Solving the Mystery of Spain's "Missing Women"

For decades, Spain presented one of the most puzzling demographic anomalies in the developed world. Official birth statistics suggested that between the mid-1970s and 2000, the country recorded an unusually high number of baby boys relative to girls, reaching a peak of 109 boys for every 100 girls in the early 1980s—the highest ratio reported among major countries at the time.

A recent IZA Discussion Paper by Manuel Bagues and Carmen Villa shows that this apparent imbalance was not a demographic reality. Instead, the study demonstrates that Spain's "missing women" were the product of data processing errors introduced during the country's transition to computerized statistical systems.

A demographic mystery

In most populations, biological factors result in a remarkably stable sex ratio at birth of roughly 105 to 106 boys per 100 girls. Spain's reported figures stood far above this range for nearly a quarter century, prompting numerous attempts to explain the anomaly.

Researchers attributed the elevated ratios to factors including maternal age, birth order, and environmental conditions such as solar radiation. More recently, some studies suggested that the spread of ultrasound technology may have enabled expectant parents to learn fetal sex, leading mothers carrying boys to receive better prenatal care and increasing male survival rates. Such interpretations implied widespread gender-related differences in prenatal investment and potentially tens of thousands of excess female fetal deaths.

Evidence of a statistical error

Bagues and Villa show that the anomaly is the result of faulty data processing rather than demographic behavior.

Comparing birth registry records with census data, they find that census cohorts display entirely normal sex ratios, with the discrepancy confined to the 1975–2000 birth registry data. The authors also identify implausible province-month fluctuations in sex ratios, far exceeding what could be expected by chance and consistent with systematic miscoding of female births as male.

Additional evidence comes from unusually high rates of missing birth-weight information in areas and periods with the most extreme sex ratios, suggesting broader data-entry problems affecting multiple variables. Finally, the anomalous patterns are largely absent from provisional statistics for the period between 1975 and 1982 and appear only in the finalized microdata, indicating that the errors were introduced during the final processing and verification stage.

Lessons for researchers and statistical agencies

The anomaly coincides with Spain's transition from manual to computerized data processing in 1975 and appears to reflect systematic coding errors—rather than simple data-entry mistakes—that were introduced during the finalization of records and persisted for roughly twenty-five years.

The findings suggest that previous explanations of Spain's unusual sex ratios may need to be reassessed. They also raise concerns for cross-country research, since Spanish birth statistics are incorporated into international databases maintained by organizations such as the UN, WHO, and Eurostat.

More broadly, the paper highlights the importance of validating administrative records. Simple consistency checks and comparisons with independent data sources could have identified the problem much earlier. The authors argue that when errors cannot be corrected, statistical agencies should clearly flag affected periods to prevent future misinterpretation.

Download the full paper here.