Arch Capital Group closed the second quarter of 2026 with the load-bearing observation an investor needs to internalize: a 2.8% sequential rise in book value per share to $68.04, accomplished on a -8.1% year-over-year decline in net premiums earned, a 230 basis point deterioration in the consolidated combined ratio to 83.5%, and a $1.2 billion repurchase of common stock that single-handedly absorbed more than the entire $1.057 billion of net income. The combined ratio deterioration is mechanical rather than fundamental: roughly 1.4 points came from higher insurance-segment catastrophe activity in the second quarter of 2026 versus a relatively light catastrophe quarter a year ago, and the rest was a mix-driven creep in the insurance excess-and-surplus property book that the company has chosen not to defend with rate cuts. The thesis that 2026 is testing is whether the combined ratio ex-catastrophe and prior-year development, a non-GAAP measure the company discloses specifically to strip out that noise, can stabilize around 82 to 83% while the company shrinks the share count through repurchases. The 82.5% print in the second quarter of 2026 is up only 1.6 points from 80.9% in the second quarter of 2025, and the gap is the cleanest read on whether the underlying underwriting margin is compressing or merely being carried by a soft top line. The next data point that puts pressure on that thesis is the third-quarter earnings release scheduled for late October 2026, which coincides with the peak of the North Atlantic hurricane season and the first full quarter of the renewal cycle on the 1/1/2027 reinsurance treaty book. A quarterly pre-tax catastrophe loss above $300 million in the third quarter, combined with any 1/1/2027 renewal commentary indicating that property catastrophe rates have fallen by more than 5%, would compress the multiple toward 1.2 times book value and put the buyback engine under scrutiny.