One point of loss ratio is 1% of earned premium. For a carrier with $25 billion to $50 billion in earned premium, that is $250 million to $500 million a year, and most portfolios do not track it as a line item.
Smaller books carry the same math at smaller scale. A $2 billion portfolio gives up $20 million for every point it fails to hold. Weather decides some of those points, and daily underwriting and claims decisions decide the rest.
This article shows how to value a point in your own portfolio, where the controllable points leak, what a 2026 survey says about carriers that closed the gap, and how to size the opportunity this week.
What Is One Point of Loss Ratio Worth?
Loss ratio is incurred losses divided by earned premium. One point is 1% of the earned premium base, so its value rises in a straight line with the size of the book.
The $250 million to $500 million in the title is the value of one point at $25 billion to $50 billion in earned premium. Your number is the same calculation on your own base.
The Arithmetic
Earned premium | Value of 1 point | Value of 0.5 point |
|---|---|---|
$500 million | $5 million | $2.5 million |
$1 billion | $10 million | $5 million |
$5 billion | $50 million | $25 million |
$25 billion | $250 million | $125 million |
$50 billion | $500 million | $250 million |
The calculation works at any size. Half a point on a $5 billion book is still $25 million a year.
Industry Scale
Verisk and APCIA estimate that US P&C net earned premium reached $953 billion in 2025, and that the industry combined ratio improved to 92.9% from 96.6% in 2024. One point on that base is about $9.5 billion.
Applied to $953 billion, the 3.7-point improvement is roughly $35 billion. That figure is a simple illustration, because combined ratio also includes expenses, so loss ratio points and combined ratio points move one for one only when expenses hold steady.
Which Points You Control
The same Verisk and APCIA commentary attributes the 2025 result mostly to unusually low catastrophe losses, not to a fundamental shift in industry risk. A good weather year lowers the ratio for everyone and teaches you nothing about your own book.
The points you control come from four places: risk selection, pricing accuracy, claims handling, and the data underneath all three. The next section maps each one.
Why a Recovered Point Compounds
A point recovered on a stable book repeats every year the improvement holds. Half a point on a $2 billion portfolio is $10 million in year one.
If premium stays flat and the gain holds, that is $50 million over five years. A single process fix can pay out for as long as it keeps working.
Where Do Controllable Loss Ratio Points Leak?
Controllable points leak through four levers, and each lever has its own measure. We call this the Controllable Points Map.
Lever | How a point leaks | What to measure |
|---|---|---|
Risk selection | Risks outside appetite are bound after a rushed or partial review | Share of bound risks outside written guidelines |
Pricing accuracy | Similar risks are quoted at different prices | Spread of quoted-to-technical price ratio |
Claims handling | Complex claims reach experienced adjusters late | Time from first notice of loss to assignment |
Data quality | Inputs differ across desks and systems | Fields re-keyed and open information requests at decision |
Risk Selection
A risk that sits outside appetite costs premium adequacy the day it is bound. The cost stays hidden until losses develop, often a year or more later.
Selection errors are more likely when files arrive incomplete or the queue is long. Our guide to the hidden cost of slow underwriting decisions shows where that preparation time goes.
Pricing Accuracy
Underpriced risks erode rate adequacy, and overpriced risks go to competitors. Both show up as a gap between the price a model suggests and the price a quote carries.
Part of that gap is deliberate judgment, and part is noise from inputs and handoffs. Our article on pricing variance separates the two.
Claims Handling
Severity management starts at intake. A complex claim identified in week one is a different file from the same claim identified in week six.
Claims analytics is still early in the industry. In the WTW 2026 Advanced Analytics and AI Survey, only 33% of insurers use advanced analytics for fraud detection and 29% for severity assessment.
Data Quality
Every lever above depends on the same inputs. When policy, claims, and billing data sit in separate systems, each decision-maker works from a different version of the account.
WTW's respondents named data concerns and IT bottlenecks as the main challenges in adopting advanced analytics. Better models do not help when the inputs disagree.
What Does the Evidence Say About Closing the Gap?
The best recent evidence is the WTW survey of 59 P&C insurers in the US and Canada, released in March 2026. It links analytics maturity to underwriting results, with limits you should know before you quote it.
What WTW Found
Insurers with more sophisticated analytics capabilities achieved combined ratios six percentage points lower and premium growth three percentage points higher than slower adopters between 2022 and 2024.
Six points on a $5 billion book is $300 million a year. That is the size of the prize the survey points to, not a forecast for any one carrier.
What It Does Not Prove
The survey shows an association, not a cause. Larger carriers also reported greater use of advanced analytics, according to Carrier Management's coverage, so size, line mix, and capital position all sit alongside the analytics result.
The sample is 59 carriers, and respondents reported their own results. Read the six points as a direction worth testing in your own book.
Where Adoption Stands
Only 16% of polled insurers currently use AI to support human underwriting, and 60% plan to prioritize it between now and 2028, according to Canadian Underwriter's report on the survey.
Most carriers are early. The gap between early movers and the rest is still open.
How Do You Size a Point in Your Own Portfolio?
Multiply earned premium by 1% for each line, then tie the non-catastrophe points to a lever. You need only your finance and actuarial reports.
Pull trailing 12-month earned premium by line of business.
Multiply each line by 1% to get the value of one point.
Pull loss ratio by line and separate catastrophe from non-catastrophe losses.
Tag the non-catastrophe points to a lever on the Controllable Points Map.
Pick the largest line and test one lever on it.
A Worked Example
The figures below are hypothetical and show the method on a $2 billion book.
Line | Earned premium | Value of 1 point | Value of 0.5 point |
|---|---|---|---|
Commercial auto | $600 million | $6 million | $3 million |
General liability | $500 million | $5 million | $2.5 million |
Property | $900 million | $9 million | $4.5 million |
Total | $2 billion | $20 million | $10 million |
If better inputs recover half a point on commercial auto, that is $3 million a year. Half a point across the whole book is $10 million.
Set a Realistic Target
Set targets in tenths of a point, per line, instead of whole points for the book. Small, defensible targets are easier to attribute to a specific change.
Loss ratios develop slowly, especially in long-tail lines. Judge early progress on leading measures such as quote spread and assignment time, then confirm on loss ratio over several quarters.
Decide Which Line to Test First
Rank lines on two measures: the dollar value of one point, and the share of the loss ratio that comes from non-catastrophe losses. The line that scores high on both is the best first test.
In the worked example, property carries the largest value per point at $9 million. Catastrophe losses take a larger share of a property ratio, so the controllable share is smaller than the headline.
Commercial auto and general liability have smaller point values, and their ratios depend more on selection, pricing, and claims handling. Those are the levers a data fix reaches first.
Where Should You Start?
Start with the lever that feeds the others, which is data quality. Selection, pricing, and claims decisions all read from the same inputs.
Start With the Inputs
Pick one line and trace one submission or claim through every system it touches. Count the fields re-keyed and the points where two systems disagree.
That count tells you how much of each lever's leak starts upstream.
Keep Judgment With the Person
Underwriters and adjusters own their decisions. Assistance belongs on preparing and surfacing information, and a person reviews every recommendation before it is final.
That structure keeps accountability clear to brokers, reinsurers, and regulators.
Review Quarterly
Re-run the sizing each quarter on the same line. When one lever improves, the largest remaining leak moves, and the next test follows from it.
Expand to a second line only after the first result holds.
How InsOps Helps
InsOps builds an insurance-trained AI that assists underwriting and claims teams with the data behind each decision. LiLa, our insurance-trained LLM, runs inside your own environment, so PII and PHI never leave controlled infrastructure. A person reviews and validates every field mapping before it is deployed.
Our Integration Gateway connects to Guidewire ClaimCenter, PolicyCenter, BillingCenter, and UnderwritingCenter, so policy, claims, and billing data from your source systems arrives in consistent Guidewire structures without custom engineering for each source.
InsOps migrates legacy data into Guidewire and keeps it flowing in real time.
If you are evaluating how to improve loss ratio performance without adding another manual review step, contact us to talk through what this could look like for your operation.
Frequently Asked Questions
What is a point of loss ratio?
A point of loss ratio is one percentage point of the ratio of incurred losses to earned premium. It equals 1% of earned premium in dollar terms.
Why does one point of loss ratio matter?
Its dollar value scales with your book. At $25 billion to $50 billion in earned premium, one point is $250 million to $500 million a year, and at $2 billion it is $20 million.
How do you calculate what a point is worth for your portfolio?
Take trailing 12-month earned premium for each line and multiply by 1%. Sum the lines for the book, then repeat the calculation on the lines where you plan to test a change.
What moves loss ratio besides weather?
Risk selection, pricing accuracy, claims handling, and data quality all move it. Pricing and claims decisions both read from the same inputs.
Which metrics show whether you are recovering points?
Track the quoted-to-technical price spread, the share of bound risks outside guidelines, time from first notice of loss to assignment, and fields re-keyed per file. Confirm on loss ratio by line over several quarters.
How can AI assist with loss ratio improvement?
AI-assisted tools prepare and standardize the data that underwriters and adjusters work from. A person reviews every recommendation and makes the final call, so accountability stays with your team.
How long does it take to see a loss ratio change?
Leading measures such as quote spread and assignment time respond within a quarter. Loss ratio itself needs several quarters, and longer in long-tail lines, because losses develop over time.
Is a loss ratio point the same as a combined ratio point?
They move one for one only when the expense ratio holds steady. Combined ratio adds expenses to losses, so an expense change moves it without touching loss ratio.