Impact of state children’s health insurance program on fertility of immigrant women

ABSTRACT

Between 1997 and 2000, all states in the United States (US) enacted the State Children’s Health Insurance Programme (SCHIP) to provide publicly funded health insurance coverage for children in low-income families. However, only 15 states including the District of Columbia initially chose to provide coverage for children of newly arrived immigrants in their SCHIP. We exploit the resulting state and time variation in the implementation of the programme in a difference-in-differences framework to estimate the effect of a publicly funded children’s health insurance benefit on immigrant women’s fertility. While estimates from full samples show that the net effect of the programme was indistinguishable from zero, we find a significant positive effect on the fertility of unmarried immigrant women, both at the extensive and at the intensive margin. Our findings have important policy implications for societies experiencing a persistent decline in fertility.

KEYWORDS:

• State children’s health insurance program
• fertility
• immigrant women
• quantity-quality tradeoff

JEL CLASSIFICATION:

• I18
• J13

Compliance with ethical standards

We hereby declare that this project did not receive specific grant from funding agencies in the public, commercial, or not-for-profit sectors. We also declare that this study does not involve any financial or personal conflict of interest.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ See Conti and Ginja (2020) for recent evidence on impacts of expanding a publicly funded health insurance program.

² For details, see Fix and Passel (1999).

³ However, US-borne children of immigrants are exempt from the five-year ban.

⁴ Due to the introduction of substantial welfare reforms by the PRWORA, the pre-1997 era represents a distinct welfare regime.

⁵ We do not control for ‘work status’ as women may often make labour supply decisions in tandem with fertility decisions. Additionally, whether women’s employment is directly impacted by generous CHIP is a pertinent question for our analysis. Therefore, we further estimate the impact of generous SCHIP on employment outcomes using the same DD framework. Results shown in Table A6indicate that employment is not an important channel through which SCHIP affects fertility.

⁶ For robustness check, we also analyse a smaller sample of women aged 17–40 where most births are concentrated. See Figure A1.

⁷ To empirically test if women select into marital relationships following the CHIP intervention, in unreported regressions, we look at the likelihood of a woman being in a marital relationship. We do not find any significant effect in the broad sample of women including immigrants and native born and also for the specific sample of immigrant women only.

⁸ Generous states include Alaska, California, Delaware, District of Columbia, Hawaii, Illinois, Massachusetts, Minnesota, Nebraska, New Jersey, New Mexico, New York, Pennsylvania, Virginia and Washington (Rosenbach et al. 2001; Olds 2016)

⁹ We also test the effect of CHIP on fertility of all women (immigrant and native) and do not find any statistically significant effect of the program. Results are available upon request.

¹⁰ Note that we estimate the Intention-to-Treat (ITT) effects rather than average treatment effects.

¹¹ Since the event study shows significant positive impact on unmarried immigrant women beginning one year after implementation, we re-estimate our DD models with policy lagged by one year to allow for an extended period between policy announcement and take-up. Our findings remain qualitatively similar to the main specification (see Table A5).

¹² See Tschoepe and Hindera (2001) for details on these changes.

¹³ Millimet and Wang (2011) also include children’s sex ratio in their model assuming that having more children belonging to the same sex can be provide certain cost advantages to households (discussed later). However, to provide a basic understanding of the quantity-quality trade-off, it is not required to account for sex ratio in the main model.