• Notes a movement away from economic determinism in mortality analysis and seeks to estimate the role of economic factors in the increase in life expectancy during the 20th century
• Must look at the responsiveness of mortality to economic change (cross-sectional analysis)
• “The increase of 24-25 years during the 20th century is essentially unaccounted for” (244)

Why national income rather than other socioeconomic variables?

• It’s the best single indicator of living standards in a country
• Income per head (the leading index of economic development) is the focus of growth models which drive policy decisions
• Income per head: an index of the GNP/capita (minus services, minus replacement of capital).  It has an indirect influence on mortality through consumption patterns

Types of Relationships (between income and mortality): existing theories, but all problematic
1) Level of Income Influences Level of Mortality at a Moment in Time

• High correlation coefficients between variables of income and infant mortality rates
• Cross-sectional relationship b/w mortality rates and levels of income is used, but questioned

• Tough to create a model to test this
• Maybe it has to do with the rate of change of income associated with level of income

• A dynamic relationship between these two is implicated in many theories…but other elements likely figure into the dynamic relationship

The Relations Re-Examined
See scatter diagrams on p. 235 for patterns of income-mortality relationships in 20th century

(1) The relationship between life expectancy and national income per head has shifted upwards

• Note that deaths in later generations  reflect diseases beyond the range of basic health care delivery to the poorest of countries…with income, they’ve "graduated" to newer illnesses

(2) Factors exogenous to a country’s current level of income probably account for 75-90 % of the growth in life expectancy in the world (1930s-60s); income growth per se = 10 to 25 %

• Income has to be important, but it can only be so important given the “massive shift” in the relation between income and life expectancy.  Other studies: income changes were of minor importance in mortality trends (238)
• They estimate exogenous factors caused 84 % of mortality decline, and income 16 %
• What of the 84%?  They ruled out literacy and nutrition
• This [the second statement] is the key, recurring point to the paper

(3) Mortality has not become progressively dissociated from standards of living at given moment

• But the shape of the relationship has changed; there are some indications that mortality has become more responsive to standard of living in low-income countries “where economic-demographic interrelations are most critical for economic prospects” (240)
• Why? New health measures specifically targeting highly responsive areas, or new health measures only exploitable in areas above a basic income level? Don’t know
• Questions/theories about how long before income-induced population growth erases income growth, but the authors think it’s irrelevant (at least 100 years is their answer)

(3a) Some of the observed scatter in cross-sectional relations is almost certainly caused by differences in national income distributions

• “The distribution of incomes is clearly a likely source of variance in the basic relation between national life expectancy and average national income and accounts for some of the scatter in the observed relationship” (242)…but data lacking for explicit analysis
• Research on the Soviet bloc: there is no strong evidence that greater income equality has raised life-expectancy above income-predicted levels

(4) Factors exogenous to a nation’s level of income per head have had a major effect on mortality trends in more developed as well as in less developed countries
Assumptions about advanced standard of living (among developed countries) and medical health support (among less developed ones) are true with qualifications.  Level of analysis hides period-specific analyses revealing influences on both levels (developed v. un/underdeveloped) --> cross-sectional analyses are better and this is why