Where a life insurer's strategy is decided before the sale

Every traditional savings policy we sell carries a pricing rate that we owe the customer for decades, and that rate is fixed on the day of sale. My work is to make sure that day's decision is taken with the following forty years in view. Since 2023 our product committee approves a savings product only after the asset-liability committee has shown that the investment team can buy assets at or above its pricing rate, with margin for expenses and adverse scenarios, at the yields available that quarter rather than at our portfolio's historic book yield. We moved most savings sales to participating products, which combine a lower promised rate with a bonus that depends on investment results, and to annuities designed around retirement income rather than early cash value. The average pricing rate of our in-force traditional reserves has fallen from 3.62 to 3.21 per cent since 2022. It will keep falling slowly, because policies written a decade ago will stay with us for decades more.

Banks remain the largest source of savings premium for China's life insurers, and for a mid-size company the terms are hard: the bank owns the customer, controls the shelf and has many insurers to choose from. The regulator's requirement, applied to the bank channel from 2023, that commissions actually paid match the expense assumptions filed for each product removed one way of competing for shelf space, but it did not make thin business profitable. I built a partner-by-partner view of what each relationship earned: new business value after commission, persistency at thirteen and twenty-five months, the solvency capital each product consumed and the share of sales that were regular rather than single premium. Of thirty-six partners at the end of 2023, eleven earned more than their cost of capital. We kept those and two city commercial banks with growing regular-premium sales, and ended the rest in 2024. Our bank managers now argue about persistency instead of volume.

I led the strategy that took Jiangheng into elderly care: the purchase of land in Wuhan in 2018, the opening of Songhe Court in October 2022 with 780 apartments including 160 nursing places, and contracts with eleven partner operators in eight cities that reserve 1,350 apartments for our policyholders. The commercial logic is that someone saving for retirement values a credible place to receive care, and that an insurer with long-dated liabilities can hold long-lived property. The risk is that rights to a place are sold faster than places exist. I designed the residence-right register that the board's risk management committee has reviewed each quarter since September 2023: letters issued, projected applications by year and age band, apartments available and expected turnover. I also wrote the rule that new letters for a community stop once projected applications within twenty years exceed the places we expect to have, whatever the sales plan says.

The second phase of China's risk-oriented solvency system took effect in 2022, with stricter rules on what counts as core capital and a more demanding treatment of some long-dated and indirect investments. For a company buying land for a retirement community while holding long-dated savings liabilities, the change mattered. I run the three-year capital plan with our chief actuary and chief financial officer. It projects core and comprehensive solvency under interest-rate, equity-market and property scenarios, sets the capital each strategic initiative must carry before the board sees it, and identifies when shareholders will need to be asked for more. We issued RMB 5 billion of capital supplementary bonds in 2021 and raised RMB 4.5 billion of new equity from existing shareholders in 2024. At the end of 2025 our comprehensive solvency ratio was 192 per cent and our core ratio 121 per cent. Each proposed community is now assessed for its capital charge as carefully as for its return.

Moving a savings insurer towards protection means building capabilities an agency force selling deposits never needed. We rebuilt our medical and critical-illness range around customers over fifty, with underwriting for common chronic conditions and claims handled through agreements with hospitals in our main cities. We also took part in city-sponsored supplementary medical schemes as a co-insurer, which taught us a great deal about claims among older people at modest premium. Our largest learning came from public long-term-care insurance. In the state pilot scheme, participating cities appoint commercial insurers to assess applicants and administer benefits; we have done so for one Hubei city since 2020 and one Hunan city since 2024, with 210 assessors and care coordinators. The fees are small. The data on how care needs develop with age are not, and they now inform both our product design and our residence-right projections.

My first thirteen years were spent on risks that arrive all at once. At Lanbei Property and Casualty I priced crop insurance for county programmes in Heilongjiang, Jilin and Inner Mongolia, built a yield database from township records and bought the reinsurance for the crop and property books. The work taught me to think about correlation before volume: a thousand farms in one river basin behave as one exposure in a flood year, however many policies they represent. Subsidised crop insurance in China depends on provincial programmes, premium subsidies and insurers able to carry bad seasons, and its economics turn on how losses are pooled, reinsured and reserved across years. I apply the same discipline at a life insurer, where the correlated risk is interest rates rather than rain. It is also what I would bring to the board of a company whose earnings depend on harvests, water or weather.