Experiments in the Projection of Mortality

Abstract

How fast will mortality fall in the decades ahead? One way of phrasing the question is in terms of past periods: will it be as fast as Canada showed in 1976-81, or only as fast as the average 1921-81, or as slow as 1926-31?...I will argue that the whole matter of projecting mortality comes down to deciding what past period describes the future.... The author concludes that any of several methods could be used, including the Brass method, but that "no trend calculation, or regression on economic variables, seems able to forecast the future, that is to [accurately account for such events as] technical advance...and new fashions in behaviour." Life tables for Canada concerning the period 1920-1982 are used as illustrations.