Skip to main content

Something in the pipe: the Flint water crisis and health at birth

Journal of Population Economics (2021)Cite this article

Abstract

In 2014, the city of Flint, MI, in the USA changed its public water source, resulting in severe water contamination and a public health crisis. Using the Flint water crisis as a natural experiment, we estimate the effect of in utero exposure to polluted water on health at birth. Matching vital statistics birth records with various sources of data, we use the synthetic control method to identify the causal impact of water pollution on key birth outcomes. Our results suggest that the crisis modestly increased the rate of low birth weight (LBW) by 1.8 percentage points (or 15.5%) but had little effect on the length of gestation or rate of prematurity. However, these effects are larger among children born to black mothers, as indicated by an increase in the rate of LBW by 2.5 percentage points (or 19%). Children born to white mothers exhibit, on average, a 30.1-g decrease in birth weight. We find little evidence that the male-to-female sex ratio declines in the overall population, suggesting that the in utero scarring effect of the Flint water crisis may dominate the channel of mortality selection. However, we observe a slight decline in the sex ratio among children born to black mothers. Finally, we find no notable change in the fertility rates of either black women or white women in Flint. These results are robust to a rich set of placebo and falsification tests.

This is a preview of subscription content, access via your institution.

Access options

Buy single article

Instant access to the full article PDF.

USD 39.95

Price includes VAT (USA)
Tax calculation will be finalised during checkout.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Notes

  1. Byproducts of chlorinated disinfectants, such as THMs, are associated with negative outcomes such as small for gestational age (SGA), low birth weight, and spontaneous abortion (Gallagher et al. 1998; Waller et al. 1998).

  2. In August and September 2014, coliform bacteria were detected in Flint water, representing a violation of the Safe Drinking Water Act. To eliminate these bacteria, additional chlorine was added; however, this caused another violation of the same Act due to the total level of THMs in the water. In October 2014, the local General Motors plant stopped using Flint tap water because the high concentration of chlorine was corroding engine parts. Ferric chloride was added to reduce the THM levels in the water, resulting in water that was 19 times more corrosive than the Detroit-supplied water. This increased level of corrosivity facilitated the leaching of lead from lead pipes into the drinking water supplied to roughly 40% of Flint homes. In six of the nine city wards, 20 to 32% of homes had tap water with lead levels greater than 15 ppb, a concentration that triggers remedial action under the Safe Drinking Water Act’s Lead and Copper Rule. In some homes, the lead levels exceeded 1000 ppb (90th percentile = 25 ppb) (Bellinger 2016).

  3. For example, if women avoided pregnancy due to their concern over water pollution, the affected infants in our sample would be fewer than the potentially affected infants if there were no pregnancy avoidance. Similarly, if pregnant women drink bottle water instead of tap water, the effect we found would be smaller than the effect without avoidance behaviors.

  4. While Michigan expanded Medicaid over this time period, we find no evidence of increase in the proportion of births covered by Medicaid. The use of prenatal care during this time period decreases. Therefore, Medicaid coverage is unlikely to attenuate our results.

  5. In the matching process, it is common to include outcomes that occurred during part of the pre-treatment period to adjust for heterogeneity in the observed and unobserved factors that affect the outcomes of interest (Abadie et al., 2003; Abadie et al. 2015). However, if we include the outcomes during each pre-treatment period, the effects of other predictors are eliminated with barely apparent weights. Therefore, we only include the birth outcomes corresponding to part of the pre-treatment period in the main analysis but implement a separate robustness check using the health outcomes during all pre-treatment periods.

  6. Flint switched back to Detroit water in October 2015, reflecting government intervention in the crisis. To exclude the effects of various government actions, we drop data recorded after October 2015.

  7. One limitation of our CDC natality data is the use of mothers’ reported residences, rather than geocoded residences, to assign resident cities. Therefore, the reported residences may involve errors. Specifically, our dataset may contain mothers who live in a municipality near Flint but not in the city Flint.

  8. Twin births account for approximately 4% of the sample.

  9. For cities in which SDWIS does not provide related water system information, we first identify all of the water systems in the respective counties using SDWIS and then match the water system to the city through a Google search.

  10. We perform three robustness checks: one includes all possible donor cities, and the other two exclude cities with health-based drinking water violations during the pre- or post-period. Our results are robust.

  11. Births without geographic identifiers occurred either in small cities or towns, which have better average birth outcomes and demographics than big cities.

  12. The polluted water pipes in Flint crisis are mainly lead pipes. The use of lead pipes in construction has been illegal since 1986. For details, see the map created by Marty Kaufman and Troy Rosencrants of the University of Michigan–Flint Geographic Information Systems Center at https://www.fondriest.com/news/amid-flint-water-crisis-gis-effort-maps-citys-pipes.htm.

  13. Although achieving a good balance regarding each of the covariates has the advantage of providing tighter bounds, Botosaru and Ferman (2019) argue that the bias of the SCM estimator remains bounded when pre-treatment outcomes are well balanced, even if the relevant covariates are not perfectly balanced. We thank an anonymous reviewer for this constructive comment, which prompted us to leverage longer pre-treatment outcome lags in robustness checks to improve our estimation procedure when Flint and synthetic Flint are not well matched in some covariates.

  14. One caveat, however, is that these results also rely on the assumption that the unobserved covariates are largely balanced.

  15. We also tested results by maternal education, but Flint and synthetic Flint do not match well in the pre-treatment period for the two subsamples, which prevents us from identifying the effect of the crisis on birth outcomes by maternal education.

  16. This differs from Grossman and Slusky (2019), who find decreases in the sex ratio by 0.7 (1.3%) and 0.9 percentage points (1.8%), respectively. Some possible explanations of the more salient effect of the crisis on sex ratios identified by Grossman and Slusky (2019) compared with our work are as follows. (1) We use a larger set of cities in the donor pool. (2) Grossman and Slusky (2019) focus on mothers who live closer to lead pipes (i.e., larger probability of lead exposure) and have a larger biological effect, while we include all mothers living in Flint. (3) Due to residential selection by socioeconomic status, households closer to the contaminated water source may be poorer and less healthy and have less means to avoid the situation. (4) The intensified exposure among those living closer to contaminated water may increase their stress, even if they are not directly affected. (5) Grossman and Slusky (2019) note that proximity to a polluted water source more likely affected the water in places of employment.

References

  • Abadie A, Gardeazabal J (2003) The economic costs of conflict: a case study of the Basque country. Am Econ Rev 93(1):113–132

    Article  Google Scholar 

  • Abadie A, Diamond A, Hainmueller J (2010) Synthetic control methods for comparative case studies: estimating the effect of California’s tobacco control program. J Am Stat Assoc 105(490):493–505

    Article  Google Scholar 

  • Abadie A, Diamond A, Hainmueller J (2015) Comparative politics and the synthetic control method. American Journal of Political Science 59(2):495–510

    Article  Google Scholar 

  • Abouk R, Adams S (2018) Birth outcomes in Flint in the early stages of the water crisis. J Public Health Policy 39(1):68–85

    Article  Google Scholar 

  • Aizer A, Currie J (2014) The intergenerational transmission of inequality: maternal disadvantage and health at birth. Science 344(6186):856–861

    Article  Google Scholar 

  • Aizer A, Currie J (2019) Lead and juvenile delinquency: new evidence from linked birth, school and juvenile detention records. Rev Econ Stat 101(4):575–587

    Article  Google Scholar 

  • Almond D, Currie J, Duque V (2017). Childhood circumstances and adult outcomes: Act II. NBER Working Paper No. 23017.

  • Bellinger D (2016) Lead contamination in Flint — an abject failure to protect public health. N Engl J Med 374:1101–1103

    Article  Google Scholar 

  • Botosaru I, Ferman B (2019) On the role of covariates in the synthetic control method. Econometrics Journal 22:117–130

    Article  Google Scholar 

  • Centers for Disease Control and Prevention. (2012). Lead in drinking water and human blood lead levels in the United States. Morbidity and Mortality Weekly Report, August 10.

  • Centers for Disease Control and Prevention (2016). Investigation: blood lead levels higher after switch to Flint river water. https://www.cdc.gov/media/releases/2016/p0624-water-lead.html. Accessed 10 Sep 2020

  • Christensen P, Keiser D, Lade GD (2018). The economic effects of environmental crises: evidence from Flint, Michigan. Available at SSRN: https://ssrn.com/abstract=3420526. Accessed 22 Mar 2019

  • Currie J, Neidell M (2004) Air pollution and infant health: what can we learn from California’s recent experience? Quart J Econ 120(3):1003–1030

    Google Scholar 

  • Currie J, Neidell M, Schmieder JF (2009) Air pollution and infant health: lessons from New Jersey. J Health Econ 28:688–703

    Article  Google Scholar 

  • Currie J, Graff Zivin J, Meckel K, Neidell M, Schlenker W (2013) Something in the water: contaminated drinking water and infant health. Can J Econ 46(3):791–810

    Article  Google Scholar 

  • Danagoulian S, Jenkins D (2021) Rolling back the gains: maternal stress undermines pregnancy health after Flint’s water switch. Health Econ 30(3):564–584

    Article  Google Scholar 

  • Dave DM, Yang M. (2020). Lead in drinking water and birth outcomes: a tale of two water treatment plants (No. w27996). National Bureau of Economic Research.

  • Duncan B, Mansour H, Rees DI (2017) It’s just a game: the super bowl and low birth weight. Journal of Human Resource 52(4):946–978

    Article  Google Scholar 

  • Edwards M (2014) Fetal death and reduced birth rates associated with exposure to lead-contaminated drinking water. Environ Sci Technol 48(1):739–746

    Article  Google Scholar 

  • Fung A, Graham M, Well D (2007) Full disclosure: the perils and promise of transparency. Cambridge University Press

    Book  Google Scholar 

  • Gallagher MD, Nuckols JR, Stallones L, Savitz DA (1998) Exposure to trihalomethanes and adverse pregnancy outcomes. Epidemiology 9(5):484–489

    Article  Google Scholar 

  • Grossman DS, Slusky DJ (2019) The impact of the Flint water crisis on fertility. Demography 56(6):2005–2031

    Article  Google Scholar 

  • Hanna-Attisha M, Lachance J, Sadler RC, Champney SA (2016) Elevated blood lead levels in children associated with the Flint drinking water crisis: a spatial analysis of risk and public health response. Am J Public Health 106(2):283–290

    Article  Google Scholar 

  • McClelland R, Sarah G (2017). The synthetic control method as a tool to understand state policy. State and Local Finance Initiative, Urban Institute.

  • Pell MB, Schneyer J (2016). The thousands of U.S. locales where lead poisoning is worse than in Flint. The Reuters. December 19.

  • Sanders HJ, Stoecker C (2015) Where have all the young men gone? using sex ratios to measure fetal death rates. J Health Econ 41:30–45

    Article  Google Scholar 

  • Sauve-Syed J. 2018. Lead exposure and student performance: a study of Flint schools. Unpublished working paper. Syracuse University.

  • Taylor CM, Golding J, Emond AM (2015) Adverse effects of maternal lead levels on birth outcomes in the ALSPAC study: a prospective birth cohort study. BJOG 122(3):322–328

    Article  Google Scholar 

  • Waller K, Swan SH, DeLorenze G, Hopkins B (1998) Trihalomethanes in drinking water and spontaneous abortion. Epidemiology 9:134–140

    Article  Google Scholar 

  • Walsh K, McCormack CA, Webster R, Pinto A, Lee S, Feng TS, Krakovsky HS, O’Grady SM, Tycko B, Champagne FA, Werner EA, Liu G, Monk C (2019) Maternal prenatal stress phenotypes associate with fetal neurodevelopment and birth outcomes. Proc Natl Acad Sci 116(48):23996–24005

    Article  Google Scholar 

  • Zhu M, Fitzgerald EF, Gelberg KH, Lin S, Druschel CM (2010) Maternal low level lead exposure and fetal growth. Environ Health Perspective 118:1471–1475

    Article  Google Scholar 

Download references

Acknowledgements

The authors thank editor Shuaizhang Feng and three anonymous reviewers for helpful comments.

Funding

Financial support from the James Tobin Research Fund at Yale Economics Department, the Yale Macmillan Center faculty research award (2017–2019), the US PEPPER Center Scholar Award (P30AG021342), the National Science Foundation in China (71602149), and a NIH/NIA grant (K01AG053408) is acknowledged.

Rui Wang is a Postdoctoral Research Associate at the National Strategic Planning and Analysis Research Center (NSPARC), Mississippi State University. Xi Chen is an Associate Professor of Health Policy and Economics at Yale University. Xun Li is an Associate Professor at the School of Economics and Management, Center for Health Economics and Management, and Institute for the US and Canadian Economies, Wuhan University. The views expressed herein and any remaining errors are the authors’ and do not represent the views of any official agency. None of the authors have potential conflicts of interests that could bias this work.

Author information

Authors and Affiliations

  1. National Strategic Planning and Analysis Research Center, Mississippi State University, Starkville, USA

    Rui Wang

  2. Department of Health Policy and Management, Yale School of Public Health, New Haven, USA

    Xi Chen

  3. Department of Economics, Yale University, New Haven, USA

    Xi Chen

  4. School of Economics and Management & Center for Health Economics and Management, Wuhan University, Wuhan, China

    Xun Li

Authors
  1. Rui Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Xi Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Xun Li
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Xi Chen or Xun Li.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Responsible Editor: Shuaizhang Feng

Appendix

Appendix

Fig. 7
figure 7

Timeline of the Flint water crisis

Full size image
Fig. 8
figure 8

The distributions of birth weight before and after the Flint water crisis. Notes: Four cities are those having positive weights using the synthetic control method (SCM). The vertical red lines correspond to birth weight being 2500 g

Full size image
Fig. 9
figure 9

Trends and SCM estimates on birth outcomes for those born to white mothers. Notes: The X-axis represents birth date relative to the Flint water crisis. The data points are measured in half-years prior to and after the water switch. The red vertical line (May 1, 2014) marks the beginning of the water switch

Full size image
Table 5 Summary statistics for newborns in Flint and other 162 American cities
Full size table

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Wang, R., Chen, X. & Li, X. Something in the pipe: the Flint water crisis and health at birth. J Popul Econ (2021). https://doi.org/10.1007/s00148-021-00876-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s00148-021-00876-9

Keywords

  • Water pollution
  • Lead exposure
  • Flint water crisis
  • Infants
  • Low birth weight

JEL classification

  • I14
  • I18
  • Q53
  • Q58