required capital dollar for dollar and does not change eligible capital. Compare the results.

Method. Add each concentration amount to base required capital, then calculate ratio and headroom.

Under the 40 add-on, required capital is $300+40=340$. The ratio is $520/340=1.5294$, or 152.9%. Headroom is $520-1.50(340)=520-510=10$.

Under the 15 alternative, required capital is $300+15=315$. The ratio is $520/315=1.6508$, or 165.1%. Headroom is $520-1.50(315)=520-472.5=47.5$. The 25 reduction in the add-on raises headroom by $1.50(25)=37.5$, consistent with the change from 10 to 47.5.

The alternative is defensible only if the exposure map captures the relevant concentrations and the evidence supports the residual amount after any recognized risk mitigants. The conclusion should identify what the add-on is intended to capture and avoid treating a lower estimate as evidence of no risk. Common mistake: subtracting 25 directly from headroom gives a change of 25. But the add-on changes required capital, and with a target multiple of 1.50 the headroom effect is $1.50(25)=37.5$.

Common Traps and a Strong Conclusion

A numerical challenge can be internally correct and still be weak if it addresses the wrong issue. Check the direction and location of the adjustment before recalculating. A haircut to eligible assets changes $K$; an added risk charge changes $C$. If both change, show both changes rather than hiding one in a net figure.

Confusing a difference with an error. A different estimate may reflect judgment, not a factual mistake. State whether the evidence shows a miscalculation, a data mismatch, or a reasonable disagreement about method.
Using evidence that does not match the assumption. A balance at a different date, a different exposure definition, or a different risk period may not validate the proposed input.
Calling an alternative “more conservative” without quantifying it. Show the input, recalculated capital, and change in ratio and headroom.
Ignoring the uncertainty that remains. If the evidence supports only a range, show a sensitivity or retain a prudent residual adjustment rather than asserting a precise result unsupported by the data.
Overstating the consequence. A calculated capital effect is not, by itself, a prediction of a rating outcome. Rating decisions may depend on other factors and agency-specific judgment.

A concise, defensible conclusion identifies the disputed treatment, traces its rationale and data, quantifies the agency case and alternative on a consistent basis, and states the limits of the evidence. The strength of the recommendation should match the strength of the evidence: correct demonstrable errors firmly, and frame judgment differences as bounded alternatives for consideration.

Check Your Understanding

For each question, state the relevant treatment, perform the calculation, and interpret the result using the specified target.

A hypothetical company has $K=450$, $C=300$, and target $q=1.40$. An agency adjustment increases required capital by 25 without changing eligible capital. Calculate the original and adjusted ratios and headroom. (Answer: Ratios 150.0% and 138.5%; headroom 30 and -5, respectively.)
A hypothetical analysis has $K=510$, $C=280$, and target $q=1.75$. An asset haircut reduces eligible capital by 18. Calculate the adjusted ratio and headroom, and compare headroom with the original. (Answer: Adjusted ratio 175.7%; adjusted headroom 2 versus original headroom 20, a decrease of 18.)
A hypothetical risk charge is exposure times a factor, with exposure 900. Other required capital is 120, eligible capital is 350, and the target is 160%. Compare a 6% factor with a proposed 9% factor, assuming additive charges and no diversification credit. (Answer: Required capital is 174 and 201; ratios are 201.1% and 174.1%; headroom is 71.6 and 28.4, respectively.)
A hypothetical analysis has $K=800$, $C=500$, and target $q=1.50$. An agency treatment reduces eligible capital by 20 and adds 30 to required capital. An evidence-supported alternative reduces eligible capital by 8 and adds 12 to required capital. Calculate both adjusted ratios and headroom, then state the headroom difference. (Answer: Agency treatment: ratio 147.2%, headroom -15. Alternative: ratio 154.7%, headroom 24. The alternative has 39 more headroom.)

What You'll Learn