SIGI vs. WRB Q3 2026: Why EPS Went Opposite Ways for Two Commercial Insurers

Selective Insurance (SIGI) and W. R. Berkley (WRB) both sell insurance mainly to businesses, and both posted a better combined ratio in Q3 2026. The combined ratio is claims plus expenses per dollar of premium earned; under 100% means the insurance itself made a profit before any investment income. The headline earnings moved in opposite directions, though: Selective's diluted EPS rose 14.1% to $2.11, while Berkley's fell 10.9% to $1.14. The gap comes from the year-ago quarters. Selective is being compared with a quarter that carried a $40 million reserve charge, and Berkley with one that carried a $79.0 million investment gain. On the underwriting itself, Berkley is far ahead. It ran a 90.8% combined ratio against Selective's 96.8%, and its book is still growing (+1.3%), while Selective is cutting back (-7.0%) to repair margins.

At a glance

• Underwriting margin: about 9 cents vs about 3 cents per premium dollar. Berkley's 90.8% combined ratio leaves roughly 9.2 cents of underwriting profit per dollar of premium; Selective's 96.8% leaves about 3.2 cents, even after a 1.8-point improvement.
• Premium growth: +1.3% vs -7.0%. Berkley is still growing, slowly. Selective is shrinking on purpose: new business fell 13% and it is letting underpriced accounts leave.
• Both lost book value to bond prices. Book value per share rose only 1.0% at Selective and 0.3% at Berkley, because higher interest rates cut the market value of the bonds both companies hold.

Side by side

Berkley writes roughly three times Selective's premiums, so the dollar figures aren't comparable; the ratios and growth rates are. Selective's figures come from its 10-Q. Berkley's come from its October 19 earnings release, since its 10-Q had not been filed when we wrote it up.

Why EPS went opposite ways

Both companies had small investment losses this quarter: $19.4 million pre-tax at Selective and $18.3 million at Berkley. What differs is the year they are measured against.

• Selective's year-ago quarter was weak. In Q3 2025 it added $40 million to reserves for older commercial and personal auto claims (reserves are the money an insurer sets aside for claims not yet paid; raising them is a charge against profit). That cost 3.3 points of loss ratio. This quarter there was "no prior year casualty reserve development in any segment or line of business." Removing that charge alone is worth more than the whole 1.8-point improvement in the combined ratio.
• Berkley's year-ago quarter was flattered. It booked a $79.0 million investment gain in Q3 2025; this year it took an $18.3 million loss, a $97.3 million swing. Excluding those gains and losses, operating income rose 3.7% and operating EPS rose 6.4%.

Operating EPS, which excludes investment gains and losses at both companies, gives a fairer comparison. On that basis Selective (+35%) did better than Berkley (+6.4%), but mostly because Selective was recovering from a bad quarter and Berkley was already running well. Buybacks helped both: diluted shares fell 1.9% at Selective and 2.2% at Berkley.

Underwriting: a turnaround vs a steady book

Strip out catastrophes and changes to old-year reserve estimates, and you get the underlying combined ratio: how current prices compare with current claims. Selective's is about 91.4% (93.2% a year ago, our calculation from its 10-Q). Berkley's accident-year combined ratio excluding catastrophes is 88.1% (88.4%). Both improved, but Berkley's starts about 3 points better, and a further 3.5 points separate their expense ratios (32.1% vs 28.6%).

The expense gap is partly a result of Selective's own choices. Its filing says "lower NPE has put pressure on our underwriting expense ratio": as it writes less business, fixed costs are spread over fewer premium dollars, and the ratio rose 1.5 points. Berkley's expense ratio barely moved (+0.1 point) on premiums earned that grew 1.8%.

Catastrophes also cost Selective more relative to its size: 5.4 points of its loss ratio versus 2.7 at Berkley. Selective's filing points to winter storms and thunderstorms in its states. Berkley's catastrophe losses rose less (to $85.9 million from $78.5 million) on a much larger book.

Inside each company, the problem lines differ:

• Selective still loses money in general liability (106.6% combined ratio, flat, despite 7.9% renewal price increases; the filing blames "social inflation," meaning larger jury awards and legal costs), commercial auto (102.0%, down from 117.6%) and personal lines (103.2%). Its excess and surplus lines (harder-to-place risks), until now its most profitable segment, worsened to 86.5% from 76.2%.
• Berkley's main Insurance segment improved to 91.7% from 92.3%. The pressure is in its Reinsurance & Monoline Excess segment: premiums fell 11.6% (casualty reinsurance -25.2%) and its combined ratio rose to 83.4% from 81.1%, though it remains the more profitable of the two segments.

Pricing and growth

Both are getting smaller price increases than a year ago. Selective's renewal price increases slowed to 5.6% from 9.6%; Berkley's rate increases (excluding workers' compensation) were about 4.0%, down from 7.6%. They are responding differently. Selective is giving up volume: its filing says lower premiums "reflect reduced new business in a competitive environment and our granular pricing actions to drive lower retention on underperforming business." Berkley's Insurance segment still grew net premiums written 3.2%, led by professional liability (+5.0%) and auto (+5.2%); the shrinking reinsurance segment held total growth to 1.3%.

Interest rates: the same tailwind, the same hit to book value

Higher bond yields help both companies' income and hurt both companies' balance sheets.

• Investment income is growing at both. Selective's after-tax net investment income rose 16% to $127.3 million, and it raised its full-year target to $490 million. Berkley's core portfolio income rose 12.0% to $369.1 million as maturing bonds are reinvested at higher yields.
• Book value is being held down at both. Selective's pre-tax unrealized bond losses rose $433 million since December. Excluding them, its adjusted book value per share rose 11% to $62.00, against the reported 1.0%. Berkley's after-tax unrealized losses widened from $317 million at June 30 to $928 million at Sept. 30, which cut book value per share 2.4% in the quarter. These are paper losses on bonds held to maturity, and they shrink as the bonds approach maturity, but they lower reported equity today.

Investment income matters more to Selective's returns. Its report attributes 15.2 points of its 15.1% ROE to investments and 3.5 points to insurance operations. At a 96.8% combined ratio, underwriting adds little. Berkley's 90.8% means its insurance operations carry a much larger share of the profit on their own.

Takeaway: The EPS headlines (Selective +14%, Berkley -11%) point the wrong way on underwriting quality. Selective's gain is mostly a weak year-ago quarter dropping out of the comparison. It still runs a 96.8% combined ratio, it is shrinking premiums 7% to get there, and investment income produces most of its return. Berkley's decline is mostly a year-ago investment gain that didn't repeat. Its underwriting held at 90.8%, and the release's own figures imply roughly no help from reserve releases, and its premiums are still growing. Per dollar of premium, Berkley's insurance business earns roughly three times as much underwriting profit.

What to watch next

• Reserves at both. Selective has had three quarters with no casualty reserve additions; year-end reviews are when its charges have appeared before. Berkley's 10-Q (expected around early November) will give its exact prior-year development by segment, including other liability and auto, where the first half showed adverse development.
• Selective's expense ratio. If premiums keep falling 5-8% while price increases slow, the expense ratio will keep rising. Selective's full-year combined ratio guidance of 96.5%-97.5% needs a fourth quarter at or below about 96% (nine months: 97.7%).
• Berkley's reinsurance mix. Its most profitable segment is shrinking fastest, and that segment's underlying loss ratio rose this quarter.

Sources: our SIGI Q3 2026 analysis (from Selective's 10-Q) and our WRB Q3 2026 analysis (from Berkley's earnings release).