Manuel Funke

Graduate School of North American Studies

Abteilung Wirtschaft

John F. Kennedy-Institut für Nordamerikastudien


Lansstraße 5
14195 Berlin

The Development of Saving Rates among U.S. Income Groups, 1984-2007

The personal saving rate in the United States has fallen dramatically over the past three decades. Average saving declined from approximately 9 percent of disposable income in the mid-1980s, to 5 percent in the 1990s, to almost zero after the turn of the millennium. Given the mixed and weak quantitative evidence on explanations referring to demographic, political, and innovation-driven changes that may have caused a large-scale shift away from saving towards consumption, the decline in the U.S. personal saving rate provokes a reconsideration of the basic economic theories on the dynamics of household saving.

The bulk of literature suggests a positive relationship between economic well-being and savings; it is widely accepted that household saving rates tend to increase with growing per-capita incomes. However, the real economic development in the United States looks different. Even though average incomes have improved by one-third over the last three decades, the personal saving rate has been constantly decreasing. This becomes even more controversial if one considers the changing distribution of incomes. Assuming a positive correlation of incomes and savings, mainstream economic theory concludes that, ceteris paribus, income shifts from low-income (low-saving) groups to high-income (high-saving) groups result in a higher aggregate saving rate. Again, despite a significant upward concentration of income in the United States over the past thirty years, the country’s aggregate household saving rate has fallen.

Considering these apparent contradictions, it seems useful to study the actual saving behavior of different income groups during the period of the savings decline in more detail. Who saved and who “dissaved” in the U.S. case? Which income group had what share in the decline of the saving rate? Was the development of the savings distribution in step with the evolution of the income distribution? And, in more general terms, do the rich in fact save more than the poor? 

I use micro-level survey data to construct a representative household dataset for the entire period of the savings decline. I extend a measure of active saving which has only been applied for the period from 1984 to 1994, until 2007, and calculate saving rates for low, medium and high income groups. This approach makes it possible to investigate household saving in multiple dimensions. For each income group, the saving rate can be viewed both in isolation and in comparison, over time and at any given point in time. Moreover, capital gains are incorporated into the analysis. The division into active and passive saving gives a deeper insight into the changing composition of household saving.

Investigating the dynamics of the U.S. personal saving rate by identifying the saving behavior in different income groups and by accounting for capital gains can help to further disentangle the savings puzzle. Only the use of a consistent micro-level dataset that is long-running and multi-dimensional allows for identifying, tracing, and comparing the changing patterns of saving behavior among different income groups.