Forming a family is not like playing blackjack
Family planning groups and economists have long debated the causes and policy implications of changes in global fertility. In 1994, Harvard economist Lant Pritchett published an influential paper making the case that fertility preferences are a much stronger driver of actual fertility levels than family planning programs. He was dismissive of the concept of “unmet need” for family planning, which has been a crucial driver of funds and attention to family planning programs.
In a paper published in Demography earlier this year, Isabel Günther and Kenneth Harttgen applied Pritchett’s methodology to a larger sample size, drawing on data collected since 1994. For the most part, their findings validated his conclusions, which is important because, as Günther and Harttgen point out:
“Some have even claimed that Pritchett’s study was one of the reasons…behind a drop in international financing for family planning efforts after the United Nations International Conference on Population and Development in 1994: between 1995 and 2008, international aid for family planning efforts dropped by 30%.”
Given the recent resurgence of attention to family planning in international policy, perhaps a refresher course is in order.
But Günther and Harttgen also found that “women in sub-Saharan Africa are less capable of translating child preferences into birth outcomes than women in other developing countries.”
Sub-Saharan Africa has long been exceptional both in terms of its desired and actual fertility. But the notion that women there are “less capable” of achieving their goals requires closer scrutiny.
First, we have to consider what “desired fertility” means. This paper discusses different ways of measuring it, such as asking the woman what her “ideal family size” is and compares her actual family size to that, or asking the woman prospectively if she wants to have another child or would prefer not to, and then scoring subsequent births as “wanted” or “unwanted” based on that answer.
Underlying all of these survey methods is an assumption that fertility preferences are a) highly specific, b) rigidly fixed, and c) a matter of strong importance to the survey respondent. Unfortunately for the demographers, we have good reason to think those assumptions are false—and worse, we’ve lost some of our ability to determine exactly how off base they are.
The analysis in the Günther and Harttgen paper, as with Pritchett’s, uses data from the USAID-funded DHS (Demographic and Health Surveys), which are conducted periodically in a selection of developing countries. Prior to 2003, the DHS included questions inquiring about how strong women’s fertility desires were: how big of a problem would it be, or how happy would the woman be, if she discovered she was pregnant in the near future?
Ilene Speizer from the University of North Carolina found that in three sub-Saharan African countries, motivation to avoid pregnancy, or the lack thereof, made a big difference:
“In Burkina Faso and Ghana, about a quarter of women who said they wanted to delay or limit childbearing also reported that it would be no problem or a small problem if they became pregnant soon. This response pattern was equally common among contraceptive users and nonusers. In Kenya, more than four in 10 women gave such ambivalent responses. Among women with an unmet need for means of delaying or limiting childbearing, 16–31% of those in Burkina Faso and Ghana, and 30–56% of those in Kenya, said that getting pregnant in the next few weeks would be no problem or a small problem.”
The questions about the strength of fertility preferences are no longer part of the DHS, although Speizer continues to advocate for more attention to this issue. Nevertheless, the limited data we have provide an important clue as to why sub-Saharan African women might not be quite as adept at “translating their child preferences into birth outcomes.” It may just be that their preferences are a bit more adaptable to a range of outcomes—and why should that be a bad thing?
Decades of fertility survey research have been based on the presumption that women approach family formation like a person playing blackjack at a casino: your goal is to hit the target (a.k.a. the “completed family”) without going over (a.k.a. “excess” or “unwanted” fertility).
It’s important to remember that these surveys are conducted by people who visit women at their homes and ask a long list of very personal questions about their sex lives, their children, and their use of family planning, among other things. If a woman does not have a preconceived notion of her “ideal family size,” the surveyor is instructed to try to elicit a numerical response, if possible.
Having extracted these numbers, demographers then turn around and grade the women on their “success” at meeting their “goals,” calling on policymakers and donors to offer assistance in this regard.
Like Pritchett before them, Günther and Harttgen found evidence suggesting that desired fertility is a good predictor of actual fertility—but without taking the intensity of that desire is also into consideration, the picture is incomplete.
Unfortunately, since family planning advocates benefit from the ability to paternalistically spin so-called “unwanted fertility” as demand for their services, these critical gaps in the data will likely remain unaddressed for the foreseeable future.