Chief Financial Officer / United States
Decisions priced for the day we might have to sell
- General-account invested assets, December 2025
- $46.8bn
- Backing fixed indexed, multi-year fixed and income annuities
- Risk-based capital ratio, December 2025
- 415%
- Company action level basis; 400 per cent a year earlier
- Annuity sales in 2025
- $6.1bn
- Through marketing organisations, banks and broker-dealers
Decisions & consequences
01Taking a minority investor's capital without handing over the portfolio
In early 2022 Wrenfield needed capital. Annuity sales were growing faster than retained earnings could support, our risk-based capital ratio was drifting towards 400 per cent, and a share issue at our valuation then would have diluted holders heavily. Thresholm Capital Partners, a fictional alternative asset manager, proposed buying 9.9 per cent of our shares for $430 million and managing $12 billion of our general account, about 28 per cent of it, in private credit it would originate. It projected around 60 basis points of extra yield on that money after fees.
Our chief executive and the lead independent director favoured the full proposal. The capital was real and the yield attractive. I recommended taking the capital but limiting the mandate to $5 billion in three asset classes we could value ourselves, with every position priced by a valuation firm our Audit Committee would choose, fees reported through every layer of each structure and a right to end the mandate after three years without penalty. My concern was not Thresholm's skill. A shareholder paid by asset volume would be originating loans for a balance sheet whose capital it also wanted to grow, and our board would see the yield long before it saw an exit price.
Thresholm accepted narrower terms at a lower price: $410 million for 9.6 per cent. We gave up about $20 million of equity proceeds and, on Thresholm's own projections, roughly $40 million a year of investment income on the $7 billion we kept. The stake stayed below the level at which our state's insurance holding company law presumes control, and our regulator reviewed the investment management agreement before it took effect.
By December 2025 Thresholm managed $5.2 billion. The mandate earned 1.4 percentage points a year over comparable public credit after fees, which is good work. In 2024 the two valuations differed by more than three points on six of its 61 positions, and one facility secured on point-of-sale consumer loans was later written down by $38 million. The Finance and Investment Committee now receives an estimated exit price for every privately priced asset class each quarter.
I would negotiate access to loan-level data in the first agreement. Getting it took us another year.
Letting $2.3 billion of annuities leave rather than paying to keep them
Between 2017 and 2019 Wrenfield sold large volumes of five-year fixed annuities and five- and seven-year indexed annuities, crediting between 2.6 and 3.4 per cent. When rates rose in 2022, competitors began offering new five-year contracts above 5 per cent. Across 2023 and 2024, $3.7 billion of our policies would reach the end of their surrender-charge periods, after which owners could leave without penalty. The assets backing them had been bought at yields near 3.8 per cent and were worth well below book value.
Our chief distribution officer proposed renewal rates of 4.9 per cent to keep most of the money, arguing that the marketing organisations selling our products would remember which carriers kept faith with their clients. He was right about the memory. I argued that 4.9 per cent could be paid only by taking credit risk we would not otherwise accept or by subsidising renewals from new business. I recommended 4.1 per cent, the rate our existing assets could honestly support, and planning for most of the money to go.
The plan began a year early. We arranged $1.5 billion of secured borrowing capacity against agency mortgage securities, shortened new purchases so that maturities fell in the expected outflow quarters, and modelled surrenders by policy cohort and by distributor. In the event $2.3 billion left over eighteen months, against our central estimate of $1.9 billion. We drew $1.2 billion of the borrowing and repaid it by December 2024, and sold $1.1 billion of bonds, realising $160 million of losses that mostly went into the interest maintenance reserve under statutory rules.
The commercial cost was visible. Two large marketing organisations moved us down their recommended lists, and 2024 sales fell 12 per cent to $5.45 billion. The spread on retained business held near 2.2 per cent, and we did not reach for riskier assets to defend it. Our product committee now limits the share of any year's sales whose surrender charges can expire in the same quarter.
I would have met the largest marketing organisations six months earlier, with our renewal arithmetic in hand. They heard the rate before they heard the reason.
Selling office mortgage loans before the appraisals arrived
At the start of 2023 Wrenfield held $6.1 billion of commercial mortgage loans, $1.7 billion of them secured on office buildings, mostly in suburban parks and secondary downtowns. Debt-service coverage still looked comfortable because leases signed before 2020 were still paying. Occupancy data told another story: our borrowers' buildings averaged about 60 per cent physical use, and nine loans with $640 million outstanding faced the expiry of more than a third of their leased space within three years.
Our head of real estate investments wanted to wait. Most loans were current, he argued, a life insurer could hold to maturity, and buyers of office loans in 2023 were demanding distressed prices. The chief risk officer and I asked for a loan-by-loan review using our own forward rents rather than the last appraisal. It identified eleven loans whose collateral would probably be worth less than the balance before maturity. I recommended selling the seven weakest and restructuring the other four with borrower paydowns and cash sweeps.
The board's Finance and Investment Committee debated the proposal over two meetings before approving it. The cost was immediate. We sold $420 million of loans in the fourth quarter at an average of 71 cents on the dollar, raised our credit loss allowance by $43 million and recognised a $165 million pre-tax charge. Statutory impairments and higher capital charges on the remaining loans took 18 points off our capital ratio. We cut the insurer's dividend to the holding company by $100 million and paused buybacks for three quarters. Some analysts called it capitulation.
By December 2025 office loans stood at $880 million. Two of the restructured loans were foreclosed in 2025 with losses of $21 million. Comparable loans that other holders sold during 2024 cleared at around 60 cents, which suggests that waiting would have cost more. Every commercial mortgage now carries a forward-rent value beside its appraisal in our quarterly risk report.
I would have acted a year sooner. The occupancy figures were available in 2022, and I treated them as noise because borrowers were still paying.
Reinsuring a closed care block at a price the actuaries disliked
Wrenfield stopped selling long-term care insurance in 2009, but we still carried 41,000 policies written between 1998 and 2009, with statutory reserves of $1.4 billion. Claims had run higher and lapses lower than the original pricing assumed, as they did across the industry. Since 2012 we had sought premium increases in 38 states; approvals varied widely by state and had raised premiums by about 60 per cent cumulatively. Each annual assumption review still moved reserves, and rating agencies charged the block heavily in their capital assessments.
In 2024 a fictional reinsurer, Sallowfield Re, offered to coinsure the whole block for the reserves plus $240 million. Our chief actuary opposed the price. He believed further rate approvals and better claims management would close most of the gap, and two directors agreed that we would be paying to remove a problem we were already solving. They were partly right: our own central case valued the block about $90 million more favourably than Sallowfield's price implied.
I recommended accepting. The block's outcome depended on morbidity, interest rates and state regulators' decisions for another thirty years, none of which a spread business is built to manage, and its capital charge was holding back the growth in income annuities the board wanted. I asked that the board paper carry the chief actuary's case in his own words, and that we test the reinsurer's security as hard as its price: the trust, its investment guidelines and our right to recapture the block if Sallowfield were downgraded.
The transaction closed in July 2025. The $240 million came out of statutory surplus at closing, but the release of required capital more than offset it, and our capital ratio ended 2025 at 415 per cent against 400 a year earlier. The rate filing programme ended for Wrenfield, and policyholders stayed with the same servicing team under the agreement. One rating agency moved our outlook to positive in October 2025.
I would have run a competitive process. By the time we negotiated, Sallowfield was the only serious bidder, and a second might have narrowed that $90 million gap.