CNSF Actuario Daños common mistakes
CNSF Actuario Daños · Understand the traps and how to avoid them.
Applying the reserve formula (e.g., reserva de riesgos en curso or reserva de siniestros pendientes) without first verifying which calculation basis the LISF and the CNSF circular actually require for that ramo, then defending the number as if the basis were irrelevant.
Candidates memorize a single formula from a textbook or study note and reuse it across ramos, forgetting that Mexican regulation prescribes specific bases (devengado vs. no devengado, per-ramo methods) and that the exam tests whether you can select the correct basis before computing.
What to do instead
Before any reserve calculation, write down the ramo, the applicable circular provision, and the required basis. Build a one-page decision table mapping each ramo to its reserve method and practice selecting the method from a scenario without doing arithmetic.
Treating frequency and severity as independent when the scenario describes a common shock (e.g., a hurricane or earthquake that simultaneously raises claim counts and average claim size), then summing variances or multiplying moments as if independence held.
Standard aggregate-loss formulas (compound Poisson, Panjer recursion) assume independence, and candidates default to them mechanically without checking the scenario's correlation structure.
What to do instead
Scan every modeling prompt for language implying common cause. If present, state explicitly that independence fails and use a copula, common-shock, or scenario-based approach; practice writing one sentence justifying the dependence assumption before choosing a model.
Computing a solvency capital requirement by summing standalone risk charges (suscription, mercado, credito, operacional) instead of applying the aggregation structure with correlations or the regulatory formula.
Candidates learn each risk module in isolation and, under time pressure, add them because addition is the simplest operation; they miss that the ERS framework aggregates with dependence assumptions and that simple summation overstates capital.
What to do instead
Memorize the aggregation logic (correlation matrix or prescribed formula) as a distinct step. Practice problems that give module-level charges and require the aggregated SCR, and always show the aggregation step separately from module calculation.
Quoting a reinsurance treaty's cession without checking whether the retention, limit, or reinstatement provisions change the net retained loss, and confusing quota share with surplus in the arithmetic.
Reinsurance problems look like simple percentage splits, so candidates apply the ceded percentage directly to the gross loss and ignore deductibles, event limits, and reinstatement premiums that the treaty terms specify.
What to do instead
For every reinsurance item, draw the layer structure (retention, ceded layer, limit) and compute gross, ceded, and net in separate lines. Do drills where the same gross loss is run through quota share, surplus, and excess-of-loss to see how the net differs.
Calculating a tariff premium by loading expenses and profit onto the pure premium without distinguishing acquisition, administration, and commission costs, or applying the loading to the wrong base (premium vs. pure premium).
Candidates conflate the expense ratio's denominator with the premium being solved for, producing a circular calculation they resolve incorrectly; they also omit that some loadings apply to gross premium and others to the pure premium.
What to do instead
Write the premium equation explicitly (P = pure premium / (1 - expense ratio - profit margin)) and verify the base of each loading. Solve one numeric example forward and backward to confirm the equation balances.
Answering ethics items by citing the general duty of care or 'acting in the client's interest' without naming the specific obligation (conflict-of-interest disclosure, suitability, confidentiality, or the CNSF conduct rule) that the scenario triggers.
Ethics questions invite generic virtue language, and candidates default to broad principles because they have not mapped the concrete normative articles to fact patterns.
What to do instead
Build a table of specific conduct obligations with one trigger phrase each (e.g., 'undisclosed commission' maps to conflict-of-interest disclosure). Answer ethics items by naming the obligation and the action it requires, not the principle.
Presenting a technical result to a regulator or board without reconciling it to the statutory basis, or reporting a figure that mixes bases (e.g., an economic-value estimate presented as a regulatory reserve).
Candidates focus on getting the number right and treat the reporting format as an afterthought, forgetting that regulatory filings require specific bases and reconciliation notes.
What to do instead
Treat every reporting answer as two parts: the figure and its basis. Practice writing a short reconciliation note (statutory vs. economic, gross vs. net) alongside each number, and check the required disclosure items before finalizing.
Selecting an investment portfolio for technical reserves by maximizing yield without checking duration matching, liquidity, or the regulatory eligible-asset and diversification limits.
Candidates treat the investment domain as a finance problem separate from insurance, so they optimize return and ignore that reserves back specific liabilities and that assets must satisfy eligibility and concentration rules.
What to do instead
For each portfolio question, first list the liability's duration and liquidity needs and the eligible-asset constraints, then optimize return within those bounds. Practice identifying which proposed asset is ineligible and why.
Misapplying the deductible and policy limit ordering (e.g., applying the limit before the deductible, or forgetting that the limit caps the insurer's payment after the deductible) when computing indemnity.
The order of operations is a small detail that candidates skip under time pressure, and different policy wordings phrase it differently, so they default to whichever order they saw last.
What to do instead
Adopt a fixed procedure: gross loss, subtract deductible, then cap at the limit, and state the order in writing. Run a set of losses straddling the deductible and the limit to confirm the procedure handles both boundaries.
Estimating incurred-but-not-reported (IBNR) claims by applying a simple development factor to paid claims without checking whether the triangle is stable or whether a change in case reserves or claims handling distorts the pattern.
Candidates learn chain-ladder as a formula and apply it to any triangle, missing that the method assumes stable development and that operational changes invalidate it.
What to do instead
Before applying a development method, inspect the triangle for outliers, changes in case-reserving practice, and diagonal effects. Practice stating the assumption explicitly and naming an alternative (Bornhuetter-Ferguson, expected-loss) when the assumption fails.