Three decisions and the members who paid for them

Withdrawing homeowners cover from seven Colorado counties while keeping the farm book

Ostervale entered Colorado in 2012, mainly through farm agencies on the eastern plains. By 2022 homeowners business in seven Front Range counties made up most of the state's 21,400 homeowners policies and US$46 million of premium. Hail had produced a five-year loss ratio of 91 per cent in that book before expenses, and our actuaries indicated a 41 per cent increase. A smaller increase approved the previous year had already cost us some of the better-maintained homes.

The stakes were concentrated. That book produced less than nine per cent of our premium but contributed US$64 million to our modelled one-in-100-year convective-storm loss.

In early 2023 I took three paths to the board: a three-year repricing with roof-age payment schedules, full withdrawal from Colorado, or a county-level withdrawal that kept farmowners and eastern homeowners. Our chief underwriting officer favoured repricing, believing the market would follow us upward. Two directors questioned whether a mutual should leave members at all. I supported the county option, because repricing would have taken at least three years to earn through while the exposure stayed on our balance sheet.

Non-renewals ran from March 2024 to August 2025, with notice longer than the state minimum. Each affected agency received its list eight months ahead. We briefed the state regulator before the first notice went out.

We removed 13,800 policies and US$29 million of premium. The modelled one-in-100 loss fell from US$212 million to US$176 million, and the 2024 reinsurance renewal cost US$3.1 million less than a like-for-like quote. Our expense ratio rose 0.6 points as fixed costs spread across less premium, and one agency moved a US$2.4 million farm book to a competitor. The board now reviews every state against written criteria each autumn. I would test the roof-age schedule in one county for a full season before deciding; that result might have moved the boundary we drew.

Accepting a higher catastrophe retention and dropping the aggregate cover in 2023

Our January 2023 catastrophe renewal arrived in the hardest property reinsurance market I had seen. The expiring programme retained US$30 million per storm, covered losses up to US$260 million and included an aggregate cover that responded after US$90 million of annual catastrophe losses. Reinsurers proposed a US$40 million retention, a risk-adjusted per-occurrence price 37 per cent higher, and an aggregate cover at a higher attachment costing US$14 million, more than double the expiring premium.

Surplus stood at US$402 million. A second weak year behind a thinner programme could pressure our financial-strength rating, on which several agencies rely.

The chief financial officer wanted to buy the aggregate at the quoted price, since frequent mid-sized storms had hurt us most. My modelling put its expected recovery near US$6 million, less than half its cost. I recommended accepting the US$40 million retention, keeping the full per-occurrence limit and dropping the aggregate, while committing the saving to roof-age and deductible changes. The risk committee chair asked us to record which surplus level would reopen the decision. We set it at US$340 million.

We bought the programme without the aggregate, filed deductible changes in all six states and reported retained catastrophe losses to the finance and risk committee every month through the storm season.

The season was severe. Retained catastrophe losses reached US$148 million, and the dropped aggregate would have recovered US$22 million, US$8 million more than it cost. Surplus fell to US$356 million, above the trigger, but only just. In 2024 and 2025, as milder seasons arrived and rate and deductible actions earned through, the combined ratio improved to 103.1 and then 98.6 per cent. The board now approves a three-year reinsurance strategy with explicit surplus triggers rather than a single-year purchase.

I presented the expected value more confidently than its uncertainty deserved. Next time I would show directors the particular seasons in which the cover pays, not only its long-run average.

Issuing a surplus note instead of reorganising or affiliating with a larger mutual

After 2023 the board asked management to restore surplus above US$400 million within two years without relying on a mild season. Two proposals arrived in early 2024. A larger mutual insurer from another region offered an affiliation with shared reinsurance and services, and some directors wanted to study a mutual holding company reorganisation that could eventually allow outside equity.

Each route would change who controlled Ostervale's future. Agents and members had been told for decades that decisions were made in Salina by a board the members elected.

Our chief executive initially preferred the affiliation, which promised about US$5 million a year of reinsurance and service savings. I argued that those savings depended on pooling terms not yet agreed, and that a reorganisation addressed a long-term equity need we did not have. I recommended a US$60 million surplus note, placed privately at 7.6 per cent, with every payment of interest or principal subject to the prior approval of our domiciliary regulator. The note ranks behind policyholder claims and counts towards statutory surplus. The board accepted the recommendation and closed the affiliation discussion, which disappointed two directors and the other mutual's leadership.

We worked with outside counsel and the state insurance department in Kansas on the note terms, prepared a ten-year repayment plan and explained the decision to our agency advisory council before any announcement.

The note closed in June 2024. Surplus reached US$398 million at the end of 2024 and US$430 million at the end of 2025, helped by the note, two better seasons and the earlier underwriting actions. Interest costs about US$4.6 million a year, and whatever the affiliation might have saved was given up. The board adopted a written capital policy with a surplus range and named actions at each boundary.

I would commission an independent review of the affiliation terms before management formed any preference. Our analysis was careful, but it was ours, and the directors who favoured affiliation deserved advice that neither side had shaped.