Official Spanish birth registry data report sex ratios well above expected levels between 1975 and 2000, peaking at 109 boys per 100 girls in the early 1980s, the highest in the world at that time. Prior research has attributed these elevated ratios to factors such as maternal age, birth order, and differential prenatal care. We show that they instead reflect systematic coding errors by the Spanish Statistical Office. Census data reveal normal sex ratios for the same cohorts. The birth registry also exhibits implausible monthly volatility and asymmetrically distributed outliers, consistent with one-directional miscoding of females as males. Additional corroborating evidence comes from provisional birth statistics, which show significantly lower sex ratios than the finalised records, and from anomalous patterns in adjacent fields on the birth registration form. Our findings underscore the responsibility of statistical agencies to validate administrative records and cross-check them against alternative sources.