Most people take it for granted that teen pregnancy causes poor outcomes. However, how do you untangle whether teen pregnancy causes bad economic outcomes or is merely a leading indicator? Arline T. Geronimus and Sanders Korenman approached the issue by looking at families that have more than one girl. Some girls will become teen mothers and yet have sisters who do not. If teen pregnancy causes poor outcomes, then these girls should do worse than their sisters who are not teen mothers. However, the sisters of teen mothers frequently had poor life outcomes themselves. This suggested that many pregnant teens were headed towards poor outcomes even before their pregnancy. Comparing sisters probably overstates the impact of teen pregnancy because sisters in the same family are not themselves the same. For a moment, think of family sitcoms. Maximizing comedic possibilities requires that one child is studious and the other just wants to spend all of her time at the mall. It is usually the intelligent child who shows more prudence in her sex life. In real life, kids with better life prospects also have a greater incentive to be sexually prudent.
Obviously, the Geronimus and Korenman paper did little to change public perception. In the New York Times, it was stated:
This does not imply that teen-age motherhood is any less of a catastrophe: Something is deeply wrong, it is safe to assume, when 15-year-old girls would rather change diapers than get on with the adventure of growing up.
Yet, if teen pregnancy causes fewer negative outcomes than you thought, then this does make it less of a catastrophe. Presumably, negative outcomes are how you rank various degrees of catastrophe.
Further evidence that teenage pregnancy did not cause catastrophe came from a natural experiment. Some teenage girls get pregnant but then have miscarriages. For the purposes of research, miscarriages are similar to some girls being randomly assigned not to have a teen pregnancy. If teenage pregnancy causes poor outcomes, then girls who have a miscarriage should do better than girls who give birth. Admittedly, miscarriages are not completely random; for example, smoking and drinking during pregnancy increases their likelihood. For the purpose at hand, however, miscarriages seemed to be random enough.
Initial evidence using miscarriages presented a mostly benign view of teen pregnancy. For example, teenage pregnancy reduced the rate of high school graduation, but it increased the likelihood of getting a GED by a similar amount. Teenage motherhood caused a substantial increase in wages during the mother's 20s. This isn't totally surprising. Older children are less of an impediment to working, and teenage motherhood does more to change the timing of children than the overall number of children. This increase in working hours during the mother's 20s decreased the overall need for public assistance. For high-risk girls, delayed pregnancy didn't eliminate the use of public assistance; it only postponed it.
However, the above results paint too positive a view of teenage pregnancy. Why? Because girls who miscarry and girls who give birth are different in regards to abortion choices. Some girls who miscarry would have had abortions instead of giving birth. There are reasons to believe that pregnant girls with better life prospects are more likely to get abortions. A recent paper in the Economic Journal adjusts for this phenomenon. It concludes that teenage motherhood does lead to worse outcomes. But the impact is modest. Kevin Lang writes:
The most consistent results are that the teen mothers are a few percentage points less likely to be married and more likely to be divorced, and a few percentage points less likely to have either a high school diploma, or to have passed a high school equivalency exam.