Why Do Harmful Norms Persist? Female Genital Cutting in Burkina Faso: Guest post by Lindsey Novak


This page in:

This is the fifteenth in our series of job market posts this year. 
For better or for worse, social norms have profound influence on many of the decisions we make—from political to personal. These norms can be particularly influential when it comes to making decisions surrounding child rearing, including the decision parents make to participate in the practice of female genital cutting (FGC). Parents living in communities that practice FGC—located primarily in parts of Africa, the Middle East, and Asia—decide whether or not their daughter will undergo FGC based on social pressure and the perceived costs and benefits of adhering to or deviating from the social norm.
The practice has no known medical benefits, and it is associated with a wide range of health complications, both physical and psychological. Women who undergo FGC are more than twice as likely to experience birthing complications (Jones et al., 1999), and are 25 percent more likely to contract sexually transmitted diseases (Wagner, 2014). In addition, women who have undergone FGC are more likely to experience depression, anxiety, and post-traumatic stress disorder (Dorkenoo, 1999;  Behrendt & Moritz, 2005). These health complications make working in and outside of the household more difficult.

Is Female Genital Cutting a Social Coordination Norm?
Despite these harmful effects, approximately 3 million girls undergo FGC each year (WHO, 2012). The prevailing theory, put forth by Mackie (1996), is that FGC is a social coordination norm—that is, no household will deviate from the norm unless a substantial proportion of the community also agrees to abandon the practice. Under this theory, if a sufficient proportion (what is “sufficient” may vary by community) of households agree to abandon FGC, then a tipping point is reached and the rate of FGC will fall to zero.
Yet recent evidence shows that the implications of this theory do not hold up to empirical scrutiny. If FGC is a social coordination norm, the persistence of the practice should be largely attributable to the community; in West Africa, however, 87% of the variation in the persistence of FGC can be attributed to individual- and household-level factors (Bellemare, Novak, and Steinmetz, 2015). Furthermore, the social coordination norm theory predicts that rates of FGC at the community level should be either very close to zero or very close to one. But a study in Sudan found rates of cutting at the community level that were almost never close to zero or one (Efferson et al., 2015).
A Theory of Heterogeneous Thresholds
In my job market paper I propose a new theory, and test that theory with observational data from Burkina Faso. I show that the rate of FGC within the community certainly affects parents’ decisions; however, each household has a different threshold, where I define the household’s threshold to be the rate of cutting in the community at which the household is indifferent between practicing and not practicing FGC. If the rate higher than the household’s threshold the household will practice FGC, and the household will refrain from practicing FGC if the rate is lower. While this may seem like a simple relaxation of an implicit assumption from the prevailing theory, I go on to show that this heterogeneity has important implications for how and why FGC persists. Drawing on Thomas Schelling’s 1978 model of critical mass, I show that if households have heterogeneous preferences, a tipping point in the rate of FGC is not guaranteed, and there may be stable interior equilibria in the rate of FGC at the community level.
Using data from the Demographic and Health Surveys that includes women born between 1949 and 1995, I show that households in Burkina Faso do in fact have heterogeneous thresholds. I further show that in some communities, a tipping point in the rate of FGC exists, and that point has likely been reached in many of those communities.
Observing a household’s threshold is not possible because the threshold is a function of household member preferences for FGC and bargaining power. Additionally, household members may not be able to articulate their individual threshold. I am, however, able to create bounds on a household’s threshold. I observe the rate of FGC in the daughter’s community and cohort (hence forth, “community-cohort”) as well as the daughter’s FGC status. I show that for a household that chooses to practice FGC, the rate of FGC in the daughter’s community-cohort is an upper bound for the household’s threshold. Similarly, if the household abstains from practicing FGC, the rate of FGC serves as a lower bound for the household’s threshold.
Figure 1 shows the cumulative distribution function (CDF) of household thresholds within Burkina Faso. The CDF for households that refrain from cutting their daughter is a lower bound for the true distribution of household thresholds, while the CDF for those who do cut their daughter is an upper bound. Thus, the true CDF of household thresholds will lie somewhere in between the two CDFs in Figure 1. The CDF approaches the 45-degree line from below, and I show that this suggests that there is a tipping point in the rate of FGC in Burkina Faso, and that tipping point is between 0.7 and 0.95. This is encouraging because it means we should see the rate of FGC in Burkina Faso continue to decline.
Figure 1. Cumulative distribution function of household thresholds

What Does This Mean for Policy?
These findings have important implications for the types of policies that should be introduced. If FGC were a social coordination norm, NGOs and activists should concentrate on interventions that change community-level preferences for FGC. Much programming surrounding FGC is currently conducted in this way. But if FGC is not a social coordination norm, but rather (as my results suggest) households have heterogeneous thresholds, these village-level interventions may have little to no effect on the rate of FGC in communities. Instead, since thresholds are heterogeneous, the set of potential policy interventions is dramatically broadened. Because households are able to deviate from the norm without waiting for a critical mass of community members to join them, interventions that target individual- or household-level preferences could be more effective.
With this in mind, I investigate how various groups within the community interact with the norm differently. Specifically, I look at whether educate parents are more likely to deviate from the norm when faced with a high rate of FGC than are uneducated parents. For a given rate of FGC within the girl’s community-cohort, a girl who has educated parents is 20 percent less likely to undergo FGC. Figure 2 shows a CDF of the proportion of girls with educated and uneducated mothers who do not undergo FGC. At every rate of FGC, there are more girls with educated mothers who do not undergo FGC than those with uneducated mothers.
Figure 2. Cumulative distribution function of household thresholds by maternal education

Going forward, focusing on interventions that aim to change a household’s or individual’s threshold as well and carefully considering the targeting of those interventions is key. These interventions should aim to incentivize households to deviate from the social norm when a larger proportion of their community is still practicing. These interventions could empower people to stand up to social pressure or alter the perceived costs and benefits associated with FGC. These are potential pathways through which educated parents are less likely to practice FGC. Additionally, interventions that discuss the potential harmful physical and psychological effects of FGC could increase its perceived costs, and interventions that debunk ideas about religious edicts that mandate FGC could reduce the perceived benefits of FGC.
Lindsey Novak is a PhD candidate at the Department of Applied Economics at the University of Minnesota. She is on the job market this year.

December 21, 2016

" Furthermore, the social coordination norm theory predicts that rates of FGC at the community level should be either very close to zero or very close to one."
I wish the author would include the source(s) for such a claim.
The early post WWII period included strong social norm pressure to avoid homosexuality. Yet Kinsey found "37% of males and 13% of females had at least some overt homosexual experience to orgasm..." (http://www.indiana.edu/~kinsey/resources/bib-homoprev.html). When we consider the criminal, economic, and social repercussions in that era when homosexual behavior was exposed (arrest, loss of employment, general social ostracizing) we have a good example that contradicts the argument about behavior being very close to either zero or one. Since the author's entire justification for a modified approach to discouraging genital mutilation is built on this unsupported statement about "social coordination norm" theory, I would contend a critique of the prevailing theory deserves a more nuanced analysis than that provided in the essay.

December 21, 2016

Mike, thank you for your very thoughtful comment. I certainly appreciate the nuance required for this topic, and it is difficult to achieve that in a short blog post. Your comment has made me consider how to express these details more effectively.
The hypothesis that rates of FGC should be very close to zero or one come from Mackie’s 1996 article (link in blog post) in which he proposes the social coordination norm theory. The two articles that I discuss in the post (Bellemare, Novak, and Steinmetz, 2015 and Efferson et al., 2015) delve much deeper into the tests of the hypothesis than I do in my job market paper.
I think that your overarching point is that people will deviate from the social norm even if there are strong social sanctions that accompany that deviation. While certainly practicing FGC and the lack of open homosexual behavior were enforced by social sanctions, I don’t think that homosexuality in the early post WWII period is a social coordination norm in the sense that Gerry Mackie proposes. It is not clear to me what homosexuals would be coordinating over. Acceptance of homosexuality is less about coordinating sexual preferences than FGC is about coordinating expectations on the marriage market. The key feature of a social coordination norm is that community members decide whether to participate in the norm based on how many other community members are participating in the practice. In the model there is somewhat of an implicit assumption that any family participating in the practice (in this case, FGC) will impose those social sanctions and families not practicing will not impose those sanctions. And I don’t think that holds for homosexuality.
I’m sure that there are other theoretical frameworks for analyzing the shift in opinions about homosexuality, but I don’t think that the social coordination norm would be a good fit for that analysis.

January 13, 2017

This makes sense. I could imagine a similar pattern going on for family planning decisions. Also because the individual-household centered interventions seem to work (see Bangladesh).
What papers are there that rigorously test what works in terms of interventions against FGM?