Citation

Wilson M, Daly M. Br. Med. J. BMJ 1997; 314(7089): 1271-1274.

Affiliation

Department of Psychology, McMaster University, Hamilton, Ontario, Canada.

Copyright

(Copyright © 1997, BMJ Publishing Group)

DOI

unavailable

PMID

9154035

PMCID

PMC2126620

Abstract

In comparisons among Chicago neighbourhoods, homicide rates in 1988-93 varied more than 100-fold, while male life expectancy at birth ranged from 54 to 77 years, even with effects of homicide mortality removed. This "cause deleted" life expectancy was highly correlated with homicide rates; a measure of economic inequality added significant additional prediction, whereas median household income did not. Deaths from internal causes (diseases) show similar age patterns, despite different absolute levels, in the best and worst neighbourhoods, whereas deaths from external causes (homicide, accident, suicide) do not. As life expectancy declines across neighbourhoods, women reproduce earlier; by age 30, however, neighbourhood no longer affects age specific fertility. These results support the hypothesis that life expectancy itself may be a psychologically salient determinant of risk taking and the timing of life transitions.US social scientists have observed that, among the urban poor, an expectation of a compressed life span and uncertain long-term survival is associated with seemingly reckless risk-taking behaviors. Under conditions of poverty and inequality, however, discounting the future may be an adaptive rather than pathological response. This dynamic was investigated further through use of demographic data for 1988-93 from 77 neighborhoods in Chicago, Illinois. Neighborhood-specific, cause-deleted male life expectancy rates at birth (range, 54.3-77.4 years) and homicide rates (range, 1.3-156/100,000/year) were significantly associated (p 0.0001), confirming the hypothesis that criminal behavior varies as a function of local life expectancy. In step-wise multiple regression analysis, economic inequality added significantly to the prediction of the homicide rate afforded by life expectancy, suggesting that mortality is exacerbated by inequality itself. Although the pattern of deaths from internal causes across the lifespan was similar between the 10 neighborhoods with the shortest life expectancies and the 10 with the longest life expectancies, patterns for deaths from external causes such as violence were substantially different. Finally, the median age at childbirth was 22.6 years in neighborhoods with low life expectancy, 25.4 years in intermediate neighborhoods, and 27.3 years in those with long life expectancy, supporting the hypothesis that early reproduction among the urban poor reflects an active decision motivated by expectations of a life course compressed in time. Suspected is a feedback loop in which local levels of homicidal violence affect expectations of future life, discount rates, and hence further violence.

Language: en