1 Introduction

A prominent literature in economics exploits state-level variation in policies governing access to contraception and abortion to identify how access to reproductive technologies shaped family formation, human capital attainment, and labor market outcomes. Economists have exploited the staggered repeal and invalidation of Comstock laws in the 1960s that made it legal to prescribe and distribute the contraceptive pill to adult women (Bailey, 2010), as well as policies governing the right of young, unmarried women to consent to the pill in the 1960s and 1970s (Goldin and Katz 2002; Bailey 2006; Guldi 2008; Hock 2007; Ananat and Hungerman 2012; Steingrimsdottir 2010; Bailey et al. 2012; Zuppann 2012b; Myers 2017). Economists also have exploited the early repeal of abortion bans in certain states to estimate the effects of abortion legalization in the early 1970s (Angrist and Evans 1996; Gruber et al. 1999; Levine et al. 1999; Ananat et al. 2009), but only a few papers additionally consider whether young, unmarried women could consent to abortion services at the time those services were legalized (Hock 2007; Guldi et al. 2008; Myers 2017).

The internal validity of the difference-in-differences research designs used in these papers requires careful and comprehensive coding so that the policy environment is fully and accurately described. It is therefore dismaying to discover that there is a great deal of dissonance in the policy coding used by different researcher teams. For example, four papers in the “power of the pill” literature that exploits policies governing young women’s legal confidential access to the pill in the sixties and seventies disagree on the year in which young women gained the right to consent to the pill for 35 of 51 states and the District of Columbia (Goldin and Katz 2002; Bailey 2006; Hock 2007; Guldi 2008). These discrepancies are neither limited to one set of authors nor are they small. Goldin and Katz (2002) and Bailey (2006) differ for 25 states by an average of 2.6 years, Goldin and Katz (2002) and Hock (2007) differ for 25 states by an average of 3.3 years, and Bailey (2006) and Hock (2007) differ for 23 states by an average of 4.0 years.Footnote 1 Similarly, the two empirical papers that had previously coded young women’s confidential access to abortion in the 1970s differ on the year minors gained confidential access to abortion for 18 states (Hock 2007; Guldi 2008). These discrepancies have the potential to substantially impact empirical estimates. In Myers (2017) I revisit the results in Goldin and Katz (2002) and Bailey (2006) and demonstrate that errors and omissions in policy coding contributed to the earlier authors’ overestimation of the effects of contraceptive policy and underestimation of the effects of abortion policy. Furthermore, in Myers (2017) I demonstrate that abortion legalization did not have homogeneous effects across states, but instead that the effect of legalization on young women’s fertility was amplified when the policy environment permitted young unmarried women confidential access to abortion.

This current paper is written for scholars seeking to understand the complex history of policies that have shaped young women’s legal access to contraception and abortion in the USA over the past 60 years and to code this policy environment accurately and comprehensively for their own empirical analyses.Footnote 2 I begin with a broad overview of the complicated interplay of common law precedents, state legislative actions, and state and federal judicial rulings that determine when and where contraception and abortion are legal and young women can confidentially access to these services without involving a parent or judge. This review covers the invalidation and repeal of Comstock laws following the introduction of the pill in 1960 and continues through the reform and repeal of abortion laws and policies governing young unmarried women’s confidential access to the pill and abortion through the late 1970s. I also consider the contemporary era of abortion policy, discussing and providing dates for “parental involvement laws” requiring parental notification or consent for minors’ abortions enforced since 1980. I then provide recommended state-by-state coding of the years these policy changes occurred. Finally, I compare this suggested coding to policy coding schemes used by various authors in prior work, document differences across the schemes, and reconcile these differences wherever possible.

For readers interested in more state-specific details and documentation, a lengthy online appendix and Myers (2022a) provide a detailed state-by-state review with complete citations of the primary and secondary sources that underpin the state policy coding. These citations reflect an extensive review of primary sources that include annotated statutes, judicial rulings, and state attorney general opinions, and of historical newspapers and medical journals providing evidence of how policy changes were implemented. I also incorporate information from snapshots of the policy environment afforded by reports from the Council of State Governments (1972, 1973), U.S. Department of Health, Education, and Welfare (1974, 1978), the Center for Adolescent Health and the Law (2006), NARAL (1989–2014), and the Guttmacher Institute (2017a, 2017b), and by a series of scholarly papers published in Family Planning Perspectives (Pilpel and Wechsler 1969, 1971; Paul et al. 1974, 1976). In addition, I review and incorporate information from Merz et al. (1995), who review minors’ access to abortion through the early 1990s, and from Bailey and Davido (2009), who provide information on the enforcement of state Comstock laws.

In doing so, I provide a thorough overview of policy changes based on primary sources, augment this with information from secondary sources, consolidate information from other reviews, and reconcile substantial disparities in previous policy codings. This exercise produces a set of unified policy codings, corrects previous errors, and clearly flags states where policy changes were ambiguous and reasonable scholars may disagree. As a result, this more clearly defined and unified set of variables will serve as a useful resource to other researchers. Scholars already have used this revised coding to explore the dynamics of state policy liberalism (Caughey and Warshaw 2016) and to estimate the effects of confidential access to contraception and abortion (Steingrimsdottir 2016; Myers 2017; Cragun 2019; Beauchamp and Pakaluk 2019; Jones and Pineda-Torres, 2021; Lindo et al. 2020; Myers and Ladd 2020; Rim 2021; Forsstrom, 2021; Farin et al., 2021). Readers who are interested in obtaining datasets with policy coding corresponding to this review can find them at Open Science Framework (Myers 2022b).Footnote 3

2 Policies governing the legality and confidentiality of contraception and abortion access since 1960

The legal rights of young women to access contraception and abortion services have depended on an interplay of laws governing adults and laws specifically targeting the age of majority and the rights of minors to consent to health care. In this section I dive into this complicated policy history, describing how it came to pass that in the early 1970s, depending on where she lived and her age, an unmarried teenager might be able to consent to neither contraception or abortion, to both, or to one but not the other.Footnote 4

I begin with an overview of the introduction of the birth control pill and policy changes in the 1960s that provided married adult women legal access to it (Sect. 2.1). I then describe the reform and repeal of abortion restrictions in the late 1960s and early 1970s, culminating in Roe v. Wade and the national legalization of abortion (Sect. 2.2). In Sect. 2.3 I turn my attention to laws specifically governing young women’s rights to access both contraception and abortion through the end of the 1970s, when a series of U.S. Supreme Court Rulings reshaped minors’ rights. Section 2.4 considers the contemporary era of minors’ access to reproductive services.

2.1 The introduction of the birth control pill

The United States Food and Drug Administration (FDA) approved Enovid, the first oral contraceptive, for the treatment of menstrual disorders in 1957. Three years later half a million women were already “on the pill” when the FDA approved it for contraceptive purposes on June 23, 1960. By 1962 approximately 1.2 million married women were on the pill, and this grew to 6.5 million married women by 1965 (Tone 2002).Footnote 5

The pill was not immediately legally available in all states. The federal Comstock Act, which had once prohibited the distribution of contraceptives across state lines, had been invalidated by the time that the pill was introduced. But many states continued to enforce their own “little Comstock” laws modeled on the Comstock Act that restricted the advertisement, sale, and/or use of contraceptives within those states. Throughout this paper, I follow convention in referring to these state laws as “Comstock laws.” The U.S. Supreme court ruling in Griswold v. Connecticut (381 US 479, 1965) recognized the right of married people to use birth control without government interference. Seven years later in Eisenstadt v. Baird (405 US 438, 1972), the court struck down a Massachusetts law restricting access to birth control for unmarried people, stating that unmarried people have the same right to privacy as married ones. The Court’s recognition of a constitutional right to privacy in contraceptive decisions altered enforcement of and compliance with state Comstock laws, and in the years following these rulings many states repealed or substantially liberalized their anti-contraception laws. After 1965, new state laws regarding contraception were generally affirmative.Footnote 6

2.2 The legalization of abortion

Abortion became legal nation-wide on January 22, 1973, when the Supreme Court ruled in Roe v. Wade (410 US 113, 1973) and Doe v. Bolton (410 US 179, 1973) that women have a fundamental constitutional right to privacy in choosing to abort a fetus. Prior to these rulings, five “repeal” states and the District of Columbia had legalized abortion.Footnote 7 In addition, thirteen “reform” states had adopted provisions resembling those set forth by the American Law Institute in the Model Penal Code (MPC). These reform laws made abortion legal if performed by a physician because of substantial risk that continuing the pregnancy would cause grave physical or mental impairment or death of the woman, or the fetus would be born with a grave physical or mental defect or in cases where the pregnancy resulted from rape or incest.Footnote 8 In the remaining states, abortion generally was prohibited except to save the life of the mother.

The Centers for Disease Control began collecting abortion surveillance data on legally induced abortions in 1969 (Smith and Bourne 1973). By the beginning of 1971, the CDC was receiving information from 17 state health departments and from one or more hospitals in 8 other states and the District of Columbia. The reported number of legally induced abortions and the abortion ratio (abortions per 1,000 live births) for these states are presented in Table 1. The variation in reported legal abortions among the reform states is notable; abortion ratios range from 13.7 in South Carolina to 277.1 in Kansas, the latter figure exceeding the abortion ratio in three repeal states. Some of the variation in legal abortion ratios among reform states likely reflects differences in reporting requirements and accuracy as well as inter-state travel from neighboring states. It likely also reflects differences in how the subjective mental health standard was applied by physicians and therapeutic abortion committees. In Maryland, one of the reform states with a high abortion ratio, “mental health” was the indication for 96 percent of legal abortions performed in the first six months of 1971 (Melton et al. 1972). In Colorado, another reform state with a high abortion ratio, the Denver General Hospital Therapeutic Abortion Board approved 62 percent of applications for therapeutic abortions, the majority for mental health reasons (Thompson et al. 1970).

Table 1 Legal abortions in 24 states and the District of Columbia, 1971

Sociologist Carole Joffe’s (1996) summary of interviews with abortion providers from this era provides anecdotal evidence of substantial inter-state and even inter-hospital variation in the ease with which physicians could obtain approval to perform abortions under mental health standards. Victor Bladheck, a physician in California at the time that the state enacted abortion reforms, recalled that after abortion reforms “the floodgates were opened…. We found three of four sympathetic doctors in the area that agreed to see these patients immediately and always agreed that the patient needed an abortion. All had the same diagnosis: ‘situational anxiety.’ These were normal women, in my opinion, with no psychiatric problems” (As quoted in Joffe 1996, pp. 121). Other providers, however, indicated that dealing with abortion committees was frustrating, time-consuming, and that many boards were extremely reluctant to approve abortions. One physician stated that he was reluctant to perform even approved abortions because he feared a zealous anti-abortion activist might still try to push prosecution jeopardizing his medical license (Joffe 1996).

2.3 Confidential access to contraception and abortion, 1960–1979

Comstock laws and abortion reforms and repeal determined the legality of prescription contraception and abortion for adult women, but minor women are subject to additional regulations determining whether they can provide legal consent to medical services. If statutory or case law has not extended this right to minors, or if a statutory parental involvement law offers a valid restriction, then minors must involve a parent in their decision to obtain contraception or abortion. For this reason, I refer to environments in which minors can provide legal consent without involving a parent as ones granting “confidential” access.

Under the system of legal precedent known as “common law,” informed consent is necessary for a physician to provide medical services, and minors are generally considered incapable of providing informed consent to medical care. Accordingly, at the time of the introduction of the birth control pill, unmarried women under the age of majority—21 in most states—who had not previously given birth generally could not provide legal consent to contraceptive services (Pilpel and Wechsler 1969).Footnote 9 Exceptions arose in states that had enacted medical consent statutes specifically granting minors capacity to consent to medical care, as well as in states in which the legislature or courts had recognized a mature minor doctrine whereby a minor can consent to medical care if she is judged capable of understanding the nature and potential consequences of treatment. In addition, in some states minors attained majority upon marriage, emancipation, or giving birth, and presumably could consent to contraception under these circumstances.

By the mid-seventies, most states had lowered the age of majority to 18, permitting women aged 18 and over to consent to contraception and, once it was legalized, to abortion as well. Confidential access to contraception and abortion for women under the age of majority continued to depend on the presence of state mature minor doctrines or other laws granting minors the right to consent to medical care without involving a parent. By 1979, 29 states had medical consent laws and/or mature minor laws that affirmed minors’ ability to access contraception legally and confidentially from any provider. Two additional states, Hawaii and Montana, had passed laws permitting minors to consent to contraceptive services, but these policies explicitly permitted physicians to choose to notify a minor’s parents.

Minors also gained increasing access to contraception through federally funded family planning clinics. On December 24, 1970, Title X of the federal Public Health Service Act was signed into law, establishing a program of federally funded family planning clinics that were required to make contraceptive services available to “all persons desiring such services…without regard to religion, creed, age, sex, parity, or marital status” (Public Health Service Act 1970; 1978). Program regulations adopted in 1972 expressly protected the confidentiality of patients (Maradiegue 2003). In 1978, Congress, which was concerned that minors were not taking advantage of the services offered by these clinics, amended the Act to explicitly mandate Title X to provide confidential contraception services to adolescents (Reimer 1986; Boonstra and Nash 2000). In 1981, Title X was again amended to “encourage family participation” under the grants. Pursuant to that language and with the encouragement of the Reagan administration, the Department of Health and Human Services adopted regulations mandating that parents be notified with ten days of the prescription of contraceptives to their minor children at Title X clinics. This rule, commonly known as the “Squeal Rule” was challenged and struck down by the U.S. Court of Appeals (Planned Parenthood Fed. of America v. Heckler, 712 F.2d 650, 1983). Currently more than 1/3 of teenagers who visit reproductive health clinics obtain services at a Title X clinic (Jones and Boonstra 2004).

In contrast to this generally affirmative trend for contraception, legislation related to minors’ access to abortion in the 1970s was more mixed. While some medical consent laws permitted minors to consent to abortion, many, particularly those passed after 1973, excluded abortion from the services to which a pregnant minor could consent. In the wake of Roe v. Wade (410 US 113, 1973), other states enacted parental notification and/or consent requirements which served to explicitly restrict minors’ confidential access.

As a result of the timing of the introduction of the contraceptive pill, the legalization of abortion and of distinctions made in states laws between minors’ ability to consent to each, in a given state in a given year between 1960 and 1976, a minor seeking reproductive services might be legally able to consent to neither contraception or abortion, to both, or to one and not the other. Three Supreme decision in the late seventies—Planned Parenthood of Central Missouri v. Danforth (428 US 52, 1976), Carey v. Population Services International (431 US 678, 1977), and Bellotti v. Baird (443 US 622, 1979)—established guidelines about the types of restrictions that could be imposed on minors seeking reproductive services. These decisions also served to clarify minor’' ability to consent absent an enabling statute.

In Planned Parenthood of Central Missouri v. Danforth (428 US 52, 1976), the court ruled that a Missouri parental consent law for abortion was unconstitutional, stating that a state does not “have the constitutional authority to give a third party an absolute, and possibly arbitrary, veto over the decision of a physician and his patient to terminate the patient’s pregnancy, regardless of the reason for withholding consent” (p. 94). This ruling established that states may not impose a blanket prohibition on minors seeking abortion, nor can they impose parental consent laws that do not have a bypass option. The ruling invalidated parental consent laws in several states that did not contain a judicial bypass option, although some of these states did not immediately cease enforcing them (DHEW 1978).

The following year in Carey v. Population Services International (431 US 678, 1977), the Supreme Court declared a New York State statute that prohibited the sale or distribution of contraceptives to minors under 16 unconstitutional with respect to non-prescription contraceptives. The court affirmed that, like adults, minors have a right to privacy in choosing whether to “bear of beget a child” (p. 678) and that, just as a state could not impose a blanket prohibition on minors seeking abortions, it similarly could not prohibit the distribution of contraception to minors. Although the ruling specifically regarded non-prescription contraceptives, which Population Services International manufactured, the Court’s reasoning suggested that its conclusions would also apply to prescription contraceptives.

In Bellotti v. Baird (443 US 622, 1979) the court offered more detailed guidance on what types of parental involvement requirements could be imposed on minors seeking abortions. It clarified that a bypass procedure must allow the judge to rule in an immature minor’s interest or to determine that a minor is mature enough to make her own decision in consultation with a physician.Footnote 10 The ruling invalidated parental consent laws enacted after Planned Parenthood v. Danforth that did not provide a confidential judicial bypass procedure that allowed a judge to determine that a minor was mature enough to make her own decision in consultation with a physician. Bellotti v. Baird also established that parental notification laws must meet similar requirements as parental consent laws, invalidating parental notification laws without judicial bypass options. The opinion written by Justice Powell for Bellotti v. Baird illustrates the Court’s reasoning in extending the right to privacy in childbearing decisions to minors:

The abortion decision differs in important ways from other decisions that may be made during minority…. The pregnant minor’s options are much different from those facing a minor in other situations, such as deciding whether to marry. A minor not permitted to marry before the age of majority is required simply to postpone her decision…. A pregnant adolescent, however, cannot preserve for long the possibility of aborting, which effectively expires in a matter of weeks from the onset of pregnancy (p. 623).

2.4 Confidential legal access to contraception and abortion, 1980–2020

By 1980, following the decisions in Planned Parenthood v. Danforth and Bellotti v. Baird, the rights of minors to consent to abortion absent a parental involvement law had been established. States could limit minors’ confidential access by passing statutes requiring parental involvement, but such laws had to include a judicial bypass option whereby a judge could either declare a minor competent to consent or make a decision in the minor’s best interest. At the beginning of 1980, only one state, Utah, had an enforceable parental involvement law on the books (Utah Code Ann. § 76–7-304, 1974), and that law had been ruled constitutional only as applied to immature minors (H.L. v. Matheson 450 US 398, 1981).

Over the following decade, several more states enacted parental involvement laws, and by the end of 1990, 13 states were actively enforcing them. However, the application of the judicial precedents established by Planned Parenthood v. Danforth and Bellotti v. Baird remained somewhat mixed, and legal challenges created a turbulent policy environment in some states. Abortion rights advocates frequently filed suit in response to each new parental involvement law, and it was and remains common for enforcement of a new law to be enjoined pending a judicial review process. Nevertheless, the number of states that were actively enforcing parental involvement laws grew, from a single state in 1980 to 20 by the end of 1991. The following year, the Supreme Court decision in Planned Parenthood of Southeastern Pennsylvania v. Casey (505 US 833, 1992) upheld several provisions of a Pennsylvania law, including a parental consent requirement. In this decision, the Supreme Court for the first time applied the “undue burden” standard to abortion regulations, which it defined as a law placing “a substantial obstacle in the path of a woman seeking an abortion before the fetus attains viability” (p. 837). In the wake of this ruling, the number of enforced parental involvement laws continued to grow, to 36 states at present in 2017.

In contrast to efforts to restrict minors’ access to abortion, there have been few efforts to restrict minors’ access to contraception since 1980. In 1983, Utah attempted to enact a parental notification requirement for minors seeking contraception (Utah Code Ann. § 76–7-325, 1983). Enforcement of the law was enjoined before it could go into effect, and it was stuck down by a district court in Planned Parenthood Association of Utah v. Matheson (582 F. Supp. 1001, 1983). The court tied its decision to judicial precedent regarding abortion:

The Court acknowledges that a decision concerning the use of contraceptives is not fraught with the time limitations inherent in a decision concerning termination of a pregnancy through abortion. Nonetheless, decision whether to accomplish or to prevent conception is among the most private and sensitive. Moreover, in contrast to the decision to marry, a decision concerning the use of contraceptives is similar to the decision whether to have an abortion in that it cannot be delayed until the minor reaches the age of majority without posing the risk of serious harm to the minor. (p. 1008).

At present, no states have laws in place that explicitly restrict the ability of private providers to provide confidential contraceptive services to minors.Footnote 11 Twenty-one states and the District of Columbia have enacted laws granting all minors the capacity to consent to healthcare services in general and/or contraceptive services in particular, and all but four states have some type of confirmatory law stating the conditions under which a minor may consent (Guttmacher Institute 2017a). Absent a law explicitly authorizing a minor to consent, the decision to provide confidential contraceptive services rests with the provider. A provider may be encouraged to provide contraceptive services to minors by legal scholars who interpret the Carey decision as affirming minors’ rights to consent to contraception.Footnote 12 The Center for Adolescent Health and the Law (2006) advises that the right of privacy extends “protection to contraceptive decisions by minors as well as adult women” (p. 7). In a report published by the same organization, English et al. (2010) advise that “even in the absence of a statute authorizing minors to consent for family planning services or contraceptive care, if there is no valid statute or case prohibiting them from doing so, it would be reasonable to conclude that minors may give their own consent for these services” (p. 8). But the American College of Obstetricians and Gynecologists (ACOG) adopts a more conservative attitude, advising that “adolescents’ legal rights to confidential contraceptive services vary by state and change over time. Where allowed, obstetrician-gynecologists should provide adolescents the opportunity to discuss [contraceptive services] without a parent or guardian for at least part of the visit” (ACOG 2017, p. 3). The advisory opinion recommends referral to a Title X health clinic if the policy environment does not permit confidential counseling (ACOG 2017).

There is little empirical evidence on whether, in practice, providers typically choose to provide confidential contraceptive services in the absence of state laws expressly authorizing them to do so. A policy report from the Guttmacher Institute asserts that physicians “commonly provide medical care to a mature minor without parental consent” (Guttmacher Institute 2017a, p. 1), and a reference manual for school health officials advises providers that contraceptive services “are” provided in states without explicit consent laws and that most providers “will use every effort to resist providing information to a parent against the wishes of the minor patient” (Cohn et al. 2005, p. 238).

2.5 Emergency contraception (EC)

When the FDA approved the emergency contraceptive Plan B in 1999, it was initially available only as a prescription contraceptive and hence subject to the same policies (or lack thereof) governing minors’ access as other prescription contraceptives. From this time until 2006, eight states chose to legislate over-the-counter access (Zuppann 2012a), and the relevant laws did not make mention of an age restriction.Footnote 13 In 2006 the FDA promulgated new rules approving Plan B for over-the-counter distribution for women aged 18 and older. In 2009, the FDA lowered the age at which Plan B could be provided without a prescription from 18 to 17. In 2013, this age restriction was lifted all together because of a court order (NARAL 2017).

2.6 Summary of dates of legal changes

In this paper I provide suggested policy coding for researchers who wish to demarcate the dates of important policy changes affecting young women’s legal and confidential access to prescription contraception and abortion. These dates can be used to implement difference-in-differences research designs estimating the average treatment effect on the treated (ATT) of reproductive policies by comparing changes in outcomes following a policy change to changes in outcomes in states where the policy is not changing. Under assumptions underlying difference-in-differences estimation, such a comparison can credibly identify the causal effect of these reproductive policies for populations who were treated. This is distinct from estimating the causal effect of the existence of the reproductive technology (i.e., the invention of the birth control pill or existence of abortion), which is not varying with these policies.

I focus on two broad categories of policy: those governing the legal provision of prescription contraceptives and abortion to adult unmarried women, and those governing the age of majority and minors’ legal rights to consent to services without a parent’s involvement.Footnote 14 Below I will describe my criteria for coding these policy changes. Readers interested in a detailed state-by-state review with citations of statutes and legal rulings should consult the online appendix or Myers (2022a).

Table 2 provides a state-level summary of the years in which young, unmarried women gained legal and confidential access to prescription contraception over the period 1960 to 2020. By “legal” I mean that physicians and pharmacists were legally permitted to distribute contraceptives to unmarried women, and by “confidential” I mean that the law explicitly allowed women in each age range to consent to contraceptive services and hence avoid involving anyone beyond healthcare providers in her decision.

Table 2 Year unmarried women gained legal and confidential access to prescription contraception, 1960–2017

The period I consider begins with the FDA approval of Enovid for contraceptive purposes in 1960. In this year women aged 21 and older were legal adults in all states, and as such legally permitted to consent to medical services. However, extant Comstock laws prohibited physicians and pharmacists from distributing the birth control pill in several states.Footnote 15 In these states, married adult women gained legal access between 1960 and 1965, as Comstock laws were repealed, struck down, or invalidated by Griswold v. Connecticut in 1965.Footnote 16 I assume that unmarried adult women gained legal access to the pill at the same time as married adult women unless a state had a Comstock law that differentiated between married and unmarried persons.Footnote 17 It is likely that in practice unmarried women did not gain legal access at the same pace as married women. Indeed, whether the right to privacy in contraceptive choices extended to unmarried adult women as well seems to have been a gray area in the law, one not firmly established until Eisenstadt v. Baird (1972). Absent a legislative statute, judicial ruling, or attorney general's opinion explicitly denying or affirming the right of unmarried adult women to consent, however, I assume that the date that birth control became legally available to married adult women also demarcates the date at which it became more available to unmarried adult women, even if the rate of increase in access was lower during subsequent years.

Columns 3–4 of Table 2 indicate the year that a legal change extended confidential access to women aged 18–20 (who were defined as adults in most states after the lowering of the age of majority) and 15–17 (defined as minors in all states). Columns 5–6 report the type of policy that granted women in each of these ages ranges confidential access. In coding these columns, I assume that women under 21 gained legal confidential access to contraception as soon as those services were available to older unmarried women, and a regulatory change affirmed the right of women under age 21 to consent to contraceptive services. I interpret affirmative changes in the legal environment to include age of majority statutes (AOM), medical consent statutes specifically granting all minors capacity to consent to reproductive services (MCL), medical consent statutes granting providers permission to provide confidential services to minors if in the provider's opinion the failure to provide services would be hazardous to a minor's health (HH), legislative or judicial mature minor doctrines (LMM and JMM), affirmative attorney general opinions (AG), and judicial rulings that affirmed minors' rights to consent (J).

I do not code laws permitting minors to consent to contraception at state-funded public health clinics as granting broad access, because these legal changes likely did not affect as broad or representative a group of women as those that permitted all types of providers to supply confidential reproductive services to minors. This is a subjective choice, and researchers may wish to explore the robustness of their analyses to making the alternative decision. The online appendix (Myers 2022a) makes note of all such laws and the accompanying data sets (Myers 2022b) include the alternative policy coding. This results in a relatively small number of changes. When looking at laws granting access to women aged 18–20, the dates are different only for Georgia (1968 instead of 1971) and Wyoming (1969 instead of 1973), both of which permitted publicly funded clinics to provide confidential services to minors before they lowered the age of majority. I also do not code as affirmative confidential access laws in Hawaii and Montana that grant minors capacity to consent to contraceptive services, but explicitly grant physicians the right to inform a minor’s parents.

For an illustrative example of the coding, consider the timing of affirmative legal changes in Iowa, which are described and documented in detail in the appendix (Myers, 2022a). Iowa did not have an extant Comstock law when Enovid was approved in 1960, and so the pill became legally available to unmarried women aged 21 and over at that time. In 1972, the Iowa legislature lowered the age of majority from 21 to 19, extending confidential legal access to women aged 19 to 20. The following year the legislature lowered the age of majority to 18, extending confidential legal access to this group. Iowa is unusual in enacting an affirmative policy governing minors’ access after 1980. (Only two other states have done so.) In 1999 the Iowa legislature passed a bill related to HIV testing that also granted minors the legal right to consent to confidential services. This is the year that I code women aged 15–17 as gaining legal and confidential access. However, I again wish to emphasize that the post-1976 policy environment governing minors’ right to consent to contraception is somewhat ambiguous in those states without affirmative laws, and many providers may choose to provide confidential services to minors.

Table 3 presents suggested policy coding for minors’ legal and confidential access to abortion over the period 1969 to 1979. With respect to adults’ legal access to abortion services, I code the earlier of the repeal of abortion restrictions or Roe v. Wade as the determining date.Footnote 18 As with the case of affirmative laws for contraception women under the age of 21 are coded as gaining confidential legal access due to the lowering of the age of majority, mature minor doctrines, and medical consent laws. It is noteworthy that in several states, medical consent laws enacted prior to Roe granted pregnant minors the right to consent to pregnancy-related medical care. Although these laws appear to have been intended to facilitate prenatal medical care, they had the probably unintended consequence of allowing pregnant minors to consent to abortion at the time it was legalized. In the wake of Roe, many states amended these laws to exclude abortion from the lists of services to which minors could consent, while others enacted parental involvement laws to explicitly restrict minors’ confidential access to abortion. In this complex policy environment, I code minors as gaining confidential legal access only if a state law explicitly granted this legal right, or if a state restriction was struck down by a court ruling. Such rulings were largely issued in the 1976–1979 policy period that I regard as ambiguous, and I caution researchers against an attempt to treat policy variation as clear or objectively defined in this span of years.

Table 3 Year women gained confidential and legal access to abortion, 1969–1979

Table 4 covers the next three decades of abortion policy, summarizing minors’ abilities to consent to abortion services from 1980 to 2020, an era in which a series of Supreme Court decisions had established minors’ default right to consent to abortion services absent a valid state restriction. At the beginning of 1980, nearly all parental involvement statues for abortion had been invalidated by the Carey and Bellotti decisions. Within a few years, legislatures around the country had moved to enact new and valid parental involvement laws, though most did not take effect until after Casey was decided in 1992.Footnote 19 In this table, I indicate enforcement of a parental involvement law if a policy was being enforced that mandated parental notification or consent for all minors under a certain age (usually 18), and if the only bypass option involved consulting a judge, independent healthcare provider, or other adult family member. A handful of laws only recommended but did not require parental involvement, or explicitly permitted the providing physician to apply a mature minor standard. I do not regard these as sufficiently restrictive to indicate enforcement in Table 4, but they are noted in the footnotes to the table and discussed in the online appendix (Myers 2022a) for researchers who wish to make a different subjective decision.

Table 4 Enforced state parental involvement laws, 1980–2020

The case of Kentucky offers an illustrative example of the shifting abortion policy environment by both Tables 3 and 4. As described in detail in the state-by-state policy appendix, in the wake of Roe, Kentucky enacted new legislation regulating the conditions under which women could seek abortions. The provisions included a spousal consent requirement for married women and a parental consent requirement for women under age 18, which was the legal age of majority. The law was challenged by two Kentucky physicians, and before it took effect a district court invalidated both consent requirements. In this ruling, the judge stated that although he was not issuing an injunction against enforcement, Kentucky would presumably “give full credence to this decision” (Wolfe v. Schroering 1974, p. 639). In Table 3, I code this as the date that minors under age 18 could legally and confidentially access abortion in Kentucky.

On August 18, 1976, six weeks after the Danforth ruling, the Sixth Circuit Court of Appeals ruled that both consent requirements in the Kentucky law were unconstitutional (Wolfe v. Schroering 1976). In January 1978, a bill was introduced in the Kentucky legislature to again implement a parental consent requirement, but the state assistant attorney general issued an advisory opinion that the provisions would not be enforceable under Danforth, and the bill died in session. In 1980, following the Supreme Court decision in Bellotti v. Baird clarifying the circumstances under which a parental involvement law might pass constitutional muster, two bills were introduced in the Kentucky legislature related to minors’ abortion access. The first, requiring a court order before a minor could obtain an abortion did not pass the Kentucky House. The second passed the House but reached the Kentucky Senate too late for a vote. This bill was reintroduced and enacted in early 1982. Before it could take effect, however, the United States District Court for the Western District of Kentucky issued a temporary restraining order, and ultimately struck down the law on the grounds that it did not specify a period in which a decision must be made in the case of a judicial bypass (Eubanks v. Brown, 604 F. Supp. 141 (W.D. Ky.1984)). The Kentucky legislature amended the law in the following legislative session, and the new version was scheduled to go into effect in July 1986. However, the district court again issued a temporary restraining order, and then issued a ruling striking and amending language related to the notification of two parents. This revised version of the parental consent law took effect for nine months in 1989, before the Sixth Circuit Court of Appeals issued a three-paragraph order instructing the state to cease enforcement, and then remanded the case to the district court (Eubanks v. Wilkinson, 937 F.2d 1118 (6th Cir. 1991)). In the meantime, and in the wake of the 1992 Danforth decision, the Kentucky legislature drafted a new parental involvement law requiring the consent of only one parent. This law—K.R.S. § 311.732 (2010)—took effect on July 15, 1994, and is still enforced. In the appendix, I also note and cite newspaper articles covering these legal changes, which suggest that providers were responsive to the policy environment.

The sum effect of all this legal wrangling in Kentucky is reported in Tables 3 and 4. Table 3 indicates that women under age 18 gained confidential legal access to abortion in Kentucky in 1974. Table 4 indicates that for the 1980–2017 period, a parental involvement law was enforced in Kentucky in 1989, and from 1994 to present.

3 Comparison to previous coding

3.1 Coding of confidential and legal access to contraception, 1960–1976

To my knowledge, four previous teams of researchers have coded the date that a legal change first granted young unmarried women confidential access to prescription contraception and hence to the birth control pill: Goldin and Katz (2002), Bailey (2006), Hock (2007), and Guldi (2008). While the paper by Hock remains unpublished, the others are influential and widely cited.Footnote 20 Table 5 reproduces that year in which I have coded a legal change first granting women under age 20 confidential access to contraception and the years in which each of these four other research teams coded these events.Footnote 21 The last two rows of this table summarize the number and magnitudes of discrepancies relative to the suggested coding in this paper.

Table 5 Comparison of coding of laws granting teenagers confidential access to prescription contraception

As reported in the introduction, the coding of the age at which unmarried teenage women could first consent to contraception is inconsistent for 35 of 51 states and the District of Columbia between Goldin and Katz (2002), Bailey (2006), Hock (2007), and Guldi (2008), all papers that used this coding to estimate causal effects of early legal access to the contraceptive pill on various outcomes. The last two rows in Table 5 summarize the number of discrepancies between each set of coding and the coding that I suggest in this paper, as well as the mean and median of the magnitude of the discrepancy. My own, independent, coding differs from that of Goldin and Katz for 27 states and by an average of 3.6 years; from Bailey for 20 states and by an average of 3.6 years, from Guldi for 16 states and by an average of 3.9 years, and from Hock for 8 states and by an average of 3.3 years. The differences cannot, for the most part, be attributed to different interpretations of the same law. Of the 20 states for which the coding differs from Bailey’s, for example, I view only 5 (Florida, Georgia, Kentucky, Mississippi, and Wyoming) as the result of differences in interpretation of an ambiguous legal environment.

Shortly after the first draft of this paper became available, a research team whose members include Martha Bailey and Melanie Guldi, two authors who had previously published papers coding access to the pill, released a working paper that reviews the coding of state-level access to the pill (Bailey et al. 2011). Their suggested coding is reproduced in the last column of Table 5. Bailey et al. corrected several errors in the coding in Bailey (2006) and Guldi (2008), but 14 discrepancies remain between our two sets of independent coding. These discrepancies are quite large in magnitude: an average of 7.4 years. For 8 of these states (Arkansas, Idaho, Illinois, Montana, Nevada, North Dakota, Oklahoma, and Utah), the differences arise because these states had established the age of majority at 18 for women and 21 for men as of 1960. I interpret this as permitting women aged 18 and older to consent to the pill at the time that it was introduced. Bailey et al., on the other hand, assume that women under the age of 21 did not gain access to the pill in these states until either the age of majority was equalized for men and women, or another legal change granted minors access. Bailey et al. justify this decision by indicating, variously, that the age of majority statutes likely was only intended to apply to marriage or that the scope of the statute did not clearly apply to medical care. They do not cite supporting evidence for these assertions.

In the case of Stanton v. Stanton (1974; 1975), the Utah Supreme Court and United States Supreme Court offer a different perspective than Bailey et al. on the intent and effects of differential age of majority statutes. In this case, a plaintiff mother challenged a child support judgment that ended child support for her daughter at age 18 (the age of majority for females in Utah) but for her son at age 21 (the age of majority for males in Utah). The Utah Supreme Court affirmed a lower court’s denial of the petition, explaining that.

“[T]he belief held by many that generally it is the man's primary responsibility to provide a home and its essentials for the family; and that however many exceptions and whatever necessary and proper variations therefrom may exist in differing circumstances, it is a salutary thing for him to get a good education and/or training before he undertakes those responsibilities…. Perhaps more important than this, there is another widely accepted idea: that girls tend generally to mature physically, emotionally, and mentally before boys, and that they generally tend to marry earlier…. we do not regard it as our judicial function to pass upon the soundness or the unsoundness of the ideas just mentioned above. What we do note is our knowledge of their existence; and that they have played an essential role in the history of the development of the law as declared in the statute under attack” (Stanton v. Stanton, 517 P.2d 1010, pp. 1012–1013 (Utah 1974)).

This suggests that the lower age-of-majority for females was based on the idea that by age 18 they matured faster and were less needing of parental resources than were males. The plaintiff mother appealed, and the United States Supreme court struck down the Utah Court’s decision on the grounds that it denied individuals aged 18–20 equal protection under the law. The Supreme Court’s decision does not explicitly address the right to consent to medical services in Utah but cite other Utah statutes providing age thresholds to vote and hold office to illustrate that the age-of-majority statute’s application to those rights not elsewhere specified was not rational (Stanton v. Stanton 1975).

The decision rule applied by Bailey et al. (2011) is that age-of-majority statutes with equal age cut-offs governed access to the pill, while those with unequal age cut-offs did not. It is not clear why this might have been the case. The Stanton v. Stanton ruling appears to imply that the scope of differential age-of-majority statutes was not limited to marriage, and the language of the statutes themselves does not suggest that such a narrow interpretation should be applied. For instance, the Arkansas legislature amended its age of majority statute in 1873 to read “Males of the age of twenty-one years, and females of the age of eighteen years, shall be considered of full age for all purposes [emphasis mine], and, until those ages are attained, they shall be considered minors.”Footnote 22 Moreover, if, as Bailey et al. (2011) argue, an age-of-majority law did not govern the ability of 18-year-old women to consent to birth control prior to its amendment, it is not clear why it would have done so after its amendment, when no other language save for the numeric age changed.

In trying to ascertain the scope of these laws, I reviewed judicial rulings that cite age of majority statutes in states where the age of majority was set at different levels for males and females. I observe that courts in different states applied the different ages of majority to a variety of rights and purposes including the need to appoint guardian ad litem, the age at which child support payments could end, and the drinking age.Footnote 23 Consider the following examples:

  • In a 1910 ruling related to the expiration of the statute of limitations, the Arkansas Supreme Court stated, “It is urged that the purpose of the lawmakers in passing the act of 1873 was to encourage early marriages by enabling females to contract marriage at an earlier age than twenty-one years without the consent of parents or guardians. That may have been a reason that appealed to the lawmakers, but there is nothing to show that this was the sole purpose of the act. The statute is broad enough to completely emancipate females at the age of 18 years” (Brakes v. Sides 1910). In subsequent rulings, the Arkansas Supreme Court applied the lower age of majority to women’s ability to redeem lands sold during their minority (Gamble v. Phillips 1913), the need to appoint guardian ad litem to represent the interests of a female minor child (Federal Bank of St. Louis v. Cottrell 1939), and the age at which child support payments could cease for female children (Jerry v. Jerry 1962).

  • In a 1973 opinion, the Illinois Attorney General indicated that where the term “minor” was used in any statute, it was defined as in The Probate Act setting the age of majority for that state (1973). The Illinois Attorney General observes that when the Illinois Liquor Control Act was passed in 1939 to prohibit serving alcohol to “minors,” this meant women under age 18 and men under age 21. The Illinois legislature subsequently amended the Liquor Control Act in 1961 to replace “minor” with “person under 21” so that the legal drinking age would be equalized.Footnote 24

  • South Dakota courts applied the differential age of majority to the termination of child support (Comstock v. Comstock 1981) and the enforcement of contracts (Gruba v. Chapman 1915).

Contemporary authors summarizing state legal environments do not adopt Bailey et al.’s (2011) approach, but rather treat statutes with differential age cut-offs for men and women as governing the ability to consent to medical services, absent other statutory language explicitly addressing consent to medical care (Paul et al. 1974, 1976).

In sum, the language of the statutes, their applications by courts, and contemporary secondary sources all support the view that when the age of majority was lower for women than for men, these were the ages that governed the rights of each sex to consent to contraceptive services. This is the approach that I adopt in my suggested coding.

3.2 Coding of confidential access to legal abortion, 1970–1975

Two researchers have coded the date that a legal change first granted confidential access to abortion in the 1970s: Guldi (2008) and Hock (2007).Footnote 25 Table 6 compares the year in which I have coded a legal change first granting women under age 18 confidential access to legal abortion and the years in which Guldi (2008) and Hock (2007) coded these events. I report dates only for the 1970–1975 period because of uncertainty about the legal environment between the Supreme Court decisions in Planned Parenthood v. Danforth (1976) and Bellotti v. Baird (1979).

Table 6 Comparison of coding of laws granting teenagers confidential access to abortion

My coding differs from that of Guldi (2008) for 14 states and of Hock (2007) for 13 states. Four of the discrepancies between my coding and that of Guldi (2008) arise for states where the legal environment was ambiguous (Florida, Hawaii, New Jersey, and North Carolina), and the remaining differences appear to be errors. Hock (2007) and I have the same coding for 8 of these 10 states. Where my coding differs from that of Hock (2007), the most frequent explanation is that I (like Guldi) treat a court ruling that invalidated a parental involvement law as confirming minors’ right to consent to abortion, whereas Hock, in general, does not. The snapshots of minors’ ability to consent to abortion provided by Paul et al. (1974, 1976) support my interpretation that in the pre-Danforth period, a court ruling enjoining enforcement of a parental involvement law de facto granted minors confidential access to abortion unless there was some other restrictive statute or judicial precedent that was not addressed by the ruling. The language of many of these court rulings also appears to strongly suggest that minors could consent. For instance, in the Florida court ruling striking down a parental consent requirement in 1973, the three-judge panel indicates that “parents cannot look to the state to prosecute and punish the physician (or other participants) who performs an abortion” (Poe v. Gerstein 1973). However, providers may still have been reluctant to provide abortions to minors without parental consent in the absence of statutory language explicitly permitting them to do so. As described in the state-by-state review, providers in Florida and Massachusetts appear to have begun changing their policies only after multiple court rulings were issued regarding those states’ respective consent requirements.

3.3 Coding of enforcement of parental involvement laws for abortion, 1980–2020

By 1980 judicial precedent established that minors can consent to abortion services absent a valid restriction. In the ensuing years, states began to pass parental involvement laws to limit minors’ confidential access to abortion. In Table 7, I reproduce the year in which I code enforced parental involvement laws in place and compare them to the coding in Levine (2003) and New (2009), and the combined codings from Sabia and Rees (2013) and Sabia and Mark Anderson (2016). As in Tables 5 and 6, I summarize the number of discrepancies.

Table 7 Comparison of coding of dates of enforcement of parental involvement laws, 1980–2017

The bottom row summarizes the number of discrepancies, counting only those within the time periods that each set of authors intends to address. My coding and that in Levine (2003) differs for 18 states. Some of these differences are minor and likely explained by the lag between the dates that new legislation is enacted and takes effect. For several other states (Connecticut, Maine, Maryland, Utah, and Wisconsin) the differences are in interpretation of the restrictiveness of laws. Researchers using a quasi-experimental approach that relies on the coding of laws in these states should explore the sensitivity of their results to each set of coding. For several other states (Arizona, Arkansas, Delaware, Missouri, Montana, Nebraska, Ohio, and Tennessee), the coding differs by more than a year and appears to be explained by error. For instance, Levine (2003) codes Ohio as enforcing a parental involvement law from 1985 to present when in fact a court-issued injunction barred enforcement of the law until it was held to be constitutional by the U.S. Supreme Court in Ohio v. Akron Center for Reproductive Health (1990).

Sabia and Rees (2013) and Sabia and Mark Anderson (2016) together cover policy changes over 1987 to 2011. Comparing my policy coding for this period, we are generally in agreement save for one substantial difference: I code Florida as enforcing a parental involvement law from 2005 to present, while these sets of authors do not.

This substantial difference in policy coding in a large state bears some examination. In 2005, the Florida legislature enacted the Parental Notification of Abortion Act, which required parents to be notified of a minor child’s abortion but allowed this average notification to be provided via telephone. Subsequently Florida legislators became concerned that the law was easy to circumvent by having someone pose as a parent on the telephone. In 2011, a new bill was introduced and passed that added requirements that notification provided via telephone be confirmed by mail and that written waivers of notice from parents be notarized. Sabia and Rees (2013) and Sabia and Mark Anderson (2016) code parental notification in Florida beginning in 2011 with the amended version of the law, whereas I code it beginning in 2005 with the first notification requirement. Any choice is clearly subjective and empirical research may wish to explore the robustness of any results based on policy variation in Florida to these alternative choices.

4 Conclusion

State policies governing young women’s legal and confidential access to abortion and prescription contraception have evolved for six decades, determined by a complex and varying interplay of U.S. Supreme Court rulings and state regulations. The resulting spatial and temporal variation young women’s legal access to reproductive technologies affords an opportunity to implement difference-in-differences research designs estimating the causal effects of reproductive control on people’s lives. It is not surprising that large and prominent literature in economics has taken advantage of these natural experiments, but the results are only as good as the policy coding used to generate them. To the extent that the coding is inaccurate due to random mistakes, estimates using incorrect policy coding are likely to suffer attenuation bias due to measurement error. Perhaps more concerning, to the extent that the policy coding does not fully reflect the legal availability of both contraception and abortion and distinguish between adult and minors’ rights to access them, then correlations between these policy changes may cause researchers to conflate the effects of various policies.

This paper services a guide for researchers seeking to exploit variation in policies governing young women’s access to reproductive technologies as a natural experiment, or simply control for these policies in pursuit of other empirical questions. I provide a broad overview of the historical context and sources of the policy variation, along with recommended policy coding and an associated supplemental dataset (Myers 2022b). I also describe and reconcile the differences between this suggested policy coding and that of earlier authors, largely by correcting prior mistakes. Where subjective choices must be made, I provide interested readers with the tools and information to draw their own conclusions and implement robustness checks in their own analyses. The accompanying online state-by-state policy appendix (Myers 2022a) provides state-by-state documentation.

My own work (Myers 2017; Myers and Ladd 2020) has demonstrated the distinctions between policies governing abortion and contraception access made in this document are relevant. Myers (2017) demonstrates that when the entire range of governing policies—those determining the legality of contraception and abortion as well as those determining whether young unmarried women could even consent to them when they were legal—are considered, there is no evidence that contraceptive policies had a substantial impact on young women’s family formation, whereas the legalization of abortion had substantial impacts that were amplified when minors could consent to abortion. This paper further demonstrates that results in earlier papers suggesting the contrary—that young unmarried women’s confidential access to the pill drove demographic and economic changes—could not be replicated after correcting errors in their legal coding and/or adding correct and complete controls for abortion access. Myers and Ladd (2020) further demonstrate that policies governing minors’ confidential access to abortion have continued to impact teen fertility through the present.

Going forward, researchers seeking to identify the effects of policies governing the legality of and rights of young women to consent to contraception and abortion in the modern era are advised to distinguish between the pre-1976 and post-1980 periods, and to avoid the 1977–1979 period all together because the legal rights of minors to consent to contraception and abortion were murky. When studying the 1960–1976 era, researchers should distinguish between policy environments in which physicians could not legally provide access to contraception and/or abortion, and environments in which contraception or contraception and abortion were both legal but the rights of young unmarried women to consent to each varied with age, state, and year. In the 1980–2020 era, both the pill and abortion are legal in all states and women aged 18 and older can consent to them. Researchers studying minors’ access to reproductive technology in this contemporary era should consider whether a parental involvement law limiting confidential access to abortion is being enforced and may additionally wish to consider whether the law grants minors the explicit right to confidentially access contraception, though the on-the-ground salience of these latter policies remains unstudied.

The “credibility revolution” in economics—the marked shift toward a focus on experimental research designs that credibly isolate and measure causal effects—has afforded our field important new insights into the causal roles policy changes play in demographic and social outcomes. The responsible use of these powerful tools requires not only understanding econometric methodology, but also having a deep and nuanced understanding of the historical and policy context that generates identifying variation, and that the variation is accurately described. It is my hope that this paper provides that context and description for future inquiries into the role of reproductive policy in shaping people’s lives.