Working Paper: NBER ID: w12485
Authors: Adam Ashcraft; Kevin Lang
Abstract: We examine the effect of teenage childbearing on the adult outcomes of a sample of women who gave birth, miscarried or had an abortion as teenagers. If miscarriages are (conditionally) random, then if all miscarriages occur before teenagers can obtain abortions, using the absence of a miscarriage as an instrument for a live birth provides a consistent estimate of the effect of teenage motherhood on women who give birth. If all abortions occur before any miscarriage can occur, OLS on the sample of women who either have a live birth or miscarry provides an unbiased estimate of this effect. Under reasonable assumptions, IV underestimates and OLS overestimates the effect of teenage motherhood on adult outcomes. For a variety of outcomes, the two estimates provide a narrow bound on the effect of teenage motherhood on adult outcomes and which is relatively modest. The bounds can also be combined to provide consistent estimates of the effects of teen motherhood. These effects are generally adverse but modest.
Keywords: teenage childbearing; adult outcomes; instrumental variables; ordinary least squares
JEL Codes: I3; J13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Teenage motherhood (J13) | Educational attainment (I21) |
| Teenage motherhood (J13) | Average educational attainment (I21) |
| Teenage motherhood (J13) | Probability of working (J29) |
| Teenage motherhood (J13) | Probability of being married (J12) |
| Teenage motherhood (J13) | Husband's earnings (J31) |
| Teenage motherhood (J13) | Educational attainment (I21) |