- Authors: Prankur Gupta, Dean Spears, Sangita Vyas
- R&R at Journal of Population Economics
- Download paper
Survey questions on ideal family size are a widely-used tool in population science, yet their validity remains contentious. We pool data on over five million women from hundreds of Demographic and Health Surveys, to shed new light on the interpretation of these reported ideals. Our research strategy uses the fact that the sex of a woman’s first birth is random, and predicts variation in her realized fertility. We ask whether this determinant of fertility outcomes influences later reported fertility preferences—and we find that it does. Specifically, women whose first-born was female later report wanting a greater ideal number of children, on average, compared to women whose first-born was male. Moreover, the preferred sex composition adjusts to match the sex mix that mothers actually have. The preference effect is immediate and grows over time after the first-born’s birth. These results hold using alternative sources of plausibly informative variation. Our findings demonstrate that survey-reported ideal fertility is not altogether causally prior to realized fertility outcomes. Methodologically, these findings challenge the interpretation of survey-reported fertility preferences.
Photo by Camille.