Bidirectional relationship between time preference and adolescent smoking and alcohol use: Evidence from longitudinal...

Scholarly interest in time preference as a potential predictor of risky health behaviors in adolescents has increased in recent years. However, most of the existing literature is limited due to the exclusive reliance on cross-sectional data, precluding the possibility of establishing the direction of causality. Using longitudinal data from the Korea Youth Panel Survey (2003-7), which followed up a nationally representative sample of 3449 adolescents aged 14years for five years, this study examines a bidirectional relationship between time preference and smoking and drinking behaviors among adolescents.

We used discrete time hazard models of smoking and drinking initiation as a function of time preference measured at the baseline and fixed-effects ordered logit model of time preference, respectively. Our measure of time preference was derived from the survey question on a hypothetical choice between immediate enjoyment today and likely higher scores on an exam tomorrow.

The overall results provide evidence on the bidirectional relationship; that is, higher time discounting (i.e., greater relative preference for present utility over future utility) results in an increased risk of engaging in smoking and drinking, and conversely, adopting such behaviors leads to a higher discount rate.

The bidirectional relationship may function as a mechanism for adolescents to engage in increased smoking and drinking or additional negative health behaviors via gateway effects, strengthening the case for preventing the initiation of risky health behaviors among adolescents.