Rieder, H.
(1996):
Estimation of Mortalities.
Collaborative Research Center 386, Discussion Paper 28

Preview 

429kB 
Abstract
If a linear regression is fit to logtransformed mortalities and the estimate is backtransformed according to the formula Ee^Y = e^{\mu + \sigma^2/2} a systematic bias occurs unless the error distribution is normal and the scale estimate is gauged to normal variance. This result is a consequence of the uniqueness theorem for the Laplace transform.
We determine the systematic bias of minimumL_2 and minimumL_1 estimation with sample variance and interquartile range of the residuals as scale estimates under a uniform and four contaminated normal error distributions. Already under innocent looking contaminations the true mortalities may be underestimated by 50% in the long run.
Moreover, the logarithmic transformation introduces an instability into the model that results in a large discrepancy between rg_Huber estimates as the tuning constant regulating the degree of robustness varies.
Contrary to the logarithm the square root stabilizes variance, diminishes the influence of outliers, automatically copes with observed zeros, allows the `nonparametric' backtransformation formula E Y^2 = \mue^2 + \sigma^2, and in the homoskedastic case avoids a systematic bias of minimumL_2 estimation with sample variance.
For the companyspecific table 3 of [Loeb94], in the age range of 2065 years, we fit a parabola to root mortalities by minimumL_2 , minimumL_1, and robust rg_Huber regression estimates, and a cubic and exponential by least squares. The fits thus obtained in the original model are excellent and practically indistinguishable by a \chi^2 goodnessoffit test.
Finally , dispensing with the transformation of observations, we employ a Poisson generalized linear model and fit an exponential and a cubic by maximum likelihood.