Home Product Pricing Blog About Team Contact
Home/Blog

Underwriting & Actuarial Insights

Technical analysis and practical perspectives from P&C insurance practitioners.

Submission scoring survivorship bias
Underwriting

Why Submission Scoring Fails When You Train on Bound Business Only

A model trained only on policies you chose to bind cannot tell you which submissions you should have declined. The data you never accepted is precisely the data that would teach the model about adverse selection.

Sarah Kim · April 8, 2025
ISO rating integration lag
Actuarial

ISO Circular Updates Take 4 Months to Reach Your Rating Tables. Here's Why That's a Problem.

The lag between ISO advisory loss cost filings and carrier implementation creates systematic mispricing. During that window, you are pricing risk against an actuarial baseline you know is wrong.

Amara Okafor · March 24, 2025
FNOL data claim severity
Claims

What FNOL Data Actually Tells You About Final Claim Severity

First Notice of Loss data contains early predictors of final claim cost that most claims systems ignore. Attorney representation at FNOL, for instance, is the single strongest predictor of final indemnity in commercial auto.

Daniel Reyes · March 10, 2025
LDF vs Bornhuetter-Ferguson
Actuarial

Loss Development Factors vs. Bornhuetter-Ferguson: When to Trust Your Own Experience

The choice between LDF and BF is a credibility question. Actuaries who default to LDF on every line without examining their development pattern stability are making a systematic error, usually toward overconfidence in thin data.

Sarah Kim · February 28, 2025
Adverse selection patterns
Underwriting

Adverse Selection Isn't Random — The Patterns Are Detectable Before You Bind

Geographic CAT concentration, occupancy class deterioration, and suppressed loss history are all visible in ACORD data weeks before they show up in your loss ratio. The question is whether anyone is looking.

Amara Okafor · February 14, 2025
Applied Epic vs Duck Creek
Technology

Applied Epic vs. Duck Creek: What the API Difference Means for Underwriting Automation

Applied Epic and Duck Creek have fundamentally different integration architectures. The difference matters for anyone building underwriting automation on top of either platform — and the mistake costs six to twelve months of rework.

Daniel Reyes · January 30, 2025
CGL risk indicators
Underwriting

CGL Risk Indicators Most Underwriters Check Last (and Should Check First)

The order in which an underwriter reviews a CGL submission produces systematic anchoring bias. Reviewing risk characteristics before the premium indication consistently produces better outcomes than the reverse.

Amara Okafor · January 15, 2025
MGA fronting carrier relationship
Industry

What Fronting Carriers Actually Want from MGAs: A Straightforward Account

Fronting carriers are not primarily worried about whether you can find business. They are worried about whether the business you find will generate losses they will have to pay while your MGA fee income flows regardless of loss experience.

Amara Okafor · December 18, 2024
ACORD 125 data quality
Technology

ACORD 125 Data Quality Problems and What They Do to Your Risk Model

ACORD 125 data quality problems are concentrated in five specific fields: SIC code, years in business, annual revenues, prior loss history, and description of operations — each one a potential source of systematic scoring error.

Daniel Reyes · December 5, 2024
RiskVert Seed Round
Company News

RiskVert Closes Seed Round to Accelerate P&C Underwriting AI Platform

RiskVert has closed its Seed Round. The funding goes toward model engineering, PAS integrations, and the ISO rating monitoring infrastructure that our pilot customers identified as their highest near-term priority.

Amara Okafor · November 12, 2024
Combined ratio lagging indicator
Actuarial

The Combined Ratio Is a Lagging Indicator. Here's What to Watch Instead.

By the time the combined ratio signals deterioration, the adverse selection is already two renewal cycles deep in your book. Risk score distribution, acceptance rate by tier, and reserve development velocity all move earlier.

Sarah Kim · October 22, 2024
Run-off claims triage
Claims

Run-Off Books Don't Need Full Actuarial Teams — But They Do Need Triage

A book in run-off concentrates the most complex claims while the premium base shrinks. Triage-based allocation separates the 15% requiring specialist attention from the 55% that can resolve through standard channels.

Daniel Reyes · September 15, 2024
Workers comp frequency vs severity
Actuarial

How Workers Compensation Frequency Trends Differ from Severity Trends

Workers comp frequency and severity trend independently and require separate actuarial treatment. Carriers that apply a single combined trend factor systematically misprice the line whenever the two components diverge — which is often.

Sarah Kim · August 7, 2024