Skip to main content

Underwriting & Combined Ratio

Updated Jun 24, 2026 at 8:22pm

Research Draft High 1,206 words

Underwriting is the core business of a property-and-casualty (P&C) insurer: pricing and selecting risks so that the premiums it collects cover the claims it eventually pays plus the cost of running the business. The combined ratio is the single number analysts use to judge whether that core business is profitable on its own terms — before any investment income is counted. It is the percentage of every premium dollar consumed by claims and expenses. The central tension is that combined ratio deliberately ignores the second engine of an insurer (investment returns on the "float"), so it can paint either a too-harsh or too-rosy picture of the whole enterprise unless read alongside that second engine.

How it's calculated / formed

The combined ratio is the sum of two ratios:

> Combined ratio = Loss ratio + Expense ratio

  • Loss ratio = (incurred losses + loss adjustment expense, LAE) / earned premiums. LAE is the cost of investigating, defending and settling claims, and is conventionally bundled into the loss ratio (per IRMI's definition).
  • Expense ratio = underwriting and operating expenses (commissions, taxes, overhead) / premiums.

A subtlety that trips up newcomers: the two pieces can use different denominators. The loss ratio always uses earned premiums. The expense ratio may be computed on written premiums ("trade basis") or earned premiums ("statutory basis"), per IRMI. The trade-basis number tends to run lower when premiums are growing fast, so comparing two insurers requires checking they're on the same basis. Some presentations also add a third line — the dividend ratio (policyholder dividends, common in mutuals) — to get a "combined ratio after dividends."

Interpretation is mechanical:

  • Below 100% → an underwriting profit. Underwriting margin ≈ 100% − combined ratio.
  • Exactly 100% → break-even on underwriting.
  • Above 100% → an underwriting loss, which must be made up by investment income for the company to earn a net profit.

The metric excludes investment income, realized/unrealized gains, and (in the statutory version) certain corporate items. It is therefore a measure of underwriting skill, not total profitability.

How it's used in practice

The combined ratio is the workhorse comparison tool for P&C insurers, reinsurers and Lloyd's syndicates. Analysts use it to:

  • Compare underwriting discipline across peers and over time. A consistently sub-100 carrier (e.g., Progressive's auto book, or many years at Berkshire's insurers) is demonstrating durable pricing skill, not luck.
  • Decompose the drivers. Splitting into loss ratio vs. expense ratio shows why a number moved — claims inflation and catastrophes hit the loss ratio; commission structure and scale hit the expense ratio. AM Best attributed roughly 8.7 points of the U.S. P&C industry's 2023 combined ratio to catastrophe losses, up from ~7.3 in 2022 — illustrating how a single year's weather can swing the headline number.
  • Read the "float" thesis. This is the connection Warren Buffett popularized. When the combined ratio is below 100, the insurer is being paid to hold policyholders' money (the float) before claims come due — Buffett calls this negative-cost float. Per Morningstar, Berkshire's P&C operations ran a 10-year average combined ratio of about 93% (and a 5-year average near 91%) while growing float to roughly $171B at year-end 2024. Even a combined ratio slightly above 100 can be acceptable if the investment yield on float exceeds the underwriting loss — Buffett's own test (2002 shareholder letter) is whether the cost of float over time is below the company's alternative cost of funds (in 2002 he put Berkshire's cost of float at about 1%).
  • Distinguish accident-year vs. calendar-year combined ratios. Calendar-year includes reserve development from prior years; accident-year strips it out to show how this year's business is pricing. A flattering calendar-year number propped up by releasing old reserves is a classic warning sign.

Adoption, debate & evidence

The combined ratio is the universal, near-uncontested standard for P&C underwriting performance — reported by AM Best, NAIC, NCCI, and every public insurer. There is little debate that it measures what it claims to measure.

The debate is over how much weight to give it. The honest base rate: underwriting losses are common, not rare. AM Best reported the U.S. P&C industry posted a 101.6 combined ratio in 2023 (a ~$21.2B net underwriting loss), improving to a 96.6 combined ratio in 2024 — its first net underwriting profit (~$22.9B) since 2020, driven mainly by premium rate increases catching up to claims rather than by lower catastrophes (cat losses held at ~8.7 points, level with 2023). In other words, the industry aggregate spends much of its history at or above 100 and relies on investment income to earn its return on equity — so judging a single insurer purely on a sub-100 combined ratio, in isolation, misreads how the business model actually works.

Genuine controversies and folklore-vs-measured points:

  • "Below 100 = good insurer" is too simple. A specialty reinsurer might rationally run a higher combined ratio because its long-tail float earns years of investment income; a fast-pay auto insurer needs a lower one because its float is short-lived.
  • Reserve manipulation. Because incurred losses include estimates of future claim payments, management has discretion. Persistent favorable reserve development can flatter calendar-year combined ratios for years before reversing — a documented source of insurer blow-ups.
  • Basis inconsistency (trade vs. statutory expense denominator) makes naive cross-company comparison unreliable.

Strengths & limitations

Strengths: simple, comparable, hard to game over the long run, and it isolates the one thing management most directly controls — pricing and risk selection. Over multi-year windows it is an excellent screen for underwriting quality.

Limitations / the #1 misuse: treating it as a profitability metric. It is not — it excludes investment income, which for long-tail lines is the larger profit engine. Using combined ratio alone to compare a long-tail reinsurer with a short-tail personal-lines carrier is the single most common error. It is also a backward-looking accounting number sensitive to reserve estimates and one-off catastrophes; a single year tells you little. Always pair it with float growth, the cost of float, accident-year vs. calendar-year splits, and the investment yield.

Sources

Confidence: high. Definitions and formula are uncontested and double-sourced (IRMI for loss ratio incl. LAE and trade- vs. statutory-basis expense ratio). Industry figures (101.6 in 2023, 96.6 in 2024, 8.7 cat points both years, ~$21.2B loss / ~$22.9B gain) are AM Best-reported and attributed. Berkshire's 10-year ~93% combined-ratio average and ~$171B float are per Morningstar; the Buffett 2002 cost-of-float test is verified against the primary shareholder letter.