It appears that the widely-observed finding in the literature showing little or no relationship between population growth (and dependency) and saving requires modification based on panel and cross-section estimation of aggregate country data. First, while it is consistent with an update of the hybrid (Keynesian, life-cycle) Leff-type model for the 1960s and 1970s, this is not true of the 1980s, when there is a "jump" in the importance of demography. Second, in a more clearly interpretable life-cycle framework (a la Mason), the role of demography increases smoothly and systematically over the 30 years, although the form which it takes (life-cycle versus lifetime level) also varies systematically over time. Overall, however, demographic factors accounted for a major portion of changes in saving across countries and over time.
