Lee-Carter model
A mortality projection model published in 1992 that summarizes how death rates at every age change through a single time index, then extends that index as a random walk.
To price an annuity or reserve for a pension plan, one must know who dies today and how fast mortality will fall. The Lee-Carter model writes the logarithm of the death rate at each age as an age-specific level plus that age's sensitivity times an index shared across all years. The index is estimated from history through a singular value decomposition, then projected as a random walk with drift, which yields a central path and confidence intervals. Its contribution was to make projection reproducible and probabilistic, which allows trend risk to be quantified. Its known weakness is that it has often underestimated gains at the oldest ages, hence the many cohort-based variants built on it since.
In the Journal of the American Statistical Association of September 1992 (vol. 87, no. 419), Ronald Lee and Lawrence Carter fit the model to US death rates from 1933 to 1987 and projected a life expectancy of 86.05 years in 2065, a gain of 10.5 years, with a band of plus 3.9 to minus 5.6 years. The Census Bureau later used their projection as its long-run benchmark.
Follow the research
One letter a week. First, the week's underwriting note, when one is published: a dated, argued decision, signed by Paul Bernardski. Then what the observatory published on cyber risk, political risk and insurance. No account, no tracking.
No spam, one-click unsubscribe, no data ever sold.