A Trend-Change Extension of the Cairns-Blake-Dowd Model
Paul Sweeting
This paper builds on the two-factor model developed by Cairns et al (2006) for
projecting future mortality. It is shown that these two factors do not follow a random
walk, as proposed by Cairns et al, but should instead be modeled as a random
fluctuation around a trend, the trend changing periodically. Projecting mortality rates
in this way suggests much greater uncertainty over future mortality improvements.
Keywords: stochastic mortality; mortality projections; mortality trends; Lee-Carter;