Underwriting Review¶

Risk-assessment review — instantiates Adverse Selection Filtering

Investigates and estimates an individual applicant's hidden risk against the pool's viability target before exposure is accepted, collecting only the evidence the risk decision actually needs.

Underwriting Review is the archetype's per-applicant risk-estimation engine. For each entrant it gathers targeted evidence, estimates the hidden risk that entrant would bring, and decides — accept, decline, or accept on adjusted terms — against what the pool can actually bear. Its distinctive move among its siblings is that it quantifies the individual: where a minimum standard applies a bright line and a prequalification workflow verifies documents, underwriting builds a reasoned estimate of one applicant's hidden type and weighs it against the pool's solvency target. Because that estimation tempts an assessor to gather everything, it carries a second discipline: collecting only the evidence the risk decision genuinely needs, and no more.

Example¶

Individual life insurance is the classic case. An applicant knows their own health; the insurer does not, and a uniform "guaranteed issue" offer would attract disproportionately sick applicants and threaten the mortality pool's solvency. Underwriting review is how the insurer closes that gap: it gathers targeted evidence — the application, medical history, sometimes an exam or fluid test — estimates the applicant's mortality risk, and decides accept, decline, or rate-up (accept at adjusted terms), always against what the pool can absorb. The review is what keeps the pool viable applicant by applicant. And it is bounded: the insurer collects what bears on mortality risk, not the applicant's entire life, because over-collection is both intrusive and legally fraught.

The output is not a price and not a bright line — it is a defensible estimate of one entrant's hidden risk, tied to an explicit judgment about whether the pool can carry it.

How it works¶

Name the hidden attribute. The review defines the specific risk it is estimating — mortality, default, loss — so evidence-gathering stays targeted.
Gather targeted evidence. It collects the documents, records, and checks that bear on that attribute, and only those.
Estimate and decide against the target. It forms a risk estimate and compares it to the pool's viability or solvency margin, yielding accept, decline, or accept-on-adjusted-terms.
Minimize what's held. Data collection is scoped to the risk question, so the review does not become a dragnet.

Tuning parameters¶

Evidence depth — simplified versus full underwriting; deeper is more accurate but slower, costlier, and more intrusive.
Decline vs. adjust threshold — how bad an estimate must be to reject rather than accept on adjusted terms.
Viability margin — how much solvency cushion the pool insists on; tighter accepts fewer marginal risks.
Data-minimization scope — how strictly collection is limited to the risk question; the privacy-versus-accuracy dial.
Automation level — how much is rules-based versus assessor judgment, trading consistency against nuance.

When it helps, and when it misleads¶

Its strength is precision at the gate: it catches the high-risk entrant a flat offer would attract and a bright line would miss, and it can accept-on-adjusted-terms an applicant a cruder filter would simply reject. It is the right tool when the cost of a single bad acceptance is high enough to justify individual assessment against a solvency target.

Its central weakness is that estimation invites over-reach: the review can slide from risk-relevant evidence into intrusive or proxy-discriminatory data, and it is expensive to run per applicant.^[1] The classic misuse is running the "risk assessment" backwards — to manufacture a justification for declining an applicant already unwanted — or collecting far more data than the risk question requires. The discipline that keeps it honest is to tie every data item to the risk it informs, hold decisions to the stated viability target, and keep a real appeal path for the applicant the estimate gets wrong.

How it implements the components¶

Underwriting Review realizes the per-applicant estimation-and-solvency side of the archetype:

hidden_type_selection_model — it builds a reasoned, per-applicant estimate of the hidden risk attribute; this is its core output.
pool_viability_target — every accept/decline decision is made against what the pool can absorb (its solvency or loss target).
privacy_and_data_minimization_rule — it collects only risk-relevant evidence, keeping the review from becoming a dragnet.

It estimates risk but does not price it — that's Risk-Adjusted Pricing, which consumes its estimate — nor does it recalibrate base rates after the fact (Claims or Outcome Experience Rating) or apply the bright-line bar (Minimum Eligibility Standard).

Instantiates: Adverse Selection Filtering — it estimates each applicant's hidden risk against the pool's viability before exposure is accepted.
Sibling mechanisms: Risk-Adjusted Pricing · Prequalification Process · Claims or Outcome Experience Rating · Deductible or Copay Schedule · Minimum Eligibility Standard · Probationary Entry · Quality Certification Requirement · Risk Tier Assignment · Seller Rating or Quality Grading · Waiting Period · Warranty or Guarantee Requirement

Notes¶

Underwriting is easily conflated with prequalification, but they do different work: prequalification gathers and verifies an applicant's documents, while underwriting quantifies the risk those documents describe and decides against a solvency target. A pool often runs both — the workflow assembles the file, the review judges it.

References¶

[1] Underwriting addresses adverse selection — the pre-contract problem of who chooses to enter — which is distinct from moral hazard, the post-contract problem of how the insured behaves once covered. The two often coexist but call for different interventions; underwriting is a gate on entry, not a control on later behavior. ↩