Does College Education Promote Healthier Behaviors? Evidence From Demographic Distortion Caused by Japanese Superstition

We investigated the causal effects of a college education on smoking, drinking alcohol, sleeping, and site-specific cancer-screening behaviors in Japan. To adjust for endogeneity, we leveraged a unique instrument-demographic distortion caused by a behavior based on a superstition. A superstition held by some Japanese people caused a demographic shift in 1967 that led to an increase in college educational attainment among a certain cohort. Using a recursive bivariate probit approach, we found that having a college education reduced smoking and increased some types of cancer-screening behaviors. Additionally, we found that the causal relationship may be mediated by an improvement in economic and social resources.