Previous studies of U.S. publicly funded family planning services have produced conflicting and sometimes confounding results. These studies have relied exclusively on single-equation estimates of family planning program effectiveness. Economic theories suggest that single-equation estimates may understate program effectiveness when the same unobserved variable affects both the fertility outcome and contraceptive behavior. To eliminate the bias that may result from single-equation estimation, I use a bivariate probit model to estimate the effect of contraceptive acceptance on the individual's probability of giving birth. I employ a unique data set created from Maryland Medicaid claims records. Results from bivariate probit estimation show that contraceptive acceptance plays a much larger role in reducing fertility than single-equation estimates would suggest.

The author is assistant professor of economics at The College of William and Mary, Williamsburg, Virginia She would like to thank William N. Evans, Edward Montgomery, Aaron Yelowitz, and two anonymous referees for a number of helpful comments. This research was supported by grant number HS08889-01 from the Agency for Health Care Policy and Research. The opinions expressed here are those of the author and do not necessarily reflect those of the granting institution. The author also acknowledges fellowship support during the completion of this research from the Robert Wood Johnson Health Policy Scholars Program at Yale University. The data used in this article were obtained from the Policy and Health Statistics Administration of the Maryland Department of Health and Mental Hygiene in Baltimore, MD.