POLICY BRIEF | SEPTEMBER, 2019 How US government restrictions on foreign aid for abortion services backfired By Grant Miller, Eran Bendavid and Nina Brooks KEY TAKEAWAYS Abortion is an issue that stirs up deeply felt passions and n The Mexico City Policy has led seems to offer little basis for compromise. But there is one to increases in abortion in a thing that both sides of the debate agree on — fewer abortions range of sub-Saharan African are better. The pro-life side opposes abortion in principle, countries. while pro-choice advocates generally hold that preventing n Those findings are based on unwanted pregnancies is preferable to terminating them. an analysis of contraception, That shared outlook could provide common ground on one of the most pregnancy, and abortion important federal initiatives concerning abortion — the Mexico City Policy. trends across the Clinton, Bush This executive order, announced in 1984 by the Reagan administration at the and Obama administrations. United Nations International Conference on Population and Development, n In countries that depend requires all foreign nongovernmental organizations that get U.S. family heavily on U.S. support planning assistance to certify they will not perform abortions or provide for family planning and counseling about the procedure. reproductive health programs, The Trump administration has greatly expanded the policy to condition contraceptive use decreased 14 almost all U.S. global health aid on compliance with these restrictions, percent, pregnancies rose 12 including HIV, malaria, and maternal and child health programs. The U.S. percent, and abortions climbed 40 percent when the policy was spent more than $7 billion on international health assistance in 2017. in effect relative to countries We recently published research (Brooks, Bendavid, and Miller 2019) indicating less reliant on U.S. support. that, rather than reducing abortion, the Mexico City Policy has had the unintended effect of significantly raising abortion rates in a set of African n The evidence suggests that countries that rely heavily on U.S. family planning and reproductive health the policy leads to a reduction aid. Many of these abortions are likely to have been performed unsafely, in contraceptive use and endangering the health of women who had them, previous studies of abortion increased pregnancies and indicate (Grimes, et al 2006). abortions. Why would abortions go up under a policy that bans promoting the procedure? We found strong evidence that, when the policy was in force from 2001 to 2009, contraceptive use dropped in countries more reliant on U.S. aid and was accompanied by increases in pregnancies and abortions. John A. and Cynthia Fry Gunn Building siepr.stanford.edu 366 Galvez Street, Stanford, CA 94305-6015 @siepr facebook.com/SIEPR/ 1 POLICY BRIEF | SEPTEMBER, 2019 That is not surprising. International organizations that House, a gap other donors did not fill. Funding resumed offer abortion counseling, such as the International under President Obama. This stop-and-start pattern Planned Parenthood Federation, are also major suppliers serves as a natural experiment testing what happens to of contraceptives and family planning information. Their abortion when the policy is in place and when it is no decision to continue offering abortion information made longer in effect. them ineligible for U.S. funding, which forced them to Our main data source for abortions and pregnancies scale back a broad range of family planning services. Our was Demographic and Health Surveys (DHS), funded research suggests that cutbacks such as these are driving by the U.S. Agency for International Development. the higher pregnancy and abortion rates we find. Information on contraception came from the United Groups opposed to abortion and those that believe it is Nations Population Division’s World Contraceptive Use a woman’s right to choose do not agree on much, but dataset. 2 We obtained data on U.S. family planning they may share the view that the jump in abortion rates and reproductive health aid by country and year associated with the Mexico City Policy is a significant from the Organization for Economic Cooperation and problem. Development and also used the World Bank’s World Development Indicator for data on a range of economic and demographic variables. In total, we examined data Shifting policy offers natural test of Mexico on pregnancies and abortions for nearly 750,000 women. City Policy effects In our statistical analysis, we separated countries into Only a few studies have rigorously examined the Mexico those with relatively high or low exposure to the Mexico City Policy’s impact (Bendavid, Avila, and Miller 2011; City Policy, which was a measure of how much countries Jones 2015). Our research stands out for the long period relied on U.S. family planning assistance, when the of time and large number of countries we examined. In Mexico City Policy’s funding restrictions were not in place. addition, we took advantage of a peculiar feature of the We then compared rates of contraception use, pregnancy, Mexico City Policy — it has alternately been imposed and abortion in high- and low-exposure countries when and lifted by successive Republican and Democratic the policy was in effect and when it was not. administrations. We looked at trends in contraception The results were striking. In the highly exposed use, pregnancy, and abortion in 26 sub-Saharan African countries, contraceptive use was 14 percent lower, while countries1 from 1995 to 2014 during the presidencies of pregnancies and abortions were, respectively, 12 percent Bill Clinton, who rescinded the policy; George W. Bush, and 40 percent higher when the policy was in force who reinstated it in January 2001; and Barack Obama, during the George W. Bush administration compared with who overturned it again in January 2009. the rates during the Clinton and Obama presidencies. As a result, the organizations that were the most Simply put, when the Mexico City Policy was active, important providers of family planning and there was less contraception and more pregnancy reproductive health services in the countries we studied and abortion. When the policy was voided and family lost U.S. funding when George W. Bush was in the White planning aid restored, those trends reversed. 1 Benin, Burkina Faso, Burundi, Comoros, Eswatini, Ethiopia, The 2 We focused on modern methods of contraception, defined as female Gambia, Ghana, Guinea, Kenya, Liberia, Lesotho, Madagascar, and male sterilization, oral hormonal pills, the intra-uterine device, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, the male condom, injectables, implantables, vaginal barrier methods, Sierra Leone, Tanzania, Uganda, Zambia, Zimbabwe the female condom, and emergency contraception. John A. and Cynthia Fry Gunn Building siepr.stanford.edu 366 Galvez Street, Stanford, CA 94305-6015 @siepr facebook.com/SIEPR/ 2 POLICY BRIEF | SEPTEMBER, 2019 Could other factors account for more Second, the Mexico City Policy could have made abortions? women more reluctant than usual to report if it altered in some fashion the legal or cultural environment for An obvious question is whether something besides abortion. Because our abortion data is reported by the Mexico City Policy might explain the abortion ups individual women rather than governments or private and downs that we found. Our investigation was not organizations, this seems less likely. But importantly, if a randomized trial in which variables can be carefully there were selective incentives for abortion reporting controlled. In a statistical study like ours, there is always a under the policy, the most likely scenario would be chance that some unobserved factor may drive the results. relatively less reporting of abortions when the policy was Nevertheless, we are confident that our interpretation active and in countries benefiting more from U.S. aid. If of the Mexico City Policy’s effects — that abortion present, this would therefore lead us to underestimate and pregnancy rates varied depending on access to the unintended consequences of the Mexico City Policy contraception — is correct. Any confounding factor that we find — or the true extent of the rise in abortions would have to vary systematically both with the sharp when the policy was in effect would be greater than our off-again, on-again pattern of the policy and with the analysis suggests. degree to which family planning programs in each African Our research did not look at the effects on women’s country benefits from U.S. assistance — an unlikely health of the changes in contraceptive use, pregnancies, scenario. and abortions under the Mexico City Policy, but it is There could also be natural concerns about abortion reasonable to believe that maternal deaths and injuries reporting. Consider two related potential issues. First, have risen, perhaps significantly. Abortion incurs risk, abortions are notoriously difficult to measure, largely and more abortions are likely to drive up mortality because many are illicit and women are uncomfortable rates. Beyond that, to the extent that international reporting these. In particular, the DHS survey data is organizations are forced to cut back family planning widely considered to underreport abortions. In addition, services, more abortions are probably performed under it can be hard to distinguish induced abortions from unsafe conditions, putting women in greater danger. spontaneous abortions in the DHS data. This study, unprecedented in scope, strengthens the case However, as with any other potential confounder, that making family planning services and contraception underreporting would have to vary with the sharp off- more widely available is an effective way to reduce again, on-again pattern of the policy and also with the abortion. From that perspective, the Mexico City Policy degree to which family planning programs in each African does exactly the opposite of what is needed to prevent country benefit from U.S. assistance. We conducted the unwanted pregnancies that prompt many women to extensive simulations to determine the extent to which seek abortions. underreporting of abortion could have influenced In this regard, it seems possible that everyone could our findings and ultimately concluded that although agree that a policy designed to curb abortion but winds underreporting is present, it did not meaningfully bias up increasing it is a failure. our study’s results. John A. and Cynthia Fry Gunn Building siepr.stanford.edu 366 Galvez Street, Stanford, CA 94305-6015 @siepr facebook.com/SIEPR/ 3 POLICY BRIEF | SEPTEMBER, 2019 References Bendavid, E., Avila, P., Miller, G. United States aid policy and induced Grant Miller is a SIEPR senior abortion in sub-Saharan Africa. Bull World Health Organ 2011; 89: 873–80. fellow and former director of the Grimes, David A., et al. Unsafe abortion: the Preventable Pandemic. institute’s King Center on Global The Lancet 2006; 368.9550: 1908-1919.” Development. He is an associate Jones, Kelly M. Contraceptive Supply and Fertility Outcomes: Evidence professor of medicine, a core from Ghana. Economic Development and Cultural Change 2015; 64(1): 31–69. faculty member at Stanford Health Policy, and a research associate at the National Bureau of Economic Research (NBER). His research focuses on health economics, development economics, and economic demography. Eran Bendavid is an associate professor of medicine and a faculty affiliate of the King Center on Global Development. He is also a core faculty member at Stanford Health Policy and his research examines how economic, political and natural environments affect population health. Nina Brooks is a PhD student in environment and resources, obtaining a joint MA in economics. Her research explores the social and environmental determinants of population health. Sam Zuckerman contributed editorial assistance to this policy brief. The Stanford Institute for Economic Policy Research (SIEPR) catalyzes and promotes evidence- based knowledge about pressing economic issues, leading to better-informed policy solutions for generations to come. We are a nonpartisan research institute, and SIEPR Policy Briefs reflect the views and ideas of the author only. John A. and Cynthia Fry Gunn Building siepr.stanford.edu 366 Galvez Street, Stanford, CA 94305-6015 @siepr facebook.com/SIEPR/ 4