Association for Public Policy Analysis & Management

There is significant research evidence indicating that social policies such as Medicaid and cash welfare (AFDC/TANF) do not have large effects on childbearing at the population level (Lopoo and Raissian; Journal of Policy Analysis and Management 2013).  However, it is possible that previous attempts to isolate these effects were hampered by heterogeneity in responsiveness to policy, and in particular, heterogeneity by birth parity.  While the decision to become a parent (i.e., the “extensive margin”) may be largely unaffected by social policy, it is possible that the decision to have a 2nd, 3rd, or 4th child (i.e., the “intensive margin”) is more sensitive to financial considerations – including public benefits. 

In this paper, we revisit the academic literature regarding the large-scale Medicaid coverage expansions to pregnant women – which began in the mid-1980s – to investigate the impacts of increased public health insurance access on birth parity, especially higher-order births.  In particular, we use the cell-based estimation approach employed by DeLeire, Lopoo, and Simon (Demography 2011) as a conceptual starting point for our empirical modeling.  We further incorporate the work of Hamersma and Kim (Journal of Health Economics2013) who argue that using state income eligibility thresholds directly reduces mismeasurement, and thus bias, that could be caused by imputing eligibility.  Through this combination of a more refined measure of access to public health care for pregnant women and an analysis of birth trends by parity, we attempt to better understand the behavioral impacts of an important, large-scale expansion of healthcare access.

Theoretically, any large reduction in the cost of childbearing should increase fertility. Given that the average cost of childbirth in the United States was $4,334 in 1989 (Health Insurance Association of America, 1989) – which is roughly $8,500 in 2016 dollars – public health insurance expansions to pregnant women should greatly reduce the costs of childbirth for the previously uninsured.  At the same time, expansions of the Medicaid program also increased access to family planning services, which may have lowered childbearing rates. Collectively, then, the overall fertility effect is ambiguous. Recent evidence from Aaronson, Lange, and Mazumder (American Economic Review 2014) shows that cost changes can affect the extensive and intensive margins differently. Thus, these pro- and anti-natalist forces from expansions of the Medicaid program may be different depending on the number of children a woman already has.

To disentangle the potential effects at subpopulation levels, we estimate a series of fixed-effects regression equations by a number of different demographic cells.  To incorporate changing norms in birth patterns in the United States – such as delayed fertility or an increase in the number of non-marital births – we examine parity across a number of different demographic cell levels, ranging from state/race/age category to state/race/age category/marital status/educational attainment.  By examining potential responses to Medicaid expansions to pregnant women under this comprehensive approach, we seek to better understand heterogeneous effects which have not yet been explored in the literature.