Building New Evidence of Historical Child Mortality Differentials for Canada

Luiza Antonie, University of Guelph
J. David Hacker, University of Minnesota
Kris Inwood, University of Guelph

Systematic evidence for standard of living and its demographic consequences is limited for pre-WWII Canada. Civil registration of births, deaths and marriages was organized independently by province and remained incomplete in most provinces until the 1920s. Consequently, few consistent measures of inequality across regions, ethnic groups, the rural-urban divide, native-born vs immigrant, etc. are available. Fortunately, the recent emergence of complete count historical census data and a new technique to estimate child mortality suggest a way to acquire new evidence. We apply the ‘missing children’ method of Hacker, Drive and Helgertz (2023) to examine child mortality differentials for Canada 1871-1921. A supervised classification model identifies several hundred thousand couples in successive census enumerations. This provides the central building block for our study of child mortality differentials. A major advantage of our datasets is the inclusion of earnings in 1901, 1911 and 1921. This allows us to examine the relationship between economic and demographic inequality in much greater detail than possible in the United States and elsewhere. Experience elsewhere leads us to expect a decline in child mortality decline for Canada in this period although the timing might well have varied across various sub-populations and communities. For example, a small but robust literature using complementary indicators (infant mortality from Thornton and Olsen 1988, 1991, 2001 and adult stature from Cranfield and Inwood 2007, 2015) suggests the possibility of a delay in mortality decline among French-descended relative to other Canadians. Also of interest is mortality change during substantial outmigration from eastern Canada in the 19th century and during western Canada’s explosive early 20th century immigration boom. On each of these topics granular census data developed using the Hacker-Dribe-Helgertz method will document differentials and change over time, and allow us to disentangle the contributions of income, class and local spatial environment.

No extended abstract or paper available

 Presented in Session 177. Building and Interpreting Censuses II