Healthiness and the economic determinants of mortality fluctuations

The intense mortality fluctuations characteristic of preindustrial populations and some poor populations today, have been linked to harvest or catch fluctuations. These fluctuations have been highlighted as one of the explanations of low population growth rates. These affirmations were statistically measured, but they were not grounded on theory.;This dissertation attempts to make explicit the theory behind the economic determinants of mortality fluctuations in poor populations. The first component of the theory is the biological connection between nutrition, epidemics and mortality. A class of models is proposed that relies on the characterization of mortality risk by using a latent variable, healthiness, that depends dynamically on the nutritional and epidemic experience of the individual. Under general conditions increased variability in nutritional intake leads to higher mortality levels for a given average intake.;The second component of the theory is behavioral. Given the negative impact on survival of fluctuating individual food consumption, and people's concern about their survival, behavior is modified in order to smooth food consumption. This behavior is limited by the available technology and the organization of society. Behavioral adaptations to risk are therefore dependent on the particular environment and society. Institutions and constraints are studied for hunter-gatherer, horticultural and agricultural populations. These practices include a whole range including the selection of crops and technology, food-sharing and reciprocity, mobility or crop storage.;Economic theory provides a valuable tool for understanding the limits and possibilities of these strategies. A model of optimal crop storage is proposed that considers the effects of consumption on healthiness and mortality. Risk aversion or survival maximization lead to similar optimal storage policies. Together with the dynamic models of healthiness, the storage model captures the dynamic relationship between harvest fluctuations and mortality as empirically measured. It introduces nonlinearities that explain the asymmetry between mortality peaks and valleys, the recurrence of peaks, and the effects of selection on the response function. At the present stage this is a theoretical model only. No numerical prediction from the model is intended to match any quantity in the real world.;Mortality time-series from British Bengal and Punjab are analyzed, which conform to the patterns expected from the theoretical analysis, and a method is proposed to measure the impact of risk on mortality levels.