Sex Preference Versus Number Preference: The Case of Korea

Abstract

This paper examines whether son preference has a strong positive effect on fertility rates via number preference. To test this hypothesis, Korea was used as a case because Korea still shows strong son preference but its fertility rates have reached below the replacement level. The hazard models with unobserved heterogeneity were used to analyze the relationships between sex preference and number preference. To reject the strong positive effect of sex preference on total fertility rates, number preference must be distinguished from sex preference. Statistical analysis shows that the existence of a son has a stronger effect on fertility behavior in Korea than the number of children. Nevertheless, the effects of son preference may not have a significant effect on the fertility rates since the probability of having only daughters becomes lower as the parities increase. The total fertility rate is defined as the sum of the hypothetical age-specific fertility rates of women, rather than the sum of the real age-specific fertility rates. If there are some important factors occuring the differences between the hypothetical age-specific fertility rates and the real age-specific fertility rates, the total fertility rates may be wrong. Because of increased birth intervals and late ages at marriage, it is difficult to say that the fertility rates below the replacement level since the late 1980 are a permanent phenomenon. In spite of the existence of the strong son preference, the ideal (or desired) number of children have decreased in Korea. Moreover, the means to control the number of children, such as abortions, have been supported by the Korean government policies. Through sex-selective technologies, abnormally higher sex ratios at birth have occurred since 1985.