Resolving the milk addiction paradox

Abstract The milk addiction paradox refers to an empirical finding in which consumption of non-addictive commodities such as appears be consistent with the theory rational addiction. This paradoxical result seems more likely when is persistent and aggregate data. Using both simulated real data, we show that disappears estimating data using AR(1) linear specification describes saddle-path solution model, instead canonical AR(2) model. able correctly discriminate between simple persistence test for main features addiction, produce unbiased estimates short long-run elasticity demand. These results hold individual aggregated they imply model a better alternative testing than