How Much Does Where You’re Born Affect Your Health at Birth?

This study investigates how much a child’s birth location causally affects birth weight and early-life health. Using 30 years of California birth records and a within-mother, mover-based design, the authors isolate the effects of place by comparing siblings born before and after residential moves. They find that moving from a below- to an above-median Zip code improves birth weight by 18 grams—comparable to Medicaid access or smoking cessation. Place effects are strongest for long-distance moves and for mothers without college degrees. Pollution, especially ozone, is the strongest predictor of these effects, suggesting environmental quality as a key mechanism.

Full Citation and Link to Article

Chyn, E., & Shenhav, N. (2024). Place Effects and Geographic Inequality in Health at Birth. American Economic Journal: Applied Economics. Advance online publication. https://doi.org/10.1257/app.20220252

Extended Summary

Central Research Question

This paper addresses a fundamental question in economics and public health: To what extent does the location in which a child is born causally affect their health at birth, particularly as measured by birth weight? While geographic disparities in infant health outcomes are widely observed across the United States, it remains unclear how much of this variation is due to place-based environmental or contextual factors versus family characteristics and self-selection. Chyn and Shenhav use a within-family, movers-based research design to estimate the causal impact of place on early-life health and decompose the sources of geographic inequality in birth outcomes.

Previous Literature

The study contributes to several streams of literature. Prior research has established that birth weight is a strong predictor of later-life outcomes such as educational attainment, health, and earnings (Black, Devereux, & Salvanes, 2007; Almond & Currie, 2011). Other work has shown that environmental factors, including pollution and climate, can influence early-life health (Chay & Greenstone, 2003; Currie et al., 2009). However, most existing studies focus on the effects of specific exposures (e.g., PM2.5, temperature) or examine aggregate area-level changes, making it difficult to isolate the total effect of place. More recent work in the economics of neighborhoods has documented the influence of residential location on longer-run outcomes (Chetty et al., 2016; Chetty & Hendren, 2018), though evidence on place effects at birth remains limited. This paper fills that gap by using sibling comparisons among mothers who move, allowing for a cleaner identification of place-based effects on birth outcomes while holding family characteristics constant.

Data

The authors use confidential birth records from the state of California spanning 1989 to 2017, comprising approximately 15.6 million births. The dataset includes detailed information on birth weight, gestational age, maternal demographics (age, education, race, insurance status), father characteristics (when available), and exact residential location at the time of birth, coded at the Zip code (ZIP-5) level. The analytic sample focuses on 3.3 million mothers who had two to four births during the observation window, allowing for within-mother comparisons. Of these, about 47 percent changed Zip codes exactly once during the period, providing a large sample of movers. The authors restrict the sample to Zip codes with at least 25 movers to ensure stable estimates. Additional variables were drawn from Census data, environmental quality indices (e.g., ozone and PM2.5 levels), and healthcare access measures.

Methods

The empirical strategy uses a fixed-effects regression model that compares siblings born to the same mother before and after she moved to a different Zip code. This design controls for time-invariant maternal characteristics, isolating the effect of the new birth location. The key parameter of interest is the location fixed effect (γj), which captures the average impact of a specific Zip code on birth weight. The authors supplement this model with event studies that check for pre-trends and other robustness tests. Additionally, they conduct variance and additive decompositions to quantify the share of geographic disparities in birth weight attributable to place versus maternal characteristics. To address potential endogeneity in moving decisions, the authors implement instrumental variables (IV) strategies and focus on quasi-random moves among military families.

Findings/Size Effects

The authors find clear evidence that birth location has a causal effect on infant health. Their main estimate shows that moving from a below- to an above-median Zip code results in an 18-gram increase in birth weight. This effect is comparable to or larger than the estimated impacts of maternal Medicaid eligibility (East et al., 2021) and reductions in prenatal smoking (Permutt & Hebel, 1989). The size of the place effect grows with the distance of the move: a county-level move from a below- to an above-median area yields a 26-gram increase in birth weight, a 44 percent larger effect than Zip-level moves.

Despite the statistically significant role of place, the authors find that place-based effects explain only a minority of the overall spatial inequality in birth weight. In a decomposition of the 109-gram difference in birth weight between above- and below-median areas, they attribute 17 percent to place effects and 83 percent to maternal characteristics. Similar patterns emerge when decomposing differences across quartiles and deciles, as well as in variance decomposition: equalizing place would eliminate 5 percent of the cross-Zip variance, while equalizing maternal characteristics would eliminate 79 percent.

Importantly, place effects are heterogeneous. Mothers without a college degree experience larger gains from moving—about 22 grams compared to just 7 grams for college-educated mothers. Accordingly, place explains about 20 percent of the birth weight gap for non-college-educated mothers but only 6 percent for their more educated counterparts. Effects are similar across White and Hispanic mothers, indicating that socioeconomic status rather than ethnicity drives heterogeneity.

Mechanism analyses show that pollution, especially ozone levels, is the strongest predictor of a Zip code’s estimated place effect. The correlation between local ozone levels and estimated place effects is about −0.33. Other significant, though smaller, correlates include physician and OB-GYN density, insurance coverage rates, and adult smoking rates. Interestingly, some social determinants like violent crime and maternal education levels are only weakly associated with place effects.

Robustness checks suggest that the estimated effects are not driven by endogenous moving or fertility decisions. Event study analyses show no pre-trends in birth weight prior to moves, supporting the assumption that changes in location precede changes in outcomes. Maternal characteristics such as insurance status and paternal presence change little after moves, explaining less than 3 percent of the observed birth weight improvements. An IV strategy using average destination quality of other movers from the same origin confirms the main results. Military moves, assumed to be quasi-random, yield even larger effects on birth weight.

Lastly, the authors estimate that the projected effect of an 18-gram increase in birth weight translates into a 0.31 percent of a standard deviation increase in test scores and a 0.06 percent increase in earnings. Thus, while location at birth has a measurable effect on long-run outcomes, the magnitude is modest, suggesting that most neighborhood effects on intergenerational mobility occur postnatally.

Conclusion

Chyn and Shenhav provide the first mover-based causal estimates of how geographic location affects infant health outcomes at birth. Their results show that place matters: residential context, especially environmental factors like air pollution, directly influences early-life health. However, place-based differences explain only a small share—approximately one-sixth—of the geographic inequality in birth weight. Most of the disparity arises from differences in maternal characteristics. The findings imply that efforts to improve early-life health may benefit from targeting both environmental quality and support for disadvantaged mothers. Furthermore, the small projected impact of birth location on long-run outcomes suggests that the well-documented importance of neighborhoods for economic mobility primarily operates through post-birth exposures. This work enhances our understanding of place-based disadvantage and offers new insights for public health and housing policy.