Abstract
This research investigates whether the positive association between college students' current and highβschool drinking is due to habit formation or whether it is solely due to the influence of unobserved individual characteristics that are correlated over time. The empirical investigation is based on individual level data from the 1997 and 1999 waves of the Harvard School of Public Health's College Alcohol Study (CAS). Amemyia's generalized least squares is used to address the heterogeneity issue in examining the relationship between past and current drinking. The results suggest that after controlling for correlation in unobserved characteristics that influence alcohol use, highβschool drinking has a significant and positive impact on college drinking, indicating the existence of habit formation. Both structural evidence of habit formation and reduced form results reveal that stricter drunk driving laws faced while in highβschool have a long lived effect, reducing drinking while in college. The results also highlight the importance of access to alcohol, both on campus and in the college neighborhood, as a determinant of college students' drinking behavior. Copyright Β© 2004 John Wiley & Sons, Ltd.