Academic journal article Demographic Research

The Early Socioeconomic Effects of Teenage Childbearing: A Propensity Score Matching Approach

Academic journal article Demographic Research

The Early Socioeconomic Effects of Teenage Childbearing: A Propensity Score Matching Approach

Article excerpt


A large body of literature has documented a negative correlation between teenage childbearing and teen mothers' socioeconomic outcomes, yet researchers continue to disagree as to whether the association represents a true causal effect. This article extends the extant literature by employing propensity score matching with a sensitivity analysis using Rosenbaum bounds. The analysis of recent cohort data from the National Longitudinal Study of Adolescent Health shows that (1) teenage childbearing has modest but significant negative effects on early socioeconomic outcomes and (2) unobserved covariates would have to be more powerful than known covariates to nullify the propensity score matching estimates. The author concludes by suggesting that more research should be done to address unobserved heterogeneity and the long-term effects of teenage childbearing for this young cohort.

(ProQuest: ... denotes formula omitted.)

1. Introduction

The longstanding literature on teen motherhood has documented its detrimental life cycle consequences. Teen mothers tend to have worse socioeconomic outcomes than other women who delay childbearing (An, Haveman, and Wolfe 1993; Hofferth and Hayes 1987). Despite this evidence, it is still unclear whether these negative outcomes among teen mothers result from the incidence of childbearing per se or from the socioeconomic disadvantages these women faced before they became teen mothers. While human capital theory holds that teenage childbearing has a real causal effect on socioeconomic outcomes because it directly interferes with adolescents' investment in human capital (Becker 1993), the selection view contends that teenage childbearing is associated with negative outcomes because it occurs mostly among disadvantaged female adolescents (Geronimus, Korenman, and Hillemeier 1994). Indeed, the concern about selection bias points out that isolating the effect of teen motherhood creates a considerable methodological challenge (Winship and Mare 1992; Winship and Morgan 1999): If both observed and unobserved preexisting characteristics of teen mothers account for the relationship between teenage childbearing and its socioeconomic consequences, assertions of causality become questionable.

A large body of research has addressed the selection bias problem by finding better comparison groups for women who give birth in their teens (Cherlin 2001; Hoffman 1998; Korenman, Kaestner, and Joyce 2001; Wu and Wolfe 2001). For example, within-family fixed-effects models compare teen mothers with their sisters who gave birth after their teenage years to control for unobserved family-level heterogeneity (Geronimus and Korenman 1992). Quasi-natural experimental approaches approximate randomization procedures by treating twin births or miscarriages as comparison cases (Grogger and Bronars 1993; Hotz, McElroy, and Sanders 1997). Finally, instrumental variables methods utilize variables that capture the exogenous component of teenage childbearing to mitigate the selection bias problem (Klepinger, Lundberg, and Plotnick 1999). Despite their intuitive appeal, all of these models have their own drawbacks. It is not uncommon to find that they are grounded on somewhat strong assumptions and/or unrepresentative samples, with mixed results at best. Thus, evaluating the "true" effects of teenage childbearing remains an elusive goal.

In this article, I extend the previous literature in three distinct ways. First, I use a propensity score matching approach to identify the early socioeconomic effects of teenage childbearing. Following the counterfactual framework (Rosenbaum and Rubin 1983; Rubin 1977), this approach matches teen mothers ("treatment" group) to those who are not teen mothers but similar in all other preexisting observed characteristics ("control" group), based on a propensity to give birth as teens. Then it compares various socioeconomic outcomes between these two groups using semi-parametric estimators. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.