OTher Challenges For The Life-Cycle Model

Other Challenges for the Life-Cycle Model
The life-cycle framework faces a number of other challenges that don’t fit naturally into our survey of smoothing at different frequencies.
First, as we have noted above, the life-cycle framework proposes to integrate many aspects of behavior in a coherent and disciplined way. Examples beyond consumption and savings include portfolio choice, fertility andretirement. Building empirical models within the life-cycle framework that successfully integrate these different aspects of behavior is obviously a daunting task. Apparent failures in this context should not lead us to a wholesale abandonment of the framework: it is still very early days.
A second area in which the life-cycle framework faces important challenges is in explaining cross nationaldifferences in savings rates (Deaton, 1992). This subject was one of the original motivations for Modigliani-style life-cycle models, but the attempt to relate aggregate savings rates to differences in population structure and economic growth has not been a success. Assessing this literature is somewhat like determining whether a glass is half-empty or half-full. Recent microdata analysis of thesavings behavior of different cohorts in different countries (Deaton and Paxson, 2000) is more favorable to life-cycle models, but they still fail to provide a complete account of the aggregate relationship between growth and savings.
A final issue is the ability of the standard model to provide an explanation for various aggregate consumption and savings episodes. Attanasio and Weber (1994) discussthe consumption boom in the United Kingdom in the late 1980s and provide plausible explanations that are founded in the life-cycle framework. On the other hand, there are no widely accepted explanations for the decline in the U.S. savings rate in the mid-1980s. It is important to note, however, that the definition of savings reflected by the National Income and Product Accounts is not the same assavings as defined by life-cycle models, because the latter includes several components—such as capital gains—which the former does not (Gale and Sabelhaus, 1999). Thus it is perhaps not surprising that life-cycle models are not entirely successful in explaining the movements of this aggregate.14
Future Research Directions for the Life-Cycle Model
The life-cycle framework is just reaching itsprime of life. In many of the areas in which the framework currently has problems—for example, savings before and after retirement—it is only very recently that good micro panel data has become available. Here, we wish to mention some of the key directions in which we think the framework, and in particular its empirical implementations, can be developed.
First, we believe that there is much to begained from modelling consumption jointly with other choices such as fertility or education. For example, Browning and Ejrnaes (2000) demonstrate that the life-cycle hump in consumption can be “explained” by the time path of fertility. But this answer begs a further question: why is the time path of income correlated with the time path of fertility? It may be that capacities for earning, fertilityand consumption are driven by the same set of biological and time constraints. For example, students who invest heavily in human capital may also be people who postpone both earnings and fertility. Completion of their training would then be associated with both rising income and fertility-driven consumption growth, a perspective that would be quite consistent with the forward- looking, optimizingbehavior posited by life-cycle models. On the other hand, it may well be that some households postpone fertility until they “have enough money to start a family.” From this perspective, capital market imperfections or a precautionary savings motive may be determining households’ decisions regarding education or fertility. In either case, modeling fertility and education along with consumption may...