Life expectancy, average age of death, and tempo-adjustment

The recent proposal by Bongaarts and Feeney for a tempo-adjustment to life expectancy is assessed using numerical techniques. I find that for a wide variety of situations no such adjustment is necessary provided mortality is observed in a given period. Bongaarts and Feeney assume the force of mortality affecting the mean age of death is not always observed. The present work brings into sharper relief the question of why this aspect of mortality is not observed, a claim of Bongaarts and Feeney not adequately justified in their paper. ❧ Studies of mortality have been at the center of formal demography since Graunt’s (1662) Bills of Mortality for London (cf. especially chapter XI, §9–11) and Halley’s (1693a,b) life table for Breslau, Silesia (now Wrocław, Poland). Life expectancy has long been a basic ingredient of mortality analysis, and is probably the most sought-after demographic statistic by non-demographers. John Bongaarts and Griffith Feeney (2002) recently proposed an intriguing tempo-adjustment for life expectancy. Their paper is interesting and important, and will potentially ∗This work was supported in part by NICHD grant T32-HD-07275-16 to the Department of Demography, University of California at Berkeley. Thanks to Ken Wachter for comments. Note added, January 2009: I last made substantial revisions to this document in 2005. I am no longer actively working on this particular strand of research, though I do maintain interest in formal demography in general and models of mortality in particular. †Department of Sociology, University of California, Irvine ‡Department of Population Health and Disease Prevention, University of California, Irvine §Health and Global Change (HGC) project, International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria.