A reply to 'On the Far Eastern pattern of mortality' by Zhongwei Zhao.

In a paper published in the previous issue of Population Studies , Zhongwei Zhao questions the uniqueness of a mortality pattern first identified by Goldman (1980) and subsequently redefined by the United Nations (1982). As originally formulated by Goldman, the Far Eastern pattern is characterized by ‘excessively high’ death rates among older men relative to death rates among both younger men and women of all ages. Thus, the Far Eastern pattern has two key features: an atypical age pattern of mortality for males and large sex differences in death rates at older ages. In the original formulation of this pattern, ‘excessively high’ was assessed by the extent of conformity of observed mortality schedules to the Coale and Demeny model life tables (Coale and Demeny 1967). The United Nations (1982) redefined the Far Eastern pattern in their collection of model life tables for less developed countries. The United Nations’ pattern embodies the age pattern described by Goldman for men, but includes populations in which this age pattern is observed among women as well as men. In his evaluation of the Far Eastern mortality pattern, Zhao does not dispute the fact that the schedules identified by Goldman (1980)—namely, Korea, Hong Kong, Singapore, and Taiwan for various periods between the 1950s and the 1970s— exhibit the key features of the pattern. However, he maintains that this pattern is not distinctive. Specifically, he claims that the same characteristics have been observed in numerous Western populations in the twentieth century. In the discussion below, I challenge Zhao’s assertion by demonstrating that, according to the criteria set forth by Goldman (1980), the Far Eastern schedules are indeed unique among countries with reliable mortality statistics. They are characterized by larger discrepancies between the mortality of young adult and older adult males and between the mortality of older males and older females than the Western populations identified by Zhao. The weaknesses in Zhao’s argument stem in part from his reliance on graphical displays, which make it difficult to discern patterns, as well as from his use of different criteria from Goldman (1980) in assessing the extent of discrepancy between observed and model schedules.