Autocorrelation,structural breaks and the predictive ability of dividend yield

Wine consumption is modelled by focusing on three decisions. The first is whether to consume any alcoholic beverages. Setting aside the teetotalers, the second decision is whether to consume wine and the third choice is how much to consume. We model these decisions using a double-hurdle approach. Our econometric results indicate that the probability a man drinks wine increases if he is under age 65, has a high income and is a high school graduate. Heavy wine consumers are over age 65, are well educated, physically active, and have high incomes.