American fertility rates vary considerably across time and space. The relative importance of public policies - welfare benefit levels, restrictions on the availability of abortion, restrictions on Medicaid funding of abortion, personal deduction and child care deductions in the federal and state tax codes local labor market conditions, and local marriage market conditions in causing this variation remains controversial. Recent work has used cross-sectional variation (either from simple cross- sectional surveys such as the Census or through short panel datasets where most of the variation is cross-sectional). Implicitly, this assumes that variation in public policies across states is exogenous. It is possible to use time-series of cross-section data and fixed effect models to control for time-invariant inter-state heterogeneity. Recent work by Moffitt and ourselves finds that doing so attenuates and often reverses the direction of the effect of public policies on demographic outcomes. Similarly, in identifying the effect of economic/employment opportunities on fertility, it is important to capture the effect of exogenous shifts in economic opportunities. Standard practice uses cross-sectional variation in earnings. Such variation is primarily due to differences in abilities and tastes rather than economic opportunities. Instead, measures of economic opportunity should be focused on identifying exogenous variation in the opportunities facing men and women. To address this critique, we propose to construct fertility rates using National Center for Health Statistics (NCHS) birth certificate data. These data provide a long time-series (1970-1988) on births disaggregated to the county. We will impute exogenous shifts in earnings opportunities using a combination of Census, Panel Study of Income Dynamics, and County Employment data. The proposal describes appropriate statistical techniques for imputing wages to the birth certificate data and then for analyzing the resulting cross-section of time-series of age and race specific county level fertility rates. Using this dataset, we will reexamine the existing literature on the determinants of American fertility. We will consider four sets of models - teenage fertility, teenage non-marital fertility, adult fertility, and adult non-marital fertility. For each sub-group-rate, we will consider the effect of public policies (listed in the first paragraph), economic conditions for men and women (long-term economic growth, short-term economic fluctuations), and other demographic considerations (e.g., marriage squeeze) on fertility.
| Leibowitz, A; Klerman, J A (1995) Explaining changes in married mothers' employment over time. Demography 32:365-78 |
