Short-run fluctuations in fertility and mortality in pre-industrial Sweden

This paper describes and interprets annual Swedish data from 1750 to 1869 on weather, harvests, real wages, birth rates, and death rates using vector autogression. Impulses due to unexplained increases in wealth, whether this occurred through increased real wages, improved agricultural yields, or warmer winters, led in the short run to increased fertility and decreased infant and non-infant mortality, and hence to increased rates of population growth. Unexplained or unanticipated fluctuations in infant mortality led to replacement cycles in fertility within one to three years, although only a negligible cumulative effect on fertility persisted after five to ten years. Fluctuations in deaths among persons older than one year evoked a fertility response several years later, but this replacement response persisted after more than a decade. Although vector autoregression is not designed to account for long-term trends and their consequences, the interrelationships found here among exogenous weather shocks and fluctuations in economic conditions and demographic rates provide support for the homeostatic mechanisms hypothesized by classical economists and discussed by Malthus. The methodology of vector autoregression appears useful for studying historical series on climatic, economic and demographic variables where we do not yet have a sufficient theoretical foundation for specifying and estimating structural models.