An insurer collects premiums first and pays claims later. The money it holds in between is called float, and it earns investment income. The cost of the float is the underwriting result: if claims and expenses exceed premiums, that is the price paid for holding the money. Whether rising or falling interest rates help an insurer or a pension plan depends on whether its assets or its liabilities are more sensitive to rates. Leveraged hedges can add a cash-shortage risk, which 2022 exposed.
Why it matters: Insurance profit comes from two sources: underwriting and investing the float.
Summary: Float is the money an insurer holds between collecting premiums and paying claims; it earns investment income, and its cost is the underwriting result. Whether rising or falling rates help an insurer or pension plan depends on whether assets or liabilities carry more dollar duration. Leveraged hedges add a liquidity risk that 2022 exposed.
- Cost of float = underwriting loss ÷ float: losing $3 million on $150 million of float is borrowing at 2%.
- With $1,100m of duration-6 assets against $1,000m of duration-10 liabilities, a 1-point rate rise lifts surplus by $34m; the flip point is asset duration 9.09.
- In 2022 the 100 largest US corporate pension plans saw liabilities fall 28.2% and funded status rise from 96.3% to 99.3%.
- The UK’s September 2022 gilt crisis was a collateral (liquidity) problem, not a solvency one.
- Defined benefit plans leave rate and longevity risk with the sponsor; defined contribution plans leave it with the employee.
Because an insurer collects premiums today but pays claims later — sometimes decades later for life insurance — it holds a large pool of collected-but-not-yet-paid-out cash called float. This float is invested (introduced conceptually in The Practitioner’s Codex Part 5’s portfolio theory, and Part 3: Sovereign Debt & Public Finance‘s bond mathematics), and the investment income earned on it is frequently as important to an insurer’s overall profitability as the underwriting result itself — sometimes more so, particularly for life insurers with very long-duration liabilities.
Float is borrowed money, and its interest rate is the underwriting result. The cost of float = underwriting loss ÷ average float. An insurer that loses $3 million on underwriting while holding $150 million of float has borrowed at $3m ÷ $150m = 2%; if Treasury bills yield more than 2%, the float still beats borrowing in the market. The insurer in Part 1.3: The Actuarial Function — Pricing Risk Before It Happens‘s combined-ratio example made a $2 million underwriting profit on the same $150 million, a cost of −$2m ÷ $150m = −1.33%: it was paid to hold other people’s money. Berkshire Hathaway’s 2024 shareholder letter puts the idea in numbers, reporting $32 billion of after-tax underwriting profit over the past two decades while float grew from $46 billion to $171 billion: two decades of negative-cost float. The letter judges that the float “has a reasonable prospect of being costless” from here. The same letter names the danger: a P&C insurer receives payment up front and learns its costs much later, so it can be losing money before its managers realize it (Part 1.4: Reserving — Provisioning for Claims Not Yet Paid).
Float is functionally similar to a bank’s deposits (Volume I’s Part 3) — money that belongs to someone else (policyholders, eventually), but that the institution can invest and earn a return on in the meantime. The crucial discipline, exactly as with an insurer’s actuarial reserves, is investing this float in assets whose duration reasonably matches when the money will actually be needed to pay claims — a serious mismatch here is precisely the vulnerability The Practitioner’s Codex Part 3 warned about for any institution holding long-duration liabilities.
Why duration matching decides who wins when rates move. An insurer’s liabilities are promises of future payments, valued (under modern accounting and solvency rules) by discounting them at current market rates. When rates rise, the present value of those promises falls, exactly as a bond’s price does; when rates fall, it rises. Whether the insurer gains or loses therefore depends on whether its liabilities or its assets are more rate-sensitive, measured in dollars: value × duration (the percentage price change for a one-point move in rates, from Volume II, Part 3: Sovereign Debt & Public Finance). This is the mechanism behind the 2022 episode traced in Part 10.2: Case One — A Global Recession, Fully Traced.
An insurer holds assets of $1,100 million with a duration of 6 against liabilities with a present value of $1,000 million and a duration of 10, so surplus is $100 million. Using the duration approximation (change ≈ −value × duration × rate change, ignoring convexity):
| Scenario | Assets | Liabilities | Surplus | Change in surplus |
|---|---|---|---|---|
| Rates fall 1 point | $1,100m + $1,100m × 6 × 1% = $1,166m | $1,000m + $1,000m × 10 × 1% = $1,100m | $66m | −$34m |
| Rates unchanged | $1,100m | $1,000m | $100m | $0 |
| Rates rise 1 point | $1,100m − $66m = $1,034m | $1,000m − $100m = $900m | $134m | +$34m |
Because the liabilities carry more dollar duration ($1,000m × 10 = $10,000m) than the assets ($1,100m × 6 = $6,600m), this insurer gains when rates rise and loses when they fall. The flip point is where dollar durations match: asset duration = $1,000m × 10 ÷ $1,100m = 9.09. Below 9.09 the insurer is betting on rising rates; above it, on falling rates; at 9.09 surplus is approximately immune to small parallel moves. Long-dated promises often leave life insurers and pension plans in exactly this position, holding less dollar duration in assets than in liabilities, so a rate rise improves their surplus; the 2022 pension data below show the effect at scale.
Pension funds face the same arithmetic. A defined benefit (DB) plan promises a formula-based income for life, so the sponsor bears investment, interest-rate and longevity risk; a defined contribution (DC) plan such as a 401(k) promises only the contributions, so the employee bears all three. In the Bureau of Labor Statistics’ March 2026 survey, 14% of private-industry workers had access to a DB plan and 70% to a DC plan, against 86% DB access for state and local government workers. DB liabilities are long (often more than a decade of duration) and are valued with a discount rate tied to bond yields, so funded status swings with rates. The household side of these plans, from vesting to payout options, is covered in Volume V.
Liability-driven investing (LDI) means choosing assets to move with the liabilities rather than to maximize return: long bonds, interest-rate swaps and inflation-linked bonds sized so that the plan’s assets gain about as much as its liabilities when rates fall. The hedge ratio is the share of liability rate sensitivity that the assets offset.
The US test. Milliman’s study of the 100 largest US corporate DB plans found that their average discount rate jumped from 2.73% at the end of fiscal 2021 to 5.18% at the end of 2022, the largest one-year move in the study’s 23-year history. Liabilities fell by $520.8 billion, from $1.848 trillion to $1.328 trillion, a 28.2% drop, even though plan assets returned −18.6%. The funded ratio (assets ÷ liabilities) rose from 96.3% to 99.3%. The drop also gives a rough duration: 28.2% ÷ 2.45 points ≈ 11.5, an upper-bound reading because the liability change also nets new accruals against benefits paid.
The UK test. UK schemes had used leverage (borrowing) to hedge more of their liabilities while keeping assets in equities. When the 30-year gilt (UK government bond) yield rose by 160 basis points in a few days in September 2022, scheme liabilities fell, but the leveraged hedges lost value immediately and demanded cash collateral, which forced gilt sales into a falling market. The Bank of England stepped in, buying £19.3 billion of gilts between September 28 and October 14, 2022. In its letter to Parliament, the Bank noted that a DB pension fund “might be better off overall” from the rise in yields. The danger was liquidity, not solvency.
Match dollar durations, then fund the hedge’s liquidity. Set asset dollar duration close to liability dollar duration (an asset duration near 9.09 in the example above) unless the board has chosen and sized a rate view. If the hedge is leveraged, then hold liquid collateral of at least exposure × duration × the stress move you must survive, which caps leverage at 1 ÷ (duration × stress): for a 20-duration hedge, 1 ÷ (20 × 1.6%) = 3.1 times to survive 160 basis points, and 1 ÷ (20 × 2.5%) = 2.0 times for 250. The stress size is a policy choice; 2022 proved 160 basis points in days is possible.
Hedging solvency risk with leverage and forgetting liquidity risk. A pooled LDI fund with $100 million of capital supporting $400 million of long-bond exposure (4 times leverage, duration 20) loses about $400m × 20 × 1.6% = $128 million on a 160-basis-point rise, more than all its capital. The scheme behind it is better off on paper, because its liabilities fell further, but unless it can wire cash within days the hedge is closed out at the worst moment. That is the spiral the Bank of England interrupted in 2022.
What is insurance float?
Money an insurer holds between collecting premiums and paying claims: unearned premiums plus reserves for unpaid losses. The insurer invests it and keeps the income. Its cost is the underwriting result, so a break-even or better underwriter holds float for free or is paid to hold it.
Do rising interest rates help or hurt insurers and pension funds?
It depends on which side carries more dollar duration (duration × value). Life insurers and pension plans, whose long-dated liabilities usually outweigh their bonds in dollar duration, see solvency improve, because liabilities fall by more than the bonds backing them. A property-casualty insurer whose bonds run longer than its short-tail claims sees surplus fall. In 2022 the 100 largest US corporate pension plans saw liabilities fall 28.2% and funded status rise from 96.3% to 99.3% despite an 18.6% asset loss. The strains are collateral calls on leveraged hedges and policyholders surrendering savings products for higher yields elsewhere.
What is the difference between a defined benefit and a defined contribution pension?
A defined benefit plan promises a set income, typically years of service × accrual rate × salary, with the employer bearing investment and longevity risk. A defined contribution plan promises only the contributions; the employee’s balance depends on what goes in and what it earns. In March 2026, 14% of US private-industry workers had access to a DB plan and 70% to a DC plan.
Float, the money held between premium and claim, costs the insurer its underwriting loss and earns investment income; strong underwriters hold it at no cost or better. Whether rate moves help depends on dollar duration: liabilities longer than assets gain when rates rise, as US pension plans did in 2022, while leveraged hedges, as in the UK gilt crisis, turn a solvency gain into a liquidity emergency.
Four questions on this chapter. Decide on your answer first, then click “Reveal Answer.”
1. An insurer loses $4 million on underwriting while holding $200 million of average float. What is its cost of float?
- 0.5%
- 4%
- 2%
- −2%
Reveal Answer
Answer: C. Cost of float = underwriting loss ÷ float = $4m ÷ $200m = 2%.
2. Liabilities have a present value of $450 million and a duration of 12; assets total $500 million. What asset duration makes surplus approximately immune to small parallel rate moves?
- 12.0
- 13.3
- 9.6
- 10.8
Reveal Answer
Answer: D. Match dollar durations: $450m × 12 ÷ $500m = 10.8.
3. A pension plan’s liabilities carry more dollar duration than its assets. What happens to its surplus if interest rates rise?
- It rises, as liabilities fall more than assets
- It falls, because liabilities are discounted less
- It is unchanged, because both sides fall
- It falls, because assets lose market value
Reveal Answer
Answer: A. A higher discount rate cuts the present value of liabilities; with more dollar duration on that side, the liability fall exceeds the asset fall.
4. What was the main danger for leveraged UK pension hedges in September 2022?
- A sharp rise in pension liability values
- Collateral calls demanding cash within days
- A wave of defaults on UK government bonds
- A fall in the funded status of most schemes
Reveal Answer
Answer: B. Liabilities fell, but leveraged hedges lost value and required cash collateral, forcing gilt sales; the problem was liquidity, not solvency.
- US Bureau of Labor Statistics, Employee Benefits in the United States, March 2026 — DB and DC access rates
- Bank of England, Letter from Jon Cunliffe on LDI (October 5, 2022) — 160-basis-point gilt move; pension funds better off overall
- Bank of England Quarterly Bulletin 2023, gilt market case study — £19.3 billion of purchases, September 28 to October 14, 2022