The Bourgeois-Pichat's biometric method and the influence of climate: new evidences from late 19th-century Italy.

This paper examines to which extent seasonal and climatic conditions might affect the reliability of the Bourgeois-Pichat's method. Other scholars have already argued on this issue, but although climate has often been claimed to explain part of the differentials in mortality figures among Italian regions, to date its impact has not actually been recognized and quantitatively evaluated. To test such hypothesis data at the regional level from late 19th-century Italy have been analyzed. Our analysis of the biometric components revealed a strong bias in the estimates of the endogenous and exogenous components in the first month of life. Variations in infant mortality among Italian regions correlated with variations in the endogenous levels rather than in the exogenous levels of infant (neonatal) mortality, as it was expected owing to the infective nature of the diseases climate might induce. Specifically, Northern and colder regions featured high figures for both neonatal mortality and the endogenous component, while the opposite scheme applied to the Southern, more temperate regions. Finally, the reasons for such misleading results were investigated. It emerged that the model's assumption of a constant and invariant proportion of neonatal exogenous deaths to the total amount of exogenous deaths was not matched by the Italian data. This situation caused the excess neonatal exogenous mortality, especially that induced by cold climate in Northern regions, to be wrongly counted in the endogenous component.