AM Best's May 18, 2026 special report puts the 2025 average approved private passenger auto rate increase at 3.7%, down from 9.7% in 2024. Homeowners came down from 13.5% to 8.3%. Read together, the two lines are the clearest industry-level confirmation yet that the hard market correction has run its course in personal lines.

The deceleration is real. It is also the arithmetic consequence of a $46 billion two-year swing in auto underwriting results, not a signal that loss costs have turned.

Key Takeaways

  • 6.0 percentage points came off the average approved auto rate increase in one year, from 9.7% in 2024 to 3.7% in 2025. Homeowners moved 5.2 points, from 13.5% to 8.3%.
  • $46 billion: auto swung from a $17 billion aggregate underwriting loss in 2023 to a gain of nearly $29 billion in 2025. That is the headroom that made smaller filings supportable.
  • 61.2 was the first-half 2025 direct auto loss ratio against 77.1 in first-half 2023, a 15.9-point improvement. The 3.7% filing average rests on that, not on lower severity.
  • Roughly 20% is what New Jersey rate filings climbed in 2025, against the 3.7% national figure. California, Nevada, New Jersey and New York are the four states AM Best names as outliers.
  • 98.9 is AM Best's projected 2027 auto combined ratio against 96.9 for 2026, with a potential breach of breakeven by 2028.

What the 3.7% Figure Actually Measures

The number is an average of rate changes approved by state insurance departments, drawn from AM Best's State Rate Filings database covering all 50 states and the District of Columbia. It is not what carriers requested, and it is not what has been earned into premium yet.

Behind it sits a two-year reversal in underwriting results large enough that smaller filings became the rational response rather than a concession.

Metric202320242025
Auto aggregate underwriting gain (loss)($17B)~$14B~$29B
Auto H1 direct loss ratio77.167.661.2
Homeowners loss ratio74.8N/A65.6
Homeowners underwriting profitLossLoss>$16B
All-lines P&C combined ratio~102~9996.2

The loss ratio path is where the mechanism is visible. Auto's first-half direct loss ratio fell from 77.1 in 2023 to 67.6 in 2024 to 61.2 in 2025. That 15.9-point improvement is the compound effect of 2023 and 2024 rate increases earning through the premium base while frequency and severity stabilized relative to their pandemic-era spikes.

Homeowners followed the same shape a year later, with the loss ratio improving from 74.8 in 2023 to 65.6 in 2025 and the segment producing an underwriting profit above $16 billion, its first in five years. The all-lines combined ratio of 96.2 was the strongest U.S. property/casualty result in 18 years.

Catch-Up Exhaustion, Not a Loss Cost Reversal

The 9.7% average approved in 2024 was not a trend rate. It carried a catch-up component correcting cumulative underpricing from 2021 through 2023, when premiums lagged repair costs, medical inflation and used vehicle prices.

That distinction is what makes 3.7% harder to work with than it looks. It sits close to an ongoing trend rate, but it still contains residual catch-up in lagging states, so it cannot be treated as either a pure trend indication or a clean competitive signal. AM Best's Chris Draghi framed the underlying condition as durable: "inflation created a new norm to which rates needed to be aligned."

If severity continues at 5% to 7% annually against a 3.7% average approved increase, the margin built during the catch-up years erodes without any deterioration in claim behavior. AM Best's own projections say as much, putting the auto combined ratio at 96.9 in 2026 and 98.9 in 2027.

The national figure also averages over a two-speed market. Rate filings in New Jersey climbed roughly 20% in 2025 alone. California's Proposition 103 prior-approval process runs the state 12 to 18 months behind file-and-use jurisdictions, the same regulatory lag we covered on the property side of California's catastrophe model shift. Nevada and New York carry their own severity and no-fault dynamics.

One more input sits under the deceleration. AM Best's David Blades attributed the homeowners turnaround to "both aggressive rate increases and enhanced pricing sophistication in states that had been generating the most adverse results." A carrier whose segmentation matches rate to risk more accurately reaches adequacy on a smaller filing, which is the pattern behind Progressive's ML pricing edge and the question we raised about algorithmic pricing as a cycle guardrail.

What the Homeowners 8.3% Is Resting On

The homeowners deceleration depends on two conditions that carry no guarantee of repeating, which makes 8.3% a weaker forward indication than the auto figure.

The first is catastrophe experience. 2025 produced no major hurricane landfalls on U.S. soil, and Q1 2026 global insured losses came in at $20 billion, 26% below the 10-year average. Gallagher Re estimates it would take $115 billion to $125 billion in insured catastrophe losses above expected levels to move property pricing materially, but a single active season can push loss ratios in affected states back through breakeven well before that threshold.

The second is reinsurance. AM Best credits a more willing reinsurance market as a contributor to the homeowners result, and Gallagher Re's April 2026 renewal data shows property catastrophe pricing down 15% to 25% for loss-free programs, a dynamic we traced in the reinsurance soft cycle analysis. Passing that relief through to filed rates leaves the net retained position more sensitive to one large event.

There is a third item sitting inside the reported loss ratios. AM Best's February 2026 market segment report noted industry reserve redundancies, with Assured Research estimating more than $20 billion. Favorable prior-year development of that size raises current-year booked results and is available only once, so a rate indication built on booked loss ratios rather than current accident year figures will read as more adequate than it is.

Tariff exposure is the input AM Best itself flags as the downside risk to its stable personal lines outlook. Imported parts and building materials feed severity directly, and a step change there would hit loss ratios a full filing cycle before rate could respond.

Further Reading

Sources