An Econometric Analysis of the Impacts of China's One-child Policy on Population Trends from 1960-2010

Abstract

This paper uses OLS regression to investigate the “successfulness” of China’s One-Child Policy, focusing on fertility rate trends, using population data from 1960 through 2010. We find that, over time, fertility rate and birth rate trended together and were cointegrated while age-dependency ratio (% old), ratio of urban-to-rural population, and ratio of male-to-female population all increased over time. Furthermore, except for the ratio of male-to-female population, all other variables were non-stationary (i.e., the mean and variance of the ratio of male-to-female population was constant over time unlike the rest of the variables). Although this statistic may suggest the opposition of the concept of socio-cultural values leading to sex selection of children with a preference of males over females, this theory does not exist in practice; we should not forget that the ratio between total male and female populations comprises of two variables whose values range in the hundreds-of-millions. The nature of the ratio between total Chinese male and female populations being stationary does not shrink the concept of sex selection with the preference for male children. We find that fertility rate and birth rate trended together and were cointegrated over time while age-dependency ratio, the ratio of urban-to-rural population, and the ratio of male-to-female population all increased over time. Interestingly, except for the ratio of male-to-female population, all other variables were non-stationary, indicating opposition towards expected socio-cultural values supporting sex selection of children with a preference for boys over girls. This phenomenon likely occurs due to China’s enormous population size, as the variables that contribute to the ratio of males-over-females variable have values in the hundreds-of-millions. The paper concludes that potential relationships exist between China’s declining fertility rate and other population-based factors, although further study with more individual data would need to be conducted to try to infer causal relationships.