Do skewed sex ratios reduce son preference?

In the early 1990s, Amartya Sen drew attention to the large-scale demographic distortion he referred to as the ‘missing women’ phenomenon, pointing to the millions of women who would have been alive if sex ratios had followed natural demographic patterns (Sen 1992). Over three decades later, the phenomenon is still highly prevalent. In 2020, there were 142.6 million missing women globally, more than double the number recorded fifty years earlier (UNFPA 2020).

Skewed sex ratios have long been studied and understood as the outcome of strong son preference. A large body of research documents how fertility behaviour, sex-selective abortion, and discrimination against women at all stages of life, including as children and the elderly, have produced persistent deficits of women (Borker et al. forthcoming, Genicot and Hernandez-de-Benito 2025). India accounts for around 46 million of those missing women (UNFPA 2020), making it a particularly important case for understanding this demographic imbalance.

Much less is known, however, about whether this relationship also holds in the opposite direction. Do highly imbalanced sex ratios, once they emerge, influence fertility preferences? In particular, can a local surplus of men weaken women's son preference? In a novel study combining various data sources from India, we provide evidence that this is indeed the case (Clément, Levasseur, and Seetahul 2025).

Sex ratios reshape fertility preferences

We study how district-level sex ratios affect women's stated preferences for sons using panel data on over 20,000 Indian women observed between 2005 and 2012 and sex ratios computed from census data. Exploiting variation across districts and over time, we find a clear pattern: women living in districts where men outnumber women are significantly less likely to express a preference for sons.

In districts where men outnumber women by 10%, women desire around 0.07 fewer sons on average, a small magnitude but a significant effect. When we focus on the sex ratio of individuals of marriageable age (15–49), and address concerns about reverse causality, this effect becomes substantially larger: a 10% surplus of men reduces the ideal number of sons by around 0.3 and increases the ideal number of daughters by a similar magnitude.

Our findings point to a ‘self-corrective’ dynamic. As women become scarcer, some of the social and economic conditions that sustain strong son preference begin to weaken. Despite India accounting for one of the largest absolute numbers of missing women globally, this effect has received little empirical attention. 

Why does a male surplus weaken son preference?

Two mechanisms appear particularly relevant in explaining this relationship in the case of India.

First, marriage markets adjust. In districts with a shortage of women, men face greater competition to get married. In our data, this is associated with women getting married at an older age and lower dowry payments by brides’ families. These reflect substantial changes in a society where early marriage and dowry payments have been documented to disadvantage daughters. These shifts reduce the economic disadvantage that households with daughters are likely to face. 

Second, women’s bargaining power improves. Where women are scarce, we observe greater decision-making power within households and lower community-level acceptance of domestic violence. Women in male-skewed districts score significantly higher on a household decision-making index and report lower tolerance of domestic violence in their communities. These patterns are consistent with broader evidence that demographic scarcity can strengthen women’s outside options, thereby increasing their bargaining power.

Addressing endogeneity concerns

An important hurdle in studying the relationship between sex ratios and son preference is endogeneity. Our analysis aims to identify whether, and to what extent, sex ratios shape son preference; however, son preference also influences sex ratios, making causal interpretation difficult.

To address this issue, we adopt an instrumental variable strategy that exploits plausibly exogenous variation in temperature at birth. Biological research shows that extreme temperatures affect the sex composition of births, as warmer temperatures damage X chromosomes in sperm (Lerchl 1998), producing more boys, while cold extremes increase male foetal mortality (Cagnacci et al. 2003). These temperature-driven shifts influence cohort-level sex ratios but are unrelated to women's later fertility preferences except through demographic composition.

This approach allows us to isolate the causal effect of sex ratios on women's stated preferences. To our knowledge, ours is among the first studies to apply such a strategy, especially in the Indian context.

Why declining son preference is not enough

Our findings should not be interpreted as evidence that gender inequality will resolve itself as women’s son preference decreases or sex ratios become more balanced.

First, the decline we observe reflects a reduction in women’s son preference. Further research is needed to assess the extent to which fertility preferences of women translate into realised outcomes in contexts where women’s bargaining power remains limited.

Second, evidence from countries such as South Korea (Yoo et al. 2017) and China (Tang and Hou 2024) shows that son preference can persist even after sex ratios stabilise, reflecting deeply entrenched gender norms that are likely to impact women in different ways across time. 

Policy implications: Beyond demographic forces

The prevalence of skewed sex ratios in various countries and concerns about their long-term consequences continue to shape policy debates. By showing that sex ratios can influence preferences, our results contribute to a more dynamic understanding of the ‘missing women’ phenomenon.

At the same time, the limited demographic normalisation process that we document highlights the importance of complementary policy interventions. In addition to direct interventions, such as bans on sex-selective abortion or cash transfers to offset the perceived ‘cost of daughters’ that have shown mixed results (Kumar and Sinha 2020), our results call for policies that target the underlying economic and marriage-market mechanisms sustaining son preference. Reforms to inheritance laws and their effective enforcement, as well as promoting women’s labour-market opportunities, can weaken the economic rationale for preferring sons by reducing parents’ dependence on them. At the same time, policies that address marriage-market distortions, such as conditional cash transfers aimed at addressing dowry practices, show promise.

References

Borker, G, J Eeckhout, N Luke, S Minz, K Munshi, and S Swaminathan (forthcoming), “Wealth, marriage, and sex selection,” American Economic Journal: Applied Economics.

Cagnacci, A, A Renzi, S Arangino, C Alessandrini, and A Volpe (2003), “The male disadvantage and the seasonal rhythm of sex ratio at the time of conception,” Human Reproduction, 18(4): 885–887.

Clément, M, P Levasseur, and S Seetahul (2025), “Sex ratio and fertility preferences in India: A longitudinal analysis,” World Development, 192: 107046.

Genicot, G, and M Hernandez-de-Benito (2025), “Firstborn daughters and family structure in sub-Saharan Africa,” Unpublished manuscript.

Kumar, S, and N Sinha (2020), “Preventing more ‘missing girls’: A review of policies to tackle son preference,” World Bank Research Observer, 35(1): 87–121.

Lerchl, A (1998), “Seasonality of sex ratio in Germany,” Human Reproduction, 13(5): 1401–1402.

Sen, A (1992), “Missing women,” British Medical Journal, 304(6827): 587.

Tang, M, and J Hou (2024), “Changes of sex ratio at birth and son preferences in China: A mixed-method study,” China Population and Development Studies, 8(1): 1–27.

UNFPA (2020), "State of world population 2020."

Yoo, S H, S R Hayford, and V Agadjanian (2017), “Old habits die hard? Lingering son preference in an era of normalizing sex ratios at birth in South Korea,” Population Research and Policy Review, 36(1): 25–54.