Bob Michael
University of Chicago

The case for intergenerational research as a priority area for the DBSB is two-fold, I think. First, much of the behavior of primary interest in population research-fertility, marriage, health and mortality, human capital investment and much more-is influenced by cross-generational circumstances. The current family or individual who is the focus of a particular analysis has a background that typically influences the actions under study. It is helpful, often essential, to know about that accumulated, prior family influence in understanding and predicting subsequent behavior or responses to changes in circumstances. Second, the study of cross-generational influences is feasible, fortunately, because there is much good micro-level data of a multi-generational nature. So the topics for study are often important ones and the data resources, that would be very expensive to create if they did not exist, are relatively available to address the study of intergenerational influences. I see this research not as an isolated or separate agenda so much as one important strategy and focus in the study of several key areas of research priority.

In many ways social demography shows that "family matters" in understanding or predicting much behavior and many preferences and attitudes. One common example is through financial resources in the form of support of children's schooling and in the form of insurance against risks taken. Another is evidence on the influence of the parent's education on the aspirations, behaviors, motivations of the next generation. There are many, many examples in the literature of background factors that influence actions and circumstances. The study of intergenerational transmission goes at least one step beyond that, however, and looks at the mechanisms through which one generation affects another. It is a strategy of making endogenous the family background factors and making more explicit in theory and evidence how that background has influence. While economists are notorious for assuming a utility function and thereafter not paying much attention to where those tastes or preferences came from, intergenerational research may be able to help economists understand why preferences are what they are, in the sense of how families help formulate those preferences and expectations and why they are so varied in some domains and so conflictual in certain collective domains. Studying how a family at one time point behaves and how that behavior influences the behavior a generation later at the comparable point in the family lifecycle, can tell us something about cross-generational transmission of attitudes or preferences or behavioral patterns or habits. When modeled more formally these studies might shed light on how families influence behaviors. When we write about the differences in observed behaviors by race or ethnicity or social class, the study of intergenerational mechanisms can deepen what we know about those correlations.

There is an influential class of models of multi-generational implicit social contracts that features the life stages of dependency and independence and explores how the implicit agreements between generations accommodate the venerability of one life stage by a quid-pro-quo with family members or society at another stage. In macro-growth models as well as micro-investment and lifecycle consumption models, these implicit contract are important. Their explication empirically through the analysis of intergenerational research is an important priority.

There is a potential for better strategies for certain policy interventions if viewed from the perspective of an intergenerational framework. If our attitudes, expectations, preferences, habits, and behaviors are much influenced by our family background and the attitudes and preferences transmitted from parent to child, then it may be useful in formulating strategies for persuading or encouraging changes, to take that fact and the ways in which the family influence has worked into account. Families greatly influenced by the lack of liquidity of assets during the Great Depression later saved money in highly liquid but low interest-yielding instruments, and those family habits affect savings behaviors even in later generations through the transmission of the habit and the orientation toward risk and illiquidity, even if the precipitating circumstance is no longer identifiable. Similarly, those whose families were scarred by war or abuse carry the attitudes and hatreds of their family's earlier generations into social interactions currently, yielding religious, cultural and class tensions that impact society. Study of culture and race, as proposed by others in this conference, can be closely related to intergenerational research.

There are several large-scale data resources that have much valuable information about cross-generational behavior, some collected in real time over the past half century. In some instances a single generation has been followed for many years so that the subject was at once the child and later the parent in the survey's questioning. In other instances, the data structure includes the children of the subjects, added as a data supplement and allowing a cross-generational study based on that condensed time frame. The PSID offers both of these alternatives, the older NLS files do as well while the Child-NLSY is a prime example of the latter data structure and the newest NSLY97 that has a concurrent interview of the youth and the parent offers some opportunity to get the two-generations measured at a single point in time. The British Birth Cohort Studies, begun with the births in a week of 1946 and followed thereafter, and the parallel National Child Development Study (NCDS) of a week's birth cohort in 1958, and another in 1970, and especially the child of the NCDS undertaken in 1991 illustrate the data resources that now are available for intergenerational research. Given the breadth and importance of the topic and the availability of data sets, this should be a high priority area of research for the DBSB over the next few years.