Fund the premiums. Discount at a rate you can defend. Let the evidence pick the mortality table.
The industry discounts future cash flows at 13–15% — more than six hundred basis points above a coherent risk build — for an asset whose payoff is among the most predictable in finance. The reason is a conceptual error: the method never funds the premiums, so the same formula gives different answers for the same policy. Take one policy — $1,000,000 face, male age 80, $40,000 first-year premium, VBT 2018 mortality — and value it four ways.
| Case | Premium reserve | Discount rate | Fair value |
|---|---|---|---|
| Industry convention | None assumed | 13% “market rate” | $250,308 |
| Case 1 — no reserve | Zero (policy lapses in year 2) | 4.20% (three-month Treasury) | $41,533 |
| Case 2 — full reserve, cash | 36 years, $3,590,543 at 0% | 6.59% | ($876,486) |
| Case 3 — LSCAPM | Optimal 15 years · Treasury ladder $589,311 | 6.59% / Treasury curve | $316,760 |
Funded with Treasuries and discounted at an ordinary 6.59%, the policy is worth 27% more than the convention pays for it. The convention needs 13% only because it pretends the premiums are free; fund them, and a true capital asset pricing model shows the asset is underpriced. That spread is not a bet on markets. It is the return for being the buyer who prices the asset correctly and funds the premiums.
Face value × annual mortality probability, discounted at the LSCAPM rate. Sample policy: $631,237 — nearly two-thirds of face.
A Treasury note for every premium, through the year where the next premium stops adding value. Sample policy: 15 years, $863,143 of premiums, ladder cost $589,311 today.
Notes still held when the policy matures are released — a second death benefit — weighted by mortality and discounted at the same rate. Sample policy: $274,834.
They fix the cost of every premium on day one, they hedge rates (a 100 bp parallel shift moves fair value about 8%, against about 24% with cash at 0%), and they come back when a policy matures early.
6.59% as of September 30, 2026 is not an opinion. It is the sum of nine components, three read directly from published markets at each quarter end and six governed by written rules.
| Component | Basis | Sept 30, 2026 |
|---|---|---|
| Risk-free rate | One-year U.S. Treasury constant-maturity yield (FRED DGS1) | 4.54% |
| Term premium | Par yield at the pool's expected duration (7-yr) less the one-year rate | 0.65% |
| Carrier default | ICE BofA AA U.S. Corporate option-adjusted spread | 0.60% |
| Six pool-specific risks | Disputed claims, regulatory change, systemic longevity, reserve adequacy, pool structure, execution — each under a written rule | 0.80% |
| LSCAPM discount rate | Sum of the nine components | 6.59% |
What is deliberately not in the rate: life-expectancy uncertainty, concentration and cost inflation. Those are already in the cash flows. Putting them in the rate too is how the industry gets to 13% — the same risk counted twice.
We do not pick one mortality curve. We carry several and let the pool's own experience vote: count the maturities each year, compute the binomial likelihood of that outcome under each curve, normalize into weights, and take the weighted average of each curve's net present value. A curve that keeps over-predicting deaths loses credibility; a year of experience shifts the weights modestly, never violently. Like a loaded-dice test — ten rolls prove little, a thousand settle it.
Market sentiment, crypto cycles and equity drawdowns have no channel into the model. The only input that changes the mark is observed mortality, plus the published rate and spread inputs refreshed each quarter.
Every assumption — mortality curves, weights, discount build, acquisition cost — is explicit and traceable to a published table or the pool's own record. The result is a Level 3 fair value that satisfies ASC 820 / IFRS 13 and can be checked line by line. Anyone with the token can re-run the mark.