Hamilton's Deep Dive Finds a 2018 Loss Still Moving Its Reserves

Hamilton Insurance Group's Bermuda segment took a $16 million casualty reserve charge in the second quarter of 2026, equal to 0.8% of net casualty reserves, after a regularly scheduled deep-dive review found roughly a third of the movement traced to fresh information on a single loss from accident year 2018 (Insurance Business, August 2026). The remainder came from 2022 and later.

The quarter's 8.2-point combined-ratio deterioration was a catastrophe story. The charge is a casualty-tail story, and it is the more durable of the two.

Key Takeaways

• $16 million on a $5 billion gross book of loss reserves, 0.8% of net casualty reserves and about 0.5% of the total net reserve position. CFO Craig Howie called the charge "pretty modest."
• $5 million traced to one accident-year-2018 loss still producing new information eight years on, surfaced by a scheduled deep dive rather than by the triangle's own signal.
• 7.8 points of catastrophe load took the combined ratio to 95.0% from 86.8%, with $45.7 million tied to the Middle East conflict. That load resolves inside the quarter it emerged.
• $13.0 million of Bermuda net unfavorable attritional prior-year development sits alongside the deep-dive charge in the same quarter, widening to $23 million year to date.

A Catastrophe Quarter With a Reserve Footnote

Hamilton's headline moved for reasons that had almost nothing to do with reserving. Gross premiums written rose 16.7% to $831.0 million, net premiums earned grew 14.6% to $586.0 million, and the group still delivered $143.8 million of net income and a 20.6% annualized return on average equity (Hamilton Insurance Group, Q2 2026 earnings release). Underwriting income fell to $29.1 million from $67.5 million, and the entire gap is catastrophe load: $49.9 million of net catastrophe losses, 7.8 points of combined ratio against 1.9 points a year earlier, with $45.7 million of it from the Middle East conflict.

By segment, the International book absorbed the larger share of that war-adjacent exposure at $33.6 million, with its combined ratio widening to 97.0% from 89.3%. Bermuda carried $16.2 million and moved to 93.0% from 84.3%. Net investment income of $141.3 million, $115.5 million of it generated by the Two Sigma Hamilton Fund, more than covered the underwriting deterioration and kept the quarter comfortably profitable.

None of that is a reserving signal. A war-driven property and marine loss reported inside the same accident quarter is current-year, and it says nothing about whether older casualty tails are adequately reserved. The $16 million charge is the part of the quarter that does, which is why it belongs separated from the 8.2-point headline rather than folded into it.

Why a 2018 Loss Is Still Moving in 2026

The charge did not come from a triangle rolling forward and throwing off an unfavorable indication in the normal quarterly re-estimation. It came from a deep dive: a periodic review that revisits assumptions, claim-by-claim detail and case reserve adequacy rather than relying on aggregate development factors. A quarterly roll-forward asks whether the triangle still supports last quarter's picks. A deep dive asks whether those picks survive re-underwriting, individual large claims included.

The contrast with the catastrophe load shows why only one of the two develops. War and terrorism losses in a property or marine treaty trigger on a defined event, with cedant notifications and initial estimates arriving within weeks or months. Reinsurers hold them against accumulation limits and realistic disaster scenarios, so the loss is recognized, reserved and largely closed inside the same reporting period. An excess-casualty claim from 2018 has none of that structure.

The vintage, not the dollar figure, is the actuarial point. A 2018 casualty loss is old enough that a triangle built on pre-2020 settlement and litigation-duration assumptions would treat it as substantially mature, and it still produced $5 million of incremental reserve need. An actuary selecting tail factors on 2015 through 2019 casualty accident years should read that as evidence the cohort's tail was understated rather than as an anomaly to exclude. Swiss Re's Social Inflation Index peaked at 7% in 2023, and nuclear verdicts above $10 million rose 52% in 2024 to a record 135 cases (social inflation and casualty reserve modeling).

Direction, not size, is what places Hamilton in the season. Travelers released $578 million with no casualty pressure cited, and Chubb's $283 million of favorable development sat alongside a North America Commercial combined ratio worsening to 85.4% from 83.5% (Q2 2026 casualty reserve divergence). Everest strengthened North America casualty by close to $200 million, CNA took a $77 million after-tax mass-tort charge, and The Hartford's $46 million of adverse general liability and $26 million of commercial auto development was offset by $110 million of favorable workers' compensation.

Growing Into the Line That Keeps Moving

The deep-dive charge is not the only prior-year movement in the quarter. Bermuda booked $13.0 million of net unfavorable attritional prior-year development, driven by casualty classes and partly offset by property, widening to $23 million year to date. Group net unfavorable prior-year development ran $17.3 million across $686.3 million of net losses and loss adjustment expenses in the first half. The $16 million measured against a $5 billion gross book is one component of a broader casualty drift rather than the whole of it.

Hamilton is expanding into that line while it drifts. Bermuda casualty reinsurance premium grew 29% in the first half of 2026, with management attributing the increase to prior-quarter business and new client opportunities and saying pricing is still supported by economic and social inflation (Hamilton Q2 2026 earnings call, August 13, 2026). Growth into a book where the company's own reserve study keeps finding adverse movement is coherent only if the new business is priced to that movement rather than to last year's picks.

The wider market is at least pricing in the same direction. U.S. casualty rates excluding workers' compensation rose 11% in the second quarter of 2026 while the global commercial composite fell 6% for an eighth consecutive quarterly decline (Marsh Global Insurance Market Index). That divergence and Hamilton's own 2018 vintage point at the same unresolved question: whether the loss picks written across the mid-2010s ever anticipated the litigation environment that followed.