Low birth weight (LBW) infants experience severe health and developmental difficulties that can impose large social costs. However, estimates of the effect of prenatal visits on the birth weight of the infant range from no effect to a small positive effect. These estimates are somewhat at odds with conventional wisdom that prenatal visits should make one's baby healthier when compared to their absence. This paper applies a test of the main identification assumption in models where the treatment variable takes multiple values and has bunching, known as the dummy test of identification from C. Caetano et al., 2021, in order to detect endogeneity in the prenatal visits model. The dummy test fails to reject in the preferred specification at the no prenatal visits bunching point with an estimated impact of over 26 grams per prenatal visit. However, a different prenatal visit bunching point fails the test, implying further endogeneity.
