13 Apr 2021

In sub-Saharan Africa, a large number of parents desire more sons than daughters: such ‘son bias’ increases risky sexual behaviour and leads to higher HIV infections. That is the main conclusion of research by Arcady Mongoue and Roland Pongou, to be presented at the Royal Economic Society's annual conference in April 2021.

Analysing data from 153 Demographic and Health Surveys (DHS) conducted in 37 sub-Saharan African countries between 1986 and 2017, their study finds that mothers who have a greater natural propensity to produce girls have a higher probability of contracting HIV. No such effect is found among fathers, although they display greater son bias than mothers.

The analysis finds that about 6.3% of women versus 5.7% of men are infected with HIV in sub-Saharan Africa. Among gender-biased parents, 60% of mothers versus 80% of fathers desire more sons than daughters. But there is significant heterogeneity in son bias, which could be explained by the diversity of family systems in African societies.

The authors find that a woman who has produced only daughters is more likely to be infected with HIV by roughly 1 percentage point compared to a mother who has produced at least one son. This effect is larger among mothers who had their first child between the ages 25 and 39. Among these mothers, bearing only daughters increases the likelihood of getting infected with HIV by 2.5 percentage points.


The consequences of son bias in Africa have not been sufficiently documented. This lack of attention could be explained by the fact that son bias does not have similar manifestations as in South and East Asia where it is more prevalent. In these latter regions, son bias has been found to lead to the abortion of female foetuses and the neglect of girls.

These consequences of son bias have not been observed in sub-Saharan Africa. But in many African societies, sons are desired as heirs and preferred for the perpetuation of the family name, as girls surrender their identity on marriage. Wives are therefore more valued when they have at least one son. Indeed, the authors find that about 90% of widows who had at least one son inherited from their late husband whereas less than 10% of widows who had only daughters had this privilege.

The authors uncover several interesting explanations for the effect of son bias on HIV. One explanation is that parents who have only daughters keep trying for a son, thereby increasing the likelihood of unprotected sex, which in turn leads to increased HIV exposure, especially when a member of a couple has other sexual partners.

In addition, among mothers who already have their ideal number of children, those who have only daughters are more than 7 percentage points more likely to desire an additional child than their counterparts who have at least one son.

A second explanation is that wives who have a greater natural propensity to bear daughters are more likely to engage in extramarital relationships. Indeed, married women are more than 5 percentage points more likely to be unfaithful when they have fewer sons than their ideal number.

A third explanation is that husbands are more likely to take an additional wife when the first wife has is more likely to bear daughters. Relatedly, women who have only daughters have a 2 percentage points greater probability of having a polygamous husband. This differential effect increases to 8 percentage points for women who are at least 45 years old.

Fourth, women who bore a daughter when unmarried are less likely to find a husband, leading to a larger number of sexual partners throughout their lifetime. Among such women, those over 45 years old who have only daughters are 18 percentage points more likely to have never been married. These findings are consistent with analysis of data from the United States.

In contrast, son preferences have no effect on a male's HIV status or sexual infidelity, except when he has low education. The findings of this study support the classical theory of sex determination according to which child sex is determined by the male sexual gametes, as well as modern generalisations of this theory that also assign a role to mothers.

This study is the first to investigate the causal effect of son bias on sexual behaviour and HIV. Its findings imply that policies aimed at curbing HIV spread in sub-Saharan Africa should also target cultural practices that value sons over daughters.

Arcady Mongoue

University of Ottawa |

Roland Pongou

University of Ottawa |