Is there a stepping stone effect in drug use? Separating state dependence from unobserved heterogeneity within and between illicit drugs

Empirically, teenagers who use soft drugs are more likely to use hard drugs in the future. This pattern can be explained by a causal effect (i.e., state dependence between drugs or stepping-stone effects) or by unobserved characteristics that make people more likely to use both soft and hard drugs (i.e., correlated unobserved heterogeneity). I estimate a dynamic discrete choice model of alcohol, marijuana and hard drug use over multiple years, and separately identify the contributions of state dependence (within and between drugs) and unobserved heterogeneity. I find statistically significant “stepping-stone” effects from softer to harder drugs, and conclude that alcohol, marijuana and hard drugs are complements in utility.