The Impact of Abortion Laws on Maternal Mortality Rates: What is the Evidence?

Paul Benjamin Linton, Esq.

This is Issue 27 of the American Reports Series. 

Executive Summary

• Advocates of legal abortion often claim that permissive state abortion laws are associated with lower maternal mortality rates than restrictive state abortion laws.
• A comparison of state maternal mortality rate changes over three time periods preceding and immediately after Roe v. Wade, however, does not show major differences between the states with permissive abortions laws versus states with restrictive laws.
• The only directly relevant real-world comparison available, therefore, fails to prove, or even suggest, that restrictive abortion laws are associated with higher maternal mortality rates than permissive laws.

Are permissive abortion laws associated with lower maternal mortality rates than restrictive laws? Advocates of legal abortion have often claimed that the answer to this question is yes.[1] Overall comparisons of the maternal mortality data for states that adopted permissive abortion laws prior to the Supreme Court’s decision in Roe v. Wade,[2] and those that had retained restrictive abortion laws, however, do not appear to show major differences in maternal mortality rates between restrictive and permissive states.[3] One comparison examines the decline in maternal mortality rates between the three-year period 1964–1966 and the three-year period 1970–1972 (before Roe was decided); the other examines the decline in maternal mortality rates between the three-year period 1970–1972 (before Roe) and the three-year period 1973–1975 (after Roe).

It should be noted, however, that although some studies have found associations between more restrictive legislative and/or regulatory state environments and higher maternal mortality rates, such studies are not as relevant as the evidence addressed in this paper for several reasons. First, because they cover pre-Dobbs periods during which substantial abortion restrictions either were not in effect or did not exist, thus limiting their capacity to even test for correlational, let alone causal, connections between abortion restrictions and maternal mortality. Second, because insofar as these studies analyzed policies directly related to abortion, such policies were primarily regulatory rather than prohibitive in nature. And third, because for the most part such studies find that a variety of independent factors in no way directly related to abortion—such as healthcare access, socioeconomic conditions, and family leave policies—were likely playing a significant, and more determinative, causal role in maternal mortality. Although the analysis set forth in this article also cannot control for a variety of confounding factors, it is nevertheless of interest as a more targeted comparison of pre- and post-Roe abortion laws, restrictive or permissive, and thus contributes to the existing literature on the issue.

Before discussing these data, some legal history is necessary to provide context. Prior to 1967, the overwhelming majority of states prohibited abortion except to save the life of the mother.[4] A small number of states allowed abortion for somewhat broader reasons,[5] but most did not. Beginning in 1967 and ending in 1970, however, sixteen states amended their abortion laws to allow abortion either on demand until late in pregnancy (four states),[6] or for the broadly stated reasons set forth in Section 230.3(2) of the Model Penal Code (twelve states),[7] allowing abortion to prevent grave impairment to “the physical or mental health of the mother,” where “the child would be born with grave physical or mental defect” or to end a pregnancy that “resulted from rape, incest, or other felonious intercourse.”[8] With the exception of Florida, which enacted an abortion law based on the Model Penal Code in 1972 after the Florida Supreme Court declared the state’s life-of-the-mother law unconstitutional in State v. Barquet,[9] no other state enacted a permissive abortion law between 1967 and the end of 1972.[10] What was the impact of the adoption of these permissive abortion laws on the incidence of maternal mortality?

Both before and after Roe v. Wade was decided, the United States Department of Health, Education & Welfare (later split into two cabinet-level departments, the Department of Health & Human Services and the Department of Education) published an annual vital statistics report which included both annual data on the number of live births by state and a moving three-year average of maternal mortality data for all fifty states.[11] The report for 1966 covered the three-year period 1964 through 1966, before any of the statutory changes mentioned in the preceding paragraph. The report for 1972 covered the three-year period 1970 through 1972. With the exception of Florida, no state enacted a permissive abortion law after the end of 1970 (and before Roe v. Wade was decided). What do the data show?

The three-year average of maternal mortality (for the entire country) for the period 1964–1966 was 31.4 deaths per 100,000 live births.[12] The three-year average for the period 1970–1972, six years later, had dropped to 19.8 deaths per 100,000 live births, a reduction of 11.6 maternal deaths per 100,000 live births.[13] For the sixteen states that had adopted permissive abortion laws between the beginning of 1967 (after the end of the first three-year period) and the end of 1970 (the end of the first year of the second three-year period), together with Texas and Wisconsin,[14] the maternal mortality rate declined from 33.0 maternal deaths per 100,000 live births during the 1964-1966 period to 20.0 maternal deaths per 100,000 live births during the 1970-1972 period, a reduction of 13.0 maternal deaths per 100,000 live births.[15]

For twenty-nine states that had retained their restrictive laws, the maternal mortality rate declined from 30.1 during the 1964-1966 period to 19.3 during the 1970-1972 period, a reduction of 10.8 maternal deaths per 100,000 live births.[16] By the end of 1972, the maternal mortality rate for the twenty-nine states that had retained their restrictive laws (19.3) was actually marginally less (albeit not significantly less) than the rate for the sixteen permissive states that had adopted permissive laws by the end of 1970, along with Texas and Wisconsin (20.0).[17] Because there was no major difference either in the change in the maternal mortality rate or in the final rates between the two categories of states, the pre-Roe maternal mortality data provide no evidence for the hypothesis that the enactment of permissive abortion laws (or the invalidation of restrictive laws) had a positive impact on maternal mortality rates compared to the retention of restrictive laws.

There is yet another comparison that sheds light on the relationship between restrictive or permissive abortion laws and changes in maternal mortality rates. That comparison examines the changes in maternal mortality rates from the three-year period preceding Roe v. Wade to the three-year period succeeding Roe. Given the argument that (all other factors being equal) permissive abortion laws will be associated with lower maternal mortality rates than restrictive laws, one might expect that there would have been a much greater reduction in maternal mortality rates in those states transitioning from restrictive pre-Roe laws to virtually no restrictions (as a result of Roe), than in those states transitioning from permissive pre-Roe laws to virtually no restrictions (again, as the result of Roe). But that is not what the data show.

During the three-year period covered by the national vital statistics report for 1975 (1973 through 1975), abortion was legal throughout the country for any reason before viability, and for virtually any reason after viability as a result of the Court’s January 22, 1973, decision. Overall, the maternal mortality rate declined from 19.8 for the three-year period 1970-1972 to 14.2 for the three-year period 1973-1975, a reduction of 5.6 maternal deaths per 100,000 live births.[18]

The maternal mortality rate for the sixteen states that had enacted permissive abortion laws between the beginning of 1967 and the end of 1970, together with Texas and Wisconsin, whose laws were not enforceable during 1971 and 1972, declined from 20.0 maternal deaths per 100,000 live births during the 1970-1972 period, to 15.3 during the 1973-1975 period, a reduction of 4.7 maternal deaths per 100,000 live births.[19] The maternal mortality rate for the  states that had restrictive laws in effect before Roe declined from 19.3 maternal deaths per 100,000 live births during the first three-year period to 13.7 during the second,[20] a reduction of 5.6 maternal deaths per 100,000 live births. The difference in the reduction of the number of maternal deaths between states with permissive laws and states with restrictive laws was less than one maternal death per 100,000 live births.[21] As in the previous comparison, this belies the idea that the kind of laws on the books before Roe was decided—permissive or restrictive—played a major role in protecting the lives of pregnant women by reducing maternal mortality.

Conclusion

Whether, and to what extent, changes in maternal mortality rates will be affected by the enactment and enforcement of restrictive abortion laws following the Supreme Court’s decision in Dobbs remains to be seen. However, the only directly relevant real-world comparison available fails to prove, or even suggest, that restrictive abortion laws are associated with higher maternal mortality rates than permissive laws.[22]

Paul Linton, Esq. is an attorney in private practice in Illinois who specializes in state and federal constitutional appellate litigation and legislative consulting. Mr. Linton, the former General Counsel for Americans United for Life, has submitted amicus curiae briefs in several Supreme Court abortion cases, including Planned Parenthood v. Casey (1992) and Dobbs v. Jackson Women’s Health Organization (2022). Mr. Linton has published more than two dozen law review articles on a variety of topics, as well as the only comprehensive analysis of abortion as a state constitutional right, Abortion under State Constitutions (3d ed. 2020) (Carolina Academic Press). Mr. Linton received his law degree from Loyola University of Chicago.

The author wishes to express his appreciation to the staff of the Charlotte Lozier Institute and, in particular, Tessa Cox, Senior Research Associate, for their review of the article and their assistance in obtaining the live birth and maternal death data set forth in the tables accompanying this article, and Joseph Pojman, Ph.D., Executive Director, Texas Alliance for Life, for his review of the article and his helpful editorial suggestions.

Table 1

Live Births & Maternal Deaths (States with Permissive Abortion Laws, Together with Texas and Wisconsin) (18 States) (1964–1966)

Total live births for three-year period 1964–1966 = 5,003,238 (18 States).

Total maternal deaths for three-year period 1964–1966 = 1,650 (18 States).

Maternal mortality rate for three-year period 1964–1966 = 33.0 maternal deaths per 100,000 live births (18 States).

Table 2

Live Births & Maternal Deaths (States with Permissive Abortion Laws, Together with Texas and Wisconsin) (18 States) (1970–1972)

Total live births for three-year period 1970–1972 = 4,348,811 (18 States).

Total maternal deaths for three-year period 1970–1972 = 870 (18 States).

Maternal mortality rate for three-year period 1970–1972 = 20.0 maternal deaths per 100,000 live births (18 States).

Table 3

Live Births & Maternal Deaths (States That, Between 1967 and 1970, Had Enacted Permissive Abortion Laws, Together with Texas and Wisconsin) (18 States) (1973–1975)

Total live births for three-year period 1973–1975 = 4,208,732 (18 States).

Total maternal deaths for three-year period 1973–1975 = 643 (18 States).

Maternal mortality rate for three-year period 1973–1975 = 15.3 maternal deaths per 100,000 live births (18 States).

Table 4

Live Births & Maternal Deaths (States with Restrictive Abortion Laws) (29 States) (1964-1966)

Total live births for three-year period 1964–1966 = 5,201,202 (29 States).

Total maternal deaths for three-year period 1964–1966 = 1,565 (29 States).

Maternal mortality rate for three-year period 1964–1966 = 30.1 maternal deaths per 100,000 live births (29 States).

Table 5

Live Births & Maternal Deaths (States with Restrictive Abortion Laws) (29 States) (1970-1972)

Total live births for three-year period 1970–1972 = 4,598,702 (29 States) (data for Connecticut, Florida, New Jersey, and Vermont are for 1970–1971 only).

Total maternal deaths for three-year period 1970–1972 = 888 (29 States).

Maternal mortality rate for three-year period 1970–1972 = 19.3 maternal deaths per 100,000 live births (29 States).

Table 6

Live Births & Maternal Deaths (States That Had Restrictive Abortion Laws in Effect at All Relevant Times Before Roe v. Wade was Decided) (29 States) (1973-1975)

Total live births for three-year period 1973–1975 = 4,360,244 (29 States).

Total maternal deaths for three-year period 1973–1975 = 598 (29 States).

Maternal mortality rate for three-year period 1973–1975 = 13.7 maternal deaths per 100,000 live births (29 States).

Summary Tables

^*All state abortion laws were restrictive before 1967, except for the three states not included in the analysis—Alabama, Massachusetts, and Pennsylvania. See footnote 16, supra.

^*As of January 22, 1973, the date on which the Supreme Court decided Roe v. Wade, 410 U.S. 113 (1973), no state could prohibit abortions performed before viability. Thus, none of the restrictive laws was enforceable as of that date.

[1] See, e.g., Harper LM, Leach JM, Robbins L, Blanchard C, Metz TD, Mazzoni S, Nash E, Szychowski J. All-Cause Mortality in Reproductive-Aged Females by State: An Analysis of the Effects of Abortion Legislation. Obstet Gynecol. 2023 Feb 1;141(2):236-242. doi: 10.1097/AOG.0000000000005035; Sherajum Monira Farin, et al., The Impact of Legal Abortion on Maternal Mortality, IZA Institute of Labor Economics Discussion Paper No. 15657 (Oct. 2022).

A detailed critique of the claims made in these articles, neither of which addresses the data set forth herein, lies outside the focus of the present article. A few observations, however, are in order. The Harper article purported to find “the presence of trigger laws [i.e., laws prohibiting abortion that were drafted to go into effect upon the overruling of Roe v. Wade] was significantly associated with … maternal mortality,” even though “none of these laws were actually in effect at the time of this study,” and, therefore, “the trigger laws prevented no abortions from actually occurring during this study period.” 141 Obstetrics & Gynecology at 248 (emphasis added). The Farin article, referring to data showing that there was a reduction in maternal mortality for non-white women when abortion was legalized in certain states, attributes the reduction “to a change in abortion-related mortality, with other causes of maternal death failing to decline.” Farin at 3-4 (emphasis added). See also, id. at 31 (“during the period of our study, legal abortion acts primarily through lower abortion-related deaths, rather than a change in pregnancy-related risk factors”) (emphasis added). However, nothing in the Farin article claims, much less proves, that restrictive abortion laws, in comparison to permissive abortion laws, adversely affected maternal mortality rates by preventing pregnant women from obtaining abortions necessary to save their lives. Finally, as the data set forth in this article reveal, maternal mortality rates declined both in states with restrictive laws and in those with permissive laws, without any major differences in mortality rates between them.

[2] 410 U.S. 113 (1973).

[3] Given the short amount of time during which states have been able to enforce abortion prohibitions with exceptions since the Supreme Court decided Dobbs v. Jackson Women’s Health Organization, 142 S. Ct. 2228 (2022), there are no data available yet comparing pre-Dobbs and post-Dobbs maternal mortality rates.

[4] See Roe, 410 U.S. at 139 (noting that, by the end of the 1950s, “a large majority of the jurisdictions banned abortion, however and whenever performed, unless done to save or preserve the life of the mother”).

[5] Alabama, by statute, 1951 Ala. Acts, p. 1630, codified as Ala. Code § 14-9 (1958), and Massachusetts, by court decisions going back to the 1940s, see Commonwealth v. Wheeler, 53 N.E.2d 4, 5 (Mass. 1944), allowed abortion for undefined (in the case of Alabama) or broadly defined (in the case of Massachusetts) reasons of health (physical or mental health). Mississippi added a rape exception in 1966, 1966 Miss. Laws, ch. 358, § 1, codified as Miss. Code § 97-3-3 (1972). Both Colorado and New Mexico allowed abortion “to preserve [or “save”] the life of the mother or to prevent serious and permanent bodily injury.” Colo. Rev. Stat. Ann. § 40-2-23 (1953), N.M. Stat. Ann. § 40-3-2 (1955). Maryland permitted a licensed physician to perform an abortion “if no other method [of treatment] [would] secure the safety of the mother.” Md. Ann. Code art. 27, § 3 (1957). In State v. Buck, 262 P.2d 495 (Or. 1953), the Oregon Supreme Court construed the Medical Practice Act and the Criminal Abortion Act in pari materia to allow a licensed physician to perform an abortion for “the relief of a woman whose health appears in peril because of her pregnant condition.”

[6] Alaska, Hawaii, New York, and Washington all adopted such laws in 1970.

[7] The text of Section 230.3 of the Model Penal Code, drafted by the American Law Institute, may be found in Appendix B to the Supreme Court’s decision in Doe v. Bolton, 410 U.S. 179, 205–06 (1973), the companion case to Roe v. Wade, 410 U.S. 113 (1973).

[8] Model Penal Code, § 230.3(2) (“Justifiable Abortion”). Arkansas, California, Colorado, Delaware, Georgia, Kansas, Maryland, New Mexico, North Carolina, Oregon, South Carolina, and Virginia all enacted one version or another of the Model Penal Code provision on abortion between 1967 and 1970. A brief description of the abortion law in effect in each state at the time Roe v. Wade was decided (including statutory citations) may be found in the author’s book, Abortion under State Constitutions (3d ed. 2020) (Carolina Academic Press).

Three of these states—Arkansas, Georgia, and North Carolina—included an undefined exception for “health” that was not limited to the pregnant woman’s physical health, but could include her mental health, as well. The other nine states included a specific exception for the woman’s mental health. All but one of those states, California, failed to define the scope of that exception. An abortion for “mental health” reasons was authorized in California if the pregnant woman “would be dangerous to herself or to the person or property of others or is in need of supervision or restraint.” Cal. Health & Safety Code § 25954 (West 1984). Despite this narrow definition (the same standard as for civil commitment), more than 60,000 abortions were performed in California in 1970, 98.2% of which were performed for alleged “mental health” reasons. People v. Barksdale, 503 P.2d 257, 265 (Cal. 1972). In Barksdale, the California Supreme Court expressed “[s]erious doubt … that such a considerable number of pregnant women could have been committed to a mental institution” as the result of becoming pregnant. Id. The experience in California demonstrates that a “mental health” exception, even one narrowly defined, effectively allows abortion for virtually any reason.

[9] 262 So.2d 431 (Fla. 1972). The law was struck down on vagueness grounds.

[10] In 31 of those 33 states, more than 150 bills were introduced to repeal or weaken their laws on abortion between 1967 and 1972. None of those bills was enacted. See Paul Benjamin Linton, Overruling Roe v. Wade: Lessons from the Death Penalty, 48 Pepperdine L. Rev. 1, 13, 23–72 (Appendix A) 2021 (listing and describing bills). To the author’s knowledge, no such bills were introduced (prior to Roe) in either Louisiana or South Dakota.

[11] The maternal mortality rate, for purposes of these reports, represents the number of maternal deaths per 100,000 live births. The term “maternal deaths,” as used in the national vital statistics reports, refers to a death from any cause related to or aggravated by pregnancy or its management (excluding accidental or incidental causes) during pregnancy and childbirth or within 42 days of termination of pregnancy, irrespective of the duration and site of the pregnancy. “Maternal deaths” thus includes deaths from abortion—legal or illegal. The Center for Disease Control (CDC) reported 24 maternal deaths from legal abortions in 1972, 26 in 1973, 27 in 1974, and 27 in 1975. Center for Disease Control, “Abortion Surveillance 1975,” p. 7 (Issued April 1977) (the CDC did not report abortion-related maternal deaths before 1972). Although the actual number of such deaths may have been much higher, the reported number represents just around 5% of the reported maternal deaths for those four years. Excluding the number of reported abortion-related maternal deaths from the number of maternal deaths would reduce the maternal mortality rate by less than one maternal death per 100,000 live births. The maternal mortality rates set forth in this article are based on the sources cited, and have not been adjusted to exclude the very small number of reported abortion-related deaths.

[12] Vital Statistics of the United States 1966, Vol. II – Mortality Part A, Table 1–17, p. 1–41 (published in 1968).

[13] Vital Statistics of the United States 1972, Vol. II – Mortality, Part A, Table 1-17, p. 1–74 (published in 1976).

[14] Although Texas and Wisconsin had restrictive laws on the books throughout the years discussed in this article, those laws were declared unconstitutional by three-judge federal district courts in 1970 and were not enforceable for any part of 1971 or 1972. See Roe v. Wade, 314 F.Supp. 1217 (N.D. Tex. 1970); Babbitz v. McCann, 310 F.Supp. 293 (E.D. Wis. 1970). Accordingly, the data for 1971 and 1972 from Texas and Wisconsin have been included with the data from states that had enacted permissive abortion laws between 1967 and 1970.

[15] See Tables 1, 2. The data on the numbers of live births set forth in the tables accompanying this article are derived from the annual vital statistics reports published by HEW for each of the years 1964–1966, 1970–1972, and 1973–1975. See Vital Statistics of the United States (applicable year), Vol. I – Natality, Table 2-2. Data on maternal deaths for 1964-1966 were obtained from NCHS files available from the National Bureau of Economic Research using ICD-7 codes 640-689: NCHS, Multiple Cause of Death File, 1964-1966. Available from https://www.nber.org/research/data/mortality-data-vital-statistics-nchs-multiple-cause-death-data. Analyses and conclusions are the author’s own. Data on maternal deaths for 1970-1975 were obtained from CDC WONDER using ICD-8 codes 630-378: Centers for Disease Control and Prevention, National Center for Health Statistics. National Vital Statistics System, Mortality: Compressed Mortality File 1968-1978. CDC WONDER Online Database, compiled from Compressed Mortality File CMF 1968-1988, Series 20, No. 2A, 2000. Available from https://wonder.cdc.gov/controller/saved/D74/D391F344.

[16] See Tables 4, 5. Excluded from both comparisons discussed in this article are data from Alabama, Massachusetts, and Pennsylvania. Data from Alabama and Massachusetts have not been included because, long before the first three–year period discussed (1964–1966), their abortion laws allowed abortion for undefined (or broadly defined) reasons of health. See footnote 5, supra. Thus, the declines in maternal mortality in those two states between the first and second three-year periods could not be attributed to the enactment of a permissive abortion law. Data from Pennsylvania have not been included because it is not clear what the Commonwealth’s abortion law permitted and what it did not permit. The law prohibited “unlawful” abortion, a term not defined by statute or in case law. Although the Illinois abortion law was declared unconstitutional by a three-judge federal court, see Doe v. Scott, 321 F.Supp. 1385 (N.D. Ill. 1971), the court’s judgment was stayed by Supreme Court Justice Marshall less than two weeks later. See Hanrahan v. Doe, 404 U.S. 1012 (1972) (noting stay order entered by Justice Marshall on February 10, 1971). Accordingly, the Illinois law remained in effect for almost the entirety of the three-year period 1970–1972.

Finally, the data for four states—Connecticut, Florida, New Jersey, and Vermont—are limited to 1970 and 1971. Data from 1972 have not been included for these states because their abortion laws were declared unconstitutional by state or federal courts in 1972 and, therefore, were not enforceable for the remainder of that year. See Abele v. Markle, 342 F.Supp. 800 (D. Conn. 1972), Abele v. Markle, 381 F.Supp. 224 (D. Conn. 1972); State v. Barquet, 262 So.2d 431 (Fla. 1972); YWCA of Princeton, New Jersey v. Kugler, 342 F.Supp. 1048 (D. N.J. 1972); Beachum v. Leahy, 287 A.2d 836 (Vt. 1972).

[17] The data from Tables 1, 2, 4, and 5 are summarized in Table 7.

[18] Vital Statistics of the United States 1975, Vol. II – Mortality, Part A, Table 1-17, p. 1–74 (published in 1979).

[19] See Tables 2, 3.

[20] See Tables 5, 6.

[21] The data from Tables 2, 3, 5, and 6 are summarized in Table 8.

[22] The author recognizes that maternal mortality is influenced by a variety of variables, including, among other variables, race, poverty, obesity, and the age at which women become pregnant. This study does not purport to control for such variables, but focuses on overall maternal mortality data for 47 of the 50 states, grouped according to the kind of abortion laws they had in effect before Roe v. Wade was decided—permissive (including two states with unenforceable restrictive laws) or restrictive.